+ ve: Saels i at tl Se RNa Hep aaa aca EAS aes ie ‘i samen sat, it pat * i as si eth i ate ih ne ; Hh ’ ie hee insid " ihe ney) rad) 2 y Aye) Digitized by the Internet Archive in 2009 with funding from University of Toronto http://www. archive.org/details/journalofexperim02wist THE JOURNAL OF EXPERIMENTAL ZOOLOGY EDITED BY WILLIAM K. BROOKS FRANK R. LILLIE Johns Hopkins University : University of Chicago WILLIAM E. CASTLE JACQUES LOEB Harvard University University of California EDWIN G. CONKLIN THOMAS H. MORGAN University of Pennsylvania Columbia University CHARLES B. DAVENPORT GEORGE H. PARKER Carnegie Institution Harvard University HERBERT S. JENNINGS CHARLES O. WHITMAN University of Pennsylvania University of Chicago EDMUND B. WILSON, Columbia University AND ROSS G. HARRISON Johns Hopkins University Manacine Epitor VOLUME II oN THE JOURNAL OF EXPERIMENTAL ZOOLOGY h VA 3 BALTIMORE J aie ‘A 0 1905 ey 4 CONTENTS. No. 1.—April, 1905. CHARLES ZELENY, Compensatory Regulation. With 29 figures.................02.0c000% if Amos W. PETERS, ehosmhoarcscence In CLenOphoresds - <2 eee cee cece be ee eee eee ie 103 IsaBEL McCracken, A Study of the Inheritance of Dichromatism in Lina Lapponica. With Mptabe ands) tistinesim Ghe CEG. ah... tls eee sn ones oes teins wei oe eles ile l7¢ C. B. DAVENPORT, COMMON MWAGROUE MULATION. o 2 s.cn oo oc cie © 6 seis dies store ove sls clsele @ cleis 137 No. 2.—May, 1905. Epwin G. ConkLin, Mosaic Development in Ascidian Eggs. With 82 figures................ 145 C. W. Haun, Dimorphism and Regeneration in Metridium. With 2 figures............ 225 CHARLES G. RoGERs, The Effect of Various Salts upon the Survival of the Invertebrate Heart. CPHL IL [BO ta hos © Seat cn ee eee ne eee a 237 C. M. Cui, Studies on Regulation. WII. Further Experiments on Form Regulation sambeproplanan) Withia4 feures..<- . a: . 22. .222s sees c tt een cee eee BOO NaouipE YATSU, The Formation of Centrosomes in Enucleated Egg-Fragments. With 8 CEMES: os nccdd oot toed Pa AR BAe crc ee ee 287 No. 3.—August, 1905. N. M. STEvEns, A Study of the Germ Cells of Aphis Rose and Aphis (Enothere. With 4 PEER ioe hese are erica Pere eee 313 N. M. Srevens anp A. M. Borne, Regeneration in Polychcerus Caudatus. Part I. Observations on Living Material. By N. M. Stevens. With 21 figures...................-+- 335 Part II. Histology. By A. M. Boring. With 2 plates and 1 figure in the PeRG oo cc Se sa Sk ha awe By wok os os ee hee ree 340 CHARLES ZELENY, The Relation of the Degree of Injury to the Rate of Regeneration. With G figures: 2. ko. Saenger 347 Epmunp B. Witson, Studies on Chromosomes. I. The Behavior of the Idiochromosomes in Hemiptera. With 7 figures.......- rrr 371 G. H. Parker, The Movements of the Swimming Plates in Ctenophores, with Reference to the Theories of Ciliary Metachronism. With 2 figures..............-- 407 Henry EpwarbD CRAMPTON, On a General Theory of Adaptation and Selection................-+-+-- 425 WarrREN Harmon LEwIs, ' Experimental Studies on the Development of the Eye in Amphibia. II. On the Comes. With 2 plates... .. 02.2022 62052000» | er 431 No. 4.—November, 1905. H. S. JENNINGS, Modifiability in Behavior.’ I. Behavior of Sea Anemones..............- 447 H. S. JENNINGS, The Method of Regulation in Behavior and in Other Fields.............. 473 T. H. Morean, “Polarity ’’ Considered as a Phenomenon of Gradation of Materials....... 495 Epmunp B. Witson, Studies on Chromosomes. II. The Paired Microchromosomes, Idiochro- mosomes and Heterotropic Chromosomes in Hemiptera.............-- 507 Cuas. W. Hararirt, Variations Among Scyphomeduse. With 1 plate and 17 figures in the WERE her sicinee < la¥eel ainsi Saisie: aie tate o's ol: ackvene ese eine leis sum Siete oleae ete ean 547 LorANDE Loss WoopRrtrfr, An Experimental Study on the Life-History of Hypotrichous Infusoria. With 3 plates and 12 figures in the text........ Wee 585 COMPENSATORY REGULATION. BY CHARLES ZELENY: Pepeetere AMINE ICEL OM never iol faictefe) Porc) welteter atest eieYstceirieree de Ie & Glad cia sca ale ue cewbens 2 TYNE. 2s s9000 090000 G00 SS OQ CRUD RG ROUS SHES AEE TOn Acta cS Ce ne ce ae aac ae ae 5 feeNbereatlets onthe Compound Wealie.. 4.4. sas casa. ns < 1;',). line) and Series III and IV in another (the dotted line). As the individual disk diameters are not exactly equivalent in the different series, it was found convenient in taking the averages for the com- bination curves to use arbitrarily disk diameters equal to whole millimeters as the points for comparison. ‘The values for such disk diameters are obtained from the separate curves constructed from the individual measurements for each series according to the method shown in Fig. 2. The average curve for Series I and IV in Fig. 2 would be one equidistant between them. ‘The combina- tior: curves for Series I and II and Series III and IV, as given in Fig. 3, are constructed on this basis. The unbroken line gives Cc ares 14 Charles Zeleny. the average of the former group (one and two arms removed) and the broken line the average of the latter group (three and four arms removed). 3. . Data: The curves show very distinctly the correlation between the rate of regeneration on the one hand and the size of the animal and the number of removed arms on the other. 1. ‘Taking up first the size correlation and using Fig. 3 as our basis of comparison, since it contains all the cases except those of Series V and therefore gives a more uniform and complete curve, we find that starting with the smaller individuals as we advance toward the larger ones there is a general increase up to a maximum at a diameter of 12 to 15 mm. This is most striking in the two later measurements, taken 33 days and 46 days after the operation. Thus in the 33-day measurement for Series I and II (Fig. 3), the regenerated length increases from 1.07 mm. for a disk diameter of 7 mm. to a maximum of 2.37 mm. for a disk diameter of 14 mm., and then goes down to .21 mm. for a 19 mm. diameter. Also for the Series III and IV at the same time the length increases from 2.04 mm. at a diameter of 7 mm. to a maximum of 3.45 mm. at a 12 mm. diameter, and down again to 1.36 mm. at a diameter of 18 mm. The medium-sized individuals thus have the maximum rate of regeneration.t 3. More striking still is the very constant difference between the regenerated lengths for Series I and those for Series IV in Fig. 2, and BeEtnen the lengths for the combination of Series I na II and those for the cba comndinn of Series III and IV in Fig. 3. This shows a very decided advantage in favor of the animals with the greater number of removed arms. The difference is évident in the upper curves of Fig. 3 from measurements taken 22 days after the operation, but becomes more striking in the 33- day and 46-day curves. For example, in the 33-day curve for a 12 mm. diameter (the diameter at which we have the maximum rate of regeneration of Series III and IV) we get a regenérated length of 2.08 mm. for Series [ and II, and of 3.45 mm. for Series 1 Dr. Hans Przibram has called my attention to the fact that the specific rate of regeneration of the arms, 7. e., the amount of regeneration per unit of disk diameter as obtained from my data, does not show this increase from the smallest up to the medium-sized individuals, but gives a fairly constant figureupto120r14mm. ‘The higher diameters then decline rapidly toward a minimum. Compensatory Regulation. 15 III and IV, an advantage of 1.37 mm. or 66 per cent in favor of the latter. Likewise, at a diameter of 14 mm. (where the Series I and II has its maximum regeneration) we get 2.37 mm. for Series I and II and 2.77 mm. for Series III ont IV, an advantage of .4 mm. or 17 per cent in favor of Series [I] and IV. Ina similar manner in the curves obtained from the 46-day measurements we get at a 12 mm. disk diameter a regenerated length of 2.46 mm. for Series I and II and 5.42 mm. for Series III and IV, and at a I5 mm. diameter 3.14 mm. for Series I and II and 3.72 mm. for Series II] and IV, which represents an advantage for the group with the greater number of removed arms of respectively 2.96 mm. (=120 per cent) and .58 mm. (= 18 per cent) for the two points named. The difference between Series [ and Series IV as represented in Fig. 2 1s still greater. “The regenerated lengths are on the whole at least twice as great in Series IV where four arms were removed as in Series I where only one arm was removed. ‘Thus at a disk diameter of 8 mm. the regenerated length in Series I is 1.05 mm. and the average regenerated length in Series IV is 2.3 mm., an increase of 1.25 mm. or I1g per cent. For the same diameter at 33 days the respective values are 1.65 mm. and 4.4 mm., an increase of 2.75 mm. or 167 percent. At 46 days the correspond- ing values are 2.0 mm. and 5.85 mm., an increase of 3.85 mm. or 192 per cent. Likewise at a 12 mm. disk diameter for 22 days the values are -Q mm. and 2.1 mm., an advantage of 1.2 mm. or 133 per cent. At 33 days the values for a 12 mm. disk diameter are 2.3 mm. and 4.2 mm., an advantage of I.g mm. or 83 per cent, and at 46 days the Be ponding values are 3.05 mm. and 6.5 mm., an adv antage of 3.45 mm. or I13 per cent. We may sum up the results on the rate of regeneration of the arms of the brittle-star, Ophioglypha lacertosa, as follows: 1. There is a definite relation between the size (7. ¢., age [ ?]) of the animal and the rate of regeneration of its arms. “he max1- mum rate is exhibited by individuals of medium size (with a disk diameter of 12 to 15 mm.). Both the smaller and the larger ones give a diminishing rate as we go away from this point. gee Lhe greater the number of removed arms (excepting the case where all are removed) the greater is the rate of regeneration of each arm. » 16 Charles Zeleny. 4. Discussion. We must, therefore, conclude that when more than one arm 1s removed the regenerative energy as expressed in the replacement of the lost arms is greatly increased. Not only is the total regen- erative energy greater in this case, but the energy expressed in each arm is greater than the total energy when only one arm 1s removed. Expressing this in mathematical form, if E, represents the regenerative energy exhibited in the replacement of the lost arm when only one is removed, assuming that increase in length 1s a measure of such energy, and E, represents the energy exhibited in regeneration when more than one arm is removed, 7 being the number of absent arms, then not only is E, > E, but also E, Dot Dera) a Ope iil Be n Therefore, when we remove 7 arms we increase the total regenera- tive energy by more than 7 times the amount exhibited when onl one is removed. The force of this statement is made especially strong when we consider that throughout the experiments the animals received no food supply whatever. Expressing the relation in still another way, let us take a brittle- star with arms 4, B, C, D and £, in-which a,,5,, ¢39¢ eee represent the respective lengths these arms will attain after a definite period of regeneration, supposing that one alone is cut off in each case. Now let us suppose instead that the first four are cut off, then after this same period of time we get for the regener- ated lengths a,>4,, b,>5,, ¢,>¢,, d,>d,. Now in the first case we cannot assume that the stimulus of removal and the resultant reaction of regeneration are purely local and concern only the tissues in the immediate vicinity of the cut surface, for we then get into difficulty as soon as we try to explain the cases where four arms are simultaneously removed. Here we find we must add a considerable quantity (r,) to each of the original single regenera- tion lengths to get the new regeneration length, e. g., a, = a, +74. Then a,+6,+¢,+d,=4,+b,+c,+d,+R, where R, (=29 represents the total response of the organism as a whole which must be added to the local effects of the operation stimulus. If, on the other hand, we consider the influence of the organism as a Compensatory Regulation. 17 whole on the regeneration of its arms as one of retardation, we must take the values a,, b,, c, and d, as representing most nearly the original local stimulus effect. anes without changing the values of r, or R, we may rearrange the formulz, making a, = a,— Tf, etc., and a,+b,+c,+d,=a,+b,+c,+d,—R,. The regeneration of the arms of Ophioglypha thus offers us a very good example of the influence of conditions away from an injured surface upon the regeneration at that surface. The result may be stated in two ways and each mode of statement may be made to lead to a separate mode of interpretation. We may say that the rate of regeneration increases with an increase in the number of removed arms. With this statement as a starting point it is natural to assume that, in the cases where more than one arm is removed, the stimulus of the additional operations or of the additional regenerating organs exerts an accelerating influence upon the regenerating tissues at any one such surface. Another mode of statement is the following: The increase in the number of removed arms is necessarily accompanied by a decrease in the number of uninjured arms present, and the rate of regeneration of a removed arm therefore increases as the number of uninjured arms still remaining decreases. If the uninjured arms exert a retarding influence upon the regenerating tissue at an injured surface we can understand why a removal op additional arms may bring about an increase in rate of regenera- tion of each. The discussion of this interpretation inv lveer the whole problem of nutrition and perhaps the whole general problem of form regulation as well. It will be best to reserve further dis- cussion until we have examined the other experiments to be described in the following pages. But whether we consider the influence of the organism as a whole to be one of acceleration or one of retardation, we must recognize in either case that the regeneration rate 1s not a matter which involves only the local conditions at the wounded surface as determined by the direct action of the operation. It seems, on the other hand, to be bound up with intricate reactions affect- ing the whole character of the activities and organization of the animal. 18 Charles Zeleny. Ill. THE OPERCULA OF SERPULIDS.- We have now considered a case (section one) in which there is a readjustment in the uninjured portions of a system as a result of their mutual interaction. ‘This interaction is not complicated by the addition of a regenerating organ. ‘The result is a new system in equilibrium, based on the resultant of the interactions of the uninjured parts. In a second section (section two) a case was considered in which it was possible to study the effect of the presence or absence of uninjured portions of the animal upon the rate of regeneration of the removed ones. ‘The reaction is here more complicated than in the first case, because there may be here an action of other regenerating surfaces upon any particular wounded surface as well as the action of the uninjured organs themselves. In the present section (section three) a case will be considered in which there are two organs, dissimilar in size, situated on mor- phologically similar opposite sides of the median line. An extremely close interaction is found to exist between these two organs so that any disturbance in one is reflected in changes in the other. This close interrelation between the opercula of the Serpulids, the organs in question, gives a good basis for the study of such interactions as were outlined in the general introduction. The opercula of the Serpulids furthermore furnish exceptionally good material for this study of compensatory regulation because of the various degrees of asymmetry present in the different species. In the following account it will be necessary to go into paths not in the line of the main discussion, but such a course cannot be ~ avoided in a study of the factors controlling the regulation of the ~ opercula. In the adult Serpulids of the genus Hydroides we have an ~ asymmetrical stable system with Te functional operculum on the right side and the rudimentary operculum on the left or vice versa. The nature of this case will first be taken up. Then the opercula of other members of the family will be described. This will be followed in turn by a description of the ontogenetic development, the regeneratory development, some speculations as to the prob- able phylogenetic dev elopment, a discussion on the comparison of regeneratory, ontogenetic and probable phylogenetic develop- Compensatory Regulation. 0) ment and finally by a discussion of the facts of compensatory regulation as here exhibited. In a separate section (section four) a special series of experi- ments on the regulation of the rate of differentiation of the oper- cula in the Serpulid, Apomatus ampullifera, will be treated. 1. Comparative Anatomy. 1. The Genus Hydroides. The opercula and branchiz of the genus Hydroides will serve as the type in our description of the anatomy of these structures throughout the family. The branchiz are brought in only incidentally as our main purpose is to get the details of the struc- ture of the opercula to serve as a basis for the regeneration and regulation experiments to be described later. Unless otherwise stated the description applies to H. dianthus. H. uncinata and H. pectinata, living in the Mediterranean at Naples, were also used in the experiments and the differences will be pointed out at the close of the special description. H. dianthus is found on the Atlantic coast of North America living attached to stones, mollusk shells and other hard materials, from low water mark to a depth of several fathoms. The worm lives in an irregularly twisted calcareous tube which is attached by its side to the supporting surface. The tube increases in size from the posterior end anteriorly and is continually being built up at the anterior end by additions from the special calcareous glands. The body of the animal is very distinctly divided into the thorax and the abdomen. The first-named region is marked by the presence of the broad, flat fold of the thoracic membrane, which is continued at both the anterior-ventral and the posterior- ventral ends as a projecting membrane. At the anterior end this membrane forms a collar which, except for a slight break on the dorsal side, completely surrounds the head end. (See Fig. 5a.) Upon the anterior surface thus enclosed by the collar are located the two semicircular rows of branchiz, one on each side of the mouth, which apparently serve on the one hand as organs of respiration and on the other through their cilia as agents for the creation of a current of water carrying food particles to the mouth. The two rows of branchiz are placed on slight ridge-like eleva- tions, the branchial ridges. These are not strictly semicircular 20 Charles Zeleny. in shape but the ends of each are curved inward. ‘This shape evidently has some connection with the proper collection of the food particles carried downward by the cilia. (See Fig. 58, c.) The number of the branchiz increases with the age of the animal and in fully grown individuals there are about fourteen on each mies Ahi Kd sy Ty \ y; se A A ma ea D we © © hb oS Fic. 5. Hydroides dianthus. A—Dorsal view of right-handed specimen, showing relations of parts. Ends of branchie and functional operculum not given (6). B,C—Diagram of anterior surface of head of left-handed and right-handed specimens (6). D—Branchia viewed from inner surface (X25). E—Rudimentary operculum (30). /—Functional operculum (> X 30). Compensatory Re guiation. 21 Each branchia consists of a long axis bearing two rows of secondary processes, the pinnules. (See Fig. 5p.) ‘The axis is continued for a short distance beyond the region of the secondary processes as a slender tapering thread. ‘The pinnule rows slope inward so as to inclose a trough-like area, V-shaped in cross section and with the cavity of the trough pointing inward, 7. e., toward the mouth. ‘The surfaces bordering this area are ciliated and it is along them that the food-bearing currents are formed. Near the dorsal end of one of the branchial ridges, not in the line of the branchiz but dorsal to it, there is a stout, naked stalk of approximately the same length as a branchia but bearing at its distal end a funnel-shaped expansion. (See Fig. 5r.) The whole organ constitutes the functional operculum. ‘The edge of the expanded portion is marked by teeth-like serrations, the hollows between which are continued for some distance down the outside. From the center of the terminal circular area within this row of serrations there arises a group of secondary pro- cesses, arranged so as to form a cup-shaped figure. ‘The ends of the processes are usually hooked and considerable foreign material often clings to them. The whole organ serves as a very eficient plug for the open end of the tube when the animal has retired within for protection. An examination of the place of attachment of the opercular stalk shows that it is located dorsal to the first branchia or sometimes nearly opposite the interval between the first and second branchiz. Near the base of the stalk there is a transverse suture varying in distinctness in different cases and which, as we shall see later, is a “breaking joint,” an important structure in the experiments. On the opposite side of the mid-dorsal line, and in a position corresponding in all respects with that of the large operculum, is a small organ consisting of a slender stalk with a slight terminal enlargement. (Fig. 5x.) It also shows a distinct line of demarca- tion between a darker colored more basal region and the lighter remainder of its body. This small organ, the “pseudopercule”’ of de St. Joseph is most commonly called the rudimentary oper- culum. A study of the relative positions of the opercula 1s interesting. An examination of 244 adult individuals of H. dianthus gave 139 Or 57 per cent with the functional operculum on the right side and 105 or 43 per cent with it on the left. The distribution between a 22 Charles Zeleny. he and left is thus fairly equal though there is a considerable ie antage in favor of those with the functional operculum on the right side and the rudimentary on the left. Similarly in H. uncei- nata out of 16 specimens ten had the functional operculum on the right side and six on the left, and in H. pectinata out of 41 speci- mens 21 were right handed and 20 left handed. | An examination of the internal structure oF the branchiz and opercula brings out a close agreement between the two in ana- tomical details. Their morphological agreement has been espe- cially emphasized by Orley and Meyer. Orley (’84) compares the internal anatomy of the branchia and the functional operculum in Serpula. He makes no mention of the rudimentary operculum. According to him an operculum corresponds morphologically with a branchial stalk, all the pinnules of which have been col- lected at the end in one bundle. He describes the presence of an axial blood vessel in both branchial and opercular stalks. -In the branchial stalk, however, he saw only one nerve trunk (the axial one) while in the opercular stalk two lateral ones were shown. Meyer (88) showed the more complete similarity of the branchia and operculum in Eupomatus uncinata ( = Hydroides uncinata), while at the same time pointing out the incompleteness of Orley’s observations and the error in his mode of homology. He describes three nerve trunks in both branchiz and operculum, although the middle one is very small in the opercular stalk and does not reach much more than halfway to the distal end. In the branchia the two lateral ones likewise are very insignificant. Meyer points out that this difference is probably due to the fact that the pinnules and ciliated groove are innervated from the middle nerve, so that this has a greater development in the branchiz where pinnules and ciliated groove are present than in the opercular stalk where they are absent. ‘The stalk of the operculum is thus made directly homologous with a branchial axis lacking its pinnules. A study of the internal structure of the branchiz and opercula of Hydroides dianthus brings out points which are in entire agree- ment with the conclusions of Meyer and which emphasize the close similarity of the branchiz and opercula. An interesting characteristic is further made out in the func- tional operculum. ‘There is a difference between the cells near the basal region and those in the middle and terminal regions of the stalk. Near the base of the stalk in the region below the Compensatory Regulation. 23 “breaking joint”’ the cells of the connective tissue have more of an embryonic character than elsewhere, having fewer and shorter processes and less intercellular material. ‘This distinction has already been noted by Orley (’84) in his description of the con- nective tissue of the opercular stalk in Serpula vermicularis. He says, speaking of this tissue, “Die Modificationen dieses Bindegewebes sind nach den Ortsverhaltnissen sehr verschieden. Im imnersten Theile des Stieles wo dieser mit dem Kiemenlappen zusammen hangt, findet man kleine weniger verzweigte Zellen in der sehr sparlichen Intercellularsubstanz. Es ahnelt sehr der embryonaler Form. twas hoher trifft man bereits Zellen an, die sich durch Grosze und durch die Zahl ihrer Auslaufe auszeichen und eine gut entwickelte Intercellularsubstanz haben.”' ‘The significance of the differences of the regions will be brought out in connection with the experiments described later in the paper (P- 55)- he rudimentary operculum of H. dianthus has two well- defined regions. The cells distal to the “breaking joint” are distinctly embryonic in form and general character. Those proximal to the breaking joint have among them well-developed supporting cells of the type found in the branchiz and functional operculum, though these cells are not as highly differentiated as in the latter organs. A comparison of the two other members of the genus Hydroides with H. dianthus brings out only slight differences in the charac- ter of the opercula and branchiez. H. pectinata, however, has pectinate secondary processes as opposed to the unbranched ones of H. uncinata and H. dianthus. Any conclusion drawn from ‘direct anatomical evidence must emphasize a very close resemblance in the internal structure as well as in the position of the opercula and branchia. A similar conclusion as regards the morphological worth of the rudimentary operculum can be reached by a recognition of similarity in position on the one hand and the nearly equal appearance of right and left- handed individuals on the other. The functional operculum in Hydroides is therefore morpho- logically a branchia which has formed an expansion at its distal end and which has at the same time lost its respiratory pinnules. 'Ntalics mine. 24 Charles Zeleny. The increase in strength of the supporting axis is a necessary con- comitant of the other changes. The rudimentary operculum cannot be compared directly with a branchia because of its bud-like appearance and embryonic tissues. In position it, however, corresponds perfectly with the functional one and therefore with the branchiz as well. 2. Other Genera of Serpulids. An examination of the different groups of Serpulids brings out the fact that we have almost all gradations between forms with no opercular modification of the branchiz and forms with the single operculum possessing scarcely any trace of a branchial character. a. Group I. No Opercular Differentiation. Examples of Serpulids with no opercular modification of the branchiz are Protula (Risso) and Protis (Ehlers). Each branchia possesses respiratory pinnules and tapers to a point at its distal end. ‘The branchiz resemble one another throughout both right and left circlets. The members of this group are able to retreat for a long distance back into the tube, in this respect resembling the Sabellids which also have no opercula. (See Fig. 6a.) b. Group II. Each Branchia ae a Terminal Enlargement. In Salmacina Dysteri Huxl. there are eight branchiz, four on each side and each has a terminal club- -shaped enlargement. The branchial stalk or axis has from fifteen to twenty pairs of ciliated pinnules. ‘The two rows of pinnules are bordered on the outside by enlarged mucous cells which near the distal end spread out along the sides of the club. ‘This enlarged region bears no pinnules. The eight branchiz are similar in their characters. It is evident that when the animal retreats into its tube these enlarged ends must collectively serve as a stopper for the opening and thus barri- cade the end more effectively than those of Protula which bear no such enlargements (Fig. 6B). c. Group III. Two Equal Opercula, Right and Left, on Ends of Branchia. Branchial Pinnules Present. Filograna implexa resembles Salmacina in having eight branchia. The dorsal one on each side is, however, tena by a small, transparent, chitinous, spoon-shaped structure obliquely attached to the side of the tip of the axis of the branchia. The other branchiz end ‘de St. Joseph, however, seems unwilling to admit an opercular function for these structures. Compensatory Regulation. 25 in short blunt points. - The two opercula are equal in size and the stalks which bear them retain the pinnules and other branchial characters (Fig. 6c). When the animal has withdrawn into its tube the branchiz are twisted in spiral form and the two opercula are superimposed, the one upon the other. The more anterior A—End of branchia of Protula. B—Club-shaped end of branchia of Salmacina (after de St. Joseph). C—One of the two opercula of Filograna (after de St. Joseph). D—Tip of non-operculate branchia of Apomatus ampullifera (X17). £, F—Tip of rudimentary operculum (E) and functional operculum (F) of same (X17). G—Distal portion of functional operculum of Serpula vermicularis ( X 19). one closes the tube after the manner of forms with but one oper- culum. The more posterior operculum, therefore, serves as a protection only in the cases where the barricade formed by the first is not effective. As compared with Salmacina, to which it is otherwise closely related, Filograna has two, more effective oper- 26 Charles Zelen ye cula instead of eight less effective club-shaped enlargements. The fact that when the animal is retracted the expanded portion of one operculum occupies a position in front of the other may be of importance in connection with a theory of the development of asymmetry in these organs in other members of the group. d. Group IV. One Functional Operculum and One Rudt- mentary Operculum. Both on ends of Branchia. Pinnules present. Examples—Apomatus, Josephella. In Apomatus the next to the dorsal branchia on either the right or the left side is expanded at its end into a globular almost trans- parent operculum. ‘The chitinous shell of the sphere itself con- tains irregularly branched blood vessels, the green-colored blood of which makes them very conspicuous. The branchia in a corresponding position on the opposite side has a small ovoid enlargement with a very pronounced network of blood vessels containing distinctly pulsating green blood. Both these opercula are placed at the ends of stalks which retain all the branchial characters in an unchanged condition.’ (Fig. 6p, £, F.) In a few cases the branchia occupying the place of the rudimen- tary operculum ended in a tapering point instead of an ovoid enlargement. ‘There are usually about twenty pairs of branchiz inthe adult. Each of the two circlets breaks off very readily along a definite breaking plane level with the anterior surface of the head. The division plane is very pronounced and the break is clean cut and takes place so readily that it is very hard to remove the animal from its tube without causing it to throw off both of the branchial circlets. e. Group V. One Functional Operculum and One Rudimen- tary Operculum. Functional Operculum with Naked Stalk. Rudimentary O perculum Not on End of Long Stalk. Examples— Serpula, Crucigera, Hydroides. The description g given above for Hydroides (p. 21) is sufficient as a general eect ren of this type. The functional oper- culum may be either on the right or on the left side, the rudimen- tary operculum in each case occupying the opposite position. ‘The opercula are not in the line of the branchiz but occupy a position ‘According to de St. Joseph (°94) the functional operculum appears on the left side and the rudimen- tary, his ‘“‘pseudopercule,” on the right. He does not mention the possibility of the reverse arrangement. The specimens which I examined at Naples showed a preponderance of the right-handed condition. (See p. 32.) . Compensatory Regulation. 27 dorsal to the first dorsal branchiz or to the interval between the first and second dorsal ones. Serpula differs from Hydroides in the entire absence of the secondary group of processes in the oper- culum (Fig. 6G). Crucigera has only four secondary processes and these are arranged in the form of a cross (de St. Joseph, ’94). Fic. 7. Spirorbis Pagenstecheri. Ventral (slightly anterior) view showing branchia and operculum with its brood chamber containing embryos (X 40). f- Group VI. One Operculum. No Rudimentary Operculum. The members of this group have only one operculum. d here 1s no rudimentary operculum. Examples are Spirorbis, Pileolaria, Ditrupa, Filogranula (?), Pomatoceros, Vermilia.. This group may be further subdivided according as the oper- culum has a position in the line with the branchie (Ditrupa, 28 Charles Zeleny. B ‘ 7 Fig, 8. ked stalked operculum in line with branchie of left side. — ture( 15). B; G=Side view and dorsal view of operculum showing highly modified terminal and lateral spines. Basa) suture Present (X10). D—Operculum of Vermilia multivaricosa, showing curved stalk and absence of a basal suture. Dorsal view—Left-handed individual (X8). A—Ditrupa subulata, showing single na Dorsal view. Note indication of basal su of Pomatoceros triquetroides, Compensatory Regulation. 29 Spirorbis, Pileolaria, Filograriula [ ?]) or dorsal to the line of the branchiz (Pomatoceros, Vermilia and others). In the latter case the operculum may further be either at one side of the median line (most species of Pomatoceros and Vermilia) or in the middle line itself (Pomatoceros elaphus, Haswell). In Ditrupa there is a large cup-shaped operculum with a naked stalk situated on the left side in line with the branchiz. It occu- pies a position on the median side of the most dorsal branchia of that side (Fig. 8a). I had only four specimens for examination. In all of these the operculum was on the left side, but whether this was merely a coincidence or not it is impossible to say. The tube of Ditrupa subulata lies freely on the sea bottom usually at a con- siderable depth. Its substance is extremely hard and difficult to break. ‘The tube is slightly curved and resembles very much the shell of the mollusk Dentalium. The Spirorbis-like forms (Spirorbis, Pileolaria, etc.) have closely coiled tubes attached by the dorsal side to a flat surface. In some the tube is attached for its whole length but in others (some species of Pileolaria) the end may rise up from the level of the surrounding surface. ‘The direction of the coil of the tube is constant for any one species but varies in the different species. Thus dextral and sinstral species are distinguished according as the tube is coiled clockwise or counter clockwise. [he dorsal side of the animal is next to the attached surface and the posterior end of the animal upon removal has a pronounced curve to the right or left according as the tube is dextral or sinstral. In the dextral species the operculum is on the right side and in the sinstral on the left so that in all cases the operculum is on the side next to the concave curve of the shell. The number of branchiz varies in the different species from five to twelve, according to Caullery and Mesnil (96). Of the two species examined by me Spirorbis Pagenstecheri had four branchiz on the left side and three plus the operculum on the right and Pileolaria sp. had five branchie on the right and four plus the operculum on the left. In both cases the operculum occupies the position of the next to the dorsal branchia on its side, 7. ¢., the right side in Spirorbis and the left side in Pileolaria. In all members of the group the operculum is in line with the branchiz. It is as a rule much smaller than the opening of the tube so that the animal can retreat to a considerable distance within the tube. ‘There is no sign in either of the two 30 Charles Zeleny. genera of any modification of the branchie to compare with the rudimentary operculum of Apomatus or Hydroides. The right or left position of the operculum is definitely correlated with the direction of curvature of the tube and as the latter is constant for any one species the former must be also. In Spirorbis Pagenstecheri the operculum serves as a brood pouch and is of a trumpet shape. The branchial pinnules are comparatively large and it is sometimes hard to tell whether a basal pinnule should or should not represent a branchial stalk. (Fig. 7.) Puileolaria sp. also uses its operculum as a brood pouch. Other species, how- ever (S. borealis, for example, according to Alex. Agassiz, 66), keep the eggs in a string within the tube on the ventral side of the body. Next comes the group in which the single operculum does not occupy the line of the branchiz but is doreal to it. First are the cases in which it is lateral. In Pomatoceros triquetroides the operculum ts very large and stout. ‘There are two lateral processes beyond which comes an expansion ending in a cap of three spines. There is a very pronounced suture near the base. The whole region below the two lateral processes is flattened dorso-ventrally. In the cases examined the operculum was always on the left side (Fig. 83, c). In Vermilia multivaricosa the operculum occupies a position corresponding with that of P. triquetroides but it may be either on the left or on the right side. The stalk of the operculum is approximately circular in cross section though possessing a corru- gated outer surface. It loops around from its point of attachment toward the median line. The terminal region is thus brought nearer to the median line than is the proximal region. ‘The ter- minal portion is very large and heavy. ‘There is a basal globular portion upon which rests a heavy cone-shaped body (Fig. 8p). Haswell (’85) describes a species of Pomatoceros (P. elaphus) with a large median operculum which is short and flattened dorso- ventrally. At the sides of the proximal portion are two wing-like lobes bearing small processes. Terminally there are three pro- cesses with Fear: like branches. In another Serpulid (Vermilia caspitosa), according to Haswell, there is also a large operculum on a short stubby eal (but not median judging by the figure, though there is no statement in the paper on this point). ‘This operculum is armed terminally with peculiar spines and serrated at a iis a > any te Compensatory Regulation. 31 _ processes and also has two lateral wings like those of Pomatoceros elaphus but not as well developed as the latter. E. Meyer (’88) concludes that these opercula have been formed by the union of two lateral ones, but the evidence as regards this point is by no means conclusive, since we have species where similar opercula are evidently lateral in position (P. triquetroides and V. multivaricosa, for example). g- Summary. ‘The principal modifications of the opercula throughout the family of Serpulids have now been passed over briefly and the general characters may be summarized. In the jirst group are the forms with no opercular modification (Protula, Protis). In the second each branchia has a small club-shaped enlargement, the combination of the enlarged ends no doubt making a more or less effective barricade against invaders when the animal has retreated into its tube (Salmacina). In the third group (Filograna) the modification is confined to the most dorsal branchia on each side. All the others lack opercular differentia- tions. [he two dorsal ones mentioned also retain their branchial characters, but in addition each has an operculum of sufhcient size to close up the opening of the tube. In the fourth group (Apomatus, Josephella) there is one functional and one rudi- mentary operculum, one on the end of each of the two next to the dorsal branchie. The stalks supporting these opercula retain their branchial pinnules and other branchial characters. In the jth group (Serpula, Crucigera, Hydroides) there are likewise one functional and one rudimentary operculum, the distribution between right-handed and left-handed forms being fairly equal in adults. The opercula, however, do not possess branchial pinnules though their internal anatomy and position indicate - branchial characters. The rudimentary operculum is not situ- ated on the end of a long stalk but is a small bud-shaped organ corresponding in position with the functional operculum. Judg- ing by their position the opercula seem to have moved down from the interval between the most dorsal and the next to the dorsal branchiz on each side. Finally 1 in the sixth group there 1s only one operculum and this retains but little of its branchial character. In some of the group, however (Ditrupa, Spirorbis, Pileolaria), It retains its position in the branchial circlet. In some cases it may be used as a brood pouch 1 (Spirorbis, Pileolaria). In other forms (Pomatoceros, Vermilia) it is a considerable distance below 32 Charles Zeleny. the branchial region and is large and massive. In some of the | species as P. triquetroides and ais multivaricosa it has a lateral position and in others, as P. elaphus, a median one. These six groups form a very complete morpbalieae series which points strikingly toward the homology of the opercula and branchia in all the forms. 3. Distribution of the Opercula between Right and Left Sides. The data upon this point will be presented under this separate heading because of their special interest in connection with later discussions. ‘The first three of the groups of Serpulids mentioned above exhibit no asymmetry in their opercula and need not be considered here. ‘The three others will be discussed in turn: a. The fourth group have one functional and one rudimentary operculum. Both opercular stalks have branchial pinnules. (Examples—Apomatus, Josephella.) b. The fijth group have one functional and one rudimentary operculum, each with a naked stalk. Rudimentary operculum not on end of a long stalk. (Examples—Serpula, Crucigera, Hydroides. ) c. The sixth group have only one operculum. No rudimen- tary operculum is present. (Examples—Ditrupa, Spirorbis, Pileolaria, Pomatoceros, Vermilia.) (a.) Inthe fourth group thirteen specimens of Apomatus were examined and of these ten had the functional operculum on the right side and the rudimentary on the left. “The other three had the reverse condition with the functional on the left and the rudi- mentary on the right. Tas.e II. F=Right | F=Left Total No: Rae | R =Right mpunaabas ampulliferde sd... 652s cake eee | 13 10 | : mae ee Eas a2. fe bb a dices’ s | — Wai | | (d.) In the fifth group Hydroides eee H. uncinata, H pectinata and one specimen of Serpula vermicularis were exam- 7 Compensatory Regulation. 33 ined for the distribution of opercula. By far the most extensive observations are on Hydroides dianthus. TaBLe III. Hydrotdes dianthus. Position of Opercula. F=Right. F=Left. Not like other i d Date. 1 No. ES as ie hetal No R=Left. R=Right. two groups. BWoods Hole, Mass. ......... 57 31 24 ok 1g01/1X/16-1X/19. Cold Spring Harbor, L. I.... 74 39 30 st 1902/VII/5-6-7. Cold Spring Harbor, L.I..... 120+ '7 69 51 Pat 1902/VIII/g to 20. 25 tty ay Dotnet =258 139 105 =I4 539% 40-77% 5-4% EXPLANATION OF TABLE. F = Functional operculum; R = Rudimentary operculum. *These two specimens had a rudimentary operculum on each side. {These irregulars come under four heads: 1. One specimen with Right = operculum missing; Left = operculum between rudimentary and functional stage. 2. One specimen with Right = between rudi- mentary and functional; Left = rudimentary operculum. 3. Two specimens with Right = functional operculum; Left = two-thirds developed functional. 4. One specimen with Right = rudimentary; Left = one-half developed functional. {In this case the number of unclassified irregular cases was unfortunately not put down. My notes give only the indefinite statement ‘‘several abnormal and incomplete ones observed are not included in the present list.” Assuming that the percentage of such cases is the same as in the other two groups, where it is 5.4 per cent of the total number, we will not be far astray in making the unknown number equal to seven. The relative relation between right-handed and _ left-handed forms is expressed to better advantage if the irregular cases are not included. Removing the last column from the former table we get the relations expressed in the following one: An examination of this table shows that 57 per cent of the “ nor- mal” cases have the functional operculum on the right and the tudimentary on the left, while 43 per cent have the opposite arrangement. ‘The striking agreement in the three sets of figures, 34 Charles Zeleny. one from Woods Hole and the other two from Cold Spring Har- bor, indicates the probability of some organic basis back of the fact. This matter will come up again later in the discussion of the development in ontogeny and during regeneration. At the same time the considerable number of cases (fourteen [?] or 5.4 per Tas_e IV. A ydrotdes dianthus. Position of Opercula (not including irregular ones). = | | = | F=Right | F=Left Locality and Date | R=Let | R=Riame | Total Woods Hole, Mass., No. 31 24 55 1900/TX/16=19-. she eo Per cent | 56.4 | 43.6 |) ee Cold Spring Harbor, L. I., | 39 | 30 | 69 FQO2/ VAT / 5-0-7 once clea ss Per cent | 56.5 | (age — Cold Spring Harbor, L. I., No. 69 | 51 120 TG) VEE G—po ers ot eter: Per cent eG 42.5 _ No. 139 105° 244 Matale cake so i) sis eee ae Percent | 257 43 _ TasBLe V. Hydrotdes uncinata. Position of Opercula. F=Right | F=Left R=Left | Reishee Locality and Date No. fe) 6 16 Maples: t902/ M0/2T $5... 20/. she] Perveent 62.5 | eae — cent of the whole number) which do not come under either of these _ groups will be taken up. Sixteen specimens of Hydroides uncinata were examined at Naples. Of these ten, or 62.5 per cent, had the functional oper } culum on the right and the rudimentary on the left, and six, o 37-5 per cent, had the reciprocal arrangement; a considerable advantage in favor of the right-handed ones. ! i \s t ; | | Compensatory Regulation. 35 Likewise 41 specimens of Hydroides pectinata were examined at Naples, and of these 21 or 51.2 per cent had the functional oper- culum on the right side and the rudimentary on the left while 48.8 per cent had the reciprocal arrangement, a slight advantage in favor of the right-handed ones. TasLe VI. Hydroides pectinata. Position of Opercula. F=Right F=Left R=Left R=Right Locality and Date Total Naples, 1902/1X/27 No. 21 20 41 1036) NAY) Oe eee ee are Per cent Sires 48.8 — In Serpula vermicularis [ examined only one specimen and this had the functional operculum on the left side and the rudimentary on the right. Our general conclusion regarding the distribution of the oper- cula in the fijth group of Serpulids, those with the naked-stalked functional and small bud-like rudimentary is that the right and left-handed forms are nearly equal in number, but there is a slight advantage in favor of the right-handed ones. (c.) Inthe sixth group there 1s only one operculum. In Ditrupa subulata only four specimens were examined as regards this point. All four of these had the operculum on the left side. In the dextrally coiled Spirorbis Pagenstecheri eleven specimens were examined and all had the operculum on the right side, while in the sinstrally coiled Pileolaria the one specimen examined had the operculum on the left. The observations of Caullery and Mesnil (’96) show that in the species with dextrally coiled tubes the operculum is always on the right side and in the sinstrally- coiled ones always on the left side. Ide St. Joseph (98) gives a description of Ditrupa arietina, O. F. Miiller (= D. subulata, Desh.) in which he mentions the operculum as a structure of the left side in agreement with the present observa- tions. Also the left side according to him has one less branchia than the right, i. e., there are 11 branchie (plus the operculum) on the left and 12 on the right. 36 Charles Zeleny. In Pomatoceros triquetroides 21 specimens were examined and all had the operculum on the left side. This makes a fairly strong probability in favor of the permanence of such a charac- teristic. As a further argument may be mentioned the statement of de St. Joseph (’94), who mentions the observation of Grube on 63 specimens of P. triqueter, L.( = P. triquetroides, D. Ch.) and of himself on many more than this number which showed the oper- culum in every case on the left side, so that we may be fairly certain that in this species the organ is a permanent structure of the left side. In Vermilia multivaricosa eleven specimens were examined. Six had the operculum on the right and five on the left side. In this case, therefore, there seems to be a fairly equal distribution between the two sides. The data for Group VI are collected in the following table: TasLe VII. Group Six. Position of Operculum. Locality: Bay of Naples. Name and Date. No. Operce. Operc. Percent Per cent Right. | Left. | Right. | Befr Dtripasubulatas i). % yi ay Se foots 4 Oo 4 o | 100 1903/I/t1o. SPIFOLbIS Eapenstechent <=)... emcysie II II fo) | 100 | fe) 1903/I to V. | | Piteolanasps (5) tice whgenine oe Sse I fe) I | o | 100 1903/III/2. Pomatoceros triquetroides .......... 22 Oo 22 fo) 100 1903/1/25, IV/3, IV/s. Mermurlia multivaricosas...- 25 s<: 0.25. II 6 5 55 45 1903/1/16, III/s. | Discussion of the Evidence from Group VI. In Spirorbis Pagenstecheri and Pileolaria sp. the position of the operculum 1Former observations on the position of the operculum in this group are those of Caullery and Mesnil (96), who state that dextrally coiled Spirorbis-like Serpulids always have the operculum on the right side (example—Spirorbis Pagenstecheri) while sinistrally coiled ones have it on the left side (example— Pileolaria sp.[?]). de St. Joseph ("98) describes the operculum of Ditrupa subulata, Desh., as a struc- ture of the left side. de St. Joseph (94) for himself and also for Grube describes Pomatoceros trique- troides as always left-handed. Compensatory Regulation. ay bears a direct relation to the direction of the coil of the tube. Caullery and Mesnil (’96) have shown that the operculum of Spirorbis and its relatives is always located on the side next to the concave surface of the coil, 7. ¢., on the right hand side in dextral tubes and on the left-hand side in sinistral ones. Whether or not any such relation can be made out in other members of Group VI It is not possible to say. In Ditrupa there is a slight curvature of a definite form in the tube but the relation of the body within the tube has not been made out definitely enough to determine the significance of the constancy of position of the operculum. In Pomatoceros triquetroides there is a definitely fixed left-handed- ness though the tube is irregularly coiled, and in Vermilia multi- varicosa there is an almost equal distribution between the two sides, the tube being again irregularly coiled. A discussion of the factors controlling the determination of the position of the opercula must be left until the ontogenetic and regeneratory development of these organs have been studied. 4. Exceptional Degrees of Development in General and One Case of a Supernumerary Operculum. The two opercula in Groups IV and V in the vast majority of cases consist of a fairly typical functional and a fairly typical rudimentary operculum. ‘There are, however, exceptions to this statement, as has already been indicated above. The main exceptions are due to the partial further development of what probably would otherwise correspond with the rudimentary oper- culum, either with or without the loss of the functional operculum. There thus arise either one functional operculum and one partly developed functional or a missing operculum and one partly developed one. In other individuals both opercula were found to be rudimentary or one rudimentary and one partly developed one were present. The discussion of these cases must be reserved until we come to the experimental part of the paper. ae One case of a supernumerary operculum was found and this 1s interesting in connection with the problem of the regulation of the opercular development. The accompanying figure (Fig. 9) gives the relations of the opercula. On the left-hand side there 1s a rudimentary operculum of the typical form in the usual position. On the right-hand side in a corresponding position is a typical 38 Charles Zeleny. functional operculum, but in addition to it there is an added rudi- mentary operculum with its point of attachment posterior to that of the functional. This added rudimentary operculum agrees in all respects with the one on the opposite side, so that we have two rudimentary opercula and one functional one. ‘The specimen indicates that the influence which determines the development of a functional or a rudimentary operculum is not always a bilater- ally differentiated one. A more complete discussion must be reserved until the experimental data have been given. 2. Development of Opercula. I. Ontogenetic Development. a. Introduction. Ina paper, entitled “A Case of Compensa- tory Regulation in the Regeneration of Hydroides Dianthus,” I described some experiments WN ; , showing that when the func- Wa tional operculum of this SHA! Serpulid is removed the cE : rudimentary operculum on the opposite side develops into a new functional oper- culum similar to the old one while in place of the old functional stalk a new rudi- mentary bud develops. In the discussion of this and similar experiments it was stated that a knowledge of the ontogenetic development Hydroides dianthus with three opercula, two rudi- of the organ 1S highly desir- mentary and one functional. Note that one of the able before we can be in a rudimentary opercula is attached near the base of the position to discuss the data functional one ( X 15). intheirfull relations. With this object in view the writer undertook to raise the larva up to the stage where both opercula have attained their normal adult development. In attempting a provisional explanation of the compensatory regulation of the opercula it was stated in the above paper that there may be a restraining influence exerted by the fully developed Fic. 9. Com pensator » Regulation. 39 operculum upon the rudimentary one which prevents the latter from attaining its full development. The removal of the func- tional operculum removes the restraining influence and the rudi- mentary continues its development. The new functional oper- culum in turn restricts the new bud developing in place of the old functional operculum and holds it at the rudimentary stage. The plausibility of this explanation is increased if it is found that in the ontogenetic development one operculum develops before the other, and therefore can hold the latter in check in the manner before indicated. With this object in view the investigation of the ontogenetic development was undertaken. b. Historical Review. ‘The first recorded observations on the development of the branchial apparatus in Serpulids which I have been able to find are those of Milne-Edwards (45), on the development of the young Protula. He saw the larve attach themselves to solid objects at the bottom and sides of his dish. Here they secreted a cylindrical tube which at first was open at both ends and shorter than the length of the larva. At about the same time two lobes were differentiated at the anterior end of the larva, one on each side of the median line. At a slightly later period he thought he saw digitations of these lobes and he took them to be the beginnings of the branchiz. Pagenstecher (’63) gives an account of the development of the branchie and operculum in Spirorbis. He states that the first traces of the branchiz are exhibited in the form of three knobs upon each of the two head lobes. “The operculum is not differentiated until a later time when there is formed “‘der Fortsatz welcher ihn tragen soll und der von den Tentakeln durch eine Runzelung oder seichte Kerbung der Oberflache ausgezeichnet war.’’ Judging by Pagenstecher’s figure there is very little difference at the above-mentioned stage between the so-called opercular outgrowth and the other branchie. ‘The figure gives two branchiz on one side and two plus the opercular outgrowth on the other. Fritz Miller (’64) noticed on the side of a glass vessel which he had on his study table a young Serpulid with three pairs of branchie, and which he took to be a member of the Protula group because of the absence of an operculum. However, a short time later he noticed that one of the branchiz had an opercular enlargement at its end though it still retained its branchial pinnules. Still Jater the branchial pinnules disappeared also and he had a Serpulid with the typical genus-Serpula- type of operculum, which had developed by a modification of a branchia. In the meantime a new pair of branchia had been added to the oral crown making three branchiz plus one operculum on one side and four branchiz on the other. ‘This is he only recorded observation of the transformation of a branchia into an operculum. 40 Charles Zeleny. In 1866 Agassiz described the development of branchiz and operculum in Spirorbis spirillum, Gould (not Lamarck), and made out an alternate appearance of the tentacles (branchiz). “‘The first tentacle appears on the right, next comes the corresponding tentacle on the left and only later the rudiment of the odd opercular tentacle (on the right side).”” The rudiment of the operculum, though at first somewhat resembling that of the tentacles, shows a difference from the start. Claparede and Mecznikow (’69), on the contrary, make out a patred mode of forma- tion of the branchiz in other species of Spirorbis. | Willemoes-Suhm (’70) speaks again of an alternate mode in Spirorbis. Giard (’76b) raised the larvae of Salmacina Dysteri. He found two lateral head lobes each of which soon showed a threefold division. ‘These divisions elongated to form the first three pairs of branchiee. On each side there were two dorsal and one ventral branchia, the latter, however, dividing into two on the fifth day, so that there were then present eight branchial trunks, four on each side. The first pinnule appeared on the eighth day on the upper third of the external dorsal branchia. This is the only notice of pinnule formation I have found. In Manayunkia, a fresh water Serpulid, Leidy (’83), describes the head lobes as showing the branchial digitations from the first trace of formation of the former. Salensky (’83) likewise states for Pileolaria sp. that four branchiz and the operculum appear at the same time from a median dorsal plate. The opercular ‘anlage”’ is from the beginning three to four times as large as the branchial “anlagen.”” In Salensky’s words: ‘‘On voit d’apres cette description, que, chez Pileolaria la formation des branchies et de |’opercule s’opére en méme temps, et « non comme Agassiz et Pagenstecher le montrent pour Spirorbis spirillum.” Meyer (88) describes the development of the branchize in Eupomatus (= Hy- droides). He makes out the appearance of the two head lobes from each of which the three processes representing the first three branchiz sprout out. The develop- ment was not carried further than this. This method of formation agrees also with that described by Roule (’85) for the larve of Dasychone. From the foregoing notes it is evident that further observa- tions on the early development of the branchie are necessary in order to clear up our ideas regarding the matter, and when we come to the opercular development we have practically nothing outside of the observations on the highly modified forms Spiror- bis and Pileolaria except the short note of Fritz Miller upon the change of a branchia into an operculum. Regarding the forma-_ tion of the rudimentary operculum there is nothing at all, and we therefore get very little aid in our study of the correlation between | | ee ee ee eee See Compensatory Regulation. 41 the two opercula at their first appearance and up to the time when they assume the final adult condition. The following observations on H. dianthus were made at the Cold Spring Harbor Biological Laboratory on Long Island, N. Y., during July, August and September, 1902. The observations on the other species, ; H. uncinata and H. pectinata, were made at the Naples Zoological Station in the winter of 1902-03. The obser- vations on the method of rearing the larve and on their general activities will be given in a separate short note.! c. Observations. When the free-swimming larva is about nine days old its body is considerably elongated and shows external signs of segmentation. ‘he apical ihe is long and two very promi- nent. reddish eyespots are present. (Fig. TOA.) As the move- ments of the animal become more and more sluggish just before its fixation to a solid object the apical cilium gradually grows smaller until it disappears entirely. At the same time three pairs of sete are formed, the first pair being especially long and promi- nent. (Fig. ros.) Fig. 1oc shows a larv a, 17 days old, with the tube covering about one-half of the body? Here ‘each of the two lateral head lobes already shows the division into three blunt pro- cesses, the forerunners of the branchiz. ‘These divisions of the head lobe appear very soon after the formation of the head lobe itself but the latter has a short separate existence before the tripartite subdivision appears. The two dorsal processes are more closely connected together than with the third and more ventral one. The relation is more clearly made out in Fig. rop, where the mutual union of the two dorsal pairs is very evident. The two larve (Fig. 1oc and Fig. 10p) are of the same age (17 days) and come from the same dish. In Fig. tok there are still the same three pairs of processes though he branchial character is more evident than before. This specimen is from the same dish as the others (16 days after fertilization). It is evident from these data that the rate of de- velopment of the different larve in a single dish varies within wide limits. At the stage represented in Fig. IOE the inner sur- faces of the branchiz are pied with very active cilia. In order *Biological Bulletin, °05, vol. viii. 2The tube secreted by the animal is at first a very short cylinder which is quite transparent and covers only a small part of the larva. The ring at first is situated near the anterior end of the body just back of the eyes. a? 42 Charles Zeleny. to get at the method in which these primary processes branch later on, I have designated them arbitrarily by the Roman numerals I, II and III. I represents the most dorsal pair, II the middle pair and III the ventral pair, R or L being prefixed to represent right or left when there is any need for distinction between the two sides. Fic. ro. Hydroides dianthus. Larve. Stages of transformation from free-swimming to sedentary life. Dorsal views. A—Free-swimming larva. Age, 9 days. Shows long apical cilium (208). B— Swimming but sluggish larva. Age, 9 days. Shows sete and shortened apical cilium (185). C—Attached larva (go). Age, 17 days. Three pairs of head lobes. Beginning of secretion of tube. D—Age, 17 days (90). E—Age, 16 days (X85). F—Age,17 days. Second of original three pairs of branchia shows secondary branches ( X 70). Compensatory Regulation. 43 In Fig. ror, also 17 days after fertilization, we find a stage con- siderably more advanced in which each of the middle branchiz has already sent out three branches, which may di be labeled according N \ | /| } f] { / | | | / / | to age, I]—1, II—2 | and II—3. Ofthese it oo ites) willbe seen that II—1 77 , \\ | | very early takes on \ Wed Re harscter of one ~ of the main branchial a trunks, so that we thus geta stage with four if pairs of branchiz. : Wi I]—2 and II—3 retain the characters of pin- Id nules of the Branchia q II. Neither I nor II has any branches at this stage. Fig. 11A shows a f later stage taken from the same dish at the Same time (17 days). Here we still have no branches of I and Ill. Branchia II has five branches and the SS, fest branch of L-II _ aap (L-II—1) a branchlet Za if i of itsown (L-I[—1—1) (pao, just appearing as a LEE I rN very small knob. It is very evident that Branchia II is rapidly outgrowing | and III in strength. Fig. 118 (19 days old) gives a still older stage. Branchia I on each side has not increased much in size and is hardly larger than UT: / Hydroides dianthus. Dorsal views. d—Age, 17 days (X47). B—Age, 19 days ( X 62). Fic. 11. 44 Charles Zeleny. the older branches of II. L-II—1 has two branchlets (= pin- nules) and R-[]—1 has one branchlet. Branchia II on each side now shows the beginning of the sixth branchlet or pinnule. Branchia III also has not increased much in size. In Fig. 12 (23 days old) Branch I1—1 with its six pinnules has very evidently taken its place as a prominent part of the branchial Fic. 12. Hydroides dianthus. Age, 23 days (75). The first secondary branch of Branch II has assumed the character of an independent main branch with branchlets of its own. The original third pair of branches is not shown. apparatus. Branchia II has now added its eighth pinnule. Branchiz I and III are still unbranched and have increased com- paratively little in size. In Fig. 13 (23 days old), in which only the right side is repre- sented as both sides are essentially similar, Branchia I for the first time shows a trace of branching, seven little branchlets (pinnules) appearing simultaneously. Branchia II has added one new branchlet making nine in all counting Branchia II—1 as the first Compensatory Regulation. 45 one. Branchia I]—1 has eight pinnules. The character of Branchia III was not made out. The operculum was seen for the first time six days later at the stage shown in Fig. 15. Fig. 14, taken five days later still (34 days old), shows an earlier stage of the operculum. Here it appears as a rounded knob on the end of Branchia L-II. ff | / ; This branchia has at a | | y | this time nine branch- /| eee at y, lets counting the one I | y; ({I—1) which has as- | | LM TT: pe SE form with the base of the cone free and the apex attached to the stalk. The corres- Dy Ss ponding branchia on INN the right side (R-II) AEN has likewise at this | \\ ie: stage nine branchlets | \\ counting the independ- \\ ent one ([I—1)or eight | dependent ones (pin- \ nules) II—2 to II—g \\ without this, but there 4 is no modification at 3: Rae at the tip of the main stalk yeas dianthus. Age, 23 days. Dorsal ane (X75). Original third pair is not shown. First pair shows beginnings as occurs on the left ~"*° P side. In Fig. 15 the opercular character of the knob on L-II is very evident. The cup of the operculum has a notched edge with eleven serrations and resembles in its character the Serpula type and not the Hydroides type. ‘The eight dependent (I]—2 to I]—9) and one independent branch are very prominently developed sumed the character | 7 ge a» Main branch. | The eight represented as pinnules are []—2 > to Il—g9. The knob at the end of the i om branchia is conical in » of secondary branches. 46 Charles Zeleny. and evidently all of them retain their respiratory character. ‘The corresponding branchia (R-II) on the right side has branchlets similar to those of the branchia (L-II) on the left side though it shows no opercular modifications. It is similar to the latter and different from all the other branchiz in one essential respect. While I, I[—1 and III show buds of new developing pinnules its pinnules (7. e., the eight dependent ones [I—2 to I1—9) are all Fic. 14. Hydroides dianthus. Age, 34 days. The next to the dorsal branchia on each side. The left one shows the beginning of the knob of the functional operculum. The right one later drops off and the rudimentary operculum regenerates from the remaining stump. well grown, indicating, no doubt, that the organ has reached its limit of development as a branchia. With the beginning of the opercular differentiation Branchia I enters on a new period. It increases in size and new branchlets (pinnules) sprout out in rapid succession, the first ones appearing — simultaneously. Thus in Fig. 15, I already has thirteen well formed pinnules with several buds crowded into a zone at the base of the terminal filament. Branchia II—1 has likewise been Compensatory Regulation. 47 increasing in size and now has 12 pinnules plusa thirteenth bud on each side Branchia III is also branched but the character of its innules is not given in the figure in order that complication may be avoided. This pair of branchie is directed away from the observer toward the ventral side of the animal. At all these stages the Branchie L-Il] f and R-I1 both retain their flexible and branchia-like char- acter in all respects except for the ter- minal enlargement im L-I1. In Fig. 16 the pin- Z nules of L-II have disappeared, leaving iV a cup-like opercu- lum with a serrated edge on the end of a long, slender, flexi- ble but bare stalk. The corresponding branchia on the other side (R-II) has dropped off, leaving 7 only a small round \ knob at the place \ of its former attach- ment. q The other Hydroides dianthus. Age, 28 days. Shows the three most dorsal branchie, I I[—1 pairs of branchia. The ventral pair directed away from observer is and EDI, all ace keep- not shown. The opercular knob of the left side is notched and its ing on with their ‘talk still retains the respiratory pinnules. The next to the dorsal Hic. 15. branchial develop- aE of the right side has eight pe but duter: icon other : branchia, except the opercular one, in the absence of new pinnule ment by continually ,,,j, adding new pinnules to the old ones. We thus have three pairs of branchize at this stage with a functional operculum of the Serpula type on the left side and a rudimentary operculum on the right side. No young Serpulids were observed that showed an exception to this order of appearance of the opercula. Numerous observations were made be 48 Charles Zeleny. upon specimens before a count was undertaken. In this count twenty individuals were noted. All without exception, the former as well as the latter, showed the functional operculum appearing on the left side. Fig. 17a shows the rudimentary operculum at a slightly later stage in which it has more definitely assumed its typical condition. Fic. 16. Hydroides dianthus. Age, 34 days. Dorsal view (38). On left side is functional operculum of Serpula type with naked stalk and cup with one row of serrations. On right side is rudimentary bud developed from the base of the cast-off second branchia of that side. The three pairs of typical branchie also are shown. The condition at this time resembles very closely that of the adult Serpula, as the functional operculum as shown in Fig. 16 is without — doubt of the Serpula type. he further changes were not followed in H. dianthus, the species at Cold Spring Harbor, but were made out in the two Naples species, H. pectinata and H. uncinata. A—H. dianthus. Age, 34 days. Primary rudimentary operculum, right side (X50). B—H. ctinata. Age, 45 days. Dorsal view. Functional operculum of Serpula type on left. Rudimentary operculum on right. Pinnules are shown in only one branchia, the other branchia being essentially ilar to this one. C—Tube of H. pectinata. Age, 45 days (6). D—Hydroides uncinata. Age, days. Ventral view. Shows Serpula type of functional operculum on left side and rudimentary euuny on right. E—Diagram of cross-section of a branchia of H. uncinata showing method of ment of pinnules. F—H. uncinata. Age, 179 days or nearly six months. Original or Primary tional operculum as appearing after it has dropped off. 50 Charles Zeleny. Evidently the adult condition with an operculum situated on either the right or the left side and having two rows of processes at its distal end: is not fully explained by the Cold Spring Harbor observations. Only three larvz out of a great many sets of eggs started at Naples lived through the period | of attachment. io of these were H. uncinata and the other H. pectinata. “These were first care- fully observed only after they had developed up to the stage corre- sponding to Fig. 16 of H. dianthus. In Fig. 178 is represented a specimen of H. pectinata with essen- tially the aoe characters as those of H. dianthus in Fig. 16. The number of branchiz has, however, meantime increased and there are here besides the opercula five branchiz on the left side and four plus the bud of a fifth on the right side. The functional operculum has the essential characters of the Serpula type (see above). [he new branchiz are being added on the ventral edge of each of the branchial ridges. Both opercula have moved down from the line of the branchie and the gap left in the line by their absence is being closed up. ‘The character of the tube at this stage is shown in Fig. 17c. The tube was so extremely irregular in shape, largely Baie it was detached from the glass frequently in order to 5 facilitate observation. Practically the same conditions are shown at this time in H. uncinata where there are on each side five branchiz plus the oper- culum: The opercula here as before have the characters of the Serpula type. (Fig. 17D.) No further changes in the opercula were noticed for a long time. Finally, six months after the fertilization of the ov a, the animals were again carefully observed, and it was noticed that the primary functional operculum (left side) had fallen off (Fig. 17F) and in its place a rudimentary one had developed, while the primary rudi- mentary operculum of the other (right) side had developed into a functional one. (Fig. 18a, B, c.) The two specimens of H. uncinata retained their simple Serpula-like operculum longer than did H. pectinata. Returning to the specimen of H. pectinata as it was found afte the reversal of the opercula we find all the adult characters except the full number of branchiz. ‘The branchiz increase in number by additions along the ventral edge of each branchial ridge. In speci mens at this stage there are beside the opercula seven branchiz Compensatory Regulation. 51 plus a very small bud on the left side and six branchiz plus a large bud on the right. The functional operculum has a basal funnel- shaped cup with a serrated edge. From the upper flat end of this cup there projects a new secondary cup, the individual serrations of which reach nearly to the base and are strongly toothed. Fic. 18. Hydroides pectinata. Age, 181 days or 6 months. 4A—Ventral (somewhat inclined) view, showing position of secondary functional and rudimentary. opercula. Pinnules of branchie are not shown in the figure (16). B—Secondary functional operculum (adult Hydroides type with two rows of pro- cesses) (32). C—Rudimentary operculum ( X 32). d. Summary of Data and Discussion. The above results con- cerning the ontogenetic development of the opercula may be sum- marized as follows: At the stage with four pairs of branchiz the next to the dorsal one on each side stops its development when it has eight long slender pinnules. The two branchiz of the third pair, counting from the dorsal side, arise originally as branches which resemble in all respects the ordinary pinnules, but instead of retaining their 52 Charles Zeleny. dependent condition increase in strength, develop secondary branches of their own and soon take their place as independent branchie coequal with the three primary pairs. After reaching its limit of growth as a branchia the next to the dorsalmost branchia on the Jef side starts a new differentiation at its end, developing a knob which rapidly increases in size and soon assumes the shape of an inverted cone. Along the edge of the upturned base notches appear so that the whole knob has the general character of the opercular cup of members of the genus Serpula. All this time, however, the stalk has retained its eight branchial filaments and the corresponding branchia on the other side has remained unchanged. ‘This stage corresponds in general with the adult of Filograna or rather with Apomatus except that only one operculum is present. ‘The branchial filaments of the stalk, however, soon disappear. Whether they drop. off or are resorbed was not made out but the former supposition is the more probable one, as they were still very long a short time before they had entirely disappeared; or, in other words, no intermediate stages of resorption were seen. Almost coincidentally with the disap- pearance of these pinnules the next to the dorsal branchia of the } right side drops off, the region of the break being near the base. From this broken stump a bud develops which in a few days has ~ reached its limit of development for the time being. This bud, which remains as the primary rudimentary operculum for several months, is a regenerated structure, a true case of phystological heteromorphosts. Furthermore, it is restricted in its development ” by some forces acting from without its own substance. At this stage the opercula remain for a considerable time (several months) . with no further change, although at the same time the animal 1 atamdin es siren meena has heneetinanhlipeenieee lite te FL A alee a te v? increasing in size, is building up its tube and new branchiz are being sales on the ventral edge of each of the branchial ridges. In its essential characters this stage is equivalent to that of adult members of the genus Serpula w ith the functional operculum on the left side. After this long period of no opercular change the primary func- tional operculum drops off, the stalk breaking near its base. Immediately the primary rudimentary operculum on the right side, no longer restricted by outside forces, starts its further devel- opment and becomes a functional opercular organ. However, 1 / does not develop into an operculum of the simple type like the pri- Compensatory Regulation. 53 mary functional one. Instead it takes on the characters of the adult Hydroides operculum with two rows of serrations. Beside the inverted cone with serrations around its upper edge there is an additional circlet of pointed and often hooked processes, which constitute the most important character of the Hydroides group as distinguished from the Serpula group. At the same time the broken stump on the left side has started to develop a knob of embryonic tissue which grows only up to the stage represented by the rudimentary operculum of the adult and is in its turn restricted in its further development by some force most likely similar to the one which in the first place restricted the original primary rudi- mentary operculum. ‘There are, therefore, at Has time a secondary functional operculum on the right side and a secondary rudi- mentary operculum on the left side. These have the essential characters of the opercula of adult specimens of Hydroides. However, one point of difference is evident in specimens taken at random from the sea. It is found that approximately the same number have the operculum on the left side as on the right though there is uniformly a slight advantage in favor of the right-handed ones (57 per cent right-handed to 43 per cent left-handed in H. dianthus), while all the larvae appeared first as left-handed ones and later by reversal changed to right-handed ones. How does this change occur? Either we must suppose that the similarity in character of all the larve was accidental or that the reversal takes place in nature during the life of the individual more than the one time described for che young animal. The first supposition seems improbable because the oe came from a great many different individuals, and moreover the order of appearance was found to be the same in the two Naples species (H. uncinata and H. pectinata). We are, therefore, forced to assume a further reversal as taking place in nature. This reversal may be a purely phy siological one, induced by the normal activities of the animal, his the first reversal already described, or may be induced by some injury to the functional operculum of the character which was found to cause such a reversal in my experiments. (See below, p. 55 7.) However, unfortunately, there is no experimental evidence to show that physiological reversal takes place in nature after the first time already described. As is mentioned elsewhere (p. 65) in this paper the experiments undertaken to determine whether worms 54 Charles Zeleny. kept in dishes in the laboratory exhibited physiological reversal were negative, although the time was in all cases too short to constitute a good test. There was no change in any case unless the functional operculum was injured. However, the very near equality between right and left-handed individuals seems to pre- clude the possibility “of all reversal being due to injury of the func- tional operculum. And we have beside the analogous case of physiological reversal in the young animals as has been empha- sized before. 2. Regeneratory Development. a. Introduction. The nature of the opercular modifications among the Serpulids has now been outlined in the discussion of the comparative anatomy and their orzgim within the individual life history has been traced in the genus Hydroides. There now remains an attempt at an experimental analysis of the factors involved in the development and maintenance of the adult charac- ters. Partly because of the imperfection of the method and partly because it is not desirable to dissect the data too minutely while presenting them, the latter will be given in the descriptive portion of this section as parts of individual experiments without perfect regard to logical development of the analysis of the factors involved. ‘The latter will be attempted more fully in the general discussions to follow the descriptive portions. In a former paper a preliminary report of some of my experi- mental results on compensatory regulation in the regeneration of Hydroides dianthus was given. Sine that time a more detailed series of experiments fae been undertaken along the same lines and the work has been extended to other species of the family. Two distinct problems have come up. In the first place is the study of the factors involved in the compensatory regulation of the opercula which, as stated above, is the main object of the present paper. In the second place a comparison of the regen- eratory development of the opercula with the ontogenetic and with the probable phylogenetic development brings up an ex- tremely interesting discussion with important bearings on the recapitulation theory. The great majority of the experiments were performed on H. dianthus and this form will be discussed first. Then will follow the other members of Group V, namely, H. uncinata, Compensatory Regulation. 55 H. pectinata and Serpula vermicularis, then a member of Group IV, Apomatus ampullifera, and finally the members of Group VI, Ditrupa subulata, Spirorbis Pagenstecheri, Pomatoceros tri- quetroides and Vermilia manltivaticosa. The adult condition of the opercula which has already been described in detail in the anatomical portion of the paper will be again briefly noted, as it must serve as the basis of our experiments. The experiments will then be described in turn, and finally the results will be discussed. b. Unoperated Condition of the O percula in H ydroides Dianthus. The character of the adult opercula has already been given on p. 21 ffl. and is also shown in Fig. 5£, F. A repetition of these data is therefore unnecessary. In all the experiments about to be described the animal was first removed from its tube and placed in a dish of sea- water, the desired operation was performed under a dissecting microscope and the animal was kept in its individual dish either with running or standing water, in the latter case the water being changed once or twice a day as required. The running water was not found as favorable as the standing because of the collection of a fine deposit on the animals notwithstanding the greatest care exercised. The dishes with standing water were found very favorable if provided with a glass cover to keep out the dust. The observations were in most cases made on the living animals. c. Operations on Functional Operculum. ‘The results may best be arranged around a description of the effect of a cross cut through the stalk three-fourths of the distance from the base to the beginning of the terminal expansion. The stump of the functional stalk remains attached to the animal for two or three days as a rule or even longer in some cases. It then breaks off from the body, separating by a clean division at a basal suture or “breaking joint’ described above (pp. 21-23). the distal end of the small stump still remaining attached to a animal a small bud now appears and gradually increases in size until it reaches the dimensions and character of the former rudi- _ mentary operculum of the opposite side. At this point it stops and proceeds no further. In order to understand the result it 1s neces- sary to look away from the immediate vicinity of the operated organ and to note the change going on ina corresponding position on ‘the other side of the animal. Even before the attached stump of the 56 Charles Zeleny. operated functional operculum has fallen off the rudiment operculum has started to enlarge and processes appear at its discal end. Gradually the structure increases in size and assumes Hina Fic. 19. Hydroides dianthus. 4, a~{—Stages in the development of a rudimentary operculum from the s of an old functional one (54). B, a-e—Stages in the regeneration of a rudimentary opercul after a cut at the region indicated by the double-headed arrow (X54). C, a-d—Stages in the t formation of the old rudimentary operculum into the new functional after removal of the old functional operculum. The final condition of the functional operculum is shown in Fig. 5r. Compensatory Regulation. 57 more and more the character of the former functional operculum. After 15 to 20 days the development is complete and we have a complete reversal of the opercula. [he former functional oper- culum is now the rudimentary and the former rudimentary has become the functional. The resulting arrangement is the exact reciprocal of the former one. The case just briefly outlined will now be taken up in more detail. Before the stalk of the functional operculum is cut it is almost impossible to pull it off from the animal. ‘The basal suture resists breaking as well as the solid material of the stalk. The hardest kind of a pull that can be given with a pairof forceps 1s not sufh- cient to dislodge the organ. Inside of a few days after the operation, how- ever, the opercular stalk as we have seen comes off of its own accord, so that great changes must be assumed to have taken place in the region. The time at which the stalk drops off varies greatly. In one case it had not come off 5 days after the operation and in ae after 6 days it was still attached. Very soon after the stalk has dropped off a bud appears at the top of the stump and this stead- ily increases in size until the ty pical form of a rudimentary operculum is ements canuverce cecuon! of scale reached. (Fig. IQ A, a—f.) It seems (X26). Exceptional case. Differentia- very probable that the breaking oft of tion at the distal end of a functional the remanent of the stalk is radioed Ses Wi ea) NESS Ges aes the distance from the base to the termi- by histological changes in the suture region which are attendant upon the beginning of the development of the new bud. Evidence in favor of this view will be given later in another place. Except in one case no differentiation of tissues took place at the cut end of the functional stalk. In this case there was an expan- sion on each of the two sides of the stalk near its tip. The stalk did not drop off for 5 days after the operation. The changes are represented in Fig. 20. Fic. 20. Opercula of Hydroides dianthus, five nal cup. The old rudimentary is also shown. 58 Charles Zeleny. The changes taking place on the other side of the body are very evident 3 or 4 days after the operation, often before the rema- nent of the functional stalk has dropped off. This fact that the rudimentary operculum begins to develop even before the func- tional has dropped off seems to argue against the mere retarding influence of the organ asa mechanical weight, though of course the weight is considerably diminished by the removal of the part of the operculum distal to the cut. The stages through which the rudimentary operculum passes in changing into a functional one are given in the accompanying figures. (Fig. 19c, a-d.) The eel points to be emphasized 1 in fie development are: first, the fact that throughout there 1s no sign of the appearance of special bran- chtal characters, and, second, that the secondary processes of the operculum appear before the primary ones. ‘The regeneratory development, therefore, differs widely from the ontogenetic or probable phylogenetic one as regards these points. A series of operations was performed on the functional opercu- lum to determine the amount of injury necessary to bring about reversal. ‘The cases overlap slightly but it is found that in general the cutting off of the distal circlet of processes does not induce reversal while cuts through the main enlarged portion of the oper- culum bring about a reversal of the opercula. In one case the stalk of the injured operculum remained attached though the rudi- mentary operculum in the meantime had reached a stage equal to three-fourths of the normal functional dev elopment. It may be concluded that a removal of the secondary circlet of processes of the junctional operculum does not as a rule cause reversal, while a stmilar injury below this point to the main portion of the cup or to the stalk of the operculum always brings about such a result. d. Operations on Rudimentary Operculum. When the rudi- mentary operculum alone is removed there is no effect upon the functional operculum and a new rudimentary develops in place of the old one which had been cut off. The cut end rounds off, a bud-like mass of new tissue appears there, and the whole, both old and new tissue together, gradually assumes the shape of — the old rudimentary operculum. The greater part of the change from the beginning is, however, accomplished by the growth of — new tissue sind ie very little by the change in form of the old. No further dev elopment takes place. The result is the same no matter what the level of the cut may be. One of the levels at =the j Compensatory Regulation. 59 which cuts were made in my experiments is shown in the accom- panying figure. In none of them did the functional operculum change its character or did any structure other than a rudimen- tary operculum develop in place of the old rudimentary. (Fig. 1QB, a-c.) e. Operations on Both Oper- cutu. When both opercula were cut off it was found that while in some cases there was a reversal, in others two functional opercula were developed, one on each side, while in still others characteristics differing from either of these two combinations were formed. An attempt was made to find the i | cause of the difference in results _ “5 y Yip and with this object in view cuts TT were made at different levels. The difference of level in the cuts in the rudimentary operculum could not be easily controlled, cae | but since in the former cases \ | sacs where only the rudimentary oper- —~ 7 ie culum was removed there was __//i__ | ; LTV no specific influence either on the A ie aoe opposite functional operculum or (aN | on the new regenerating one, it is Fig. 21. supposed that these differences of — Hydroides dianthus (X30). Operations made level have here also no influence on both opercula simultaneously. I, I, I, IV upon the character of the result. epee the regions of functional ee in In every Easel Care was aalbera ay which the cuts of the four groups of experiments bring the cut well down toward mentary operculum between the two dotted the base of the rudimentary ines includes all the cuts on this side. operculum. ‘The different levels on the functional operculum are indicated in Fig. 21. A summary of the results is made in Table VIII. If we neglect the differences in the levels of the cuts in the rudi- mentary operculum and divide those of the functional one into LfT were made. The shaded portion of the rudi- 60 Charles Zeleny. four groups we get a very interesting correlation between the regions and the corresponding results of the operation. In this way we get four fairly well marked groups. Group I consists of accurately located cases near the distal end of the stalk where it expands into the cup. Group II has the earlier cases located from description alone, 1 in most cases stating that the “functional stalk was cut near its middle.” Group III has accurately determined cases located about one-fifth of the length of the opercular stalk from the basal suture. Finally, Group IV contains the cuts made just distal to the basal suture. (Fig. 21.) TasLe VIII. Hydrotdes dianthus. Both Opercula Removed. Groupie ihc o renner Oo 3 fo) fo) fe) Group #0 ones I I 3 I I Group llth ae fo) I 4 fo) fe) Group, EVin772 22:2 fo) 5 fe) ) fo) F,= Original functional operculum. R2= Resultant rudimentary operculum. R,= Oniginal rudimentary operculum. S = Functional stalk remains attached. F.= Resultant functional operculum. r’,r’= Small undifferentiated buds of new tissue. In Group I the three valid cases (see Table VIII) all showed a reversal of the opercula, the old functional becoming the new rudi- mentary and the old rudimentary becoming the new functional. In Group II where the region of the cut was not so accurately located the results are scattering. Seven of the cases give results valid for our purpose. Of these three developed two functional opercula, one showed a reversal similar to that of Group I, one showed the development of a new functional in place of the old and a new rudimentary in place of the old rudimentary, in one the functional stalk did not become detached and the old rudimentary developed into a functional, and in still another case there were rudimentary buds on both sides, neither of which reached a stage beyond a rounded knob and were, therefore, not developed even up to the rudimentary stage proper. | | | | a me Compensatory Regulation. 61 In Group III, consisting of accurately determined levels, four of the five cases showing valid results developed two functional oper- cula, the fifth one showed a reversal of the opercula. In Group IV, also consisting of accurately determined levels, five cases showed a clear result and all of them had a reversal of the opercula. The result is a peculiar one in that the most distal and the most proxtmal groups agree in giving rise to a reversal of the opercula, while the intermediate two groups give rise in a majority of the cases to two functional opercula. An attempt at an explanation is hazardous and can be little more thana guess. Such a provisional attempt may, however, be made, for by so doing some light may be thrown on the regulation of the “normal” condition in the animal. An examination of the whole number of cases where both opercula are cut off shows that in all but two the rudimentary operculum after regeneration did not stop at the rudimentary stage but kept on developing until it reached the functional stage. [he difference in the results is then due to differences in the regeneratory development of the old func- tional operculum. What factor or factors hold it in the rudimen- tary stage in some cases, while in others it is allowed to develop into a full-sized functional organ? ‘Iwo factors are to be con- sidered: first, the influence of the position of the cut upon the initial stages of change in the embryonic tissue in the neighbor- hood of the basal suture and, second, the possibility of a retard- ing influence emanating from the new functional operculum which is rapidly developing on the opposite side from the stump of the old rudimentary. First Factor. It has been shown above (p. 58), in the series of experiments with an uninjured rudimentary operculum, that injury to the extreme distal portion of the functional operculum does not lead to the dropping off of the injured organ or to the development of the opposite rudimentary operculum into a func- tional. Further, from the same series of experiments it 1s seen that the development of the opposite rudimentary operculum is more easily induced by a terminal injury to the old functional than is the dropping off of the old functional stalk. The latter point 1s well illustrated by several cases in which the injured stalk remained attached to the animal while the opposite rudimentary had already developed into a full sized new functional. From these data we 62 Charles Zeleny. may assume that the distal cuts do not affect the embryonic tissue at the basal suture as quickly as do the more proximal cuts and for this reason the tissue will have had only a small start when the opposite rudimentary already has a very considerable one. ‘The latter may then restrict the further development of the bud after the old stalk has fallen off. In Group IV, on the other hand, the cut is so near the embryonic tissue of the suture itself that it may . directly injure the cells which are to give rise to the mechanism of a cleaving plane or else leave such short leverage in the distal por- tion of the stalk as to compel the growing tissue to do all the work in pushing off the useless portion and thus to retard its growth. It is possible also that in some of the cases in Group IV the short por- tion distal to the suture is not cast off, but that the tissues are re-formed and thus give rise to the growing bud. However, no observations were made on this last point and it is merely put down as a possibility in view of some of the later experiments on Pomatoceros (ps 70): In the above paragraph an attempt has been made to show how the embryonic tissues at the basal suture 1n Group I and in Group IV may develop up to the stage of a rudimentary operculum less rapidly than those of Groups II and II]. Second Factor. Admitting this greater development of the bud of the old functional side in Groups II and III than in Groups I and IV, and further assuming the uniform development of the bud of the old rudimentary side in all four groups, we are led to the consideration that if one of the opercula when well developed com- pels the other to stay in a rudimentary stage such a cause may act in Groups I and IV and not in Groups II and III. Therefore, in the former cases, the result of the simultaneous operation on both opercula 1s a reversal of the original condition, while in the latter tt 1s the production of two “functional” opercula. f. Body Cut in Two. Regeneration of Opercula at the Anterior End. When the body is cut in two in the thoracic region two opercula and groups of branchiz are regenerated on the two sides of the median line but the opercula instead of being differen- tiated into a large one and a small one are both of the large func- tional type. We must assume in this case that since both had an equal start in development the retarding influence of the one upon the other which occurs in other cases did not occur here. The newly developed opercula were in some cases exact dupli- Compensatory Regulation. 63 cates, the one of the other, but in others one operculum was considerably larger than the other, though both showed the true “functional’’ characters. “The extremes of the different cases are given in the accompanying figure. (Fig. 22.) The resultant opercula usually differed from the normal func- tional one in being shorter and stubbier than the latter. It is to be noted that two a developed opercula of the kind indicated can a a Fic. 22. Hydroides dianthus. Opercula as regenerated at the anterior end of the posterior piece after trans- verse section in the thoracic region (X 17). Left, case with equal opercula. Right, case with unequal opercula. be of little value in closing up the opening of the tube as each one stands in the way of the other. The animals in the experiments were, however, not kept in their tubes so that the actions under such circumstances were not observed. It seems that the anterior missing segments were not regenerated in any case. Whether such regeneration would be possible under favorable conditions cannot be said. In my specimens bacteria and infusoria developed on the tender regenerating tissues and the growth was retarded and finally stopped. The main point as regards the opercula made out in the group of experiments where the body was cut in two in the thoracic region is this: When the opercular buds have an equal start in development both develop into functional opercula. 64 Charles Zeleny. The relation of the developing opercula and branchiz of each side to the nerve cord of that side is very interesting. ‘This is most noticeable in the regeneration of these organs from a cut near the posterior end of the thorax where the nerve cords are widely separated. In the Serpulide it will be remembered the nerve cords do not come together ventrally as in most Annelids but remain widely separated, forming two latero-ventral trunks. ‘The principal blood vessels, however, do not have this arrangement. The branchial circlets, each with its operculum, regenerating from a cut near the posterior end of the thorax, are always widely sepa- rated and seem to be located in intimate relation with the nerve trunks of the corresponding sides. ‘This fact agrees very well with the data as made out by Morgan (’02) for the regeneration of the head of the earthworm which: showed that the regenerating head always develops in connection with the anterior cut end of the nerve cord. A similar relation has been made out for other forms. A further discussion of this and other cases of nervous control in regeneration is reserved for a future time. The results of a transverse cut in the abdominal region were in every case negative as far as the posterior piece is concerned. Its anterior cut Sie in every case healed over and no regeneration took place. The piece lived for a considerable time but did not show any signs of regenerating tissue. In this connection two other groups of experiments may be described. The first concerns the regeneration and regulation following the longitudinal dorso-ventral division of the body into equal right and left parts. In this group a dorso-ventral longitudinal cut divided the body into approximately equal right and ‘jeft halves. Fourteen speci- mens were operated on in this way and of these several showed traces of the regeneration of knob-like elevations near the anterior end of the cut pare ‘Two of these showed especially clear structures which correspond very well with young regenerating branchial circlets from the anterior end of a posterior piece after @! transverse section of the thorax. It is probable that the new structures may be located at a cut end of a nerve cord. In one of the cases the new circlet in question was a considerable dis- tance behind the old branchial circlet. In no case did the animal live long enough to allow of a full development of the new organs. ot, iM TAD dy ey eee Compensatory Regulation. 65 The effect of the cut upon the old organs is as follows: The rudimentary operculum in several cases showed a slight develop- ment though only one advanced to the stage with bork rows of opercular processes present. Usually the rudimentary operculum remained unchanged or a slight dev elopment of the secondary processes took place. No changes were observed under similar circumstances in the functional operculum, The second group of experiments concerns the regeneration from a posterior cut surface of a half (right or eer Serpulid. Iwo of the cases of regeneration at the posterior cut surface after longitudinal dorso-ventral section of the whole organism showed very clearly the character of the new tail bud regenerating there. After transverse section of the whole body in the abdominal region the new tail bud is always very evidently double. In the present experiments, however, where only half of the animal was used the regenerating bud always showed a single tail knob. As the two components of the ventral nerve-cord in Serpulids and Sabellids are widely separated this singleness of structure may be correlated with the presence of only one of these nerve cords at the posterior end when the animal is cut in two longitudinally before the cross cut is made. g. Progressive Changes in Opercula. Progressive changes in the opercula of adult specimens were not observed. Several groups of specimens were kept under observation for varying periods of time but in no case was evidence of such a change noted. It must, however, be stated that the periods were all rela- tively short, not over a month at most. The indirect arguments in favor of the occurrence of such changes as furnished by speci- mens in nature with intermediate stages of reversal, etc., are given elsewhere. (p. 33.) h. Ex periments on Group V. (See p. 26 for definition.) A series of experiments on H. uncinata was undertaken with the object of determining the relative capacity for regeneration at different levels in the body. The most posterior region showing regeneration of branchial and opercular structures was an anterior cut surface located between the next to the last (sixth) and the last (seventh) thoracic segments. Several posterior pieces back of this point lived for a sufficient length of time to allow of regenera- tion if it were to occur at all, but all these healed up at the cut surface and showed no regeneration of head structures. We 66 Charles Zeleny. D Fic. 23. A, B—Hydroides uncinata. Regenerating branchie and opercula at anterior cut surface after trans- verse section between fourth and fifth thoracic segments (93). 4—13 days after operation. B— 14 days after operation. C—H.uncinata. Regenerating branchie and opercula at anterior cut surfa after transverse section between first and second thoracic segments, 18 days after operation ( X 60). D—Dorsal view of Apomatus ampullifera showing functional and rudimentary opercula ( X 5). Pinnules not represented. (See also Fig. 6 p, £, F). E, F, G—Apomatus ampullifera. Regenerating branchie and opercula at anterior cut surface after transverse section between the third and fourth thoracic segments, 23 days after operation (X 60). E—Dorsal view of both branchial groups. F—Left group as viewed from right side. G—Right group as viewed from left side. Compensatory Regulation. 67 must, therefore, for lack of positive evidence to the contrary decide that in H. uncinata the power to regenerate head _struc- tures is found only in the thoracic region and that anterior surfaces of the posterior pieces after transverse section through the abdomen do not possess this power. The manner in which the regeneration takes place is extremely interesting when compared with the dev elopment of the same structures in ontogeny. It was found that in each of the cases at the earliest stages three branchial buds and one opercular bud were present on each side. (Fig. 23a, B, c.) This corresponds with the number present in the ontogenetic development of H ydrotdes at the first appearance of the opercula. A similar relation holds for the regeneration of the branchial circlets of Apomatus after a transverse cut in the thoracic region. A series of experiments was performed on H. pectinata to deter- mine whether the cutting of the animal in two by a transverse cut through the second and third segments of the thorax would cause any changes in the opercula remaining at the anterior end. In this series the opercula were not disturbed. ‘The result was not completely satisfactory because most of the specimens died at an early stage but it was found that the cut did not cause the opercula to change. A severe bodily injury, therefore, need not cause a reversal. Another series was undertaken in the hope of finding the influence of sectioning of the thorax upon the differentiation of the opercula after the functional stalk had been cut at its middle. It was found that the separation of the region of the body back of the fourth thoracic segment from the rest does not retard the changes of reversal in the opercula which usually take place after a section of the functional stalk at its middle. A single specimen of Serpula vermicularis was operated on. Both opercula were cut off, the functional one at its middle. ‘The result was a reversal of the former condition. “The animal was kept in its dish unobserved for about three months and was then found to have reversed back again to its original condition. 1. Experiments on Group IV. Ann Sea ampullifera. “Iwo characteristics of the branchie and opercula of Apomatus need to be taken into account before going on with a description of the experiments. (See also description, p. 26.) 68 Charles Zeleny. In the first place there are two opercula, one a large spherical body and the other a very small terminal enlargement, each at the end of the branchial stalk occupying the next to the dorsal position in its branchial circlet. This stalk is in each case a typical branchia except for the opercular enlargement and apparently carries on its full respiratory as well as its opercular function. (Bie 2230.) In the second place each branchial semicirclet taken as a whole breaks off very readily along a definite line at its base so that all the branchiz including the opercular one come off together. Thus a very slight irritation is sufficient to cause the animal to throw off the whole branchial apparatus, including the opercula. The fission plane is in a very definite region at the base of the branchial circlet and after coming off the whole branchial crown holds together in one piece because of the union of the branchiz near their bases. “The right and left branchial circlets act independ- ently in the matter since it often happens that only one 1s cast off. Usually, however, both are thrown off. Such an operation as the removal of the animal from its tube usually brings about this autotomy of the branchiz. Out of 42 specimens removed from their tubes on November 6, 1902, thirty lost both branchial circlets, 6 lost one of the circlets and only 6 retained the whole branchial crown. For this reason it was not possible to repeat the ordinary operculum reversal experiments on Apomatus as after such an operation the branchial circlet was cast off. After the casting off of the branchial crowns in these cases a regeneration of two functional opercula usually followed, though one was often larger than the other. Probably correlated with the differentiation of a “breaking joint” at the base of the circlet is the fact that the regeneration of the branchial crown does not show only three branchiz plus the operculum at the first differen- tiation as in Hydroides but at once brings out several branchiz on each side. One of these may show fhe opercular differentiation from the start, while in other cases it dev elops first as a branchia and only later shows the opercular character. The regeneration at the anterior end of a posterior piece after transverse section through the thoracic region showed a less highly differentiated character of the new organ at the start than when the regeneration took place from the * *breaking joint.” A great number of operations were made, but the anal proved very Compensatory Regulation. 69 sensitive to the injury and nearly all the individuals died very early. However, the beginning of the process of regeneration was observed in a few cases. In one of these which had been cut through the third thoracic segment there was, 23 days after the operation, a distinct midication of the young branchial circlets at the anterior end of the posterior piece in the form of three branchial knobs on the left side and jour on the right. (Fig. 23£, F,G.) Of the latter four the ventral one was very small and the next to the dorsal one evidently larger than the others and showing thus early its opercular character. The regenerating tissue appears first as an undifferentiated mound. When dif- ferentiation does occur it takes on the form of three or four knobs in each mound, corresponding evidently with those of Hydroides dianthus after a similar section and reminding one strongly of the first branchial differentiation in the young of the latter species where, as we have seen, each branchial circlet appears first as three processes, one of which divides at its base, forming four in all. The early differentiation of the opercular knob after a thoracic cut in Apomatus as in Hydroides, however, brings in a point of dif- ference as compared with the ontogenetic development. }. Experiments on Group VI. In four specimens of Ditrupa subulata the functional stalk was cut in two just below the terminal cup. The embryonic tissues at the base of the stalk increase in bulk and bulge out, showing an oblique suture. ‘The stalk drops off a few days after the operation and a new operculum develops from its stump. Evidently the increase in the embryonic tissues serves as a mechanical stimulus for the dropping off of the old stalk. See Fig. 244, B All the specimens of Spirorbis Pagenstecheri in which the operculum was removed died. ‘There is, however, evidence that regeneration of the operculum takes place. A considerable number of the specimens just removed from their tubes showed stages of growth of the operculum from a small bud to a large full-sized operculum. Whether the process is a direct physiolog- ical one or is due to injury cannot of course be definitely stated. A periodic replacement may be connected with a possible periodic injury during the breaking out of the embryos from the brood pouch, though evidence is also lacking as to the length of life of the animals. Several experiments on the regeneration of the opercula in 70 Charles Zeleny. Pomatoceros triquetroides were started. that the operculum has a distinct basal suture. It will be remembered In about half of the specimens which were removed from their tubes it was & Fic. 24. A—Ditrupa subulata. Stalk, two days after operation, showing pro- jection of new tissue at one side of basal portion below breaking joint (X20). B—Regenerating operculum of D. subulata, 9 days after operation (X 20). C, D, E—Pomatoceros triquetroides. Stages in regeneration of new operculum from breaking joint level (10). C—Operculum just pulled off. D—3 days after operation. E—8 days after operation. F—Vermilia multivaricosa. Stalk of operculum two days after removal of cup (X16). Note projecting knob of new tissue at side of stalk. found that the oper- culum had _ been thrown off at this basal suture, the distal portion of the operculum remain- ing in the tube. The plane of the fracture is not a straight one but the middle is pointed forward so as to give in a_ dorsal view the form of an inverted A. Figs. 8B, C} 24¢7 pDyeee Three series of op- erations were Car- ried out. In the jirst the operculum was cut in two distal to the basal suture. Here it seems that the part of the operculum above the suture did not drop off but the regeneration took place by a growth from the cut surface. The first sign of this regeneration was a swelling of the terminal region from which three knobs developed, which evidently became the terminal processes of the new oper- culum. ‘The evidence for this change and for the later changes in general is not complete as the later stages were not followed out. S ————— ss eeee™”™—SS"S—-—~—“—~—~—~—~—~—~—~—~—~——”—OO Compensatory Regulation. 71 The interesting general point is that regeneration takes place from the cut surface without a breaking off at the basal suture. When the operculum was pulled off at the time of removal of the animal from its tube the break always took place at the A-shaped suture and the regeneration then naturally followed from this level. In a third set of experiments the animal was cut in two in the thoracic region. All such specimens, however, died before the appearance of regeneratory changes. The operculum was removed in five specimens of Ver- milia multivaricosa. In no case was there any regeneration of the organ. In one individual the cut was made through the narrow portion of the stalk just below the terminal cup. In this case (Fig. 24F), two days after the operation, there was a protruding knob on the median side of the stalk which may represent the beginning of an opercular regeneration, such as that shown in the case of Ditrupa. (Fig. 24.) In the other four specimens the cut was through the cup portion of the operculum. In all of these there was no regeneration, though three of them lived more than eleven days after the operation. k. Discussion of the Data. It has been seen that the char- acter of the regeneratory process varies according to the loca- tion of the cut. When the regeneration takes place from the breaking joint of the operculum (Hydroides, etc.) or of the branchial circlet (Apomatus) the regeneration is highly special- ized and the stages do not follow the ontogenetic ones very closely. When, however, the regeneration is from a thoracic cut, where the branchial and opercular tissues are not as highly special- ized with respect to the mechanism of regeneration, the organs pass through a stage which may very well be compared with a corre- sponding stage in the ontogeny of Hydroides. However, even here the regeneratory development does not follow the other closely because the operculum is very evidently differentiated as such from the start in regeneration though not in ontogeny. Our general conclusion may, therefore, be that when there is no definite mechanism for the autotomy of a region of the body the regenerating tissue may in its various stages resemble ontogenetic Stages quite closely, but where a definite mechanism Is present the resemblances are much less close, the development being hastened in the latter as compared with the former and both being hastened, 72 Charles Zeleny. though in different degrees, as compared with the ontogenetic development. The discussion of the regulation of the process may be referred to the general discussie of compensatory regulation in the group of Serpulids, as given on p. 76. 3. Probable Phylogenetic Development. The opercula and branchiz of the family Serpulidz furnish as good a case of a morphological series as can be found within the animal kingdom. ‘There are all gradations between species with no modification of the branchia up to those with a degree of opercular modification so great that no branchial characters can be made out in the organ. Riathernipres: in the ontogeny of the one form studied (Hydroides), in which there is a high degree of modifi- cation, each of the two opercula passes through a stage in which it is to all appearances a functional branchia. ‘The paleontological evidence, however, is fragmentary. Our only knowledge 1s obtained from the calcareous tubes and it is not always possible to decide whether the animal inhabitant was or was not operculate. Tubes evidently belonging to the genus Spirorbis are, however, found as low down as the upper Silurian. The morphological and ontogenetic evidence leads us to the probable conclusion that the ancestors of the present day opercu- late Serpulids were non-operculate forms and that the opercula arose in the course of phylogeny by the development of enlarge- ments upon the branchiz which served to close the opening of the tube in which the animal lived. Some speculations as to the origin of the asymmetry of the opercula in the Serpulids may be permissible if it is recognized that the course of the probable phylogeny can at present be no more than guessed at. ‘The existence of a morphological series running om forms with no opercular modification of the branchize (Protula) through forms with a terminal enlargement at the end of each branchia (Salmacina), others with two equal opercular knobs one on each side of the median line attached to stalks still retaining respiratory pinnules (Filograna), to still others with a large operculum on one side and a small one on the other (Hydroides, etc.) or with one operculum and that lateral in posi- tion (Ditrupa, etc.) indicates that the early differentiations of the q ; ¢ : § % ‘ j Compensatory Regulation. 7B operculum were symmetrically arranged with respect to the median line. However, the ontogeny of Hydroides shows an asymmetry from the very first appearance of the opercular modification. Furthermore, the fact that this earliest development always occurs on the left side indicates some correlation between the character of the tube and the position of the organ. In Hydroides, how- ever, there is an irregularity in the coiling of the tube from the very start, so that we get no evidence here of such a relation. An examination of several Serpulids brings out the following relation between the adult position of the functional operculum and the character of the coils of the tube. A tabulation of the result is given below: TABLE IX. Genus Funct. Operculum Tube 0 Always tight.........<. Dextral coil. BERN ATIAN © Aiec sc aia ole, > PS SAV SOLELEnajcinrciecoie =.0: sxe) Sinistral coil. LUST PN ayaiyis GLE tb eared, Sa levoran 65 Definite curve. Relation to animal not made out. omatoceros......-..:.. Pinan Ss lelbes dotanis../ = on Irregular. MA ce ee es RAP EAOR LEE: (ers die. d'e, alors Irregular. BEMGTMIGES.. ... cs a 2s PRIPINE OG MIELE «sel cg < 62). E—Operation: Same as D, except that thoracic cut is at second segment. Figure shows the regenerating left branchial circlet days after operation. Note beginning of opercular enlargement. Compensatory Regulation. 79 because there is a very definite line of cleavage along which the break always takes place, and similar materials may be assumed to exist at the regenerating surface at the time of the operation in all individuals of a set of experiments. If, therefore, the body is cut in two at various levels and the Wentichize are thrown off at the “breaking joint” any differences in the regeneration of the branchiz and opercula may be considered as due to differences in the posterior body operations. In the first lot of Apomatus the branchiz were removed in the manner mentioned but there was no operation on the body. In the second lot the body was cut between the third and fourth thoracic segments in addition to the removal of the branchiz, and in a third lot the thoracic cut was made between the first and second segments. ‘he last two operations in a great number of the cases naturally caused the death of the animals, but in general a very interesting result was obtained. In the cases where the body was cut in two in the thorax as mentioned the opercular differentiation appeared much earlier than in those in which the body was intact. When the body 1s uninjured except for the removal of the bran- chial circlets the branchial buds to the number of eight or nine on each side appear simultaneously or nearly so, although there is a slight gradation in size from dorsal to ventral edge of the branchial ridge from the beginning. The few remaining buds to be developed are added from the ventral edge. ‘These bud-like processes increase rapidly in length and soon appear as long slender filiform processes which usually take on the See tee pinnules before the appearance of the first traces of opercular differentiation. ‘The opercular differentiation then appears as a vesicular enlargement in the next to the dorsal branchia on each side. In the two- -segment and four-segment thoracic pieces, left by the thoracic cuts in lots two and three, the development does not follow this course. The opercular bud is from the start very evidently different from the others, being larger and more spherical than the branchial buds proper. This is well shown in the figures. (Figs. 25 and 26.) The branchiz in this case, however, are also thicker and shorter than the corresponding ones where the body remains intact. It is very evident that while in the one case where the body is intact the operculum passes through a distinct knob, in the other case where only the anterior two or four seg- ments remain the operculum appears as such from the start. Unfortunately the animals in the last two lots did not live long enough to show whether or no the final outcome would have been the same in the two cases. In fact none of them showed any pinnules on the branchiz at the time of death. Notwithstanding 8o Charles Zeleny branchial stage before showing even a beginning of an opercular Fic. 26. Apomatus ampullifera. Regenerating branchial circlets. Nine days after operation ( X 40). Operation: Autotomy of both branchial circlets at breaking joint and removal of body posterior to second thoracic segment. Note pronounced opercular differentiation on both sides. these limitations it is evident that we have a definite acceleration of the rate of differentiation of the opercula. Two probable factors may be eee as concerned in the bringing on of the acceleration. The shock of the transverse — division of the body may lead to cach an increase in rapidity of differentiation. 2. The small size of the piece itself may directly | influence the process and bring about such a differentiation. ‘This action may result because of the difference in the interactions of the organs in the one case as compared with the other. | Compensatory Regulation. SI VY. REGULATION OF THE RATE OF GROWTH AND NATURE OF DIF- FERENTIATION DURING REGENERATION OF THE CHELAE OF GELASIMUS AND ALPHEUS. 1. Introduction. The general problem to be taken up in the experiments on the chelz of the two Decapod Crustaceans mentioned corresponds with that already given for the Serpulids. The interactions of the two chelz naturally constitutes the principal point of study. Likewise the influence of the removal of one or both chele upon the rate of moulting of the animals will be discussed and some further incidental points will be touched. In Gelasimus pugilator the two chelz are of nearly the same size and character in the female but differ widely in the male. In Alpheus dentipes the chelz differ both in size and character in both male and female. Dis Gelasimus Pugilator. In Gelasimus the male has one of the two chelae enormously developed. ‘This large chela is nearly equally distributed between right and left sides in a group of individuals taken atrandom. In the female the two chelz are small and equal in size. The animals readily autotomize their legs if a needle is inserted between two of the joints distal to the “breaking joint” so as to touch the nerve. In the following experiments the animals were made to throw off their chelz in the way mentioned. ‘They were kept in glass dishes with just sufficient water to keep them moist and fed with bits of the horse-mussel, Mytilus. Under these conditions they lived very well, though unfortunately the growth of the new legs was extremely slow and the experiments could not be com- pleted as satisfactorily as was wished. 1. Experiments on Males. a. Large Chela Alone Removed. ‘The first object of the work was to determine whether reversal of the character of that of Hydroides takes place in these forms. The large chela was autoto- mized in the manner already indicated. In the great majority of the cases the animals lived through the 62 days after the opera- tion, but in only a few did a moult oie place so that the results are 82 Charles Zeleny. not entirely satisfactory. “lwenty specimens, ten with the large chela on the right side, and ten with it on the left, were treated in this way. Five specimens had moulted at the end of 62 days after the operation when the experiment was closed. The first one moulted 54 days after the operation and in this the regen- erated chela ( = former large one) was as yet smaller than the other ( = former small one). The old small one had no pro- nounced change as a result of the moult. In the four other specimens, one of which moulted 59 and the other three 62 days after the operation, the new regenerated chela was in each case larger than the opposite old one, though it had not as yet attained the full size and characteristics of the typical large one. It may be safely concluded from the above observation that no reversal of the chelz in the sense of the reversal of opercula in Hydroides takes place 1 in the males of Gelasimus after removal of the large chela, for it seems evident that in the first case men- tioned, where the regenerated ‘chela was as yet smaller than the old small one of the « opposite side, it had not yet reached its full growth. In further support of this view is the fact that no pro- nounced change in the old chela was noticed. b. Small Chela Alone Removed. ‘The following results were obtained when the small chela alone was removed: Only four of the ten specimens moulted before the end of the 62 days, con- stituting the limit of the experiments. In all of these the newly regenerated chela were much smaller than the opposite large ones and approached i in character the ordinary small chela. ‘The four specimens mentioned moulted, respectively, 48, 61, 61 and 62 days after the operation. ‘Therefore, here also there is no reversal of the chelz. c. Both Chele Removed. In this set of experiments both chelz were autotomized. ‘Ten specimens were kept for 62 days and eighteen for 42 days. Seven of the former moulted and showed the characters of the regenerated chela. In each of the seven a large chela was regenerated in place of the former large one and a small chela in place of the former small one. ‘There was no reversal. The chelz after the first moult did not of course as yet have the full size of the old ones but the difference in size was very evident. In one of the seven cases mentioned here as having moulted so as to show the characters of the regenerated chelz one specimen Compensatory Regulation. 83 showed an abnormality in that the smaller regenerated chela had two pinchers at its end. This case will be neeereaa in a separate note at another time.! d. Two normal male specimens kept in glass dishes for 62 days did not moult or show any changes. They were fed on fragments of Mytilus in the same manner as the others. 2. Experiments on Females. a. Removal of One Chela (right or left). Two of the six specimens moulted before the completion of the experiment. In one of these the regenerated chela was a trifle smaller than the opposite one. In the other the two chelz, the old and the regener- ated one, were nearly equal in size after the moult. One of these specimens moulted 56 days, the other 62 days after the operation. b. Removal of Both Chele. Only three specimens were oper- ated on in this way. All had moulted within 46 days after the operation. The new animals regenerated two new and equal chele. ‘The regenerated structures, therefore, repeat the char- acter of the removed appendages. 3. The Rate of Regeneration and of Moulting After the Opera- tion. (Male and Female.) The data show very plainly that the moulting takes place sooner in the cases where both of the chelz are removed than in the cases where only one or none are removed. In fact all three mem- bers of the female set with both chela removed moulted before any of those with only one chela removed had done so. It does not seem possible that the matter of accident can come in here as there are too many cases both as regards this point and as regards other related ones. A general comparison in both males and females of the cases in which both chelae were removed with those in which only one or none were removed bring out an interesting result. I. Time of moulting. The specimens meh both chele removed moulted sooner than those with only one chela removed. 2 The regenerating buds in the specimens where both chele had been removed were in general larger than the corresponding buds "Biol. Bull., ’o5. 84 Charles Zeleny. where only one chela was removed. ‘Vhis statement is of course not a definite quantitative one on account of the difficulty in esti- mating relative sizes, but becomes interesting in connection with the following results on Alpheus (p. 85, jf.) and the former cases already described. T he results, therefore, indicate that where two chele are removed the time of moulting 1s hastened and the rate of regeneration of each chela 1s increased as compared with the cases where only one chela ts removed. 4. Rate of Moulting after Removal of One or Both Eyestalks. (Male and Female.) Three sets of experiments were performed. a. In one set of three females the right eyestalk was cut off near its base. Pigment very soon collected near the cut end so that the region Heearnied a color much darker than the normal eye color. A similar change took place in all the other experiments after the cutting of the eyestalk. b. In another set there were three males, one with the large chela on the left and two with it on the right side. The night eye stalk was cut off near its base. ‘The first noticeable change as in the last set was a collection of very dense pigment at the cut surface. All three of the specimens moulted before the close of the experi- ment, one 56 days, another 58 days and the third 61 days after the — operation. At moulting the dark pigment of the cut surface dis- appeared and the end of the new eyestalk was rounded and resembled the old eyestalk except that it was shorter. c. Ina third set of four males and one female both eyestalks | were cut off near their bases. “Two of the specimens lived for more than a few days. One moulted 14 days after the operation | but was found dead and broken after the moult. In the other one the cut surfaces of the eyestalks were black with pigment at this time and also five days later, 19 days after the operation. - This specimen moulted 23 days after the operation, at which time the eyestalk ends were rounded but there was no appreciable increase in eyestalk length. ‘The very dark pigment disappeared — with the moult as in the last series. The animal died 40 days” after the operation. ( d. Conclusion. In the specimens with an eyestalk operation | the time of moulting was hastened as compared with unoperated | Compensatory Regulation. 85 specimens. In the cases where both eyestalks were removed the moults came sooner (14 and 23 days) than when only one was removed (56, 58 and 61 days). Here again, therefore, there is a hastening of the physiological process in the cases of a greater disturbance as compared with those of a lesser feturbonce of the organism. 3. Alpheus Dentipes. 1. Introduction and Review of Former Work. The reversal of the cutting and snapping chelz of Alpheus has been demonstrated and the process studied by Przibram (’o1,’02) and recently by Wilson (’03) and | | mrucs (03, 04). It has been \ shown that when the snapping \ : claw is removed, the cutting chela of the opposite side is differen- tiated into a snapping chela, while in place of the removed snapping chela a new cutting chela is developed. When the cutting chela alone is removed there 1s no reversal, a new cutting chela developing in place of the old. hese \) ~— When both chele are removed VE GIN hts there is again no reversal, a new Cee m snapping chela being regenerated ) in place of the old snapping and La me Chele of Alpheus dentipes (<4). Left-hand anew cutting chela in place of the figure: cutting chela. Right-hand figure: snap- oldcuttingone. Inthe latter case Przibram, workingon A. dentipes and A. platyrrhynchus mentions the fact that the newly regenerated cutting chela is relatively larger as compared with the snapping chela than in the animal before the operation. Wilson’s result on the Beaufort, N. C., species (H. heterochelis) agrees with that of Przibram except that in the case where both diel were removed the new cutting chela does not approach the new snapping chela as nearly as in the species upon which Przibram worked. Wilson further added the interesting result that when the snapping chela is removed and the nerve leading to the cutting chela is cut below the breaking joint ping chela. Dotted lines represent the lengths measured (see text, p. 91). 86 Charles Zeleny. Post-spin us thoracic length in millimeters. 3:0 = 40 5.0 6.0 7.0 8.0 90 10.0 11.0 ea Bae ais a > s =] 4 = Ss o 5 33 N rs 3 ~_ @ bs | =| ° ° = 2 ° 2 = 3 > Fe ° — =] L om] 5 8 —— zo e+ — ves aay ei ia ream eee a aia al JN) o to o % ae a ne R -! as, () G Time of moulting in days after operation, 3.0 4.0 — 5.0 6.0 7.0 8.0 9.0 10:0. ea Post-spinous thoracic length in millimeters. Fic. 28. Alpheus dentipes. Lower data—Time of moult in days after operation. Upper data—Interv: between first and second moult in days. Abscisse are post-spinous thoracic lengths. © = Cutting chela (Cu) alone removed. © = Snapping chela (Sn) alone removed. = Both chele (Cu+ Sn removed) and the lower line to fit the last group (both chele removed). Compensatory Regulation. 87 “the reversal in some cases at least does not take place or is incom- plete.” Brues (Wilson, ’03, p. 210) adds the interesting fact that in A. heterochelis the nerves supplying the two chelz and the ganglionic centers from which they proceed do not differ percep- tibly in size. 2 othe Waa: The experiments about to be described in the present section of the paper were performed at the Naples Zodlogical Station in the winter of 1902-03. ‘They confirm the general facts of reversal of the chelze as given above. ‘Their main object, however, was the determination first, of the effect of the removal of one or both chelz upon the rate of moulting of the animal, and, second, of the influence of the presence or absence of the opposite chela upon the rate of regeneration of a chela. a. The Influence of the Removal of One or Both Chele upon the Rate of Moulting of the Animal. ‘Three sets of specimens were operated on. In one set (Sn) the snapping chela alone was removed. Ina second (Cw) the cutting chela alone was removed. Inathird (Sn + Cu) both snapping and cutting chelz were removed. The animals were kept in isolated dishes for 59 days after the operation and were fed either every day or every other day on small pieces of fresh fish meat. W ithout taking into account the cases where the legs were accidentally nutoronized a second time during the experiment, or in which other disturbances occurred, we have the following relation between the time of moulting and the post-spinous thoracic length of the animals without reference to the character of the operation: On the codrdinate paper (Fig. 28, p. 86) the abscisse represent the thoracic lengths in millimeters and the vertical columns (ordinates) the days ier the operation when moulting occurred. The animals were killed 59 days after the operation. The data from the first set (Cu) are represented by the symbol ©, those of the second set (Sm) by the symbol © and of the third set (Sn + Cu) by x. It will be seen that in general the moulting interval increases with the size of the animal as repre- sented by the thoracic length. On pp. 88 and 89 the data are put in Tables X, XI and XII, each of the sets being placed by itself. The interval of time in days between the frst moult and the second moult is put down in a separate column. Upon averaging this interval in the three 88 Charles Zeleny. sets separately it.is found that the interval decreases from 29.6 days jor the Cu set and 28.7 days for the Sn set to 22.9 days for the Sn + Cu set. This result is represented on codrdinate paper on p. 86, Fig. 28. When two chelz are removed there is, therefore, a shortening of the period between the first two moults as compared with the cases where only one chela is removed. It will be seen that there is only a slight difference between the moult period in the two single chela cases. This result agrees perfectly with that obtained. for the time of appearance of the first moult in Gelasimus, where TaBLeE X. Alpheus dentipes. Time of Moulting. Cutting Chela (R or L) Removed. i d Go Nes Thoracic Date of Ist 2d 3 Interval Length. Operation. moult. moult. moult. —_ Ist—2d. 1903 | 562 6.6 1/8 5 24 — 19 565 Oey, 1/8 2 29 52 27 569 8.0 1/8 20 = = 39+ 573 8.5 I/9 22 48 — 26 576 6.5 I/9 16 48 — 2 579 eu I/9 I 26 56 25 582 77 I/9 20 == *1/7 4 20 36 59 16 - a I No N ‘oO Summary or Tasies X, XI, XII. Comparison of Interval Between First and Second Moults. Cu Removed =29.6days. (Av. of 7 cases.) Sn Removed = 28.7 days. (Av. of 8 cases.) Cu+Sn Removed = 22.9 days. (Av. of 10 cases.) go Charles Zeleny. Original cutting chela length in millimeters. 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 SPSied ae eh ae a 7 oles aoe ea d += 6.0 6.0 E a (o} | : See = 5.0 2 4.0 4.0 : : Rie = 3.0 3.0 =) 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 o a g 3.0 4.0 5.0 10.0 o Die fee CCC a 4 : || | | aaa oO 3 7.0 7.0 : | 7a oO i=] Oo Ey fan} 6.0 6.0 Second moult. on o 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 Original cutting chela length in millimeters. Fic. 29. Alpheus dentipes. Lengths of regenerating cutting chele at the end of the first and second moults. Ordinates = Regenerated cutting chela lengths in mm. Abscisse = Original cutting chela lengths. = Cutting chela (Cu) alone removed. © = Snapping chela (Sn) alone removed. > = Both chele (Cu+ Sn) removed. The broken lines fit the last group (both chele removed) and the unbroken lines fit the first two groups (one chela alone removed). Compensatory Regulation. QI b. The Influence of the Presence or Absence of the Opposite Chela upon the Rate of Regeneration of a Removed Chela. A comparison was made of the regenerated lengths of the cutting chela in cases where only one chela was removed with cases where both chele were thrown off. When the cutting chela alone was autotomized a new cutting chela was regenerated in its place. When the snapping Bho alone was removed there was a reversal and the old cutting chela was differentiated into the new large snapping chela, while in place of the removed snapping chela a typical cutting chela was developed. In the second case this new cutting chela is the one taken in our measurements. In a third case both chelz were thrown off, a new cutting chela regenerating in place of the old cutting chela and a new snapping chela in place of the old snap- ping one. The lengths are taken from the moulted casts of the animal, the original length being taken from the cast of the first moult, the first moult condition from the cast of the second moult, etc. The final measurements are taken from the alcoholic speci- mens of the animals killed 57 days in each case after the operation. The lengths measured in the data about to be described are the greatest lengths of the cutting chela, 7, ¢., the distances from the tip of the pincher process of the fourth podomere to the farthest corer of the base. (Fig. 27.) The chelz or their casts were in every case drawn carefully to scale by the aid of a camera lucida and the measurements are taken from these drawings. Out of 29 specimens kept for 59 days only 12 can serve for first moult data and 13 for second moult data. ‘The others are not valid because of the death or escape of the animal or the accidental secondary autotomy of one or both of its appendages. The relation of the regenerated chela lengths to the original lengths is shown on coérdinate paper (Fig. 29) for both the first and the second moult. The number of individual cases is small, but it is evident that the regenerated lengths of the cutting chela in the cases where both phe were reniaved have a distinct advan- tage over the others. This is especially clear for the first moult. The relation comes out very clearly when we take the ratio between the regenerated cutting chela length and the original length in that specimen as our basis for comparison, for we see that the regenerated length increases as we go from sm: ll original lengths to large original lengths. As the cases given on the coor- g2 Charles Zeleny. dinate paper (p. 90, Fig. 29) show, the correlation is not a perfect one, 1. é., it is positive but equal to slightly less than one. The results of such a comparison are given for both the first and the second moults in Tables XIJI-XVI. TasLe XIII. Alpheus dentipes. Length of Regenerated Cutting Chela. Cutting Chela Removed. Cat. Original Reg. Cu. Lg. Regen. Cu. Reg. Cu. Lg. Reg. Lg. r eno See No. | Cu. Lg. Ist moult. Orig. Cu. 2d moult. Orig. Lg. 573| 7-0 3:8 54-3 4.9 70.0 578 «9.8 6.3 64.3 7-7 78.6 a ane = 59-3 a 74-3 TasB_e XIV. Alpheus dentipes. Length of Regenerated Cutting Chela. Snapping Chela Removed. | | Cat. Original Reg. Cu. Lg. Regen. Cu a Reg. Cu. Lg. Reg. Lg. we No. | Cu: Lg: Ist moult. | Orig. Cu. _ 2dmoult. Orig. Lg. 561 8.8 6.7 76.1 | 7.7 87.8 563 Qist 6.0 65.9 | 6.9 75.8 575 9.7 6.3 64.9 6.7 69.1 578 5-9 a) 57-6 4.0 | 67.8 == == = 66.1 | = | fie The tables show that Cu + Sn has a very distinct advantage over Cu alone or Sn alone. This advantage amounts to 19.1 per cent for the first moult and 16.5 per cent for the second moult. Just after the first moult the cutting chela regenerated from the breaking joint surface of what was formerly the snapping claw has a distinct advantage over the cutting chela regenerated from a removed cutting chela but this advantage is nearly overcome ~ at the time of the second moult. Compensatory Regulation. 93 TaBLeE XV. Alpheus dentipes. Length of Regenerated Cutting Chela. Both Chelea Removed. at. Original | Reg. Cu. Lg Beeeube ee Reg. Cu. Lg Reg Lg. eres No. Cu.Lg. | Istmoult. Orig. Lg. 2d moult. Orig. Lg. ee 5-7 | 3-7 64.9 ) 4.2 Be 567 Whee 6.4 85.3 6.9 92.0 574 79 Gi, F252 6.4 81.0 wae or | 3-9 68.4 4.5 78.9 580 | 3.2 = = BUA 106.2 Gog) 4.6 a0 78 .3 4.0 86.9 584 5-0 |. 3.9 78-0 4.5 go.o oo. |... |. = = | — 74-5 = 87.0 TasLe XVI. Summary of Tables XIII, XIV and XV. First Moult. | Second Moult. JI Original data. peeenetated Ig. Ce | Cu ue 7 3 Original Lg. Sn I Te) Cu + Sn 725 87.0 Comparisons. (Cu + Sn) — Cu +15.2 eee, Absolute difference ....... Jo) (Cul Sn) = Sn + 8.4 | +11.9 {| Sn=Cu + 6.8 | + .8 |e te Sy eee toe 6 sel | Cu Per cent of increase ...... (Cu + Sn) — Sn +12.7 fo Sie Sn Sn — Cu Cu 94 Charles Zeleny. The general result is clear. The regenerated cutting chela is larger in the case where two chelz are regenerating than where one alone is to be replaced. ‘This result is emphasized by the fact that the time between the first two moults is shorter in the specimens with two chele removed than in those with only one gone. (see p. 87.) 3. Discussion. The significance of the data may be emphasized by bringing them out in two ways, one of which lays special stress upon the fact of removal of a certain organ or organs and the necessity of a certain amount of regulation in restoring the normal form, and the other of which emphasizes rather the interactions of the two chelz as parts of a system normally stable at the condition with a large snapping chela on one side and a small cutting chela on the other. The first point of view in which the total necessary amount of regulation is compared with the rate of regeneration of a part of the whole will frst be taken up. The three series of experiments may be compared in the following way: 1. With the cutting chela alone gone the animal has merely to accomplish the regeneration of this organ in order to regain its normal condition. 2. With the snapping chela alone gone the animal has not only to regenerate a cutting chela in place of the old snapping one but also to differentiate the tissue of the old cutting chela into the new snapping chela. 3.. With both chelz gone the animal has not only to regenerate a new cutting chela in place of the old one but also to regenerate a new snapping chela. Taking the ratios given in Table XVI, p. 93, we see that in the first case where the least work is to be accomplished, at the end of the first moult the cutting chela has regained but 59.3 per cent of its original size, while in the second case it has reached 66.1 per cent, and in the third case 74.5 per cent of its original size. At the end of the second moult likewise we have, respectively, 74.3 75.1 and 87.0 per cent for the three cases in question. Therefore, the amount of actual regeneration accomplished in the cutting chela is greater the greater the amount of other work of a similar character to be accomplished at the same time. Compensatory Regulation. 95 But if we consider the matter from the second point of view, namely, of the influence of the presence of another similar and opposite organ upon the regenerating tissue, the apparent anomaly of the case is cleared up to a great extent. For it is seen that in the first case (Cu alone remov red) we have an uninjured opposite large snapping chela to retard the growth of the regenerating cutting chela. In the second case (Sin alone remov ed) we have as the retarding agent at the beginning merely the smaller Cu chela which, however, gradually “undergoes the changes in size and character leading up to the large snapping chela. Finally, in the third case (Cu + Sn removed) we have at the beginning no retarding agent, though gradually the snapping chela develops from this point. Expressing this in concise form we have the following relation referring to the retarding agent as indicated by the size and com- plexity of differentiation of the chela situated on the side opposite to the developing cutting chela: The snapping chela in the group where the cutting chela alone is removed is greater than the cutting chela developing into a snap- ping chela, as in the group where the snapping chela alone 1s removed and this in turn is greater than the retarding agent where both chelz are removed, and which amounts to zero at the beginning with a gradual Aeadleprn ie up to a snapping chela condition. And referring to the corresponding regenerated lengths of the cutting chela we have: The amount of regeneration of the cutting chela in the first group (Cu alone removed) is less than the amount of regeneration of the cutting chela in the second group (Sn alone removed), and this in turn is less than the amount of regeneration of the cutting chela in the third group where both aS are removed. In graphic form this may be represented as follows: Snapping chela [as in Cu group] > Cutting chela (— snapping chela) [as in Sn group] > Zero (— snapping chela) [as in (Cu + Sn) group]. Correspondingly, Amount of regeneration of cutting chela in Cu group < Amount of regeneration of cutting chela in Sn group < Amount of regeneration of cutting chela in Cu + Sn group. 96 Charles Zeleny. Evidently the differences between the retarding influences of the three members are greatest near the beginning of the experiments, 1. é., immediately after the operation, and gradually decrease as we go away from this point. Correspondingly, the results show a greater comparative difference in regenerated material at the end of the first moult than at the end of che second moult. This result agrees very well with the experiments on the fiddler- crab Gelasimus (p. 81) and on the brittle-star Ophioglypha, (P- 7)- s moulting involves not only an increase in bulk of the animal but also a complicated degree of differentiation of materials before it can be accomplished, we may likewise compare the acceleration of moulting in Alpheus and Gelasimus with the acceleration of the rate of differentiation of the opercula in Apomatus when the pos- terior region of the body is also removed as compared with the cases ines this region is uninjured. GENERAL DIscusSION. The following discussion does not serve as a summary of the data of the preceding sections. For this the reader must be referred to the summaries of the individual sections which are complete entities in themselves. It is the writer’s purpose in the general discussion to show the manner in which the various data, at first sight seeming to have little in common, can be brought under a common point of view. In the introduction it was stated that the standpoint of the present paper would be the consideration of the organism as a system made up of mutually interacting parts, the relations of which were to be studied by noting the disturbances produced as a result of the removal of one or more of the parts. In the paper on the dimensional relations of the members of compound leaves the relations of the parts of a system were studied in which the removal of one member was not followed by its regeneration but resulted in changes in size and position of the remainder. ‘The chief reactions were the following: In the five- leaved forms in which an asymmetrically placed leaflet was removed the other four leaflets tended to rotate to a position such that the new system was a symmetrical four-leaved one. Like- Compensatory Regulation. Q7 wise in the three-leaved system after removal of one of the asym- metrically placed leaflets the two remaining leaflets tended to take up a position so as to form a symmetrical two-leaved system. From the position reactions it 1s evident that the parts of the normal compound leaf are exerting a continual influence upon each other which when resolved into its resultants gives rise to a configuration very definite for a given species. The removal of one of the parts changes the whole system of reactions, and we have a tendency toward the formation of a new stable symmetrical system, with one less leaflet than the original number, the completeness of the new symmetry being only imited by the rigidity of the leaflets. In Ophioglypha we have a radial system in which the removed arms are regenerated. ‘lhe experiments on the rate of regenera- tion bring out the presence of an unsuspected interaction between the arms which must naturally be correlated with some interac- tion present in the perfect, unmutilated animal. ‘The data of the experiments show that (leaving out of consideration the cases where all five arms are removed and which cannot be used because of the early death of the animals) the rate of regeneration of an arm is greater the greater the number of other arms removed at the same time. ‘This indicates an interesting interaction of the arms upon each other for the presence of unremoved arms seems to retard the rate at which the removed ones are regenerated, for it is not probable that the increase in rate in the one case is due entirely to the increase in stimulus to regeneration produced by the added injuries. The two members of a pair of appendages 1 in bilateral animals have been shown by the present experiments to have a profound influence upon each other. Inthose Serpulids, for example, which have one large functional and one small rudimentary operculum it has been shown that either organ originally has the potentiality of developing into a functional operculum, which is to be developed in this way, depending upon the matter of an early start. When one side gets a start over the other the development of the latter is Poernicted to a rudimentary stage, while the former develops to a full functional size. Also, when the functional operculum is removed its restricting influence being removed at the same time, the rudimentary operculum immediately develops into a functional one, which in turn restricts the developing new bud of the other side. When both develop at 98 Charles Zeleny. the same time, as from the anterior cut surface of the thorax, two functional opercula are formed. Similarly, in the Decapod Crustaceans the two chele# have a profound influence upon each other. In Alpheus there are two chelz, one a larger “snapping” chela and the other a smaller “cutting”? chela. The snapping chela seems to hold the cutting chela in check, for as soon as the former is thrown off the cutting chela changes over to the snapping chela by a qualitative and quantitative change combined. The new organ regenerating in place of the old snapping chela comes into a system no longer relatively like the old, so that the inter- action of parts forces it into a different niche in the new order of things. ‘The reversal, here as in the Serpulids, is easily under- stood in the way mentioned, if we consider the systems as asym- metrical interacting systems such that the removal of one part can lead only to the development of a certain definite structure. The removal of the organ (functional operculum or snapping chela) brings about an instability in the system which because of the reactions between the parts tends to assume the condition of a new stable system. ‘This new system reacts now in a different way on the regenerating organ, causing it to develop into a different structure. The readjustment in the old material and in the regenerating material is further complicated by the fact that both processes go on at the same time, the final outcome being the resultant of both. In Alpheus as also in Gelasimus we have an interesting relation between the two chelz, in that when both are removed the rate of regeneration is greater in each than when one alone is removed. Evidently this comes under the Ophioglypha relation that the presence of an unremoved organ retards the rate of regeneration of a removed one. Likewise if we consider the cutting chela of Alpheus as a stage in the development of the snapping chela (Wilson, ’03) and the rudimentary operculum as a stage in the development of the functional, we can say that the presence of the larger organ retards the differentiation of the smaller one. This comes into relation with the series of experiments on the rate of differentiation of the regenerating opercula in Apomatus, in which it was found that the absence of the posterior region of the body back of the second or fourth thoracic segment accelerates the rate of differentiation of the regenerating opercula. ort ie li Compensatory Regulation. 99 From the point of view of the retarding influence exerted by one organ upon another the data that have been brought out in the present paper may be collected in the following concise form: 1. Opbhioglypha. Arms a, }, c,d, e. ai = Rate of regeneration of an arm when it alone is cut off. a2 = Rate of regeneration of an arm when two arms are cut off, etc. Ta, = Retardation as result of influence of remaining arms on a, etc. Then Qs = da + Tag = 43 +1az = G2 + Tag = Ai + Tay arms remaining = none, e, dande, c,dande, b,c,dande. Imtheabove a> a > a > @ > a and (oy << Tas << Tas << Tao

) co meets XAT Oo | 5 5. fo) 95 | 74| 169 | all fo) 47S 319 4 Zot xe) Gomi s5 1) ter | all | o Metal 2c... : 56 56 = 864 | 753 | 1617 | all fe) From a comparison of Tables II and III it follows, therefore, that here the differentiating color characters S and B follow different lines of behavior in heredity, B breeding pure in second generation from random parentage, and S continuing to breed some pure and some mixed broods in the second generation, fol- lowing Mendelian expectations for recessives and dominants. Experiment 3. In order to determine how the color types were inherited in the third generation (first generation bred from 126 Isabel McCracken. similars of random selection, second bred from similars of pure selection) certain broods were chosen from those designated in Table Il, for B x B crossings, and in Table III, from broods true to parents for S x S crossings. Males of one brood were allowed to breed freely with females of another brood as in previous experiments. Great mortality prevailed at this time, materially affecting the number of S broods available. ‘Tables IV and V show results in the aggregate. Table IV shows no reappearance of S in the fourth generation from B x B. TaBLE V—S IL@IRTAL ce OB See eel oe Epc Cet RRS ane ee eee 200 Pam Nenta eb latexangnsense OLpansy, 5 aisle in afin taincieies efeveus Sessions «eles ae asyes 203 PEP NOLO CUOLG Spt beeps cers PN ale wis Be nm Se eS Nera e a clee ee shee Ries teens 204 faa isles sarc lesenchyme rr aptersietetsis) c= Oss oie ee lteys os Sivele «iF Fieve) s,> sietecin a) acetaseverse 205 BY. ermlation.an) Ascidian Ege and Embryo ......-.--+-.-----s+:+-+---- Lees ae ct Ae 206 Sanargll Cameingtins 2a sascdieeaacmeic cries 6 Ure Seater tae eraser et ne ire Ree earn ee 209 [ UDihtael DOGS Sqlsin nas 5 ooo Se oS appa! 6 DBRAD Ean Ser Ble Ad bene e hoe eee a arise 209 Pale abizd ONO OGplasmic SMbStANGES a2. ar 9- ee cielo) oleic incate eis wae ee ve byniale een cee wes 210 i. Cleavane final Iberenllris@ll 5 Bannon poE od BOOSnne SOS an OF Sees Ose SA Onoee enna acon erar Big 4. Determinate and Indeterminate Cleavage and Development ..................-... 214 OTLEY 15228600708 poghHotodeBecros 6 06facciGh COCR BE Ea Og tia OH anno Anninn eae ici 216 221 larvee the substance of the unsegmented egg is ee, undiffer- entiated and the cleavage cells are so nearly equal and homo- feneous that it has not been possible to trace the lineage of individ- 146 Edwin G. Conklin. ual blastomeres throughout the development. The most notable exception to this rule is found in the case of ascidians. ‘That the cleavage of the egg in these animals is constant in form and differ- ential in character and that specific blastomeres are destined in the course of normal development to give rise to specific parts of the larva has been demonstrated by Van Beneden and Julin, Chabry, Castle, and many others. Chabry (’87) also showed, in one of the earliest experimental investigations dealing with the potency of cleavage cells, that individual blastomeres of Ascidia aspersa always develop into those parts of the larva which they would produce under normal conditions. On the other hand, Driesch (95) discovered, some eight years later, that in Phallus mammilata individual blastomeres up to the 4-cell stage at least are capable of giving rise to entire larve and this conclusion was afterward confirmed by Crampton (’97) in the case of Molgula manhattensis. Since the results of Chabry were thus flatly con- tradicted by these later investigators and as they have been de- fended by no one who has actually experimented on these eggs* these results have been generally discredited and the ascidians are now commonly regarded as belonging to that group of animals in which the early cleavage cells are equipotential. ‘The ascidians, therefore, should afford an excellent opportunity of determining the exact method by which an egg fragment or isolated blasto- mere gives rise to an entire larva, since in this case it is possible to follow the lineage of individual cells until they enter into larval organs; furthermore, they should afford means of testing the justice of the distinction which has been proposed (Conklin, ’97) between determinate and indeterminate types of cleavage, and finally they should throw light upon the significance of the high degree of differentiation which is known to exist in the early development of these animals. I. NORMAL DEVELOPMENT. I have recently (’05') shown that these differentiations of the ascidian egg are much greater than has heretofore been supposed} in the unsegmented egg of Cynthia (Styela) partita at least five distinct kinds of odplasm can be recognized. ‘These are, (1) the ‘Several persons, viz: O. Hertwig (’92), Roux (’92), Weismann (92), Barfurth (°93) have discussed Chabry’s work from a critical point of view. Mosaic Development in Ascidian Eggs. 147 deep yellow protoplasm which later enters into the muscle cells of the tail of the larva; (2) the light yellow material which becomes mesenchyme; (3) the light gray material which forms the chorda and neural plate; (4) the slate gray substance which becomes endoderm, and (5) the clear transparent protoplasm which gives rise to the general ectoderm. All of these substances are recog- nizable in the egg before the first cleavage and immediately after that cleavage they all occupy their definitive positions in the ege, the yellow protoplasm forming a yellow crescent around the pos- terior side of the egg just dorsal to the equator, the light gray substance forming a gray crescent around the anterior border of the egg, the slate gray substance lying at the middle of the dorsal hemisphere and between the two crescents, while the transparent protoplasm is chiefly localized in the ventral hemisphere of the egg. In these positions and from these substances the organs and germinal layers specified arise. At the first cleavage of the egg all of these substances and areas are equally divided, since this cleavage lies in the plane of bilateral symmetry of the egg and future embryo. ‘The second cleavage plane is perpendicular to the first and separates the gray crescent in front from the yellow crescent behind; the cells of the anterior quadrants are therefore very unlike the posterior ones and the two can always be distinguished ata glance. (Fig.1.) ‘The third cleavage is equatorial and separates four clear ventral cells from four dorsal ones which contain the yellow and gray crescents and Bie! deep gray material. (Fig.2.) The ectoplasm i is now com- pletely segregated in the four ventral cells but the other ooplasmic substances are not as yet located in separate cells, though from the time of the first cleavage onward their locations and boundaries are perfectly sharp and distinct. At the fourth cleavage each of the eight cells divides, thus giving Tise to sixteen cells (Fig. 3) and at the fifth cleavage these are increased to thirty-two. During the fifth cleavage the substance of the gray crescent is segregated into four cells (A*’, A®*, Fig. 4)* at the anterior border of the egg, while the yellow crescent comes 1The system of cell nomenclature employed in this paper is similar to that used by Castle (96) and is fully explained in my work on the cell-lineage (05!); in brief A and a designate cells of the anterior half of the egg, B and b those of the posterior half, the capitals being used for cells of the vegetal (dorsal) hemisphere, the lower case for those of the animal (ventral) hemisphere. Corresponding cells of the | right and left sides receive the same designation, except that those of the right side are underscored. 148 Edwin G. Conklin. Norma DEVELOPMENT OF CYNTHIA PARTITA, 4-CELL TO 64-CELL STAGES; X 333. The yellow crescent which surrounds the posterior half of the egg dorsal to the equator is stippled. The gray crescent around the anterior border of the egg is left unshaded. The boundary between the clear protoplasm and the yolk is indicated by a crenated line. The polar bodies (shaded by vertical lines) lie at the animal or ectodermal pole. Fig. 1. Four-cell stage from the animal pole, the yellow crescent showing through the egg. Fig.2. Telophase of the third cleavage (8-cell stage), from the left side. Fig. 3. Twenty-cell stage from the animal (ventral) pole. Fig. 4. Twenty cells, transitional to the 24-cell stage, from the vegetal (dorsal) pole. The gray crescent is now segregated in the two pairs of cells A®?, A®4; the yellow crescent will be localized in separate cells at the close of the division which has already begun in the cells B*.1. Figs. 5 and 6. Ventral and dorsal views of the same egg in the 64-cell stage. The yellow and the gray crescents each consist of a double arc of cells; the anterior arc of the gray crescent (A?-4, A7.8) is composed of neural plate cells, the posterior arc (A7-3, A7-7), of chorda cells; only two pairs of cells in the yellow crescent (B7-4, B7-8) are muscle cells, the others are mesenchyme. The pair of cells A7.6 also gives rise to mesenchyme. All the other cells of the dorsal hemisphere (Fig. 6) are endodermal. All the cells shown in Fig 5, except those of the yellow and gray crescents, are ectodermal. Mosaic Development in Ascidian Eggs. 149 150 Edwin G. Conklin. to occupy six cells (B®, B°*, B®) around the posterior border (the spindles which lead to the formation of these six cells are indicated in Fig. 4). These thirty-two cells are increased to sixty-four at the next cleavage (Figs. 5 and 6); during this cleavage four chorda cells ee A™7) are separated from the four neural plate cells (A74, A*’, Fig. 6), while the six cells of the yellow crescent have given rise to twelve, four of which are muscle cells (B74, B7*) and eight mesenchyme (B7*, B77, B7*, B7°).- At the same time am additional pair of secene a me cells (A7*) is separated from a pair of endoderm cells in the anterior quadrants. ‘This is the only mesenchyme cell derived from the anterior quadrants. At this stage all the substances of the germ layers and of the principal organs of the larva are gathered into separate cells, but although this segregation into separate cells comes relatively late in the cleavage these substances have been definitely localized in certain regions of the egg from the time of the first cleavage. Subsequent cleavages lead to changes in the shape of the embryo but produce no changes in this localization. In the gastrulation the endoderm cells are depressed and are overgrown in front by the chorda cells and these in turn are covered by the neural plate cells; similarly the mesenchyme cells overgrow the endoderm at the posterior border of the blastopore, while the mesenchyme cells are overgrown by the muscle cells, and finally the latter by the ectoderm. (Figs. 7-10.) Inthe closure of the blastopore the anterior (dorsal) lip grows posteriorly until it covers most of the dorsal face, while the muscle cells form the lateral boundaries of the blastopore. (Figs. 9, 10.) In this over- growth of the dorsal lip the chorda cells which originally lay at the anterior border of the egg are carried back into the posterior half of the embryo, where by interdigitation they form the chorda. The neural plate cells are also carried back with the chorda nearly to the posterior end of the embryo. The ventral (posterior) lip of the blastopore then grows forward over the remnant of the blastopore and the neural plate is rolled up into a tube which closes from behind forward. The muscle cells become arranged in three rows on each side of the chorda; in front of the muscle cells is a mass of small mesenchyme cells, while a double row of endoderm cells ventral to the chorda constitutes the cord of ventral or caudal endoderm. (Figs. 11 and 12.) Finally the tail of the larva elongates greatly and becomes coiled around the body of the Mosaic Development in Ascidian Eggs. 151 larva within the egg membranes, and about twelve hours after the fertilization of the egg the larva may hatch and become free swimming. However, in a considerable proportion of cases the larva never hatches but undergoes its metamorphosis within the egg membranes. Il. OBJECTS AND METHODS OF EXPERIMENT. ‘This brief review of the normal development! shows that there is a remarkable degree of differentiation and localization of the substances of the egg and embryoand it seems to render necessary some further explanation of the results of the experiments of Driesch and Crampton; certain it is that the egg is highly differ- entiated and if portions of this differentiated ooplasm may give rise to portions of the larva which they would never produce under normal conditions it is important to know the steps by which this is accomplished. With this object in view I spent the summer of 1904 at the Ma- tine Biological Laboratory at Woods Hole, Mass., experimenting on the eges of Cynthia (Styela) partita and of Molgula man- hattensis; | was unable to obtain Ciona intestinalis, the normal development of which I had studied during the previous summer, and my experimental work is therefore limited to the two species first named. Most of my work was done on the egg of Cynthia, which 1s a better object for experimental work chert Pena of Mol- gula, owing to its greater size and the more brilliant coloring of its different odplasmic substances. Enough work was done on Mol- gula, however, to show that the dev elopment of isolated blasto- meres is the same in this genus as in Cynthia. All the experiments performed had for their purpose the testing of the potencies of the various substances and blastomeres of the egg. Injuries to the unsegmented egg of whatever nature, whether produced by sticking, cutting or shaking the eggs, invariably inhibited all father development. II Baas therefore been unable to test the dev elopmental ere of the different kinds of odplasm of the unsegmented eg But inasmuch as these substances are the same in appearance ae localization before and 1For a more detailed account of the normal development of these ascidians the reader is referred to my previous papers on the “Organization and Cell-Lineage of the Ascidian Egg” (05"), and on “Organ-Forming Substances in the Eggs of Ascidians” (7os*). Loe Edwin G. Contin NorMAL DEVELOPMENT OF CYNTHIA PARTITA, GASTRULA TO TADPOLE; X 333- The neural plate or tube is finely stippled, the chorda coarsely stippled; muscle cells are shaded by vertical lines, mesenchyme by transverse lines. Figs. 7 and 8. Ventral and dorsal views of a gastrula (180-cell stage), showing T-shaped blastopore, neural and chorda plates, mesenchyme and muscle cells. Most of the cleavage cells are in the ninth generation. Figs. 9 and 10. Two views of the same gastrula from the dorsal pole; Fig. 9, showing the super- ficial cells, Fig. 10, those at a deeper level. The overgrowth of the dorsal lip of the blastopore and the approximation of the muscle cells of each side toward the median plane have reduced the blastopore — to a longitudinal groove in the posterior half of the embryo. The ectoderm cells are in the tenth genera- tion and there are in the entire embryo about 360 cells. Fig. 11. Dorsal view of an embryo in which the neural plate (n. p.) is closing to form the neural tube (n.t.) Beneath the nerve tube is the notochord and on each side of the latter is shown a row of muscle cells (ms.) At the posterior end of the muscle rows is the caudal mesenchyme, at their anterior end the trunk mesenchyme (m’ch.) Fig. 12. Young tadpole viewed from the left side, showing three rows of large muscle cells (ms.) — along the side of the notochord (ch.); dorsal to the latter is the nerve tube (n. t.); anterior to the muscle rows is the trunk mesenchyme (m‘ch.); ventral to them is the ventral or caudal endodem (v. end.) 154 Edwin G. Conklin. after cleavage begins it can scarcely be doubted that their poten- cies are also the same. Hundreds of experiments involving many thousands of eggs were made upon the various cleavage stages. Lhe methods of experimenting which I employed were essentially like those used by Driesch and Crampton, viz: the eggs in the 2-cell, 4-cell, 8-cell or later stages were strongly spurted with a pipette, or were shaken in a vial, and thereby some of the blasto- meres were frequently injured while others were uninjured and continued to develop. The injured blastomeres were rarely killed, as was shown by the fact that they remained transparent and entire for a day or more, whereas dead cells soon become opaque and disintegrate. These injured cells never again divide and sections show that their nuclei are frequently broken and their chromosomes scattered. Cells are more likely to be injured during nuclear division than during rest. The fact that these injured cells never again divide though they remain whole within the chorion and preserve their characteristic color and structure makes it possible to determine at all stages just what cell or cells have been injured. Whether or not the presence of these injured cells within the chorion may influence the development of the uninjured cells will be considered later. Attempts to completely separate individual blastomeres by the use of Herbst’s calcium- free sea water were not successful, probably owing to the presence of the chorion and to the close union between the blastomeres. In addition to this method of experimentation which yielded hundreds and thousands of eggs in which one or more of the blasto- meres had been injured I also cut eggs and embryos in two with knives made from small needles. In no single instance was | able to get fragments of unsegmented eggs to dey elop; in the gastrula stages | was more successful, being able to cut gastrula in two in the manner described by Driesch (’03) and observe the subsequent development. I have not attempted to repeat the various ingenious methods of injuring blastomeres which were devised and employed by Chabry, since they are necessarily slow and difficult of application and yield but a small number of injured eggs, whereas by simply spurting or shaking the eggs one may injure blastomeres in an enormous number of eggs which can then be sorted out and classi- fied according to the character of the injury; furthermore the ease and certainty with which the identity of injured blastomeres of Mosatc Development in Ascidian Eggs. 155 Cynthia may always be determined renders unnecessary such experiments as Chabry’s on the individual cleavage cells. If one desires to trace with accuracy the lineage of individual blastomeres, whether in normal or experimentally altered develop- ment, it is essential that a large quantity of material should be available. In even the most favorable material the lineage of the later stages can be successfully studied only by the aid of fixed and stained material and without a large number of eggs it 1s difficult if not impossible to secure all the stages of development. Furthermore it is desirable that a considerable number of eggs of every stage be available for study, since the liability to error decreases with the number of cases studied. Accordingly, in addition to the study of living eggs during successive stages after their injury, many eggs were also fixed at brief intervals and were afterward stained and mounted entire or sectioned. For this purpose I have found Kleinenberg’s picro-sulphuric acid followed by my picro-hematoxylin to give the best results. Entire eggs so prepared show cell outlines, nuclei and karyokinetic figures much more plainly than in the living condition; on the other hand the yellow crescent is less distinct since the yellow pigment is extracted by alcohol; nevertheless this crescent may always be recognized by its peculiar staining qualities and it therefore affords a never failing aid in orientation. Wt. RESULTS OF EXPERIMENTS. In undertaking this work it seemed to me scarcely possible that all of these strikingly different kinds of odplasm, each with its own peculiar developmental history and destiny, were neverthe- less morphogenetically alike, as might be concluded from the results of Driesch and Crampton. On the other hand a possible escape from this conclusion was suggested by the fact that although the cleavage cells are strikingly different from one another, the isolation of the odplasmic substances in them is not quite com- plete; almost all of the yellow protoplasm is contained in the yellow crescent; but a small amount of it is found around the nuclei of all the cells; most of the gray substance is contained within the dorsal hemisphere, but a small amount of it occurs in the ventral cells also; most of the clear protoplasm is found in the ventral hemisphere but a small quantity is also found in the dorsal cells. 156 Edwin G. Conklin. It therefore seemed possible that the production of a complete larva from any one or two of the first four cells might be due to the replacing of a missing substance by the greater development of the trace of that substance contained in the cells in question. “Thus the anterior quadrants which lack the yellow crescent might, perhaps, regenerate it from the small amount of yellow perinuclear protoplasm “which they contain, and correspondingly the posterior quadrants might regenerate the lacking gray crescent from the small amount of gray substance which they contain. In the light | of the work of Driesch and Crampton either there must be such regeneration, or the substances which appear so different must after all be each and all totipotent. However the solution of this problem has turned out to be much simpler than I had supposed possible, viz: 1solated blastomeres do not give rise to entire larve, as claimed by Driesch and Crampton, but on the contrary each blastomere produces only those parts of a larva which would arise from it under normal conditions. The development 1 USS in short, a “mosaic work.” Since the first cleavage is bilaterally symmetrical each of the first two blastomeres con- tains one-half of each and all of the substances of the egg and correspondingly the half larva which develops from one of these blastomeres contains portions of every larval organ. Owing to the fact that the cells which arise from an ‘isolated blastomere close over the injured surface these partial embryos are rounded in form and many of the one-half larve resemble superficially whole larve of half size, but in no case are they complete. When the anterior or posterior quadrants of the 4-cell stage are killed nothing even remotely resembling a normal larva is ever pro- duced. My results are cheretones directly opposed to those of Driesch and they agree in all essential respects with those of Chabry. The partial embryos and larve obtained in these experiments may be classified as right or left, anterior or posterior, dorsal or ventral, or composite forms. Furthermore they may be known as half, quarter, eighth, sixteenth, etc., embryos, according as they are produced from blastomeres of the 2, 4, 8, 16, etc., cell stages; however, the character of the embryo depends entirely upon the region from which the isolated blastomeres come and not upon the number of such blastomeres. Mosaic Development in Ascidian Eggs. Gn ~s 1. Right or Left Half Embryos (Figs. 13-33, 36-46). a. Cleavage. When the right or left half of an egg is injured in the 2, 4 or 8-cell stage, the other half continues to segment in a normal manner, provided it was not also injured. | ae traced the cell- lineage of these right or left half embryos up to the eighth genera- tion of cleavage cells (the 112-cell stage of normal eggs), while I have determined the lineage of many individual cells as late as the ninth or tenth generation (218-360 cell-stage). The cell-lineage of these half embryos is essentially like the right or left half of a normal egg, except that the direction of division and consequently the position and size of some of the blastomeres may be slightly altered. This alteration in the direction of cleavage is most evident in cases where the egg was injured in the 2-cell stage, and it is prob- ably due to the fact that the uninjured blastomere in such cases becomes nearly spherical in shape, and does not remain hemi- spherical as inthe normal egg. Owing to this fact the median pole of certain cleavage spindles, 7. ¢., the one next to the original median plane, is shifted toward the middle of that plane. The resulting mass of cells is, therefore, more nearly spherical than in the half of a normal embryo. (Figs. 13-20.) If the injury occurs in the 4-cell stage or later, the change in the direction of the early cleavages is not so evident as when it takes place in the 2- -cell stage. In case one of the blastomeres was injured at the close of the first cleavage, the direction of the karyokinetic spindles of the second and third cleav ages are entirely normal, since in both these cases they lie parallel with the first cleavage plane, Fig. 13; but in the fourth cleavage in which one pole of the spindles lies nearer that plane than ‘Whe other, the median pole is shifted toward the middle of that plane and consequently the cells formed along the median plane come into closer contact with one Another and the cell aggregate is more nearly spherical than in the right or. left half of a normal 16-cell stage. (Figs. MGs, 21, 22.) These results entirely agree with those of Chabry and Crampton. The fifth cleavage of the right or left half embryo is also like the normal except in the direction of a few of the divisions; e. Ges Fig. 16 is nearly normal but in Fig. 17 the division of the cell 158 Edwin G. Conklin. DEVELOPMENT OF RIGHT BLASTOMERE OF 2-CELL STAGE. Figs. 13-20. Successive stages in the development of the same right half embryo, the left blasto- mere having been injured in the 2-cell stage; drawn at intervals of about five minutes. Here and elsewhere the yellow protoplasm is indicated by coarse stipples. Fig. 13. Right half of 8-cell stage, posterior view. A small amount of yellow protoplasm surrounds the nucleus of the ectoderm cell b*-2._ The position of the cells shows that the ventral ends of the third cleavage spindles diverged from the first cleavage plane in the posterior quadrant and converged toward that plane in the anterior quadrant, just as in the normal egg. (See Conklin, ’o5!.) Fig. 14. Right half of 16-cell stage, anterior view. The yellow crescent is seen through the cell B*-1. In the normal egg of this stage the cells A®-! and a®-3 lie more nearly in front of the cells A®-? and a*.4. Fig. 15. Same stage as preceding posterior view. In normal eggs the cells B®? and b*-‘ lie nearly behind the cells B>“! and b®-3 and not on their median side. Fig. 16. Right half of 30-cell stage, dorsal view. A®? and A®4 are cells of the gray crescent; Bé.? and B®.?, cells of the yellow crescert. Fig. 17. Right half of 34-cell stage, posterior view. In normal eggs the cell B* lies on the lateral border of B®.3, Fig. 18. Same stage as preceding, dorsal view. The cell B®-! normally lies between B®.3 and A®.!. 160 Edwin G. Conklin. DEVELOPMENT oF RicHT BLASTOMERE OF THE 2-CELL STAGE; ALSO OF Ricut AND Lert BLASTOMERES OF THE 4-CELL STAGE. / « Figs. 19, 20. Same embryo as that shown in Figs. 13-18. Fig. 19. Right half of 46-cell stage, — the yellow crescent cells are not quite normal in position. Fig. 20. Right half of © posterior view; 48-cell stage, dorsal view. The caudal endoderm cells (B’-' and B?.2) have been shoved away from the © median plane by the cell B7-°. Figs. 21,22. Fixed and stained preparations of half embryos in the 16-cell stage. Fig. 21. Right half embryo, posterior view. Fig. 22. Left half embryo, ventral-posterior view. Figs. 23,24. Successive stages of one and the same half embryo, the left half having been injured — in the 4-cell stage, dorsal view. Fig. 23. Right half of 16-cell stage. Fig. 24. Right half of 32-cell — stage. The cleavage is like the right half of a normai egg in every respect. rE q Mosaic Development in Ascidian Eggs. 161 162 Edwin G. Conklin. B®? into B®? and B** is almost at right angles to its normal direction. In other cases, as is shown in Fig. 24, this cleavage is normal in direction, and | am, therefore, of the opinion that the condition shown in Fig. 17 and the later stage of this same egg shown in Fig. 19 may be due to some slight injury to the developing half of this egg. In Fig. 18, which is a dorsal view of the same egg in the same stage as Fig. 17, the cells A*? and A** have moved in toward the median plane as compared with Fig. 16, though in this respect, also, the corresponding stage shown in Fig. 24 is quite normal. ‘This shifting of the anterior dorsal cells toward the median plane is shown again at the next cleavage (the sixth), of thisegg. (Fig. 20.) The seventh cleavage, which is shown in Figs. 25 and 26, is also normal except for the direction of a few of the divisions. The cells which constitute the yellow and gray crescents are in all respects like the right half of a normal egg. However the position of the cells A71 and A7?, Fig. 25, and the direction of division in several of the ectoderm cells shown in Fig. 26 are not quite normal. In conclusion therefore it may be said that the cleavage of one of the blastomeres of the 2-cell stage or of the right or left blasto- meres of the 4-cell stage, is like that of the corresponding half of a normal egg, except in minor details. Even these minor differences are not always present and when they are they do not alter the localization of the odplasmic substances. In every case the dis- tribution of the yellow, the gray and the clear substances to the ~ different blastomeres is the same as in the right or left half of a normal egg; the cells of the yellow crescent, for example, form only the right or left half of a ncrmal crescent, and the same is true of the gray crescent and of the other substances of the egg. Even the small amount of yellow protoplasm which is found around the nuclei of the posterior ectoderm cells b*?, Fig. 13, is perfectly normal in its occurrence and subsequent iseibacon I have elsewhere (’05") shown that the localization of different ooplasmic substances in the ascidian egg precedes cleavage and that cleavage and localization are here aie ely independent of each other; these experiments show that in both cleavage and localization the development of the right or left half of an ascidian egg is a “mosaic work,” for the slight amount of regulation, which is manifested in the changes in the direction of certain cleavages, and the consequent closing of the embryo in no way alters the Mosaic Development in Ascidian Eggs. 163 histological character of the cleavage cells nor their developmental tendencies. b. Gastrulation. In the development of the right or left half of an egg the process of gastrulation sometimes occurs in an unusual manner. The most frequent modification of the normal process is that shown in Figs. 27, 29, 30, where the endoderm cells are not infolded but come to protrude above the level of the other cells, thus forming exogastrulz. In later stages these endoderm cells must become infolded for it is a rare thing to see exogastrulz or any indication of an original evagination of endoderm cells in any of the cultures of older embryos. By what process these exogastrulz right them- selves I have not been able to observe, but [ think it probable that this like’ normal gastrulation 1s accomplished by overgrowth of the ectoderm cells and change of shape of the endoderm cells. Sometimes when the endoderm cells are evaginated other por- tions of the blastula wall invaginate. In this way false gastrulz may arise in which the infolded cells are not endodermal but ectodermal, as is clearly shown by their histological structure. (Hig. 03.) While some embryos in the gastrula stage show such abnormali- ties as those which have just been described in other cases the gastrula is strictly a half one, as is shown in Fig. 31, and it seems to me probable that exogastrulz or false gastrule only arise when the surviving half of the egg has been slightly injured. ‘These half gastrulz contain just one- -half of all the cells of the normal gastrula and the position of the various cells and organ bases is essentially like that which occurs in the right or left half of-a normal gastrula; the cells of the yellow crescent lie along one side only of the blastopore groove; the neural plate and chorda cells each form half of the arc which is normally present in the anterior lip of the blastopore, while the closing of the open side of the gastrula, which is turned toward the injured cell, is chiefly accom- plished by the overgrowth of the ectoderm cells of the ventral side. (Fig. 31.) Except, therefore, for this tendency of the cells along the injured side to come together, these half gastrulz are strictly partial and the gastrulation no less than the cleavage may be regarded as an illustration of mosaic development. 164 Edwin G. Conklin. Ricut or Lerr Harr Emepryos; 64-Cetrs to GAsTRULA. Figs. 25,26. Fixed and stained half embryos; spurted in the 2-cell stage and fixed 2 hours later. Fig. 25. Right half of 64-76-cell stage, dorsal view. The neural plate cells (A%-7, AS-8, A815, AS.16) have just divided, the chorda cells (A%-3, A7-7) are dividing. The position of the cells A%-1, A%-? is slightly abnormal. (v. Fig. 6.) Fig.26. Left half of 64-76-cell stage, ventral-posterior view. Figs. 27,28. Right half of embryo in about 180-cell stage; spurted in the 4-cell stage and fixed 23 hours later. Fig. 27. Dorsal view; the large endoderm cells lie above the level of the other cells and form an exogastrula; some of the yellow cells (stippled) still lie at the surface while others are covered by endoderm cells. Fig. 28. Ventral view of similar embryo. Figs. 29, 30. Living right half embryos, dorsal view, showing the endoderm cells forming exogas- trule and the yellow crescent cells at the surface. 165 Eggs. 1c Development in Ascidian Mosa 166 Edwin G. Conklin. Ricut or Lerr Harr or Turee-QuartTer Empryos; GastrurA To TApPpoLrt. DRaAwN FROM Fixep AND STArneD MATERIAL. Fig. 31. Right half gastrula of about 220-cell stage; spurted in the 4-cell stage and fixed 3 hours later. The neural plate, chorda and mesoderm cells are present only on the right side and in their normal positions and numbers. Fig. 32. Left half of young tadpole, dorsal view; spurted in the 4-cell stage, fixed 5 hours later. The notochord is normal except for size and number of cells; the muscle and mesenchyme cells are present only on one side; the neural plate is abnormal in form but not in position. Fig. 33. Right half of young tadpole, dorsal view; spurted in the 4-cell stage, fixed 44 hours later (slightly younger stage than Fig. 32). The notochord consists of a small number of cells which are interdigitating; muscle cells and mesenchyme lie on the right side of the notochord, but not on the left, though the muscle cells have begun to grow around to the left side; the neural plate is normal in posi- tion but not in form. Fig. 34. Right-posterior three-quarter embryo, from the right side. The left anterior cells (A*-!, a‘-?) were killed in the 8-cell stage and the embryo fixed 5 hours later. The posterior half of the embryo is normal, but the left half of the anterior part is lacking and the neural plate is abnormal and has not formed a tube though sense spots are present. Fig. 35. Left-anterior three-quarter embryo, dorsal view; the right posterior quadrant (B*) was killed in the 4-cell stage and the embryo fixed 6 hours later. The anterior half of the embryo is entirely normal. The muscle cells are lacking on the right side though they have begun to grow around the hinder end of the notochord. The posterior portion of the trunk mesenchyme is found only on the left side, but its anterior portion, which is derived from the cells 47-6 and A’. (Fig. 6) of the anterior quad- rants is present on both sides. In the region of the injured cell the notochord and neural tube are curved away from that cell. Fig. 36. Left half embryo, from left side; spurted in the 4-cell stage, fixed 6 hours later. The dorsal lip of the blastopore is being overgrown by the ventral (posterior) lip. Muscle cells and mesenchyme are found only on the left side. The neural plate is abnormally folded, but still open; sense spots are present. 167 A g } Mosaic Development in Ascidian Eg 168 Edwin G. Conklin. c. Formation of Larva. A considerably later stage in the development of the half embryo is shown in Figs. 32, 33 and 36 (Figs. 34 and 35 are three-quarter embryos and will be described later); of these stages Fig. 33 is the youngest and Fig. 36 the oldest. In all of these figures the blasto- pore has already Heeed and the chorda cells have given rise to a fusiform notochord, which lies in the posterior half of the embryo. The blastopore closes chiefly by the posterior growth of the dorsal (anterior) lip, as in the normal gastrula. With the formation of the notochord the posterior half of the embryo becomes elongated and narrower than the anterior half and the developing tail bends around toward the injured side. (Figs. 32, 33.) The anterior half remains large, the posterior half becomes long and narrow; the latter portion contains the notochord and muscle cells, the former the gastral endoderm, mesenchyme and most of the neural plate. al he general superficial appearance of an embryo of this stage is very similar to a normal one, but a more detailed study shows many differences. (1) Neural Plate. The neural plate occupies in the main its normal position, that is, it lies along the first cleavage plane on the dorsal side, next to the injured cell. In this position the plate becomes folded and ultimately comes to contain a vesicle (the sense vesicle) though the steps by which this vesicle is formed are always irregular and abnormal. (Figs. 36-40.) ‘The anterior por- tion of ie plate is usually doubled over posteriorly while the posterior portion is folded forward (Figs. 36, 39, 40) and in this way a vesicle is finally formed. The tail of the embryo grows around toward the injured side so that the concave side of the embryo is median or dorsal, the convex side being lateral or ventral. In the younger, normal larve the concave side is ventral, the convex dorsal. In these half larve the nerve plate lies along the concave side, a con- dition which is the reverse of what is found in the normal larva. (cj. Figs. 12 and 36.) In the older half larve there 1s almost always found one or more pigmented sense spots in the neural plate or sense vesicle. (Figs. 36-40, 45,46.) These pigment-spots appear within cells of the neural plate and, as Lam well convinced, always within definite cells, though owing to the abnormal foldings of the neural plate they do not Ske ays occupy exactly the same Mosaic Development in Ascidian Eggs. 16g 5 positions. Furthermore these sense spots may be more numerous than in the normal larva, as shown in Figs. 45 and 46, probably owing to the fact that the cells which form the pigment and which normally lie on the margins of the neural plate do not come to- gether to form two spots as in normal larva, but remain separated so that several such spots are formed. (2) Notochord. The chorda cells grow back into the posterior half of the embryo and the cells here interdigitate in the normal manner, finally forming a linear series of cells. (Figs. 32-46.) The notochord, which is at first relatively short and thick, Fig. 33: becomes later very much longer and more slender, Fig. 40, and in all respects it has the appearance of a normal notochord, save that it evidently contains a smaller number of cells. “The position of the notochord of the half larva is always slightly abnormal; it never lies along the original median plane (first cleavage) as in normal larvze, but its anterior end is diverted away from that plane and toward the lateral border of the larva. (Figs. 32, 33> 37,41.) This position is that which the chorda cells, which arise in the anterior lip of the blastopore and which grow posteriorly around the mar- gin of the blastopore, would natur rally assume. (cf. Figs. 31 and 33.) What it is which causes the chorda cells to interdigitate i in their characteristic manner is a question difficult to answer; it certainly is not dependent upon the crowding together of chorda cells from the right and left sides since it occurs normally when the cells of one side only are present; on the other hand it must depend upon a certain amount of lateral compression of the chorda cells since it occurs very rarely if at all in the anterior half larve in which the ectoderm and mesoderm of the tail are lacking. (3) Muscles and Mesenchyme. In these right or left half embryos and larve the muscle and mesenchyme “cells are present on one side of the notochord; here they occupy their normal posi- tions, the muscle cells giving rise to three rows of cells along the lateral border of the notochord and the mesenchyme forming a group of small cells anterior to the muscle rows. (Figs. 32, 33, 36.) In later stages the muscle cells slowly extend over to the side of the tail on which they were originally lacking; this takes place espe- cially at the hinder end of ‘the tail, the overgrowth taking place around the end of the notochord and over its ventral side. In this way the right or left half embryo or larva tends to become com- plete, but I have never seen a case in which three rows of muscle 170 Edwin G. Conklin. cells were found on both sides of the notochord. Indeed, I am not at all sure that this extension of the muscle cells around the end of the notochord is accompanied by any increase whatever in the number of muscle cells or in the number of rows of cells. The latest stage in which I| can positively identify the three rows of muscle cells is shown in Fig. 36. In this larva the muscle rows lie nearer the ventral side than in iontnal larve (see Fig. 12), and they are evidently extending over the ventral surface toward the opposite side. In later stages the muscle cells become much elongated, but I have not been able to determine the number of rows present. I have found it still more difficult to decide whether the trunk mesenchyme ever extends over to the side on which it was originally lacking, but I believe that this takes place only to a limited extent, if at all, and that Chabry was right when he afhirmed that only one atrial invagination 1s formed in these right or left half embryos. 2. Three-Quarter Embryos (Figs. 34-35). In connection with the right or left half embryos I shall here consider three-quarter embryos, which, of course, include the whole of the right or left half. “wo such embryos are shown in Figs. 34 and 35. In the former the left anterior quadrant was killed in the 8-cell stage; in the latter the right posterior quadrant in the 4-cell stage. “The embryo in which the cells of the anterior quadrants were uninjured (Fig. 35) is perfectly normal in its anterior half; its posterior half, however, lacks those parts which would have developed from the cell which was injured. ‘This embryo is younger than the one shown in Fig. 34 and no sense spots are present, but the sense vesicle is closing in a normal manner. This figure well shows that a part of the trunk mesen- chyme is derived from the anterior quadrants, and indeed from the pair of cells A**, Fig. 6, while a portion of it comes from the posterior quadents, as may be seen by comparing the nght and left sides of Fig. 35. The muscle cells are entirely lacking on the right side, the substance which would have formed them being located in the injured cell B*; they are shown growing around the end of the notochord as in the half embryo shown in Fig. 33: The notochord and nerve tube are apparently full sized, which is explained by the fact that they come from the anterior quadrants, but owing to the lack of the right side of the tail they are somewhat distorted in form. ; Mosaic Development in Ascidian Egg & 17! Rieut or Lerr Harr Larvar. Fixep anp Stainep MatTERIAL. | { Figs. 37-40. Four half larve from eggs which were spurted in the 2-cell or 4-cell stage and fixed 22 hours later. These larve are still within the egg membranes though at a corresponding age normal | larve are undergoing metamorphosis. Fig. 37. Right half larva, ventral view. The tail which is elongated is turned down toward the dorsal side; the sense vesicle also lies on the dorsal side and is here seen through the embryo. The muscle cells are chiefly on one side of the notochord but have grown over to the other side at the posterior end. Fig. 38. Left half larva, dorsal view. The neural plate with sense spots is partly covered by the end of the tail. The mesenchyme is found only on the left side. | Fig. 39. Left half larva from the left side. The neural plate is folded so as to form a nearly closed sense vesicle, in which are two sense spots. } Fig. 40. Left half larva viewed from the left side. The neural plate is partially closed, but is abnormal inform. In all of‘these larve the neural plate lies on the concave side. 172 , Edwin G Conklin. Ricut or Lerr Harr Larvar Drawn From Livinc SpeciMENS FROM 12 Hours (Fics. 41, 42) To 2¢ Hours (Fies. 45, 46) Arrer THE Injury or ONE oF THE First Two BLasTOMERES. Fig. 41. Posterior-dorsal view. Fig. 42. Same embryo, posterior ventral view. In both these figures the muscle cells are found chiefly on one side of the notochord, but they have grown over to the opposite side at the end of the tail. Fig. 43. Right half larva from right side. Fig. 44. Left half larva from left dorsal side. The yellow crescent on the injured blastomere apparently occupies dif- ferent positions with respect to the larva in these two figures, but it is by no means certain that the convex side of the larva is morphologically the same in the two figures. Figs. 45, 46. Two views of one and the same left half larva. Fig. 45, from the dorsal side; Fig. 46, from the left side, showing two sense spots on the dorsal and two on the ventral sides. The neural plate is continuous between these spots on the side next to the injured blastomere. Mosatc Development in Ascidian E 174 Edwin G. Conklin. In Fig. 34 the left anterior quadrant was killed and the posterior portion of this embryo is normal save only for the fact that the notochord and nerve tube are smaller than usual, which is ex- plained by the fact that the substance of these organs is derived from the anterior quadrants; three rows of muscle cells are found on both sides of the tail. ‘The anterior half of this embryo, on the other hand, is quite defective; the neural plate is irregularly folded and has not formed a sense vesicle, although sense spots are present. I have seen and studied many three-quarter embryos sim- ilar to those shown in Figs. 34 and 35 and they all show, as do the right and left half embry os, that where part of the substance which would normally form an organ is destroyed the organ which develops is defective, whereas if alk or any organ- forming! substance is lacking the organ to which it would ‘normally give rise is also lacking. So far as I have observed these partial larve never escape from the egg membrane, and in this my observations accord with those of Chabry and Driesch, and although I have kept them alive until a period after the normal larve have undergone metamorphosis [ have never observed this transformation in ‘eae In conclusion then I find that the cleavage and gastrulation of these half or three-quarter embryos is partial and the resulting larva incomplete although the notochord is well formed and there is a tendency on the part ‘of some of the cells to grow over and close up the open side of the larva. However, this regulation never leads to the formation of a complete larva; the neural plate may close, but it forms an abnormal sense vesicle: at the end of the tail the muscle cells extend over toward the ceed side, but they do not form three rows of cells on each side of the notochord as in the normal larva; the mesenchyme likewise does not develop along the injured side and it is probable that only one atrial invagination is formed. Furthermore not a single cleavage cell nor any one of the odplas- mic substances ever gives rise to parts or organs which it would not normally produce; the notochord, jor example, invariably comes from the chorda cells, the sense vesicle from the neural plate cells and both these structures from the material of the gray crescent, the muscles always come from the muscle cells and these from the sub- stance of the yellow crescent; the ectoderm, fromthe ectoderm cells and ultimately from the clear protoplasm; the endoderm, from the Mosaic Development in Ascidian Eggs. SS wn endoderm cells and these from the deep gray material of the eg In spite therefore of the regulation which is apparent in the ce of the open side of the embryo, and in the formation of a whole notochord and of an imperfect sense vesicle, the various odplas- mic substances of the unsegmented egg and of the different blastomeres are not totipotent but each shows in these experi- ments, as well as in normal development, that it is differentiated to give rise to one, and only one, particular kind of tissue. 3. Anterior Half Embryos (Figs. 47-52). The anterior and posterior half embryos show even more clearly than do the lateral ones the mosaic character of the develop- ment of these eggs. When the posterior half of an egg is killed in the 4-cell or 8-cell stage the anterior half continues to develop as if the posterior half were still living. ‘The cleavage is in all respects like that of the anterior half of a normal egg; the gastrula- tion is essentially the same, but the later development 1s modified in many important particulars. Figs. 47 and 48 are ventral and dorsal views, respectively, of one and the same living embryo of the 76-cell stage, in which the posterior dorsal cells, B**, containing the yellow crescent, were killed in the 8-cell stage. None of the cells of the ventral hemi- sphere were injured and consequently the cleavage of these cells is quite normal; thirty-two ectoderm cells are present, all of which have entirely normal positions, shapes and sizes. (cj. Figs. 5 and 47.) ‘The anterior half of the dorsal hemisphere is also entirely normal (c}. Figs. 6 and 48); eight chorda cells are shown forming an arc which bounds anteriorly the six endoderm cells and which is flanked on each side by the anterior mesenchyme cell, A’. ‘The number, size and position of each and all of these cells is the exact counterpart of what is found in the normal embryo, and, although the outlines of the neural plate cells were so indistinct in the living specimen from which this figure was made that I could not draw them, there is every reason to suppose that these cells like all the others in this embryo conform to the normal type. In the posterior half of the dorsal hemisphere all the parts which would have developed from the cells B*' and 5** are entirely lacking; there are neither mesenchyme, caudal endoderm, 176 Edwin G. Conklin. Anterior Hatr aND THreE-QUARTER Empryos; 76 Cetts To METAMORPHOSIS. Figs. 47, 48. Anterior-ventral three-quarter embryo of the 76-cell stage (v. Figs. 5 and 6); the dorsal posterior cells B*-!, containing all of the yellow crescent, were killed in the 8-cell stage. The ventral ectoderm cells (Fig. 47) are quite normal both in position and number (cf. Figs. 5 and 47); the anterior dorsal cells are also normal, but the posterior dorsal cells (muscle, mesenchyme and caudal endoderm) are entirely lacking. (cf. Figs. 6 and 48.) Figs. 49-51. Three views of one and the same anterior half embryo of about the 250-cell stage; spurted in the 4-cell stage and fixed 2 hours later. Fig. 49. Dorsal view, superficial focus, showing the neural plate. Fig. 50. Dorsal view, deeper focus, showing two rows of chorda cells besides several ectoderm and endoderm cells. Fig. 51. Dorsal view, still deeper focus, showing the cells of the ventral ectoderm. Fig. 52. Arterior half embryo, dorsal view. Spurted in the 4-cell stage, fixed 22 hours later. The yellow crescent is plainly visible in the injured cells. Sense spots are present but the neurdl plate never forms atube. The chorda cells lie in a heap at the left side. There is no trace of muscle sub- tance or of a tail in this anterior half embryo. This embryo is from the same experiment as Figs. 37-40; normal larve of this stage are undergoing metamorphosis. Mosaic Development in Ascidian Eggs. 177 178 Edwin G. Conklin. nor muscle cells. Unfortunately this particular embryo was not followed through the various stages of development until it gave rise to a larva and none of the older stages which I have studied have shown precisely this ty pe of injury, 7. e., the destruction of the yellow crescent w ithout injury to the ectoderm cells of the posterior half. In many other cases which | have seen all of the posterior half of the egg was injured in the 4-cell stage. I have followed the dev elopment of the surviving anterior hale es of such eggs as late as the stage of the metamorphosis of the normal larvz; the develop- ment of such blastomeres is alw ays partial. Figs. 49, 50 and 51 represent three views of one and the same anterior half embryo of about the 250-cell stage; in all the figures the embryo is viewed from the dorsal side, but in Fig. 49 the focus is high and only the ectoderm and neural plate cells of the dorsal surface are shown; Fig. 50 is a median optical section showing chorda and endoderm cells surrounded on the anterior side by ectoderm; Fig. 51 represents the ectoderm of the ventral surface which is visible at a deep focus. This half embryo is exactly like the anterior half of a normal one in the formation of the neural plate, the chorda plate, the general ectoderm and gastral endoderm, in the overgrowth of the dorsal lip of the blastopore, even in the position, shape and size of the individual cells. (c7. Figs. 9 and ro.) Finally in Fig. 52 there is represented an anterior half embryo 22 hours after the posterior cells were killed, and at a stage when normal larvae of corresponding age have already under- gone metamorphosis. The ectoderm has not yet inclosed the embryo on the side next the injured cells, and this rarely happens in anterior or posterior half embryos. ‘The neural plate has not rolled up nor invaginated to form a tube, though it is slightly depressed along its median line; two sense spots are present though there is no sense vesicle. ‘The large rounded chorda cells are irregularly scattered along the posterior border of the embryo, where they project beyond the ectoderm; they never form a noto- chord. There is no trace of yellow crescent substance nor of muscle cells in these anterior larvze and no indication whatever of a tail. hey are, therefore, altogether unlike the normal larve and they afford complete and convincing evidence that the anterior blastomeres of the ascidian egg are not totipotent but rather that the development is a mosaic work. Mosaic Development in Ascidian Eggs. 179 4. Posterior Half Embryos (Figs. BG 5O): All that has been said of the mosaic-like development of the anterior half of the egg is equally true of the posterior half. The cleavage progresses in Gormal fashion up to the time of the closure of the blastopore. Figs. 53 and 54 represent posterior half em- bryos of the 32-cell and 70-cell stages, respectively. “The former is entirely normal and the latter is normal in all respects save that a single pair of cells, B**, is larger than in the normal embryo. The clear, the yellow and the gray Fennec of the egg are distributed exactly as in the posterior half of a normal embryo. ‘The clear ectoderm cells lie on the ventral side and only two of them appear in the dorsal view shown in Fig. 54 (the two clear cells at the pos- terior pole). In Fig. 53 the gray endoplasm is contained in two cells (B°*) and in Fig. 54, in four (B71, B7?); these cells give rise to the strand of caudal endoderm. ‘The yellow crescent consists at the 32-cell stage of a single arc of yellow cells (Fig. 53) which then, by division, become a double arc of fourteen cells (Fig. 54); the inner arc consists of eight mesenchyme cells and the outer of six muscle cells. In all these respects these posterior half embryos are entirely like the posterior half of a normal embryo. But while the pregastrular stages of these posterior half embryos are like the normal, the gastrula and later stages show many interesting modifications. Figs. 55, 56, 57 are three views of one and the same posterior half embryo, the normal embryos of the same stage being young tadpoles like Fig. 11. In all of these figures the embryo i is viewed from the dorsal side; Fig. 55 shows the ectoderm cells which cover the dorsal surface; ce 5, the muscle cells which lie below the ectoderm on the dorsal side; Fig. 57 is an optical section at a still deeper level showing the caudal endoderm.and mesenchyme. Fig. 58 is another posterior half embryo of similar age seen from the ventral side, showing the yellow mesoderm cells on each side of the caudal endoderm. The gastrulation occurs between the stages shown in Figs. 54 and 55. he caudal endoderm and the surrounding arc of mesen- chyme, shown in Fig. 54, invaginates; the muscle cells come to lie above (dorsal to) the mesenchyme cells and finally the latter are overgrown by the ectoderm in the manner shown in Prev.) in normal embryos the posterior part of the blastopore is closed chiefly by the growth of the anterior lip; in the latter stages of 180 Edwin G. Conklin. Posterior Harr Empryos;32 Certs to TappoLe Stace. Fixep AND STAINED PREPARATIONS. Fig. 53. Posterior half of 32-cell stage, dorsal view. The cleavage is altogether normal. Spurted in the 4-cell stage, fixed 1 hour later. Fig. 54. Posterior half of 76-cell stage (cf. Fig. 6); spurted in the 4-cell stage, fixed 2 hours later. Two rows of yellow crescent cells are present, the inner being mesenchyme, the outer muscle cells; the anterior pair of mesenchyme cells (B*-®) are larger than normal. There are two pairs of caudal endoderm cells (B%-! and B7.2). A pair of ventral ectoderm cells is visible in the midline behind. Figs. 55-57. Three views of one and the same embryo; spurted in the 4-cell stage, fixed 4 hours later, normal embryos being in the stage represented by Fig. 11. Fig. 55. Dorsal view of the super- ficial ectoderm. The notch in front represents the notch in the ventral lip of the blastopore. Fig. 56. - Same view, deeper focus, showing the muscle cells beneath the ectoderm; these cells are continuous from side to side, there being no chorda inthe midline. Fig. 57. Same view, still deeper focus, showing the double row of ventral endoderm cells in the midline, and on each side of this a mass of mesenchyme cells. Fig. 58. Ventral view of posterior half embryo of the same stage as the preceding, showing the muscle and mesenchyme cells beneath the ectoderm and on each side of the strand of ventral endoderm. 182 Edwin G. Conklin. gastrulation a blastopore groove is left in the posterior half of the embryo, on each side of which lie the muscle cells. (Fig. 9.) By the continued growth of the anterior lip this groove is shoved to the posterior end of the embryo and the rows of muscle cells are tilted up from an antero-posterior to a vertical position. Later, when the notochord 1s formed, the muscle cells come to lie alongside of it, thus forming the three rows of muscle cells on each side. Finally the ectoderm of the posterior lip of the blastopore, which has, up to this stage, formed a notch at the end of the blastopore groove, grows forward and reduces this groove to a minute pore. Owing to the absence of the anterior lip of the blastopore, and of the notochord and the neural plate, the later stages in the develop- ment of these posterior half embryos is much altered. In the first place the blastopore groove and the muscle cells are not pushed to the posterior end of the embryo. ‘Then the muscle cells on each side of the blastopore groove are not kept apart by the notochord but come into contact forming a continuous layer of muscle cells across the dorsal side. (Fig. 50.) The blastopore groove, therefore, disappears by the fusion of the lateral lips of the groove and the ectoderm cells grow over the whole dorsal surface; the only trace of the blastopore groove which is left is a slight notch in the ante- rior border of the embryo. (Figs. 55, 56.) |The ectoderm never entirely incloses the posterior half embryo on the side next the injured cells, but the endoderm here comes to the surface as shown in Figs. 57 and 58. No trace of notochord, neural plate nor sense spots ever appears in these posterior half embryos, and what is more remarkable a tail is never formed but the embryo always remains rounded in form, as shown in Figs. 55-58. It is quite evident that the elonga- tion of the tail of the normal larva, together with the elongation of the individual muscle cells and perhaps also the arrangement of these cells in three rows on each side, is dependent upon the pres- ence and elongation of the notochord. Perhaps one reason why a normal notochord is never formed in the anterior half embryo is due to the fact that the ectoderm does not completely inclose the embryo, so that the chorda cells in their growth crowd out of the open side and hence become free and scattered. In conclusion, the study of anterior or posterior half embryos establishes in a most convincing manner the fact that the develop- ment of individual blastomeres of the ascidian egg 1s a mosaic work. Mosaic Development in Ascidian Eggs. 183 These blastomeres give rise only to those tissues and parts of an embryo which would come from them normally. Nothing even remotely resembling a complete normal larva is ever produced from the anterior or posterior quadrants of the egg. 5. Quarter Embryos (Figs. 59-70). The development of individual blastomeres of the 4-cell stage furnishes additional confirmation of the mosaic theory as applied to ascidian eggs; in every instance individual blastomeres give rise only to those parts or organs which they would produce in normal embryos. Quarter embryos generally show more abnor- malities and variations than half embry os,—probably owing to the more severe injury which they have suffered, which often Liiees the surviving quarter of the egg. The cleavage of these quarter eggs is normal in every detail, save that the position of the cells is sometimes slightly altered; the thythm of cleavage and the size and quality of the cells is the same as in the corresponding quarter of a normal egg. In Fig. 59, which corresponds to the 16-cell stage of the normal egg, aoa of the surviving quadrants has dnd twice; in Fig. 60 the left posterior quadrant of a 44-cell stage 1s shown and in both of these figures the size, quality and position of the cells as well as the rhythm of division and the distribution of the different odplasmic sub- stances is entirely normal. Fig. 61, which is the right anterior quadrant of the 76-cell stage, is agueal Hee ery respect, save forthe position of the endoderm cells which are here displaced toward the first cleavage plane. ‘The mesoderm cells in the right poste- rior quadrant, shown in Fig. 62, are not normal in position; the two caudal endoderm cells (lying next the first cleavage plane) _ are, however, normal and the ectoderm cells are normal save that they show a tendency to grow inward at the first and second cleavage furrows and thus surround the embryo. In particular, attention should be directed to the yellow crescent and caudal endoderm cells in Fig. 60, and to the neural plate and chorda arcs in Fig. 61, which are similar in every respect to the quarter of a normal embryo at these stages. I have already commented upon the fact that the quarter embryo shown in Fig. 63 is a “false gastrula” since the invaginated cells are ectodermal, probably neural plate cells, while the Targer endo- 184 Edwin G. Conklin. Quarter Empryos; 16 Certs to YounGc Tappore Stace. Fixep AND STAINED PREPARATIONS. Fig. 59. Left anterior and right posterior (diagonal) quarter embryos of the 16-cell stage, ventral view. Fig. 60. Left posterior quarter embryo of the 44-cell stage, posterior view. Fig. 61. Right anterior quarter embryo of the 76-cell stage, dorsal view, showing the neural plate and chorda cells of the right side. Fig. 62. Right posterior quarter embryo of about the 18o-cell stage, dorsal view (cf. Figs. 7, 8); spurted in the 4-cell stage, fixed 24 hours later, showing 6 muscle and 2 caudal endoderm cells. Fig. 63. Left anterior quarter embryo, dorsal view; spurted in the 4-cell stage, fixed 5 hours later. An invagination of the ectoderm cells has the appearance of a gastrula, but is probably the invagination of the neural plate. Fig. 64. Left anterior and right posterior (diagonal) quarter embryos, dorsal view; spurted in the 4-cell stage, fixed 5 hours later. Muscle cells are found only in the posterior quarter. 186 Edwin G. Conklin. Quarter Emeryos; YounG Tappote To MetaMorpuHosis StaGes. Fixep AND STAINED PREPARATIONS. Fig. 65. Leit anterior and right posterior (diagonal) quarter embryos, dorsal view; spurted in the 4-cell stage, fixed 5 hours later. The anterior quarter shows thickened ectoderm cells, probably neural plate. around the endoderm cells; in the posterior quarter are 8 muscle and 3 caudal endoderm cells. Fig. 66. Left anterior and right posterior (diagonal) quarter embryos from the right anterior side, the dorsal pole being above; spurted in the 4-cell stage, fixed 22 hours later. In the posterior quarter the muscle and mesenchyme cells form a solid mass; in the anterior quarter the chorda cells project freely over the dorsal surface and the neural plate is partially infolded and contains three sense spots. Fig. 67. Right anterior and left posterior (diagonal) quarter embryos, dorsal view; spurted in 4-cell stage, fixed 22 hours later. Fig. 68. Left anterior and right posterior (diagonal) quarter embryos, dorsal view; spurted in 4-cell stage, fixed 22 hours later. In this and the preceding figure the chorda cells (Ch.), neural plate (n. p.) and sense spots are found only in the anterior quarters; the muscle, mesenchyme and caudal endoderm cells, only in the posterior quarters. Figs. 69, 70. Right anterior quarter embryos, dorsal side above; spurted in.4-cell stage, fixed 12 hours later. These embryos show free chorda cells, neural plate and sense spots, but not a trace of muscle cells. Mosaic Development in Ascidian E 188 Edwin G. Conklin. derm cells remain on the rounded surface of the embryo. I have not observed in detail the process of gastrulation in any of these quarter embryos, but it is evident chat there is no considerable gastrula cavity and that the endoderm cells are chiefly overgrown by the ectoderm, as shownin Fig.65. Ultimately the endoderm and mesoderm are largely overgrown, though in this case, as inthe half embryos, the ectoderm does not entirely inclose the embryo on the side next to the injured cells and through the opening thus left some of the endoderm cells may protrude. Although the localization of ooplasmic substances and of organ bases 1s usually the same as in the quarter of an entire embryo, in some cases there are dislocations of these substances and bases which are probably due to injury of the surviving quar- ter. ‘Thus in the left anterior quarter, shown in Fig. 64, large endoderm cells lie at the surface next to the first cleavage plane; in the same quadrant of another egg shown in Fig. 65 the neural plate cells lie at the periphery of thie quadrant and chiefly on the left side, instead of along the median plane as in normal embryos. I have seen many other instances of such dislocations but they are all of such nature that they can be interpreted as due to slight injury to the surviving blastomeres. In not a single instance are parts derived from a blastomere which would nor- mally have come from another cell. The anterior quarter embryos are always recognizable by the presence of the neural plate and, in later stages, of ‘the sense spots. The neural plate usually remains at the ore es and is not infolded, but in some cases it is invaginated through at least a portion of its area, though a sense vesicle is not formed. (Figs. 63, 66.) In all later stages one or more sense spots appear in the plate. (Figs. 66-70.) ‘The neural plate always lies along the dorsal side of the Anibty o, though it may be shifted more or less from the median plane. (Figs. 65-70.) The chorda cells are found exclusively in the anterior quadrants and in later stages they protrude to the exterior along the injured side where Te are found as scattered cells in the perivitelline space. (Figs. 66- 70.) In no case, save one, have I seen any indication that these cells form a rod- -shaped notochord, and this case (Fig. 72) was that of a living embryo in which it 1s possible that the notochord-like structure was really composed of gastral endoderm and hence not a true notochord at > = = all. It is evident that the chorda cells are unable to give rise to a a Mosaic Development in Ascidian Eggs. 189 notochord when once they have escaped and have become free, a certain amount of compression being necessary to bring about the characteristic interdigitation which leads to the formation of a rod- shaped notochord. ! The posterior quadrants can be distinguished in all eggs at all stages by the presence of the yellow crescent substance or cells. In early stages, as I have shown, these crescent cells are normal in position and character; in later stages the yellow cells fill the whole interior of the embryo. When once these cells have been inclosed by the ectoderm | have been unable to recognize any constancy in their position and arrangement. As in the posterior half embryos, a tail is never formed in these posterior quarter embryos and the muscle cells are never elongated, both these features evidently depending upon the presence of a notochord. The caudal endoderm cells are found in most, if not all, of these posterior quarter embryos as a single row of yolk laden cells which lie along the first cleavage plane (Figs. 65-68), the position which they normally occupy. These quarter embryos show in the most unmistakable manner that the development 1s strictly partial, and that an individual blastomere never gives rise to parts which it would not produce in the entire embryo. Among the hundreds of quarter embryos which I have studied both in the living condition and as stained and mounted preparations I have never seen a single one which even remotely resembled a normal larva. 6. Eighth or Sixteenth Embryos (Figs. 71-76). When eggs are spurted or shaken in the 8-cell and 16-cell stages a great variety of abnormal forms are produced, a few of which are shown in Figs. 71 and 73-76. Without exception, however, the same principles apply here as in the case of half and quarter embryos, viz: a given blastomere or group of blastomeres produces only those parts “of an embryo or larva which would develop from it under normal conditions. Fig. 71 represents an embryo derived from the dorsal anterior eighth of an egg (the cell 4*") 14 hours after the injury. Normally this eighth gives rise to neural plate, chorda, gastral endoderm, and a small amount of 1Chabry, however, figures (his Fig. 18) a partial embryo with a rod-shaped notochord lying outside the embryo in the perivitelline space. 190 Edwin G. Conklin Partiat Empryos FrroM IsoLtateD BLAsTOMERES OF 8-CELL oR 16-CeLL StaGes. DRAWN FROM Livine SPECIMENS. Fig. 71. Right anterior dorsal eighth embryo, 14 hours after injury, showing endoderm; chorda, and neural plate cells with sense spots. Fig. 72. Right anterior quarter embryo, 14 hours after injury, showing chorda, neural plate, and sense spots. Fig. 73. Posterior ventral quarter embryo derived from the cells b‘-?, b'-2 and containing no endo- derm and only a small amount of yellow protoplasm which was derived from the perinuclear plasm of the cells b*-?, Fig. 13. Fig. 74-76. Three views of a partial embryo derived from 7 cells of the 20-cell stage, viz: 2 (B*?), 2 (b*-#), 1 (a5), 2 (a5-3). (cf. Figs. 3 and 4.) The embryo consists of an outer layer of clear ectoderm and of a mass of yellow mesenchyme cells derived from the cells B®”, but it is wholly without endoderm. Fig. 74, Ventral view; Fig. 75, Posterior; Fig. 76, Postero-dorsal. 192 Edwin G. Conklin. mesenchyme derived from the cell A*®.. Inthe embryo shown in Fig. 71 the neural plate cells are clearly shown around the periphery of the figure and two of the cells contain sense spots. ‘The chorda, endoderm, and mesenchyme cells are shown internal to the neural plate, but | am unable to distinguish in this embryo between these three kinds of cells; they are ail more or less yolk-laden as in the normal egg. Owing probably to the fact that no ventral ecto- derm cells are present the neural plate 1s not pushed up onto the dorsal face and there are no evidences of gastrulation, although normal embryos of a corresponding age have already reached the full larval development. That this failure to eastrulate is not due to the slower development of the egg fragments as compared with the entire egg is shown by the degree of histological differen- tiation of the neural plate and sense spots, the latter appearing normally only in the fully formed larve. Fig. 72 is a quarter embryo of the same age as the preceding, derived from the cells 4*', a*? of the right anterior quadrant. The ventral ectoderm cells have here pushed the neural plate cells up onto the dorsal face of the embryo, while the chorda cells (?) lie along the median and transverse furrows. Four sense spots are present in the neural plate. Fig. 73 is also a quarter embryo of the same age as the pre- ceding, derived from the two posterior ventral cells b*?, b#. This embryo consists entirely of ectoderm which is arranged in a single layer of cells around a central cavity, the blastocoel. There fas been no gastrulation and the embryo contains neither endo- derm nor ereccdemn. A few of the ectoderm cells next to the cell 4*! contain yellow pigment, exactly as in the normal embryo. Figs. 74 to 76 are three views of one embryo, about 20 hours after ric egg was spurted in the 20-cell stage. By the spurting all the cells were killed except seven from which this embryo has developed, viz: a pair of mesenchyme cells B*?, and five ectoderm cells, b*4, b*4, a®4, a®* and a*4. (See Fig. 3.) This embryo consists entirely cn an outer layer of clear Seer cells, inclosing at its posterior end a mass of small mesenchyme cells; it contains no endoderm. It 1s an interesting fact that the mesenchyme cells are here inclosed by the poodeen showing that some process in the nature of gastrulation must have taken place. A great many other partial embryos, produced from one or more blastomeres of the 8, 16 or 32-cell stages, have been studied Mosaic Development in Ascidian Eggs. 193 but they all illustrate the principle that a blastomere never gives rise to any other structures than those which it would produce in a normal embryo. 7. Anterior and Posterior Half Gastrule (Figs. 77-82). In a recent publication Driesch ('03) has maintained that an alteration in the capacity for regulation occurs in the ascidian development between the early and the late gastrula stages. When the open cup-shaped gastrulz of Phallusia were cut in two transversely into anterior and posterior halves, each of these halves developed into “einer vollstandigen kleinen Appendi- cularie, welcher Organe niederer Bedeutung (Otolith, Augenfleck) eventuell fehlten.”” However, when the elongated gastrulz were cut in two transversely a head developed from one piece and a tail from the other, “so deutlich und sharf begrenzt und ausgebildet, als habe man eine fertige Appendicularie scharf durchschnitten.” Considering the results which I have obtained on the develop- ment of the two anterior or two posterior cells of the 4-cell stage of Cynthia the conclusions of Driesch seemed most remarkable and I therefore undertook to repeat his experiments upon Cynthia. Gastrulz of the stage shown in Fig. 8 were cut in two with a sharp knife made from a needle, under a Zeiss binocular dissecting microscope. With the power used the individual cells of the yellow crescent could be plainly seen and it was always easy to determine the exact boundary between the anterior and posterior halves. In every instance the section was made as close as possi- ble to this boundary (second cleavage plane) and so as to leave all of the yellow cells in one of the pieces. Owing to the presence of the chorion the experiment was not an easy one to perform, since the chorion would frequently slip under the knife, or the egg move within the chorion. Nevertheless in one day I succee Bed in cutting in two about thirty of these early eastrule; ten of these lived for twenty hours or longer after the operation, the others were too badly crushed to survive. Four of these which survived the operation are shown in Figs. 77-82, the drawings having been made from nineteen to twenty hours after the operation. Every one of these ten surviving embryos was a partial one and, although I was unable to Boeeraaic their structure with the same amount "of detail as in the case of stained and mounted preparations, it was 194 | Pdi Go-Carts ANTERIOR AND Posterior Hatr GaAsTRULAE. Figs. 77-82. Partial embryos derived from gastrulz of the stage shown in Fig. 8, which were cut in two transversely so as to leave the whole of the yellow crescent in one half. The chorion is shown as a line around the embryos. Figs. 77, 78. Dorsal and ventral views respectively of one and the same embryo, drawn 19 hours after the operation. A mass of cellular debris lies between the two half embryos; the endoderm cells are chiefly contained in the anterior half, the mesenchyme and muscle cells are entirely confined to the posterior half. Neither half at all resembles a normal embryo or larva. Figs. 79, 80. Ventral and postero-dorsal views of another embryo, 193 hours after the operation. The crescent of yellow cells is entirely confined to the posterior half and neither half resembles a normal larva. Fig. 81. Anterior and posterior half embryos 20 hours after the operation. Fig. 82. Pos- terior half embryo from the postero-dorsal side, 20 hours after the operation. The anterior half is degenerating and is shown only in dotted outline; the posterior half contains all of the yellow cells and practically no endoderm. At the stages represented by all these figures the normal embryos have already undergone their metamorphoses. 196 Edwin G. Conklin. quite evident that not one of them resembled in any respect what- ever a normal larva. In some cases both halves survived, as shown in Figs. 77-81, in other cases one half only survived. In all cases the surv iving halves became rounded in form after the operation, the more Stree injured cells being crowded out of the embryo and forming a cellular mass of Reyes within the cho- rion. In every instance the surviv ing halves remained within the chorion, which was sometimes iafoled as shown in Figs. 77-80. Each half was surrounded by a layer of clear ectoderm cells; the vellow cells were always found exclusiv ely in the posterior half, the gray endoderm cells largely in the anterior half. Nothing resembling a notochord or neural tube ever dev eloped in either half and no structure resembling a tail was ever formed. In fact these half embryos produced by cutting the early gastrulz in two were altogether like the anterior and posterior half embry os which I have already described. (c7. Figs. 77-82 and Figs. 47-58.) These results were so defines nal conclusive that I did not continue the experiments and I regret now that I did not also cut gastrulz in two along the median plane, though there is no reason to doubt that the Peale would be the same as in cases where one of the first two blastomeres 1s killed. Comparing these results with those of Driesch, only one of two explanations is possible. Either Phallusia must differ most fundamentally from Cynthia, or Driesch must have mistaken the median for the transverse plane in these cup- shaped gastrulz. That the former possibility is not probable is evidenced by the fact that the cell-lineage of all ascidians so far studied is essen- tially the same; furthermore my results as to the development of anterior and posterior halves of the egg of Cynthia are confirmed by my experiments on Molgula, as well as by Chabry’s experi- ments on Ascidia aspersa. There is ev ery reason to believe that what is true of these three genera is also true of Phallusia. On the other hand there are certain evidences that Driesch may have mistaken the transverse plane for the median; on p. 56 he says, “ Aber auch an der Bechergastrula kann man die kunftige Mediane und also auch die Hauptrichtungen senkrecht zu ihr unterschieaeel es verlaufen namlich die Zelltheilungsgrenzen des Ektoderms dieser Objecte so, dass sie gerade in der Medianen eine uber die ganze Oberflache fortgesetzte nur sehr wenig gebrochene Ein- heitslinie bilden (S. z. B. Castle, Fig. 62, 71) welche ohne Weiteres ¢ . ve q - i Mosaic Development in Ascidian Eges. 197 foxe schon bei schwacher Vergrosserung kenntlich ist; schneidet man also in der Mitte und senkrecht zu dieser Linie, so zerlegt man auch die Bechergastrula in ‘vorn’ und ‘hinten. +f, It is true that the median plane is marked out by a nearly straight line, though Castle’s figures to which Driesch refers show this line between endoderm and not between ectoderm cells, but any one who has studied these embryos knows how difficult it is to determine the median plane in this way, especially in living material. Even in stained and mounted preparations it would not be a sure guide, much less could it be relied upon in the study of living gastrula. Whether the median plane appears as a straight line or not depends entirely upon whether that plane lies directly in the line of vision, and conversely some of the transverse planes of cleavage may appear as straight lines if they lie in the line of sight. This Fig. 7 shows several transverse rows of ecto- derm cells which in the hinder part of the embryo are curved back in the middle and forward at the sides, but if the embryo were rotated forward so that the polar body were brought to the highest point these transverse rows would appear nearly straight. I am convinced therefore that the half gastrulae from which Driesch obtained apparently normal larve were right or left halves and not anterior and posterior ones as he supposed. Whether these larve were really normal, 7. ¢., whether they had the organs of both the right and left sides, cannot be determined from Driesch’s figures or descriptions, since he seems to have considered that fe only evidence required to show that a larva is complete is that it should have a head and a tail. The fact that Driesch always obtained partial larvae from the anterior and posterior halves of an elongated gastrula, where the chief axis is unmistakable, requires no comment. IV. OTHER EXPERIMENTAL WORK ON THE ASCIDIAN EGG. Chabry’s (’87) contribution on the normal and teratological embryology of ascidians contains not only the most careful and complete experimental work which has ever been done on the ascidian egg but it is at the same time such an excellent analytical treatment of the normal development that it deserves to rank as an embryological classic. The experimental part of his paper was based upon an unusual knowledge of the normal and patho- 198 Edwin G. Conklin. logical development of this species and it was carried out with a delicacy and precision of method which has never been surpassed. Add to this the fact that the work was undertaken with clear insight into the principal problems involved and at a time when almost no other work of this sort had ever been done! and its right to rank as one of the great works in experimental embryology seems assured. Considering these facts it 1s surprising that this work should have received so little attention and that it should have been so widely misunderstood or discredited. Chabry’s extensive experiments deal with right and left half embryos, anterior and posterior two-quarter embryos, and various forms of three-quarter, one-quarter and two- quarter diagonal embryos, and in all of these I find that my results are in the main in accord with his. The points in which my work is more detailed than his concern the presence and distribution of the various odplasmic substances and the more accurate study of some of the later stages, made possible by the use of fixed and stained material. That the substance of the mesodermal crescent was seen by Chabry as early as the 32-cell stage is evident from his description of the mesoderm cells, which in Aecds aspersa are greenish (“‘verdatre’’) in color and which he recognized when only chee were present on each side. Neither Driesch nor Crampton speak of having observed any of these odplasmic substances and neither of them studied the later stages by means of fixed and stained material. 1. Cleavage. Chabry showed that in rhythm of cleavage and in the size and character of the daughter cells the isolated Beemer of Ascidia behave as if they were still part of the normal ege, while he described in great detail the changes which take place in the facets between cells, Crampton’s co nclasione are very similar; he found that “an isolated blastomere of the Molgula ege segments as if still forming a corresponding part of an entire embryo. ‘The cleavage phenomena are strictly partial, as regards the origin of cells, the inclination of cleavage planes, and especially in respect to the thythm of segmentation.” Driesch, on the other hand, found in Phallusia that there was no fixed relation between the iSee Roux, Ges. Abhand I, p. 958. ~ le eet, + : : Mosaic Development in Ascidian Eggs. 199 cleavage planes of the surviving half and the dead blastomere; that after the third cleavage the cells occupy very different posi- tions from the normal (Tetraeder, Halbtetraeder); that divisions may be equal or unequal at the fourth cleavage, and finally that the cleavage could not be regarded as partial (“halb”’) nor entire (“ ganz’ *) but “regellos-solid.”” “The evidence which Driesch brings in support of this conclusion 1s of little value since it is plain that he was unable to orient these cleavage forms and did not know from what part of the original egg they came nor from what pole they were viewed. My observations on the cleavage of isolated uninjured blastomeres of the egg of Cynthia “en iee and extend the conclusions of Chabry and Crampton that the cleavage of such blastomeres is unaltered save for slight changes in the direction of some of the divisions; they are opposed to the conclusions of Driesch that the cleavage of. such blastomeres is inconstant and irregular. 2. Gastrula. Chabry figures four gastrulz from isolated blastomeres, viz: his Figs. 108, 114, 129 and 130. O. Hertwig, who copies Fig. 129 in his book, “Die Zelle”’ (’98), says that it is a normal typical gastrula. Similarly Korschelt and Heider, who also copy this figure in their text-book (’02), affirm that it is a normal small gastrula. However, these authors bring no particle of evidence to the support of this bare assertion; Chabry himself nowhere says that any of the gastrulz figured by him are normal and the figures themselves do not show that such is the case. On the other hand I can positively affirm that a normal entire gastrula is never formed from an isolated blastomere of the egg of Cynthia. In the absence of any evidence in favor of Hertwig’s and Korschelt and Heider’s interpretation and in the face of this positive evidence against it I think it may safely be assumed that Chabry’s figures are not those of normal typical gastrule. Crampton expressly says that he did not carefully gieeee the process of gastrulation in the embryos derived from isolated blastomeres of the Molgula egg, but Driesch says that the process of gastrulation may be easily observed in Phallusia, that a typical ascidian gastrula is formed and that the closure of the blastopore takes place i in the normal manner. “Alles sind verkleinerte Aehnlichkeitsbilder der Processe an normalen Eiern, welche stets vergleichen wurden.”’ 200 Edwin G. Conklin. However, it is quite evident from the observations of Van Beneden and Julin, Chabry, Castle and many others that something more than a mere invagination 1s necessary to constitute a normal gastrula. The ascidian gastrula is bilaterally symmetrical and its anterior and posterior portions are very unlike; furthermore all the principal organs of the larva are here represented by cells of peculiar structure and localization. In order to determine whether a gastrula is normal or not all of these features have to be considered, and this Driesch has not done. 3- Larva. It is somewhat surprising that doubt should have been expressed as to whether Chabry obtained half embryos or whole embryos of half size from one-half of the ascidian egg. He again and again declares that lesion of a single cell up to ‘the 16-cell and probably up to the 32-cell stage always causes a “hemiterie,” or monster. (Chabry, pp. 246, 249, 250, 257, 258, ae etc.) He even enters into a calculation of the number of kinds of monsters which may be produced by injuries to the cleavage cells. He says that if at the 8-cell stage each cell is capable of four different kinds of modifiea- tion (certainly less than the reality), the number of modalities of this stage is 4° (= 65536) of which only one is normal. In this way there arises that “admirable and infinite v ariety of monsters”’ to which he repeatedly refers. He says expressly, p. 289, “De la on tire aisément la conclusion (que je ne crois valable que pour l Ascidie et les animaux, dont les b!astoméres sont différenciés de bonne heure), que chaque blastomeére contient en puissance cer- taines parties dont sa mort entraine la perte irrémédiable et que les différentes parties de l’animal sont préformées dans les diftér- entes parties de l’oeuf.’’ Again on p. 299 he says, “On ne saurait donc conclure avec sécurité de l’oeuf d’Ascidie a celui des autres animaux, mais, en ce qui concerne celui-ci, il est exact de dire qu'il se comporte comme s’il contenait en puissance un seul adulte déterminé et que chaque partie de |’oeuf contint une partie de cet adulte.”” This same conclusion is repeated again and again so that as Barfurth (’93) and Driesch (’95) have said there can be no. question as to what Chabry believed that his observations and experiments proved. ied tio Mosaic Development in Ascidian Eggs. 201 The statements of Driesch and Crampton are even more posi- tive and explicit that whole larvae are formed from any one or more of the first four blastomeres. Driesch (p. 405 in sum- marizing his results uses, in part, the very words of his Ae ae regarding the value of the cleavage cells in the echinoderm eg “Aus isolirt tberlebenden Blastomeren des Ascidieneies ae wickelt sich nicht ein halber (resp. viertel, drei viertel) rechter oder linker (resp. vorderer oder hinterer) Embryo, sondern stets ein ganzer von halber Grosse, dem allerdings (meist) gewisse Organe von minderen Bedeutung (Otolith, ein Haftorgan) fehlen.’’ Crampton neither figures nor describes the larve obtained from isolated blastomeres rok Molgu'a, but he says, p. 55, “Enough of the later development has been ascertained o prove that a eres arises wh ch resembles the normal larva, except as regards its smaller size and certain minor defects. My results, thereto: e, are entirely confirmato y of those of Driesch upon Phallusia.”’ Chabry fist discovered that larve from one of the first two blastomeres were superficially like normal larve in that they had head and tail, notochord, neural plate and sense spots, but he showed that they also lacked the organs distinctive of the missing side, viz: one papilla, one or more sense spots and one aie invagination. It ‘s surprising therefore that neither Driesch nor Crampton undertook to prove that the larve obtained by them from one of the first two blastomeres were really complete. One looks in vain in their papers for any evidence that the organs characteristic of that side which would have developed from the dead half (muscles, mesenchyme, papilla, atrial inv agination) are present in the surviving half. Chabry further showed that the type of embryo derived from the anterior or posterior two-quarters of the egg@ was very unlike that derived from the right or left two-quarters, while the one- quarter embryos were still more unlike the normal; n each of these cases he found that the development was strictly partial, only those parts arising from a blastomere which would develop from it in the normal “embryo. In the face of these conclusions of Chabry’s neither Driesch nor Crampton advance any evidence in favor of their claim that the anterior and posterior ‘quadrants of _the egg as well as the right or left may give rise to a larva. Cha- bry’s figures and descriptions show plainly what my work proves that nothing even remotely resembling a normal larva is ever pro- 202 Edwin G. Conklin. duced from any portion of an egg which does not include the whole of the right or left half. In my opinion Driesch and Crampton have not a eaied nor taken any account of anterior or posterior half embryos, but only of right or left ones. The question whether these embryos were actually complete will be considered when we come to deal with the various larval organs. Both Driesch and Crampton make the claim that single blasto- meres of the 4- cell stage of the ascidian egg may give rise to entire larvze.° Phisias a caucial test of their views, for while it is possible and | believe practically certain that all their “complete larvz of half size”? were derived from the right or left halves of the egg and so included portions of all the various odplasmic substances, this explanation could not apply to their quarter embryos. Driesch figures a larva with all the principal organs (his Fig. 16), which he says is derived from one of the first four blastomeres. How- ever, in size it is as large as any of the half larvz which he figures, and I have no doubt that it is such. Crampton figures correctly the early cleavages of one of the anterior quadrants and he gives two figures of quarter larvae, probably of an advanced stage; these figures, however, show no structure whatever save that there is an outer layer around the embryo. ‘There is absolutely no evidence that these embryos are complete. Crampton calls attention to the fact that the long axes of these quarter embryos “are approximately parallel to the principal dorso-ventral axis of the original egg,” a fact which I alsocanconfirm. (See my Figs. 66,69, 70.) He does not, however, determine the fact, which he apparently assumes, that the long axes of these quarter embryos correspond to the long axis of a normal embryo. ‘This is actually not true, as | have shown; the long axes of the quarter embryos are not antero-posterior in direction but dorso-ventral and there has not therefore been any shifting of the axes nor of the odplasmic substances of these quarter embryos. Whether a larva derived from the right or left half of the egg is complete or not can be determined only by a study of the various systems of larval organs. It is evident that parts of all organs which are normally famed along the median plane (first cleavage plane) would appear in an embryo derived from one of the first two cleavage cells, even if the dev elopment were strictly partial; the really decisive test as to whether such an embryo is complete Mosaic Development in Ascidian Eggs. 203 or not must be found in the study of those organs which do not lie along the median plane. a. Neural Plate and Sense Organs. Chabry says that he never saw a partial embryo in which the neural plate had invaginated ; on the contrary the nervous system always remains spread out in the form of a layer or plate; this plate occupies the face of the embryo which is morphologically median in position (its normal location), while the sense spots consist of pigmented cells which are superficial in position and which lie near the base of the tail. ‘This agrees very closely with my observations, though I have frequently seen the neural plate invaginate by an irregular process. ‘The eye is said by Chabry to be formed on the right side normally, but the fact that it may appear in the left half embryo leads him to conclude that its rudi- ment exists in the left half of the egg also. He thinks that the otolith comes only from the right posterior cell. JI have not determined the exact cell origin of the sense organs in the normal larva, but in the partial larve they are formed only from the anterior quadrants and from either the right or left sides. 1 have not been able to distinguish between the eye and the otolith in the partial embryos of Cynthia. Driesch says nothing of the neural plate nor of the manner in which the nervous system is formed in his small larve, though he mentions the fact that ‘the sense vesicle with the eye and otolith are not formed in the typically clear manner characteristic of the normal development.” He found the eye spot almost always present, the otolith very seldom and he concludes that it makes no difference in the presence or absence of the sense organs w hether the embryo has developed from certain cells of the 4-cell stage rather than from others. Since Driesch expressly states that He never raised a quarter embryo beyond the stage of his Fig. 16, at which stage the sense organs have not appeared, and since ecither his figures nor descriptions give any evidence that he has distin- guished anterior or posterior quadrants from right or left ones, it would be interesting to know how he could dercemine that sense organs might be formed from any quadrant of the egg—a result entirely contrary to my observations. 204 Edwin G. Conklin. b. Notochord. Chabry supposed that the notochord arose from both the anterior and posterior quadrants of the egg. Castle (’96) held that a single pair of cells of the posterior quadrants, B**, ‘the posterior enone fundament,”’ were the only cells of the posterior quadrants which entered into the formation of the notochord. I am of the opinion that this cell is a mesenchyme and not a chorda cell (see Conklin, ’05'), but even if it should be found to be a chorda cell it is only one cell of nine on each side of the mid line which give rise to that structure, while eight ‘pairs of chorda cells come from the anterior quadrants. Certain it is that no trace of a notochord ever arises from the posterior cells when they are isolated, whereas chorda cells always arise from isolated anterior cells, though a notochord is rarely formed in such cases. Chabry describes (p. 294, Fig. 118) an anterior two-quarters embryo in which a naked chorda was seen in the perivitelline space outside the body of the embryo; such a case somewhat resembles the one shown in my Fig. 72. However, in every other instance which I have observed the chorda cells of an anterior embryo do not give rise to a notochord, but after escaping from the body of the embryo lie free in the perivitelline space as scattered cells. (Figs: 625 66— foe) But while a notochord is rarely or perhaps never formed in an anterior embryo and never in a posterior one, it is invariably found ina right or left one, and the figures of Chabry and Driesch as well as my own show that the process of formation is essentially the same as in a normal embryo. Chabry indeed believed that the notochord was primitively double and that half of it arose from each lateral half of the egg. He speaks of the fact that in Ascidia and Botryllus it is composed of a double row of cells and Crampton also refers to the fact that in the normal ascidian tad- pole there are two rows of chorda cells, whereas Driesch has well said that in its fully formed condition the ascidian notochord is normally composed ofa single row of cells. I find, as did Driesch, that the notochord of a iafeeal embryo is formed by interdigitation, just as in the normal embryo, but [ also find, as opposed to Driesch that the notochord is never formed from any cells save the chorda cells which come from the posterior part of the gray crescent. Furthermore, my observations show, as did Chabry’s, that the Mosaic Development in Ascidian E gg. 205 formation of a tail is dependent upon the development of a noto- chord. c. Muscles and Mesenchyme. Chabry paid no particular attention to the number and location of the muscle cells in his partial larvae, though he frequently speaks of their presence as being proved by the ‘twitchings of the tail; these movements are less energetic than in normal larve and, as a consequence, partial larve do not escape from the egg mem- branes. Driesch also found that partial larve rarely hatch, probably because of their weak muscular movements, but he, too, paid no attention to the number and position of the muscle cells. Owing to the brilliant color of these cells in Cynthia they are recognizable at all stages; in the partial larve they are found only along one side of the notochord, where they form the characteristic three rows of cells, whereas the muscle cells of the opposite side are entirely lacking. In the oldest larve a few of the muscle cells extend around the end of the notochord to the side on which they were lacking. [ have not been able to determine whether the num- ber of muscle cells is actually increased during this process or merely rearranged, but | believe that the mohole process consists in the moving oe certain cells over to the side on which they were lacking, without any increase in their number. ‘This is part of that process of regulation which begins with the rounding up of the surviving lastsmerc after the other one has been billed: In fact, this very extension of the muscle cells around the end of the notochord begins in this rounding up of the surviving blastomere and in that slight change in the direction of division which causes the median cells of the yellow crescent to ae nearer the middle of the first cleavage plane than in the normal egg. (Fig. 15.) Chabry found (p. 308, Fig. 132) only one atrial inv agination and one organ of fixation (papilla) i in right or left half embry Os. Driesch did not determine the number of atrial inv aginations but he does call attention to the fact that but one papilla i is present in embryos from isolated blastomeres. I have not observed the formation of the atrial invaginations or of the papillz in Cy nthia; even in the normal larvee they are inconspicuous at the time of the metamorphosis and I have not studied them before that period. However, the areas of trunk mesenchyme in which the atrial invaginations appear, are conspicuous areas of clear, slightly 206 Edwin G. Conklin. yellow, protoplasm in front of the muscle rows on each side of the tail; these areas may be recognized in the early cleavage stages mad ques case are both these mesenchyme areas present in right or left half embryos. It is almost certain, therefore, that only one atrial invagination is formed in such embryos. We find, therefore, that those parts of the larva which normally lie on the right side are missing in a left half embryo and those which normally lie on the left side are not found in a right half embryo, whereas unpaired organs which lie along the median plane are represented in Boek lateral half embryos. This is exactly what might be expected from a study of the organization of the egg since the substances, which give rise to median organs, are found along the median plane in both right and left blasto- meres, whereas the materials which give rise to organs of the right side are found only in the right blastomere, chose’ which give rise to organs of the left side, in corresponding positions in the left blastomere. Neither Driesch nor Crampton attempt to show that a larva from the right half of an egg has the organs of the left side and this is the whole question at issue; if it does have these organs it is a complete embryo; if it lacks them it is a partial embryo, even if it does have a head and a tail. Chabry found that a larva from one of the first two blastomeres had a head and tail and median organs, but that it did not have the organs of the missing side and this conclusion I can entirely confirm. All of the muscle substance (myoplasm) and most of the mesen- chyme (chymoplasm) is localized in the posterior half of the egg, and corresponding to this distribution we find that an anterior half embryo entirely lacks muscles, though it may have a small amount of mesenchyme (that derived fa the cell A**), whereas a posterior half embryo contains a large number (probably the full normal number) of muscle cells and most of the mesenchyme. V. REGULATION IN THE ASCIDIAN EGG AND EMBRYO. It is well known from the work of Loeb (’92) and L. Schultze (‘99) that the brain of Ciona will be regenerated when extirpated in the adult animal, and that the siphons will be restored when they are cut off. Driesch (’02) has also shown that Clavellina has extraordinary powers of regenerating almost all lost parts. Mosatc Development in Ascidian Eggs. 207 This power of regeneration in the adult is in striking contrast with its lack in the egg and embryo and requires some explanation. It should not be overlooked that such injuries to the egg and embryo as have been described in the preceding pages are prob- ably more extensive and far-reaching than any which are capable of being repaired in the adult. As Chabry says the destruction of one of the first two blastomeres is the same in its effect as the destruction of the right or left half of the body of an adult. The destruction of the anterior half of the egg is similar to the total.loss of the nervous system and notochord of the larva; while the death of the posterior half corresponds to the destruction of the whole of the muscular system and most of the mesenchyme of the larva, since in each case the specific substance which alone gives rise to these organs is destroyed. ‘Therefore these injuries are probably much more extensive than any which have been practiced on the adult animal. Furthermore, I am of the opinion that the extremely rapid development of the ascidian egg and embryo may itself act as a check on regulation. In Cynthia and Ciona the fully formed larval stage is reached in about twelve hours after the fertilization of the egg, and these larva usually undergo metamorphosis into the adult form within the next twelve hours. In Molgula the development is even more rapid. It seems to me probable that the restoration of the parts of the missing right or left half of a larva might be fully accomplished if the larval life were longer. In a right or left half larva one day old the ectoderm cells have closed over the injured side, the notochord is complete, the neural plate has invaginated, although abnormally, and the muscle cells have begun to grow over from the uninjured to the injured side. There is here evidence of considerable regulative ability and it seems to me possible that, with more time before the metamor- phosis, complete rows of muscle cells might be found on both sides of the tail and that the mesenchmye cells might grow over to the side on which they are lacking and an atrial invagination appear in them. Inasmuch as the only form of regulation shown by the ascidian egg or embryo is this overgrowth of cells from the uninjured to the injured side, it is probable that no amount of time would ever suffice to produce an entire larva from the anterior or posterior half of an egg or from a quarter or any smaller portion. As a 208 Edwin G. Conklin. matter of fact there is not the slightest indication in an anterior half embryo of any attempt to restore the missing myoplasm or muscle cells, nor does a posterior half embryo show any tendency to form chorda-neuroplasm or neural plate or chorda cells. So far as observation and experiment show, each ooplasmic substance is capable of giving rise only to one particular kind of organ or tissue. ‘The question may be raised whether the presence of the injured blastomere within the chorion may not influence the development of the surviving cells and possibly prevent regeneration. In this and in all previous experimental work on che ascidian egg these injured cells have been left within the chorion in contact with the surviving cells and in this respect all work on these eggs has been done under similar conditions. Owing to the presence of the chorion it is practically impossible to remove the injured cells, and I am therefore unable to furnish an experimental test of the influence or lack of influence of these cells upon the surviving ones. However, there 1s sufficient evidence, I think, to show that it is not the presence of these cells which prevents regeneration. Contact with the injured cell might be expected to hinder or pre- vent the closing of the surviv ing half along the injured side, but it is just this aoe of regulation, a this only, which is manifested by these eggs. The presence of the injured igen: nothing to do with the failure of the anterior half embryo to form a tail, or the posterior half embryo, a head; on the other hand, I have shown conclusively that the development of a tail is dependent upon the presence of a notochord, and the formation of a head upon the presence of the gastral endoderm and neural plate. The only possible influence of the injured cell upon the surviving one would be to limit the form- regulation; but as I have said this it does not do. It is inconceivable that the presence of the injured cell should prevent the myoplasm from giving rise to other organs than muscles, or the chorda- -neuroplasm to other organs than chorda and neural plate. These injured cells are rarely killed, but they remain transparent and entire, although quiescent; they do not decay and form a nidus for bacteria and T am convinced that their presence does not materially influence the development of the surviving half nor limit its powers of regulation. . { j Mosaic Development in Ascidian Eggs. 209 VI. GENERAL CONCLUSIONS. The conclusions which follow from these experiments are so obvious that they need but little emphasis here. Not only is the fact established that individual blastomeres give rise only to those parts of an embryo which they would produce under normal conditions, but the cause of this is clearly indicated. The devel- opment of the ascidian egg is a mosaic work because individual blastomeres are composed of different kinds of odplasmic material; this mosaic work is not merely a cleavage mosaic but also a mosaic of germinal substances, several of which are recognizable before cleavage begins. 1. Organ-Forming Substances. I have elsewhere shown that at least five distinct kinds of ooplasm are recognizable in the egg of Cynthia before the first cleavage and that all of these substances are localized in their final positions as early as the close of that cleavage. In these experiments | have not been able to isolate the different odplasmic substances in the unsegmented egg, but after the second or third cleavages several of these substances may be isolated and in such cases each substance gives rise only to a definite kind of tissue or organ, and apparently it has no power to produce any other kind. The myoplasm produces muscle cells only; the chorda-neuro- plasm, only chorda and neural plate cells; the chymoplasm, only mesenchyme; the endoplasm and ectoplasm only endoderm and ectoderm, respectively. Whenever an isolated blastomere lacks any of these substances, the embryo which develops from that blastomere lacks the corresponding organs. Accordingly the potencies of individual blastomeres are dependent upon the ‘ooplas- mic substances which they contain; the prospective value of any blastomere is not primarily a function of its position, but rather of its material substance. _ The reason that the anterior quadrants of the egg never produce muscle cells is evidently due to the fact that they orally lack the yellow myoplasm; the fact that the posterior quadrants never produce a neural plate or chorda, is evidently due to the complete absence of the chorda-neuroplasm in these quadrants; the cells of the ventral (animal) pole produce only ectoderm, without a trace of endoderm or mesoderm,—evidently because these cells are composed almost entirely of clear ectoplasm. 210 Edwin G. Conklin. Experiment confirms, therefore, what observation of the normal development plainly indicates that these strikingly dijjerent odplas- mic substances are not toti potent, but that as early as the close of the jirst cleavage and probably much earlier, they are differentiated jor particular ends, and that if they develop at all they give rise to organs of a particular kind. These materials are, therefore, “organ- forming substances’ > and the areas of the egg in which they are localized are “organ -forming regions. I need not here point out the similarity between this conclusion and the well-known theories of Sachs and His, nor the differences between my results and the commonly accepted view that the egg is composed of ‘simple undifferentiated protoplasm” or that ‘“‘cleavage is a mere sundering of homogeneous materials capable of any fate,” or that “the prospective value of a blastomere is a function of its position.”” Whatever may be true of other animals these things are certainly not true of ascidians. While there are few, if any, other cases known in which the differentiations of the ooplasm are so striking or so numerous as in the egg of Cynthia there can be no doubt that organ-forming sub- stances are present in the eggs of many animals. In particular the works of Fischel (’97, ’98, ’03) on the Ctenophore, of Boveri (01) on Strongylocentrotus; of Wilson (’04) on Dentalium and Patella and of Conklin (’03) on Physa, Planorbis and Limnza have shown that distinct kinds of protoplasm are present in these eggs which are destined in the course of development to-give rise to particular germ layers or organs. In the light of these discoy- eries it can scarcely be doubted that the general cause of mosaic development is to be found in the presence in the egg or blasto- meres of distinct kinds of protoplasm, or of organ-forming substances. 2. Localization of Oéplasmic Substances. The three principal substances in the egg of Cynthia, viz: the clear, the yellow and the gray, are already present and localized in the oocyte before it escapes from the ovary. The yellow (mesoplasm) forms a peripheral layer around the entire egg; the clear (ectoplasm) is the clear achromatic substance within the germinal vesicle; the gray (endoplasm) constitutes most of the Mosaic Development in Ascidian Eggs. DTT oS remainder of the egg.' For the sake of brevity this earliest form of localization may be described as concentric or spherical, although the germinal vesicle does not lie exactly in the center of the egg but is slightly eccentric toward one pole. During maturation and fertilization this concentric localization gives place to a polar or radial form. Immediately after the entrance of the spermatozoon into the egg the peripheral layer of yellow mesoplasm flows rapidly to the lower pole where it collects in the form of a cap; the clear ectoplasm which escapes from the germinal vesicle at first lies at the animal pole where it surrounds the maturation spindles but after the entrance of the spermatozoon it also flows to the lower pole where it collects into a layer or stratum just above the mesoplasm; the gray endoplasm after these movements occupies almost all of the upper half of the egg. The egg at this stage appears to be radially symmetrical, the three principal substances being arranged in strata at right angles to the egg axis. Soon after the entrance of the spermatozoon this radial form of localization gives place to a bilateral one; the sperm nucleus and aster move up to the equator of the egg along one meridian which further development shows to be the median plane on the posterior side; the clear and yellow substances also move to the posterior pole along with the sperm nucleus and the yellow substance here forms a crescent around the posterior side of the egg, just below the equator. At this stage the egg is bilaterally symmetrical, there being but one plane which will divide equally all of the ooplasmic aro onecs Finally during the first cleavage this early bilateral localization is changed into the definitive localization which 1s characteristic of all stages up to the late gastrula. ‘The yellow crescent remains in the position which it occupied before the first cleavage and here gives rise to muscle and mesenchyme cells; the clear “protoplasm comes to occupy most of the ventral hemisphere and gives rise to =) . . ectoderm; the gray substance occupies the dorsal hemisphere in Although I have not been able to isolate these various odplasmic substances before cleavage begins and, therefore, can bring no experimental evidence to prove that they are organ-forming substances at this early stage, it nevertheless seems probable that materials which are identical in color and texture with the organ-forming substances of later stages, to which they directly give rise, are also similar in potency. There is no apparent reason for believing that these strikingly different kinds of odplasm of the ovarian egg are any less distinct or more nearly totipotent than during the cleavage stages. PHP? Edwin G. Conklin. front of the yellow crescent and its anterior portion becomes the gray crescent of chorda- neuroplasm, while its posterior portion is the deep gray endoplasm which gives rise to the gastral endoderm. The form of localization of these substances, therefore, undergoes marked changes during the fertilization and first cleavage; it 1s concentric in “the odcyte, polar or radial immediately after the entrance of the sperm, bilateral just before the first cleavage, and definitive at the close of the first cleavage. I have elsewhere (’05') shown reason for believing that even in the stage of radial localization in the egg of Cynthia there is prob- ably some structural peculiarity of the egg which determines that the path of the sperm shall lie in one meridian rather than in another and therefore that the median plane of the embryo and its posterior pole are not bears by the chance movements of the sperm within the eg Similarly the basis for polar or radial localization 1s eae in the ovarian egg in the slight eccentricity of the germinal vesicle toward the sitter pole, though the ooplas- mic abaeances are largely localized in concentric form at this stage. [am unable to determine whether any structural basis for bilateral localization exists in the ovarian eggs of ascidians, but inasmuch as the localization invariably becomes bilateral at a later stage it seems necessary to suppose that there is some such intrinsic determinative factor. In almost every group of animals the chief axis of the egg is already marked out in the odcyte, the pole toward which the ger- minal vesicle is eccentric becoming later the animal pole of the egg and the ectodermal pole of the embryo. Despite this eccen- tricity of the germinal vesicle the localization of odplasmic sub- stances in the oocyte of ctenophores, nemertines, echinoderms and ascidians is chiefly concentric, the polar localization of these substances first appearing during the maturation and fertilization. On the other hand Wilson (’ 04) has found a markedly polar localization of the ooplasm 1 in the odcyte of Dentalium; while it is probable that in the odcytes of insects and cephalopods the local- ization is bilateral in form. Boveri (’01) found that distortion of the egg of Strongylocen- trotus after the formation of the equatorial zone produced no change in the polar stratification of the egg nor in the potencies of its different substances. Wilson (’03), “Yatsu (04) and Zeleny (04) have discovered that fragments from any part of the egg of Mosatc Development in Ascidian Eg ggs. 213 Cerebratulus before maturation may give rise to entire larve; whereas this is not usually the case after maturation and fertiliza- tion, the potencies of the substances at the animal and vegetal poles being different. It is evident that during the concentric stage of localization section of an egg in any plane would leave samples of all the ooplasmic substances in each piece; in the stage of polar-radial localization any section of the egg parallel with He egg axis would leave samples of all the odplasmic substances in each piece; in the bilateral stage, only the right or left halves would contain parts of all the substances. Probably one important rea- son why parts of eggs may give rise to whole embryos in some cases and not in mriers may be found in the fact that at the time of the experiment the form of localization may have been concen- tric in some cases, radial in others and bilateral in still others. (See Boveri, ’01; Wilson, ’04".) 3. Cleavage and Localization. In previous publications (’05', ’05*) I have pointed out the fact that localization precedes cleavage in the ascidian egg and that the localization pattern does not closely coincide wrth: the cleavage pattern. On the other hand there 1s normally a constant relation between particular cleavage planes and the various odplasmic substances. The first cleavage always lies in the median plane and bisects all the odplasmic neces the second is transverse to the median plane and separates the yellow crescent from the gray one; the third cleavage is at right angles to the two preceding ones and separates the clear ectoplasm of the ventral hemisphere from the different substances of the dorsal hemisphere. Probably in no other animal is the cleavage so constant and so perfectly bilateral as in the ascidians and yet even here the position and direction of the cleavage planes is less constant than the form of localization. Among annelids and mollusks, as is well known, the cleavage 1s typically spiral and in many cases it is radially sy mmetrical. This radial symmetry of cleav age does not indicate, however, that the localization of odplasmic Saenenee is also radially symmet- trical, for in some cases this localization is known 2 be bilateral and this is probably true in all cases. (See Conklin, ’05', pp. 90-92.) The relation of the cleavage planes to this esa organization 214 Edwin G. Conklin. is very different in cases of spiral and of bilateral cleavage, and consequently the results of killing any one or more of the first four blastomeres may vary in deen cases; in general there is less likelihood of obtaining an entire embryo from an isolated blasto- mere of spiral cleavage than from one of the first two blastomeres in bilateral cleavage. In other cases the cleavage planes bear no constant relation to the planes of localization. ‘Thus in the frog’s egg the first cleavage may lie in the median plane or at varying angles to this plane and Brachet (04) has recently shown that the character of an embryo derived from one of the first two blastomeres depends entirely upon the relation between the first cleavage plane and the median plane of organization. It is probable that the bilaterality of organization is no more perfect in ascidians than in annelids, mollusks or amphibians, but the bilaterality of cleavage is much more perfect. Accordingly, each of the first two Segeonses of the ascidian egg always con- tains half of ev ery odplasmic substance, in the frog’ s egg it may or may not contain half of these substances, in the annelid or mollusk it never does. I agree therefore with Brachet (’04) and Wilson (041, ’04?) that the varying results of experiments on the potencies of blasto- meres are due in part to the varying relations of cleavage to local- ization, and in part also to the different types of localization (concentric, radial, bilateral) in different eggs. 4. Determinate and Indeterminate Cleavage and Development. In a great many animals belonging to phyla as widely separate as Ctenophora, Poly clada, Nemertinea, Nematoda, Rotifera, Annelida, Mollusca, Arthropoda and Tunicata the cleavage of the egg is constant in form and differential in character and under normal conditions, definite cleavage cells always give rise to defin- ite structures of the embryo or larva. For this type of cleavage I proposed several years ago (97, 98) the designation “ deter- minate.” In a few animals the cleavage is known to be extremely irregular, as in Pennaria (Hargitt, ‘04), Renilla (Wilson, *84), and probably also in planarians (Hallez, ’87; Stevens, ’o4), while in other cases it is unknown whether the cleavage is normally con- stant and differential or not (Echinoderms); in still other cases —————— Mosaic Development in Ascidian Eggs. 215 the planes of cleavage bear no constant relation to the planes of localization, as in the eggs of some of the vertebrates (frog, fish). For all such cleavages I proposed the name « sideterminate, ’ but at the same time I was careful to state that this was “‘to be under- stood as applying only to the cleavage, for in its main features and results the development of all animals is determinate; that is, predictable. Even in cnidaria, echinoderms and _ vertebrates there appears successively a blastula, gastrula, larva, and adult of determinate form and character” (’98, p. 21). But while the cleavage is indeterminate in some cases there is reason to suppose that there is a definite organization of the egg in all animals—in short that the organization of the individual is determinate at all stages from the egg to the adult. Even in such an egg as that of Pennaria it is certain that there must be deter- minative factors somewhere, if not in the cy toplasm then in the nucleus, which determine that the egg shall develop into a Pennaria rather than into some other animal; and it is further evident that . these determinative factors must be present in the cytoplasm at a relatively early stage, if not at the very beginning of dev elopment. In the echinoderm egg, which was at one time supposed to be homogeneous or isotropic, Boveri (’01) has shown that a polar- radial localization of at least three distinct morphogenetic sub- stances takes place immediately after maturation, and in this case, as in the ascidians it is probable that there is an earlier concentric localization of these substances in the odcyte. Since these three substances are localized in zones or strata, one above the other, around the chief axis of the egg, they are all present in each of the first four blastomeres of the egg, each of which may give rise to an entire embryo; but when they are isolated each is found to be strictly limited in its potentialities. While therefore there are several groups of animals in which the cleavage is indeterminate there are few or none in which the ooplasm is isotropic; on the contrary in almost every phylum the eggs and blastomeres show differentiations and localizations of the ooplasm which are of morphogenetic value. “Everywhere,” as Fischel (03) has well said, “the fundamental principle of normal development 1 is a mosaic work.”’ But while Fischel supposes that “only the materials for the primitive organs of the embryo are preformed in the egg cell and that the material substratum for the differentiation of the special organs is probably first formed during 216 Edwin G. Conklin. the later stages,” I find that in the ascidian egg all the principal organs of the larva are represented by distinct organ-forming substances which are localized in their definitive positions and proportions < as early as the close of the first cleavage. There is a world-wide difference between such results as these and those which were reached by Driesch and some of the earlier workers in this field. Wilson (’04) has recently expressed the opinion that “had the experimental analysis of cleavage been first undertaken in the case of such a determinate type as that of the gasteropod or annelid and had Roux not handicapped his theory with a purely speculative hypothesis of differentiation, which proved to be untenable, the whole discussion would have taken a different course; and | believe it would from the first have been recognized that the mosaic principle holds true in greater or less degree for every type of dev elopment, not excepting the most ‘indeterminate’ forms of cleav age.’’ Considering the fact that such highly determinate forms as the ascidian and the cteno- phore were studied in some of the earliest experiments on the potency of cleavage cells, | am of the opinion that the course which this discussion took was not primarily due to the fact that work began on relatively indeterminate forms. On the other hand I am convinced that the whole trend of opinion on the organization of the egg and on the potency of cleavage cells would have been A perene ie those who did this work had been more familiar with the normal development of the forms studied, and in their zeal for the experimental method had not discarded the old and approved method of observation. It has taken such careful observers of normal processes as Boveri and Wilson to apply most successfully the experimental method to the problem of the organization of the egg, and the results of such work constitute a well-deserved tribute to the permanent value of the work of Roux. SUMMARY. I. Normal Development. 1. Inthe ovarian egg of Cynthia (Styela) partita there are three strikingly different kinds of odplasm, viz: a superficial yellow layer, a central. gray area, and a large transparent germinal vesicle. At this stage the localization of these. substances is approximately concentric. ——— Mosaic Development in Ascidian Eggs. 5 2. During maturation and fertilization the yellow substance flows rapidly to the vegetal pole where it forms a superficial layer or cap; the clear substance escapes from the germinal vesicle and flows toward the vegetal pole where it forms a stratum above the yellow cap; the gray substance occupies the animal half of the egg. The localization at this stage is polar-radial. 3. The sperm nucleus which les in the center of the yellow cap moves to the hae pole of the egg and the yellow and clear substances move with it. The yellow material here forms a crescent which lies with its center at the posterior pole and its arms extending forward on each side about halfway around the egg; the clear substance forms a band just above and scents to the crescent; the gray substance occupies the remainder of the eg At this stage the igo ons is bilateral. 4. The first cleavage furrow appears in the plane of bilateral symmetry and divides each of the odplasmic substances equally. At the close of this cleavage the clear substance occupies the ani- mal (ventral) half of the egg; the gray substance lies at the middle of the vegetal (dorsal) pole while around the posterior border of the dorsal hemisphere 1 is the yellow crescent and around its anterior border is a light gray crescent. ‘his is the definitive localization of these substances, and in these positions the clear material gives tise to ectoderm, the gray to endoderm, the yellow crescent to muscles and mesenchyme, and the gray crescent to chorda and neural plate. 5. Ihe second cleavage is transverse to the antero- posterior axis and separates the gray crescent from the yellow; the third cleavage separates the clear protoplasm of the ventral hemisphere from the various substances of the dorsal hemisphere. Te: Experiments. 6. Individual blastomeres were injured by spurting or shaking the eggs in the 2, 4, 8, or 16-cell stages. “The surviving blastomeres were then studied both in the living condition and after being stained and mounted. 7. Cleavage. Isolated blastomeres always segment as if they still formed part of the whole, except that the direction of some of the cleavages is slightly altered so that the resulting cell mass is more nearly spherical than in the normal egg. T hese alterations 218 Edwin G. Conklin. in the direction of cleavage and the consequent closing of the injured side are more apparent in isolated blastomeres of the 2-cell stage than in those of later stages. 8. Gastrulation. In right or left or anterior halves, gastrula- tion usually proceeds as if the fragment still formed part of the whole; even though the gastrula may be rounded in form the location of the different substances shows that it is strictly partial. Not infrequently isolated blastomeres give rise to exogastrule, which ultimately right themselves. In posterior halves and in quarter embryos, gastrulation does not at all resemble the normal process, either in methods or results. Q. Right or Lejt Half Embryos. A lateral half embryo is usually closed along the injured side; it has a head and a tail; a ty pical notochord, Se is formed only from the chorda cells of the surviv ing side, and which is therefore composed of half the normal number of cells; an atypical neural plate and sense vesicle, formed only from the typical neural plate cells of the surviving side; a typical mesenchyme area in which the atrial invagination of one side is formed and three typical rows of muscle cells on one side of the notochord, but none along the injured side. In the latest stages to which these lateral embryos were reared (corre- sponding to the period of metamorphosis in normal larve) the muscle cells have begun to grow around the hinder end of the noto- chord to the side on a ee they were lacking; but in no case are the three rows of the normal embryo present on this side. Prob- ably only one atrial invagination and one papilla are ever formed in these lateral embryos. These are therefore half embryos in which some cells have grown over from the uninjured to the injured side, but in which absoliitels no change has taken place in the potency of the individual cells or of the different odplasmic sub- stances. 10. Anterior Half Embryos. Embryos derived from the two anterior quadrants of the egg have no trace of muscle cells nor of muscle substance; although the normal number of chorda cells are present they rarely if ever form a notochord but usually escape from the body of the embryo and lie scattered in the perivitelline space; the neural plate cells are present in normal number and position but the plate rarely, if ever, invaginates to form a sense vesicle; in late stages sense spots are formed from certain cells of the neural plate; cells of the gastral endoderm and general ecto- - 1 : : i ) Mosatc Development in Ascidian Eggs. 219 derm are frequently present in their normal positions and num- bers. A tail is never formed in these anterior embryos and they bear no resemblance whatever to a normal larva. 11. Posterior Half Embryos. Embryos derived from the two posterior quadrants have no trace of notochord or of chorda cells, neural plate, sense vesicle, sense spots, or gastral endoderm; they contain a mass of muscle and mesenchyme cells and a double row of caudal endoderm cells, as in the normal embryo. ‘There is no indication of a tail or head, the embryo remaining rounded in form as long as it lives. 12. Three-Quarter Embryos. Embryos derived from three of the first four blastomeres are more nearly perfect than are half embryos, but they always show defects corresponding to the missing blastomere. If an anterior blastomere is killed, the neural plate and sense vesicle of the resulting larva are atypical and the notochord lacks the normal number of cells; if a posterior cell is killed, the muscle and mesenchyme cells are lacking along one side of the tail. 13. One-Quarter Embryos; Two-Quarter Diagonal Embryos. Embryos derived from any one quadrant or from two diagonal quadrants of the egg are always very defective. “They never eee a notochord, though if they come from anterior quadrants they may give rise to scattered chorda cells in the perivitelline space; there 1s never a sense vesicle, though if they are from an anterior quadrant a neural plate and sense spots are present. The poste- rior quadrants always contain muscle, mesenchyme and caudal endoderm cells, but never a trace of notochord, neural plate nor sense spots. The embryos are always rounded, there being no distinction of head and tail, and in no respect do they resemble normal larve. I4. Eighth and S1xteenth eee When blastomeres are injured in the 8-cell or 16-cell stages a great variety of abnormal forms are produced. Ventral blastomeres give rise only to rounded masses of ectoderm cells in which fiere is no trace of endoderm or mesoderm; posterior dorsal cells give rise only to muscle, mesenchyme, and caudal endoderm; anterior dorsal cells to neural plate, chorda, and gastral endoderm. 15. Anterior and Posterior Half Gastrule. When cup-shaped gastrulz of the stage shown in Fig. 8 are cut in two transv ue so as to leave all of the yellow celia in one half and all of the chorda 220 Edwin G. Conklin. and neural plate cells in the other a notochord, sense vesicle, or tail is never formed and nothing resembling 2 normal larva develops from either half. The anterior half never contains muscle cells; the posterior half contains many muscle and mesen- chyme cells, but evidently no chorda or neural plate cells. PT: Conclusions. 16. My results confirm and extend those of Chabry, but they are fundamentally unlike those of Driesch; I agree with the work of Crampton as to the cleavage of isolated blactourane but cannot agree with him that whole embry os or larvz are ever formed from isolated blastomeres of the ascidian egg. 17. Regulation in the ascidian egg and embryo 1s limited to the closing of the embryo and the consequent extension of certain cells from the uninjured to the injured side; and also to the forma- tion of a typical notochord and an atypical sense vesicle in right or left half embryos. One odplasmic substance never gives rise to another nor does a given type of cell ever produce cells of another type or organs of a diferent kind than those which would arise from it ina Rota embryo. ‘The fact that the power of regulation is apparently greater in the adult ascidian than in the egg or embryo may be deceptive; the injury to the egg which wipes out completely certain odplasmic substances may be really greater than any which may be repaired in the adult. Furthermore it is possible that the very rapid development of ascidians may act as a check on regulation. 18. ‘These results ee that at least five of the substances which are present in the egg at the close of the first cleavage, viz: ectoplasm, endoplasm, my yoplasm, chymoplasm, and chorda- neuroplasm, are organ-forming substances. They develop, if they develop at all, into the organs which they would normally produce; and conversely, embryos which lack these substances, lack also the organs w hich would form from them. Although I have been unable to test the potencies of these substances before cleavage begins, there seems to be no reason for supposing that they are ever totipotent. Three of these substances are clearly distinguishable i in the ovarian egg and I do not doubt that even at this stage they are Aerated for particular ends. Mosaic Development in Ascidian Eggs. 221 19. A possible explanation of the fact that all the fragments of an immature egg may give rise to entire larve, ea the pro- portion which gives rise to larvz steadily decreases after matura- tion and fertilization, may be found in the fact that before matura- tion the localization of odplasmic substances is usually concentric, after maturation and fertilization, polar-radial; while just before or shortly after the first cleavage the localization may become bilateral. Also the fact that isolated blastomeres may give rise to whole embryos in some animals and to partial ones in others may be due to the varying relations of cleavage planes to localiza- tion planes. If at the close of the second cleavage the localization is still radially symmetrical, each of the first four blastomeres would probably be capable of giving rise to an entire larva; if the first cleavage invariably lies in the plane of bilateral symmetry, as in ascidians, each of the first two blastomeres might be capable of giving rise to an entire larva (though my work one that this would not necessarily happen); if the cleavage planes do not coincide with the planes of localization, as in mollusks and anne- lids, isolated blastomeres would not give rise to entire larve. (See Wilson, ’041, ’04?.) , 20. The development of ascidians is a mosaic work because there are definitely localized organ-forming substances in the egg; in fact the mosaic is one of organ-forming substances rather than of cleavage cells. The study of ctenophores, nemertines, anne- lids, mollusks, ascidians and amphibians (the frog) shows that the same is probably true of all these forms and it suggests that the mosaic principle may apply to all animals. (cj. Fischel, Wilson.) LITERATURE CITED. BarFurtH, D., ’93.—Halbbildung oder Ganzbildung von halber Grosse? Anat. Anz., Viil. BENEDEN, VAN ET JULIN, '84.—La segmentation chez les ascidiens dans ses rapports avec l’organization de lalarve. Arch. de Biol., v Bovert, Tu., ’o1.—Ueber die Polaritat des Seeigeleies. Verh. Phys. Med. Ges., Wirzburg, xxxiv. ’o1.—Die Polaritat von Ovocyte, Ei und Larve des Strongylocentrotus lividus. Zool. Jahrb., xiv. Bracuet, A., ’04.—Recherches expérimentales sur l’oeuf de Rana fusca. Arch. de Biol., xxi. 279 Edwin G. Conklin. Caste, W. E., ’96.—The Early Embryology of Ciona intestinalis, Flemming L. Bull. Mus. Comp. Zool., xxvii. Cuasry, L., ’87.—Contribution a |’embryologie normale et teratologique des Ascidies simples. Journ. Anat. et Physiol., xxii. ConkLin, E. G.,’97.—The Embryology of Crepidula. Jour. Morph., xii. ’98.—Cleavage and Differentiation. Woods Hole Biol. Lect. ’03.—The Earliest Differentiations of the Egg. Science, xvii. ’05'.—The Organization and Cell-Lineage of the Ascidian Egg. Jour. Acad. Nat. Sci., Philad., xi. ’05°.—Organ-Forming Substances in the Eggs of Ascidians. Biol. Bull., viii. Crampton, H. E., ’97——The Ascidian Half-Embryo. Ann. New York Aca. DCi., X. : Driescu, H., ’95.—Von der Entwicklung einzelner Ascidienblastomeren. Arch. Entw. Mech., i. ’o2.—Studien tuber das Regulationsvermégen der Organismen. 6 Die Restitutionen der Clavellina lepadiformis. Arch. Entw. Mech., XIV. *03.—Ueber Anderung der Regulationsfahigkeiten im Verlauf der Ent- wicklung bei Ascidien. Arch. Entw. Mech., xvii. FiscHet, A., ’97, ’98.—Experimentelle Untersuchungen am Ctenophorenei. I and II Arch. Entw. Mech., vi, vii. ’03.—Entwicklung und Organ-Differenzirung. Arch. Entw. Mech., xv. Hattez, P., ’87—Embryogénie des Dendrocoeles d’eau douce. Paris. Haraitt, C. W., ’04.—The Early Development of Pennaria tiarella. McCr. Arch. Entw. Mech., xviii. Hertwic, O.,’92.—Urmund und Spina Bifida. Arch. Mik. Anat., xxxix. *98.—Die Zelle, II. Jena. KorscHELT UND HErpER, ’02.—Lehrbuch der vergleichenden Entwicklungs- geschichte. Allgemeine Theil. Jena. Logs, J., ’92.—Untersuchungen zur Physiologischen Morphologie der Thiere, vi and Vii. Roux, W., ’95.—Gesammelte Abhandlungen iiber Entwicklungsmechanik der Organismen. Bad. II, Leipzig. ’92.Ueber das Entwicklungsmechanische Vermogen jeder der beiden ersten Furchungszellen des Eies. Verh. d. Anat. Ges. zu Wien. ’03.—Ueber die Ursachen der Bestimmung der Hauptrichtungen des Embryo im Froschei. Anat. Anz., xxi. Mosaic Development in Ascidian Eggs. 222 ScHULTZE, L. S., ’99.—Die Regeneration des Ganglions von Ciona intestinalis L. und uber das Verhaltnis der Regeneration und Knospung zur Keimblatterlehre. Jena. Zeit., xxxiii. Stevens, N. M., ’04.—On the Germ Cells and the Embryology of Planaria Sim- plicissima. Proc. Acad. Nat. Sci., Philad. Weismanv, A., ’92.—Das Keimplasma. Jena. Witson, E. B., ’84.—The Development of Renilla. Philos. Trans., clxxiv. ’03.—Experiments on Cleavage and Localization in the Nemertine Egg. Arch. Entw. Mech., xvi. ’04'.—Experimental Studies on Germinal Localization, I and I]. Jour. Exp. Zool., 1. ’04?,—Mosaic Development in the Annelid Egg. Science, xx. Yatsu, N., ’04.—Experiments on the Development of Egg Fragments in Cere- bratulus. Biol. Bull., vi. ZELENY, C., ’04.—Experiments on the Localization of Developmental Factors in the Nemertine Egg. Jour. Exp. Zodl., 1. CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. E. L. MARK, Director—No. 164. DIMORPHISM AND REGENERATION IN METRIDIUM. BY C2 Wie ECAEUN® WitH 2 Ficures. The experiments on which this paper is based were begun at Harvard University in the winter of 1901-02 and continued at the United States Fish Commission Laboratory at Woods Hole in the summers of 1902 and 1903. To Prof. E. L. Mark, Director of the Harvard Zoological Laboratory, and Drs. H. M. Smith and Francis B. Sumner, directors in successive years of the Fish Commission Laboratory, I wish to express thanks for the accommodations kindly provided for the work. At the suggestion of Dr. W. E. Castle, to whom I am deeply indebted for guidance and help throughout this investigation, | undertook to discover if the dimorphism which occurs in Metri- dium marginatum Muilne-Edwards, is perpetuated in accordance with some hereditary law, perhaps related to the law of Mendel. So far at least as concerns asexual reproduction, it soon became evident that this is not the case. Further studies have shown that the dimorphism is due to a peculiar method of development in asexual reproduction. The dimorphism of actinians was noticed and described more or less completely by Thorell (°59), Dixon (’88), MeMurrich (’89), and Carlgren (’93), but the first full discussion of it was pee by Parker ('97), and the only extensive experimental study thus far made with a view to discov ering the exact nature of the anomaly is that of Carlgren (’04). The dimorphic condition referred to is this: The polyps of a given species may have either one or two (rarely three) siphono- glyphs. With each siphonoglyph, as shown for Metridium by Parker (’97), there is invariably associated a pair of directive mesenteries. 226 C. W. Habn. The siphonoglyph is a groove in the cesophagus, covered with cilia, which tend to set up a current of water into the polyp. The relative abundance of monoglyphic, diglyphic and_ triglyphic Metridia, as indicated by observations of Torrey ('98), Parker (99), and myself, is shown in Table I. Monoglyphic polyps, it will be observed, predominate 1 in all localities examined, though the proportion varies within wide limits. ‘The large and small TABLE I. ; Total No. Number Number Per cent EEN Counted. Diglyphic. _Monoglyphic. Monoglyphic. INewport, Role (Parker) 2... -/- 131 53 Thee 59 Oakland) Gala(lormey 24. O22 200 43 157 78.5 SallemmuWViasset so% wo deeu ar ecce 465 57 399 85.8 yas Misses syeee tee oe ss 670 53 611 92.6 Woods Hole; Mass’. <2 52. 22 40. 123 15 | 108 87.8 poly ps from the same locality occur with about the same pro- portion of monoglyphic to diglyphic forms. [here is no dis- coverable Horrelation with any variation in color or in structure, save the constantly associated pair of directive mesenteries and perhaps certain irregularities of the non-directive mesenteries. These structures have been investigated by Parker (’97, 99). He finds the diglyphic character to be associated more often with regularity in the number and arrangement of the non-directive mesenteries. But the variations in the mesenteries are not con- formable to any known laws and cannot be regarded as of specific importance. My first experiments were directed toward discovering whether monoglyphic and diglyphic polyps produce, in asexual Teproduc- tion, each its own sort only. The result obtained gave an emphatic negative to’ this hypothesis. Each sort was found to produce, by asexual methods, both monoglyphic and diglyphic polyps, monoglyphic individuals predominating among the young in both cases. Two methods of inducing multiplication were employed. The anemones were, in some cases, divided into halves by a vertical cut, so as to produce artificially specimens equivalent to those Dimorphism and Regeneration in Metridium. 227 supposedly produced by fission in nature. ‘The greatest success, however, was attained when fragments were cut from the basal portion of a polyp, as in the natural process of basal fragmen- tation. In either case the number of siphonoglyphs in the parent polyp was determined, so that the fragments from monoglyphic and diglyphic parents might be kept separate. Diglyphic and monoglyphic parent polyps could be distinguished either by inspection externally, when they were fully expanded, or by cutting the contracted polyp across after removing frag- ments from its base. ‘The fully regenerated young were stupefied by means of magnesium sulphate and fixed in chromic acid (1 per cent) then imbedded in paraffin and sectioned. Staining the sections with hematoxylin and eosin made it possible, with a low power of the microscope, to decide the various questions on which the interpretation of a polyp’s structure depends. Fragments cut from monoglyphic polyps produced twenty- eight monoglyphic and five diglyphic individuals. (See Table IT.) Three of the diglyphic polyps were simple, 1. ¢., having a single cesophagus to which the two pairs of directive mesenteries were attached; two were double, 1. ¢., having a divided cesophagus, each branch of which was connected with a different pair of direc- tive mesenteries, as if a portion of cesophagus had been produced in connection with each pair of directives, but the two portions had failed to unite. . There were also twelve polyps regenerated from fragments cut from monoglyphic parents, which up to the time when they were examined had developed no mesenteries which could be recognized as directives. It is probable, however, for reasons which will presently appear, that most, if not all, of these would have become unmistakably monoglyphic polyps had they been given a longer time for regeneration. Fragments cut from diglyphic polyps produced in all thirty- six monoglyphic and ten diglyphic polyps, all simple, as well as two polyps whose character was indeterminable. These experiments show conclusively that each sort of polyp can produce the other by asexual methods; but, it will be observed, the diglyphic parents produced a somewhat greater proportion of diglyphic young, and the question at once arose whether this might not be due to an hereditary influence. A more careful study, however, of the regenerated polyps showed that such was 228 C. Ws Habn- not the case. It was found that in the monoglyphic polyps the directive mesenteries arose regularly from the mew or regenerated portion of the body-wall, and that in the diglyphic poly PS, likewise, one of the two pairs of directives arose in nthe same position, but the other arose from the old or parental tissue. This observa- tion at once suggested a different explanation of the dimorphism, viz: that it was due not to the monoglyphic or diglyphic nature ar the parent polyp, but to the character of the fragment from which regeneration took place. If this fragment contained a pair of directive mesenteries the polyp produced would be diglyphic, since it would retain the pair of directives derived from the parent and gain a second pair through regeneration. If, on the other hand, the parent fragment contained no directive mesenteries, then the regenerated polyp would be monoglyphic, containing only the pair ‘of directives produced in the new tissue. To test this hy pothesis a more detailed examination of the polyps was made, the monoglyphic polyps of different parentage being compared with each other, and, likewise, the two lots of digly) phic polyps with each other. In determining which is the regenerated portion of a polyp no single Beeicees can be relied upon. Usually considerable familiarity is necessary to enable one to distinguish regenerated from parent tissue. ‘The latter, as seen in cross- section, 1s usually characterized by deep folds of the ectoderm into which small V-shaped points of the mesogloea extend. (Figs. 1 and 2.) The external surface in this side is more regular and evenly curved; the mesenteries are quite regular, especially is this true of the secondaries and tertiaries in relation to the primaries. The primaries arising from the old portion of the body-wall are longer than those arising from the regenerated part. The mesogloea on the regenerated side is not of uniform thickness or contour and does not conform as regularly with the folds of the ectoderm, when such exist. Regenerated directives and old directives usually differ in length and thickness when viewed in cross- sections, the former usually being short and thick. (Figs. 1 and 2.) These conditions, however, vary greatly, but when the evidence from one criterion is uncertain that furnished by other criteria is usually conclusive. The results of the detailed examination made are incorporated in Table Il. From this it will be seen that the monoglyphic A NN Dimorphism and Regeneration in Metridium. 2 polyps derived from monoglyphic parents had the directives developed in new tissue in all except one of twenty-six cases, in which the limits of old and new tissue could be recognized. In that one case the directives clearly were attached to the old or Fic. 1. Cross section of a diglyphic polyp produced by a fragment cut from the foot-disc of a diglyphic parent in such a way as to include a pair of directive mesenteries. The parental directives lie in the upper half of the figure, the regenerated directives in the lower half. The ectoderm is stippled in the regenerated portion of the body-wall. (From a camera lucida drawing, histological details being omitted.) Fic. 2. Cross section of a monoglyphic polyp produced by a fragment cut from the foot-disc of a diglyphic polyp so as not to include directive mesenteries. The regenerated portion of the body-wall is indicated, as in Fig. 1, by stippling of the ectoderm. The single short pair of directive mesenteries is attached to this portion of the body-wall. (From a camera lucida drawing, histological details being omitted.) parental part of the body-wall. It-would seem, accordingly, that in this exceptional case no directives had been produced in the new tissue. Had this taken place the polyp would have been diglyphic, with the directives arranged exactly as in the twelve 230 C. W. Habn. diglyphic polyps mentioned in the fifth column of Table II. This peculiar monoglyphic polyp may be considered the result Tas_e II. - a i ror % a = 5 a i ay = aie 2 Beee| € 2 o ge as a eee, > 2 || cone ie ay ee eon a B 2 © ig'.8e] Parents. Young. ere o 2 ‘one y 22 6 2 eo +s Ss | Totals it ae ae MQ = eB] 5 5 z ae = ae SI ee) 5 é =. 2 | 2z 20 |S3320) 3 s3—|fe-aeo ee = Ss = 4 MI J M 25 I = a= 2 | 28 | D — = z 2 (double)) — | 5 f M 24 I — — to eee 36 Deer | | \ D —- | — 9 1? | — | 10 Character of polyps produced by basal fragments cut from monoglyphic and from diglyphic parents respectively, the relation of the cut to the parental directives being unknown in a majority of cases. M = monoglyphic; D = diglyphic. of the union of the cut edges of the parental fragment without regeneration of a directive system. Such a result occurs not infrequently in the case of fragments containing only non-direc- tive mesenteries, as well as of oss like this one, which contain a pair of directives. In the former case wholly aglyphic polyps are produced. ‘Three such were observed in these experiments. Carlgren (04) has shown, in the case of Sagartia, that such a result 1s obtained most often when the parental fragments are of relatively large size. It is probable that the same 1s true in Metridium, though precise observations on this point are wanting. The monogly phic polyps of diglyphic parentage, like those of monoglyphic parentage, had the directive mesenteries in all cases except one (out of a total of twenty-five) in the regenerated area. In that one case the single pair of directives was clearly received directly from the parent polyp and no new directives had been regenerated. Likewise in the case of the diglyphic polyps no difference was recognizable between those of monoglyphic parentage and those of diglyphic parentage. In all cases except possibly one there was a pair of directives in the old tissue, and one in the new. ‘There were three diglyphic polyps of this sort Dimor phism and Regeneration in Metridium. ZT derived from monoglyphic parents, and nine derived from diglyphic parents. [he possible exception mentioned was a diglyphic polyp of diglyphic parentage which had two pairs of Hea mesenteries, both apparently in the new tissue. Yet the limits of the old tissue could not in this case be located with certainty, and it is possible that one of the two pairs had really been derived directly from the parent fragment. Otherwise we must suppose that regeneration had taken place in such a way as to produce simultaneously out of new tissue two pairs of directive mesen- teries. [That sucha thing probably occurs sometimes is indicated by the observation once in a great while of a trig/yphic individual, a condition which would be reached if a fragment already con- taining a pair of directive mesenteries acquired two more by regeneration. The triglyphic condition may, however, arise in a different way, vx: by laceration of a diglyphic polyp, which then produces in the area of regeneration a new or third siphono- glyph system. It still remains to account for the fact shown in Table II that more diglyphic polyps are produced by digylphic than by mono- glyphic parents. A moment’s reflection will show that this is not dificult. If pieces are cut at random from the bases of polyps without reference to the position of the directive mesenteries, it is evident that directives are likely to be included in the fragment removed, twice as often when the parent polyp is diglyphic as ates it is monoglyphic, since the diglyphic polyp contains fwo directive sys- tems on opposite sides of the body, whereas the monoglyphic polyp contains only one. Accordingly we should expect the proportion of diglyphic polyps regenerated to be about twice as great in one case as in the other. The observed proportions are not greatly at variance with this expectation. In order to test more fully and directly the hypothesis already presented,—that the condition of a regenerated polyp, whether monoglyphic or diglyphic, depends on whether the parental frag- ment did or did not contain portions of the directive mesenteries,— advantage was taken of the fact that in the experiments summarized in Table I] certain fragments had been cut from the bases of parent polyps in such a way as to include a pair of directive mesenteries, and others had been cut in such a way as not to include directives in the fragment removed, the two lots having been reared separately. In the former lot unfortunately the 232 GaR> EWabn. mortality was high, because of unfavorable conditions in the aquarium in which they were placed, and only three polyps survived. Further, two of these were insufficiently regenerated to show the character of the new mesenteries, but the third was clearly a diglyphic polyp with one pair of directives in.the old tissue and one in the new, as expected. The fragments cut so as to exclude directives did somewhat better. ‘Ten polyps were reared from them. Eight of the ten were clearly monoglyphic, with the directives always in the new tissue; a ninth polyp was insufficiently regenerated, but it had a pair of mesenteries on the regenerated side, which gave some indications of being directives. If so, this polyp is similar in character to the eight previously mentioned; if not, it 1s aglyphic. The tenth polyp was diglyphic, but quite asymmetrical in char- acter, one of the two pairs of directive mesenteries arising close to the boundary between the old and the new tissue. It 1s impos- sible to say w hether a pair of directives was accidentally included in the fragment from which this polyp developed or whether there arose simultaneously two areas of regeneration, each of which produced a pair of directive mesenteries. This direct experiment, incomplete though it is, supports the hypothesis based on the experiments previously described. It indicates that the dimorphism found in Metridia asexually pro- duced is not dependent upon the monoglyphic or diglyphic char- acter of the parent poly p, but upon we erHier the parent fragment does or does not contain a portion of a siphonoglyph system. It harmonizes, likewise, with the observations of Carlgren (’04) on regeneration in Sagartia and other actinians and supports the idea advanced earlier by Carlgren (’93) and supported by Parker (97), that the dimorphism of actinians 46 an ancidenee: een reproduction. As a control of the experiments examination was made of nine spontaneously regenerated polyps collected at Lynn, Mass. This yielded iis closely similar to those obtained from the artificially regenerated polyps. One of the polyps was inde- terminable; one was diglyphic, with one pair of directives in the old and one in the new tissue; and seven were monoglyphic. Of the seven monoglyphic polyps, five had the directives attached to what was unmistakably the regenerated portion of the body- wall, while in the remaining two old and new tissue could not be Dimor phism and Regeneration in Metridium. 233 distinguished on account of the advanced state of regeneration of the polyps. The frequent occurrence of asexual reproduction in Metridium explains the prevailing asymmetry of individuals in this species, regenerated diglyphic polyps in particular being rarely sym- metrical. It is usual to find the mesenteries < arranged with more primaries and secondaries on one side of the plane passing through the siphonoglyphs than on the other. ‘This condition is to be explained by the fact that fragments, either spontaneously pro- duced or formed artificially by random cuts from the base of the foot-disc, arise without any definite reference to the parental mesenteries which traverse that region. Hence, if the directive mesenteries chance to be nearer one end of a fragment than the other or if the new directives are formed nearer one edge of the regenerated area than the other, an asymmetrical polyp results. The idea which has been advanced in the foregoing pages is capable of giving an explanation also of the great variation in the numerical proportions of monoglyphic and diglyphic polyps in different localities. (See Table I.) If we suppose that in cer- tain localities, like Newport, R. I., or at particular seasons of the year sexual reproduction is favored, regular hexamerous diglyphic polyps should at such places or seasons be relatively more abun- dant. Torrey (02) correctly explains as due to asexual repro- duction patches of similarly colored sea- anemones, but the occur- rence of diglyphic individuals among polyps asexually produced does not, as he supposes, show that the diglyphic character has been inherited as such, but rather that in these particular cases the parental fragments happened to include a directive system. The diglyphic hexamerous poly p of Aiptasia, described by Andres and cited by Torrey (’02) as “evidence to be explained,” may be explained on the same basis. The dimorphism which, according to Torrey (02), occurs in polyps produced by budding from the column of sea-anemones is doubtless capable of epelaneuen ina similar way. The production in the experiments here dekcibed of polyps with a divided cesophagus and perhaps, in other cases, of two directive systems formed simultaneously in the regenerated tissue are worthy of notice as giving a clew to the origin of double mon- sters and of triglyphic polyps. Both of these abnormal conditions well known in collections of polyps, doubtless arise in spontaneous 234 GW: Thabne asexual reproduction, since the processes leading up to them have been observed in regeneration artificially induced. In view of these facts it is improbable, as has been supposed by several investigators, that double Metridia area stage in a process of repro- Waeued by longitudinal fission. The older view that they are genuine monstrosities seems better supported, but not the view that they are due to coalescence, as was once thought to be the case. From long strips cut from the margin of the foot-dise and including a half or more of its circumference, polyps with two or three distinct oral discs have several times been obtained in these experiments. ‘This result throws still further light on the origin of double monsters. SUMMARY. The dimorphism which occurs in Metridium is due not to alternative inheritance of the diglyphic and monoglyphic condi- tions, but to the frequent occurrence of asexual reproduction. This takes place spontaneously by basal fragmentation and may readily be induced by cutting off pieces of the foot-disc. Whether a particular fragment produces a monoglyphic or a diglyphic polyp depends, not on the monoglyphic or diglyphic condition of the parent polyp, but upon whether the fragment does or does not contain some portion of a directive system, for a directive system is regularly produced in the regenerated portion of the young polyp. Accordingly, if the portion derived from the parent already contained a directive system, the young polyp will have two such systems and will be diglyphic. But if the parental fragment contained no directives, the young polyp will have only one directive system, that produced in regeneration, and will be monoglyphic. Not only the dimorphism of Metridium, but also its prevailing asymmetry and extreme variability in number and arrangement of mesenteries can be explained by its method of development in asexual reproduction. ‘Trigly phic polyps and those with two or more oral discs or with double or branched cesophagus or devoid of siphonoglyphs are abnormalities due probably to regeneration from fragments of unusual size or shape, as compared with the fragments normally produced in spontaneous basal fragmentation. Dimorphism and Regeneration in Metridium. 235 BIBLIOGRAPHY. CaRLGREN, O., ’93.—Studien tber nordische Actinien. Kongl. svenska Vet.- Akad., Handl., N. F., Bd. 25, No. 10, 148 pp., 10 Taf. ’o4.—Studien itiber Regenerations- und Regulations-Erscheinungen. Kongl. svenska Vet.-Akad., Handl., N. F., Bd. 38, No. 8, 84 pp. 2D vig. TO) Wats Drxon, G. Y., ’88.—Remarks on Sagartia venusta and Sagartia nivea. Sci. Proc. Roy. Dublin Soc., N. S., vol. 6, pp. 111-127. McMvraricy, J. P., 89.—The Actinaria of the Bahama Islands, W. I. Jour. of Morph., vol. 3, no. 1, pp. 1-80, pl. 1-4. Parker, G. H., ’97.—The Mesenteries and Siphonoglyphs in Metridium margina- tum Milne-Edwards. Bull. Mus. Comp. Zool. Harvard Coll., vol. 30, no. 5, pp. 259-273, I pl. 99.—Longitudinal Fission in Metridium marginatum Milne-Edwards. Bull. Mus. Comp. Zool., Harvard Coll., vol. 35, no. 3, pp. 41-56, 3pl. \ THorELL, T., °59.—Om den inre byggnaden af Actinia plumosa Mull. Ofversigt Kongl. Vet.-Akad., Férhandl., Arg. 15, 1858, pp. 7-25, Tab. r. Torrey, H. B., ’98.—Observations on Monogenesis in Metridium. Proc. Cal. Acad. Sci., ser. 3, Zodl., vol. 1, pp: 345-360, pl. 21. *02.—Papers from the Harriman Alaska Expedition. XXX. Anemones, with a Discussion of Variation in Metridium. Proc. Wash. Acad. Sci., vol. 4, pp. 373-410, pl. 34, 35. From the Rudolph Spreckles Physiological Laboratory of the University of California. ae EPFECT OF VARIOUS SALTS UPON THE SUR- wiv A OF THE INVERTEBRATE. HEART. BY CHARLES G. ROGERS. WitH I PLatTe. The cause which underlies the rhythmic contraction of the heart has been the subject of much controversy. In recent years the importance of the inorganic compounds of the blood has been acknowledged by most phy siologists, but the role of each of these different salts in originating and maintaining rhythmic contrac- tions has caused much discussion. Up to the present time almost all of the work done upon the physiology of the heart has concerned vertebrates alone. It was, therefore, a pleasure to follow the kind suggestion of Dr. Loeb and use the heart of an invertebrate as the subject upon which to conduct a series of experiments which may furnish an answer to the following questions: What is the influence of the various salts, found in the blood, upon heart action? And, is this influence the same in the crab as in the heart of the verte- brates ? I wish to express my thanks to Dr. Loeb for suggesting the problem and for many kind criticisms during the course of the work. METHODS. In the study of the problem the hearts of the marine czab Brachynotus nudus were employed. ‘This crab is found very abundantly along the western shore of San Francisco Bay, be- neath rocks, between tide marks. The crabs may be kept in the laboratory for a considerable period without serious deterioration and hence prove to be an admirable form upon which to work. In carrying out these experiments three general methods of procedure have been employed, of which two, however, were 238 Charles G. Rogers. soon discarded. ‘The first consisted in carefully isolating the hearts in watch glasses, each containing about ten cc. of the solution to be tested, and making observations upon the number and quality of the beats dev eloped. This method was employed only during the early stages of the work as it did not permit of any accurate estimate of the amount of work done by the heart. The second method was that of carefully suspending the hearts by means of delicate glass hooks in connection with light recording levers and allowing them to trace upon smoked papers the records of their contractions. In these experiments the hearts were moistened with the solutions by means of camel’s hair brushes. Oxygen, of course, is easily taken up from the air, but the effects of the various constituents of the solutions are difficult to deter- mine with this method as the amount of solution in contact with the heart is small and variable. One can not be sure at any given time that the solution has replaced the body liquid normally present. In view of this fact a third method was employed. This was similar to the second except that the hearts were im- mersed in tubes, each containing about 30 cc. of the solution. While the general results of the first method agree with those of the third they are disregarded in the final summing up of the work as lacking in sufficient accuracy. The second method failed to give uniform results. “The author feels that the results obtained by the third method are far more reliable than those obtained by either of the preceding ones, hence they form the basis of the following report. EXPERIMENTS. A. What is the Optimum Concentration of Salt Solution which will Favor the Rhythmic Activity of the Heart? Botazzi! has measured cryoscopically the osmotic pressure of the blood of many of the marine invertebrates and has found that it is practically the same as that of the sea water. The average depression of the freezing point in the body liquids of inverte- brates is given by Hamburger’ as —2°.29, and the depression of the freezing point of the sea water is given by Hober as —2°.3. 1Botazzi. Archives Italiennes de Biologie, xxviii, 1897, p. 67. *Hamburger. Osmotischer Druck und Ionenlehre in den Medicinischen Wissenschaften. Effect of Salts Upon the Invertebrate Heart. 239 The NaCl solution having the same osmotic pressure contains 3-783 per cent of NaCl, or is a solution of about § m. concentra- tion. In bays and in the mouths of rivers the Banat pressure of the sea water becomes greatly modified. Dr. Loeb! has shown that animals taken from the waters of San Francisco Bay thrive in solutions with an osmotic pressure approximately equal to that of a2 m. NaCl solution. ‘The animals upon which the present Dek was carried out were collected at a point about three miles north of the Golden Gate where the water sweeps by at both flood and ebb of tide in strong currents. At flood tide the water has nearly the same concentration as the water of the open ocean, but at ebb tide it is much freshened by the water of the Sacra- mento River. A series of experiments was made to determine the concentra- tion of a solution of NaCl which would favor the rhythmic con- traction of the heart. For this purpose a 2 m. NaCl solution was employed and this was diluted with distilled water in varying amounts. It might be expected that the beats, if any at all appeared, would continue longest in that concentration of NaCl which most nearly approaches the normal concentration of the blood of the animal. As a result of a long series of trials it was found that a 2 m. and a ;& m. solution of NaCl acted most favorably. Both of these concentrations were employed in the further work. B. Is NaCl Essential for Maintaining Rhythmic Contractions? In considering the question whether any substance is essential for the origination of rhythmic contractions the following con- dition must be kept in mind: In order to demonstrate that a substance is necessary for the origination of rhythmic beats in a muscle we must employ a muscle which does not beat rhythmi- cally when it is removed from the body. Dr. Loeb* has shown that this is true in the case of the center of the bell of the medusa Gonionemus, and was able to demonstrate that NaCl is essential for the origination of rhythmic contractions in this muscle. ‘The Loeb, J. Pfluger’s Archiv., vol. xcvii, 1903, p. 394. Also University of California Publications, Physiology, vol. i, No. 7, pp. 55-69. "Loeb, J. American Journal of Physiology, vol. ili, 1900, p. 383. 240 Charles G. Rogers. ventricle of the heart of the turtle also does not exhibit rhythmic contractions when it is removed from the body of the animal and in this case also Lingle' was able to show that the development of rhythmic contractions depended upon the presence of NaCl. Lingle also emphasized the necessity of a large supply of oxygen. Overton? working independently, and apparently not having seen the reports of the work already mentioned found that in the absence of NaCl, e. g., in a pure sugar solution muscle does not respond to electrical stimuli. In the present work we are dealing with a heart of a single chamber and one that continues to beat when it is removed from the body of the animal. We can, hence, only raise the question whether the heart of the crab will continue to beat for a long time in the absence of NaCl while with this salt present it will continue to beat for a much longer period. We may also raise the question whether when the contractions of the heart have ceased in some solution lacking in NaCl the addition of NaCl will cause rhythmic contractions again to take place. In order to show whether the hearts of the crabs depend upon the presence of NaCl to maintain rhythmic contractions they were immersed in solutions lacking in this salt, but in which the osmotic pressure was kept approximately at the normal height by means of cane sugar. In some experiments no oxygen was added to the solution in others hydrogen-peroxide was added: following the experiments of Lingle, and in others a current of gaseous oxygen was allowed to bubble through the solution. As the result of these experiments it was found that the fresh hearts of the crab do not cease beating at once when placed in a pure sugar solution and that the length of time during which such contractions may continue is somewhat extended by the presence of oxygen. In no case, however, did the heart in such a solution continue to beat for more than one hour and in the very great majority of cases not more than twenty minutes. ‘There seemed to be no great difference in the action of the hearts when the concentration of the solutions varied between 2m. and 2 m. In a sugar solution the beats are at first not weakened but very soon they lose in strength and soon cease altogether. When no oxygen is ‘Lingle, D. American Journal of Physiology, vol. viii, p. 75 ff. 2Overton. Pfluger’s Archiv., Bd. 92. Ejfect of Salts U pon the Invertebrate Heart. 241 added to the solution the beats are of regularly decreasing ampli- tude until finally no contraction is visible. When fee oxygen supply is ample it frequently happens that the last beats have perhaps one-third of the amplitude of the normal contraction, but they occur at regularly increasing intervals until they cease altogether. In some cases irregularities of beat occur. “These may be due to injuries received by the heart when it was removed from the body of the animal or from a deficient supply of oxygen. A very marked effect of the cane sugar is the great increase . of muscular tone which occurs in all heats amnerted 4 in such solutions. If, as has been held by some, NaCl is the substance which is necessary for the development of rhythmic contractions we should find upon adding NaCl to the sugar solution that the length of time during which a heart will continue to beat will be lengthened as we increase the amount of the salt, up to the limit of the con- centration in which this salt exists in the sea water. In order to test this varying amounts of 2 m. NaCl were added to a 3m. cane sugar solution and it was found that as the proportion of NaCl in the solution increased the hearts beat for a longer time. The following examples will illustrate: m. NaCl plus 75 cc. 2 m. cane sugar beat for 26 minutes. P /) 8 5 m. NaCl plus 50 cc. $ m. cane sugar beat for 35 minutes. m. NaCl plus 25 cc. $m. cane sugar beat for 70 minutes. No. 222—25 cc. No. 213—50 cc. No. 228—75 cc. Ico anles anjoo (The above experiments were made without adding any extra oxygen to the solutions.) In a pure 2 m. NaCl solution the hearts beat on the whole longer than in the mixtures of NaCl and cane sugar. It now becomes of interest to know whether other substances than the NaCl have the power to aid in the maintenance of rhythmic contractions when added to a solution of cane sugar. On account of the importance of calcium, potassium and magne- sium for marine animals we naturally turned first to these in order to answer the question. Small and vary ing amounts of the chlorides of these metals were added to Slicions of cane sugar and records made of the heart contractions under the influence of these solutions. 242 Charles G. Rogers. C. The Effect of the Addition of Calcium Chloride to a Sugar Solution. Small amounts (.5 cc. to 3.00 cc.) of a 5 m., $m., or 7 m. calcium chloride solution added to 100 cc. of $ m. cane sugar solution in which a heart may be immersed modify very materially the action of the heart. ‘The beats become more uni- form in quality and occur at more regular intervals than when the heart is immersed in a pure sugar solution. At the same time it seems probable that the length of time during which a heart will continue to beat in such a solution is somewhat lengthened. It is dificult to make a definite statement with regard to this last point, however, on account of large individual variations in the actions of different hearts. A very characteristic effect of the addition of calcium is seen in the gradual retardation of the beats, the contractions coming at regularly increasing intervals until they stop altogether. In some cases instead of this retardation we may find a progressive decrease in the amplitude of the beats while the rate remains fairly constant. If larger amounts of calcium chloride be added to the solution there is a very evident poisonous effect exerted upon the heart by the salt and in a pure calcium chloride solution the effect becomes very marked, the hearts continuing to beat for only a very few minutes even though a large supply of oxygen be available. D. The Effect of the Addition of Potassium Chloride to a Sugar Solution. The addition of small amounts of a ,°, m. solution of potas- sium chloride to a solution of cane sugar in which hearts are immersed brings about a marked increase in the amplitude of the contractions. At first the beats occur much more rapidly than in the pure sugar solution and these contractions are very powerful. After the first series of very strong contractions, which lasts for only a few minutes (eight or nine at the most) comes a series of contractions of nearly normal amplitude but somewhat more rapid than usual. Following these but coming more slowly is another series of* exceedingly _ strong contractions which is finally followed by a rapid decline in the amplitude of the beats. Effect of Salts Upon the Invertebrate Heart. 24:3 Coincident with this decline is the characteristic increase of muscular tone generally associated with the action of the sugar solutions. A particular feature of the action of solutions con- taining potassium is that a single muscle twitch occupies much less time than when the muscle is immersed in a pure sugar solution, or even in other solutions in which the amount of potas- sium is much less. E. The Effect of the Addition of Magnesium Chloride to a Sugar Solution. When small amounts of a ~, m. solution of magnesium chloride are added to a solution of cane sugar in which the hearts are immersed it is found that the quality of the contraction becomes modified although the length of time during which the heart will continue to beat is not altered. ‘The first contractions are usually stronger than normal. After a short series of these powerful beats the beats lost strength and became somewhat irregular, and finally were weak and rapid, with occasional strong contractions scattered among the much weaker ones. F. The Effects of the Addition of Calcium and Magnesium to a Sugar Solution. oO Small amounts of 2 m. CaCl, and a 3m. MgCl, solution added to a solution of cane sugar have a marked influence upon the action of a heart pe eeedcs in such a solution. ‘The beats become more regular as to time and intensity and the average length of time during which the heart will continue to beat is greater than in the solution with either one salt alone. G. The Effect of Sodium Chloride and Calcium Chloride. While in a pure NaCl solution the heart tracings are very similar to the fatigue curves of voluntary muscle, as has been noted by other observers upon other hearts, the addition of a slight amount of a 2? or ;; m. solution of calcium chloride exerts a profound influence upon the heart action. ‘The cardiac 244 Charles G. Rogers. contractions become more regular in time and amplitude and last for a longer time than when the heart is immersed in pure NaCl alone. This may be due to one of two causes: either the calcium is necessary in itself for the long continuance of the con- tractions or it may be necessary to counteract the poisonous effects of the NaCl. A record of a ‘single experiment will indicate the trend of results. The heart was immersed in a solution contain- ing 100 cc. 7 m. NaCl plus 3 cc. 3 m. CaCl,. These propor- tions are the same as regards these two salts as were used later in the optimum solution. In this solution the beats continued fora period of more than two hours, probably for more than three but owing to a mechanical defect the heart tracing is imperfect. The record indicates however the main fact which is to be demon- strated—that the addition of calcium chloride to a solution con- taining sodium chloride renders that solution less harmful. In no case did a heart continue to beat for so long a time in a pure sodium chloride solution as in the experiment just mentioned. Whether any other salt may be found to fully take the place of the calcium chloride in the solutions I am at present unable to say. Up to this time none has been found. H. Will NaCl Restore to Activity Hearts Which Have Ceased Beating in. Other Solutions? A heart immersed in a solution containing 100 cc. 3? m. cane sugar and .5 cc. 73, m. CaCl, beat regularly fora period of fifteen minutes, the beats becoming gradually retarded during that time. During the next fifteen minutes only one contraction was recorded, At the end of half an hour the heart was immersed in a ae NaCl solution and rhythmic contractions began at once and con- tinued for about an hour and a half, becoming more rapid and of less amplitude toward the end of the series. It ought to be stated that in this case the response was unusually prompt and long continued. In another case an immersion for 50 minutes i a solution con- taining 100 cc. 3 m. cane sugar, .5 cc. 33; m. CaCl, .75 cc. 2m. MgSO,, and a slight amount of hydrogen pee suthced to bring the heart to a standstill. The heart was then immersed in a solution of , m. NaCl and after a latent period of twelve Effect of Salts Upon the Invertebrate Heart. 245 minutes contractions were resumed. At first these contractions were very weak and of little amplitude, but they gradually became stronger, and later diminished in the manner common to hearts immersed in a pure solution of sodium chloride. The substitution of what is termed later in the paper the “‘opti- mum solution”’ failed to restore rhythmic contractions in hearts which had ceased beating in a sugar solution. A large number of experiments were made to discover if hearts which had ceased beating in a pure NaCl solution could be made to beat again by some other solution. In no case were beats resumed after they had stopped in a pure NaCl solution. This might seem to indicate that in this case irreversible compounds are Se eaned j in the tissues of the heart under the influence of the sodium chloride which will not allow rhythmic activity to proceed. Ne. Tbe Ejfect of Hydrogen Peroxide and of Oxygen in Solutions. During the course of the experiments it became evident that in some cases at least the failure of the heart to respond to the solutions was due to an insufficient supply of oxygen. Even when well aerated the solutions contain less oxygen than does the blood by which the hearts are normally surrounded. Lingle’ found that the addition of small amounts of hydrogen peroxide to his solutions aided very materially in the long continuance of the heart beats. My own experiences confirm THe results in this re- gard. In fact it seems safe to say that without a good supply of oxygen the heart beats are impossible. In many experiments made with sodium chloride as the princi- pal agent there was noticed a very marked loss of tone as the heart Peatiaued to beat. At first this was attributed entirely to the action of the NaCl but later it seemed more probable that the loss of tone was, partly at least, due to the lack of oxygen. Hearts beating 1 ina solution lacking in oxygen show panked. fatigue in a short time and finally cease entirely to beat. Such Hearne may be revived and again caused to contract rhythmically by simply adding to the solution in which the heart is immersed a little hydrogen peroxide. The following experiment will illustrate the point in question: When a heart was immersed in what I have ‘Lingle. Loc. cit. 246 . Charles G. Rogers. shown elsewhere to be the “optimum solution” plus hydrogen peroxide or plus gaseous oxygen it would continue to beat for a period ranging from 20 to 30 hours. ‘To show the effect of the oxygen a heart was immersed in such a solution lacking in oxygen. At first the beats were quite strong but became weaker rapidly and within forty minutes had ceased entirely. When the heart had been in the solution for fifty minutes it was immersed in another solution of the same composition but containing hydrogen peroxide. After a latent period of about an hour and a half con- tractions were again resumed, becoming gradually stronger till they had reached a maximum which was steadily maintained. When this heart had been beating steadily for one and three- fourths hours it was again immersed in the solution lacking in oxygen. ‘The contractions were almost immediately slowed and were later reduced in amplitude. After being in this solution for fifteen minutes and the beats had become very feeble the heart was again placed in the solution containing the hydrogen peroxide. After a few minutes of weak contractions it again recovered and continued to give maximum contractions for some hours. The exact length of time during which the beats continued was not taken. If instead of adding hydrogen peroxide to the solution we allow a current of gaseous oxygen to bubble through it, taking care that the bubbles do not cause sufficient agitation of the solution to mechanically stimulate the heart to contraction, we find that the heart will make use of the oxygen held in the solution and con- tinue to beat for a long time. (See Fig. 4.) Ke a he Effect of Van't Hoff’s Solution. Van’t Hoff has given us the formula showing the relative pro- portions of the various salts in the sea water. Calcium 1s, ac- cording to his statement, the only considerable variant. The other sale exist in the following proportions: NaCl 100, KCI 2.2, MeCl.e7:8, MesO, 3.8. If, as has been supposed, these salts exist in the blood and body liquids of the crab in the same proportions in which they are found in the sea water and the heart of the crab derives its stimulus from such a solution, then an artificial solution containing these salts ="? Effect of Salts U pon the Invertebrate Heart. 247 in the proportions mentioned should be as favorable for the con- tinuance of rhythmic contractions as the blood itself provided we add to it the amount of calcium which 1s usually found in the sea water in which the animal occurs. Analyses of the sea water showed that the amount of calcium chloride present in the sea water of San Francisco Bay is about one for every one hundred of sodium chloride. ‘The addition then of a corresponding amount of calcium chloride to the Van’t Hoff solution should render that solution favorable for the life of the hearts. A series of experi- ments with solutions containing sodium, potassium and mag- nesium in the proportions stated and with calcium as a variant were made. The amount of calcium chloride used ranged from -5 cc. to 3.25 cc. for every 100 cc. of NaCl of the same molecular concentration (,%; m.) As the result of these experiments it seems safe to say that the heart of the crab will continue to beat for a long time only in a solution which contains a greater propor- tion of calcium chloride than the sea water. In the following experiments the sodium chloride, magnesium chloride, magnesium sulphate, and potassium chloride were employed in the propor- tions stated above, the concentration of the solutions being {4 m. Various amounts of ,; m. calcium chloride were Added as indicated: Exp. 349—1.0 cc. CaCl, to 100 cc. NaCl, heart beat for 50 minutes. Exp. 384—1.5 cc. go minutes. Exp. 287-——2-0\Ce; Exp. 513—3.0 ce. “ee “ “ “e “ee “ee ce 18 hours. 6 hrs. or more. << (SG ““ “ (SAG “ O Exp. 515—3.25 cc. 3 (See Figs. 1, 2 and 3.) The above figures do not give an adequate idea of the improve- ment in the action of the hearts caused by the increase in the amount of calcium in the solution. ‘There is a very marked 1m- provement in the quality of the beats, and the regularity of the contractions in addition to the increase in the length i time during which the contractions may continue as the greater amounts of calc1um are added. ‘That a part of the effect of the calcium con- sists in neutralizing the poisonous effect of the KCl in the solution was shown when ite amount of KCl employed was less than that called for by the formula. In such cases the amount of calcium 248 Charles G. Rogers. chloride needed to make a well balanced solution was less than when the full amount (2.2 cc.) was used. When we employ a solution containing sodium, potassium and magnesium in the proportions in which ches exist in the sea water al add a sufficient amount of calcium to neutralize the poisonous effects of these salts we find that the beats exhibit a remarkable uniformity of contraction which is long continued. When more than 3.0 cc. of calcium chloride is added to 100 of sodium chloride we End that the amplitude of the contractions 1s lessened, but the beats are slower, more nearly the normal rate, and continue through a longer period than in any of the solutions containing less of the calcium. L. The Effect of Sea Water as a Nutrient Solution. In Van’t Hoft’s solution of ; m. concentration we have been using the various salts in the proportions in which they exist in the sea water, as it is found in the bay. If the concentration of the various salts in the blood of the animal is the same as in the sea water by which they are normally surrounded we should find that when a heart is immersed in sea water it would beat as well as in our artificial solution. “The water used in this series of experiments was taken from the open ocean and hence of higher concentration than the water of the bay. In order to reduce the osmotic pressure of this water to about that of the water of the bay it was diluted with distilled water. It was found that the most satisfactory results were obtained when 85 cc. of sea water and 15 cc. of distilled water were used. But even this dilution did not give a solution which was so favorable for the action of the hearts as was the artificial solution. In the diluted sea water all the hearts behaved like hearts immersed in solutions containing too much NaCl or too little CaCl,. As we have already shown the sea water contains about one part of calcium chloride for every one hundred of sodium chloride. But the artificial solution which had been found most favorable for the long continued heart action contained at least three parts of calcium chloride to every one hundred of sodium chloride. If now we add to the diluted sea water small amounts of calcium chloride so as to raise the proportion of this salt to about that which we have in our Ejject of Salts Upon the Invertebrate Heart. 249 artificial solution we should have a solution which would be equally as good as the artificial solution in its action upon the heart of the crab. “This was indeed found to be the case. The heart beats became more regular and the length of time during which the heart would continue to beat was much increased. The solution now proved to be in every particular the equivalent of the artificial solution. ‘This suggests the possibility that the con- centration of the CaCl, in the blood of the crab, and also the concentration of this same salt necessary for long continued heart action is higher than that of the sea water. M. The Role of Sodium Bicarbonate and Sodium Hydrate in Artificial Solutions. It was found during the course of the experiments that the addi- tion of small amounts of sodium bicarbonate to the solutions em- ployed had a very beneficial effect upon the action of the hearts. For a time it was thought that this substance was in itself a neces- sary component of the liquids intended to favor rhythmic activity. Dr. Loeb! has shown that the sea water is practically neutral in reaction. How the bicarbonate could affect the action of the heart was a puzzle until it was remembered that very small amounts of free acids in artificial solutions may exert very in- jurious effects. ‘The role of the bicarbonate in neutralizing any free acid that may be present in the solutions throws a new light upon the subject and makes its presence desirable. It has the power to neutralize acids and yet is not itself alkaline in reaction. It is therefore possible to have solutions containing an excess of this substance without affecting the neutrality of the solution. By thus neutralizing any free eid which may be present we make the buaditions © most favorable for heart activity, and oe for the proof of the fact that the body liquids are neutral 1 reaction. The addition of small amounts of ~, NaOH will have exactly the same effect. Care has, of course, to be taken not to add too much of this substance as the solution must not be too alkaline in reaction. 'Loeb, J. Archiv. fiir die gessammte Physiologie, Bd. 99, 1903, p. 637. 250 Charles G. Rogers. N. Lhe Substitution of Another Metal jor Sodium. Dr. Loeb! has shown that in the case of the skeletal muscles which are made to give rhythmic contractions by means of elec- trolytes, it is possible to substitute in the place of the sodium another metal, especially lithium. Up to the present time no one has succeeded in making such a substitution in the case of cardiac muscle. A large ainubes of experiments were made using lithium chloride in the place of sodium chloride in the Van’t Hoff solution. In no case was it found that rhythmic contractions would con- tinue in such a solution for a longer time than they would in a pure sugar solution. SUMMARY AND CONCLUSIONS. . The blood of the crab studied, Brachynotus nudus, probably = the same concentration as the average of the sea water of the bay. 2. On account of the fact that the heart does not cease beating when it is removed from the body of the animal it is impossible to determine that any substance is essential for the origination of rhythmic contractions. It has been demonstrated, however, that such contractions will not long continue when NaCl is absent. Calcium, potassium and magnesium each have an important influence upon the heart contraction. 3. The balance between the salts entering into the composition of the artificial solution, and presumably of the blood also, is a very delicate one and can be determined with great accuracy. 4. Sodium chloride has the power to restore rhythmic contrac- tions in hearts which have ceased beating in some other solutions. 5. Hearts which have ceased beating in a pure NaCl solution do not again beat when placed in a solution lacking in NaCl. 6. The presence of a supply of oxgyen in the coleomes is neces- sary for rhythmic contractions. Oxygen may be supplied by adding small amounts of hydrogen peroxide to the solution or by aoe oxygen gas to bubble through it. . The solution most favorable for rhythmic contractions of ihe ‘cfab’s heart was found to have the following composition: Loeb, J. Festschrift fiir Professor Fick, 1899. Effect of Salt Upon the Invertebrate Heart. 251 100 parts NaCl, 7.8 parts MgCl, 3.8 parts MgsO,, 22 paris KC 3.25 parts CaCl,, all of ; m. concentration and oxygen. In case the oxygen is added in the form of the peroxide, sodium bicarbonate or sodium hydrate must be added to neutralize the acid introduced with the peroxide. 8. Sea water to which has been added calcium chloride acts in the same way as does the artificial solution. g. The normal circulating fluid of the crab must contain a larger proportion of calcium than does the sea water. 10. Lithium chloride can not be substituted for sodium chloride in the artificial solutions. 252 Charles G. Rogers. EXPLANATION OF PLATE. All records read from right to left. The time marker indicates intervals of thirty seconds. Fig. 1. Record of a heart beating in a solution containing roo cc. m. NaCl, 2.2 cc. $ m. KCl, 7-8 cc. $m. MgCls, 3.8 cc. $ m. MgSOu, 1 cc. $ m. CaCle, .75 cc. $§ m. NaHCO; and oxygen. It will be noticed that the first beats were strong, but became rapidly; weaker, then slower, finally ceasing in less than an hour from the beginning of the experiment. The curve is characteristic of all hearts beating in solutions in which the amount of Na is too great, or the Ca too small. Fig. 2. Record of a heart beating in a solution similar to that mentioned above, except that 2 cc. 2 CaCl were employed. This heart beat for a much longer period than that in the previous experiment. Section b of the record was taken two hours from the start a; section c four hours and section d six hours. Fig. 3. Record of a heart beating in a solution similar to those used in the above experiments except that it contains a larger amount of CaCls, viz: 3 cc. $ m. solution. The beats maintain their first strength and rate of contraction for a long period. Section 4 is taken four hours from the beginning of the record, section c seven hours, section d eleven hours, section e eighteen hours and section 7 twenty hours. At the point x in section d a couple of drops of hydrogen peroxide were added to the solution in which the heart was immersed. Its effect is shown in the ensuing stronger and more rapid contractions. Fig. 4. Portion of the record made by a heart beating in a solution similar to those above, but con- taining 3.25 cc. $m.CaClo. The beats shown at a are characteristic for hearts beating in such a solution and under ordinary conditions will continue for periods of thirty hours or more. At the point b the solution containing oxygen was replaced by one which had been heated to expel the gases in solution and then cooled to the same temperature as the solution first employed. The effect of the lack of oxygen is shown in the regularly decreasing force of contraction. At c the heart was again placed in the solution containing oxygen, and after a short time regained its former strength and continued beating for a number of hours. This experiment was repeated frequently with similar results. ae q - = . : = 4 ‘ 1d ; as “ aa, ad = iP . \ “ a * 4 ) * SUUDIES ON REGULATION: VII. FURTHER EXPERIMENTS ON FORM-REGULATION IN LEPTOPLANA. BY C! Me CriED: Wir 34 Ficures. The present paper is devoted to a consideration of certain phases of the process of form-regulation in Leptoplana, which, although of great interest when viewed in the light of the conclu- sions reached in previous papers on Leptoplana (Child, ’o4a, ’o4b, ’04c), do not in themselves afford sufficient data for these conclusions. Considered at this time they serve to confirm and extend the conclusions already drawn from other data. A. TYPICAL CHANGES IN PROPORTION DURING REGULATION. During the process of form-regulation of pieces in many of the lower animals certain changes of form occur which inv olve not only the new parts but Ae the old fully differentiated regions. Under normal conditions these changes consist in approximation to the typical proportions of the species. A description of these changes in Planaria, where they are considerable, has been given by Morgan (’00) who has applied to them the term “ morphal- laxis.” I have shown recently (Child, ‘02, °03) that similar changes in Stenostoma are at least in part the result of traction upon the parts in certain directions in consequence of the charac- teristic motor activity and I have obtained strong evidence as yet unpublished, that the same factors are concerned in Planaria. But the term “morphallaxis”’ has been applied to phenomena which, in my opinion, are wholly diverse and therefore, although I have employed it in certain previous papers (Child, ’02, ‘03a, ’03b), it seems preferable to use some less vague term. Driesch (or) considers “morphallaxis” identical with his “Restitution durch Umdifferenzierung,” but morphallaxis may occur without 254 C. M. Child. any “redifferentiation”’ as in the case of Gonionemus (Morgan, °99)- Some term is necessary to denote those changes of form in the Turbellaria and other groups which are primarily mechanical and connected with motor activity. There is no fundamental difference between such changes in the new parts and in the old. Both, as well as many other regulative phenomena, may be included under the head of mechanical regulations (Child, 2). The fact that the typical proportions ae result is merely incidental. I have shown (02) that in the absence of the loco- motor tensions the result may be exactly the reverse. Until opportunity offers for a more extended discussion of this matter I preter to designate these changes merely as changes in propor- tion. During form-regulation in epianiens changes in proportion similar to those occurring in Planaria and Stenostoma occur. In pieces containing the cephalic ganglia changes are considerable, though not as great nor as rapid as in Plana and Stenostoma, a difference which is ev idently due to the fact that the tissues of Leptoplana are less soft and plastic than those of the other forms mentioned. [he changes consist in relative elongation and re- duction of the transverse diameter, especially toward the posterior end. In order to make clear my point of view in these experiments it 1s necessary to refer briefly to earlier experiments on Stenos- toma and Leptoplana (Child, ’02, ’03a). In the case of Stenos- toma I found that the changes in form and proportion of the pieces during regulation, 7. ¢., fie elongation and the change from cy Peal to conoidal form were adie primarily to the tension upon the tissues consequent upon the use during locomotion of the posterior end as an organ of attachment. It was possible to inhibit or retard the change in proportions by preventing the pieces from attaching themselves to the substratum (Child, ’o3a). In Stenostoma the changes i in proportion are much more rapid than in other forms examined, being completed in many cases in twenty- four to thirty hours. In my first paper on Leptoplana (Child, *’o4a) a brief descrip- tion of the*method of locomotion was given, certain points of which must be recalled to mind. In creeping, Leptoplana uses the margins of the head region for drawing ihe body forward, a iis ee Studies on Regulation. 255 while the margins of the body from about the middle or a little anterior to it, to the posterior end are employ ed as organs of attachment, the most posterior part, the “tail’’ being most fre- quently used in this manner. The body of the animal is fre- quently subjected to tension during creeping and I believe that, as in Stenostoma, this tension 1s ee considerable importance in determining the general form. Its effect upon the newly formed regenerating parts has already been discussed (Child, ‘oab, ’o4c). e have now to consider the changes in proportion of the old parts during regulation and to discuss the role of mechanical factors in these changes. Since the changes are slow it is imprac- ticable to control them by preventing the animals from attaching themselves to the substratum, as was done in the case of Stenos- toma, but modification of the changes is possible by certain meth- ods which will be described below. In Stenostoma (Child, ’02) the posterior region of the body is subjected most frequently and in greatest degree to tension, but in Leptoplana the lateral margins of the body as well as the posterior end are used for dtiachmiene: so that very frequently only the anterior portions of the body are subjected to tension, the whole posterior portion being attached. However, the posterior end is usually the first part to attach itself and the last to be released, hence in the long run it is undoubtedly more stretched than other parts. It is probable that the outline of the body of Leptoplana is due in large part to the mechanical factors connected with attachment and locomotion. Saez a & 2 z Sees ai a, il Sea loa ee a oa sect chefs = = = on CZ j ; ral o Bi Sal sl se: iS eu ceil) LD Series. Time. | Sal eve || wes s 3 S| eee o @ = castes °7) BS “oS S Ss] Fy] Sa ce Sle o S| RS & > 2 a oO matey ~ ab Pus} oo os Ms a0 ao 2 Se Eos ae 3 Seca = F Bi Se)-|) OES || SS a oe) 2 -a © | wa O bo ] ev Pw Po vo | 2 al Ss ot 4H |v fe 2) Ie OL = S es Time of section. 4 — 3 _- | 3 3 — I 1G) GE oeappee 5 gS Nee 3 Ons) 63 3 3 223 2 a DTOAYS Aes: 5: 62) || 3 51,3 BSN HSS | 92.5 2 neg? || Wak 3 3.0) GEN Songer GG | 2 et. || keg |p ite: tay) |) aean 4 Noa days: .<. ++. megelec2 cia) Tage! 3 Teeth MOG. nee jeter Nhat 5 7G Ce copson. 3-8 | 2.3 | 1.2 | 1.6 | 1.4 | 0.6 | 1.2 | 1 0.9 6 Time of section. 9g = AS etat |patyets Bak | SHS) Shed = 7 67. ig) GE Soo nlodae Ouse mle seal sasuilt 530 es 3-3 | 3 DeiSailiez 8 DIG MCAYS nie = ac. Gate || ReteGF MINS) 7-tse [oO || eps 3 Daly ||| ails |) 16g) 9 a7) Gey Seneeaae SHO merit lea. hull 4-20 alll 2,10 Hees 204) 2 Te 10 Time of section. 10 = Ants WG 5 Pls 2a) | ezak = II 1 CEN S6 es ee OSE" | atic l|) Peay |) a7 Ge Des oh e\ Tey Il pigs) 12 pe DSWAAYS 1.7. /c- =: Onsieetes: |i Br Si Wl eans 2 2 vise on 13 AVEGAYS ores 1-12 Daal |f ey I Devel na: 1.7 2 ie I 14 Grd ay seis. The eG; Te Dah |) 4 Tiei5 ez) || wa I 15 (5 Gy Roneeaen 5 I I 1.8 2-5" | Ded tet °.8 0.7 16 (| Time of section.) 11 — 2 4 6 2 2 1.8 = 17 || 13 davs....... 12 I reer || Yee | 16 1.8 2 V6)! x 18 (ees days........ WL De\| italy aaah alll Gey | MO 1e/ (7) || Stee 19 59--- \| 37 dayss-ct- :: | 1x Ay || ite?) 1.7 | 2.7 Ded Ti Tey 1.4 I.I 20 Gi idays..-2)...; | 9260) 1) 13 Wes. | Asti tats ial TA5 1-2 |) 0-9 21 QGrdays=.-...- | fetal 13) 2 4 3 Ty) 8 0.8 22 Wats On days. -.0)..... Re aware | orgt|| nae. ||. 2.7 0.9 O.9)9 | 05715)|) 0.6 23 each series, however, give the dimensions of the piece as they were at the time of section. Certain of the headings of the vertical columns require brief explanation; the eighth column is headed “To level where decrease in width begins;” the anterior region of the body is the widest part 258 C. M. Child. and under ordinary conditions the width of the body 1 is uniform to about the level of the eyes; during the changes in proportion accompanying regulation the deereaeee in width may begin a con- siderable dace anterior to the eyes, hence the importance of giving this measurement. The followi ing column “Greatest width” gives the width of this widest anterior region. ‘The tenth and eleventh columns also require a word of explanation. The cut surface contracts after section in all cases and thus the extreme posterior end of the old tissue 1s more or less reduced in width. When the new tissue arises its width 1s the same as that of the con- tracted cut surface, hence the arc of the cut surface is equal to the width of the new tissue at its anterior end, the eleventh column of the table. It is desirable, however, to determine not only the width at this point, 7. ¢., the arc of the contracted cut surface, but also the width of the body just anterior to the region where it is effected by the local contraction. “These measurements are given in the tenth column under the heading, “ Width at posterior end of old tissue.”’ In Fig. 2 the difference in level of the two measure- ments is indicated by the transverse dotted lines. As is evident from the figure the “width at the posterior end of the old tissue” is measured on the tangent to the contracted cut surface at right angles to the longitudinal axis and the width at the anterior end of the new tissue 1s the arc of the cut surface. As in Fig. 2, there is usually a marked difference between these two measurements except in later stages and this difference represents the contraction of the cut surface. The last column of the table gives the number of the figure representing each stage, in order that comparison between the table and the figures may be readily made. The intervals are not exactly the same for Series 27 as for the other series, but the difference is not sufficiently great to prevent comparison. A few data are given to supplement the figures and table. Series 27 (Figs. 1-6). ‘The level of section 4 mm. from the anterior end is indicated in Fig. 1. The short anterior piece resulting from section contains nothing but the head region and a short portion posterior to the ganglia. ‘The margins of this part of the body are used chiefly for drawing the body forward and for the flying movements, not, like the posterior end, for holding to the substratum during locomotion. As a matter of fact this piece Studies on Regulation. 259 was incapable before regeneration occurred of adhering to the substratum by its posterior end. It advanced by means of its cilia and alternating extensions and contractions of the antero- lateral margins of the head but its posterior end did not adhere. yf Regeneration of the posterior portion began in the usual manner from the somewhat contracted cut surface and in the course of a few days the new part began to be used for attachment as the ani- mal advanced. ‘The first result of this method of use was the 260 C. M. Ghild. change in form of the new tissue from a rounded mass with convex margins to the more elongated condition with concave margins shown in Fig. 2. It is clearly evident that in this piece with short posterior end the mechanical conditions connected with locomo- tion are very different from those existing before section (Fig. 1). Before section this piece was continuous across the whole posterior end with the parts posterior to it, and therefore any tension to which it was subjected in consequence of attachment of the pos- terior parts of the animal must have been approximately parallel to the longitudinal axis. But after the regenerating posterior portion has begun to serve as an organ of attachment (Fig. 2, two days after section) the lines of tension are no longer nearly parallel to the longitudinal axis but converge toward the posterior end (Fig. 2), this being the part most frequently used for attachment. If the tissues of the body are at all plastic their relations must be altered to a greater or less extent by these new conditions. The effect of the tension must bring about elongation and decrease in width of the body, most marked posteriorly and decreasing toward the anterior end. And this is exactly what occurs. At the stage of Fig. 2 attachment by the posterior end has been possible only a Sinores time and has not as yet affected any marked change in the form of the piece. In Fig. 3 (twenty-seven days), however, the form is greatly altered. ae only has reduction of size in the old part occurred but its proportions are different (see table of measurements). Its length is 25 per cent and its greatest width 33 per cent less than originally; moreover, it is now much narrower at the posterior end than at the anterior end, whereas originally its width was about the same at both ends. The Figs. 4-6 show later stages in the process. ‘The change of form is not great after the stage of Fig. 3, but in consequence of the more rapid reduction in size of the old part as compared with the new, some change does occur. It is of interest to note that as the piece oradually becomes smaller and less active the length of the old part decreases more rapidly than its width, 7. ¢., it becomes relatively wider (Figs. 4- 6 and table). This change 1S always more or less evident in pieces of this kind and is exactly the reverse of what might be expected if the change in form were a “reduction to approximately normal proportions. ” Pieces of this kind always show a decrease in motor activity after several Studies on Regulation. 261 weeks without food. At the stage shown in Fig. 6 (seventy-five days) the specimen was much ese active than at the stage of Fig. 3. Its movements were slow and it did not adhere very closely to the substratum. The relative decrease in length of the old part is probably simply the result of the decrease in longitudinal ten- sion accompanying decrease in motor activity. ake new part does not show this change to such an extent since it is still in- creasing in amount, 7. ¢., relatively, at the expense of the old tissue, as is evident from the Figs. 3-6 and from the measurements of these stages in the table. The piece died about eighty days after section. Series 57 (Figs. 7-10). In this case section occurred near the middle of the body, 9 mm. from the anterior end. In consequence of contraction of the animal during section the cut was oblique on the left side (Fig. 8), but this does not affect the value of the results. This piece included portions of the region where the margins are used for attachment, and so was able to hold to the substratum and creep in the normal manner after section and before the regen- erating part became functional, 7. ¢., the regenerating posterior end was not the only posterior organ of attachment, as in the pre- ceding series. [he small protruding piece on the left side was much used for attachment even after the new tissue appeared. In consequence of the ability of the piece to attach itself by the margins and also because of the greater length of the piece the change in direction of the tension is much less than in the pre- ceding series. It might be expected therefore that the change in form would be less rapid as well as less in amount than in the preceding series. Figs. g (twenty-five days) and ro (thirty-seven days) represent later stages of the piece and it is evident that the change of proportion is relatively slight. At the stages of Figs. 8 and 9 the width of the posterior region of the old tissue 1s rela- tively less than at the time of section (see table) but in Fig. 10 the proportions are almost the same as at the time of section. Exam- ination of the figures and measurements will show that the width at the posterior end underwent a relative decrease during earlier stages, while in later stages a relative increase appears. In other words the piece first acquired a somewhat tapering form without much reduction in length but later reduction in length brought about a return to approximately the original proportions. 262 C. My Child: The apparent change in position of the pharynx 1s of some interest. ‘his is probably due on the one hand to reduction of the old anterior portion and on the other to regeneration of a 7 new posterior portion. That any extensive actual shifting in the position of the pharynx has occurred seems doubtful, although some degree of shifting of the internal organs within the body wall is perhaps possible. Studies on Regulation. 263 In general the change in proportions in this piece was less marked than in other similar pieces. The relative decrease in width at the posterior end of the old part was less than usual. Possibly the tissues of this individual were less plastic than in most cases, but I think it more probable that an individual pecu- liarity in the activity of parts of the margin during locomotion is EL. responsible. Throughout the course of the experiment it was noted that the small protruding part on the left side was very frequently and closely attached to the substratum, 7. ¢., 1t was used in much the same manneras atail, doubtless a consequence of its form and relation to the substratum during movement. But the attach- ment of this part was usually fallawed or accompanied by the attachment of the other parts of the lateral margins in this region of the body on both sides. “This method of use of the parts must 264 C. M. Child. retard or prevent the decrease in width in this part of the body to a certain extent, hence the unusual width even in late stages (Fig. 10) of the posterior part of the old tissue. ‘This, of course, determines the width of the new tissue at the level of the boundary between old and new. Consequently this also is wider than usual (compare Fig. 10 with Fig. 4). This piece was accidentally killed at the stage of Fig. Io. i | an Series 58 (Figs. 11-16). In this series a specimen somewhat smaller than the preceding was cut transversely near the posterior end of the pharynx (Fig. 11). Figs. 12-16 show the changes in proportion which occurred, Fig. 14 is probably not quite fully extended. ‘The decrease in width at the posterior end of the old part is marked in Figs. 12-14. “Toward the end of the experiment (Figs. 15-16) the change is an almost proportional reduction in size, though with approaching exhaustion there is some relative Studies on Regulation. 265 decrease in length. Fig. 16 represents the specimen ninety-six days after section. A few days later it died. Series 59 (Figs. 17-23). ‘The level of section in this series was some distance posterior to the posterior end of the pharynx (Fig. 17). Figs. 18-23 show the changes in form during one hundred and thirty-six days. ‘The changes in form of the old part are relatively slight in this case as might be expected since only the posterior end is removed and the changes in the direction of tension are only slight. But a relative reduction in the width of the body in the posterior part of the old tissue does occur during the first sixty-one days of the experiment as the figures and the table show, 7. ¢., during this time the old part has become rela- tively longer and more tapering. As the activity of the piece decreases during the later part of the experiment, however, a change in the reverse direction occurs; the piece becomes relatively shorter and broader until at the last stage measured (Fig. 23, 136 days after section) the posterior width of the old part is slightly g greater in proportion to the length than it was at the time of section. Circumstances made it necessary to conclude the experiment at the stage of Fig. 23, but there is no doubt that further relative decrease in length would have occurred had the piece been kept. 2. Discussion of the Experiments. In all the four series of experiments described the results are similar; in each piece the old part becomes relatively longer and more slender during the earlier part of the experiment. ‘The greatest change in all cases is in the width at the posterior end of the old part which undergoes greater reduction than the width at the anterior end, 1. e., the body assumes a more tapering form. ‘These changes 1 in proportion differ in degree according to the level at which section is made, being in general greatest in short pieces. As I have shown for Stenostoma (Child, 02) this 1s exactly what must be expected if the changes in form are the result of the changes in tension connected wah locomotion. It is evident that if these changes 1 in proportion continued the piece would finally attain proportions approximating those of the specimen from which it originated, 7. ¢., the characteristic proportions of the species. ‘This also is to be expected if these changes are due to 266 CHV Ghild: mechanical factors since in the typical animal under typical con- ditions those factors constitute a characteristic complex. But in all of the four cases described these changes are followed in later stages of the experiment by changes in the reverse direc- tion: the pieces become relatively sheer and broader until in some cases the width of the body 1s relatively greater than before section. [hese “inverse’ ’ changes i in proportion carry the piece farther and farther away from its original proportions. As was noted above the motor activity of the pieces decreases in marked degree long before death occurs and these changes are undoubtedly the result of the decreased longitudinal tension consequent upon the decrease in motor activity. As the tension decreases the effect of various internal physical conditions, capillarity, surface-tension, etc., becomes manifest and it may be also that a reaction to the altered conditions leads to reduction of the elongated parts, though this reduction may be in part mechanical. Whe occurrence of these changes of proportion in opposite directions disposes effectually of tie idea that these pieces possess some inherent capacity for assuming the characteristic proportions of the species. The piece will attain the characteristic propor- tions at least approximately provided the characteristic complex of conditions is present: changes in this complex result in changes in proportion and changes in the reverse direction may occur under certain conditions as I have shown The changes in proportion are not as rapid nor as marked as in Stenostoma or Planaria but they are similar in kind. ‘The tissues of Leptoplana are much firmer than those of the other forms men- tioned and are consequently less readily affected by mechanical conditions. Reverse changes in proportion have already been described for Stenostoma (Child, ’o2) and I have also found them in Planaria and other species. These four series of experiments on Leptoplana are sufhcient to illustrate the character of the changes 1 in pieces containing the cephalic ganglia. All other experiments of the same kind—some twenty series—afforded similar results. “The validity of the results can scarcely be questioned since those of each series are in a sense independent of the others. ‘The measurements of the various stages and specimens were not tabulated and compared until long after the experiments were concluded; thus the results obtained in a given series were not influenced by the results of other experi- Studies on Regulation. 267 ments. his fact renders the agreement between the different series all the more convincing. Moreover, the experiments show, I think, that it is not impossible to obtain fairly accurate series of measurements, even of animals so changeable in form as these. These four series also afford some interesting data bearing upon the questions discussed in my second paper upon Leptoplana (Child, ’o4b). As ees the amount of regeneration Series 27 (Figs. 1-6) and Series 57 (Figs. 7-10) are nearly equal but in Series 58 (Figs. 11-16) a amount of regeneration is much less and in Series 59 (Figs. 17-23) still less. As regards Series 27, 58 and 59 the result agrees with the conclusion reached regarding this point in the paper above mentioned, viz: that the amount of pos- terior regeneration is proportional to the size of the part removed. According to this Series 58 might be expected to show less regen- eration than Series 27 but as a matter of fact the amount of regen- eration is greater than in any other piece among the hundreds examined. This case is an individual exception but the only one observed. At all events comparison of these pieces from different series shows that the amount of regeneration is much less when the level of section is in the posterior half than when it is in the anterior half. The bearings of this fact were discussed in the paper referred to above. The form of the regenerated part requires only brief considera- tion. It is so manifestly determined in large degree by the me- chanical conditions connected with locomotion ae discussion is scarcely necessary. The decrease in width of the new part without corresponding increase in length which is most evident in Series 27 (Figs. 3-5) can scarcely be accounted for except as the result of longitudinal tension. The differences in the rapidity and amount of intestinal regen- eration are also well illustrated by the four series. ‘The rapidity and amount of regeneration is greatest in Series 27 (Figs. 2-6), slightly less in Series 57 (Figs. 8-10) much less in Series 58 (Figs. 12-16) and scarcely perceptible in Series 59 (Figs. 18-23). The difference is not merely absolute and dependent upon the size of the new part but is relative, the extent of the intestine in the new tissue being relatively greatest in Series 27 and decreasing with approach of the level of section to the posterior end. Moreover, in the later stages of Series 58 (Figs. 14-16) and Series 59 (Figs. 21-23) the intestinal branches in the new tissue 268 C. M. Child. gradually disappear. The branches do not simply become invisible or difficult to see because of loss of contents but the distal portions actually disintegrate. The disintegration occurs not only in the new parts but in the old as well, though not indicated in the figures, so that in these stages only the more central parts of the intestine remain. Some consideration of these changes has been given in previous papers (Child, ’o4b, ’o4c) but other species are more favorable for experiment along this line. ‘These cases sup- port the suggest ons reached in the other papers on Leptoplana regarding the effect of internal intestinal pressure on form and extent of the intestinal branches. B. THE EXPERIMENTAL PRODUCTION OF ABNORMAL FORMS. A considerable variety of abnormal forms may be produced experimentally in Leptoplana. Duplication of the anterior or posterior end may be obtained by the usual method, viz: partial longitudinal splitting from oneend or the other. Repeated cutting of ee parts is always necessary in order to obtain duplication and in most cases in spite of all precautions the parts unite, or else the repeated cutting causes the death of the specimen. Since the cephalic ganglia do not regenerate from other parts of the nervous system, duplication of the head is possible only when each part contains a considerable portion of one of the ganglia, 7. e., only when the cut lies very near the median plane. ‘Tails can be pro- duced from a cut surface facing more or less posteriorly anywhere along the lateral margin of the body provided the part can be kept ont uniting with es other parts. A full description of the various experiments is unnecessary since both method and results are similar to those described for other species of Turbellaria. A few of the experiments, however, are of interest and are given brief consideration. 1. Formation of a Tail Between Two Heads. This specimen was one of a series of ten in which the attempt was made to duplicate the head region by splitting the body from the anterior end along the median line to the pharyngeal region. Five good cases of duplication were obtained, after cutting several times, some with heads of- equal size, others with one head Studies on Regulation. 269 larger than the other. In each case where the cut separated the ganglia into halves or nearly so each half regenerated the other half and each of the two heads was similar to other cases of lateral regeneration in this region (Child, ’o4c). It was necessary to repeat the operation of longitudinal splitting from three to five times in order to prevent union of the two parts. Fig. 24 shows the anterior portion of the specimen in question after the operation of splitting had been performed three times. At this stage the operation was repeated for the last time. At the next examination, a week later, it was found that the cut surfaces resulting from the last opera- tion had united in a somewhat unusual manner (Fig. 25). The right margin of the left head overlapped the left margin of che right head. It is evident from Fig. 25 that the left margin of the right head is giving rise to an outgrowth in the posterior direc- tion which 1s situated ventral to the original body; in other words, from that part of the cut surface on the right head which did not unite with the opposing surface, a tail is regenerating. In Fig. 26 the new posterior end has elongated still further. It functions in all respects like the 25 typical posterior end of the species. [he animal uses it for attachment in creeping in the same manner as the posterior end of the main body. Since the right head is the dominant head of the specimen the functional activ ity of this accessory posterior end is fairly codrdinated with that of the other parts of this head. It will be noted from the figure that the right margin of the left head continues ventrally on ‘lie right margin of the new posterior end and the left margin ‘of the right head is continuous with its left margin. ‘The peculiar relations of this part to the other parts of the specimen make it especially interesting. It is highly probable that the attachment to the substratum of this part of the margin 270 C. M. Child. of the right head and the longitudinal tension exerted upon it are the most important factors for the formation of the tail, as was shown to be the case in Stenostoma (Child, ’03a). It is difficult otherwise to understand why a bilaterally symmetrical posterior end should arise from the extreme lateral portion of the body. The cases of formation of supernumerary posterior ends in Planaria and other forms are without doubt similar in character. An account of certain experiments along this line will be given at another time. 2. Experimental Duplication of the Posterior End. Attempts at duplication of the posterior end by partial longitudi- nal splitting succeed only rarely because the use of the margins and posterior ends for attachment during locomotion is such as to press the two cut surfaces closely together and union almost invariably occurs within a few days, no matter how often the operation is repeated. A few cases showing some degree of dupli- cation of the posterior end were obtained, but one case was of special interest since it indicates the importance of the mechanical tension as a factor in the regeneration of the posterior end. This case was one of a series of eight specimens each of which had been cut transversely through the middle of the body and then the anterior piece split longitudinally nearly to the ganglia (Fig. 27). After the first operation all the pieces united again, but after the second operation one piece was found in which the union was not complete (Fig. 28). In this piece the contraction of the longi- tudinal cut surfaces was so great that each half of the transverse cut surface which originally formed the posterior end of the piece had been drawn in into a position facing the median plane, 7. e., at right angles to its original position. In consequence of this contraction a part of the lateral margin of each half of the speci- men formed the actual posterior end. The posterior portion of the specimens with the two parts separated as widely as possible is shown in Fig. 29 on a larger scale (x 14, the other figures x 7). Here it is seen that the longitudinal cut surfaces have united except for a short distance at ‘the posterior end. ‘The originally transverse cut surfaces, though now nearly longitudinal, can be distinguished from the original longitudinal cut ‘surfaces by their concavity and by the amount of regeneration which has taken . Studies on Regulation. 271 place upon them. From each of these cut surfaces new tissue has grown out at right angles and in much larger amount than on the longitudinal cut snes | just anterior to these. It should be noted that the position shown in Fig. 29 was never taken by the specimen and has been used in the figure merely to show the parts without overlapping. The position of the two posterior ends usually approached that shown in Fig. 28, though when the animal was holding tightly to the pabereaciin the over- lapping was much greater than in the figure. Most commonly during ordinary creeping the two regenerated “posterior” ends 27 St were apposed and turned dorsally so that neither of them touched the substratum. ‘The new tissue was not much used by the speci- men for attachment to the substratum, the most posterior parts of the lateral margins being employed instead. ‘This functional substitution of the lateral margin for the posterior end 1s in itself interesting and determines certain other important features. ‘The parts of the lateral margins which formed the actual posterior end of the piece reacted to contact with the substratum in much the same manner as the posterior end in normal animals. If one of the regenerated “‘tails’’ happened to be in contact with the sub- stratum it often adhered to some extent. While the other tail 272 C. M. Child. applied itself to the dorsal surface of the first (Fig. 28). But the margins of the body bent across the posterior end adhered so much more closely that the regenerating tails were not subjected to the characteristic tension. Even when one or both of the posterior ends underwent temporary contraction and the new tails were stretched the contraction took place in a curve parallel with the curved margin of each half as indicated by the direction of the arrows below Fig. 28. It is easy to see that this peculiar form of contraction follows from the course of the longitudinal muscles in the curved posterior parts. From this description of the movements it is clear that the ten- sion to which the regenerating tissue on the originally posterior surfaces 1s subjected 1 is slight ‘compared with that in the typical case of posterior regeneration. The incurved lateral margins of the old part perform the function which in typical cases is per- formed by the new tissue on the posterior cut surface. And sec- ondly, since all muscular contraction of these posterior ends follows the curve indicated by the arrows in Fig. 28 it is clear that any tension to which the new tissue may be subjected in consequence of attachment during such contraction is approximately perpen-_ dicular to the cut surfaces. Attachment of one of the new tails Was sometimes observed under these conditions and actual stretch- ing perpendicular to the cut surface was visible. If we compare the amount of regeneration in this specimen with that in other pieces in which the cut surfaces had united we find that there is a marked difference. Figs. 28 and 30, both drawn to the same scale, represent two pieces of the same series at the same time after the operations. In the case shown in Fig. 30 the cut surfaces united fully, and the new tissue was subjected to the ty pical longitudinal tension and posterior regeneration occurred inthe typical manner and amount. Inthe other case the new tissue on the (originally) posterior cut surfaces could not be used in the typical manner, hence was subjected to slight tension only. The amount of regeneration in this case is only a small fraction of the amount in Fig. 30. ‘This case constitutes almost an experimental demonstration of the correctness of the conclusions reached in pre- vious papers (Child, ‘o4b, ’o4c), viz: that the mechanical conditions are important factors in determining the amount of regeneration. Another feature of importance in this case is the direction of regeneration from the posterior cut surfaces. On each side the Studies on Regulation. 273 new tissue grows out perpendicularly to the cut surface, 7. ¢., at right angles to the longitudinal axis of the body when the parts are in their usual position (Fig. 28). In my first paper on Lepto- plana. (Child, ’o4a) I showed that the direction of outgrowth of new tissue was determined, at least in part, by the direction of the mechanical tension to which it was subjected. As regards the present case it was pointed out above that when these outgrowths of new tissue are subjected to tension it is approximately perpen- dicular to the cut surface. Thus, in this case the direction of the outgrowth 1 is determined by the direction of the tension to which it is subjected. If the specimen were capable of holding the two ends in the position indicated in Fig. 29 and if the new tissue were much used for attachment there is little doubt that each outgrowth would soon become oblique with respect to the surface from which it arose and approximately parallel to the longitudinal axis of the body. But in the case under consideration the functional sub- stitution of the margins of the old part for the tail became more and more complete as time went on. The animal seemed to become more and more accustomed to the altered positions of parts and coordination apparently became more perfect. The form and relative size of the new parts did not change appreciably from the condition represented in Fig. 28. This case, like others described in previous papers, indicates how readily the course of regulation may be altered when the really essential conditions are changed. It also shows very clearly that the power of a piece to attain the characteristic form of the species is dependent upon characteristic functional activity. Repeated attempts to obtain other specimens of the same sort were unsuccessful. Usually the contraction of the cut surfaces was not sufficient to bring the transverse surface into line with the longitudinal. Sr Duplication of the Posterior End by Protruston of the Pharynx from the Cut Surface. Like the preceding, this case is interesting, not merely as an abnormality, but as adding to the evidence regarding the factors concerned in the growth oi new parts. The individual, a worm of average size, was cut transversely through the anterior end of the phary nx leaving only a small part 274 G. Me Child. of the old pharynx in the anterior piece (Fig. 31). Six days after section the anterior piece presented the appearance of Fig. 32. The regenerating posterior region was incompletely divided into two parts by the mass of pharyngeal tissue that protruded in the median line. It will be observed that the separation of the two tails is much more complete ventrally than dorsally. On the dorsal side a thin membrane joins them back almost to the tip of the protruding pharyngeal mass, while on the ventral surface they are distinct to the level of the old tissue. Fig. 33 is an enlarged diagram showing the conditions at the level of the dotted line in Fig. 32. The stippled mass in the middle represents the old pharyngeal tissue, on each side the portions of the tails in contact with the substratum are indicated and continuous with these is the thin membrane dorsal to the old pharynx. A few days later the old pharyngeal tissue dropped off, but the walls of the old pharyngeal pouch had already united with the body-wall so that when the pharynx fell away an opening facing postero-ventrally remained. The condition of the piece twenty-one days after section 1s shown in Fig. 34. ‘The regenerating posterior region has elongated considerably and only its posterior third is double. The regen- erated pharynx, like the body, is duplicated posteriorly. The relations of parts are more clearly shown in the enlarged diagrams 35 and 36. In Fig. 35, a dorsal view, it is seen that union between Studies on Regulation. 275 the two tails extends further posteriorly on the dorsal side than on the ventral; the region between the two longitudinal dotted lines in Fig. 35 is not in contact with the substratum as the animal creeps and at the anterior end of this space between the two tails the opening still persists. Fig. 36 is similar to Fig. 33 and indicates the relations of parts in the dorso-ventral plane at the level of the transverse dotted line in Fig. 35. ‘The two tails are less widely separated than in -earlier stages, hence the thin membrane between them is thrown into a dorsal fold as indicated in the figure. The postero-ventral opening in the space between the two tails was apparently connected with the pharyngeal pouch of the new pharynx and may therefore be regarded as a mouth. In the living animal I was unable to find any other mouth on the ventral surface. It was my intention to keep the specimen alive as long as possi- ble in order to determine whether this duplication was gradually obliterated and then to fix the piece and study by means of sec- tions. During the six weeks following the stage shown in Fig. 34 no marked changes except redticoon: in size Soecnered: At the end of this time the piece was lost. It is evident that the protru- sion of the mass of old pharyngeal tissue from the cut surface was the condition which originally determined duplication of the end. If this be admitted several questions arise at once. Of these we may consider first why the region dorsal to the old pharynx does not regenerate as rapidly as the regions lateral to it. ‘The pro- trusion of the old pharynx in a postero-ventral direction offers no obstacle to growth in the posterior direction of the parts dorsal to it. Evidently the dorsal region does regenerate, for the thin membrane uniting the two tails dorsal to the pharynx represents the regeneration from this region. But regeneration here is less rapid than in the regions lateral to the pharynx so that duplication appears dorsally as well as ventrally, though to a less extent. The factors which have served in so many other cases, viz: attachment to the substratum and tension consequent upon loco- motion are in my opinion the chief factors concerned here. It 1s evident that the protrusion of the mass of pharyngeal tissue in the postero-ventral direction prevents the new tissue which arises dorsal to it from coming into contact with the substratum. This 276 C. M. Child. was clearly seen to be the case by observation of the specimen during movement. Only the regions lateral to the pharynx could be used in the characteristic manner. The result of the duplica- tion in the complex of functional conditions is the duplication in structure. [he thin membrane dorsal to the pharynx elongates only as the tension exerted upon the two tails is transferred to it and so remains behind these parts in growth. But it may appear at first glance that the loss of the old pharynx a few days later alters eonditions and that after this there is noth- ing to prevent the median region from coming into contact with the substratum and elongating as rapidly as other parts. This is not the case, however, for as long as the two tails are used in the typical manner (Figs. 34 and 35) the median connecting region cannot come into contact with the substratum but must remain as a dorsal fold (Fig. 36). Indeed the fold is frequently more marked than before the es of the old pharynx since the two tails often lie nearer together than was possible when they were separated by the old phary nx. Thus, even after the loss of the old pharyngeal tissue the median region is subjected to tension only as the lateral parts are strongly stretched and so maintains the relations with other parts which were originally determined by the presence of the mass of pharyngeal tissue. Incidentally this case shows very clearly that the terminal por- tions are formed first in the regeneration of the posterior end. The postero-ventral “mouth” bemveen the two tails which was at first situated at the level of the cut surface is carried posteriorly as regeneration occurs and at the stage of Fig. 34 lies far behind the cut surface. Occasionally when the animal drew back suddenly from some object which its head had touched the two tails became more or less extended laterally each forming an angle of about 45° with the longitudinal axis. Under these conditions the median region was sometimes in contact with the substratum, but since this position was not often taken and was never maintained for more than a short time no appreciable effect could be expected. This case is interesting chiefly because it constitutes valuable evidence in favor of the view that the functional conditions and among them the tension to which parts are subjected are important “formative” factors. Studies on Regulation. i) al sel D. REGULATION AND EMBRYONIC DEVELOPMENT. The problem of regulation must be regarded as a part of the problem of development; indeed, as has already been abundantly demonstrated the investigation of regulatory phenomena and processes 1s of endaniental importance as a means of throwing light upon the problems of embryonic development. On the other hand the phenomena of regulation and of ontogeny are in certain respects so different that caution is always necessary in extending conclusions from the one field to the other. Probably the most important held of investigation in connection with regulatory phenomena is the Actenmindtion by experimental methods of the conditions and _ processes of morphogenesis. Exact knowledge of these conditions and processes is of funda- mental importance in biology, not only directly as an addition to _the data-of science but airecily as well, since it affords the only means by which we can ever hope to attack intelligently certain other problems, such for example as those of Heredien and evolu- tion. It is clearly impossible to obtain any intelligent conception regarding the nature of the germ cells before we have determined the relation between the adult organism and the cells from which it arises. It is scarcely too much to say that the only satisfactory method of determining what is inherited is the method of elimi- nation: at any rate if we can determine experimentally what 1s not inherited we shall be in a far better position to discuss the nature of inheritance. Objection to these statements may be made on the ground that it is possible to determine by direct experiment, 7. ¢., by hybridization or other forms of breeding, that certain “characters,” ¢. g., a color or a structural feature are inherited in certain cases. Such-an objection rests, however, on a total misunderstanding of my position. The point | wish to make is that the color or the structure is not itself inherited, but only an unknown something that can give rise to the one or the other under certain conditions. ‘Uhis is of course familiar to all, yet it seems to be forgotten again and again. Qualities, relations, and reactive capacities of protoplasm not “characters” are inherited. Theories of heredity which regard the germ cell as containing a multitude of elements, each representing some character or group of characters in the developed organism are the monstrous offspring of a morphology divorced from physiology 278 C. M. Child. and can serve only to delay the advance of biology. Exact experi- mental data concerning the phy siology of dev elopment | constitute the most effective weapons for combating and overcoming errors of this kind. Within the last few years they have been forced repeatedly to retire from one defense to another. But there is evident in much of the work of recent years upon developmental physiology a certain inclination to regard the problems of morphogenesis as at present insoluble. This mani- fests itself in Driesch’s later work in the form of a vitalistic or “autonomistic’’ hypothesis based on certain phenomena of form-regulation, another interesting though perhaps logical conse- quence of the separation of morphology and physiology. As I have shown in several papers (Child, ’02, 03, ’o4a, ’o4b) certain of the phenomena of regulation which appear so mysterious to Driesch and others are so only because they are wrongly inter- preted. Others, like Morgan, who are less extreme hold that while the problem of morphogenesis is fundamentally a physico- chemical problem yet it 1s at present and perhaps will always be insoluble. These views undoubtedly take their origin from the morpho- logical conception of life, the belief that the essential feature of organic development is the production of structure. When we cease to consider “‘the tendency to return to normal proportions,” “form-entelechies’’ and other similar and fundamentally mor- phological abstractions, and regard the organism as a complex functioning in a characteristic manner morphogenesis appears as one of the results of this functional activ ity of the complex. The term function is used here to include all the activities of the organ- ism, all transference and transformation of energy. Undoubtedly qualitative differences must exist as a basis for complex function. The point of importance is that the organism is primarily a fune- tional rather than a morphological complex. It is the qualities, 1. e., the capacities for functional activity that are transmitted from individual to individual and from period to period. ‘The form of the organism is in general the result of its own activity under characteristic external and internal conditions. . Roux has recognized two stages in development, an organ- forming and a functional stage. During the first period the vari- ous organs develop without “functional” activity, while during the second increasing “functional”’ activity and interdependence Studies on Regulation. 279 appear. In other words, during the first period the machine is constructed and during the second it functions. Here again the morphological conception of development appears as the basis of this analysis. In my opinion, all stages of development are to be regarded as functional though the kind of function and its visible results differ. It is important, moreover, to distinguish between the visible substances or those which future investigation may prove to exist in the nucleus or cytoplasm of the germ-cell and the structures into which they develop. Suppose’a certain substance or region of the ege can be followed to the entoderm of the larva. It is not con- ceivable that this substance or region if isolated can form a typical intestine, although its cells may differentiate into typical intestinal cells. In other words this region may continue to function in the characteristic manner after isolation, and each of its cell units may undergo the differentiation corresponding to this function, but the typical form of the whole does not appear because the typical relations of the elements to each other, and to the environment are not established. ‘The differentiation of the cell units is doubt- less in large part the result of their chemical constitution while the formation of a characteristic organ, the intestine, is largely the result of physical factors, the natural pressures and movements of cells, surface tensions, tensions due to conditions in other parts, pressure of fluids, etc. Development of the typical form is due to these conditions as well as to the character of the substance itself. And these conditions have been ignored to a large extent in the study of development. Morphogenesis is very commonly regarded as the result of the composition of the substances in the germ-cell. From this point of view have arisen the hypotheses which regard morphogenesis as analogous to crystallization, and the theories of formative stuffs which, though perhaps correct for certain elements of structure, are, Be rilicliss, without general significance because they are based on inadequate conceptions. ‘The relation between the composition of the germ-cell and the structure of the developed organism cannot be a fagrect one, but Is rather exceedingly remote. ‘To look for the equiv alents of mor- phological characters in the germ-cell must involve us in many difficulties, because most so-called morphological characters are primarily typical space-relations of masses and these necessarily 280 Cou Ghild: differ from anything in the germ-cell. When, however, we con- sider the matter from the physiological side we can trace a certain relation between the cell and the fully formed organism; both exhibit characteristic functional activities and frequently we can trace a given functional complex from the cell through on- togeny by its effect upon organization. ‘This, I think, is the real significance of His’ “‘organbildende Keimbezirke”’ and Wilson’s (04) formative substances in_ the ege-cy toplasm. But the designation of these regions as formative! regions or substances is a morphological mode of expression which seems to me misleading. All living complexes are formative, or may be under proper condi- tions; but they are formative because they are functional. Let us consider briefly a case In point, viz: the typical form of the pos- terior end in Leptoplana or Stenostoma. ‘This is a characteristic of the species, yet I have shown that it depends largely upon mechanical conditions of tension for its formation. The tension is the result of typical activity on tissues of a typical physical quality 1 in a typical environment. ‘The typical activity depends again on a ty pical constitution, and so on. What is transmitted from the previous generation? Certainly not tail-germs, nor tail- forming subseanees: nor anything that is directly related to the tail of the adult. In this case thee “tail” is primarily a physical arrangement of material resulting from a characteristic complex of phy sical and other conditions. In other cases the analysis may proceed on widely different lines but the result must be similar. I believe it can be shown that morphological conceptions of development are all anthropomorphic in character and related to our conceptions of man-made structures composed by adding element to element. It becomes more and more evident as our knowledge increases that these conceptions must be discarded. But we cannot as yet substitute for them any adequate conception; we can compare life to nothing but itself. It is perfectly evident that with the physics and chemistry of the past and present we can never hope to interpret the phenomena of life, but we are entering on a period which promises that our physical and chemical conceptions will be as profoundly transformed by increasing knowledge of the processes of living organisms as our conceptions of life ever were by the adoption of physico-chemical hypotheses. Studies on Regulation. 281 The phenomena of regulation in the broadest sense constitute at present one of the most important and promising fields for work. Within the last few years many of our conceptions regarding devel- opment have been profoundly modified by the bestiles of experi- mental work along these lines and there can be little doubt that they are destined to still greater modification. Whatever modifies our theories of development must alter our ideas regarding inherit- ance and the nature of the germ-cell, both problems which are receiving much attention from morphologists. ‘The physiological investigation of development will probably afford in future far more numerous and exact data regarding the nature of inheritance and other fundamental problems of biology than any other field of research. The phenomena of regulation differ from those of embryonic development as the conditions differ in the two cases. Many parts of the field are accessible even at present to experi- mental methods and there can be no doubt that in future it will be possible to extend control of them much farther. Roux (95), maintains, however, that two distinct categories of development “typical” and ‘ ‘regulatory ”’ must be recognized and that the mechanisms concerned in the formation of a given struc- ture in typical development may be different from those which come into play in regulation. Hypotheses of this kind only increase instead of dimniaish our difficulties and, moreover, they are based on theoretical considerations and not on observation and experiment. If we admit instead that form and structure are results of reactions to conditions it is evident that changed condi- tions may change both the processes and the results. Regeneration of a part removed may differ widely from the ontogenetic develop- ment of the same part, but, as I have attempted to Shona in this and preceding papers on Leptoplana (Child, ’o4a, ’o4b, ’o4c) the con- ditions to which the regenerating tissue 1s subjected are different from those to which the part is subjected in ontogeny; in the one case the tissue arising from the cut surface 1s Eoaneara with a fully developed part, in the other all parts are developing together. It is to be expected therefore that the course of regeneration will be briefer than that of ontogeny and, moreover, that it will differ in various respects, according as particular conditions differ. Notwithstanding these difrerenacss indeed often because of them, the phenomena of regeneration are of great importance in the physiology of development. But other methods of regulation 282 Ci Ne Ghild: occur, one of the most important being what Driesch (‘o1) has designated as redifferentiation. I think it probable that in many cases the so-called ‘“redifferentiation” when subjected to a closer study will turn out to be merely differentiation. Cases of this sort in which the formation of new tissue is involved, such for example as the formation of the new pharynx in the old tissue in Planaria, differ from regeneration proper, 1. €., the outgrowth of new tissue from the cut surface, chiefly in that growth is not localized at the cut surface but is distributed through a larger or smaller portion of the old tissue. But why should the new growth be localized in the one case and not in the other? In attempting to answer this question it is necessary to anticipate somewhat and state certain conclusions from my experiments for which the data have not yet been fully given. These will serve merely as suggestions to make clear my point of view. In cases where regeneration, 1. €., outgrowth from the cut surface, takes place it will be found that the old part remains essentially a part as regards function; functional substitution for the part removed does not occur. But when proliferation at the cut surface begins as the direct result of the altered conditions the functional conditions to which this region is subjected are more or less similar to those characteristic of the part removed and growth, 7. ¢., regeneration of this region and its differ- entiation into a part more or less like that removed occur. If, on the other hand, the old part is capable of performing more or less perfectly the functions of the part removed, that is to say, if its reactions are modified by the changed conditions so as to resemble those of the part remov ed, corresponding changes will occur in the structure of that portion which resembles functionally the part removed and it will be “redifferentiated” into a part like that removed. As an example let us compare the case of Leptoplana (Child, ’o4a, ’o4b, ’o4c) with that of Stenostoma (Child, ’02, 03): Regulation after removal of the posterior end in Leptoplana is wholly or almost wholly regeneration, 7. e., growth from the cut surface, while in Stenostoma the posterior region of the old part “redifferentiates”’ into the new tail. When we compare the behavior of the pieces after removal of the posterior end, we find that in Leptoplana functional substitution of the posterior end of the piece for the original posterior end does not occur or occurs only in slight degree, while in Stenostoma the posterior end of the piece functions almost perfectly as a tail. This brief comparison Studies on Regulation. 283 is perhaps sufhcient to illustrate the point. I believe that the absence or occurrence of functional substitution of other parts for a part removed are important conditions determining whether regeneration or “‘redifferentiation”’ shall occur. In certain cases of regulation, as for example one form of regula- tion occurring in Clavellina (Driesch, ’o2) the old structure dis- appears more or less completely and the piece seems to return to the embryonic condition. Out of this apparently undifferentiated mass a new complete individual arises by processes more or less similar to those of ontogenetic differentiation. ‘This is apparently a case of true redifferentiation. Our knowledge of cases of this sort is very incomplete as yet; we do not even know exactly to what extent the old structure disappears, but it is evident that we have here something widely different from regeneration and approach- ing more closely to embryonic development. In the first place, according to my point of view, the disappearance of the original structure of the piece is only incidentally a part of the regulative process; the old structure disappears simply because the conditions which maintained it are no longer present. ‘The previous differ- entiation has not destroyed the capacity of the tissues for reacting to altered conditions and the first effect of this altered reaction is the disappearance of the old structure. “The transformation of the part into a “whole” is probably identical with the loss of the specification which resulted from the particular conditions to which it was subjected; it is thus a negative rather than a positive change, the loss of visible differentiation rather than the acquisi- tion of new potentialities. [he development of the new individual from the undifferentiated mass occurs in much the same manner as in typical ontogeny. ‘The different origin of the cell-mass in the two cases does not constitute a fundamental difference, though it may be found to determine some differences in detail. From what has been said it follows that the greater the degree of differentiation, or in other words the greater tie specialization of conditions in different parts of the organism, the greater will be the difference in the parts and the less ‘he capacity ae altering the reactions in correspondence with altered conditions. Hence we may expect “redifferentiation” to occur only in relatively simple forms while regeneration, or destruction of other parts followed by regeneration, or finally destruction without regeneration may follow removal of a part in more complex forms. 284 C. M. Child. SUMMARY. 1. During regulation typical changes in proportion of the old parts occur, consisting in a relative decrease in width and increase in length of the body. ‘These changes in proportion are greatest in the. posterior region of the piece involved and are greater in short than in long pieces. In cases where a marked decrease in motor activity occurs, as for example during the later stages of regulation of pieces without food, changes of proportion in the reverse direction occur. 2. These changes in proportion are primarily due to mechanical factors. The eae elongation and decrease in width are largely the result of the tension consequent upon the use of the regenerating posterior end as an organ of attachment during locomotion. ‘The change in direction of the tension differs according to the length of the piece, being greatest in short pieces. The reverse changes in proportion are the result of marked reduction in the longitudinal tensions consequent upon a decrease in motor activity. Under these conditions the piece approaches more or less a rounded form in consequence of internal pressures, surface tension, capillarity, etc. 3. Certain cases of experimental duplication of the anterior or posterior end afford strong evidence in favor of the view that the direction and amount of posterior regeneration and the form of the regenerated part are determined in large degree by the func- tional conditions connected with motor activity, the mechanical tension being probably the chief factor. 4. The experimental analysis of regulative phenomena consti- tutes one of the most effective methods of attacking the problem of morphogenesis and affords valuable data for the problems of heredity. The results of this feld of work indicate that physio- logical conceptions and hypotheses must be substituted for mor- phological. Hull Zodlogical Laboratory, University of Chicago. May, 1904. Studies on Regulation. 285 BIBLIOGRAPHY. Cuiip, C. M., ’02.—Studies on Regulation. I. Fission and Regulation in Stenos- toma. Arch. f. Entwickelungsmech., Bd. xv, 2 and 3 H., 1902. ’03.—Studies on Regulation. II. Experimental Control of Form-Regula- tion in Zooids and Pieces of Stenostoma. Arch. f. Entwickelungs- mech., Bd. xv, 4 H., 1903. : ’03b.— Studies on Regulation. III. Regulative Destruction of Zooids and Parts of Zooids in Stenostoma. Arch. f. Entwickelungsmech., Bd. xvii, 1 H., 1903. : *o4a.—Studies on Regulation. IV. Some Experimental Modifications of Form-Regulation in Leptoplana. Journ. of Exper. Zool., vol. i, No. 1, 1904. *o4b.—Studies on Regulation. V. The Relation between the Central Nervous System and Regeneration in Leptoplana; Posterior Regeneration. Journ. of Exper. Zool., vol. i, No. 3, 1904. *o4c.—Studies on Regulation. VI. The Relation between the Central Nervous System and Regeneration in Leptoplana; Anterior and Lateral Regeneration. Journ. of Exper. Zool., vol. i, No. 4, 1904. Driescu, H., ’o1.—Die Organischen Regulationen. Leipzig, 1901. Morean, T. H., ’99.—Regeneration in the Hydromedusa Gonionemus vertens. Amer. Nat., vol. xxxii, 1899. °oo.—Regeneration in Planarians. Archiv. f. Entwickelungsmech., Bd. x, 1900. Roux, W., ’95—Gesammelte Abhandlungen. Leipzig, 1895. Witson, E. B., ’04.—Experimental Studies in Germinal Localization. I. The Germ-Regions in the Egg of Dentalium. Journ. of Exper. Zool., vol. i, No. 1, 1904. THE FORMATION OF CENTROSOMES IN ENUCLEATED EGG-FRAGMENTS: BY NAOHIDE YATSU. Witnu 8 Ficures Pee MET OAN CON Ms erat Tere soo sister io cisatstr Poe Malice se pa oe aes eaite ewes cecees 287 II. Solutions and Precautions Against Accidental Fertilization................00.-00.. 000 289 IIT. Experiments on the Egg at the Metaphase of the First Maturation Mitosis................ 291 A. Enucleated Fragments Treated with CaCle Solution, Heat-sterilization............ 291 dome Method Sqrtvarstuccasrtere ice mecinie Noe cre re NS Tare a He ele Ee eo. 04 MOOR EOE 291 Gra Cy tasters, sou died cinder ect yt ae rese oxecye ete veteta eos s) ts oa 2 ids oravoroidin sic Classe ve 291 Grey tasters: studied im) SeChOnS en... cei seieie = = 15 een Pier state) Slee eve cciarassee ais 293 d. Nature of the Central Granules, and Development of the Cytasters......... 299 B. Enucleated: Fragments Treated with CaCl» Solution, Time-sterilization............ 301 C. Enucleated Fragments Treated with MgCls Solution, Heat-sterilization............ 301 IV. Experiments on the Egg before the Dissolution of the Germinal Vesicle; CaCls Solution, RG Ate SLEW c/a td OMe reese eye te ne ase) sya telea ete ones debe oslo ors ovsyoiey heise) cies. «idol ard ise. slereye ore 303 Pee NevicaOieleitelahune pte yrs sieve ayepayerseretsyrc tare isis Aesevere seceiias eae aies etere so taelehs apts BO 304 WL. (C@nIGSHOINS Geeta as ole Sierras Oey CREE eae ISS CO a ar ae a 308 WIL, SUNIRZET? 66 do 4.0.08 Bo OS OC ERS ERI Eo CR Gaeta Setar eer 310 WUE, LDARERATINE 2: 5 Gia oc, Oren cI COTE NEO Ck CHR eee coe eye ean ay 311 I. INTRODUCTION: Whether or not an aster containing a centrosome (centriole) may arise in the egg-cytoplasm independent of preexisting centers is a cytological problem of high interest. [his seemingly difficult question may be decided by a simple experiment. If we are able to produce an aster with the centriole in an egg-fragment con- taining no preexisting centrioles, we cannot escape the conclusion that these structures may be formed de novo in the egg-cytoplasm.’ 1The main part of the present paper was carried on in the summer, of 1904 under a grant from the Carnegie Institution, to which I wish to express my sincere thanks. To Prof. E. B. Wilson I acknowl- edge my great indebtedness for his kindly advice and criticism during the progress of this work. I am also under obligation to Prof. J. S. Kingsley, Director of the Harpswell Laboratory, for his kindness extended to me in many ways during my stay at his laboratory. *The presence or absence of the centrosome, i. e., the larger body surrounding the central granule or centriole, need not concern us here; for according to the latest work (for example, of Vejdovsky and Mrazek, Meves, Bouin), the centrosome is a periodical accumulation of the special substance, centro- plasm, around the centriole, which alone can be considered as a constant and autonomous structure. 288 Naohidé Yatsu. This method of examining the problem was first employed by Wilson in 1901 by shaking to pieces the unfertilized eggs of Toxopneustes and treating the enucleated fragments thus obtained with MgCl, solution. Le found, upon studyi ing the living frag- ments, chat not only do asters appear in such fragments, but also that some of them have the power of division like the normal asters. In sections he observed that the asters thus formed con- tain in some cases the centriole. He, therefore, drew the con- clusion that these centrioles must have been formed de novo. Subsequently Meves and Wassilieff almost simultaneously raised objections against Wilson’s method and cast doubt upon his results on a priori ground. Meves thinks that by shaking the egg center may be ehtowm out into the cytoplasm, while Wascihet pilees the view that shaking may bring about the flowing out of the nuclear fluid from the ege- nucleus. To meet these criticisms Professor Wilson suggested to me in February, 1903, to carry out similar experiments on enucleated fragments obtained by cutting eggs individually. In the summer of 1903 I made series of experi- ments at the Harpswell Laboratory on the eggs of Cerebratulus lacteus and obtained clear evidence that cytasters do appear in such fragments, provided the egg be cut after the first maturation figure is formed (which in this egg occurs before fertilization); no cytasters are found, however, if ie operation be performed before the germinal vesicle has faded. [ tried similar experiments on the egg of Echinarachnius parma and found that here cytasters arise in nucleated fragments from the matured egg. In the meantime Petrunkevitsch independently attacked the same problem in enucleated fragments obtained both by shaking and by cutting the eggs singly. In none of the fragments did he find asters containing a centrosome. He was, themrone: led to the con- clusion that in the whole eggs the centrosomes in the cytasters are the division products of ‘the egg center. “The obvious inade- quacy of his evidence has been pointed out by Wilson in a brief rejoinder (’04).1 In 1904 I undertook a repetition and extension of the experi- ments of the previous year during my stay at the Harpswell Laboratory, confining my attention to the ege of Cerebratulus lacteus. Fortunately. I obtained very consistent and constant 1A more detailed historical review of literature I shall take up later on (see p. 304, et seq.) Centrosome in Enucleated Ege-Fragments. 289 results; in fact, cytasters appeared in almost all cases. In sec- tions of the enucleated fragments, in which asters were produced, I found that in all the cytasters the centrioles were present. The egg of Cerebratulus is better suited to our present purpose than that of sea-urchin. The egg, when removed, has a large germinal vesicle. In some twenty minutes the nuclear membrane fades away, and in an hour or so the egg being still unfertilized, the first maturation mitosis reaches the metaphase. The mitosis, however, does not proceed beyond this stage, unless fertilization takes place. £— Onevean; therefore, i in the nemertine egg trace the progressive changes of condition in the cytoplasm from the primary oocyte onward. Moreover, the asters of Cerebratulus have very well defined centrioles, which seem to have, even in the division stages, far greater power to resist the action of various fixing fluids than those of the sea-urchin egg. To avoid confusion I shall follow sicily Boveri's definitions of centriole, centrosome and centroplasm, using the term “‘aster” for the whole structure including the ray system, archiplasm, centroplasm and centriole. ‘The aster without the centriole (if such aster exist) I shall call “pseudaster”’ (= Boveri's pseudo- sphere). In accordance with Wilson I use the term “cytaster”’ for an aster or pseudaster which is unconnected with nuclear matter. Il. SOLUTIONS AND PRECAUTIONS AGAINST ACCIDENTAL FERTILIZATION. The means of producing cytasters that were tried were: shaking, ether, MgCl,, CaCl,, KCl, and NaCl. Of these the first two did not cause any perceptible change on the eggs, while the other four modified mitoses and produced the cytasters in various degrees. The following solution of CaCl, proved best of all; it was, there- fore, used almost exclusively for the experiments on enucleated fragments: 1] did not succeed in producing the normal polar bodies artificially. Although, as a matter of fact, in a small percentage of the CaCl eggs and the CaClo+KCl eggs (the latter being Professor Wilson’s material) one or two polar bodies were extruded, yet in these cases the mode of maturation was so abnor- mal that it was thought undesirable to use the eggs thus matured for other purposes. 290 Naohidé Yatsu. 55 per cent solot Cal, (= 5 m.-CaCl) oe en eee I part. DEARWATET oon Mite e es eter eet ee pe g parts. The two solutions found to be the next best were: 14:6 per cent-of NaCl (=2%m: NaCl)... 22 ee I part. DOA -WARED J 2-52: 50 855.4 ee Henk oa soe ee ee 2 parts. Losospercent of KCl (=22ms Cl). een «ce aie I part. SEAS WALCR eo nso cheese ars_0h0/)g/epereinis sete ieee = eee ee 2 parts. it may be stated that there are only slight differences in action of these three solutions, while that of MeCl, solution is totally different from the others. The precautions taken against accidental contamination of the eggs with spermatozoa were as follows: Sea-water was first heated to 80° C. for twenty minutes, cooled down to the original temperature (20° C.) and well-water was added to make up the loss by evaporation. ‘The bottle containing the water thus steril- ized was aerated by violent shaking. The female worms were kept for two days in an aquarium separate from the males. To obtain the eggs for experiment a piece about an inch long was cut out of a ripe oemale with a pair of sterilized scissors and was sub- merged in fresh water for five minutes. After that the piece was washed with sterilized sea-water to get rid of the eggs, which had been squeezed out in fresh water, since the eggs thus discharged might have undergone some pathological changes. The piece thus treated was chopped up with a pair of scissors sterilized with fresh water. Of course, all the dishes used for experiment were washed with fresh water. For each experiment with the cytasters two lots of eggs from the same piece were used as controls to see if the particular lots of eggs were healthy. A. One lot of eggs was examined after five hours’ sojourn in the sterilized sea-water.. Most of them were at the metaphase of the first maturation mitosis, while in a few eggs the germinal vesicle was found intact. B. The other lot of eggs was mixed with sperm-water:ten minutes after release. Fertilization and subsequent development took piace normally as in those eggs kept in ordinary sea-water. 1Dry crystals dissolved in well water. Centrosome in Enucleated Egg-Fragments. 291 Ill. EXPERIMENTS ON THE EGG AT THE METAPHASE OF THE FIRST MATURATION MITOSIS. A. Enucleated Fragments Treated with the CaCl, Solution Heat-Sterilization. a. Methods. ‘The eggs were released in the sterilized sea- water, the precautions already stated being taken. After an hour and half the operation was begun. Had the eggs been fertilized they would have reached the two or four-cell stage by that time. As a matter of fact, none went beyond the metaphase of the first maturation mitosis. [he eggs were cut singly into two with a lancet known as Jaeger’s straight keratomy knife. Both the nucleated fragment (7. e., the one containing the mitotic figure) and enucleated one were put for five minutes in separate dishes with the sterilized sea-water so as to give them time to round up, since sudden contact of fresh cut surface witha salt solution seemed injurious. ‘lhe enucleated fragments were examined and drawn Then they were transferred into the solution of CaCl, prepared with the sterilized sea-water. (See p. 290.) In it they were kept for an hour or sometimes a little oe (This length of time was found to be the right one from the experiments ae on the entire eggs.) [he enucleated fragments thus treated were then put back into the sterilized sea-water. [he water was changed once. After from five to ten minutes the fragments were studied in a compressorium, but not compressed. Some of the fragments subjected to the above treatment were fixed with acetic sublimate (saturated solution of sublimate 98 parts plus glacial acetic acid 2 parts). When they reached go per cent alcohol they were stained with erythrosin and, after clearing, they were fastened on a piece of ulva by means of celloidin-clove-oil. ‘The iron-alum- hematoxylin method was used for all sections. While the enucleated fragments were in the CaCl, solution, the nucleated ones were stained with aceto-carmine to see that the two ends of the first maturation spindle were not injured by the operation. b. Cytasters Studied in Life. First I shall take up one particu- lar case of the formation of cytasters as anexample. ‘The egg was cut in a plane a little below the equator. In the animal half one could see a dumbbell- shaped clear area, indicating the first matura- tion figure (Fig. 1a), while in the vegetative half not a single clear spot was present (Fig. Ic). The animal half was stained with 292 Naohidé Yatsu. aceto-carmine. As is shown in Fig. 1b, there came into view a complete mitotic figure. The enucleated fragment, after five minutes’ stay in the sterilized sea-water, was transferred into the CaCl, solution, where it was first plasmolyzed. (Fig. 1d.) In Ib Fic. I (X 330). Ia, Nucleated half with first maturation figure. 1b, the same stained with acetocarmine. Ic, enucleated half of the same egg before CaClo treatment. sd, the same plasmolyzed in CaCl» solution. Ze, the same transferred to sterilized sea-water. zf, the same after the appearance of the clear area containing cytasters (section shown in Fig. 8). ? Centrosome in Enucleatcd Ege-Fragments. 293 the cytoplasm no visible change took place while in the salt solu- tion, nor could even the primary radiation be seen. After an hour it was put back into the sterilized sea-water, afterward the water was changed. The fragment became perfectly spherical (Fig. re), but there was no visible indication of the cytaster forma- tion. After about ten minutes a clear spot with rays around it appeared near the center of the fragment. (Fig. 1f.) This cen- tral area grew very rapidly, reaching after half an hour almost the size of the germinal vesicle. It | to enlarge at this stage. During the growing period the rays became meetin, so that the clear area gave the appearance of a vesicle. It should here be noted that he size of the fragment did not change perceptibly during the formation of the clear area. (cf. Fig. te and 1f.) Gis above case with a large spherical ential area is the com- monest mode of appearance of the cytasters, while quite often the clear area has an irregular outline or a deep indentation on one side. Fig. 2a shows an enucleated fragment, in which, under the same treatment, appeared two clear areas with fine radiation. These two made their appearance simultaneously at two separate spots. They, therefore, are not the division product of one original aster. [he stained nucleated half from the same egg is represented in Fig. 2b. In one enucleated fragment three clear spaces of about the same size appeared, as 1s shown in Fig. 3. In another three areas of different shape were found. (Fig. 4.) In still another case a splendid display of the cytasters was seen (Fig. 5a), dozens of small asters being scattered throughout the fragment; the enucleated half of this egg is drawn in Fig. sb. c. Cytasters Studied in Sections. INGE enucleated fragments subjected to the CaCl, treatment were fixed after from twenty to thirty minutes’ sojourn in the sterilized sea-water and cut into sections. One of them had no asters in it; the mode of appear- ance of the cytasters studied in the remaining eight fragments may conveniently be classified in three categories: a, the cytas- ters found throughout the cytoplasm; , one single large aster at the center of awe fragment, and c, a group of cytasters in a large central clear area. a. In four fragments several cytasters made thier appearance 1This fragment has been drawn greatly compressed. Some twenty asters are left out in this figure owing to the difficulty of drawing all of them in perspective. IE (X 330). Fic. nucleated fragment from the same egg st. 2b, h three cytasters. 4, 2a, Enucleated fragment with two cytasters. enucleated fragment with h many cytasters. 5), nucl t wi 3, enucleated fragment ine. h acetocarmi cytasters; three of the wit wit - 5a, enucleated fragment m have fused fragment from the same egg. Centrosome’ in Enucleated Egg-Fragments. 295 throughout the cytoplasm, as is shown in Fig. 6, corresponding to the pieces studied in living state. (Fig. 2a, 3, 4 anil 5a.) Inone of the fragments as many as twelve cytasters were found. In case a few cytasters are produced they have a tendency to come together near the center of the fragment. Another point to be noted is that the larger the number of cytasters in a fragment the smaller the aster. ‘The central portion of the fragment stains lighter than the outer. Besides fine yolk granules large ones of various sizes are found. ‘The latter kind" of granules is more thickly disposed near the periphery than the center. b. In one fragment a single aster is found near 6 the center with an en- larged centrosome. (Fig. 7.) The general charac- ter of the cytoplasm is the same as those of the four fragments described un- der a. c. In three fragments I found a group of cytas- tersin a central clear area. This type occurs more frequently than any other, as I learned from the study of living fragments. Fig. 8 is a section through Fic. III (X 913). the same fragment repre- 6, Section of an enucleated fragment, in which several sented in Fig. bie he cytasters have been produced by CaCle solution; four Creve —C y GO p ] asm is cytasters are seen in the section; the centrioles of the two packed with large yolk cytasters are in the next section. granules, which look as if they were crowded together by the central accumulation of hyalo- plasm. ‘They stain very dark and resist extraction by the iron- alum solution for a considerable length of time. In fact, when these granules become dark blue both the rays and centrioles are found completely decolorized. At the center there is a large clear area of almost the size of the germinal vesicle. In it some two or three dozens of cytasters are seen, most of them being situated near the periphery. Each aster bulges out a little tow ard 296 Naohidé Yatsu. Fic. IV (X 903). 7, Section of an enucleated fragment, in which a single cytaster has made its appearance. Notice enlarged centrosome and many centrioles in it. 8, section of the enucleated frag- ment, represented in Fig. 1f, with central clear area containing many cytasters and yolk-islands. the yolk part, so that in sections as many inden- tations are present as the number of intervals between two asters. It should be noted that the yolk granules near the clear space in over extracted preparations shows a radial arrange- ment (Dotterstrahlung of Hacker), and clear streaks run between the rows of yolk gran- ules. ‘These streaks, however, do not go far into the yolk layer. In both the types a and c the structure of the cytasters is very similar, so much so that it is hardly neces- sary to describe them separately. The only difference between them lies in that in the type a the rays are much stronger than those of the type c. Some of the cytasters are drawn with a higher power. Fig. ga and gb are the same ones shown jn Fig. 6; the former is the upper one and the lat- ter the lower one. Fig. gc is from a section of another fragment in which only three large 1 Centrosome in Enucleated Egg-Fragments. 297 cytasters have arisen. At the center of the cytasters there is always a little accumulation of centroplasm; in Fig. gc this has enlarged a good deal. The ray system around the centroplasm is exactly the same as that of the normal asters. Many rays extend quite far into the yolk region. A few stronger rays go through the centroplasm and reach the center, while most of them start from the periphery of the centroplasm. No marked archiplasmic dif- ferentiation can be detected, but there is a little difference in the nature of rays be- tween the part of rays running through the yolk layer and that within the central portion of the cytaster free from yolk gran- ules. This relation is clearly seen in Fig. 6. At the center of the cytaster are always found a few dark granules of various sizes. Insmall cytas- ters, where there is very little centro- plasm, the granules are crowded together (Figs. ga and gb), ; ga and b, Cytasters from “Es section shown in Fig. 6, more while ia ease Hae Sane highly enlarged. 9c, cytaster with enlarged centrosome and several troplasm is enlarged the granules are found farther apart from one another. In no case I was able to find at the center either a single granule or the granule at the division stage. In the section represented in Fig. 7 only one large cytaster occupies the central part of the fragment and the centroplasm Fic. V (X 2284). centrioles in it, showing intermediate stage between the cytasters Qa and 5, and the one shown in Fig. 7. is of enormous size with many dark granules. The rays are comparatively short and not very straight. They intercross one another, giving a felt-like layer. —_ ts id oS 298 Naohidé Yatsu. Fic. VI (X 2182). 10, First maturation figure (CaClz entire egg), showing abnormal multiplication of centrioles i enlarged centrosome. JZ, central aster of first maturation figure (CaCl. entire egg) with centroso Js . . . . of enormous size and many centrioles in it. Centrosome in Enucleated Egg-Fragments. 299 d. Nature of the Central Granule and Development of the Cytaster. It 1s highly i important to determine whether the dark granules found in the cytaster of the enucleated fragments now finder consideration are real centrioles or not. “The nelineet com- parison of these cytasters with the asters of the normal egg is, I think, not just for the reason that, even if the normal aster did appear in the enucleated fragment, it would have been acted at the same time by the salt solution. The more reasonable way, it seems to me, would be to compare our cytasters with the asters found in the entire CaCl, eggs. Fig. 10 shows a portion from a section of an entire ege shaken fifteen times, treated for an hour with the CaCl, solution and fixed after ten minutes’ sojourn in sea-water.' In the section one observes a large centrosome at either end of the first maturation-spindle; single at the right, and double at the left—the latter undoubtedly a division-product of one original centrosome, as shown by the course of the spindle fibers. Besides these, four asters are found in the vicinity. Noteworthy is the abnormal growth of the centrosome and extraordinarily rapid multiplication of the centrioles in the centrosomes. (<7. Wilson, ’o1, Figs. 24, 25 and 34.) Another section shown in Fig. 11 illustrates these points very clearly. This is from an egg shaken fifteen times, treated with the CaCl, solution for an hour and killed after five minutes’ sojourn in sea-water. Only the central aster of the first matura- tion mitosis is pictured here. (cf Moreans09, Pl: ro; Fig. 67-) The centroplasm of colossal size is surrounded by irreular rays, and in it one sees a number of dark stained oranules. That these granules are really the centrioles can clearly be demonstrated in Morgan’s figures (PI. 10, Figs. 68, 70 and 60B), each granule having acquired a new ray system about it. From the above two examples it will be seen that CaCl, has the power to call forth two independent phenomena at the same 1This lot of eggs was originally intended for the study of cytasters in enucleated fragments obtained by shaking, but quite a number of eggs escaped from being broken. It is noteworthy that shaking has no effect at all on the mitotic figure, nor are the cytasters produced by it. We may, therefore, safely look upon all the changes that have taken place in the section about to be described, as due to the action of the CaClp solution. Although it may be claimed that these changes are caused by the combined action of shaking and CaClo, yet I think the comparison does not lose its validity for our present purpose, since the operation of cutting might give the egg as strong a shock as shaking does. 300 Naohidé Yatsu. time, one an acceleration of the division rate of the centriole and the other an enlargement of the centrosome. One who 1s familiar with the literature soon finds that number, shape and size are not the criteria of the centriole. Meves describes a group of cen- trioles in the first maturation mitosis of the oligopyrenous sperma- tozoon of Paludina (’03, Pl. 3, Figs. 70 and 78). Heidenhain’s microcenters in the leucocytes are another example, although these may be brought about by some pathological conditions, as suggested by Boveri (OI, pp. 21 and 22). The spongy or pluricorpuscular centrosomes have been observed by Wilson in the MgCl, egg of Toxopneustes (’o1, Figs. 70 and 82). ‘These three examples will sufhce for the present to show that the cen- trioles may vary in number. Generally speaking, the centriole has constant size to a particular kind of cells of an animal, yet considerable periodical fluctuation in size was noticed by the writer in the egg of Cerebratulus. ‘The centriole is as a rule spherical, yet quite often we meet rod-shaped ones among the “normal” eggs. In abnormally treated eggs the size and shape of the centriole are exceedingly variable as is shown in the MgCl, eggs (Toxopneustes) and CaCl, eggs (Cerebratulus). (Fig. 10.) From this it will be seen that Vejdov sky and Mrazek’s conclusion that “die Centriolen in allen Fallen dieselbe Grosse und Beschaf- fenheit zeigen”’ seems untenable especially in abnormal cases. Now let us consider how the cytasters which have appeared in enucleated fragments differ from the normal aster. The size and number of the dark granules are not constant in the cytaster, while in the normal aster the centrioles are almost of the same size and never exceed two in number. In some cytasters an enormous accumulation of the centroplasm takes place, while in the normal case the growth of the centrosome is limited. As — we have already seen, all these abnormalities of the cytasters — occur in the whole egg treated with the CaCl, solution. We are, — therefore, led to ‘ne cepncltision that the dark granules at the center of the cytaster of the enucleated fragments are not mere ~ metaplasmic g granules, but real centrioles. i. My experiments and sections of the enucleated fragments are — not numerous enough to ascertain the development of the cytaster. I may, however, be able to construct the history out of the materia | at hand without great error. According to the distribution of the — centers in enucleated fragments we shall get two different types | —- Centrosome in Enucleated Egg-Fragments. 301 of appearance of the cytasters; in one case the cytasters will dev elop throughout the fragment as in Fig. 6, while in the other one, two or three cytasters will appear near the center. Suppose in the latter case these cytasters grow to a considerable size, accom- panied by the multiplication of the centrioles, then we shall have a condition somewhat similar to that shown in Fig. 7 (in this case it should be mentioned only one cytaster has been formed). Meanwhile the rays degenerate, leaving radiating line of yolk granules behind. ‘The granules are pushed out as the centrosome grows. In case two or three cytasters appear they finally fuse together, giving rise to a huge central space. ‘The yolk granules found as islands in the clear area may be the remnant of the interastral spaces. In fact, in some cases the yolk island is con- nected by a narrow bridge with the peripheral yolk layer. ‘Then most of the centrioles move out toward the periphery of the clear space. [Each centriole acquires a daughter ray system around it. The condition shown in Fig. 8 is thus reached. (Morgan, ’gg, Pl. 10, Figs. 67, 68, 70 and 60B.) B. Enucleated Fragments Treated with the CaCl, Solution, T ime-Sterilization. To test whether the formation of the cytasters in the enucleated fragments be due to the action of the CaCl, solution or to the heat-sterilized sea-water 1. made a few experiments using sea- water kept for two days, the precautions and subsequent treat- ment being the same as Experiment A. The cytasters appeared exactly in the same way as the experi- ments in which the heat-sterilized sea-water was used. [wo fragments are shown in Fig. 12a and 12b. One piece was cut into sections. ‘The cytological characters of the cytasters were similar to those of Fig. 8. From this experiment it will be seen that the formation of the cytaster is entirely due to the action of CaCl,. C. Enucleated Fragments Treated with the MgCl, Solution, Heat-Stertlization. An enucleated fragment was put in the MgCl, solution used by Morgan (3.5 per cent of MgCl, in the sterilized sea-water). When I examined this fragment after half an hour a clear area 302 Naohidé Yatsu. had been developed to a fairly large size. (Fig. 13a.) The nucleated half of the same egg was stained and the mitotic figure was found intact. (Fig. 13b.) The enucleated fragment was cut into sections. The general character of the cytoplasm is very similar to that of Fig. 7, while the large aster (Fig. 14a) at the center shows totally different features from any other that 12a 126 134 (Fic. VII (X330)- 12a and b, Enucleated fragments containing cytasters produced by CaCl solution, time-sterilization. I3a, enucleated fragment with cytaster produced by MgCl solution, heat-sterilization. 3b, nucleated fragment from the same egg as 73a, stained with acetocarmine. come under my examination. ‘The center (centrosome?) is elon- gated; no central granules could be made out. Around this center long strong rays radiate. ‘They look very brittle, judging from the fact that some rays show jagged broken edges. For comparison I reproduce a central maturation aster (Fig. 14b) — from an entire egg kept in 3.5 per cent solution of MgCl, in sea- Centrosome in Enucleated Egg-Fragments. 303 water for twenty-two minutes after the first maturation mitosis reached the metaphase. The centriole is here obscured by the strong rays. Striking is the similarity between Figs. 16 and as (qe organ, (99, Pl ro. Figs. 54c, 55, 57-MeCL, and Fig. 63-NaCl.) The strong-rayed asters seem to be due to the peculiar action of MgCl, on the egg of Cerebratulus. Further experiments are necessary to find out whether the centrioles are produced by MgC... I4a Fic. VIII (X 903). I4a, Cytaster from section of the fragment shown in Fig. 13a. 14b, central aster of the first maturation figure, modified by MgCl» sclution. LV. EXPERIMENTS ON THE EGG BEFORE THE- DISSOLUTION OF THE GERMINAL VESICLE. Gick SOLUTION, HEAT-STER- ILIZATION. To determine whether the cytasters are produced by the CaCl, solution before the dissolution of the germinal vesicle the follow- ing experiments were carried out: In 1903 ten enucleated fragments were cut from the eggs just released into water and were treated with the CaCl, solution. No cytasters appeared in any of these fragments. In 1904 three parallel experiments were made on the egg fragments taken from one individual. I. Five enucleated fragments were cut from eggs immediately after they were released into the sterilized sea-water. ‘These fragments were kept in the water and then transferred into the CaCl, solution. After an hour’s sojourn in this solution they were put back to the sterilized water, which was changed once. No asters appeared. 304 Naohidé Yatsu. II. Five enucleated fragments were cut from the egg imme- diately after they were released and kept for an hour in the ster- ilized water; first, in order to give the enucleated fragments more time to ripen, so to speak, and second, for the sake of uni- formity with the following experiment. Then the fragments were transferred into the CaCl, solution. After an hour they were put back into the sterilized sea-water. In none of the jragments did cytasters appear. III. (Control) Eggs were kept for an hour in the sterilized sea-water. Meanwhile the germinal vesicle faded and the first maturation mitosis reached the metaphase. Five enucleated fragments were cut and treated in the same way as Experiment A,a. In all the fragments cytasters were found. The above three experiments! show clearly that the cytasters do not appear in the enucleated fragments from the egg immedi- ately after release. Ve) REVIEW OF- CECE RARURIE. An experimental study of the cytasters was for the first time made by Morgan. In 1893 he saw refractile drops in the egg of Arbacia treated with the sea-water to which a little NaCl (2 percent) had been added. Inthe winter of 1894-95 he extended his experiments on the egg of Sphzrechinus to see if the refrac- tile drops, which he later found to be the cytasters, cause the division of cytoplasm. Although his expectation failed, yet, from the studies along this line, he reached important results which may be summarized as follows: 1The following objection might be raised. The enucleated fragments for Experiments I and II were smaller than those for Experiment ITI, and cytasters might not have been able to develop for this reason. In fact, however, the former were only a little smaller than the latter; 7. e., about the size of a fragment represented in Fig. 12a. The minimal size of the cytoplasm which can produce cytasters is, I think, by far smaller than any piece I used for the above experiments. In this connection I might cite a case in which an egg was cut into three pieces, one nucleated and the other two enucleated. In one of the enucleated fragments the aster was found. Another point: the eggs for experiment I and II were cut, as I stated expressly, immediately after release. Sections of the normal eggs clearly show that a few asters do rarely appear after from twelve to fifteen minutes’ stay in sea-water, in spite of the fact that the nuclear membrane remains apparently intact. The asters thus developed prior to the dissolution of the germinal vesicle, lie usually on or close to the nucleus and very rarely far away from it. Centrosome in Enucleated Egg-Fragments. 05 1. Asters and pseudasters (7. ¢., asters without the centriole) are produced de novo in the cytoplasm by the action of some salts. The presence or absence of the centriole in the aster is not due to the action of fixing fluid (’00, p. 522). The centriole in the cytaster is sometimes single, sometimes a group of granules (’96, P- 343): 2. [he first step to the cytaster formation is a local accumula- tion of hyaloplasm, rays are formed in it, and then the centrioles develop at the center (’99, pp. 477 and 513). 3. he cytasters become more distinct when the nuclear mem- brane fades, while they become less marked when the nuclei come into the resting stage ('99, pp. 468, 469 and 517). 4. In the unfertilized egg cytasters develop more slowly and are less distinct than those in the fertilized egg (96, p. 344; 99) P- 473): 5. No cytasters appear before the dissolution of the germinal vesicle in the egg of Sphzrechinus (96, pp. 348 and 349), and in Sipunculus (99, p. 502). In his paper on the nature of the centrosome Boveri (01) touches on the question of the cytaster in several places, although he has no observations of his own. He distinguishes two kinds of cytasters: one assumed to be descendants of the ovocenter, and the other artificial asters (p. 169). Central bodies may be present in the latter, yet their identification as centrioles is doubt- ful, unless their division is actually observed. In other words, he does not accept the formation de novo of the centrosome and seems to incline to the conclusion that every centrosome in the egg is a division product of the ovocenter.’ In eggs of Toxopneustes treated with solutions of MgCl, Wilson (or) confirmed the formation de novo of the a nrtiole from the study of sections as well as living eggs. Moreover, he, for the first time, proved the above fact experimentally in enucleated fragments. His results are as follows: Tt is perhaps worth pointing out that R. Hertwig (’02) misquotes Morgan’s experiment, stating that he obtained cytasters in enucleated egg-fragments (p. 19). This is an error, the experiment having been done for the first time by Wilson (’or). Fischer and Ostwald (’o5) cite Morgan’s merogony experiment as giving the same result (p. 253), but this is also erroneous and in the papers they referred to no experiment giving this result is described. “After the appearance of Wilson’s paper (’01), Boveri accepted the formation de novo of the centro- some (’02, p. 40). 306 Naobhidé Yatsu. I. Asters having the power of division arise in the cytoplasm independent of the nucleus. ‘These asters first appear simul- taneously im situ scattered through the cytoplasm and, though plainly visible in the living eggs, show no evidence of genetic connection with one snathier. “At a later period, however, they multiply by division synchronously with the division of the nuclear asters. 2. At first vague clear spots appear in the cytoplasm, which gradually become surrounded by radiating lines of granules and finally assume the form of asters. In sections central granules appear in the accumulations of hyaloplasm and afterward rays are formed about them. 3. In the cytasters there is a central granule which 1s a true centriole formed de novo in cy toplasm. The.central bodies dale as in the ordinary asters and thus give rise to the centers of the daughter aster. Sometimes two centrioles are found in a een trosome (p. 561). 4. In enucleated fragments obtained by shaking the unfer- tilized eggs and treated with MgCl, the typical cytasters often containing the centrioles are found. Moreover, these asters may multiply by division (p. 581). WassiliefF (02) made interesting experiments on the egg of Strongylocentrotus lividus. ‘The centrale: he claims, is formed by the interaction of the nuclear fluid and cytoplasm. “Der Kern sondert in das Protoplasma eine gewisse Substanz ab, welche zur Bildung eines Centrums in Protoplasma Veranlas- sung giebt und um dieses letzteres herum lagert sich die pro- toplasmatische Strahlung ab” (p. 769). I perfectly agree with him, so far as this conclusion is Sana, though his evidence was not strong enough to establish it. WMorome he fails to consider what seems to be of prime nee he insists that the nuclear fluid flows out as the egg nucleus fades. If so, why should the cytasters in some cases appear, while the egg nucleus is intact? He states that the cytasters must have originally been connected with the nuclear aster. To bear out this view he gives a case in which a cytaster is connected with the nucleus. The connection seems to me to be merely secondary one. He raises objections to Wilson’s results on the formation de novo of the cen- trioles in the enucleated fragments obtained by shaking on the eround that if the eggs are so violently shaken that they break Centrosome in Enucleated Egg-Fragments. 307 up into fragments, the membrane of the egg-nucleus may be torn and consequently the cytasters are formed by the intermingling of the nuclear fluid and cytoplasm. Meves (02, a, 6, and ’03) thinks that the cytasters may arise in the following way: Numerous centrioles may be handed down to the egg from the last division of the multiplication period somewhat as in the formation of the oligopyrenous spermatozoon in Paludina. He, therefore, holds the view that cytasters may be derived from preexisting centrioles which have acquired a new ray system around them by the action of salt solution (’02, a, 155). He criticises Wilson’s experiment on enucleated frag- ments on a ground slightly different from Wassilieff’s objection, assuming that, even if there be no preéxisting centrioles in the cytoplasm, the egg center may by shaking be thrown off in the enucleated fragments (’02, a, p. Ta 5Ne It was in order to test the Ler two possibilities that Professor Wilson suggested to me two years ago to repeat his experiment on enucleated fragments by cutting unfertilized eggs in two singly and to treat the enucleated piece with some sale solution. This I tried both in the egg of Cerebratulus lacteus and of Echina- rachnius parma’ in the summer of that year. In the meantime appeared Petrunkevitsch’s paper on artificial parthenogenesis (’04). He took up the same form as that studied by Wassilieff, Strongylocentrotus lividus, in which I fully realize how difficult the fixation of the centriole is. Surprisingly enough Petrunkevitsch was led to the conclusion that 1m the egg of 5 sea-urchin there 1s no centriole at all (p. 32).2, His whole argu- ment, therefore, applies to the centrosome not to the centriole. He denies the formation de novo of the centrosome and tries to rescue Boveri’s idea of continuity of the centrosome. He came to this conclusion from the study of sections of the eggs, the “stages” of which were selected arbitrarily. Despite “this he insists that his view regarding the origin of cytasters is thus con- In the egg of this echinoid I used the following solution: 11.8 per cent of MgCle (=*,° m. MeClo), I part: sea-water, I part. In two cases out of eighteen Pattee the cytasters were formed. Total preparations of these two pieces showed that they had no nucleus in them. In passing I should state that the following solution is the best to induce parthenogenesis in the egg of Echinarachnius: 18.6 per cent of KC] (=2° m. KCl), 15 parts; sea-water, 85 parts. Noteworthy is the fact that, judging from his figures, he actually saw the centrioles, but mistook them for reduced centrosomes (not in Boveri's sense), e. g., Pl. 2, Fig. 24. 308 Naohidé Yatsu. firmed by the fact “in glanzender Weise” (p. 45). Besides repeating Wilson’s experiment on enucleated fragments obtained by shaking he also made cutting experiments. In both cases he very seldom saw cytasters; none of them had centers and they faded earlier than the true asters in the control eggs (p. 36). Centrosomes were never observed in the enucleated fragments. His general conclusion 1s, therefore, exactly in agreement with the position taken by Boveri in his Zellstudien IV, 7. ¢., there are two kinds of cytasters, one containing the centrosome, the other devoid of a centrosome. ‘The centrosome of the former are sup- posed to arise solely as division products of pre€xisting ones, and only the latter can be produced de novo in the cytoplasm. Wilson (’04) in his rejoinder to Petrunkevitsch’s paper points out that the evidence given in that paper does not sustain this conclusion and that the negative result is insufficient to disprove the formation de novo of the centrosome. ‘The results brought forward in the present paper fully sustain this position. VI. CONCLUSIONS. My experiments consist in cutting singly unfertilized eggs by horizontal section at two different ‘periods and in treating the enucleated fragments thus obtained with a solution of CaCl,. By these experiments I[ think I have established the facts, (a) that at the period of the metaphase of the first maturation mitosis cytasters can arise at any point of the egg,’ but (b) that prior to the fading of the germinal vesicle cytasters never arise. (cf. foot- note on p. 304.) In all the cytasters developed in enucleated fragments there is a central group of dark staining bodies, which I do not hesitate to identify as multiplied centrioles. It is to be regretted that I did not find in any enucleated fragment either a single centriole or one in division. This, however, does not invalidate the general conclusion for the reason that centers of exactly the same nature as those in enucleated fragments are found in the nuclear division figure in whole CaCl, egg. ‘My experiments show that cytasters appear in the vegetative half. It was, however, impossible to test experimentally whether the cytasters develop in the vicinity of the first maturation mitotic figure. Nevertheless it will not be unreasonable to infer that they may so arise there from the fact that the whole CaCls egg has many cytasters near the animal pole as well. Centrosome in Enucleated Egg-Fragments. 309 My cutting experiments were performed at two periods, one immediately after release and the other an hour and a half later. From these we can by no means determine exactly when the cytoplasm acquires the power of producing the centrioles and ray system. What brings about the change in the characters of the cytoplasm during this interval? In all probability the inter- mingling of the nuclear fluid and cytoplasm during the time of fading of the germinal vesicle gives to the cytoplasm the aster producing power. A striking difference between the matured and immature cytoplasms has been described by many observers. Delage emphasizes the fact that during this period, when the cytoplasmic maturation takes place, the eggs become fecundable both in Strongylocentrotus (’99) and in Asterias (’o1.) Wilson verifies this phenomenon in the eggs of Cerebratulus (03, p. 417). Spermatozoa can enter immature eggs freely, but they remain undeveloped. (O.and R. Hertwig,’ 89, p- 199; Wilson, 96, p- 149.) In immature cytoplasm not only is the development of the sperm nucleus and ray system inhibited, but also the centrioles do not arise in eggs, even if they are treated by salt solutions. Morgan (’99) noticed that cytasters did not dev elop either in the ege of Sphzrechinus or of Sipunculus before maturation begins, and I was told by Professor Wilson that he observed the same fact in the MgCl, egg of Toxopneustes. Leaving open for the present the question how the nuclear fluid acts upon the cytoplasm we can at least say that the matured cytoplasm is ready to produce or, in other words, has the power to form centrioles as well as rays as a result of certain stimuli, this being in our case a CaCl, solution. As to the origin of the centrioles in the cytasters there are two possibilities besides the one just mentioned. First, as suggested by Meves the centrioles might multiply in the cytoplasm during the growth period of the egg and become the centers of the cytas- ters under the action of a salt solution. Such an assumption is not in contradiction with what has just been said, that asters do not develop in unmatured cytoplasm even when the spermatozoon brings a centriole into the egg, since the centrioles might be present, but incapable of producing asters until the germinal vesicle fades. There is another possibility similar to the aboy e, namely, that centrioles may be present as such in the nucleus and, at the dissolution of the germinal vesicle, escape into the cytoplasm 310 Naohidé Yatsu. where they acquire rays and thus give rise to the cytasters. Apart from the fact that neither of these assumptions is supported by any direct observations they contradict the definition of the cen- trosome as given by Boveri (’o1, pp. 132, 162, etc.) that the organ is single (or double by anticipation). A multiplication of cen- trigles. capable of producing centrosomes 1s nowhere known to take place unless it be in abnormal or degenerating cell, such as the giant cells or the oligopyrenous spermatozoa. It may be said that centrosomes (centrioles) arise by the enlargement of ultra- microscopical granules or plastids that coexist with the visible astral centriole. ‘This is quite possible, but if visible centrioles may thus arise in addition to the visible ones already existing and independently of them centrosome formation de novo 1n the ordinary SENSE of this ex pression is demonstrated none the less. “The results of my cutting experiments, therefore, I believe, lead us to the unavoidable conclusion that the centrioles are formed de novo, as Wilson maintained. In conclusion one word about the nature of the sperm centriole. One might be readily led to infer from what I have said that the sperm centriole in the normally fertilized egg may arise in the same manner as those found in the cytasters. In Cerebratulus, at least, this is not the case, for I have been able to show that the centriole, as such, is actually brought into the egg in the middle piece of the spermatozoon. Detailed evidence in support of this statement will be published hereafter. VII. SUMMARY. 1. When subjected to the action of a solution of CaCl, enu- cleated fragments of unfertilized egg of Cerebratulus lacteus, obtained by cutting the eggs singly at the metaphase of the first maturation mitosis, develop true asters containing central bodies. The corresponding nucleated fragments show the typical matura- tion spindle. 2. Cytasters do not, however, appear in enucleated fragments from unfertilized eggs before the fading of the germinal vesicle. 3. The central bodies of the cytasters developed in enucleated fragments are centrioles identical in structure with those in the nuclear asters of whole eggs similarly treated. . Centrosome in Enucleated Egg-Fragments. 311 4. Centrioles, therefore, can be produced de novo in the matured cytoplasm (7. e., after the dissolution of the germinal vesicle). Zodlogical Laboratory, Columbia University. January 23, 1905. VIII. LITERATURE. Bovert, Tu., ’01.—Ueber die Natur der Centrosomen: Zellenstudien, Heft 4. *o2.—Das Problem der Befruchtung. DeLaGE, YVES, ’99.—Etude sur la mérogonie: Arch. Zool. exp. (Ser. 3), 7. ’o1.—Etudes expérimentales sur la maturation cytoplasmique et sur la parthénogenese artificielle chez les échinodermes: Arch. Zool. exp. (Ser. 3), 9. FiscuHer, M. H., anp OstwaLp, W., ’05.—Zur physikalisch-chemischen Theorie der Befruchtung: Arch. f. d. ges. Physiologie. 106. Hertwic, O. anv R., ’87.—Ueber den Befruchtungs- und Teilungsvorgang des tierischen Eies unter dem Einfluss ausserer Agentien: Jen. Zeit. 20. Hertwic, R., ’02.—Die Protozoen und die Zelltheorie: Arch. Protistenkunde 1. Meves, F., ’02.—Ueber die Frage ob, die Centrosomen Boveri’s als allgemeine und dauernde Zellorgane aufzufassen sind: (a) Verhandl. der anat Gesell. in Halle. b) Mitteilungen Aerzt. Verein Schlesw.-Holst. Jg. 10, Nr. 6. *03.— Ueber oligopyrene und apyrene Spermien und uber ihre Entstehung, nach Beobachtungen an Paludina und Pygaera: Arch. mikr. Anat. 41. Morean, T. H., ’96a.—On the production of artificial archoplasmic centers: Rept. of the Am. Morph. Soc., Science N.S. 3, No. 54. *96b.—The production of artificial astrospheres: Arch. Entwm. 3. *98.—The effect of salt-solutions on the unfertilized eggs of Arbacia: Science N.S. 7, No. 164. *99.—The action of salt-solutions on the unfertilized and fertilized eggs of Arbacia and of other animals: Arch. Entwm. 8. *oo.—Further studies on the action of salt-solutions and other agents on the eggs of Arbacia: Arch. Entwm. Io. PETRUNKEVITSCH, A., ’04.—Kiinstliche Parthenogenese: Zool. ‘Jahrb. Supple- ment 7. Veypovsky, F. anp MrAzex, A., ’03.—Umbildung des Cytoplasma wahrend der Befruchtung und Zellteilung, nach den Untersuchungen am Rhynchelmis-Eie: Arch. mikr. Anat. 62. gue Naohidé Yatsu. WassILiEFF, A., °02.—Ueber kiinstliche Parthenogenesis des Seeigel-Eies: Biol. Centbl. 22 No. 24. . Witson, E. B., ’96.—The Cell: 1st Ed. New York. ’o1.—A cytological study of artificial parthenogenesis in sea-urchin eggs. Experimental Studies in Cytology I: Arch. Entwm. 12. *o3.—Experiments on cleavage and localization in the nemertine egg: Arch. Entwm. 16. *o4.—Cytasters and centrosomes in artificial parthenogenesis: Zool. Anz. 28. Yatsu. N., ’04.—Aster formation in enucleated egg-fragments of Cerebratulus: Science N. S. 20, No. 521. REGENERATION IN POLYCHGRUS CAUDATUS BY N. M:. SHEVENS AND’ A. M. BORING. PART I. OBSERVATIONS ON LIVING MATERIAL. N. M. STEVENS. WITH 21 Ficures. While enjoying the hospitality of the Hopkins Seaside Labora- tory at Pacific Grove, Cal., the past summer, I made a few experi- ments to test the powers a: regeneration of the red accelous flat- worm, Polychcerus caudatus, which abounds there in shallow tide-pools on the underside of stones and shells and on Ulva. The object of the experiments was a comparison of the regenera- tion of this form which has no definitely differentiated organs— eyes, central nervous system, pharynx, etc.—with the more highly organized fresh-water ‘Planarians, as well as with the ean’: of Schultz (’02) and Child (’04) on Leptoplana and other marine forms which show very incomplete anterior regeneration. Method. In most of the experiments, the worms were cut into three nearly equal parts as in Fig. A, a—b, c—d. These parts will be spoken of as head-pieces, middle-pieces and tail-pieces. The material was kept in covered glass dishes, somewhat shaded, and the sea-water was changed morning and evening. Regeneration in general was much slower than in fresh-water Planarians. The Animals are very sluggish normally, and the pieces moved but little even when disturbed by changing the water, the head-pieces, however, being much more active than the middle-pieces and tail-pieces. ane tail-pieces continued to deposit eggs for several days as freely as did the entire worms, and the eggs developed normally. N. M. Stevens and A. M. Boring. Cc D b iy a ‘ o a es SS) 3 M Se a) -b O -d es =e G : ae R c SSO ae : . (ee = Be aS, d A.—Whole worm showing planes of section. B—D.—Head-pieces after 2 weeks’ regeneration. E.—Head-piece after 4 weeks’ regeneration. F.—Middle-piece after 2 weeks, showing ventral union of anterior edges (e—f), V of new tissue (g—e—h), and posterior regeneration. G.—Middle-piece after 2 weeks, showing anterior regeneration where the edges have not united as in F. H.—Middle-piece after 19 days, showing heteromorphic tail. L.—Middle-piece after 4 weeks, showing more advanced anterior regeneration of the type shown in F. M.—Middle-piece after 4 weeks, showing anterior regeneration of the type shown in G. N.—Posterior regeneration of middle-pieces, showing super- numerary appendages. O—P.—Regeneration in tail-pieces, 2 weeks. R—S.—Lateral regeneration, 4 weeks. T—X.—Young worms, still in jelly, with appendages just developing. Regeneration 111 Polycherus Caudatus. B27 Head-pieces. These pieces very soon began to produce new tissue at the cut surface as in other Planarians. Among the 40-50 pieces in a series, at the end of two weeks, the stages cor posterior regeneration shown in Figs. B, C and D were found with all imreeme dite stages. A rounded mass of new tissue of considerable size forms posterior to the cut surface, a—b, before the characteristic notch and appendage appear. Continued regeneration adds to the length of the new part while the old part decreases in width and the whole piece gradually assumes the typical form. ‘The notch, at first broad and shallow, becomes deeper and narrower, and the appendage longer. The new part assumes the characteristic pigmentation of the adult tail-region, and a digestive region forms anterior to the line of section, a—b. Regeneration of these pieces was not followed longer than four weeks, when most of the pieces had assumed the form shown in Fig. E, where, if one compares with Figs. A and B, morphallaxis is very apparent. Middle-pteces. In these pieces posterior regeneration proceeded somewhat differently. New tissue appeared along the whole of the cut surface, but was so distributed as to fon a median notch from a very early stage. One or more appendages appeared earlier than in the regeneration of head-pieces. Figs. F and G show the usual amount of posterior regeneration after two weeks, and Figs. H, L, M and N after four weeks. In all of these pieces the notch is still much broader and more widely open than in the typical form A. The multiple appendages shown in the figures were at first thought to be a peculiarity connected with regeneration; but examination of many normal worms showed Thee though one appendage is the typical structure, still all the variations observed in regeneration are to be found in normal adult worms. These variations are, however, far more frequent in regeneration, and more frequent in middle-pieces than in head-pieces, where, as a rule, only one appendage develops. ‘These observations sug- gested a comparison with the formation of the tail-region in the embryo. Figs. T and X show two young worms, ten to twelve days after the eggs were laid, and still in the jelly which enveloped the eggs. The appendage has appeared but not the characteristic 338 N. M. Stevens and A. M. Boring. notch. Posterior regeneration in head-pieces (Figs. B and C) follows more nearly the embryonic method of tail development than does that of middle-pieces, where regeneration from the beginning seems to be based on the adult form of the tail-region which has been removed. Anterior regeneration varied greatly in different lots of material and in different pieces of the same series. [here are, however, two distinct types. In most cases the cut anterior end, a—b, folded together ventrally and the portions on either side of the median line united as shown in Fig. F, e—j. In the first set of pieces no anterior regeneration ocenried while the material was under observation. In another set, in which all the pieces regen- erated better, a few at the end of two weeks showed a V of new tissue between the united cut edges, Fig. F, g—e—h. At the end of four weeks such pieces had developed as in Fig. L, g—e—A, and later some of them produced typical worms. As the union of the cut edges, as in Fig. F, e—/, appeared to hinder regeneration in many cases, an attempt was made to prevent the union of the edges or to remove the hindrance later on. Pieces were cut as in Fig. A, x—y, or with a sharper angle, but the cut edges still curled under and united as before. Cutting the line of union was equally unsuccessful. “There were a few pieces which contracted at the anterior end without folding under and uniting; these regenerated as shown in Figs. G and M, and in due time produced worms of typical form. Anterior regeneration was, however, in all cases less rapid than posterior. One piece produced a hetero- morphic tail, Fig. H. ‘This individual did not crawl normally, but half crawled, half swam with great difficulty on its back or side. This was the only case of heteromorphosis observed. T ail- pieces. Anterior regeneration of tail-pieces was of the two general types described for middle-pieces and illustrated in Figs. O and P. In general it was less rapid and less complete than in middle- pieces. Lateral Regeneration. A few worms were cut longitudinally in various ways. Regen- eration occurred along the whole cut surface as in other forms, the new material being distributed in proportion to the amount Regeneration in Polycherus Caudatus. 339 temoved. Figs. R and S show the result in two cases of diagonal section, as in Fig. A, o—p, and leas after four weeks. In Fig. R the posterior end on the regenerating side is in approximately the same condition as in cases of entire posterior regeneration. In Fig. S, an abnormal notch and appendage has dev eloped near s, as though the notch and appendage were a necessary accompani- ment of posterior regeneration without regard to the presence of the same structure in ihe old part. ‘This phenomenon also recalls the supplementary heads and tails described by Morgan (’ or) and others, as appearing on long obliquely or longitud: nally cut surfaces. General Discussion. The results of the experiments show that in Polychcerus cauda- tus anterior regeneration at different levels may proceed much as in many fresh-water forms (Figs. G, M and P), or it may be pre- vented or delayed, not by muscular contraction and union of the muscle bands, as described by Schultz (’02), but by a folding under and union of the cut edges. (Fig. F, e—j.) That such union of the cut edges 1 is not an insuperable h ndrance to regenera- tion in this form is proved by such cases as are shown in Figs. E L and O, where regeneration begins with the formation of a Vv of new tissue and ends with the production of a typical head- region. In Polychcerus there is no axial gut (Bardeen, ’o1), nor is there a central nervous system to influence regeneration (Lillie, ’oo; Child, ’o4). ‘The fact that head-pieces, which are more active, regenerate more rapidly than middle-pieces or tail-pieces, might be held to support Child’s theory that “there is a close parallelism between the rapidity, amount and completeness of regeneration and the characteristic activity of the part concerned;”’ but the difference in rate of regeneration and morphallaxis is not propor- tionate to the difference in activity, for head- pieces are easily stimulated into activity by changing the water or jarring the dish, while middle-pieces and tail-pieces hardly move at all during the first two weeks unless violently disturbed. ‘The difference in activity is great, while the difference in rate of regeneration 1s comparatively small. So far as regeneration in Polychcerus has been tested by these experiments, it seems to be largely a question of “organization” 340 N.M. Stevens and A. M. Boring. 3 and “‘totipotence” of material (Morgan, ‘04) modifed in many cases by the folding under and uniting of the anterior cut surfaces. It is the intention of the authors to supplement this work with further experiments during the coming summer. PART Il, HIStPOLOGyY- BY A. M. BORING. Wirth 2 PLates AND I FIGURE IN THE TEXT. After working on the external features of the regeneration of Polychcerus caudatus in California during the past summer, Miss Stevens brought back to Bryn Mawr some preserved material— the whole flatworms and pieces that had regenerated for varying lengths of time. ‘The simplicity of structure and the lack of any great differentiation of tissue, made it a matter of interest to work out the histological side of the regeneration of this form, in order to see whether it differs in any essential points from the method of regeneration in more highly differentiated forms, such as Pea simplicissima, eccabed by Stevens (’o1), and Planaria maculata, worked out by Curtis (’ 02) and Thacher (02). Technique. The material had been fixed in a mixture of corrosive sublimate and acetic acid, the regenerating pieces at the end of one, two, five, seven, ten, fourteen, and twenty- eight days. After being hardened in the alcohols, and embedded in parafhne, the whole worms were sectioned in transverse and sagittal planes, and the regenerated pieces in transverse, sagittal, and frontal planes. The sections were stained in Belaheld’ s hematoxylin, followed by orange G. This combination gives a good differentiation of the various tissues. [he reproductive cells stain purple, the mucus blue, the nuclei of the parenchyma cells brown, the parenchyma itself pale yellow, the muscle cells deeper } alee and the cilia usually form a slightly stained border at the margin of the sections, in parts of which the separate cilia can be distinguished. Regeneration 1n Polycherus Caudatus. 341 Normal Structure. Before describing the process of regeneration, it seems necessary to describe the normal structure, as this differs essentially from that of other Planarians, and has not been described in detail. Fig. K is a sagittal section of a whole worm showing the general outline of the form and the location of the different openings; x 1s the digestive opening, r the female repro- ductive opening, and p the penis. It also shows the position of the cells that secrete the jelly in which the eggs are laid 7. Fig. I 1s a transverse section taken near the anterior end of the worm (Fig. K, a—), showing the testis cells ¢, maturing spermat- ozoa s, mucus m, the nance homie: nuclei 7, and the ciliac. Fig. 2 is a transverse section through the middle-region (Fig. K, c—d), showing in addition, egg cells o, the irregular digestive region d, containing some food ip and the digestive opening x. Fig. 3 1s a transverse section near the posterior end (Fig. K, e—f), showing besides the foregoing features, the female reproductive opening r, and the jellygland 7. In these three sections, certain meaneions between the dorsal and ventral sides can be seen. Most of the mucus lies on the dorsal side. “There are more nuclei on the ventral side than on the dorsal, and there is a marked aggregation of nuclei at the lateral edges of the ventral side. By comparing Fig. 4, a piece of the dorsal margin of a transverse section (similar to Fig. 2), with Fig. 5, a piece of the ventral margin, an additional difference appears, that of ie arrangement of the muscle fibers. On the dorsal side, they are more regularly arranged, forming an outer circular and an inner longi- tudinal layer, while on the ventral side, dpa 342 N. M. Stevens and A. M. Boring. are no distinct layers. ‘The apparent difference in the length of the cilia in these two figures (4 and 5) may be due to their being matted together in ace ‘There is no defnite ectoderm or endoderm. ‘The cells com- posing the mass of the body are the parenchyma cells, irregularly spindle-shaped, with large nuclei. (Fig. 4, 7. .) In many places, the outlines of these cells are so indefinite that it appears as though they merged into one another, forming a syncytium studded here and there with nuclei. Among these parenchy ma cells are mucous cells, which have similar nuclei, but contain masses of a blue-staining secretion. (Fig. 4, m.) On the outer edge, where one would expect to find a definite ectoderm, these parenchyma cells are ciliated (Fig. 4, c), and stain a little more deeply, perhaps due to a cuticular secretion; but in no other way is the outer layer Socsis essere ite ans ee making up the mass of the body. This outer layer is not even arranged regularly, for the nuclei are at varying distances from the poe of Ai cilia, and at irregular distances apart. The cells of the digestive region (Fig. 2, d)—1t is not definite enough to be called a digestive tract—do not differ in any respect from the other parenchyma cells. In places pieces of crustaceans, which have been taken in as food, are found in between the cells near the digestive region (Fig. 2, 7), showing that this cavity is continuous with the spaces between the loose parenchyma cells. At the opening of the digestive region (Fig. 2, x), a few of the cells are sometimes Shad: (Fig. 2, €) tikes ectodermal parenchyma cells. Muscle fibers are scattered throughout the parenchyma, but are accumulated especially among tne ectodermal parenchyma cells (Fig. 4, g), around the female reproductive opening, and in the penis, of which they are the chief constituent. They vary much in size, in fact, so much that in sections stained with iron hzmatoxylin and orange G, some take the black and some the yellow color. The reproductive cells are more distinctly differentiated than the other cells in these- flatworms. ‘They are not grouped into ovaries or testes, but they lie in definite positions among the paren- chyma cells, and are discharged through definite openings, guarded by muscle cells having a sphincter-like arrangement. ‘The testis cells (Fig. 2, 7) extend “along the lateral edge on near the anterior end to the penis which is an external Saireenee organ. (Fig. Regeneration in Polycherus Caudatus. 343 Kp): The ege cells (Fig. 2, 0) lie on each side of the median ventral line, extending from the region of the digestive opening back to the female reproductive pore. Just in front of this pore lie the cells which secrete the jelly in which the eggs are laid. (Fig. 3, 7.) This form has no central nervous system, no eyes or other sense organs, and no excretory system. Regeneration. The regeneration of this form is as simple as its structure. The worms were cut into three pieces as stated in Part I, a head-piece, a middle-piece, and a tail-piece. In the regeneration of Planaria simplicissima and of Planaria maculata, the old ectoderm stretches over the cut surface in a thin layer, but the regenerative process in Polychcerus caudatus is more like the regeneration after natural fission in Planaria maculata, as fecoanedl by Curtis (’02), where the exposed surface simply heals over and embryonic cells migrate to that region and form the new tissue. In Polychcerus, the ‘cells at the cut end secrete a cuticular substance and develop cilia. Sections of most of the pieces fixed two days after being cut, show short cilia at the cut end (Fig. 6, c,) and the cells stain a little more deeply at the base of the cilia. In the five day sections, the cilia have reached their normal length. (Fig. 7, c.) By this time there is also a decided accumulation of nuclei at the regenerating end. Fig. 7 shows this, and a comparison of Fig. 7 with Fig. 6 clearly shows the progress of regeneration. This accumulation is not due to cell division, either in the regenerating end, or the old part. Cell division has been carefully looked for throughout the work, and the one or two cases which might possibly be interpreted as prophases of mitosis lose all significance from their rarity and the entire absence of actual mitoses; neither has any evidence of amitosis been discovered. Many of the nuclei in Fig. 7 have their long axes pointed toward the end, and the cells, as far as their outline can be made out, point in the same direction, indicating a streaming of parenchyma cells toward the regenerating region. In the whole worm, the parenchyma nuclei are accumulated on the ventral side and especially toward the lateral edge. In Fig. 9, a sagittal section some distance lateral to the median line, the accumulation of nuclei on the ventral side is continuous with the 344 N. M. Stevens and A. M. Boring. accumulation at the regenerating end 7, suggesting this accumu- lation as the chief source of the walls? in the new part. Some of the cells come from the dorsal side, but the evidence from the exami- nation of many sections is convincing that the majority come from the ventral side. The muscle cells must develop from the parenchyma cells im situ, as they appear below the ectodermal parenchyma only in pieces which have been regenerating several days. (Fig.8, g.) In Fig. 6, a section of a piece before the accumulation of nuclei had begun, some fibers appear scattered irregularly through the parenchy ma near the end, but these are probably old fibers, as this section shows no definite layer of muscle fibers below the ecto- dermal parenchyma at the regenerating end. The seven-day and ten- day sections show an increase in the length of the new part, but no other new points. In two weeks, most of the new tissue has taken on the loose parenchymatous character of the old part, as shown by the spaces in the tissue and the more scattered position of the nuclei in Fig. 10, only the extreme end of the regenerated tissue still having the nuclei in close proximity and the cells densely packed together. (The dotted lines in Figs. 9-12 show approximately the boundary between old and new parts. In the regeneration of one of the oldest head-pieces, a new digestive opening has formed. (Fig. 11, x.) It is in all respects like the opening in a full sized worm, being situated about halfway between the anterior and posterior ends, and opening directly from the digestive region to the exterior. Middle-pieces of this age have the old digestive opening, but some distance posterior to this, at the base of the new tissue, there is an accumulation of parenchyma and muscle cells, as in Fig. 12, 7, which can be recognized as the anlage of the penis, for the sperm has moved down near to this anlage. Anterior to the penis is a slight indenta- tion r which may grlicaie the anlage of the female genital pore. Sections show that anterior regeneration 1s always slower than posterior; there is less new tissue at the anterior end than at the posterior, and it keeps its compact character and accumulation of nuclei longer than the posterior, as shown by comparing Fig. 7, an anterior end, with Fig. 8, a posterior end of the same age. In some pieces in which the anterior end folded under to form a pocket, as described in Part I, no regeneration can be seen in the Regeneration in Polycherus Caudatus. 345 oldest stages, but in a few of these, the growth of new material between the united edges can be seen in section. (Fig. 13, v.) By studying the whole series of sections, this region can be identi- fied as the place where the cut edges united. ‘The accumula- tion of nuclei shows new tissue to be regenerating on both sides of the line of union. A few cases of lateral regeneration were studied, but the sec- tions showed no divergence from anterior and posterior regenera- tion. The regeneration of Polychcerus caudatus is an excellent example of the remolding of the old tissue in a piece of an organ- ism, into the tissues and form of the whole organism, without the assistance of cell division by mitosis or amitosis. “Uhis is what Morgan calls morphallaxis. Other flatworms in which regenera- tion has been worked out histologically, Planaria simplicissima and Planaria maculata, show a proliferation of new cells at the cut end, as well as the changes of form due to morphallaxis, but in Polychcerus the new part is formed-wholly of cells which migrate from the old part. Regeneration in this form is, therefore, an example of morphallaxis, pure and simple. Bryn Mawr College, Pa. April 19, 1905. LITERATURE. BarDEEN, C. R., ’o1.—On the Physiology of Planaria maculata with Especial Reference to the Phenomena of Regeneration. Am. Journ. of Physiolog CuiLp, C. M., ’o4a.—Studies on Regulation. IV. Some Experimental Modifica- tions of Form Regulation in Leptoplana. The Journal of Exp. Zoology, vol. 1, No. I, 1904. ’o4b.—Studies in Regulation. V. The Relation between the Central Nervous System and Regeneration in Leptoplana: Posterior Regeneration. Jbid., No. 3, 1904. ’o4c.—Studies in Regulation. VI. The Relation between the Central y, vol. v, Igor. Nervous System and Regulation in Leptoplana: Anterior and Lateral Regeneration. Ibid., No. 4, 1904. Curtis, W. C., ’02.—Life History, Normal Fission, and Reproductive Organs of Planaria maculata. Proc. Boston Soc. of Nat. Hist., vol. xxx, No. 7, 1902. 346 N. M. Stevens and A. M. Boring. Litutk, F. R., ’01.—Notes on Regeneration and Regulation in Planarians. Am. Journal of Physiology, vol. vi, 1go1. Morean, T. H., ’o1.—Regeneration. Columbia Univ. Biol. Series, No. 7. New York. Igol. °o4.—An Analysis of the Phenomena of Organic ‘Polarity.’ Science, N.5S., vol. xx, No. 518. Stevens, N. M., ’o1.—Notes on Regeneration Planaria lugubris (simplicissima). Archiv. fur Entw., Bd. xii, H. 3, 1got. Tuacuer, H. F., ’02.—Regeneration of a Pharynx in Planaria maculata. Am. Naturalist, vol. xxxvi, No. 428, 1902. EXPLANATION OF PLATES. Figs. 1, 2, 3, 9, 10, 11, 12, 13 were drawn with Leitz oc. 2, obj. 3, camera lucida. Figs. 4, 5, 6, 7, 8 were drawn with Leitz oc. 2, obj. 1-12, camera lucida. Figs. 1 to 8 are reduced one-half. The following lettering is used in all the figures: c, cilia; c, cilia half developed; d, digestive region; f, food in digestive region; g, muscle fibers; 7, jelly gland; m, mucus; 7, nuclei of parenchyma cells; 0, OVa; p, penis; g, food among parenchyma cells; r, female reproductive opening; s, sperm; t, testis cells; v, hew material; x, opening to digestive region; y, appendage. ; Prate I. Fig. 1. Transverse section of whole worm near anterior end. (Fig. K, a—b.) Fig. 2. Transverse section of whole worm near middle, through the digestive opening. (Fig. K, c—d.) Fig. 3. Transverse section of whole worm toward posterior end, through female genital pore. (Fig. K, e—f.) Fig. 4. Portion of dorsal margin of transverse section. Fig. 5. Portion of ventral margin of transverse section. Fig. 6. Sagittal section of anterior end after 2 days’ regeneration, showing developing cilia. Fig. 7. Sagittal section of anterior end after 5 days’ regeneration, showing accumulation of nuclei. Fig. 8. Sagittal section of posterior end after 5 days’ regeneration, showing appearance of muscle fibers. Pirate I. Fig. 9. Lateral sagittal section of middle-piece after 5 days’ regeneration, showing accumulation of nuclei, 7. Fig. 10. Median sagittal section of middle-piece after 2 weeks’ regeneration, showing tissue with the loose character of the old. (Exceptionally rapid regeneration.) Fig. 11. Sagittal section of head-piece after 4 weeks’ regeneration, showing the new digestive opening. Fig. 12. Sagittal section of middle-piece after 4 weeks’ regeneration, showing anlage of penis and of female genital pore. Fig. 13. Transverse section of middle-piece (4 weeks) with a ‘‘pocket,” showing triangle of new material, v. REGENERATION IN POLYCH@RUS CAUDATUS. N. M. Stevens anp A. M. Borine. PLATE I. cS é Soe WH eS Oe py ole ss ee ote ey A.M. Boring del. Tue Journat or ExPerIMENTAL ZOOLOGY, vol. ii. REGENERATION IN POLYCHGRUS CAUDATUS. N. M. Stevens ano A.M. Borinc. PLATE IL. “4, =. soy A. M. Boring del. Tue Journat or ExperiMeNTAL ZoOLocy, vol. ii. PS REWATION OF THE DEGREE OF INJURY TO THE RATE OF REGENERATION? BY CHARLES ZELENY. I. INTRODUCTION. It is a common belief that an increase in the degree of 1 injury to an animal lowers its vitality and thereby diminishes its capacity for repairing sustained injuries. It is certainly true that if an animal is mutilated to a degree so great that it can barely survive the operation a rapid rate of regeneration of the parts is not to be expected, though there is little direct evidence in favor of this statement. ‘he general view that injury to an increased number of organs implies a decrease in the rate of regeneration of each, however apparent it may seem at first sight, needs further exam- ination. [he data to be given below prove very conclusively that the view 1s an erroneous one, for it is shown that the animal with the greater number of removed parts regenerates each part more rapidly than does the one with the lesser number of removed parts. In the summer of 1902 the author performed some experiments on the fiddler crab, Gelasimus, which showed that when both chela are removed each of the regenerating buds grows more rapidly than does the single one in the cases where only one chela is removed. ‘The rate of moulting of the animals is likewise greater in the individuals of the former group than in those of the latter. The difference was naturally more plainly made out in the female individuals which have chelz of equal size than in the male individuals which have chelz of unequal size. ‘The results are, however, not as conclusive as they might have been, had the number of individuals been greater and had a greater length of time been available for the experiment. 1Contribution from the Zoélogical Laboratory of Indiana University, No. 68. 348 Charles Zeleny. In the winter and spring of 1902-03 with the above results in mind two groups of experiments were undertaken to further test this point. A comparison of the rate of regeneration of the arms in five series of the brittle-star, Ophiogly pha, with one, two, three, four and five removed arms respectively, showed that excepting the case where all five arms are removed and in which the animals were dead or dying before the completion of the experiment, a series with a greater number of removed arms regenerates each arm faster than does a series with a smaller number of removed arms. [hus with an increase in the degree of injury there 1s more than a corresponding increase in the total amount of regeneration In a given time. In the Crustacean, Alpheus, a result similar to that for Gela- simus was found but with the addition of a quantitative deter- mination of the actual rate which was not possible for Gelasimus because of the slow rate of moulting in the latter. The Alpheus data are, however, complicated by the fact that the two chelz are of unequal size and undergo a reversal upon removal of the larger one.t. The number of individuals available for the final comparison was likewise small because a large proportion of the specimens cast their chele accidentally during the course of the experiment.’ It seemed desirable, therefore, to test the results in a more con- clusive way upon a form which does not have the complications found in Alpheus and Gelasimus. The common crayhsh, Cam- barus propinquus, has chele which fulfill the requirements of such aform. ‘They are equal 1 in size and similar in character, are cast off at a definite breaking joint upon injury to their nerves and the animal does not Sr throw off its appendages as a result of the necessary handling incidental to the course of the experi- ment. In one series the ‘right chela alone was removed. In the other series the two chele and the last two pairs of walking legs were removed. The resultant data show very conclusively that in the series with the greater degree of injury each chela regenerates more rapidly than the single removed chela of the series with the 1Przibram, ’o1, Arch. Entw. Mech., xi; Wilson, ’03, Biol. Bull., iv; Brues, ’o4, Biol. Bull., vi; Zeleny, ’05, Journ. Exp. Zodl., 11. ?The description of the preceding experiments is given in Journ. Exp. Zodl., vol. 11, No. 1, Apr., 1905, pp- I-102. Rate of Regeneration. 349 lesser degree of 1 injury. Likewise the members of the series with the greater injury moult more rapidly than those of the series with the lesser 1 injury. 2. METHOD. The specimens used in the experiment were collected in a small brook about a mile and a half from the Indiana University campus at Bloomington. ‘They were all taken from a part of the brook not exceeding two hundred feet in length and it is probable that the general conditions of the environment to which they had been subjected were similar for all up to the time of capture on October II, 1904. About 150 specimens were obtained at this time and Fig. 1. Diagram showing a bottle as arranged for the reception of one of the crayfish used in the experi- ment. See text description on page 350. from this lot 77 of the individuals ranging in thoracic length from ten to twenty millimeters were selected and divided into two groups which were made as nearly as possible equivalent in point of size of individuals. In series A which comprised 36 individuals the right chela was removed at its breaking joint. In series B which comprised 41 individuals the two chele and the last two pairs of walking legs were removed in a similar manner. Except for this difference in the degree of injury the two series were consistently treated alike throughout the whole course of the experiment. The series with the greater injury (Series B) was purposely given the greater number of individuals in anticipation of a greater death rate in this series. Both males and females were included in each series. 350 Charles Zeleny. The crayfish were kept in individual wide-mouthed bottles, which were inclined at a slight angle to the horizontal and were covered with pieces of cheese cloth held in place by rubber bands (see Fig. 1). The crayfish were fed every fifth day on frog meat or beef, the water being changed immediately after the meal. Under these conditions no difficulty was experienced in keeping the animals alive. “The few deaths recorded during the course of the experiment were for the most part due to neglect in changing the water immediately after the meal. Such a suspension of ordinary care is especially liable to be fatal when occurring soon after a moult. The operation on the majority of the crayfish was performed October 12, 1904, and the experiment was closed April 20 after an interval of 181 days. A small minority comprising 16 individuals was operated on two days later and kept until April 22, the interval being likewise 181 days. 3. DATA. The records of the experiment include the sex of the animals, the date of moulting, and the length in millimeters of the thorax and of the chelz after each moult. ‘The size of the regenerating walking legs of each individual in Series B is approximately expressed in my notes in fraction of the legs which are being replaced. ‘The latter data are, however, not given in the tables reproduced in the present paper. In these tables (pp. 352 to 358) the moulting time is given in days after the operation. The thoracic length i is the distance in millimeters between the posterior edge of the thorax and the base of the thoracic spine. ‘The chela length is the greatest length in millimeters of the next to the last segment of the chela, the propodite. The data are given in Tables I, I], 1] and IV. The males and females are separated because the rate of moulting and of regener- ation was found to be different in the two sexes. ‘The individuals in each table are arranged in order of thoracic length after the first moult. In the columns giving the original lengths, 7. ¢., the lengths before the operation, blank spaces indicate that the meas- urements were not taken. In the other columns a blank space indicates that the animal had not moulted when the experiment was closed, 181 days after the operation. In the last column the Rate of Regeneration. 351 number after “died” is the interval in days between the operation and the time of death. A comparison of the rate of moulting in the two series is given in Table V. This table is derived from Tables I to IV and gives the number of male and female individuals which had moulted -in each series 95 days, 130 days, and 181 days after the operation. The first column under each moult gives the number of individuals which have moulted, the second the number which have not moulted, the third the number which have died, and the fourth the per cent of the living which have moulted. The data for the rate of regeneration as derived from Tables I to IV are given in Tables VI and VII. ‘Table VI gives the male individuals of the two series and Table VII the female individuals. In these two tables the individuals are arranged in order of moult- ing, those moulting first being put at the head of the list. The specific amount of regeneration (Sp. Amt.) is the amount per unit of thoracic length at the end of the first moult. ‘The specific rate of regeneration (Sp. Rate) is the amount of regeneration per unit of thoracic length per day. EXPLANATION OF TABLES. Series A = Series with right chela alone removed. Series B = Series with the two chele and the last two pairs of walking legs removed. In Tables I, IT, III and IV the individuals are arranged in order of the thoracic length as determined after the first moult. In Tables VI and VII they are arranged in order of the date of the first moult. In Series B the specific amount of regeneration and the specific rate are the averages of these quantities for the two chele of the individual. 352 Charles Zeleny. TasBLe I.—Series A. Males (181 days ajter operation). | Original. | First Moult. | Second Moult. a oe al “ | at. = 2si|| & = a3 = = Sz = 2s)lec0 Remarks. | “*| & |} 85] A] & | 25/46! 6 | a | 25) 6 » er bere [ae | 71 | 10.9| 4.7 | 6.7|| * | — | — | — |] Died 157. 806 | 86|11.5| 4.6 | 6.7|| 164 | 112] 5.0 | 66 72) Re aerial | Phe — || ? | 14) 5-3 1°73 | 797 | -- | -- |) 57 | 12-21 5-7. | 7-6) a4g |e26el 5:8) | jee Fae | 12.6:| 7.0 | 108 | 13.3 | 5.9 | 7.8 | 173°) 33-5. | 0-Do|eaea wag | .. | co W 58 |ag7163 | 86ll * | 22) 225) eee FOUN PERE 72. | 14:2 | 6.0 | 9.7 |) 143 | 148) 7.60) sore ; QO! SATE Reece tea ete aeaa|| Bog. .| G29)! 2 Hig? 14-0 | 6.6. || 0:04) a en eee 762 | 14.1 | 9.1 || 107 15.2 6.85°| .9-9:|] 170. | 25-1 | °7-5 eo i eee Ce | 105 | 15-5} 6.7 | 8.0] TAL \ PLO OS NN BSF 1 F5-0-7-3)" | 208 7354 1520)|( P10 437 | 10.0 7:0- | 12.47 754 | 15-1 | 12.0 || 106 | 16.1 7-0: +] 14 OW ene — | — | =|) Bredsers 761 | 16.1 | 11.0 |} 116 | 16.6| 7.0 | 11.8 | 52s | 10.2) | 2.5 ] rz | 17-780 | 13.0 768 | 15.7 | 13.6 | 166 | 19.4 | 5-25 | eeaall Rate of Regeneration. 353 TasLe I].—Series A. Females (181 days after operation). Original First Moult. Second Moult. Cat. 5 20 aS 5 So| ecu S 5 Sol eu Remarks. Sele }ec | 6 |gd\sel é | é |28| 3 Foe) 10-3 | 6:4 || 135 | 11.0} 4.6 | 6. Beguleies) 5.3) 140 | 11.3 3:8 | 5.6 Homes | 8:5 || 119 -|.14:0 | 6:0. |+ 9:1 | | ( *3d Moult. 786 | iy) 14.4 = | 8.8) || me 14.8. | 6:2 2 1 eae 785 8.0 || 104 | 14.6] 6.3 | 8.3 || 180 | 14.0] 6.4 | 7.9 ; FASE | As || £OO 15.0 | 6:45), QeX|) WSs 5-4. 0'7.0..| 9.1 784 | 13-9| 9.1 || 163 | 5.2") 6:0; | 19:6 738 | 14.6| 7.6|| 108 | 15.4| 6.9 °| 9.0 760 | 14.6| 9.6|| 165 | 15.7| 6.0 | 9.4 799 | 15.2| 9.8 || 142 | 15.9] 6.8 | 10.4 GOGn PES. | 10.2 || 133 |-16.2 | 6.7 | 10.6 783 15.3 | 10.2 || * | — age ta Died 154. ateletse7)| 10.5 || 153 | 17-0.) 6.2 | 11.1 798 | 16.0 | 10.2 ce on 805 | 16.2; 11.1 || 167 | 17-0| 6.1 | 10.5 Fag eee AZO |e |e: ae FjOu\el7-© | 10.2 || 181 | 18.0 | 7.0. | 10.0 Joy) agee Se en ieee yn re 759 | 16.9 | 11.1 || 144 18.8 | Pout jt? 354 Charles Zeleny. Tasie Il].—Sertes B. Males (181 days after operation). | tal Ong | First Moult. Second Moult. | Third Moult. | eal ; > | = Cat.| & é| 2 las|oall 2 flee!

  • | 12.8] 5: | 5-5 *Lost in moulting. 794 37 -=- | 53 | 50 || 103: | 33-0] 5-6°| 5-7 742 31 | 12.6] 5-5 | 5-4 * SS SS SS | re 802 en — se ee ee ered re 788 | I = ee 48 | 14.2] 5-5 | 5-8 || 128 | 14.7| 64°] 64 747 | 31 6.2 | 6.1 84 | 14.2 | 7.3 | 7.0 * — | — | — || *Died 166. 795 | 35 | 5-6 | 5.4 || 81 | 14.1] 6.0 | 60 || 161 | #* | — | — || *Died 166. | **Lost. 7387 32 5-6 | 5-7 || 143 | 14.0] 68 | 6.9 * = SS SS Dee #0o |)... 37 | 13-9| 6.1 | 5.9 || 114 | 14.8] 7.0 | 6.9 =e 772 | ¥3-1 | 118 | 14.0| 6.1 | 6.3 || 142 | 15.1 | 6.7 | 6.7 * — |) — | — || *Diediaa6- 793 33 14-0) 5-3) 5-4 | 116 | 144/63 6.9 779 | 14-7 | 108 | 15.0} 5-6 | 5.6 || 148 | 16.2| 7.6 | 7.6 769 | .- GEE | EM S58) | ir Coe fae EON PS Ay Pty PC 774 | 15-2 | 142 | 15-9] -- - 168 | 16.6) 5.6 | 6.4 73% | 15-3 | 134 | 16.1] 6.8 6.9 || 174 | 16.8! 7.7 | 738 749 117 | 16.4| 6.1 2 | 150 | 16.9| 7-6 | 7-7 763.| =. 52 | 16.9} 6.7 | 6.9 || 107 | 17.0] 7-4 | 7-4 755 | 16.1 | 121 | 17.3 6.8 | 7.0 || 169 | 17.9| 8.0 | 8.0 757 34 | 17-3| 5-6 | 7.0 || 109 | 18.2/| 8.8 | 8.8 78r | .- | 104 17-5 | 7-0 | §.2 || 129 | 18.7| 8.8 | 8.6 || 180] 19.0} 9.3 | 9.1 756 | 16.6 143 7.3) 7.0 | 6.9 os oe oe ae 771 | 16.4 115 17.8 | 7-2 | 7-0 || 138 | 18.7] 8.0 | 8.2 782 | 18.0 148 | 18.8 | 7.0 7-4 - e 750 | 17-6| 147 | 19.0| 6.8 | 6.3 || 179 | 20.2] 8.9 | 8.7 766 | 19.7 161 19.8) 6.6 6.1 350 Charles Zeleny. TaBLe V.—Rate of Moulting. 95 Days after Operation. First Moult. Second Moult. : : ~_. : Be le = 2) + ome E Bo) = Ez 3 3 a Os = 3 = O53 ° ono S 5 5 S 6 Oo o & 5 ° 3) ov = Ze a P= = Ze a P= penes AvGls.2n, 7 fe) fo) 41.0 fe) 16 I fe) Series Big’ 2). -.- 15 Oo Oo | 100.0 I 13 I GB Senes A-975..2- I 18 O ioe O 19 O fe) Series BQ ..... 12 12 I 52.0 ti 30 2. |g 130 Days ajter Operation. First Moult.** Second Moult. 7 = > = — — % = E Sale) eee E 2) ee 3 = z Os 3 3 = Os 2) ro We o 5S o re) S ko = Ze a P= = Ze a ae penies AG) ne — -— — — I 14 2 |, “Ges Sehes Big ee = — — — 7 6 2) eae SeriesA Q ..... 5 14 o 2023. |) oe 18 fe) 5-3 Series B 9 ff...) 19 6 I 70.0) |e 12 2: 50.0 181 Days ajter Operation. Second Moult. Third Moult. : = = Res= = ; = g Ss) a | Bee 3) 2 ee S| ace | SB). ) 2 SCs ei ee ae ce 2. || gee o 5 oS ° roe) 3 & oS = Ze fa) a Ae A = z a a i ha 1] | | menes AG’ ..... Geen’ 3 | 42.9 14 3 : is < | Senes Big! ...\... KO] 3 2 | 76.9 I 10 4 g.1 penes AO: ..\. - 3 15 I 10s. Hook 17 I 5.6 senes BQ ..... 21 3 2 i A || ae: 17 6 15.0 7 Without having moulted. * Of living individuals. ** A comparison of the males is not valid here because all those of Series B had already moulted 95 days after the operation. TT One individual of Series BQ moulted a third time 128 days after the operation. Rate of Regeneration. 357 TasLeE VI.—Rate of Regeneration (1). Males—First Moult. Series A. | Series B. Cat. : s Ei : | Cat. : iS E 5 a ue <— % na & 2 : ee IN| a = 2 3 Bele = ay ae | : A = & & *746 3 ae a 4: | 1733 I 2 es Foley. (\ 12-2 \., .467 5002 Hi e7gOMess3) (12-2 |) 9.434 | O32 oeeseey | 13-7) -400) |, -O070)|| 28780 1-34) IIo) .418 | 50123 alee | 10.0 |. *.431 FOOOM | erOOln | 42 |). 1375 393 0094 730 | 72 | 14.2) .423 | .0059 || 1758 | 44 | -- ou 806 | 86 | 11.5 | .400 .0047 || 780] 44 | 14.8 473 .O107 804 | 92 | 14.9 AOR |) sOG5Ox|Il Sons) -o44= | 0304. 388 | .0088 Toe etoS, | 15.5 432 JOOAT NE —7000|) 48 || 15 5 432 .009O Faeroe" | 16en |. 472 | <0045 || Bor | 65 | 14-3 .469 | -.0072 FOR NW iO7? |* 15.2 451 10042) ll" 7320), 69° 10-5 385 .0056 744 | 108 | 13.3 | -444 | .oo41 || 765 | 73 | 17-5] -471 | .0065 G2 NeLI2. \| 3727 452 .0040 TAO? nO We E47: 459 .0060 foleito: | 1656 |) 2422 .0036 745) 03) 105.8 402 .0048 Vi7 | ¥37 | 15-9 | -459 | -0034 || 764 | 84 |) 14.7 | .483 | .0057 735° | 137 | 16.0 | .437 | .0032 773 | 95 | 16.9 | .450 | .0047 Be 70Ss|4100" | 19.4 776 | 181+ Divers | Ps < 444 .0049 | Rte im x 435 .0080 + .003 |+ .0003 +.006 |+ .0005 For explanation of Tables VI and VII, see pp. 350, 351. *, 1 No visible regeneration has taken place. ** The stump of the operated leg was diseased for a long time after the operation and the data are therefore not included with the others. + The plus sign after 181 indicates that the animal had not moulted when the experiment was closed. TT The animal moulted a second time before measurements for the first moult were taken. 358 Charles Zeleny. TasBLe VII.—Rate of Regeneration (2). Females—First Moult. Series A. Series B. | 4 5 Cat & E 3 Cat “5 | E & = No = & 4 a. No ae & c. i a ay a A | & n n +786 heel ie Fae +788 ro}. | ie! 785 | 104 | 14.6 432 0042 724: | 7 eases 415 .O154 743 | 106 | 15.0 427 0040 744 | (29 Gtk 395 0136 738 | 108. | 15.4 448 0042 742 | - 33 (12.6 433 .O140 775 | 119 | 14.0 .42 0036 747 | 3% | 13-5 | -450 | .on47 790"| 193° | 16.2 414 0031 769 | 32) see 408 | .0127 7GE | 125" | tre 418 0031 787) | “32a dere, 412 0129 807 | 140 | 11.3 336 0024 793 | -33- | 14/0 |, eae o116 799 | 142 | 15.9 | .427 | .0030 || 757 | 34 | 17-3) = : 759 | 144 | 18.8 378 .0026 795:\. 35 "| 1335 407 o116 75i | 153. We a7e0 365 0024 7944) - 37.) coe 415 OII2 784 | 163 | 15.2 395 0024 $00 | 37. || 12890 432 O1I7 FOO: | 105.1] 5c 382 0023 763 | 52 | 16.9 | .402 0077 8a5-| 167° *|) 170 359 | -0021 Far. | To4.” | ars 5: 5s 770 | 181 | 18.0 389 0021 779 | TOS 4| 51520 373 | . 005 783 154+ 771 | 115 |-17.8 | =399°) e@aRs 798 | 181+ 749 | 117 -| 16.4 375 .0032 753 | 181-+4| 772 | 118 | 14.0 443 .0038 767 | 181+ 755 | 121 | 17-3 | -399 | .0033 731 | 134 | 16.1 | .425 | Joage 774 | 142 | 15.9 pin im 756 | 143 |.17-8 | _.390%) }ege7 750 | 147 19.0 | .345 0023 | 782 | 148 | 18.8 383 0026 766 | 161 _— 19.8 a a 802) 17+ Agee): | PIE = ZOO .0030° || Av. 20] -.< a: .403 | .0083 easesi| -.< .. |+.006 |+.0001 eases’ |) 2. .. |+.004 |+ 0007 T No visible regeneration has taken place. * These cases are not included in the general result because one chela in each is much smaller than the other probably as a result of secondary injury. ** Chele deformed and small. No measurement taken. + (Plus sign), see under Table VI. Rate of Regeneration. 359 4.) RESULTS. Rate of Moulting. A comparison of Series A with Series B shows that the individ- uals of the latter, the ones with the greater degree of injury, moult sooner than those of the former, the ones with the lesser degree of injury. ‘The data for the rate of moulting are collected in SToable V in which 1s given the number of individuals of each series which “ewe ee . W 4 Ric. 2. Fie. 3 Diagrams to illustrate the comparative degree of injury in Series A (Fig. 2) and Series B (Fig. 3). CC! = chele. Wi to W4, W!1 to W'4 = walking legs. Plain lines = uninjured legs. Barred lines = removed legs. had moulted once, twice or three times 95, 130, and 181 days after the operation. ‘The males and females are considered separately in each case because the rate was found to differ in the two. ‘The more rapid rate of moulting of the series with the greater degree of injury is evident throughout. Ninety-five days after the operation only seven out of the seventeen male members of the series with W 1 W 3 360 Charles Zeleny. the lesser injury (Series A), or 41 per cent, had moulted while at the same time all those of Series B had moulted. Likewise 95 days after the operation only one out of the nineteen female mem- bers of Series A or 5.3 per cent had moulted while thirteen out of the 25 living members of Series B, or 52.0 per cent, had done so. At the same time the only individuals that had moulted a second time belonged to the series with the greater injury. Of these one is a male and the other four are females. One hundred and thirty days after the operation Series B shows a similar advantage over Series A. Only five out of the nineteen females, or 26.3 per cent, of Series A had moulted once or more as against nineteen out of twenty-five, or 76.0 per cent, of the living females of Series B. At the same time only 6.7 per cent of the males and 5.3 per cent of the females of Series A had moulted twice while 53.8 per cent of the males and 50.0 per cent of the females of Series B had done so. ‘The one individual that had moulted a third time belongs to the series with the greater injury in accordance with the general rule for the other moults. The final data as collected 181 days after the operation, when the experiment was closed, show the same advantage of Series B over Series A. ‘Thus only 42.9 per cent of the living ‘males and 16.7 per cent of the living females have moulted twice while 76.9 per cent of the living males and 87.5 per cent of the living females of Series B have done so. For the third moult at the same time there are none of the males and only one female, or 5.6 per cent of the living females, in Series A as against one male, or 9.1 per cent of the living males, and three females, or 15.0 per cent of those living, in series B. The general result is very clear. The individuals of Series B moult more rapidly than those of Series A. Emphasis must be again laid on the fact that Series B differs from Series A only in the greater degree of injury in the former. All other conditions are as nearly alike as possible in the two cases. Specific Amount of Regeneration. The amount of regeneration of the right chela at the end of the first moult divided by the thoracic length gives a quotient which may be called the specific amount of regeneration or the amount per unit of thoracic length. It is a fairly constant quantity for the Rate of Regeneration. 361 individuals of one sex in a series and is equal in the two series. (See Tables VI and VII, pp. 357 and 358.) The amount of regeneration of the right chela at the end of the first moult is therefore the same no matter what the degree of injury may be. The average specific amount of regeneration for the males of Series A at the end of the first moult is .444 (+.003). For the males of Series B at the same time it 1s .435 (+.006). ‘The differ- ence between the two is just equal to the sum of the probable errors and therefore cannot be considered as significant. Like- wise the females of Series A have an average specific amount of regeneration equal to .400 ( +.006) and those of Series B an aver- age of .403 (+.004). ‘The difference is less than the sums of the probable errors and is therefore not significant. These results show very definitely “that the specific amount of regeneration of a remov ed chela at the end of the first moult after the operation is a constant which is not affected by the time of the moult, the size of the animal, or the degree of other i injuries to the individual. Four of the individuals, eh moulted very soon after the operation, three within the first day and one in three days, are not included in this statement. None of the individuals moulted in the interval between three and twenty-seven days after the operation so that it is not possible to say to what degree the state- ment holds true for this period. For all periods AN 27 days up to 181 days when the experiment was closed the specific amount of regeneration is a fairly constant quantity for the first moult after the operation. Specific Rate of Regeneration. The specific amount of regeneration of the right chela divided by the number of days between the date of operation and the first moult gives the specific rate of regeneration. The specific rate of regeneration is the amount of regeneration per unit of thoracic length per day. The average specific rate of regeneration of the two chelz in the series with the greater injury (Series B) 1s greater than that of the one removed chela in Series A. This is brought out very definitely in Tables VI and VII (pp. 357 and 358). For the males the values of the specific rate are .0049 ( +.0003) for Series A and 0080 ( 4.0005) for Series B. For the females the corresponding 362 Charles Zeleny. values are .0030 ( +.0001) for Series A and .0083 ( +.0007) for Series B. In ae instance there is a very striking advantage of the series with the greater injury over the one with the lesser injury. This amounts to 63 per cent for the males and 177 per cent for the females. The individuals with the two chelz and the last two pairs of walking legs removed as compared with the individuals in which the right chee alone is removed regenerate the right chela more rapidly ee do the latter. This takes place notwithstanding the fact that at the same time they have also to regenerate he left chela at the same rate as the right one and also the last two pairs of walking legs. ‘The individual therefore which has the greater amount of material to regenerate regenerates each part faster than does the individual with the smaller amount of removed material. Relation between Rate of Moulting and Rate of Regeneration. The fact that the specific amount of regeneration is a constant for all individuals of a sex makes the relation between the rate of regeneration and the rate of moulting a very close one. One of three possibilities in the relation of the two must be the true one. The more rapid rate of regeneration of the limbs may be the cause of the acceleration of the moulting or the opposite may be the case or finally the two phenomena may be co-ordinate and only indirectly related. If the first is the case the growing limb-buds in pressing against their chitinous envelopes more vigorously 1 in Series B must be supposed to act as the stimuli for the increase in the rate of moulting. If the second possibility is true the first result of the operation is an acceleration of the rate of moulting which secondarily affects the rate of regeneration. Finally it is possible that the stimulus of the removal of the limbs acts directly and independently upon both regeneration and moulting processes. 5- DISCUSSION. In opposition to the very common belief that an increase in the degree of injury to an individual implies a lowering of its ability to repair the sustained injuries the experiments on the several forms mentioned above have shown that with an increase in the 1Gelasimus, Alpheus, Ophioglypha, Cambarus (see pp. 347-362). Rate of Regeneration. 363 number of removed legs or arms there is an increase and not a decrease in the rate of regeneration of each. ‘This striking fact must be reckoned with in any theory bearing on the nature of regeneration. It would be premature to attempt to build up a constructive theory on the basis of the few facts so far discovered. Enough, however, is clear to make profitable the mention of the bearing of the facts on some of the more common theories of regeneration. 1. Observers who have had to do with the responses of animals to adverse conditions have pointed out again and again the pecu- liar fact that the response to such adverse conditions is in a direc- tion advantageous to the animal. ‘The results of the present ex- periments may therefore be taken as another instance of such a response. [he crayfish with the greater number of removed legs and the brittle-star with the greater number of removed arms respond to the greater injury by an increase in the rate of regenera- tion of each member. Obviously the animal with the greater number of removed appendages has more need of the absent organs than does the other. It therefore may seem very plain to some that the rate of regeneration is greater in the one case because there is more need of a rapid replacement in that case. It is but a step further to the familiar statement that the cause of any need in an organism may be taken as a sufhcient cause for the fulfll- ment of that need. ‘There are unfortunately a few who have been and will continue to be satished with superficial explanations of this character. For these the problem of regeneration is con- sidered solved when the naive statement is made that if a part of an animal is removed it is obviously more advantageous to the animal to regenerate a new part than not to regenerate one. 2. The suggestion may be made that the animal with the greater number of appendages gone, exercises the regenerating ones more vigorously than does the animal with the smaller num- ber gone. As a result of this greater activity the regenerating appendages erow more rapidly in the former case than in the latter. In support of this idea it may be said, for instance, that the animal which still has one uninjured chela concentrates its chela-functions upon that organ, thereby lessening the activity of the regenerating bud. On the other hand the Haima with both chelze gone having no uninjured chela upon which to concentrate its chela-functions exercises to their full extent the developing 364 Charles Zeleny. functions of the new buds. ‘Thus each of the new buds grows faster than the single bud of the other animal because each gets more exercise of its parts than does the latter. Unfortunately observations made upon the individuals of the two series did not show any difference between them as regards the activity of either the old or the new parts. It is, however! very hard to judge differences in activity in animals like the crayfish in the present experiment, for the specimens are observed only when disturbed by the presence of the observer. ‘The individuals with one remain- ing chela under these circumstances naturally often put themselves in a defensive attitude threatening the observer with their unin- jured chela. The members of the other series having no chela cannot do this. A strong individuality was found in the members of both series. Daily ohsery ation of the individuals in the experi- ment for 181 days with but a few gaps enabled me to make out striking differences in the activities of members of a single series. ‘These andi idual differences in function were not correlated with any differences in the resulting regeneration as far as I was able to decide. Though there is no evidence one way or the other from the present instance, the comparative activities of the organs in future experiments of a similar character should be carefully observed. On the other hand there seems to be great danger in carrying the idea too far for it is inconceivable to me how the attempt of an animal to exercise a function for which it has no morphological background can lead to the formation of a structure furnishing the necessary background. Does not such a state- ment of the case come dangerously near to the other statement that “the cause for the existence of a need is a sufficient cause for the fulfillment of that need?” A mystical attempt to function resulting from the need to function has been supplied, that 1s all. 3. [he difference in the mechanical redistribution of food materials which results from the difference in the extent of the injury may be supposed to cause directly the greater rate in the one series. A discussion of the assumptions which must be made in order to explain the facts on this basis will be interesting. Before going on it will be well to recognize the fact that the difference in activity of the parts in the two series as formulated above under the second suggestion (pp. 363, 364) is supposed to lead to a difference in the distribution of food materials which in turn brings about the difference in rate of growth. ‘The same must be said of all Fic. 4. ee aes , Fie. 5. Fic. 6. Diagrams to illustrate the distribution of food materials of a constant amount (K), to the limbs of an unoperated crayfish (Fig. 4), a crayfish of Series A (Fig. 5), and a crayfish of Series B (Fig. 6). 366 Charles Zeleny. other attempts at explanation. ‘The factor that they bring up may be supposed in every case to cause a difference in food- material distribution which in turn causes the difference in rate of erowth. Under the present head it is therefore pertinent to dis- cuss only the supposition that the direct mechanical disturbance of the channels of food-distribution shunts the materials off in such proportions into the new channels as to make probable an explanation purely on these grounds. At the very beginning the pure assumption must be made that the total amount of food materials elaborated and involved in chela-building in our case is a constant (K) in all individuals regardless of the character of the injury. The source and distribution of the food stuffs may be indicated by the diagrams shown in Figs. 4, 5 and 6. ‘The facts to be explained are that chela C in Series A Ges 5) regenerates less rapidly than either chela C or C* in Series B (Fig. 6). The assumption according to the hypothesis now being tested is that the rate of growth and regeneration is determined by the amount of food riences an increase in the amount of food material going to a part determining the increase in rate of growth or regenera- tion of that part. The assumption is also made that the same kind of material is used in the growth of an uninjured chela as in the regeneration of a removed one and that the total amount of this food-material being distributed is a constant (K). The distribution of the food-material may be represented for the three cases given in Figs. 4, 5 and 6 by the following formulz: K = C+C'+D+D'+E+E'. (Unoperated series, Fig. 4.) = rep. bud C+C'+D+D!+E+E (Series A, Fig. 5.) = reg. bud C+ reg. bud C'+ D+ D'+reg. bud E+reg. bud E'. (Series B, Fig.6). Now the rate of regeneration of chela C in Series B is greater than that of chela C in Series A. Therefore according to the pres- ent hypothesis the amount of food material going to the former chela bud is greater than that going to the latter or reg. bud C (in series B) > reg. bud C (in Series A). Therefore reg. bud C'+ D + D'+ reg. bud E+reg. bud Et > C'+ D+ D!+EFE* (Series B). (Series A). But D + D' is the same in the two series. Therefore reg. bud C'+reg. bud E+ reg. bud E1< C'+ E+ E} Rate of Regeneration. 307 or the regenerating buds of the chelz and walking legs receive less food material than do the same organs when uninjured. As the total amount of food material to be distributed is by hypothesis constant it follows that when a greater number of appendages is removed the surplus of material is greater. The surplus of material is therefore greater in Series B than in Series A and crowds upon both the regenerating parts more vigorously than in the latter. [he regenerating buds in Series B, the series with the greater injury therefore grow more rapidly than those of Series A. Evidently when the degree of injury becomes great enough to dis- turb the mechanism of production of food material so that the amount of the latter is diminished and there is no longer a con- stant quantity K, the present statement cannot hold. 4. The results of the experiments on the relation between the degree of injury and the rate of regeneration of the crayfish and the brittle-star bring out very strongly an essential difference between crystals and organisms. In the former no matter what the number of removed parts the growth of each in a nutrient solu- tion is entirely independent of the number or character of other removed parts in the same crystal. In the organism on the other hand there is no such independence. No part of the organism can be removed without affecting all other parts. ‘This difference may, however, be due to the difference in the nature of the food- supply and not to an essential difference in the structures them- selves. In crystals the food material is external and practically inexhaustible. Each part is thus independent of restrictions due to amount of food material. 5. The stimulation of the nerve of a leg or arm as a result of the injury to that member may be supposed to induce the processes leading to its regeneration. If it is assumed that an increase in stimulation causes more than a corresponding increase in intensity of such processes there will result a greater rate of regeneration in an animal with a greater injury than in one with a lesser 3 injury. Physiologists have found that in general the curves for increasing stimulus and increasing response are not parallel. In some cases and this is especially true near the lower limits of the curves, the response curve runs up faster than does the stimulation curve. The organism possesses some inertia in every case and it 1s neces- sary to overcome this before any response at allis obtained. Hav- ing passed the lower limit, however, there is a very rapid upward 368 Charles Zeleny. rise of the response curve in many instances. May not the acceleration of the regeneration rate with an increase in injury to the animal be explained on a similar basis? 6. In a series of experiments on the opercula of the Serpulid worms it was shown that when the large functional operculum is cut off near the middle of its stalk the small rudimentary oper- culum of the opposite side develops into a functional operculum while the remaining stalk of the old functional drops off at its base and in place of it a new rudimentary operculum is developed. The final result of the operation is therefore a reversal in position of the opercula. When the rudimentary operculum is cut off a new rudimentary is regenerated in its place. When the whole head region of the animal is cut off the two opercula which are regenerated are equal in size and resemble the old functional one.’ It seems therefore that when one of the regenerating buds gets a start over the other it holds the latter back at the rudimentary stage. On the other hand when both buds have an equal start two opercula of equal size are developed. A retardation stimulus must be assumed to be given out by the functional operculum which holds the rudimentary operculum in check. When the organ is removed the retardation stimulus is likewise removed and the rudimentary operculum is enabled to develop into a functional one. ‘The same method of reasoning may be applied to eee case of the regenerating chelz of the crayfish. Each uninjured chela may be assumed to exert a retarding influence upon the growth or regeneration of all the others. When only one chela is removed the number of uninjured limbs remaining is greater than when the other chela and the last two pairs of walking legs are also removed. ‘The retardation influence in the former case is there- fore greater than it is in the latter and correspondingly the rate of regeneration in the former case is smaller than it is in the latter. The term “retardation influence or stimulus” is undoubtedly a very vague one. It may perhaps be best considered as a nervous 'The results obtained for the chele of Alpheus are essentially similar except in the case with both chele removed where the regenerating chele are not alike. The difference here is probably due to the fact that the removal of both chele at their breaking joints leaves a basal stump on each side and is not a total removal as in the corresponding case of the Serpulid opercula. However, the resulting chele even here show an approach toward similarity. (See Przibram, ’o1, Arch. Entw. Mech., xii; Wilson, ’03, Biol. Bull., iv; Zeleny, ’05, Journ. Exp. Zodl., ii.) Rate of Regeneration. 369 influence exerted either upon the other organ or organs directly or else upon the mechanism for carrying food materials to those organs. The retardation influence 1s brought out in a somewhat different light when considered as a manifestation of the inertia of the organism. This idea also tends to unite the two suggestions of positive stimulation as a result of injury to the animal and the negative or retardation stimulus exerted by the uninjured organs. The former acts after the operation in overcoming the inertia of the organism. The latter on the other hand is merely a mani- festation of the same inertia acting before removal. The crayfish data when taken alone furnish no evidence in favor of either one of these two views as opposed to the other. The experiments on the Serpulid opercula and Alpheus chelz, however, cannot be as well explained by the positive stimulation theory as ‘by the retardation view. The foregoing speculations are evidently of but small direct value. ‘Their purpose is accomplished if they have emphasized the importance of the discovered relation between the degree of injury and the rate of regeneration in any general theory of regen- eration. 6. SUMMARY. A comparison. was made of the rate of regeneration and the rate of moulting in two series of crayfish with different degrees of injury. In one series the right chela alone was removed. In the other series the two chelz and the last two pairs of walking legs were removed. It was found that the rate of regeneration of each chela in the series with the greater injury is greater than that of the single removed chela in the series with the lesser injury. Likewise the rate of moulting of the animals is greater in the former series than in the latter. Indiana University, May 31, 1905. BR 2 BAGS - —— ee CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. E. L. MARK, Director.—No, 169. THE MOVEMENTS OF THE SWIMMING-PLATES IN CPENOPHORES, WITH REFERENCE TO THE THEORIES OF CILIARY METACHRONISM. BY Ga ES PARKER: WitH 2 Ficures. i INTRODUCTION: Since the publication in 1880 of Chun’s elaborate monograph on the ctenophores, it has been generally admitted, contrary to the opinion of many of the older investigators, that the swimming- plates of these animals are their principal organs of locomotion. Moreover, the ciliary nature of these organs may now be regarded as well established, and their relatiy ely enormous size has already made them favored objects with investigators of ciliary phenom- ena. As is well known, these swimming-plates are arranged in rows and the members of each row, like ordinary eles beat metachronally, not synchronally. he explanation of this pecu- liarity has collletl grin era comer opposing views. According to the first of these, which has been developed chiefly by Engel- mann ('68, p. 475; °79, p- 388), it is maintained that one element beats immediately after its next neighbor in a given order because of a nerve-like impulse that is supposed to pass from cell to cell and thus to bring into action in regular sequence the overlying elements. ‘This may be called the neiaia theory of ciliary action. According to the second view, advanced in the main by Verworn (90, p. 175), the cause of metachronal action is not to be sought for in the cell-body proper, but rather in the mechanical Sines of one cilium on another, in that the action of one cilium mechanically stimulates the next one to action. “This may be called the me- chanical theory of ciliary action. Because of the minute size of ordinary cilia, experimental tests of these two theories are not easily carried out; hence the anatomical conditions presented in ctenophores are of unusual importance. It is the principal object 408 G. H. Parker. of this paper to discuss the cause of metachronism in ciliary action as exemplified in the swimming-plates of these animals. The material upon which I worked consisted almost entirely of the common summer ctenophore of the New England coast, Mnemiopsis leidyi A. Agassiz, though I also made some observa- tions on the winter species Pleurobrachia rhododactyla L. Agassiz. ‘The work was done for the most part during the last few summers at the Wood’s Hole Laboratory of the United States Bureau of Fisheries, to the officers of which I am under obligations for many kindnesses shown me. II. OBSERVATIONS. Anatomical. Mnemiuopsis leidyi is a lobate ctenophore measuring often as much as seven or eight centi- metres in length. Its external form is shown in Fig. 1, which is a view of the animal so placed that its sagittal plane corresponds to the plane of the paper. The mouth is directed downward and the two large lobes that charac- terize this group of ctenophores are seen at the right and left of it. The aboral pole is pointed upward and four of he eight rows of swimming-plates are shown converging toward it. Their relation to the sense body at the aboral pole can be seen ae eae clearly in Fig. 2, where it will ai pine cnepuee tar ie Mss: be observed that from the most the aboral pole is uppermost. Two short sub- aboral plate of each row a nar- lone eaheriee ewimming plates and two row band extends to the sense ong su Sagitta ones are seen converging to- ward the aboral pole. The subsagittal ‘Tows body. These bands, before they ae vibratile lines far over the surface of reach the sense body, unite in pairs and enter that organ as four bands. As will be seen by comparing Figs. 1 and 2, the rows of swimming-plates are either long or short and the pairs The Movements of the Swimming-Plates in Ctenophores. 409 formed by the unions into the bands consist always of a long row combined with a short one. Since the long rows lie near the sagittal plane and the short ones near the transverse, they have been called, respectively, subsagittal and subtransverse. ‘The combination of a long subsagittal row with a short subtransverse one to form a pair has long been known to be one of the structural characteristics of the lobate ctenophores, and, as will be shown later, this feature is not without its physiological significance. Each such pair, as can be seen in Fig. 2, is restricted to a quad- rant of the animal’s body. ‘The number of swimming-plates in the subsagittal and the sub- transverse rows varies more or less with the size of the animals. Thus in a small specimen eight millimeters long the subsagittal rows contained each about 19 plates, the subtransverse ones about 12; while in a large individual sixty millimeters long, there were about 73 plates in each subsagittal row and about 39 in the subtransverse ones. In the speci- men from which Figs. 1 and 2 were drawn, there were about 29 plates in each subsagittal row and about 17 in each subtransverse one. In Mnemiopsis the bands that lead from the sense body to the swimming- ee “1: : Aboral view of Mnemiopsis leidyi. plates are ciliated, as 1n other cteno- The four subsagittal rows of swimming- phores, and, as Samassa (92, p- 229) plates, two from each lobe, and the has shown for other lobate forms, a ee ca bares band of cilia connects plate with plate. In this species, however, the spaces between the plates seem to be much more sparsely provided with cilia than in other lobate ctenophores, if, in fact, cilia are not sometimes entirely absent from these regions. The second species upon which [ worked, Pleurobrachia thododactyla, was of simpler structure than Mnemiuopsis. It belongs to the Cydippidz and has the typical form of an oblong spheroid. Its eight rows of swimming-plates are of about equal length and can be readily distinguished as subsagittal or sub- transverse only by their relations to other parts in the animal’s 410 G. H. Parker. body. Ina specimen of average length, about sixteen millimetres, there were approximately 40 plates i in each row. Physiological. The resting position of the swimming-plates in both Mnemiop- sis and Pleurobrachia is one /n which the individual plate 1 is turned close to the body of the animal and with its tip directed orally. In action the plate makes a vigorous stroke aborally and then returns to its resting position. In consequence of such movements carried out more or less simultaneously by certain plates in each row, the animal’s body is moved through the water with the mouth forward. ‘The plates in any one row strike one after another beginning at the aboral end, 7. e., to use the term proposed by Verworn (90, p. 152), they beat metachronally. Ordinarily the first plate to strike is the most aboral one and the others follow in sequence giving rise, by the order of their beat, to a wave-like appearance w hich progresses, of course, in an oral direction. Chun (80, p. 172) has shown that when in a normal animal a wave starts over one row, a like wave also starts over the other row of the same quadrant, 7. ¢., the two rows of any quadrant act in union. ‘This relation was observed by Verworn (91, p- 456) in all the ctenophores that he studied, and it ts certainly an invariable occurrence in Mnemiopsis, but not in Pleurobrachia. In Pleuro- brachia, though the rows of plates on the same quadrant are often seen to beat in unison, they also frequently beat independ- ently. ‘That their beating in unison is not a mere matter of accident is seen from the fact that, whereas rows on’ the same quadrant often beat in unison, adjacent ones belonging to different quadrants do not beat in this manner. ‘There can be no question, I believe, that the rule laid down by Chun, to the effect that rows on the same quadrant always beat in unison, has its exceptions, for in Pleurobrachia the two rows of any quadrant may beat independently. As might be expected from the researches of Chun (’80, p. 172), the removal of the sense body from Mnemiop- sis or from Pleurobrachia is invariably followed by a complete loss of the partial or perfect unison of action between rows of plates. Since there is an agreement in the metachronism of the two rows of plates on any quadrant in Mnemiopsis, there should The Movements of the Swimming-Plates in Ctenophores. 411 be a synchronism in the action of the corresponding plates in these two rows, and such proves to be the case. ‘This condition is very noticeable when at the beginning of a series of swimming- plate movements, the waves run at varying rates, for when a wave passes rapidly or slowly over one row, it passes at the same rate over the other row of the same quadrant. Similar conditions were observed in Pleurobrachia when its swimming-plates were acting In unison. The reversed action of the swimming-plates in ctenophores has been stated to occur by numerous observers, but the expression reversed action in this connection is undoubtedly somewhat ambiguous. In the so-called reversal of cilia and other like organs at least two kinds of reversal are possible: a reversal of the propagation wave, “ Reizwelle”’ of Engelmann, and a reversal of the effective stroke of the cilia (Parker, ’05, p. 9). In the first instance the question turns on the sequence in which the cilia beat; thus in the normal action of a series of cilia, element a may beat first and z last, the propagation wave passing from a to z; while in reversed action zx would beat first and a last, the wave passing in the reverse direction. In the second instance only the effective stroke of the cilium is concerned; this may be normally toward x or reversed toward a irrespective Bf the sequence in which the cilia of the series act. In ctenophores a reversal of the direction of the propagation wave has often been observed. ‘This was early noticed on frag- ments of Beroé by Eimer (80, p. 226), an observation confirmed on this and other species by Chun (’80, p. 182) and by Verworn (ge. p= 107; O1, p. 459), though the latter is misquoted i in this respect by Putter (03, p. 35). “Reversal of the propagation wave occurs occasionally in Pleurobrachia. When a rapid wave from the aboral end of the animal reaches the oral limit of a row of plates, it may be reflected aborally over the row again, but it seldom retraces its course for more than one-third the whole length of the row. As Verworn (’90, p. 167; 91, p. 440) observed, these reversed waves can often be induced by stimulating mechani- cally the oral end of a row of swimming-plates. In Mnemiopsis I have never observed unquestionably reversed waves, nor have I been able to induce them by special stimulation. Some slight evidence of reversal has been seen when a relatively slowly moving wave near the oral end of its course 1s overtaken 412 G. H. Parker. by a more rapid one. ‘This is seen to sweep over the slower wave and, as it does so, what seems to be a reversed wave starts from the point of collision and runs aborally over not more than six or eight plates at most. This short wave is the only evidence of reversal that I have found in Mnemiopsis; it is my belief that this reversal of the swimming-plate action, so common in many ctenophores, 1s almost entirely absent from this species. According to previous investigators, ctenophores can reverse the effective stroke of the swimming-plates as well as change the direction of the propagation wave. Under ordinary conditions, the effective stroke carries the animal with the oral end forward; when this is reversed, the animal moves with the aboral end ahead. Chun (80, p. 181) mentions that this reversed form of locomotion is a regular though rare occurrence with all cteno- phores, especially hen by normal locomotion their oral ends collide with some fixed body. Verworn (’91, p. 432) states that he has on rare occasions observed this reversed swimming in Eucharis and Callianira, but not in other species of ctenophores. I have never seen any evidence of the reversal of the effective stroke in either Pleurobrachia or Mnemiopsis and I am inclined to believe that Chun’s statement that the effective stroke can be reversed in all ctenophores, may be a mistake based upon a con- fusion of this form of reversal with the reversal of the propagation wave. In Pleurobrachia it can be easily shown that when the propagation wave reverses from an oral to an aboral direction the swimming-plates continue their effective stroke in an aboral direction as before. When a row of swimming-plates in Pleurobrachia or Mnemiop- sis 1s cut through so as to divide it into oral and aboral portions, the plates in both parts cease to move for a short time and when they resume their activity, the two parts are found to beat differ- ently, 7 1. ¢., their propagation waves are found to be independent. In this respect the American species agrees with the European forms experimented upon by Eimer (’80, p. 227), Verworn (’go p. 167), and others. If a row in Mnemiopsis is cut with care, the aboral part almost immediately begins to beat metachronally with reference to its fellow of the same quadrant, and the oral part reestablishes independent movements in a few minutes or even seconds. In the quickness of recovery of the oral part Mnemiop- sis 1s in strong contrast with Beroé and Eucharis, in which, accord- The Movements of the Swimming-Plates in Ctenophores. 413 ing to Krukenberg (’80, p. 2) and Verworn (90, p. 156), the activ- ity of the oral part may not return for an hour or so after the row is cut. [he waves in the oral part of Mnemiopsis always proceed from near the cut end of the row orally; the most aboral plate to show motion, however, is not the one next the wound but usually the third or fourth from it. The oral portion will thus move its swimming-plates for hours without relation to the movements of the aboral part. I have never seen any evidence of the reéstablishment of harmony in the two parts of a severed row such as has been described by Eimer (’80, p. 229) and Verworn (’90, p. 167; ’9I, p- 463). When a swimming-plate band is cut through, not where ae are swimming-plates but between the most aboral plate and the sense body (compare Fig. 2), the whole row in its movements becomes independent of the sense body and if the sense body is destroyed, the co6drdination of the four pairs of rows entirely disappears, as has already been shown by Verworn (’gI, pp. 457-459) and others for several European species. When a Mnemiopsis is cut in two transversely, the parts of rows on the aboral portion retain their coordination as in a normal animal; those on the oral part, as might be expected, lose all signs of such relations. It is clear from this and the preceding experi- ments that the coordinating influences proceed from the aboral pole, and, when this is lost, coordination disappears. In this respect my observations confirm those of Krukenberg (80, p. 2) on Beroé and are opposed to those of Eimer (’80, p. 231), who stated that the oral half of Beroé is indistinguishable in the move- ments of its plates from a whole animal. When a Mnemiopsis is shaken in sea-water, it can be broken easily into fragments and the plates attached to these pieces will continue to beat rhythmically and metachronally for from one to two days. As Verworn (’90, p. 157) has shown for Cestus, so also in Mnemiopsis, even a oa plate with a small basal piece of protoplasm will beat rhythmically for a long time. ‘This condi- tion led Verworn to believe that each plate possessed a certain degree of autonomy, which was seen in the continued activity of the isolated plates and must be imagined to be counteracted by some influence when the plate as a member of a row was quiescent. But in my opinion the swimming-plate, when it beats, does so because it 1s stimulated, and its quiescence is evidence of the absence of appropriate stimuli. When it beats normally on 414 G. H. Parker. a whole animal, it does so in response to a wave of stimulation from the aboral pole, the cessation of which is followed by the cessation of the movement in the plate. When one plate with a small amount of protoplasm attached continues to beat, as it often will for hours, it does so because the fragmentary condition of its base exposes this part to continual stimulation. I see no reason to assume that the plates possess an autonomy that is inhibited much of the time by the animal. A fragment of a swimming-plate of Mnemiopsis made by splitting the plate lengthwise will continue to beat if a small basal mass of protoplasm is still attached to it. Whole swimming- plates or fragments of plates cease to beat when the base is trimmed off to such an extent that only the swimming-plate proper is left. In this respect the plates of Mnemiopsis resemble those of the ctenophores on which Verworn (90, pp. 158, 161) experimented. This failure of the isolated plates to vibrate has generally been attributed to the loss of a stimulus normally received from the basal protoplasm, but Putter (’03, p. 42) has suggested that it may be due to the rapid death of ae plates after isolation from the living substance of the animal. That this is not so in Mnemi- opsis is seen from the fact that a fragment of a plate cut off from its basal protoplasm and kept in sea water half an hour trembles and curves when a little picric acid is applied to it just as the living plates do on a whole animal when this reagent is poured on them. I therefore believe that the quiescence of isolated plates is due to the absence of a stimulus to contraction and not to early death. It is evident from what has preceded that the rows of swimming- plates of ctenophores ordinarily beat in pairs corresponding to the quadrants of the animal’s body and that the plates of any row beat metachronally beginning ordinarily at the aboral end. As Chun (’80, p. 172) long ; ago pointed out, that which regulates their beat proceeds usually from the region of the aboral pole and here four centers must be assumed, one for each quadrant of the ani- mal’s body. It is also evident that the regulating influence in its passage from the aboral pole is strictly limited to the bands leading from the sense body to the rows of plates, and to the rows of plates themselves, and that, though the waves usually start from the aboral end and progress toward the oral one, they may in some species reverse and run some distance aborally. All these The Movements of the Swimming-Plates in Ctenophores. 415 facts are explainable on either the theory of neuroid transmission as advocated by Engelmann, or on that of mechanical transmis- sion as put forward by Verworn; for on the former assumption the band of epithelium leading from the sense body to the oral end of a row of swimming-plates may serve as a transmitting tract, and on the latter the ciliated bands leading from the sense body to the rows of plates may serve to transmit the mechanical disturbance from the center to the plates. | propose now to turn to certain observations that are, in my opinion, inconsistent with one or other of these theories. Since in accordance with the idea of mechanical transmission the mechanical activity of the vibratile elements is a necessary ‘accompaniment of transmission, it follows that any means of bringing this activity to a standstill ought to check transmission. It might be supposed that the cutting ‘off of one or more plates would produce such an effect. When this is done in Cestus, according to Verworn (’g0, p. 173), and in Mnemuiops‘s, according to my own observations, the waves still pass regularly over the whole row of plates and are not interrupted by the interval from which the plates have been removed. Since, however, the spaces between the plates in Cestus, as well as in the lobate ctenophores, have been shown by Samassa (’92, p. 229) to be ciliated, it might be assumed that these cilia in their vibrations transmit the mechanical disturbance over the whole row. ‘The assumption that in the absence of plates the cilia may transmit the disturbance is, however, in my opinion improbable, for the space made by the removal of a plate is so considerable in comparison with the length of the cilia that, unless we assume as Verworn (’90, p. 173) does that the whole base of the plate is surrounded by cilia, [ see no way by which the mechanical disturbance made by the cilia on one side of the root of the plate could influence those on the other side and thus effect transmission. As I have never seen in Mnem- lopsis any reason to believe that the plates are surrounded at their bases by cilia, I do not believe that transmission can be accounted for in the present experiments by the mechanical theory, even admitting the presence of cilia between the plates. A modification of the experiment just described has been employed by Verworn (’90, p. 171) with the view of testing further the nature of transmission. Ties experiment consisted in restrain- ing a plate from beating instead of cutting it off and then ascer- 416 G. Eo Parker. taining whether waves passed beyond it. When a plate in Beroé is Farued aborally by a lancet point, waves from the aboral end fail to pass this plate. If only the tip of the plate is held and the base is allowed to move, the wave passes onward to the oral portion of the row. These observations led Verworn (’g0, p. 171) to conclude that the mechanical vibrations of the plates were neces- sary for transmission, and he drew this conclusion notwithstanding the fact that in Cestus he (’g0, p. 172) found that the holding or even the pulling out of a plate did not interfere with transmission. Verworn confessed to have been astonished at the conditions found in this species, but, as already stated, he believed that they might be explained on the assumption that the base of each plate is more or less surrounded by cilia which after the removal of the plate continue to transmit mechanically. Unfortunately I have been unable to try the experiment of restraining plates in Mnem- iopsis, for the rows of plates in this species, like those in Beroé, as pointed out by Krukenberg (’80, p. 10), are so sensitive to mechani- cal stimulation that the moment they are touched they are drawn down into the animal’s body to such an extent as to make experi- ments of this kind very unsatisfactory , if not impossible. Although the great sensitiveness of the rows of plates in Mnem- iopsis prev ented me from trying the experiment of holding plates individually, it afforded a very natural means of checking their action. As Verworn CoO; py 70) has shown, when the muddle of a row of plates is touched, the row in that region becomes depressed and the edges of the depression fold over and cover the plates. Thus in Mnemiopsis half a dozen plates may become so much restrained that they will not show the least motion and yet waves that arrive at the aboral entrance to this depression emerge from its oral end with the greatest regularity. This may happen while the covered region is under dose inspection through a lens and gives not the least sign of plate or ciliary movement. I am, Be refares forced to pentlude that, contrary to the statement made by Verworn (’90, p, 170), such restrained tracts transmit with perfect regularity even in the absence of observable ciliary and plate motion. Kraft (’90, p. 223) in his study of the ciliated epithelia of verte- brates, showed that, though low temperature may bring cilia to a standstill, it does not oreatly check the transmitting power of the tissue. It ought, therefore, to be possible to check the action of The Movements of the Swimming-Plates in Ctenophores. 417 plates in a ctenophore by cold and yet leave the transmitting power of the row unimpaired. To test this proposition, I passed a small curved metal tube through the substance of a Mnemiopsis directly under one of its rows oF swimming-plates and at right angles to the direction of the row. The animal was anchored by being pinned in a small aquarium of sea water whose temperature was 21°C. Normal waves of action were seen to course over the row of swimming-plates under which the metal tube had been placed. I now passed through the tube water of a temperature between Aen 5 CA steady flow was kept up to insure as complete a eine as possible of that portion of the row under which the tube went. The chilled plates soon ceased to move and the waves appeared to jump from the aboral side from which they approached the chilled region to the oral one beyond it. Sometimes half a dozen waves in rapid succession appeared thus to jump this chilled region. But the best evidence was obtained when the waves ran at considerable intervals, at which times the correspond- ence between the parts of the wave in front of and behind the chilled region was most striking. To be certain that there was no movement of cilia or plates in the chilled region, a small amount of powdered carmine in sea water was placed on the plates of that portion. ‘The carmine remained motionless while wave after wave ran over the aboral and the oral parts of the row. At the close of the experiment the chilled region was allowed to regain its normal temperature, whereupon its plates became vibratile again and the waves passed without interruption. ‘This experiment was repeated on six different individuals and with constant results. In one instance the temperature of the water used for chilling the tissue was 8.5° C. and under this condition the cessation of movement was only partial, but in all other experi- ments the temperature was kept at 5° C. or lower with the result that complete cessation of movement invariably followed. Hence it is fair to conclude that in Mnemiopsis a temperature of 5° C. or less will check the movement of the swimming-plates without essentially altering the transmitting power of the row. In handling ctenophores in the experiments last described, I noticed that when the row of plates under which the metal tube passed was subjected to a little local stretching by the awkward manipulation of the tube, the plates often ceased to vibrate in the stretched region. On repeating this operation I found that as a 418 G. A. Parker: rule the slight stretching of a region would bring the plates of that part to a standstill, though i it did not interfere seriously with trans- mission. But it must be noted that in such an operation much care must be used not to overstrain the tissue, for otherwise a permanent cessation of action will follow. Avoiding this difhculty, however, mechanical strain, like low temperature, may be made to check motion without interfering with transmission. Ill. THEORETIC CONSIDERATIONS. The results of the experiments just described, in which the swimming-plates of ctenophores were removed or restrained, or the row chilled or stretched locally, afford good grounds for deny- ing to the mechanical theory any essential part in the explanation of ciliary metachronism. If ordinary transmission is really depen- dent upon the mechanical action of one element on the next, it 1s inconceivable how such a process can be accomplished when these elements for any reason cease to move. ‘That transmission does take place after the swimming-plates have been brought to a stand- still by physical restraint, cold, etc., is unquestionable. Verworn (90, p. 172) admitted surprise when he found that in Cestus Seas occurred even after the removal of a plate and he was led to assume a continuous band of cilia to account for this condition. In Mnemiopsis no such band is present and yet transmission takes place even after the removal of a plate. The failure of a wave to pass when the plates in Beroé are restrained from moving is not, as Verworn believed, a satisfactory test of the nature of transmission, for, notwithstanding the care used in restraining the plate, the operation may influence the deeper parts of the tissue and thus check transmission as well as plate movement. The fact that transmission does occur in Mnemuiopsis when the plates are restrained, shows how treacher- ous such negative evidence is. These facts, together with the evidence from chilled and stretched rows, show, | believe, that the mechanical theory is not a necessary part of the explanation of ciliary metachronism. Although the mechanical theory may not be the correct explana- tion of transmission, its rejection does not imply a rejection of the idea that the plates are open to mechanical stimulation. Every- one who has worked with ctenophores knows how sensitive the The Movements of the Swimming-Plates in Ctenophores. 419 plates are in this respect. The slightest touch will often cause them to vibrate and will even give rise to a wave which, beginning with the plate stimulated, runs orally over the row. ‘This condition is undoubtedly suggestive of such a view as that advanced by Verworn; his (’90, p. 168) ingenious experiment of attaching plate to plate by cotton filaments and thus obtaining a form of mechanical transmission shows how this idea may fad applica- tion. When, however, it is remembered that in rest the plates point orally, that the propagation wave ordinarily proceeds from the aboral end of the row, and that the effective stroke of the plate is made in the aboral direction, it is clear that each plate as it goes into action does not strike toward the next plate to act but away from it and hence in a direction unfavorable for mechanical stimu- lation. When the propagation wave is reversed, as happens in Pleurobrachia and probably in many other ctenophores, the action of the plates is such that an oral one may well stimulate mechani- cally the next in turn, and, while I believe that the normal wave depends for its propagation upon a neuroid transmission, | am inclined to the opinion that the reversed waves may depend largely on mechanical transmission. As is well known, these neverecd waves seldom extend far and are always insignificant as compared with the normal ones. Hence I do not believe that mechanical stimulation plays any really important part in transmitting the normal wave. Direct stimulation seems to be a possible means of transmission over a cut in a rowof plates. Since both Eimer (’80, p. 229) and Verworn (’90, p. 167) have recorded the occurrence of this form of transmission in Eiooea species, it might be looked for in other forms,though I have beenunable to find any evidence of it in Mnem- iopsis or Pleurobrachia. However, I see no reason why the vibra- tion of a plate on the aboral side of a cut may not stimulate to action a plate on the oral side of the same cut provided the two plates are brought close enough together. ‘The subject is worthy of further investigation. Most of the observations that have been brought forward against the mechanical theory might now be urged in favor of the neuroid theory, for transmission without ciliary or plate motion is what Is implied by this view. ‘The idea that the movement of the swimming-plates is controlled by nerves was held by some of the older investigators such as Eimer (’73, p. 45) and Krukenberg 420 G AE oRarker: (80, p. 5), though on insufhcient grounds, for no one has ever demonstrated that nerves are connected with these plates. ao mann (’87, p. 442) has used the expression “innervated” reference to the rows of swimming-plates, but it is perce evident from other statements in his account (p. 440) that this term is used in a physiological sense and not in an anatomical one, and that he consistently adheres to his original idea (’79, p- 395) of epithelial transmission. Chun (’80, p. 173) made perhaps the best brief statement of the mechanism of transmission in ctenophores when he declared that the rows of epithelial cells served as nerves. It is in my opinion an open question whether in any instance cilia are really controlled by nerves. Such a con- trol is denied by Verworn (’95, p. 251), though Piitter ('03, p. 98) in his recent survey of the whole subject of ciliary activity states that in the larve of certain annelids such control occurs. It seems to me that Putter’s grounds are insufficient for such a conclusion; but, however this question may stand for annelids, in the cteno- phores not the least histological evidence has ever been advanced to show that their rows of plates are accompanied by nerves. Samassa (92, p. 226), who has studied this matter with care, denies that ctenophores have any nervous system properly so called and points (p. 230) to the epithelial bands in Beroé as the transmitting organs. ‘There thus seems to be good reason for believing that Tie epithelial cells on the rows of swimming-plates in ctenophores transmit impulses that control the metachronism of these plates; in other words the neuroid theory, contrary to the statement made by Verworn (’g0, p. 175), is tenable. Such a conclusion is entirely consistent with the results of Gruetzner (82) and of Kraft (go) in their experiments on transmission in the ciliated epithelia of the higher animals, for both investigators found it necessary to assume a deep-seated cellular transmission to explain the spread of ciliary disturbances in active and in quies- cent fields of cilia. Although the results of my experiments make me confident that the metachronism of the swimming-plates of ctenophores is due to neuroid transmission, I do not believe that the facts warrant the extreme position taken by Engelmann (’87, p. 440) that no form of mechanical transmission obeune It seems to me much more likely that, as Chun (’80, p. 174) has declared, mechanical action is a subordinate though real factor in transmission. In my The Movements of the Swimming-Plates in Ctenophores. 421 opinion, this factor would never of itself result in giving rise even to a single full wave, though it might, if vigorously started, carry a wave over a small number of plates. Its influence at most would be only of a subordinate character. Chun’s view that both neu- roid and mechanical factors take part in transmission has been adopted by Putter (’03, p. 98). While mechanical transmission may be of only subordinate importance, mechanical stimulation must be regarded as of no small significance. It has already been pointed out that a plate, if mechanically stimulated, may become the point of origin of a wave which will be transmitted over the row of plates in all respects normally. Hence mechanical stimulation will not only bring a plate into action but will induce the formation of a normal neuroid propagation wave. If the ciliated epithelia of the higher animals and such special- ized structures as the swimming-plates of the ctenophores are con- trolled by impulses that are passed in a definite direction and within circumscribed limits from cell to cell, it seems highly probable that many of the coordinated responses of the lower metazoans and of the early larval stages of the higher forms may depend upon this form of mechanism rather chan on any kind of true nervous organization. ‘Thus it may well be that the slow but uniform responses of sponges to local stimulation may be due to neuroid transmission through their epithelial layers and not through true nervous tissue, aah, as is well known, has been sought for in vain in these animals. ‘The exact orientation to light of larval sea urchins at the blastula or gastrula stage involves a certain coordinated beat of the cilia which, in the absence of ner- vous elements, may well be due to neuroid transmission. ‘Thus animals in such early stagess of growth may carry out by means of their epithelia reactions which in later stages would be performed by a true nervous mechanism. Conditions of this kind lead me to believe that before primitive metazoans possessed any nervous organs whatever, they probably had in their epithelia organs which exhibited the most fundamental property of nervous tissue, namely, a capacity to transmit in a prescribed direction impulses to motion. From epithelia of this kind sense organs and central nervous organs were probably evolved, and yet dhis evolution did not bring about the entire suppression of these primitive pre- nervous mechanisms; for the ciliated epitheila of the highest 422 Ge Hf. Parker. a animals, as well as the swimming-plates of the ctenophores, still possess the power of neuroid transmission. IV. SUMMARY. 1. In Mnemiopsis and Pleurobrachia the swimming-plates normally beat metachronally beginning at the aboral ends of the COWS. 2. In Mnemuiopsis the two rows of plates belonging to the same quadrant of the animal’s body beat in unison. In Pleurobrachia this is also often true, but all eight rows in this ctenophore may beat independently. 3. The propagation wave (“ Reizwelle” of Engelmann) shows scarcely any evidence of reversal in Mnemuiopsis, but often reverses in Pleurobrachia. 4. Reversal of the effective stroke of the plates was never observed in Mnemuopsis or in Pleurobrachia. 5. On cutting a row of plates in Mnemiopsis transversely, the oral part quickly recovers and begins beating, but not in unison with any other part; the aboral part Le recovers soon and beats in unison with the other row of its quadrant. 6. A single isolated plate will beat if it retains a small amount of basal: protoplasm. 7. Plates without basal protoplasm will not beat, though they are not dead. 8. Loss of a plate in a row does not prevent the passage of a wave even in Mnemiopsis where the cilia on the rows do not always form continuous bands. g. When the plates on part of a row in Mnemiopsis are re- strained from moving, an impulse to plate movement may still be transmitted. 10. Cooling a part of a row with water at 5° C. will bring the movement of the plates to a standstill, but not interrupt trans- mission. 11. Stretching part of a row will cause local cessation of move- ment, but will not interrupt transmission. 12. The metachronism of the plates in ctenophores cannot be explained as a result of the mechanical influence of one plate on its neighbor, but the facts observed necessitate the assumption of a deep-seated transmission from cell to cell, nerve-like in character. The Movements of the Swimming-Plates in Ctenophores. 423 13. his neuroid transmission 1s probably supplemented by mechanical transmission, which of itself is insufhcient to carry forward a normal wave. : 14. Phylogenetically an epithelium with neuroid transmission probably preceded true nervous structures and such an epithelium is in_all likelihood the only means of transmission in many animals at their earliest larval stages (blastula, gastrule, etc.) and in such primitive forms as sponges. BIBLIOGRAPHY. Cuun, C., '80.—Die Ctenophoren des Golfes von Neapel. Fauna und Flora des Golfes von Neapel. I. Monographie. xviii, 313 pp., 18 Taf. Emer, T., ’73.—Ueber Beroé ovatus. Zoologische Studien auf Capri. I. g2 pp., g Taf. *80.—Versuche tuber kiinstliche Theilbarkeit von Beroé ovatus. Arch. f. mikr. Anat., Bd. 17, pp. 213-240. ENGELMANN, |. W., ’68.—Ueber die Flimmerbewegung. Jena Zeitschr., Bd. 4, pp- 321-478. 79.—Physiologie der Protoplasma- und Flimmerbewegung. Hermann, L., Handbuch der Physiologie, Bd. 1, Theil 1, pp. 343-408. °87.— Ueber die Function der Otolithen. Zool. Anz., Jahrg. 10, No. 258, PP- 439-444- GRUETZNER, P., ’82.—Zur Physiologie des Flimmerepithels. Physiol. Studien, ppsi—32. Kraft, H., ’90.—Zur Physiologie des Flimmerepithels bei Wirbelthieren. Arch. f. ges. Physiol., Bd. 47, Hefte 4, 5, pp. 196-235. KRUKENBERG, C. F. W., So=Der Schlag der Schwingplaettchen bei Beroé ovatus. Vergleichend-physiologische Studien zu Tunis, Mentone und Palermo, Abt. 3, pp. 1-23. ParKeER, G. H., ’05.—The Reversal of Ciliary Movements in Metazoans. Amer. Jour. Physiol., vol. 12, No. 1, pp. 1-16. Putter, A., ’03.—Die Flimmerbewegung. Ergeb. Physiol., Jahrg. 2, Abt. 2, pp. I-I02. Samassa, P., ’92.—Zur Histologie der Ctenophoren. Arch. f. mikr. Anat., Bd. 40, ppals7—243, Lat. 6-12. Verworn, M., ’90.—Studien zur Physiologie der Flimmerbewegung. Arch. f ges. Physiol., Bd. 48, Hefte 3, 4, pp. 149-180. ’91.—Gleichgewicht und Otolithenorgan. Arch. f. ges. Physiol., Bd. 50, Hefte 9, 10, pp. 423-472. ’95.—Allgemeine Physiologie. Jena, 8vo, xii+, 584 pp. Cie GENERAL THECKRY OF ADAPTATION AND SELECTION. BY HENRY EDWARD CRAMPTON, Pu.D. It is the purpose of the present brief article to state in general and non-mathematical form some of the results of statistical and experimental studies, dealing with lepidoptera, that have been in progress for nearly six years, and in the second place to develop on the foundation of these results a generalized conception of natural selection and its actual basis during the segregation of the fit or adapted and the unfit or unadapted individuals of a species. I am led to offer this statement in this form and at this time because many months must elapse before the statistical results of recent studies may be put in final form for publication, though the more general conclusions to be drawn from them may be scared in simple | terms; and it is my desire, furthermore, to present the theory so that it may bear the scrutiny of the numerous investi- gators now at work upon the problems of variation and selection, in order that its validity may be tested in connection with different kinds of biological material. Ina word, this paper is an outline of a fuller discussion that must be deferred until the complete statistical results of my own investigations may be brought into relation with those of other inv estigators of the problem of natural selection and of its actual basis. I: In 1899 a brief preliminary investigation was begun upon the pupz of a saturnid moth, Philosamia cynthia, in order to ascertain whether there was a definite relation between the variability of various pupal structures and the elimination that took place after the larve spun their cocoons in the fall of the year and pupated. The same question was also examined with reference to the reduction in numbers that occurred when the pupz emerged in 426 Henry Edward Crampton. the following spring, when many individuals proved to be unable to form perfect moths, and thus offered a comparison with the pupz that metamorphosed successfully and perfectly. The material possessed a peculiar interest for the reason that the pupa does not “use’’ many of its structures at all during i its long period of quiescence but remains practically inactive until the time for emergence approaches. To state the results of this investigation in richest form, it was found in many instances that the pupz which died before metamorphosis were of somewhat different types and were more variable than the surviving individuals; and the same relation between elimination at the time of metamor- phosis and structural variation appeared on the basis of a statis- tical examination of the two groups of pupz distinguished at that time.! Since 1900, the same species has been much more extensively examined to see if similar phenomena were exhibited in other years, and other forms of related moths have also been investi- gated—z. e., cecropia, promethea, ruber, jorulla, etc. The results have been confirmatory, in that in every series where successful pupz have been compared with unsuccessful pupz some of the characters exhibit the stated relation between variation and elimination. Nevertheless, the conclusion that this relation was a general and a final one did not appear to me to be justified, for there were many cases where the reverse was true, where, that 1s, the survivors were of the same type as the others but were more variable, or they were of the same type and variability. In addi- tion, the characters that exhibited “selection,” where it was indi- cated, were of such a nature that they could hardly have served the pupa either advantageously or disadvantageously. The pupa does not “‘use”’ its antenna, and yet the pupz that succeeded in producing perfect moths often possessed antennz that were cer- tainly selectively different from those of the contrasted unsuc- cessful group. It was inferred, therefore, that selection was “indirect,” and that the real basis for selection was not to be sought in the series of individual characters themselves but in the condition of correlation between and among these char- acters. The conclusion of the preliminary study reads as follows: 1See Biometrika, vol. iii. On a General Theory of Adaptation and Selection. 427 ce . the test of fitness or unfitness has reference to the physio- logical and morphological co-ordination or correlation among the constituents of the whole organism, and . . . any relaxation in either series, in a formative sense or otherwise, results in an instability which may culminate in death, and which expresses itself in structural deviation as well as ina higher degree of vari- ability.” It is implied in the foregoing that a distinction may be made between formative correlicion ‘and functional correlation. In a later instalment! the subject is discussed at length, and it is con- tended that the condition of correlation exhibited by the structures of the pupa 1s dependent upon the correlation of the formative factors or agencies which control the manufacture of the pupa by the larva, while the immediate functional elements are concerned scarcely if at all. The case is therefore quite different from that of the moth, where indeed formative factors of the general condi- tion of correlation must be operative, but where the physiological co-ordination of the imaginal structures has a large share in deter- mining the “fitness” Ae the organism. But entirely aside from the Ee intive values to be assigned to these two classes of factors, the point is that the separate “characters”? do not serve directly as adaptive or unadaptive elements of the organism, but they do so only in so far as they exist in close or loose correlation with other structural and functional characteristics. IL. The truth of the conclusion stated was next put to the test of quantitative determination. The co-efficients of correlation were determined, according to the familiar methods, in the case of the characters that had been previously treated individually in their relation to elimination, and the co-efficients of correlation of the two groups were compared, with positive results. While it 1s true that the general condition of correlation, regarded as the basis for elimination, 1 is only imperfectly indicated by the degree of correla- tion between any two characters and that the co- Aiea of multi- ple correlation involving three or more characters would be more reliable as indices of this condition, yet if the principle be true 1Now in press, Biometrika. 428 Henry Edward Crampton. the advantage in favor of the surviving or more successful group of pupz shoud appear more clearly tee the co-eficients of correlation are used than where the comparison with the eliminated group is based upon the types and variabilities of the individual characters concerned. Such is, indeed, the case. While the former group is not invariably the superior in correlation, there is a smaller proportion of negative cases than where the individual characters are taken singly; so that definite confirmation is found for the conclusion reached at first entirely by inference. HE It now becomes the task to develop the principle of “the corre- lative basis for selection’? so as to cover the wider range, over which, I believe, it extends. And I may state at the outset that statistical results have already been obtained proving in part that the wider range is indeed covered, though in the nature of the case, as will appear, complete mathematical demonstration is impos- sible. So far, the general condition of correlation, which it is contended serves as the basis for elimination, has been regarded as deter- mined by the whole series of internal or organismal characters taken together. We may next attempt to bring the series of en- vironmental conditions or influences into the case by taking as an illustration the correlation between a single internal character as representing the whole series of internal characters and a single external character as a representative of that series. The first is “length of pupal period”’ in days, and the second is the “¢trme of the year.”’ Neither of these varieties is simple, it is true. The time of emergence will depend upon a number of things, upon the time of pupation, upon the weight of the whole organism, which, it is found, is indeed correlated with the type character; while in the second place the time of emergence is dependent also upon the time of the year, as increase of temperature hastens metamorphosis. But the point is rather, that when a given series of pupz is kept under natural conditions of temperature, their times of emergence, even when they are members of a single family, will form a curve of error, like that of structural cneeeree such as antenna length, weight, etc. Likewise, the time of the year reckoned as so many days from an arbitrary date such as January first, will form a On a General Theory of Adaptation and Selection. 429 curve; and this external “character’’ too, is compound or at least representative of a series of external influences that affect lepi- doptera, for not only will temperature conditions agree with calen- dar time, but food-supply and many other things will follow in a general way the temporal curve. It 1s clear, I think, that a certain degree of correlation between time of the year and metamorphosis will be adaptive, while a low degree will be unadaptive. Those individuals that mature too early will, even if they find mates, produce eggs and larva that will find poor food-supply, while those that emerge too late, supposing that they too find similar mates, will produce larve that will not have time to become full fed before cold weather will kill them and cut off their food-supply. Facts might be cited, showing still further that those that differ most from the average as regards the time of metamorphosis, vary also in unadaptive directions in internal characters, produc- ing few eggs, possessing imperfect wings, and in other ways. It is needless to amplify the disadvantages that a lack of correla- tion with external influences or conditions would entail. In brief, then, we find that the principle of correlative basis for selection involves not only the whole constitution of the organism itself, but the whole series of graded external influences as well, be these inorganic or organic, homogeneric or heterogeneric. IV. A few words are necessary with regard to the relations of the conception presented above. In the first place, it differs from the general theories hitherto brought forward in having a concrete basis in the results of statistical investigations of correlation and variation, and secondly, in that it places the series of external con- ditions on the same plane as the series of internal conditions, in their relation to the final welfare of the organism, regarding them also as varying according to the familiar laws of error. How far it may be justifiable to extend this principle over the external world, remains for future investigation; but it will be possible, as I belive, to utilize statistical methods in such inves- tigations. Selection is not regarded as in any way originative but only as judicial, so to speak. As the members of any species present themselves at the bar, “selection”’ decides the question of survival 430 Henry Edward Crampton. or destruction on the basis of the condition of correlation that is exhibited. Adaptation receives a precise definition: it means a degree of correlation, capable of numerical determination; and the question as to the utility of any given character becomes sub- ordinate to the question of the effectiveness of any given combina- tion of unit-characters, working in a functional of formative complex. Finally, to possess an evolutionary: value, this conception must be taken in connection with the view that the heritable characters of an organism are congenitally determined. ‘The heritable nature of fluctuations as contrasted with mutations, however, is not a matter that is necessarily brought into court. Barnard College, Columbia University, June 10, 1905. EXPERIMENTAL STUDIES ON THE DEVELOPMENT OF THE EYE IN AMPHIBIA. i ON DHE CORNEA. BY WARREN HARMON LEWIS. Associate Professor of Anatomy, fohns Hopkins University. WituH 2 Prares. INTRODUCTION. With the introduction of the binocular dissecting microscope the possibilities of investigating the subject of correlative embry- ology have been greatly enhanced. With its aid, as I have already pointed out in my paper on the origin of the lane Rana palus- tris, one can make with very delicate instruments exceedingly minute dissections of the living amphibian embryo, and by trans- plantation and extirpation of organs and tissues can gain an insight into the influences of intra-organic environment in develop- ment. Spemann’s experiments on Rana fusca and my own on Rana palustris establish without doubt the correlative character of the origin of the lens in these two species and my unpublished work on Rana sylvatica and Amblystoma punctatum show that in these species likewise the lens is dependent for its origin on the influence exerted by the optic vesicle on the overlying ectoderm. The cornea is likewise a correlative product, but of a quite dif- ferent nature from the lens as will appear in the following pages. While the lens is apparently dependent upon specific influences from the optic cup, the cornea or rather corneal changes of the ectoderm may be brought about by such different structures as the lens alone or of the optic cup alone. Spemann! concludes that the clearing of the corneal ectoderm is dependent in Rana fusca on the presence beneath the skin of the eye with its lens. In ay own experiments on Rana palustris? it was noted that in 1 Verhand. d. Anat. Gesel., 1901. 2 Am. Jour. of Anat., vol. 111, Fig. 9, p. 512. 432 Warren Harmon Lewts. this species likewise corneal clearing of the ectoderm fails when the eye is wanting. In this paper it will be clearly shown, how- ever, that Sel clearing of the ectoderm in Amblystoma will occur over a naked lens or over the optic cup without the lens, rovided the lens or cup are close to the overlying ectoderm. If this be true for Amblystoma it probably holds also for other amphibia and consequently Spemann’s conclusion for Rana fusca should be modified to this extent. In a preliminary communication before the Association of American Anatomists, at Philadelphia, 1903,' concerning my experimental studies on the development of the eye in amphibia, the following conclusions were given for the cornea:» | (ye cornea fails to develop when the optic vesicle is entirely removed. (2) Over the regenerated eye with lens a cornea develops, nor- mally, except for size, which is small to correspond to the small regenerated eye. (3) If the optic cup 1s torn out after the lens fae separated from the skin, a small area of clear epithelium will develop immediately over the undisturbed lens. (4) Such clear- ing for the cornea will also dev elop over an optic cup, from which thie lens has been extracted, but not in all cases. These conclusions were based mainly upon experiments on Amblystoma punctatum and Rana palustris. More recent experi- ments on Rana sylvatica show that in this species also the cornea fails to develop when the eye is wanting. The present paper 1s concerned more especially with the conditions in Amblystoma, which is a more favorable form for the study of corneal forma- tion than Rana. The experiments enable me to consider only the early stages of corneal formation, namely: (1) The thinning of the ectoderm of the corneal area; (2) the clearing of this ectoderm and loss of its pigment; (3), the formation of the endo- thelial layer of the cornea, which is in reality the anterior wall of the anterior chamber of the eye. In addition to the above conclusions some of my more recent experiments show that the cornea will form from ectoderm other than that which normally gives rise to the cornea. This spring I have repeated most of the experiments on Ambly- stoma and find that it is easy to confirm all of the conclusions stated above. Some new experiments show that even after the 1 Am. Jour. of Anat., vol. ii. Experimental Studies on the Development of the Eye. 433 cornea is formed it will disappear almost completely if the optic cup and lens are entirely removed, without injury to the over- lying cornea. The experiments in this paper with the sole exception of MD, have been selected each from a series of several similar ones. METHODS. Embryos of various ages were operated upon under the binocu- lar microscope. Either ordinary tap water or a 0.2 per cent salt solution was used. The older embryos were first anes- thetized with acetone-chloroform. ‘The embry os were held with a pair of fine forceps, and the incisions were made with a very small pair of scissors, the points of which were ground with great care. Ordinary needles complete the instruments needed. “The manipulation requires considerable practice, but one soon finds experiments possible, which at the beginning seemed beyond such methods. ‘The method possesses great advantages over the use of the hat needle or electric cautery. A new and very wide feld of work is opened by means of this dissection method and it will undoubtedly throw much light upon dev elopmental processes. The embryos were Killed: in Zenker’s fluid, cut into serial sec- tions 5 # or 10 » in thickness and stained in hematoxylin and Congo red. The operations were all performed on the right side. The figures are all from photo-micrographs of transverse sections through the region of the right eye. EXPERIMENTS. A. Non-development of the Cornea after Total Extirpation of the Eye. The numerous experiments on Rana palustris where the optic vesicle was completely removed at an early stage before lens for- mation and consequently long before there are any indications of corneal formation, have all failed to show corneal changes in the ectoderm which under normal conditions would have fennel corneas. Even days after the operation and long after the cornea on the normal side of the head was well dev eloped all traces of 434 Warren Harmon Lews. corneal clearing of the ectoderm were wanting in the skin cover- ing the region from which the eye had been taken. I have sieeade given an experiment of this nature in my article on the ori- gin of the lens in Rana palustris.!. In this embryo a skin-flap was turned forward from over the optic vesicle, the latter cut away and the flap returned to its original position without injury to the corneal area of the ectoderm. athe embryo was killed eleven days after the operation but showed no signs of corneal changes on the side operated on, while upon the normal side there was a well- developed cornea. In other experiments similarly performed only a portion of the optic vesicle was cut away and over the regenerated eye corneal formation took place provided the regenerated eye was of sufficient size to come into contact with the skin. It is evident then that the lack of corneal formation after complete extirpation of the eye was not due to the turning forward of the skin-flap. If a skin- flap 1 is turned forward from over the eye and then replaced without injury to either it or the eye, perfectly normal develop- ment of the eye, lens and cornea will ensue. A deeply situated regenerated eye separated from the ectoderm by mesenchyme will not cause corneal formation. In similar experiments on Rana sylvatica like results follow total or partial extirpation of the eye. In Amblystoma punctatum the corneal changes are much easier to follow than in Rana as the ordinary ectoderm is of con- siderable thickness and the contrast between it and the cornea much greater than in the frog. In Ambly stoma as in frog larve the early total extirpation of the optic vesicle, before the: period of lens formation, results in the failure of corneal dev elopment. For example, if in an embryo at this stage a flap of skin is carefully turned forward from over the eye, the optic vesicle completely cut away, and the uninjured skin-flap returned to its original position, it readily heals in place but no traces of corneal formation are to be observed even sixteen days after the operation. Fig. 1 gives an accurate idea of the con- ditions in such experiments. The lens also is entirely wanting while on the normal unoperated side optic cup, lens and cornea are present. [he endothelial layer of the cornea likewise fails to dev aoe without the presence of the optic cup. ‘Am. Jour. of Anat., vol. iu, p. 512, Fig. 9. Experimental Studies on the Development of the Eye. 435 If at a much later stage, namely, shortly after separation of the lens from the ectoderm, both optic cup and lens are taken out in a manner similar to that just described, corneal changes fail to develop, even if the embryos are allowed to live from twelve to eighteen days after the operation. (See Fig. 1, from Experiment X1V yn) The end result as regards non-development of the cornea 1s the same in each embryo, whether the eye is taken out before the lens begins to form or shortly after its separation from the skin. In Amblystoma, as in Rana, however, corneal formation occurs after partial extirpation of the eye, whether this 1s done before the lens has formed or shortly after its separation from the ecto- derm, provided the regenerated eye comes into contact with the ectoderm. The mere lifting of the skin-flap here as in Rana does not interfere with corneal formation, so that the lack of corneal development after complete extirpation of the eye is not due to the lifting of the skin-flap but must be in some manner associated with the absence of the eye. It is evident that the cornea 1s not a self-differentiating structure. |b Rudimentary Corneal Area ajter Late Extir pation of the Eye. If in Amblystoma the optic cup and lens are taken out some- time after the separation of the lens from the ectoderm but before corneal changes are visible on the surface of the embryo a small clear corneal area will develop in the region where the large normal cornea would have formed. An examination of a normal embryo of the same stage at which the operation was performed ° shows that corneal changes have begun and consist in a slight thinning of the ectoderm over the eye. The endothelial layer is also in the process of formation. ‘The operations consist in making an incision about the caudal side of the eye and then carefully turning forward the skin-flap from over it without injury to the corneal region. [he eye and lens were then cut out and the skin-flap turned back into place where it readily healed. A few days after the operation a small clear corneal area appears in the ectoderm of the corneal region. An examination of the sections shows that this clear area differs from the ordinary ectoderm surrounding it in that it is much thinner, the pigment is wanting and the aodeemal cells have lost or not acquired the 436 Warren Harmon Lewis. usual vacuolization. It is in general like the normal corneal ectoderm except in being somewhat thicker. “The rudimentary corneal area, however, never seems to become much larger than that pictured in Fig. 2, even twenty days after the operation. This experiment (Mn,) is only one of several in which almost exactly similar results were obtained as regards the development of the small corneal area. In most of these experiments the rudimentary cornea is at the bottom of a depression in the ectoderm as in Experiment Mn, (Fig. 2). Such depressed areas occur, however, without corneal changes. (See Fig. 3, Experiment ME,.) In this experiment (ME,) the optic cup was removed some time before the separa- tion of the lens from the ectoderm. ‘The lens has been pinched off from the skin and has become separated from the latter by mesenchyme. The lens is, however, very much smaller than the normal one on the other side of the head. [he embryo was killed 16 days after the operation and there is no trace of corneal formation even at the bottom of the depressed ectodermal area. The depression of the ectoderm then can scarcely be looked upon as a factor in its clearing. In explanation of these rudimentary corneal areas it may be that the influences causing corneal formation had, at the time of the removal of the eye, not been acting long enough on the ecto- derm to enable the clearing process to go on independently and form the normal sized cornea. A longer continued influence of the eye is evidently necessary for normal corneal formation and as will be shown later on it is necessary for the eye to be present even after the cornea is well formed if the latter is to continue its existence. It may be that the cornea can never maintain itself independently of the eye; however, farther experimentation 1s necessary to determine this point. Here it again becomes apparent that the cornea of normal size is not a self-differentiating structure. C. The Size of the Cornea ts Dependent upon the Size of the Eye. In the various experiments where portions of the eye have been cut away the regenerated eyes even thirty days after the operation fail to reach rhe size of the one on the normal side, unless the amount cut away is very small. The size of the regenerated eye 1s somewhat in proportion to the amount of eye stuff left Experimental Studies on the Development of the Eye. 437 behind so that the new eye is to be viewed more as a re-formation than a regeneration. Such a regenerated eye coming into con- tact with the epidermis causes corneal clearing, the area of which varies in size with that of the eye. If but a earl portion of the optic vesicle is cut away the regenerated eye will be of nearly normal size, with a cornea in all respects normal except of slightly less diameter to correspond to the smaller diameter of the eye. If more of the optic vesicle 1s cut away a still smaller regenerated eye and cornea will result. In Experiment VII,,, (Fig. 4), the somewhat irregular optic cup is about three-quarters of the diam- eter of the normal one and the cornea about two-thirds of the diameter of the normal cornea. In other respects the cornea is like the normal one. The lens which is nearly as large as the normal one is still adherent to the retina and fills the entire posterior cham- ber of the eye. This is not an uncommon condition of the lens in the regenerated eyes even as late as eighteen days after the operation. If a still greater portion of the optic cup 1s cut away a regenerated eye less than one-half the diameter of the normal one may develop with a correspondingly small cornea. In Experiment WAITERS (Fig. 5), the small irregular optic cup contains a large lens w hich fills completely the cup cavity. The endothelial layer stretches over the lens and is for the most part in contact with it. The cornea although but one-half the normal diameter is in other respects like the normal one on the opposite side. Even more of the optic vesicle may be cut away than in the preceding experiments, yet if the small regenerated eye remains in contact with the skin a very small corneal area will develop. In the above experiments an incision was made around the caudal part of the eye and the skin-flap with lens attached turned forward. Varying amounts of the optic cup were cut away and the skin-flap with the lens attached turned back into the original position where it readily healed. In another series of experiments both lens and optic cup were turned forward with the skin-flap and then varying amounts of the deep portion of the eye cut away. The skin- flap with the lens and remainder of the optic cup were then turned back into position. The re-formed eyes vary in size according to the amount left attached to the skin-flap, and the cornea in each embryo also varies in size with the size of the re-formed eye. 438 Warren Harmon Lewis. In still another series of experiments a portion of the optic cup together with the lens and all of the epidermis over the eye were cut away. In these experiments new epidermis soon covers the remainder of the eye and from it the cornea develops, and here, too, the size of the cornea varies with the size of the eye. ‘This formation of corneal from strange ectoderm will be treated more fully in another section. The area of the corneal clearing of the epidermis over the optic cup alone or over the naked lens is likewise in proportion to the size of the area of contact of these organs with the skin. A large optic cup may be so situated that only one small corner of it is superficial and in contact with the epidermis. The size of the corneal clearing is in proportion to this area of contact and not in proportion to the size of the eye. D. Corneal Formation with the O ptic Cup Alone and Without the Lens. In order to analyze more completely the influence of the eye on corneal formation I have in the following series of experiments excluded the possible influence of the lens and find that the early stages of corneal formation will develop without the presence of a lens or without a lens ever having formed from the skin. This last point is illustrated by a single fortunate experiment (MD,) on Amblystoma. A skin-flap was turned forward from over the eye in the usual manner at a time when in normal embryos of the same stage the skin is just beginning to show signs of thick- ening for lens formation. A portion of the shallow optic cup was cut out and the skin-flap replaced. ‘The sections show that for some reason the regenerated eye failed to cause lens forma- tion, but nevertheless corneal formation is present (Fig. 6). The optic cup is contracted and the cavity much reduced in size, the pupil is small and the endothelial layer of the anterior chamber reduced in area. ‘The transparent corneal area is smaller than normal and slightly thicker. If at a somewhat later stage after the lens has formed and separated from the skin, but before there is any corneal clearing, a skin-flap is turned forward and the lens with part of the cup cut out, a small cornea will form over the small optic cup provided the latter is close under the skin. (See Fig. 7, from Experiment Experimental Studies on the Development of the Eye. 439 MF,.) The contracted cup with small cavity and pupil pre- sents a similar appearance to the condition seen in Experiment MD,. The extent of the endothelium and of the cornea cor- respond in size with the cup. The cornea 1s somewhat thicker than the normal one on the opposite side of the head, but other- wise similar to It. I have numerous other similar experiments giving like results. In these experiments the optic cup or its endothelial membrane lie close to the corneal clearing, which corresponds in size with the area of contact. If, however, the optic cup and its endo- thelial membrane lie somewhat deeply buried and separated from the ectoderm by mesenchyme the corneal clearing fails to develop. The formation of the cornea is then neither dependent upon the formation of a lens nor upon the presence of the lens. E. Small Corneal Clearing over the Superficial Naked Lens. If in Amblystoma the optic cup is taken out about the time of, or shortly after, the separation of the lens from the skin and the lens left in position close against the skin, a small clear corneal area will develop immediately over the naked lens. In such experiments there was at the time of the operation no trace of cornea and if both optic cup and lens are taken out the small area of corneal clearing does not appear. An incision was made about the caudal two-thirds of the eye and the whole eye and skin-flap turned forward together. The optic cup was then carefully removed, leaving the lens im situ and the skin-flap with the lens attached turned back into its normal position. At about the time when corneal clearing appears on the normal side the skin over the lens clears also, but is limited to the area immediately over the lens. As both epidermis and lens become perfectly transparent one can look down into the depths of the head in the living animal. The lens in most of the embryos 1s considerably aaailler than the one on the normal side and often shows degeneration changes. The endothelial layer does not develop about the naked lens. The corneal area does not seem to spread beyond the extent of the contact area of the lens, nor 440 Warren Harmon Lewis. does the skin become as thin as that of the normal cornea. The pigment disappears and the large vacuolated cells which are present in the surrounding ectoderm are absent. (See Fig. 8, Experiment MG,.) If, however, the lens is disturbed by the operation so that mesenchyme grows in between it and the skin, the corneal changes do not occur. (See Fig. 9, from Experiment Ma,, and Fig. 3, from Experiment ME.) If after the optic cup is taken out the lens is transplanted a short distance from the normal position, the mesenchyme often separates the lens from ectoderm and in such experiments the corneal clearing likewise fails to develop, and a condition similar to that seen in Fig. g is present. F. Corneal Formation from Strange Ectoderm. If the ectoderm covering the optic cup and lens is completely torn away at a stage shortly after the separation of the lens from the ectoderm but before there are any visible corneal changes, the wound thus formed will heal by the ingrowth of ectoderm from the sides of the denuded area. In many of the experiments there was more or less disintegration of the optic cup before the wound healed. In some almost the entire optic vesicle disappears; in others but little of it is lost. In the experiments where the optic vesicle remains of sufficient size to come into contact with the new ectoderm true corneal formation follows, the size of the cornea varying with the size of the eye. In Experiment De eo ( ig. 10), there was very little loss of optic vesicle substance and the new ectoderm soon covered the entire denuded area. The cornea with its endothelial membrane is apparently normal except in size being of less extent than the normal in proportion as the eye is smaller than normal. The new cornea was not well developed until about four days after the normal one on the left side had become perfectly clear. The difference in time in the formation of the corneas on the normal and operated sides may be even much greater than this, as when there is considerable disintegration of the eye, after the skin is torn off from over it, healing may be delayed a day or so and the eye much reduced in size. In some of these experiments the corneal clearing may be delayed from four to eight days after the one on the normal side is Experimental Studies on the Development of the Eye. 441 perfectly clear. And more than eight days may elapse before all of the pigment cells are gone. In another experiment performed in a similar way (XVop5q) there was considerable disintegration of the eye before the new skin completely covered over the denuded area. ‘The eye is about one-half the diameter of the normal one and the clear corneal area even smaller in size. [he various layers of the retina are irregular and the lens also. ‘The latter fills the pos- terior chamber of the eye and has a process projecting into the pupil. (Fig. 11.) In these experiments as in those in which a portion of the optic cup was cut away without injury to the overlying skin, the size of the corneal area is in direct proportion to the area of contact of the underlying eye. In these experiments the skin that grows over the eye is at first opaque and shows no signs of clearing. Pigment cells are scattered through it as in the ordinary epidermis. ‘This con- dition often remains until long after the cornea on the normal side is well formed and clear. The clearing of the new epidermis which has grown over the eye is usually a slow process, and a few pigment cells are especially prone to remain even for a long time after the skin has cleared. If not only the skin from over the eye is cut away but with it 1s taken the lens and even the lens and part of the optic cup the adjoining skin will slowly cover the optic cup and after consider- able delay a cornea will form over this optic cup, the lens being absent. The size of the cornea varies with the size of the re-formed eye. Here, too, the corneal formation is much retarded and only appears days after the one on the normal side is well formed. If too much of the optic cup is taken away with the lens, what remains may be so deeply buried that it does not come into con- tact with the skin. In such embryos the cornea does not develop. (See Fig. 12, from Experiment DOV TT) If skin, lens, and optic cup are completely removed, new epidermis will cover over the large wound but corneal changes fail to appear even four weeks after the cornea has developed on the normal side. So we can scarcely look upon the cornea in the above experiments as a product of regeneration, but must con- sider it as a new product from skin other than that which normally gives rise to a cornea. 442 Warren Harmon Lewis. G. Degeneration of the Cornea after Extirpation of the Optic Cup and Lens. If after the cornea is well formed an incision is made dorsal to the eye and the optic cup with the lens taken out, care being taken not to injure the cornea, the large cornea will gradually disappear. At first instead of forming a bulge on the surface of the head there results a depressed area owing to the absence of the optic cup. At the bottom of this area is the corneal clearing. This corneal area gradually contracts and pigment cells invade it from the adjoining skin, and in some of the experiments there is scarcely a trace of the cornea left 30 days after the operation. CONCLUSION. It seems very probable that the optic vesicle brings about lens formation through a specific influence. The cornea, however, iniesO. fal AS Ats early stages are concerned, namely, the thinning and clearing of the skin and loss of pigment can hardly be ascribed to a specific influence. “That the mechanical pressure of the eye, or cup or lens may in some way be accountable for the corneal changes is a possibility. I am more inclined to the view, however, that the changes are due to another and quite different reason. ‘The contact of either the eye or cup or lens with the epidermal cells must of necessity alter the environment of the overlying cells as regards their relation to the mesenchyme. It is possible that this exclusion from contact with the mesenchyme cells may be responsible for changes in the metabolism of the epidermal cells and cause them thereby to alter their mode of development, such alteration leading to corneal changes. That such apparently slight alterations in environment are responsible for other important changes in the history of embryonic ecto- dermal cells I think quite probable. From some experiments already completed it seems highly probable that the central nervous system is in part at leat dependent for its origin and differentiation on the difference of environment of cells whee at one time possessed the possibilities of producing either ordinary epidermal cells or of producing nerve cells. as uN Ww Experimental Studies on the Development of the Eye. SUMMARY (AMB LYSTOMA 5) 1. A normal cornea will not develop without the eye. 2. The size of the cornea varies with the size of the eye, with the area of contact between it and the skin. 3. Contact between eye and skin is a necessary factor, as an eye separated from the skin by mesenchyme will not cause corneal formation. 4. [he optic cup alone (without the lens) can cause corneal formation. 5. The lens alone (without the optic cup) can cause corneal formation, provided, as is the case with the optic cup, it is in contact with the skin. 6. The size of the corneal area over the optic cup or lens 1s dependent upon the area of contact between these structures and the ectoderm. 7. It is not necessary that the lens should be first formed from the skin in order to have corneal formation. 8. The cornea will develop from strange epidermis other than that which normally forms the cornea. g. After the cornea is once formed it degenerates and disap- pears after extirpation of the rest of the eye. 10. [The cornea is_ neither predetermined nor self-differ- entiating. 11. The cornea is dependent upon the correlation between the eye and the overlying ectoderm for its origin. 444 Warren Harmon Lewis. EXPLANATION OF PLATES. Pirate I. Fig. 1. Experiment XIVas. Right optic cup and lens taken out shortly after separation of the latter from the skin. Embryo killed 11 days after the operation. Figure from section through right eye region. A bit of the optic cup with nerve is deeply buried near optic foramen. No traces of corneal formation. The normal left side has a well developed cornea. X 80 diameters. Fig. 2. Experiment Mn. Right optic cup and lens taken out at a somewhat later period than the above, but before there were visible corneal changes. Embryo killed 12 days after the operation. Figure from section through the right eye region. A small depressed area of corneal clearing is seen, which is scarcely ;/; the diameter of the normal cornea on the left side. 80 diameters. Fig. 3. Experiment ME2. Right optic cup removed shortly before the separation of the lens from the skin. Embryo killed 16 days after the operation. Figure from section through the right eye region. The lens, much smaller than normal, is separated from the ectoderm by mesenchyme. There are no corneal changes in the epidermis. On the left side the cornea is large and well formed. X 80 diameters. Fig. 4. Experiment VII3«:. A portion of the right optic cup cut away before corneal formation. Embryo killed 9 days after the operation. Figure shows irregular eye smaller than normal with cor- respondingly small cornea. 80 diameters. Fig. 5. Experiment VII;. Operation as above except that more of the optic vesicle was cut away. Figure shows small eye and cornea. % 80 diameters. Fig. 6. Experiment MD,. Portion of eye cut away before lens formation. Embryo killed 12 days after the operation. Figure shows absence of lens, small eye with small cavity and pupil, and corneal formation over the optic cup. X 80 diameters. meee PVAIE I: W. H. Lewis. EXPERIMENTAL STUDIES ON THE DEVELOPMENT OF THE EYE. Tue Journar or ExperiMEentAL Zo6xoGy, vol. ii. bio we 446 Warren Harmon Lewis. Pirate II. Fig. 7. Experiment MF:. Lens and part of optic cup removed shortly after the separation of the lens from the skin. Embryo killed 14 days after the operation. Figure from section through right eye region, shows the regular reformed optic cup with small cavity and pupil, and overlying cornea with its endothelium. 80 diameters. Fig. 8. Experiment MG,4. Right optic cup removed shortly after separation of lens from skin, but before corneal formation. Embryo killed 13 days after the operation. Figure from section through the right eye region, shows a lens smaller than normal pressed close against the skin, where there is distinct corneal clearing, but no endothelial formation. 80 diameters. Fig. 9. Experiment Ma;. Operation as above except that the lens was disturbed. Embryo killed 13 days after the operation. Figure shows the small lens separated from the skin by mesenchyme. There are no corneal changes in this region nor endothelial formation. A small mass of eye-cells lies medial to the lens. X 80 diameters. Fig. 10. Experiment XV3e2. Epidermis from over the entire right eye cut off before traces of corneal formation present. Embryo killed 11 days after the operation. Figure from a transverse section of the right eye region shows how new skin has completely covered the eye and become trans- formed into cornea. X 80 diameters. Fig. 11. Experiment XV3¢¢. All of the skin over the eye and a portion of the optic cup removed. Embryo killed 11 days after the operation. Figure shows small eye with irregular lens and corneal formation from the new skin. % 80 diameters. Fig. 12. Experiment XVIIzig. All of the skin over the eye, the lens and part of the optic cup cut away before corneal formation. Embryo killed 11 days after the operation. Figure shows deep optic cup without pupil or cavity separated from skin by mesenchyme. There are no traces of corneal formation in the new ectoderm. X 80 diameters. PLAGE SIT. W.H. Lewis. EXPERIMENTAL STUDIES ON THE DEVELOPMENT OF THE EYE. Tue Journar or Experimenta Zoo.oey, vol. ii. | | ] MODIFIABILITY IN BEHAVIOR. I. BEHAVIOR OF SEA ANEMONES. BY H. S. JENNINGS. A thorough study of the modifiability of reactions to external stimuli in lower organisms seems at present one of the great desiderata in the study of animal behavior. Recent work has been devoted largely to the study of sharply defined forms of reaction and to the discovery of conditions under which these forms appear in the typical way. As a result there 1s a wide- spread impression that the behavior of lower organisms 1s com- posed of invariable reflexes, occurring always in the same way under the same external circumstances. ‘This is far from the truth and leads, as it seems to the writer, to a fundamentally false conception of the nature of animal behavior. Inner states and changes are fully as important in determining behavior as are external stimuli, modifying fundamentally the reactions which the latter produce. ‘The present studies are devoted to an analysis of some of these modifying factors; in other words to some of the inner factors in behavior. The study of the behavior of sea anemones herewith presented was made possible by a stay at the Carnegie Research Laboratory at the Tortugas. I am under great obligations to the Carnegie Institution and to the director of the laboratory, Dr. A. G. Maye ae for opportunity to carry on the work, and for supplying every facility that could assist it. The Tortugas laboratory furnishes an ideal situation for carrying on such investigations. An indef- nite number of species of sea anemones and corals can be procured at a few moments notice, and they live as well in the laboratory as in the sea, since the water becomes cooler instead of warmer when brought into the house. 448 1. S. “‘fennings. I. CHANGES IN BEHAVIOR DUE TO VARYING STATES OF METABOLISM. Nagel (’92) and Parker (’96) have shown that the food reaction of actinians toward weak stimuli becomes changed on repetition of the stimulation. A Metridium or an Adamsia at first readily takes filter paper soaked in dilute juice of crab meat. But after this has been fed several times in alternation with pieces of meat, the reaction to the filter paper becomes slower, and finally ceases, while the meat is taken as readily as before. Torrey (04) shows that in Sagartia the state of hunger or satiety determines largely the reaction to small solid bodies. A very hungry Sagartia readily swallows inert bodies, such as filter paper and sand grains, while a fairly well fed one rejects these, though it takes meat. Has this effect of hunger and satiety any connection with the changes observed by Nagel and Parker, or are these of a different character? What relation have they to the changes due to experience in higher animals? The whole problem of the changes induced in behavior by changing metabolic states is one of the greatest importance for an understanding of the adjust- ment or regulation produced in behavior. I have attempted to study this matter carefully in a number of sea anemones, and to distinguish modifications due to this cause from those which result from other factors. Stoichactis Helianthus. This large sea anemone has often a disk 10 to 15 cm. in diameter. This is covered closely with short tentacles of uniform size, about 8 mm. in length.t. Stoichactis is voracious; it is usually when captured ready to take large quantities of crushed crab appen- dages. ‘To three specimens I fed piece by piece nearly all of three good sized ghost crabs (Ocypode). The food reaction depends on contact with the meat itself—that is, on chemical stimuli in combination with contact. Hard parts of the crab, or other indifferent objects, are usually not taken, though in rare cases even filter paper is swallowed. Food is taken in the following wav: If a piece of crab’s leg, with some of the flesh exposed, ss placed on the disk of a hungry ‘For photographs of the disk of Stoichactis, see Duerden, 1902, Pl. 1. Modipfiability in Behavior. 449 specimen, the tentacles immediately surrounding it (including many not in contact with it) begin suddenly to wave back and forth. After an instant this aoueilly ceases, and all is absolutely quiet for a few seconds. Then the movement begins again. All the tentacles that in their waving motion come in contact with the food, bend over against it and shrink, in such a way as to hold it down against the disk. Now that portion of the disk bearing the food begins to sink inward, by a folding of the surface. The eee eh may be 4 or 5 cm. distant, begins to open, and the walls of the esophagus protrude from the mack as large bladdery lobes. The region between the mouth and the food body begins to contract, the tentacles borne here collapsing and almost completely effacing themselves. By this contraction the mouth and food approach each other, the intervening region disappearing. Meanwhile other parts of the disk swell and their tentacles become plump and enlarged; this appears to be a secondary phenomenon due to the squeezing of the internal fluid from the contracted region to other parts. ‘The esophageal lobes increase in size, becoming 2 to 4 cm. long, and half as thick; they extend toward the food, finally reaching it. By the contractions and expansions already mentioned the mouth. may be moved from the center of a disk 10 cm. in diameter to within I cm. of the edge. By this time mouth and food may be hidden beneath the surface of the contracted disk, though in other cases they lie on the surface in plain view. Now the esophageal lobes extend over and around the food, while the tentacles progressively withdraw from it until the food body is lying on the contracted portion of the disk, completely covered by the esophageal lobes. Next that part of the disk beneath the food withdraws, involving an enlargement and further displacement of the mouth, till there is nothing beneath the food body, and it is pressed by the esophageal lobes into the internal cavity. The whole reaction is thus very complex. Twenty or more pieces of crab, including entire large append- ages, may thus be successively taken, till the body of the anemone has become a mere stretched sack full of crab appendages. But in the later reactions of a series the process of food-taking becomes much slower, the animal seeming to become eradually satiated. The food may be taken by the tentacles and held for a long time before it is finally moved to the mouth. In other cases the ten- 450 H. S. ‘Fennings. tacles do not react for some minutes, the food lying on the disk undisturbed, until finally it is slowly taken. Sometimes there is an interesting combination of the positive food reaction and the negative reaction (to be described later). ‘The food is taken by the tentacles and carried very slowly to the mouth, in the way above described, while the mouth opens and the esophageal lobes are protruded. But when the food body reaches the lobes, or sometimes before, the process stops. ‘The food is released by the tentacles, and is finally carried away and rejected, in the way to be described. Finally, when the animal seems fully satiated, the piece of crab meat may be rejected as soon as it comes in contact with the disk. But after one or more pieces have been rejected one may sometimes see another piece accepted. The internal state is in a condition of most unstable equilibrium, and may easily incline toward the positive or the negative reaction. Thus it is clear that in Stoichactis the reaction to a given stimulus is by no means a set, invariable property of the organism, but depends on the state of the internal processes. “To the same stimulus we may get a quick positive reaction or a quick negative reaction; a slow and deferred positive reaction or a combination of the positive and negative reactions. Peculiar effects are observed when several pieces of meat are placed at the same time on different parts of the disk. If the animal is hungry all are carried to the mouth; the entire disk folds inward and the pieces are swallowed simultaneously or successively. I have seen six pieces, placed as far apart on the large disk as possible, thus ingested. When the animal is less hungry the results are different. In some cases, when two pieces of meat are placed on the disk, one is swallowed while the other is rejected. If the rejected piece is again placed on the disk after the first piece has been disposed of, it will sometimes be swallowed. Adding new pieces while swallowing is in progress often pro- duces interference. “Thus, in one case two pieces of meat, a and b, were placed near opposite edges of the disk. Both began to approach the mouth in the usual food reaction. Now two new pieces, c and d, were placed near the edge midway between a and 8. Thereupon the reaction to a and 4 ceased, while d was transported to the edge of the disk (about 2 cm.) and dropped off. Now the food reaction was resumed,.a, ) and c traveling toward the mouth. Modifiability in Behavior. 451 Piece d was now replaced on the disk. ‘The reaction to the other pieces was suspended, and d was carried to the mouth. Here it came against the middle of the esophageal lobe that was extend- ing toward a,—in such a way that d could not well be ingested without a rearrangement of the lobes. ‘Thereupon d was again carried away from the mouth and once more dropped over the edge of the disk. ‘The other pieces were now successively swal- lowed. Piece d was readily swallowed when given to another specimen. The Rejecting Reaction.—After Stoichactis has become satiated, it rejects food, as we have seen. ‘The rejecting reaction presents a number of points of much interest. By this same reaction the disk is kept clean when débris falls upon it. If a mass of waste matter of any sort (as a mass of dead plankton or a quantity of sand) is placed on the disk of Stoichactis, measures are set in operation which result, within ten or fifteen minutes, in removing this material and leaving the disk free. ‘The behavior in bring- ing about this result is complex and the operation may be accom- plished in more than one way. The tentacles bearing the débris or the rejected food body collapse, becoming thin and slender, and ly1 ing flat against the disk. At the same time ache disk surface in this region begins to stretch, separating the collapsed tentacles widely. “As a result the waste mass is left on a smooth, exposed surface, the tentacles here having practically disappeared—though under usual conditions they form a close investment almost completely hiding the surface ‘of the disk. ‘Uhus the waste mass is fully exposed to the action of waves or currents, and the slightest disturbance in the water washes it off. Under natural conditions this must usually result in an immediate removal of the débris. If this does not occur at once, often the region on which the débris is resting begins to swell, and becomes a strongly convex, smooth elevation, thus rendering the washing away of the mass still easier. But the process may go much farther. If the débris is not removed in the way just described, new reactions set in. If the mass is nearer one edge of the disk this edge usually begins to sink, while at the same time the tentacles between the edge and the waste object collapse and practically efface themselves. ‘Thus a smooth, sloping surface is produced and the waste mass slides off the disk. If this does not occur at once, after a little time the 452 Jehieye Fennings. region lying behind the mass (between it and the center of the disk) begins to swell, producing a high, rounded elevation, with foniacles: plump and swollen. ‘Che waste mass is now on a steep slope, and is bound soon to slide down and over the edge. Some- times by a continuation of this process the entire disk comes to take a strongly inclined position, with the side bearing the débris below. Often one portion of the edge of the disk after another is lowered in this way, till all the waste matter has been removed. The disk then resumes its horizontal position, with nearly flat or slightly concave surface. Sasneatines the edge bearing the débris cannot be lowered, owing to the fact that it is almost against an elevation in the irregular rock to which the anemone is attached. In this case, after perhaps an attempt to bend the edge downward, the part between the edge and the waste body swells and rises, rolling the mass toward the center, while at the same time the region between it and the center sinks down. ‘The sinking continues till it reaches the opposite edge, so that the mass is rolled across the disk to the opposite side and there dropped off the disk. ‘The process is slow, often taking fifteen minutes to half an hour. The rejecting reaction is characterized by great flexibility and variability. [he debris or refused food sets in operation cer- tain activities; if these do not remove the source of stimulation, other activities are induced until one is successful. Thus in Stoichactis the same stimulus—crab’s meat—may in the same individual produce sometimes the long train of activities resulting in the ingestion of food; in other cases the complicated and variable behavior resulting in rejection, in still others a com- bination of the two. The deciding factor is internal—the con- dition of the metabolic processes. A iptasia. Two species of Aiptasia were studied. One was Aiptasia annulata Les.; the other a smaller and darker species, with shorter tentacles, which I have been unable to identify with certainty. I shall call it Aiptasia No. 2. Both came from the moat surround- ing Fort Jefferson. Rather small specimens, with columns 4 to 10 cm. in length, were used in most of the work. Modipiability in Behavtor. 453 The species of Aiptasia are relatively active and quick-moving anemones. Especially is this true of Aiptasia annulata. If the tip of one of the long tentacles is touched, the whole disk and column shrinks with a sudden quick contraction, reminding one of the rapid contraction of a medusa. To the eye all parts of the body appear to contract at once. Often the disk and column have contracted strongly before the actual contraction wave has made any apparent progress from the tip of the long tentacle to the disk. Certainly in this animal the general contraction does not appear to be due to a spreading of an actual contraction wave from one part of the animal to another, through the actual pulling of one region upon the neighboring one, as it does in Hydra, and according to Torrey (’04), in Sagartia. On the contrary, there seems certainly to exist some rapid method of conduction, suggest- ing nervous action. In Aiptasia annulata the use of India ale indicates the presence of cilia driving a current away from the mouth and toward the tip of the tentacles, as in Metridium. Aiptasia annulata usually takes crab meat or filter paper soaked in the juices of such meat, but refuses neutral bodies, such as plain filter paper or sand. Aiptasia No. 2, on the other hand, 1s usually prepared to swallow readily balls of plain filter paper and other small neutral bodies, as well as crab meat. ‘This furnishes opportunity for some interesting comparative experiments. Food is taken in the following way: If a small object comes in contact with a tentacle it adheres to the surface, and the tentacle contracts strongly, the whole animal usually contracting at the same time. [hen the tentacle bends over and places the food with considerable precision on the mouth. ‘The tentacles near by likewise bend over and are applied to the food body, holding it down against the mouth. ‘This happens even when the body is quite neutral, as plain filter paper, so that the bending of the neighboring tentacles is clearly due to some influence transmitted from the one tentacle in contact with the body. The mouth now opens, the lips protruding a little and seizing the food, while the ten- tacles may release it and bend away. But sometimes the tentacles follow the food into the mouth and their tips remain enclosed for some time. The actual swallowing of the food is mainly due to the activities of the lips and esophagus; it may occur without any intervention of the tentacles, when the food is placed directly on 454 H. S. ‘fennings. the mouth. A piece of meat or filter paper may be completely enclosed by either species within ten seconds of the time it comes in contact with a tentacle. With these two species of Aiptasia the experiments of Nagel and Parker, mentioned on page 448, were repeated and varied, with somewhat peculiar results. Pieces of crab meat and of filter paper (plain or soaked in juice of crab meat) were given alternately to the individual under experimentation. In Metri- dium and Adamsia, as we have noted, the animal soon comes to reject the filter paper, while still accepting the meat. In Aiptasia annulata a typical experiment is as follows: The animal is fed alternately filter paper soaked in crab juice and crab meat. Both are taken readily till four pieces of each have been ingested. At the fifth piece of paper—the ninth piece of the whole series—the animal balks and rejects it. But it likewise rejects the immediately following fifth piece of meat! It has evidently lost its hunger, and refuses to take any thing. ‘This is the usual result with Aiptasia annulata. In Aiptasia No. 2 plain filter paper (not soaked in crab juice) was given alternately with pieces of crab meat. In a typical experiment six pieces of filter paper and six of meat were taken in regular alternation. But the seventh piece of paper and the immediately following seventh piece of meat were rejected. The results above given are the usual ones. But sometimes, though rarely, results are reached which are analogous to those attained in Metridium by Parker. Thus, i In one case a specimen of Aiptasia annulata accepted the first piece of plain paper, but thereafter refused paper consistently, while accepting meat offered in regular alternation with it. For all these results the following explanation suggests itself: The animals when hungry take both meat and filter paper; when satiated they take neither. Usually the tendency to take both ceases at the same point, but sometimes the reaction to the weaker stimulus (filter paper) cease before that to the stronger stimulus— as a higher animal that is not hungry may ae most things, while accepting peculiarly tempting morsels. If the degree of hunger is thus the determining factor, then it should be possible to produce the rejection of the filter paper by feeding meat alone. ‘This turns out to be the case. Indeed, usually the rejection of filter paper may be induced more readily Modipfiability in Behavior. 455 by feeding meat alone than by feeding the two alternately, or than even by feeding filter paper alone. ‘Thus, two specimens of Aiptasia No. 2, which we will call A and B, living side by side, were both found to take plain filter paper readily. “hen A was fed alternately meat and filter paper, while B was fed successive pieces of meat. After eight pieces had thus been fed to each, A still took filter paper (though slowly), while B refused it abso- lutely—though B would still slowly take a piece of meat. ‘Thus B, through satisfying its hunger with meat, had come to reject filter paper, while A still accepted it after devouring several pieces. Apparently meat is more satisfying to sea anemones than is filter paper! In another case a specimen of the same species was fed filter paper alone. It swallowed ten pieces in succession, till the body was puffed out with them, meanwhile ejecting some of the pieces already swallowed, in the intervals between the taking of new ones. In Aiptasia annulata similar relations were found. ‘The animal could be caused to reject filter paper soaked in crab juice much more readily by feeding it meat alone than by feeding soaked paper alone, or by feeding the two in alternation. A large number of comparative experiments were tried, showing this result to be general. It is therefore clear that the state of hunger or satiety is the essential factor in this behavior, in Aiptasia. ‘The experiments showed further that it is not the mere mechani- cal fulness of the digestive cavity that determines acceptance or rejection, but some change in the metabolic processes themselves. Filling the digestive cavity with filter paper does not have the same effect in producing rejection as does filling it with meat. Even when the cavity is so filled that pieces of paper are repeatedly disgorged, new pieces are readily taken. In Aiptasia No. 2, a piece of paper that has been disgorged after remaining some time in the cavity, is usually swallowed again immediately, iiedteis returned to the disk. As the animal becomes less hungry the details of the behavior toward food bodies change greatly. In a hungry specimen, as we have seen, the food reaction is rapid, often requiring but ten to fifteen seconds. After several pieces of meat have been ingested the reaction of all parts becomes much slower and less precise. The tentacles touched by the food may not react at all for several seconds; then they bend in a rather languid way toward the 456 H. S. fennings. mouth, while the surrounding tentacles may quite omit their reaction. The food body is not placed so accurately upon the mouth as in the hungry individual. Ata further stage toward satia- tion, a piece of crab meat applied to the tentacles induces either no reaction at all or a straight withdrawal—a negative reaction; they may then bend back from the disk along the column. If the meat is placed directly on the disk, in contact with the mouth, the latter may very slowly open and in a languid way partly or entirely enclose the food, even when there is no reaction of the tentacles. The mouth is thus usually readier to give the food reaction than are the tentacles. In this condition of approaching satiation some peculiar com- binations and alternations of positive and negative reactions may be observed. In a specimen of Aiptasia No. 2 after five pieces of alternate meat and paper had been taken, another piece of paper was swallowed, then after one and one-half minutes this was disgorged. ‘The disgorged piece lay on the disk for a few seconds, then the mouth opened and began swallowing it again. But after it was about half enclosed, it was again rejected. Now it was — grasped again and partly re-swallowed, then again rejected. This performance was repeated once more before this piece of paper was definitely rejected. A fresh piece of paper presented imme- diately after was slowly swallowed, then in two minutes disgorged. The anemone presented exactly the spectacle which we should interpret in a higher organism as a struggle between desire and repugnance for the available food. In another case a piece of meat was presented after six pieces had been swallowed. ‘The tentacles reacted only very slowly, but finally deposited the piece of meat on the disk, and withdrew. The mouth opened part way, then closed again without ingesting the food. Later it opened again a very little and enclosed a minute shred of the meat between its lips. The piece was thus quietly held for ten minutes, when it was seen to be sinking imper- ceptibly. Fifteen minutes after it was given it was completely enclosed. Many other cases were seen of partial rejection and acceptance of the same piece of meat. At times after one piece has been rejected, another is accepted. In Adamsia and Metridium, according to Nagel (92) and Parker (’96), after the tentacles of a certain region of the disk have through repeated trials come to reject soaked filter paper, Modifiability in Behavior. 457 those of another part of the same disk will still carry it to the mouth. ‘This shows clearly that a general lack of hunger on the part of the organism as a whole cannot be the only factor involved. In Aiptasia No. 2 I tried experiments to determine whether there was the same independence in the tentacles of different regions. Crab meat was given to the tentacles of the left side; these carried it to the mouth, where it was swallowed, the tentacles of the right side playing no part in the reaction. After the tentacles of the left side had taken five pieces they reacted very slowly, a piece of meat resting against them for several seconds before it was seized. When it was finally carried to the mouth, however, it was swallowed readily. ‘The next piece of meat, not being seized at once by the left tentacles, was trans- ferred to those of the right side. ‘They seized it instantly and quickly carried it to the mouth. ‘Thus it is clear that the experi- ence of the individual tentacles plays some part in the behavior; either from fatigue or some other cause, tentacles frequently stimulated gradually lose the tendency to respond. ‘The fact that this result is produced by meat, the purest form of food, seems to indicate that fatigue may be the cause. But the rest of the experiment indicates that this plays only a minor part in the change of behavior. After a short rest the giving of food to the tentacles of the left side was resumed. ‘They continued to carry it slowly and with much delay to the mouth, where it was very slowly swallowed. After taking four more pieces, the tentacles of the left side absolutely refused to carry any more food to the mouth. ‘The mouth had now almost ceased taking food when directly applied to it, though after some minutes the food was finally ingested. Now a piece of meat was given to the tentacles of the right side, which had only reacted once, and that more than fifteen minutes ago. Yet they behaved in exactly the same way as did the others, refusing to react at all, save by hanging back from the disk along the column. Thus it is clear that the animal is asunit so far as hunger and satiety are concerned. If the satiety has arisen through the activ- ity of the tentacles of one side, the tentacles of the other side are equally affected by it. It is the general progress of metabolism that is the chief factor in determining the reactions to food. As Torrey (’04) has already noted for Sagartia, the reactions of satiated sea anemones differ i in many other w ays from those of 458 H. 8. fFennings. hungry specimens. ‘The well fed animal reacts much less readily and strongly to simple mechanical shock. If touched with a needle the well fed individual of Aiptasia either does not react at all, or contracts very slightly, while the hungry specimen reacts suddenly and powerfully. A slight disturbance in the water has no effect on the well fed individual, while the hungry one contracts strongly. To chemical stimuli the same relations apply. A much stronger solution of any given chemical is required in order to produce contraction in the well fed individual, as compared with the hungry one. The bearing of such facts on quantitative determinations in reaction work is evident. If we should attempt to determine the strength of a given chemical which causes con- traction in Aiptasia, we should obtain totally different results, according as we used specimens that were very hungry, moderately hungry, or thoroughly satiated. No “normal” concentration for causing reaction could be determined for even a single given specimen, ioe the state of metabolism, and with it the tendency to react, is continually changing. It is, of course, clear that the change due to varying metabolic states cannot be interpreted alone as a general increase or decrease of sensitiveness. Much more significant is the complete qualita- tive change in the nature of fhe reaction to a certain stimulus, due to this cause, which we have seen both in Stoichactis and in Aiptasia. 2. ACCLIMATIZATION TO STIMULI. Sea anemones show acclimatization to stimuli in the same way as do the protozoan Stentor and many other low organisms. A light stimulus that is not injurious may cause at first a strong reaction, then on repetition produce no reaction at all, or a very slight one. ‘This is easily shown with Aiptasia annulata in the following way: A specimen is selected with outspread disk close beneath the surface of the water. From a height of about 30 cm. a drop of water is allowed to fall on the water surface just above the disk. At once the animal contracts strongly. Waiting till it has expanded again, another drop is allowed to fall in the same way. Asarule there is no reaction to this or to succeeding drops. Sometimes there is a response to the first two or even three drops, but usually there is no reaction after the first one. A slight Modipfiability in Behavior. 459 reaction of a different sort, that often comes on later, will be mentioned in the next section. Experiment shows that the failure to respond is practically universal if the drops fall three minutes or less apart. With drops five minutes apart there is still marked evidence of acclimat- ization, though irregularities appear. With drops falling at intervals of more than five minutes I was unable to satisfy myself with certainty that acclimatization occurs. Related to the present subject are changes in the reaction to light. Auiptasia annulata is very sensitive to light, expanding in darkness, but contracting after a few seconds when exposed to strong light. In ordinary daylight the animal remains contracted for some hours, but after such a period most specimens extend in spite of the light. In comparative darkness the animals direct the disk toward the source of light, through a contraction on the side of the column exposed to the light. After remaining undis- turbed for a long time in an aquarium that is fairly well lighted, the animals give up their orientation with respect to the strongest source of light; with less light they retain it. 3- REACTIONS MODIFIED AS A RESULT OF THE PAST EXPERIENCES OF THE ORGANISM. Under this head will be considered all positive changes in reaction, due to former stimuli or former reactions of the organism, aside from those due to changes in metabolism. We have already described certain cases belonging here. In the reaction by which the disk is kept clean in Stoichactis we find that a mass of débris on the disk causes first one reaction, then another, till one of these or a combination of several rids the animal of the stimulating agent (see p. 451). In this case either the continua- tion of the same stimulus, or the fact that a certain reaction has been given, induces a new reaction, without change in the external conditions. A similar phenomenon is often seen in the experiments with falling drops of water, described above. ‘To the first drop the animal responds by a sudden sharp contraction, then to a consider- able number of drops there is no response. Now if the drops con- tinue, the animal usually begins to shrink slowly away from the region where the drops are falling, so that in the course of time the 460 H. 8S. fFennings. disk has been withdrawn some distance below the surface, though no decided reaction has occurred to any one stimulus. ‘These facts are precisely parallel to those which I have described in a previous paper (1902, p. 50) for the infusorian Stentor. More marked changes result when the animal is stimulated by light strokes of a rae At the first stroke on the disk Aiptasia contracts strongly. It then extends in the same direction as before. When it is fully extended the stimulus is repeated. The animal responds in the same way as at first. ‘This is usually con- tinued for about ten or fifteen stimulations, the animal each time extending in the same direction as at first. But at length, when stimulated anew, the animal contracts, bends over to one side, and extends in a new direction. Under natural conditions, where stimulation at every extension would usually be due to some fixed object, this would of course put an end to the series of stimuli. If, however, the stimuli are still continued after each extension, the animal repeats for a number of times the extension in the new direction, then finally turns again and tries a new position. This may be repeated many times. But in the course of time the reaction becomes changed i in a still different manner. The anemone releases its foothold and moves to a new region. ‘This result I have not succeeded in attaining by striking the animal with a rod each time it extends; the time required is evidently to be measured in hours. But obstructions may be so placed that every time the animal extends, the disk strikes against a solid body. In such a case it is usually found after a few hours that the animal has moved to a new region. Thus to the same stimulus when repeated many times the anemone reacts first by contraction, then by turning repeatedly into new positions, then by moving away. The phenomena are parallel to those described by the present author (’02) for the infusorian Stentor, and by Wagner (05) for Hydra. Beyond doubt other stimuli would here, as in Hydra and Stentor, produce the same series of reactions. In the behavior just described there are at times certain phe- nomena which bear a striking resemblance to the formation of new habits. Aiptasia annulata frequently extends its body in most awkward turns, the column retaining an irregular and crooked form. ‘This is evidently due to its method ofa life. The animal lives in irregular crevices and crannies beneath stones or in the Modipfiability in Behavior. 461 hollows of the coral reefs. In order that its disk may protrude into the free water, it is often compelled to extend in the irregular way mentioned, and to retain the crooked forms thus Fone heal When removed from the natural habitat it still retains these irregularities of form and action. ‘The lower part of the column may Sean at right angles to the upper part, or there may be permanent S-shaped bends, or still more irregular forms. It would appear that these must have arisen as a nenule of the way in which it extends in its natural habitat. The peculiar methods of extension found 1 given individuals could then hardly be characterized Pi ae than as habits, the peculiarities of form being the structural cor- relates of the habits. In searching for experiments that would test the possibility of the formation of new habits in sea anemones, the following sug- gested itself. It should be possible to produce new habits in Aiptasia by so arranging the surroundings as to compel the animal to extend in a new way whenever it extends, and to retain the new form thus induced. If the animal when thus compelled by obstacles to extend in a new direction, still extends in the same direction after the obstacles are removed, one would be inclined to hold that a new habit had been formed. I supposed that this result would require a long period of time. But some preliminary experiments showed it to be attained, in some cases, with such absolute ease as to raise the doubt whether we have here anything that can be called habit formation. ‘Thus an individual attached to a plane horizontal glass surface was bent in extension far over to the left. Stimulating it repeatedly, it con- tracted at each stimulation, then bent, in extending, again to the left. This continued for fifteen stimulations, one succeeding another as soon as the animal had become fully extended. At the next contraction the animal turned and bent over to the mght. Now when stimulated it contracted as before, then bent regularly, in extending, over to the right. It seemed to have acquired a new habit—bending to the right instead of to the left. Attentive examination showed that when the animal contracted in response to stimulation, the concave side of the column con- tracted a little more than the rest, so that that side remained a little shorter. In other words, the animal did not take on an entirely symmetrical structure, but the region which was most con- tracted in extension remained most contracted also in the con- 462 H. S. Fennings. tracted animal. Now on expanding, all parts extended more or less proportionately to their extension in the contracted animal, so that the original curved form was regained. In other words the structural conditions resulting in the curved form were not really given up even in contraction, and were only made evident when extension occurred. If the animal was compelled by repeated strong stimulation to contract maximally in all parts, then in extension there was no greater tendency to bend in the direction previously occupied than in any other. And in about half the individuals this result followed (after once the first habitual position found in nature had been given up) even after a single stimulation, so that there was no indication of anything like the formation of a new habit. What is the interpretation to be given to the numerous cases in which bending in a certain direction when extended does induce, in the way set conn above, bending in the same direction on a new extension? Is this the formation of a habit? It is certainly a condition of affairs that gives the same result as habit formation. ‘The anemone might indeed be looked upon as a sort of structural model, illustrating the principles on which habit formation might occur. A certain action (extension in a certain direction) leaves structural peculiarities, persisting even in the intervals of action (in the contracted state), which result in a repetition of the same action. Is not this the picture that we commonly make for our- selves of the real nature of habit formation? In the sea anemone this seems to occur in a relatively gross way, but it appears difficult to point out any difference in principle between this and habit formation. Ifthe persisting structural peculiarities were of such a nature as to be hidden from observation, there would be no ground for hesitation in calling these phenomena the formation of habits. There can hardly be loabe that the striking individual peculiarities of action and structure, described above, ‘have arisen in precisely this way, so that it plays the part taken by habit formation in higher animals. Tt would be well if the study of this matter could be extended to the same individual for a long time, beginning with a young, still regular, Bu aeunten compelling i it to live in a position where it would have to extend in a definite irregular way. In this way the development of the structural correlates of the habits (?) could doubtless be observed. Modipiability in Behavior. 463 The facts may be summed up for the anemone as follows: _ Performance of a certain action involves the assumption of certain structural conditions. ‘These conditions persist in a slight degree even in the intervals between the actions. At a new action they show their influence by causing it to take place in the same way as the former one. ‘This gives the same results as what we are accus- tomed to call habit. 4. GENERAL AND COMPARATIVE The sea anemones are among the lowest of the Metazoa, and their behavior, when compared with that of most other animals, is of a very simple character. Yet it is evident that even in these low organisms the reaction to a given external stimulus depends upon many things beside the nature of the stimulus itself. Vary- ing states of metabolism induce totally different reactions to the same stimulus, one state producing the long train of actions look- ing toward the ingestion of food, another inducing the equally long and variable chain of activities resulting in rejection. [he same factors cause marked changes in reaction to other stimuli than possible food. Past stimuli received and past reactions per- formed likewise determine the reaction to a given external con- dition, resulting sometimes in a cessation of reaction, in other cases in a complete change in its character. Certain simple con- ditions produce a tendency in the organism to perform more readily an act previously performed (bending, on extension, in a certain direction). Examination of the conditions under which the animals live shows clearly that all the usual reactions and modifications of the reactions are such as to assist in adapting the organism to its environment. In other words, they aid the physiological processes of which the organism is the seat. Aiptasia annulata, for example, lives in crevices beneath and among stones or coral rocks. It 1s, of course, evident that its food reactions maintain its metabolic processes, which would necessarily cease in their absence, that the rejecting reaction keeps the surface clean, so that respiration may take place uninterruptedly, and obstacles or injurious substances be avoided. The transformation of the food reaction into the rejecting reaction after the animal is satiated with food 1s of course as much to the interest of the sea anemone as it is to that of higher 464 H. S. fennings. animals. If the food reaction were an invariable reflex, occurring whenever food is present, without regard to internal conditions, . the results would be disastrous. ‘The fact that the very hungry animal will take indifferent bodies that would otherwise be rejected is of course likewise adaptive; as Torrey (’04) remarks “substances with a very small food value must be of some impor- tance to a starving polyp although they would not be desirable as food to a well Rouriched animal.” The tendency of Aiptasia to remain in the dark and to contract when strongly lighted keeps it in the crevices where it finds pro- tection for its soft body. ‘The fact that it faces and bends toward the lighted side keeps its tentacles and disk directed toward the entrance to the crevice, where food may be captured; if they were directed toward the darkest part of the crevice little or no food would be obtained. While the contraction under light 1s protec- tive, it would result, if continued indefinitely by a lighted polyp, in starvation; we find that after a considerable period of light the animal extends. In correlation with its life in irregular crevices or under stones we find that Aiptasia does not take any definite position with reference to gravity, as some other anemones do. Such a reaction would render its usual habitat impossible. The tendency to react by a quick contraction when there is a slight dis- turbance in the water is undoubtedly protective. Yet such a dis- turbance when not followed by an attack from its author is not harmful and the animal under such circumstances quickly resumes its usual behavior, even though the disturbance continues. But such a disturbance maintained indefinitely would result in loss of opportunity for obtaining food, and the animal after a time shrinks gradually away fom such a disturbed region. Injurious stimuli, interfering with the natural physiological processes of the polyp, cause contraction—the animal withdrawing from the field of action for atime. But this continued indefinitely would result in a loss of food and doubtless other i injurious effects. We find that the animal has recourse then to extension in another direction, and finally to creeping away and establishing itself elsewhere. Located in an irregular crevice, we find that the polyp extends in various directions, Sari it finds a direction in which its disk and tentacles are unimpeded in their spreading to form a trap for prey. It then continues to extend in this manner, even though this may require the body to bend at right angles or to take other irregular Modipfiability in Behavior. 465 forms. It continues to extend in this manner even when removed from its irregular crevice, and the body is found to have become structurally modified, so that a collection of Aiptasias shows many crooked and zigzag shapes, each being an adaptation to the crevice in which the animal lived. The formation of such habitual methods of extension can be imitated and modified in the labora- tory. All together, the activities and their modifications are clearly such as to directly adjust the organism to its environment, enabling the physiological processes to continue under all sorts of conditions. It has become the fashion to neglect such facts, but they fairly force themselves on the attention of the careful student of the behavior, and their existence can hardly be held to be accidental. To remove such an organism to the artificial conditions of the laboratory and then endeavor to understand its behavior is like dissecting an internal organ out of the body and trying to under- stand its functions when thus separated from the other structures with which it interacts. Almost everything the animal does has a direct relation to something in its usual environment, and when cut off from this environment, its activities are likely to become unintelligible. One can hardly resist the belief that the fact that these activities do assist the physiological processes of the organ- isms has determined their selection and retention from among other possible activities. This adaptation and adaptive modifiability of behavior in sea anemones and their relatives has not been explicitly set forth in most works dealing with their reactions. Yet when other careful accounts of behavior in such organisms are analyzed we can dis- cover such relations as clearly as in Aiptasia. Let us look for example at the cases of Hydra, studied by Wagner (’05), and of Cerianthus, as described in the classical papers of Loeb (’91). It will be found instructive to consider the conditions on which the retention of a certain position depends. Hydra and the sea anemones tend as a rule to retain a position at rest, with the foot attached and head free. ‘This usual position is often said to be due to a reaction to gravity, or to contact, or to some other simple stimulus. But when we examine into the matter closely, we find that it is not an entirely simple one. Let us take first the case of Hydra. Suppose the animal is placed on a horizontal surface with head downward and foot upward. It does not retain this 466 He se fennings. position, but bends the body, placing the foot against the bottom, releases its head, and straightens upward. Aiptasia shows the same reaction. In neither of these animals is the reaction due to a tendency to keep the body in a certain position with reference to gravity, for both keep the body indifferently in any position with reference to the pull of gravity, provided that the foot is attached and the disk and tentacles can be spread freely. To what then is the reaction due? Evidently there is a tendency to keep the foot in contact with a surface, for the body of the inverted Hydra is bent till the foot comes in contact. ‘There is likewise a tendency to keep the head free, for it is released. But this is not all, for now the body is straightened, then the tentacles are spread out symmetrically in all duecaeae It is clear that the reaction is directed toward getting the organism into a position that may be called * “normal,” and this normal position has various factors— attachment of foot, freedom of head, comparative straightness of body, and tentacles outspread. Suppose now that our Hydra has reached this position, and all the conditions remain constant; is this sufficient? We find that it is not. If the conditions remain so constant that no food is obtained, the Hydra becomes restless and changes the position of its body repeatedly, though still retaining its attachment by the foot. Later even this is given up, and the animal, of its own internal impulse, quite reverses the position attained through the ‘righting reaction.” It now bends the body, attaches the head, and releases its foot, thus bringing it back into the inverted position. Is this because the irritability of head and foot have become reversed, so that the head now tends to remain attached, the foot free? Apparently not, for no sooner has the animal taken the inverted position than it draws its foot forward and now performs the “righting reaction”’ again, so that it stands once more on its foot. alee alternations of behavior are repeated, and we find that by this means the animal is moving from place to place (see Wagner, 1905, Fig. 3). It seems clearly impossible to refer each of these acts or the whole behavior to any particular present external stimulus. An internal state—hunger—drives the Hydra to move to another region, and these different opposite acts are the means by which another region is reached. Each phase of the locomotion is Modipfiability in Behavior. 467 evidently partly determined by the fact that a certain other phase has just been performed, partly by the general state of hunger. The same behavior is shown by Hydra mnie continued injurious stimuli of different sorts. In speaking of mghting reactions, it is often said that the organism js forced by the different irritabilities of diverse parts of the body to take a certain orientation with reference to gravity or to the surface of contact (see for example Loeb, 1900, p. 184). The facts just brought out (taken from Wagner) show that we cannot in Hydra consider this orientation forced, save in the general sense that all things which occur may be considered forced— including of course the behavior of man. Man takes sometimes a sitting position, sometimes a standing one, sometimes a reclining one, depending upon his “physiological state’? and past history, and the facts are quite parallel for Hydra. So far as objective evidence shows, the behavior is not forced in Hydra in any other sense than it isin man. Both organisms take that position which seems best adapted to the requirements of their physiological processes; these requirements vary from time to time. In the sea anemone Cerianthus the conditions for staying in a certain position are somewhat more complex than in Hydra, accord- ing to the account given by Loeb (1891). Cerianthus is usually found in an upright position, inhabiting a tube made of mucus and imbedded in the sand. If placed head downward in a test tube, it rights itself in the same way as Hydra and Aiptasia, freeing the head, bringing the foot into contact, and straightening the body. But in Cerianthus Loeb showed clearly that*gravity plays a part in the behavior. If the animal is placed on its side on a wire screen of large mesh, it bends its foot down through the meshes, lifts up its head, and takes its usual position with reference to gravity. If now the screen is turned over, the animal again directs its head upward, its foot downward—as a human being under similar circumstances would do if possible. It may thus weave itself in and out through the meshes. But to be in line with gravity, with head above and free, is not the only requirement for “Cena ue. Loeb found that it would not remain indefinitely in this position on the wire screen, as it does inthe sand. After a day or so it pulls 1 its foot out of the wire and seeks a new abode. Only when it can get the surface of the body in contact with something, as is the case when it 1s imbedded 408 H. S. ‘fennings. in the sand—in its natural habitat—is it at rest. If this condition is fulfilled, the requirement of the usual position in line with gravity may be neglected. Loeb found that when the animal is placed in a test tube, so that its body is in contact with the sides, it remains here indefinitely, even though the tube 1s placed in a horizontal position (Loeb, 1891, p. 54). The head is bent upward, but the body remains transverse to the direction of gravity. Examples of the fact that a certain orientation with reference to gravity is not a rigid requirement even in animals that usually or at times react to ie agent, are common among sea anemones and other lower organisms. ‘hus, Torrey (’04) shows that Sagartia, though it usually maintains an upright position, may ofttimes take a position on the surface film, with head downward. In the rejecting reaction of Stoichactis, described on p. 451, we have clearly a-reaction with reference to gravity, though one which even the most sanguine could hardly denominate a fixed tropism. The situation “waste - matter -on - the - disk - not - removed - by - the - first - (usual) - reaction” is responded to by taking such a position with reference to gravity as results in removing the waste; then the reaction to gravity ceases. ‘This is somewhat analogous to the reaction to gravity described by Bohn (1903) in the hermit crab. While investigating a shell which it may adopt as a home if fitting, this animal Pie up a certain position with reference to gravity— namely, with the body on the steepest slope of the shell, and head downward; it then turns the shell over and ceases to react with reference to gravity. Of a different but equally significant char- acter are the variations shown in the reactions to eravity by the low acelous flatworm Convoluta, as described by Bohn (’03b) and Gamble and Keeble (03). Under conditions that are favor- able Convoluta remains on the surface of the sand. But when the sun becomes hot, or when the tide rises, so that the animal 1s likely to be washed away, it becomes “positively geotropic,” going downward in the sand, where it is protected. When the tide falls again Convoluta becomes “negatively geotropic,” thus reaching the surface of the sand, where it obtains food and carries on its usual activities. “These alternations of reaction become a fixed habit with Convoluta, so that when removed to an aquarium it still goes downward at high tide, upward at low tide, though the con- ditions surrounding it remain constant; it may thus be used for a time as an in- igor tide indicator. Gradually, however, when RY fe Modifiability in Behavior. 469 removed for a long time from the influence of the tides, this alter- nation of reactions to gravity ceases, showing it to be a true habit, resulting from individual experience. Many other instances of reactions to gravity, of the most diverse sorts and variable charac- ter, could be given. Gravity affects organisms in many diverse ways—determining the distribution of internal substances of dif- fering specific gravity, causing differences in the ease of move- ments in diverse directions, inducing strains or pressure in unac- customed parts of the body when an unusual position is taken— indeed, influencing the life processes in almost every detail. Any of the points at which it comes in contact with the life processes may serve as the basis for a reaction, so that we find behavior induced by relations to gravity in different organisms to be of the most diverse character. We have been assured by various writers that the reaction to gravity must be explained in the same way in all cases, but this is evidently said rather in the capacity of a seer or prophet, than in the capacity of a man of science whose con- clusions are inductions from observation and experiment. Returning to Cerianthus, we find, according to Loeb, that even the usual position in line with gravity and with sides in contact, does not satisfy the animal indefinitely, if left quite undisturbed. If it secures no food it again leaves its place and seeks another region. Thus in order that Cerianthus may remain quiet in a given position, a considerable number of conditions should be fulfilled, constituting the usual, and perhaps what we may call the “normal” state of affairs for this animal. ‘hese conditions are the following: (1) The foot should be in contact; (2) the head should be free; (3) the body should be straight; (4) the axis of the body should be in line with gravity, with fhe head above; (5) the general body surface should hei in contact; (6) food should be received at inter- vals. If these conditions are largely unfulfilled, the animal becomes restless, moves about, and finds a new position. But no one of these conditions is an absolute requirement at all times, unless it be that of having the head free. In the wire screen the animal remains for a day or two in the required position with reference to gravity, even though foot and body surface are not in contact. In pene horizontal Ae it remains with foot and surface in contact, though the body is not straight nor in line with gravity. If all conditions are fulfilled save that of food, the animal remains for a time, then moves away. 47° ses Fennings. Clearly, the holding of any given position depends, not on the relation of the body to any one or two sources of stimulation, but on the proper maintenance of the natural physiological processes of the organism. ‘The actinian does not always maintain a certain position with relation to gravity, nor does it always keep its body straight, nor its foot in contact, nor its body surface in contact. It does not at all times receive food. It may remain quiet for considerable periods with one or more conditions lacking. The organism tends on the whole to take such a position as is most favorable to the unimpeded course of its natural physiological processes. Certain usually required conditions may be dispensed with provided other favorable ones are present. The behavior, like that of higher animals, represents a compromise of the various needs imposed upon the animal by its physiological processes. Examination of the literature shows that throughout the Ccelen- terates there is a similar dependence of behavior on the progress of the internal physiological processes, particularly those of metab- olism. The state of metabolism decides whether Hydra shall creep upward tothe surface or shall sink to the bottom (Wilson 91); how it shail react to chemical and to solid objects (Wagner ’05), whether it shall remain quiet in a certain position, or shall reverse this position and undertake a laborious tour of exploration. In the sea anemones it determines, as we have seen, even the details of long trains of reaction. ‘The state of the metabolic processes appears to be the most important determining factor in the behavior of Ccelenterates. The same dependence of behavior on the internal physiological processes is found in other groups, even in those much lower than the Coelenterates—the Protozoa, and particularly the Bacteria. This is brought out especially in some of the work of Engelmann. A number of examples of this relation will be given in the paper which follows the present one, so that they may be omitted here. The fact that in higher animals behavior depends largely on hunger and satiety is, of course, so well known that it need not detain us. The relation of behavior to the internal physiological processes, of which we have given some examples in the foregoing pages, is manifestly of the greatest significance for the understanding of behavior. The facts adduced show directly that in many cases the determining factor in reactions to stimuli is not the anatomical configuration of the body, taken in connection with simple laws Modifiability in Behavior. 471 of conduction, but is the relation of the action of the external agent to the internal processes. ‘The problem presented by the fact that the same stimulus, in the same intensity, applied to the same part of the body, produces qualitatively Bee ee and even opposite results, depending on the inner metabolic states, seems not to have received the attention it deserves. It evidently places marked difficulties in the way of a simple mechanical conception of the reflex process, based merely on the anatomical structure of the organism. ‘he internal physiological state determines in some way which of various courses within the body the transmitted stimulus shall follow and what organs it shall arouse to activity. The organism cannot be looked upon as a static structure, on which external agents must act in a simple invariable way. The organism 1s a process, and some of the chief determining factors in behavior are given by the relation of the internal to the external processes. As the internal processes change, the reaction to external agents changes correspondingly. We find that reactions which assist the existing internal processes are continued or repeated, while those which oppose them are changed. This gives one of the chief bases for the regulatory character of behavior, as I shall attempt to set forth in farther detail in the paper which follows the present one. ‘The metabolic processes, while the most striking of those taking place 1 in the lower organisms, are of course not the only ones occurring in animals. ae immense number of other processes are in progress, and the relation of external agents to these processes may and does equally determine behavior. This gives the phenomena of behavior their complexity, prevent- ing them from being in relations of simple dependence on external agents, as they are often represented of late. Such a view quite underestimates the difficulty of the problem of behavior. ‘The dependence on external agents exists, but is complex, and can usually not be predicted without a knowledge of the present internal state of the organism—this depending on its past history and the course of its various internal processes. It would of course be more convenient if the problems of behavior were as simple as they are often proclaimed to be. Work revealing their complexity 1s naturally not receiv ed with the acclaim that greets the announcement that all these things are simple and easy. But if our object is really to obtain control of the vital processes, then we must face them in all their com- 472 H. S. fennings. plexity. To control animal behavior it 1s necessary to study animal nature, in much the same way that it is necessary to study human nature in order to control human behavior. It is neces- sary to know the past history of the organisms, and what is going on within them, in order to predict what they will do. He who expects even the lower animals to behave always in certain simple invariable ways when acted upon by the various forces of nature has many disappointments in store, when he comes to make a thorough study of the matter. The internal modifying conditions must be made the object of deliberate and extended investigation in lower animals as well as in higher ones, before the study of Ss behavior can be placed on a really scientific basis. LITERATURE, (ChlED. Boun, G., ’03.—De l’évolution des connaissances chez les animaux marins littoraux. Bull. Institut Général Psychologique, No. 6; 67 pages. ’03.—b. Les Convoluta roscoffensis et la théorie des causes actuelles. Bull. Mus. d’Histoire Naturelle, pp. 352-364. DueErDEN, J. E., ’02.—Report on the Actinians of Porto Rico. Bull. U.S. Fish Com., vol. xx, second part, pp. 323-374, 12 pl. GampsLe, F. W.,anD KEEBLE, F., ’03.—The Bionomics of Convoluta roscoffensis, with Special Reference to Its Green Cells. Quart. Journ. Micr. Sci., vol. xlii, pp. 363-431. Jennincs, H. S., ’02.—Studies on Reactions to Stimuli in Unicellular Organisms. IX. On the Behavior of Fixed Infusoria (Stentor and Vorticella), with Special Reference to the Modifiability of Protozoan Reactions. Amer. Journ. Physiol., vol. viti, pp. 23-60. Lors, J., ’91.—Untersuchungen zur physiologischen Morphologie der Thiere. I. Ueber Heteromorphose, pp. 48-59. ’o0.—Comparative Physiology of the Brain and Comparative Psychology. 309 pages. New York. NaceEL, W., ’92.—Das Geschmacksinn der SSO Ee EE =k .O15 = .002 Pcabable ercox of standard deviation: «<2. ~ +c Ls2meece ene eee + .o1! + .o1! PINETARC MENIATOT A PaseoRls Gos LNs < «se ca sag BORGO E En see ee eee -626 -641 ‘For calculating the factors of this curve I am under obligations to my colleague, Dr. Smallwood. Variations Among Scyphomeduse. 555 A study of the table will naturally give rise to the question as to how the several variations in these marginal organs are to be explained. For example it will be observed that twenty-three specimens had more rhopalia than marginal lobes. A reference to Figs. 1 to 3 will quickly afford an explanation of this phenome- non. The figures also show double and twin rhopalia, several cases of which were found. Similar cases are cited by Browne (op. cit., p. gO), and in connection with their discussion he ven- foes the suggestion that perhaps in later development and durin metamorphosis, by a growth of the margin these so-called ‘twin thopalia may become separated thus giving rise to an independent lobe. ‘This seems to me to be extremely doubtful. As a matter of fact it is well known that during growth following metamor- Be) en Fig. 1. Diagram of marginal lobe showing twin rhopalia. Fig. 2. Compound marginal lobe, one of which contains twin rhopalia. Fig. 3. Compound marginal lobes the central one notched at the tip. phosis, increase of the marginal dimension takes place entirely by growth of the inter-rhopalial areas. ‘This is very easily seen by a study of the appearance of new marginal tentacles and by the origin and multiplication of the branching canals. It may therefore be accepted as practically certain that the twin rhopalia of the ephyra continue such in the adult medusa. It is also almost certain that a similar condition is involved in the case of such double rhopalia as are shown in Figs. 3 and 4. In these respects, as in others, as cited by Heckel (op. cit.) we may, | believe, regard 1 it as undoubtedly true that the larval variations are carried over into the adult through the several phases of meta- morphosis. Furthermore, this view is confirmed by the facts clearly established, that the ratio of variation found in the adults is essentially the same as that found in the ephyre. 550 Chas. W. Hargitt. As suggested above, it seems reasonably clear that the excess of rhopalia may be accounted for in the manner already proposed. But it remains to consider those cases in which the number of marginal lobes is in excess of the number of rhopalia. Of these there were found seventeen cases, as against twenty-three of the former. It is quite obvious that for piss a different explanation must be found. We are here limited to two alternatives, namely, either there are cases in which for some reason there has been a failure of a given lobe to develop its usual organ; or on the other hand there may possibly be a subsequent and independent origin of an extra lobe. While I have found undoubted cases of the occurrence of the former condition, and am constrained to regard it as the more usual and probable explanation, at the same time I have found an occasional case in which a belated lobe appears to arise after the ephyra has become free from the strobila, but at the same time it must be admitted that in these cases there is usually found the accompanying rhopalium, though this is not always true. I am therefore constrained to consider both alter- natives as possible, though giving to the first the larger probability. A few unusual features in the marginal and rhopalial variations call for a merely passing notice. Fig. 1, showing an ordinary twin rhopalium, calls for little note aside from the stiremeneant eae it plainly occupies the position and relation of a typical organ, namely, the terminus of a single canal. And in this connection it may be well to recall that in the early ephyra- -life all the canals are simple, that is, unbranched. It is only during the progress of metamorphosis that the complexity of the adult canal system is gradually differentiated. Figs. 2, 3, show a series of extremely interesting variations of eraduated complexity. The first shows a trifid condition of the otherwise normal ephyra lobe, though with the added abnormality of twin rhopalia in one of the notches. In the second there is shown a quadrifid lobe, the lappets of the median pair being small and not particularly remarkable, while in each notch of the outer lappets is a normal sensory body. Several other figures show similar features. Among 486 ephyre taken in the “eel pond,” Woods Hole, in April, 1902, the variants numbered 144, or 29.6 per cent. ce 2 er Variations Among Scyphomeduse. Sy, 6 Fig. 4. Hexamerous ephyra, one lobe of which is compound. Fig. 5. Hexamerous ephyra, with two-lobed mouth, and two gastric lobes. Fig. 6. Ephyra with small supernumerary rhopalium, several examples of which were found. Fig. 7. Ephyra with an adradial rhopalium, and one of the normal lobes devoid of an organ, perhaps due to injury. co8 Chas. W. Hargitt. 3s Taste III. =< ‘ | No. Specimens. | Gastric Loses. Ora Lopes. RHOPALIA. Mareinat Loses. I : 2 F 2 6 6 I 3 3 i, 7 I 3 + 7 7 S 3 + 9 9 I 4 3 10 10 I 4 3 12 12 78 | 4 . 9 9 I | 4 4 8 9 35 4 4 IO 10 2 4 4 fe) fe) 8 4 4 II II 7 4 4 2 12 I 5 5 10 10 I 6 6 II II I 6 6 12 I2 I 6 6 13 13 I 4 - I4 I4 I 7 7 10 10 Each of these rhopalia occupied a single octant of the ephyra margin, and differed but little in size from the others. It should be stated that each was found on a different specimen. In two specimens were found a very rare feature among these varied rhopalial phenomena, namely, the presence of a rhopalium in an iter'obular, or adradial position, as shown in Figs. 6, 7. One of such cases I also discovered in connection with the study of the development of Cyanea. Here we probably have the origin of the condition which eventuates in the equally rare occur- rence of an adradia! rhopalium in the adult medusa, cases of which will be considered in a later connection. Oral and Gastric Organs. As compared with the marginal system that of the vegetative shows comparatively little variation, at least in the ephyra stage, though the present data are far from complete. In the first place Variations Among Scyphomeduse. 559 the number of specimens tabulated was less than one-third of the entire number in the preceding series. ‘This is due in part to the poorly differentiated stage of these organs in the early ephyra, the gonads being entirely lacking, and the mouth-lobes being often so contracted as to render certain determination im- possible. In Table II is shown the range of variation so far as accurate data are at command, including as in lable I divergencies or lack of correlation between the two sets of organs. As will be noted, out of a total of 486 specimens, only 12 or 2.68 per cent vary from the normal. In several figures are shown illustrations of these aspects of our Fig. 8. Ephyra with seven gastric and oral subject. In Fig. 5 is shown a __ lobes, and ten marginal lobes. sketch of one of these, in which there are but two oral divisions and two gastric pouches. One might suppose the directly opposite relations of these organs as figured to be unusual, but when compared with Figs. 6 and 7 it will be seen to be quite in keeping with that found in almost every case, that is, the angles of the mouth correspond with those of the stomach so that the pouches of the latter of course occupy inter- mediate positions. In other words, the angles of the mouth occupy the perradi of the body while the gastric pouches or lobes occupy the interradu. A comparison of figures will readily show the lack of correlation of these organs with those of the marginal system, and at the same time the close correlations between themselves, the latter being likewise evident in the data of the table. Additional data of a similar sort will be presented in connection with the second series, a comparison of which will still further emphasize the small ratio of variation as compared with that of the marginal organs. <60 Chas. W. Hargitt. A Comparison of the Variations Exhibited by Ephyre During Metamorphosis. During the current summer, 1905, I was able to secure a collec- tion of about 1000 ephyre from Waquoit Bay, a body of water some ten miles east of Woods Hole, from which had also been secured a portion: of the previous series 1n 1gOI. Among these were found 392 specimens which were just emerging into young medusze. They varied in size from 7 to 14 mm. in diameter, the radial canals were well differentiated, and the gastric pouches easily distinguishable. There were also 218 ephyre among the number which were entirely devoid of any indications of meta- morphosis, indeed, apparently but recently escaped from the strobila. I was particularly glad to have an opportunity to study a series of this character from a strictly local environment and from the same brood, so to speak, since it afforded an opportunity to test a feature of variation already referred to, namely, whether varietal features existing in one stage are carried over into another, or whether during a period of metamorphism there was at work any selective processes. While the numbers examined in these cases are too small to afford conclusive data on a problem of this character, they may at any rate afford a fair indication as to probabilities, and when taken in comparison with similar series in larger numbers, as in the former case, and also in connection with die observations of Browne (op. cit.), they may become correspondingly more convincing. A comparison of the data presented in Tables IV and V will show, both in relation to the percentage variation and to the question of the persistence of varietal features during the several phases of metamorphism, rather striking points of likeness. So far as the ratio of variation is concerned it will be seen at a glance that it is so nearly the same in the several cases as to pre- clude the probability of anything more than slight and incidental differences. For example, in “Table I the total per cent of variation is 26; in Table V it is 22.2; while in Table IV it is 22.9. Compared with the ratio obtained by Browne, which for 359 ephyre taken in 1893 was 22.6, and for 1116 speci- mens taken in 1894 was 20.9, the results become still more conclusive. Variations Among Scyphomeduse. 561 Of 218 ephyrz taken at Waquoit Bay in April, 1904, there were 50 variants, or 22.9 per cent. The chief features are tabulated as follows: Tas_e IV. No. or Specimens.) Gastric Loses. Orat Loses. Ruorauia. | Mararnar Loses. I 3 3 6 6 J 3 3 7 7 2 4 4 | 6 6 2 4 4 | 7 7 31 - - | 9 9 4 4 4 | fe) IO I ys 4 | II II I 4 4 | I2 | 12 I 5 5 | II II I 5 5 | 12 | 12 I 6 3 | 12 12 3 6 6 | 12 | 12 I 6 | 7 | 12 | Te Now if we compare these results with those shown in Tables VI and VII, in which are presented in detail the variations found in 226 specimens of young Aurelia flavidula taken at Waquoit Bay in May, 1905, and in a collection of adults made at New Bedford about the same time it will be seen that they all tend to confirm the general propositions under consideration, namely, that varietal features found in the ephyra persist in the adult, and furthermore, that there is no evidence of any selective process involved during these several changes in ontogeny. Of 392 ephyree i in process of metamorphosis, taken at Waquoit Bay in April, 1904, there were 87 specimens, or 22.2 per cent., which showed various aspects of variation, the principal features of which are classified in the following tabulated form: 562 Chas. W. Hargitt. TABLE V. _ = a ) No. or | Gastric Orar | Ruopatia. CaNnaL SysTtEM. SpeciMeNns.| Loses. Loses. | f (Perradiale: eee O! AE" tier I 2 2 6 : it Interradiall=y- see O° Aber | Perradial) Soe eee 1 a ; = + u \ Interradial-3- eee OT 1 f iPerradiallnsee eee oO: 7 aa a | 4 4 7 \ Internadiall- oo -seeeee T) “Tp GTaSE Z J Perradial’ 2 2-2 ee i VT eee : 4 + : \ Inferradialiee > eee Oo 1 Sror f Perradialy see eee 2 eh 45 } 4 4 7 \ Interxadialieeeee eens DOW uaa If Rerradiallje.erpiseeti 2 2 teat = 4 zs \ interradiall--e eros fee te oh ee e | { Perradial 22 2-- eee 2 1 ew + si * 1\ Interradiallas- eee fo Tow = f Perradialle cee pmo eit 1 I | 4 - I\ Interradial)= 3.2 -see be MAW “th *T = ii Perradial) 2 see ences 2 ee 3 a + 1\ Interradiall=- 2s eee Meni ot hi ‘ - f Perradial@yon see 2- 2 tas 4 43 \ Interradialle acess je Pai ‘ re | f Perradial ..........- 2 2 ee : + I\ Interradial.......... Pik oe : . i Perradiale enero 2. 2,82) + + 3 I\ Internadiall/ soe 2 Tee if Perradialivae cose jg. 2 I : ¥ : "5 | Interradialiee---e ¥, I it | if Perradial®.= sais ae 2 Ta I 6 5 II | : I\ Internadialiess eerie BP Sa" tee 6 6 6 - if ee fly oc pos Lv De ae nt Imterradial jects estes OA) 1) ae ; 6 6 2 f coe ee Ha A I \0. eel | \ interradial/=-eeae--- oe ‘Ges et Ed te 5) : 6 6 a J oe ee ee i Tere eee i Unterradialiecseeieck 2) 1 eee if iPerradialitaac asset 2 Sepia I 6 5 13 . : i Interradiall@ acs seeore 1 I) Benen I 6 6 13 Ibid. I 6 8 (double) 15 Cf. Fig. 10. : 8 8 o f ae Bhai s meee ° 1 Teer Se il Inferradialee eee ee ° Gee Variations Among Scyphomeduse. 563 Variations in the Adults. As in the study of the ephyrae attention was directed chiefly to the marginal or sensory bodies, and to the central or vegetative, so likewise in the study of the adults the same systems have received primary consideration, though in the latter including also the canal system, as a correlating medium between the others. Attention was also directed to the problem of the probable influence of local conditions in determining variations. As favorable localities from which to secure specimens more or less subject to a definite environment, New Bedford harbor and Waquoit Bay were selected, the latter serving moreover, the further end of ascertain- ing the variations exhibited by ephyrz and adults under the same environment. In addition to authorities cited in a previous part of this paper attention may be directed to the observations of Bateson and Romanes on the variations of Aurelia aurita. ‘The latter has described in some detail variations found in this medusa and has illustrated by diagrams many of the features described. In both the illustrations and the analysis of the facts there is an apparent effort of the author to reduce the variations to as few symmetrical types as possible. As I have pointed out in another connection in reference to the work of Ehrenberg and Agassiz, these attempts to discover a law of symmetry, or perhaps better in modern phrase, a law of regulation, in the diverse variations encountered, have apparently been only partially justified. While it is doubt less true that in many cases, perhaps in a majority, some form of regulation may be distinguished, there are too many cases in which this is lacking to be considered as merely exceptions to such a law. A study of the following facts and illustrations, will I believe justify this view. As Bateson in commenting upon Romanes’ work has remarked, “Tt is impossible in regular threes, sixes, etc., to say that any particular segment is missing or added rather than another.” And if this be the case with an organism like Aurelia, in which the several organs are so sharply differentiated as to be easily distinguished at a glance, it is much more likely to be true in organisms of more complex structure and less sharpness of differentiation. In an attempt to ascertain the comparative frequency of certain 564. Chas. W. Hargitt. variations in Aurelia, Bateson examined 1763 adult specimens taken on the Northumberland coast in 1892. In the tabulation of his results he presents details of only the gonads and oral lobes. Of these there were but 28 abnormal specimens, or a variation of only 1.6 percent. Of the 28 abnormal individuals 19 he considers as “ symmetrical varieties, ” and observes that the other 9g speci- mens, or 33 per cent are “irregular varieties” and are seen “for the most part in single specimens only.” Here Bateson apparently falls into the same error which he has criticized in Romanes, namely, the attempt to reduce the variations to “symmetrical varieties,” regarding “irregular varieties” as exceptional. But the presence of 33 per cent of the so-called “irregular varieties” is too large a proportion to be designated as exceptions. As is w Fall known Aurelia is an octamerous medusa, each octo- mere being characterized by a single, more or less dichotomously branched radial canal, at the terminus of the central stem of which is located the sensory body, or rhopalium, and separated from the adjacent octomere by an unbranched canal, as shown in several of the diagrams. In normal individuals this arrangement is very sy mmetrical, and easily distinguishable. It must not be inferred, however, that the several branching canals are exactly similar, or symmetrical. Indeed it may be safely said that probably no two in a given individual are exactly alike, any more than are two leaves of a given plant. Still, the differences are usually slight, striking variations occurring chiefly in those cases where departures Sour the typical arrangement are considerable. For convenience in following readily the subsequent discussions, it may be well to briefly remind the reader that for descriptive purposes the several canals have been designated by the special names, perradial, signifying those canals arising between the gastric pouches, or mouth angles; interradial, indicating those occupying intermediate positions, or emerging from the outer median portion of the gastric pouches; while the term adradzal refers to the unbranched canals alternating with the other two series. Gastric and Reproductive Organs. Among the most conspicuous variations from the typical con- dition just described are those involving a numerical, or meristic departure. ‘This will be readily understood when it is remenneneal Variations Among Scyphomeduse. 565 that these organs are large and conspicuous, four in number, and usually the first to attract attention. It is doubtless on this account that so many of the earlier observations concerning variation in these medusz dealt almost exclusively with this feature. Among the commonest variation is the hexamerous form, where there are six each of the gastric lobes, gonads, and oral arms. This will be observed at a elance by comparing the several tables, especially Nos. [V and V. Next in frequency is the pentamerous type, where there is a symmetrical arrangement of the organs upon the plan of five. As the several details of these variations are specified so far as their numerical aspects are concerned it is only necessary to refer to the tables already cited. It may be well to notice briefly a few features not capable of tabulation. Among these are the not infrequent occurrence of signs of atrophy, as shown in Figs. 11 and 12. In the former it will be observed that associated with the small size of the pouch and gonad is the entire absence of the interradial canal and its marginal organ. In the latter will be observed the presence of a mere rudiment of a regres- sive gonad in one of the pouches, while in the opposite compound pouch there are two gonads, and in this case the absence of the perradial canal system. Associated with variations in the number, is that of variation in the size and relations of the organs, as already pointed out in the figures cited. Attention was directed to the compound character of the organs. This is a very common occurrence, and probably is indicative of the manner of the origin of supernumerary organs of this character. However, the pentamerous and hexamerous condition is frequently distinguishable in the ephyra, and seem to be quite distinct from the beginning And I have found in Cyanea that occasionally trimerous poly ps occur, and probably give rise to trimerous ephy re and later trimerous medusz. It may not be improbable that the suggestion of Ehrenberg (op. cit.), that the circular gonads which he observed were te result of fusion of what may have been earlier distinct organs, is quite as likely as that above. It may be suggested in (his connection that I have never seen a case such as that cited by Ehrenberg, though its occurrence does not seem improbable, but in every case w hich has come under my observation of a compound gastric pouch, the gonads have been more or less distinct, as indicated in Fics 12: 566 Chas. W. Hargitt. Of 226 small adults collected at Waquoit Bay, May, 1905, there were 55 variants, or 24.3 per cent. The general features of variation are tabulated as follows: TasLe VI. | Pra ae sickae ce | Ruoratia.| Cana System. J Perradial Pe ee i Gi : 2 3 | 2 | Interradial I pea | 2 ih Perradial eres omg, ie 3 : 4 4 / (Interradial ... ae ee | | | Permadiallioa es P. eeeaar 3 4 4 7 | Interradial aps O “1S | i Petradial $252 2. Bae At 4 4 9 \Interradial ... t ae | | J Perradial ae rt 2th 7 | + 4 | 9 \ Interradial . 2 SC eae | J Perradial Bie 2. aa 5 + Z as \Interradial ... Pe Wt |"| Perradiali saz: 2 Tae s | + z we eee te ie pe to | J Perradial ss Ae 3) or eae 2 4 4 10 5 | Interradial I UT aaa | Perradial 2. 2) 2S eGTeay ‘ + 4 i, eee Td Goes | ¥ ; id : Pesradial 3423 2. 2 \ovam 7 as Interradial ... Tee : j | ; a: ner: Bae 2 “De ieee | Interradial ... 2.1 Seer ; , | i i oe as 2 eee Interradial Tay a : 6 | 6 7 eee Wee: O ih ee Interradial. 1 ‘E JT” ee : 6 6 Pe oa 2 eee I. ‘T 2 ee Interradial .. 21 1 “1 jue Variations Among Scyphomeduse. 567 Of 129 large adults taken at New Bedford in May, 1905, there were 29 variants, or 22.5 per cent. The chief features of variation are shown in the following table: TABLE VII. No. or | Gastric Orat | R a Cc Sy SPECIMENS. Loses. Tonrge te ee ANSE SO ESTEM: Rermadiale oe. oat i I 3 3 6 Iinterradiall) = ose. i 3h i 5 Rerradta lee im it alee 4 4 ; 7 Interradrale ss 6) it ie Rerradialie eer Desa ie ed I ie te 4 3 9 Interradiall” 75.4. Te te oh Reradiall= nee Tipe Tie “Ye al 2 4 4 : 9 Intersadiale sass. te 1 ol erradialaseee ee RD BY it OU I 4 4 10 ; Interradiall aaa Legh 2 YI perradialsenee see De Cle, Oa I i 4 IO E Interradial’ =. . Pas Re Ce Rennadialig asa Si inet I 4 4 IC Interradiall = 444. a aT: | Dal Rerradialia ee Ghats. vie At 2 2 4 II : | Interradial rp eae De To Bite oF (Perradial ........ 5S AG ASD 9) 2 4 4 12 Interradialiae == oe ict pear Rerraditall a2. y= re eas Gee! I 4 4 12 Interradial! 92.5). oe Calpe ma Rerradial ee (oy i ate ah ite wl I 6 6 II ; Imtenradial’ 2s5—)- Igy Ale vie ck) Ale al Rerradiales sere [ee Teaeiee hes Th eel 2 6 6 12 \Maitenradiall 25-5. ee CGR Ss So fa I have frequently been able to confirm the observation of Bateson, (op. cit.), that where there are cases of marked dif- ference in the size of the gonads of a given quadrant or of adjacent quadrants, that there is frequently ; a noticeable decrease in the size of the corresponding portion of the bell, as shown in Place 1, Fig. 5. 568 Chas. W. Hargitt. As compared with the v ariations occurring in other organs, par- ticularly the marginal organs and canals, the per cent is extremely small, as will be seen at a glance in comparing the several tables. My observations on this point confirm those of both Bateson and Browne. The former found but 1.6 per cent, while the latter found 1 itas large as2.4 percent. My observations gave the average of 2.75 per cent, as the total variations to be detected in Aurelia flav alas As already suggested in connection with Ehrenberg’s observa- tions, in which he claimed that variation reached 10 per cent, either this must be taken to include the total, in which case it is evidently too low, or if it refer to the vegetative organs alone it is certainly too high, unless indeed it may be possible that the Aurelia aurita of the Red Sea differs very greatly from the species in other waters, or from our own species. Rhopalia and Radial Canals. An examination of the several tables will show that there is a general variation in the direction of an increase in the number of both rhopalia and radial canals. This has been shown to be the case in Aurelia aurita by both Ballowitz (op. cit.), and Browne (op. cit.). While confirming for the most part the results obtained by both these observers, there are points of difference which must be reviewed with some detail, and other points wherein I am unable to accept the conclusions of either in all par- ticulars. Some of these will be considered in their appropriate connections. Concerning the number of rhopalia little’need be said further than to direct attention, as above, to the tabulated facts. A brief word or two in explanation of the Tables V to VII will suffice to render their meaning clear. Beginning with the first, or left hand column, there is listed the number of specimens having the characters given in the following columns. For example, in the second is given the number of gastric pouches, or lobes, in the third the number of oral lobes! in the fourth the number of rhopalia, and finally in the last the canal system not including the adradial system, since it bears for the most part, no direct relation to the rhopalia. The numbers following in each case refer to that of the rhopalia, and where less than the normal Variations Among Scyphomeduse. 569 number is given, as in the first line of Table V the fact in relation to the canals is shown by the 0, indicating their absence. Where a larger number than the normal ts present, as in the fifth line and those following, the figure 2 in the perradial canals indicating that the extra rhopalium is perradially located. In other words, the figures in the columns under rhopalia and canal system serve to show the correlation of the two sets of organs. As indicated above, the tables give no account of the adradial canals, since normally they sustain no direct correlation with the Fig. 9. Ephyra with three compound marginal lobes, quite comparable with those of Figs. 1 to 3. Fig. 10. Ephyra with two mouths, and opposite compound marginal lobes. other canals or with the rhopalia. It must be said, however, that there are definite exceptions to this rule. And while not suth- ciently numerous to call for tabulation along with the others, they occur in too many instances to warrant the somewhat ultra pronouncement of Browne (01, p. 100) to the contrary. Whether a given canal shall be called adradial, interradial, or perradial depends not alone on its position or whether it be branched or otherwise, but upon both its position and relations to the other canals. The name signifies nothing in itself but that of relation- ship. ‘There is no intrinsic reason why an adradial canal should 570 Chas. W. Hargitt. be unbranched, and as Browne admits in another connection (p. ror), there is good reason to believe that its position may shift considerably. Figures 12, 14 and 15, are careful drawings of variant canal features, in each of which there is shown at a an adradial rhopalium, while in one case, Fig. 14, the canal is some- what branched, but taken in its relations with both the other canals I see no other alternative than to regard it as adradial. This is also shown at 4, Plate I, Fig. 2. Both Ehrenberg and Ballowitz (op. cit.), have figured similar cases of undoubted adradial rhopalia. The chief variations found in rhopalia are those of number and position, the later of which has just been noticed. ‘The smallest number found was five, only a single specimen among the entire lot studied having so few. Six were found having but six rhopalia. Others having larger numbers are tabulated in their appropriate places in the various tables. As will be seen the largest number found was fifteen, and in but a single specimen. This specimen was further peculiar in having two fully developed and functional mouths, as shown in Fig. 10. This duplex oral condition served to give the animal a somewhat ovoid shape and at the same time a more or less bilateral aspect, the latter being further accentuated by the presence on almost exactly opposite sides of two compound marginal lobes and rhopalia, as shown in the figure. A similar condition so far as the marginal lobes and rhopalia are concerned is shown in Fig. g. In this case there are three compound lobes, but they do not tend in any way toward either bilateralism or even a trimerous form. An additional compound lobe would have rendered the variation a strikingly sym- metrical one. But like the former and several others of a similar character which came under observation there was seldom exact symmetry. The effect of less or more than the normal number of thopalia may, or may not, disturb the general radial symmetry of the umbrella. For example, in Plate I, Fig. 1, is shown a medusa with but seven rhopalia, yet the general symmetry seems quite normal. By a careful inspection it is not difficult, however, to discover that one of the interradial systems is entirely lacking, Again in Plate I, Fig. 6, of a hexamerous specimen, there are Variations Among Scyphomeduse. 571 12 Fig. 11. Single octomere of adult medusa showing entire absence of the interradial canal, and the contiguity of two adradials. Fig. 12. Diagram of medusa showing adradial rhopalium at a and lack of perradials in the octo- meres drawn. Very small degenerate gonad in upper left hand pouch. Fig. 13. Branched adradial canal at a, and at x the complete blending of the several canals of the segment into one system. 572 Chas. W. Hargitt. six gonads, six oral arms, and eleven rhopalia, the twelfth being absent and its perradial canal system likewise lacking, as shown at P, and still the general symmetry is hardly affected. On the other hand, in several of the photographs, particularly Figs. 4 and 5, it will be seen at a glance that the symmetry is more or less seriously disturbed. In still others, while the general symmetry might not appear to be seriously affected, when atten- tion is directed to the marginal symmetry it will be seen to have suffered quite definitely, in one case three rhopalia instead of one, occupying a single octant. In such a case, which is not rare Fig. 14. Diagram showing a branched adradial canal and rhopalium shown at a sketched from Fig. 2 of plate. Fig. 15. Diagram of the specimen shown in Plate I, Fig. 4, adradial canal and organ at a, perradial system at p, interradial absent or fused with the former. the variation would seem to have been restricted wholly to that single segment, the other seven remaining normal. _ And thus it is throughout; variations in one organ involving in many cases more or less distortion of the correlated organs, or even the entire organism. In other cases the variation has been associated with a regulative adjustment which has more or less served to maintain a fairly definite symmetry of both the immediate organs and the entire animal symmetry. In connection with the study of ephyrz already given, attention was directed to the occurrence of twin rhopalia in several instances some of which are illustrated in several of the figures. Such double structures have been observed in adults, though unless a a en V artations Among Scyphomeduse. - Ss they be searched for critically they are seldom seen, since the hoods and lappets serve to screen them from ordinary observation. It may be of some interest in this connection to refer to the occasional appearance of twin rhopalia in regeneration, an extended account of which I have elsewhere described.'. It has been found that occasionally double rhopalia are produced in regeneration where originally there was only a single one which had been excised in the experiment. Rarely also during such experiments a super- 16 i7 a Fig. 16. Drawing showing three perradial canals and rhopalia in a given octomere, p. Fig. 17. Sketch showing two octomeres a, a, with extremely simple canal systems. The positions of the rhopalia are adradial. numerary rhopalium may develop at unusual points, as has been described in the paper just cited. It is not at all improbable, therefore, that in cases of injury the marginal portion of the umbrella involved may be more or less modified during the process of regeneration, and that marginal organs may appear in some- what unusual places. It is not improbable that the somewhat anomalous appearance shown in Plate I, Fig. 4, may be due ‘Regeneration in Rhizostoma pulmo. Jour. Exp. Zodl., vol. i, p. 85. 574 Chas. W. Hargittt. to regeneration following a marginal injury. A close inspection of the notch just beyond the complex network of canals will reveal the presence of a small thopalium, indistinct in the illustration, but very distinct in the specimen. ‘To Browne’s suggestion that the absence of a marginal body may be due to injury it will suffice to have called attention to the fact that these organs are promptly regenerated and therefore would not probably be long lacking in - any case in which sufficient time had elapsed for the injury to heal. Concerning variations in the canals of Aurelia it remains to call attention to some few points not hitherto considered. In several of the photographic figures are shown varying features of anas- tomosis among the canals, chiefly in the peripheral portions. In addition to the case just referred to others are shown in Figs. 1 and 6. In that of Fig. 1 is shown at various points the usual type of anastomosis, W chile | in Fig. 6 of the hexamerous specimen is shown a most complicated type affecting only one-half of the umbrella. A few other cases are figured in the diagrams in which only a single segment may be involved. This phenomenon of anastomosis I have found about as common in the small adults as in the larger, though Browne considers it as quite rare in small specimens. I have also found it frequently affecting the terminal portion of the adradial canals, especially in those rather rare cases in which the rhopalium is adradial. The branching of the adradial canals has already been referred to ina brief way. It is only necessary to again call attention to Me matter, and refer to several of the figures in which it is shown, , Figs. 13, a, and 14, a. An interesting and unusual condition is cones in Fig. 11, 7, where the interradial system is wholly lack- ing and the two edeadials thus brought into contiguous relations. CYANEA AND DACTYLOMETRA. Several incidental references have been made in the preceding pages to variations observed in ephyre of Cyanea. It has long been well known that Cyanea arctica is a remarkably variable species, so much so that Professor Agassiz recognized some of them as distinct species, and described as such Cyanea fulva, and Cyanea versicolor. But the names have long since passed into the limbo of synonomy, the forms so designated not having even a varietal recognition. In the present instance, however, they may Variations Among Scyphomeduse. 575 serve to suggest the fact that the species varies greatly, but chiefly in the nonessentials of color and size, the southern forms being usually much smaller than those of more northern range. In structural features I have found that this species exhibits very similar variations to those found in Aurelia. While it has not been within the scope of the present paper to enter upon any large survey of the problem as it relates to Cyanea, and no exact data have been accumulated, I have examined considerable numbers of both the ephyre and adults, and find considerable variation in the number of rhopalia, the gastric and oral lobes, and the less important matter of coloration to which reference has been made above. Similar observations have also been made upon our species of Dactylometra, and to the same effect. In general aspects it varies less than does Cyanea, as perhaps both vary jess than does Aurelia, but concerning the fact of considerable variation there can be no doubt. Ina paper upon the structure and development of Dactylometra, Mayer! has stated that the tertiary tentacles arise invariably on either side of the ocularlappets. While I have had no opportunity to examine any considerable number of these medusz, I have nevertheless, found considerable disparity on this point. In several specimens examined in 1902 I found these tentacles arising at points intermediate between the primary and secondary series. I have also found considerable variation in the number of the thopalia and other marginal organs. Since, however, the data are too few to warrant any definite attempt at estimating the quan- titative variations of either of these species, it must suffice merely to note the facts in a qualitative way and leave to another time, or other observer, the further consideration of details. RHIZOSTOMA PULMO. Among about fifty specimens of this medusa which I had occasion to examine critically in the progress of experiments upon regeneration, an account of which has been published elsewhere, and perhaps half as many others examined in the large aquaria less critically, about 15 per cent showed features of variation in Bull. Mus. Comp. Zoél., vol. xxxii. 570 Chas. W. Hargitt. one or more organs. Some of these it may be worth while to briefly consider in this connection. As in the former series, the chief variations noted were those of the marginal organs, in which the range was from five, observed in two specimens, to twelve, found in only a single specimen; and in those of the vegetative organs—gonads, gastric and oral appendages. In these the variation was similar to the former series, though there was not close correlation, between the organs of the two series. [here was, however, almost perfect correlation between the members of the same series. In other words, the gastric lobes and oral pendants were the same in number, and almost always of similar size. In a single case, and that a speci- men having twelve marginal lobes and rhopalia, there was a perfect correlation of all the organs, including gastric and oral. In the usual crowding of the oral arms due to the large number present, two of these organs had been forced into the center of the group, the others forming a closely crowded circle about them. An incidental feature of two of the oral arms was that of the branch- ing or bifurcation of the terminal portions. In the one case the lobes being unequal, while in the other they were quite uniform in size, though with the tips organically fused, or grown together. I have frequently found branched tentacles and oral arms, but it is unusual to find a subsequent union of the branches. It is rarely found in Hydra, and other hydroids, only a single case having come under my own observations. Variations among the Rhizostomata have been recorded inci- dentally by several other observers, among whom are Haeckel,! Keller? and Lendenfeld.? The latter has given much more critical attention to the details of the problem than either of the former. This is, however, in relation to the larval, or ephyra history rather than to that of the adult. Indeed, concerning the adult this observer has recorded but a single case amony many specimens of Crambressa mosaica. Among specimens of Phyllorhiza punctata he records having noted those with more than the normal number of marginal bodies, but believes these supernumerary organs to have been a result of abnormal growth following injury. ‘Haeckel, E., Das System der Medusen, 1879. *Keller, C., Zeits. f. wiss. Zo6l., Bd. xxxviii, S. 641. : % SLendenfeld, R. von., ibid, Bd. xlvii, S. 260, et seq. Variations Among Scyphomeduse. S77 His most remarkable observations are concerned with the marginal lobes and sensory bodies of the ephyrae. He finds these to vary in number from eight, the normal, to twenty-four. While this extreme of variation is large it is not, however, improb- able. ‘The most remarkable feature of the case is the interpreta- tion which Lendenfeld gives. He claims that during metamor- phosis these larvz pass foul the ordinary condition RE octamerism through stages of, first twenty-four, later sixteen and finally emerge to the adult stage with the normal eight-merous condition. “So habe ich gefunden, dass die Ephyren von Phyllorhiza punctata acht, spatere Stadien vierundzwanzig, noch spatere sechzehn und endlich die ausgebildeten Medusen wicdés bloss acht Randkorper besitzen.”’ Though it is not expressly claimed that these phenomena are involved in the observed ontogeny of these medusz, it is neverthe- less, clearly implied. If this inference be correct then the results must be accepted unless other observations may serve to discredit the account given. If, on the other hand, as [am strongly inclined to believe, these several stages have been observed in connection with the general phases of metamorphosis such as one might find among a given series of larvze, then the conclusion | should incline to draw is that these phenomena are but larval variations similar to those described in the earlier portions of this paper. Moreover, when we are farther advised that these peculiar variations are due to injuries the suspicion is still greater that the observed variations are not normal processes during metamorphosis, but in fact, true variations as above suggested. Furthermore, my own experiments upon regeneration in Rhizostoma (op. “ie, would seem to preclude the factor of injury as of any importance in relation to variation. In these experi- ments there was not the slightest evidence to support the view pro- posed by Lendenfeld. We may, therefore, accept these instances as but further extension of our lgno led: of the wide prevalence of variation among the Scyphomedusz, which the more critical attention given to the subject during the past few years has brought to light. FOSSIL MEDUS. It will not be without interest in this connect’on to call attention to the occurrence of variations among fossil medusz. Ina recent 578 Chas. W. Hargitt. monograph on “Fossil Medusze”’ Walcott! has described a large number of fossil medusz, among which several cases of remark- able variation are recorded. In some the variation was so general as to render difficult specific diagnosis. For example, in the description of Brooksella alternata the author says: “The varia- tion is so great in this species that a brief diagnosis is of little value.” . . . “The umbrella lobes vary in number from 6 to 20 or more, and in form from broad, slightly rounded to narrow and strongly rounded. ‘There is no regular sequence of 6, 8, 12, etc., on the contrary the irregular numbers 5 and 7 are largely represented, and 6 and 8 are abundant.” Again in his description of Laotiara cambria he says: “Its variations are greater than in Brooksella.” The lobation of the exumbrella is from the simple four-lobed variety, through series of 5, 6, 7, to 8 or more, to what he designates as the compound type, which are apparently medusz in process of fission. In this species, as in the preceding, there seems to be no definite regularity or sequence in numerical order. In the words of the author: “Tn many individuals there is no regularity, and in the extreme forms there is an irregular network of subumbrella lobes and oral arms.” From the numerous figures of this well-illustrated monograph it is quite evident that similar, if not equally extensive variation is also present in many other of the species described. It is not, however, within the scope of this paper to undertake an extensive review of the entire subject, and hence further details concerning this phase will not be submitted. Citation of the foregoing facts may serve to direct attention of those interested to a phase of the general problem hitherto little considered. INFLUENCE OF ENVIRONMENT ON VARIATION. As is well known, there is a more or less currently accepted belief in the influence of environment as a modifying factor in the variation of organisms. Reference has already been made to Haeckel’s views as to the influence of such factors in relation to the abnormalities arising in medusz reared under artificial con- ditions, and to his arennee suggestions concerning the probable influences of similar conditions in nature. 1Mono. Unit. States Geol. Surv., vol. xxx, Wash., 1898. a a Variations Among Scyphomeduse. 579 It was with these in mind that in securing material for my investigations [ endeavored to have it polecred from points some- what remote, yet in the same general region, and also from environments so definite and yet- distinctly different, as to afford a means of estimating the probable effects traceable to direct and determining factors. It was an unexpected bit of good fortune that brought me into possession of material from an environment apparently quite likely to afford just the desired conditions suitable to a test. ‘This was the occurrence of ephyre of Aurelia in considerable numbers in April, 1902, in a small more or less isolated, and polluted pool, known at the “eel pond” located at Woods Hole, and connected with the waters of the harbor by a very small inlet, sufhcient to admit tide-water daily. The pond has served in some measure as a general dumping ground for various waste and sewage from the village. From this pond I obtained 486 ephyra, all quite young, and many of which I was able to examine alive soon after their capture. A few specimens were obtained from strobilating polyps kept in aquaria where they thrived quite well for several weeks. A second series was obtained from Waquoit Bay, a large bay opening directly into Vineyard Sound, and some ten miles east of Woods Hole. ‘This collection was made in April rtgo1, and contained 1026 specimens. Still a third series was collected at Waquoit in 1904, numbering about 1000. ‘They were obtained in May, and were mostly in process of metamorphosis into young medusz. In 1905 a collection of adults were obtained from somewhat similar environments, one series, indeed, from Waquoit. The other numbering about 200—though on account of poor preserva- tion only 129 were available for accurate study—were collected at the mouth of the Acushnet River, in New Bedford harbor. This environment was as unlike that of Waquoit as is the latter from the “eel pond.” New Bedford harbor receives the sewage and other pollution of the city, as well as the constant influx of fresh water from the river, thus constituting an environment at once more or less local and peculiar, being about 17 miles west of Woods Hole and therefore nearly thirty miles from Waquoit. Observations made upon the ephyre of Cyanea during their development, a brief account of which has been given elsewhere, in which considerable variation was discovered and found to be 580 Chas. W. Hargitt. due in some measure to the artificial conditions under which they were reared, led me to anticipate similar results in ephyre obtained from an environment like the “eel pond.” In this, however, I was somewhat disappointed. For while the ratio of variation found was somewhat larger than that in either of the series from Waquoit, as will be seen “by a comparison of the tables, still it was far less than had been anticipated. The total number found among the 486 specimens taken in the eel pond which showed variant features was 144, or 29.6 per cent. In the collections from Waquoit the series of 1901 comprising 1026 specimens the variants were 24.9 per cent; of the series of 1904, the 218 ephyrz showed 22.9 per cent of variants. While the difference in favor of the eel pond series is appreciable, it is still small, too small indeed, to warrant a final conclusion as to the influence of any given factor as a determining condition. Again, it must not be overlooked that the number of specimens under consideration was likewise comparatively small. Moreover, when we come to compare the data obtained relative to series of adults the uncertainty is greatly accentuated. Com- paring the data of Table VI with those of Table VII, wherein are shown the several features of variation, it will be seen that those from the New Bedford environment, within which they were doubtless bred and reared, have a lower per cent than those from Waquoit, the exact figures of the two being 22.5 per cent for the former, and 24.3 per cent for the latter. Therefore when a careful analysis of the available data is + 7 we are compelled to admit that the evidence concerning the influence of environment so far as the present organisms are con- cerned is not convincing. And until further and more extended comparisons can be made in these or similar circumstances the answer to the general problem must be regarded as negative. SIGNIFICANCE OF THE VARIATIONS IN RELATION TO NATURAL SELECTION: From the foregoing review of the history of variation as it pertains to Aurelia in particular, and to a less extent to other Scyphomedusz also, it must be quite evident that the phenomena are numerous and involve almost every part of the organism. Furthermore, so far as Aurelia is concerned, variations have been Variations Among Scyphomeduse. 581 more or less continuous from the earliest records of von Baer and Ehrenberg to the present time precluding any probability of the operation of simply incidental factors. An inspection of the tabulated records of more recent times will show that the tendency has been constantly toward a more or less definite increase of the several organs, particularly the marginal or sensory, though including also the central or vegetative. If one had taken occasion to construct a curve representing the various phases it would have shown that the variations had been preponderatingly upon one side of the modal line. In the absence of the curve it may facilitate a ready appreciation of the situation to submit percentage values of the variations above and below the normal. Results obtained by Browne covering a period of about five years and including an examination of several thousands of speci- mens presented in percentage figures are as follows: NorMAL. Axsove NorMat. Betow NorMat. LRiGpRUR) couseqgboOrgOgOeae 78.71 per cent. 16.74 per cent. 4.55 per cent. WEREPALIV OR forevcilaciwis'eys sivisis1s,5 97-6 per cent. 1.8 per cent. 0.6 per cent. Results obtained by my own observations covering a period of four years and an examination of about 2500 specimens are as follows: Normal. AxBove NorMAL. Betow NorMat. LRORRUEL 56066000 HORACE 75.07 per cent. 22.97 per cent. 1.96 per cent. WGRa 605.460 pO ede COOOUEe 97-24 per cent. 2.2 per cent. 0.56 per cent. The significance of this line of rather definite and continuous variation is somewhat doubtful. Without specific details in the work of Baer and Ehrenberg it is impossible to formulate con- clusions as to the ratio of variation in Aurelia aurita as observed by them, but the more recent observations of Bateson, Browne, Ballowitz and those herein described, make it perfectly certain that for at least two species of meduse from widely separated regions variation has been remarkably active and continuous. But at the same time it seems equally certain that so far as one is able to see there has been no evidence of the operation of any- thing like natural selection at work. ‘The variant forms do not appear to be more numerous than formerly, nor does the v ariation seem to be appreciably larger in one species than in the other, if 582 Chas. W. Hargitt. indeed we really have in Aurelia aurita and Aurelia flavidula definitely distinct species, a query which has frequently forced itself upon my attention during the present research. In this connection has naturally arisen the question as to the operation of any process akin to mutation.” Mayer" in a recent paper on “The Variations of a Newly Arisen Species of Medusa,” p. 4, reviewing the Vv ariations in Aurelia remarks: “It is evident that sy noeeneat sports,” or discontinuous variations of Aurelia, are continually being produced, and yet the form of the species as a whole remains unchanged.” I have previously discussed the matter of the symmetry of these variations, and need not take it up again, further than to say that it seems to me unfortunate to contend for the dominance of the idea of symmetry in the sense referred to, and that so far as it appears to me there is little in these variations which can be regarded as “discontinuous,” or mutative. They seem on the contrary to be definitely continuous, and somewhat of the nature of fluctuating variations. Browne’s suggestion (op. cit., p. 100), that “if a very slow and gradual change i is taking places in the number of tentaculocysts, fen the tendency is toward the estab- lishment of a race with ten tentaculocysts, due to an increase of two opposite perradial tentaculocysts,’ hardly seems warranted from the facts as known. I can hardly see that there is any such predetermined variation as would be called for by his suggestion. So far as I am aware, the only case of variation among medusz which might seem to be of the nature of mutation is that of Pseudoclytia pentata, Mayer (op. cit.). Of this case we have only the records of a single series of observations. Whether sub- sequent evidence will clearly confirm Mayer’s conclusions remains to be seen. Furthermore, the results would have to be followed through several generations of medusz in order to clearly estab- lish the case as one of definite mutation, and this the future must determine. *Mayer: The Variations of a Newly Arisen Species of Medusa. Bull. Mus. Brooklyn Inst. Arts and Sciences, vol. i, 1901. ‘ a - : * i : ~~ s :) i Z in 7 D * ; \. t ‘ 2 me . Chas. W. Hargitt. Wi oa) + EXPLANATION OF PLATE. All figures were photographed natural size directly from nature. : Fig. 1. At I is shown the missing segment, of an otherwise normal specimen. Toward the per- iphery may be seen typical anastomoses of the radial canals. Fig. 2. Medusa with nine rhopalia, the extra one at A occupying an adradial position, at the ter- minus of an adradial canal, unusual in its terminal branching. Fig. 3. Medusa with ten rhopalia, the two extra ones occupying the same perradial segment, at P. Fig. 4. Medusa with only seven rhopalia, one at A being at the terminus of the much curved adradial canal. At P is shown, at the margin of a complex perradial segment, a very small rhopa- lium. ‘The interradial rhopalium is lacking in the adjacent segment to the left. Fig. 5. Medusa having but three oral lobes, which were excised before the photograph was made, ~ two of the gonads much smaller, and the umbrella of that side also appreciably narrower. Fig. 6. Hexamerous meduse, having eleven rhopalia, the twelfth at P with the entire perradial system of that segment lacking. On the lower right-hand side may be seen the very complex anasto- moses of the canal systems of that region. ; VARIATIONS AMONG SCYPHOMEDUS#. C. W. Hareirr. PEARE Is Tue Journat or ExPeriMENTAL ZoOLoGy, vol. ul. PPE RIMENTAL STUDY ON THE LIFE-HISTORY Of HYPOTRICHOUS, INFUSORIA: BY LORANDE LOSS WOODRUFF. With 3 Prares anp 12 Ficures 1N THE Text. i, Tino NGaehocenmode SACeOnC Teno TREE CeO anc een aoe nn nop ren ar anicon: 585 Hem Genera lel Victnodsr and whechniquesy.cts\o)-t1-|cvoveye(=|<)<)e)-¥e (clare etsls/-totersietale loley stele re ei orelersictotele el aiere 587 CPD ESET TIOMsOlnt Ae NCULLUTESS: a e.jae nate, opate) == Wlovs la\erelercters svaleve rs terete aie Oerneieteis case siaiele eters 590 Teme Oy trichaprallaxs iGultunevA's ts; yye.s/cvsias eieses jens. ci=veletayer sieve stelesereisrareieisiaiere teverene ees 590 a, Ossniotelnaisicilbs & (Cnlatiqeal bh pppoe pope beoUumeeboond oops dbomacteoonoacscnT 594 qe leurotiachaylanceolata, Culture Ay si-)icrereitchalecilnicte ev elsiar the cipteltaeiisieetercicts 594 Ameenrotrichaylanceolata, Culture Be cris oc eieiacls el-imy-iieireleleltleieteteiee eee ciesine 596 PEE aStrOstylarstemmits: Culenre Ac) <:.:. leicrsis cies crore oie olelstelcdele eve tere exch sioieisieierstore eerste 596 ieee Discucsroniouthe! Datasor the Culturess: 260. en. -joeqeiae cite «rin emleieilsieisens aeleeiee 601 1. Rhythmical and Cyclical Variation in the Rate of Division. ................... 601 Dem AG CHICIA MINE) UVENESCENCES) (a1 of-y-tosalelnicle oie Jo) role lele) solutions, are plotted in Diagram XII. This shows that KBr in both dilutions had on the whole a very slight accelerating effect on the division-rate, and also that the greatest variation from the control was caused by the 72, solution. ‘The chief effect of KBr, however, seems to have been to change the rhythm of division as shown when plotted in periods of five days. ‘The daily applica- tion of this salt also was deleterious, and I had especial difficulty in maintaining the culture for more than two days when subjected to daily stimulations, which accounts for the omission of the daily curve in three out of the four periods of the experiment. 5. Comparison of Results. Comparing the results of all the experiments with the seven salts when used as initial stimuli, it is clear that K,HPO, >35 ae the greatest acceleration of the division-rate, while NaCl ats produced tbe greatest slowing of the rate. The largest variation from ae control, when plotted in five-day periods, was shown by KCl ;%,. All the salts tested agreed in ses a marked deleterious effect when*employed daily: K,HPO, 72,5 being slightly the most active in this regard. ‘The table on the opposite page gives the actual status of each experiment in relation to the control for each five-day period of the work. Calkins tried stimulating his Paramoecium cultures with various salts, among them the dike potassium phosphate’ and found that it not only produced an acceleration of the division-rate, but also that there were less fluctuations in the rate. His results show far more uniformity with this salt than do my own. Greeley (’04) investigated the effects of a number of salts on the physical struc- ture of protoplasm and incidentally on the division-rate of Para- moecium, and he arrived at the general conclusion that “with Parameecia from alkaline cultures, anions or liquefying agents stimulate cell-division, cathions or coagulating agents inhibit it. Thus I have frequently observed in my experiments that when the liquefying solution is too weak seriously to modify the structure The Life-History of Hypotrichous Infusorta. 623 TABULATED Resutts or SALT EXPERIMENTS. Five-pay PeEriops. Poqars| Nee yom ‘Avawe SALT SoLu- l Varia- | Errect ae AGE UseEp. TION. : TION IN |IN Four ee Net Ist 2d 3d 4th Four | Lines. ae EFFEcr. Lines. Oy wis ° —2 +9 3 14 + 4 33 +1 KH2PO, ———-— ae ee he | Lad LE a =4\ =3 | 10 — 3 20 fe) 5 ° a0 3 —6 +2 —2 13 —9 3% —24 KsHPO, Per toes een tS |) sae eae to Say crag oe +8 +6 -—7 —22 43 =15 103 —32 KCl ae +9 +1 +1 —23 34. «| 12 84 =o Too ae 722 eT aha aa) 31 —21 Ta ih Sm NaCl |) Se | SK 24 aay Lt Seta), meet mete | ee =s5 See wots ee +4 —I + 2 — 3 10 +2 24 ap K2SOx4 ate sl fe) + 4 = 2 14 +10 34 +24 ae ° | —8 + 1 -— I 10 — 8 24 —2 MgSO, — | N | a “5 Fe | ee $20 +9 —=3 8 6 26 8 63 | 2 Bre | ard —7 +7 fo) +4 | 1 | +4 43 +1 KBr — oem SAP Se Lr | ———— — = | ° +7 = ae 16 + 6 4 +14 Record of the variation in the number of divisions of each initial stimulus experiment from the control, during each five-day period; and also the net effect for the whole twenty days of the experiment. For example: the KH2PO4 ," culture, during the first five-day period, divided exactly the same number of times as the control; during the second period, two times less; during the third, nine times more; and during the fourth, three times less than the control. For the four per- iods of the experiment, then, there was a total variation of fourteen divisions, or a “net effect”’ of four more divisions than the control. 624 Lorande Loss Woodruff. of the protoplasm it will however, greatly increase the motility of the protoplasm and the rate of cell-division.” Among the electrolytes employed by Greeley are three of the salts which I have used: KCl and MgSO, with predominant cathions and NaCl with the anion predominant. He found that KCl #, and MgSO, =3,5 each exerted an inhibiting influence on the fission- rate, through a coagulating of the protoplasm. Referring to these salts he remarks that “the less active solutions, such as KCI and MgSO, do not produce quite so dense a coagulum as the others, and the reaction is considerably slower.” As already stated, my work with an initial stimulation of thirty minutes with KCl 3, and MgSO, s3; produced a quickening of the rate of fission for the first five days or more; the total result, however, for the twenty days of the experiment showed an inhibiting influence. With NaCl 3, Greeley found an increase in the rate and this agrees with the first period of my NaCl experiment, but here again I found a slowing of the rate for the total twenty days. It is impossible, though, to make a direct comparison of Greeley’s results with my own, both on account of the great difference in the methods employed and because he gives no details of the individual experiments. Peters (’04) describes some experiments with KC] on Stentor in which he found that initial stimulation for ten minutes produced an increased division-rate for the three days over which the longer experiments extended. This accords with my results for the early periods of stimulation with the #, and with the ;,; solutions of this salt. To draw any general conclusions from my experiments with salts on the division-rate of Gastrostyla, I think, would be hazar- dous. Before this can be safely done it will be necessary to per- form many experiments on different forms. Work on this subject up to the present time, while affording a nucleus of data as a basis for future investigation, is too meagre and the methods employed by different workers too varied to make the results at all compar- able. As Towle (04) aptly remarks: “the first step toward a clearing of the haze that envelops the subject will be found, I believe, when an effort is made to unify the conditions under which different investigators are working.” From this work on the Protozoa, I am persuaded that the most adequate method of attacking the problem is by breeding long cultures of Infusoria on a fixed diet. While this is a tedious process, it is the only way The Life-History of Hypotrichous Infusorta. 625 in which it is possible to know with any degree of certainty exactly what the pedigree of the subjects of the experimentation is, and unless one has the daily record of the ancestry of each protozoon and knows its status in the life-cycle, any results obtained lose a large part of their value. Nothing emphasizes this point more forcibly than the record of my experiments with the dibasic potassium phosphate. Vile EhbeECr OF SliGHi (ON TLE RATE JOR) DiVviStONe. Maupas (’88) made some interesting experiments on the effect of light on the division-rate of various Infusoria, by keeping cultures for one month in the light and then for one month in the dark and then comparing the rate of division during the two periods. But it would seem that his method is open to criticism for it is clearly impossible to keep the conditions absolutely con- stant during the two months of the experiments, not to mention the fact that, according to Maupas himself, “senescence” is increasing. Consequently it is impossible to say that the dif- ference, or absence of difference, in the rate during two consecu- tive months shows the effect, or non-effect, of light on bipartition. I would call attention to the fact that he found less difference in light and darkness than my records show for any two consecutive months of any of the cultures when light and all other factors have been apparently constant. With this in mind I made an experiment on the effect of light on the division-rate of Oxytricha fallax, and endeavored to elimi- nate the factors which seem to vitiate Maupas’s experiments. This was accomplished by isolating an individual from each line of Oxytricha A-culture, and starting with them a second culture (designated A’) in absolute darkness.t' By this method the light and dark series were carried on simultaneously and this ruled out the question of relative “senescence’’; and at the same time varia- tion in the food was reduced to a minimum, since the same infusion was supplied to both cultures simultaneously. “T’empera- ture differences were avoided also. It would seem, therefore, that light was the only factor removed in the case of culture A’, and 1The culture was necessarily, of course, subjected to light for two or three minutes each day when the record of divisions was being taken. “ee 626 Lorande Loss Woodruff. that this had a very insignificant influence on the fission-rate is shown by the accompanying table. The experiment certainly substantiates Maupas’s result, however obtained, that light 1s of little or no direct importance in the economy of the ciliate. Nes ese Oe See 23 divisions. : Ye tae Oe Se 8 “ Oxytricha A (light) ........------+-+- . Z VNC Geren snc 20 “ | ) No es sea cee 22 divisions = “ Oxytricha A! (darkness) .......------ a a ei ae i = eR SOr eT I Pee oc ae 8“ Total, 82 divisions. Excess in light, 6 divisions. VIII. SUMMARY. 1. The chief object of the work was to ascertain if the life- history of hypotrichous Infusoria is characterized by “cycles,” and if so, the cytological changes which occur and the effect pro- duced on the cycles by changes in environment. 2. [wo cultures of Oxytricha fallax, two of Pleurotricha lan- ceolata, and one of Gastrostyla steinit have been carried on. Oxytricha culture A extended from October 26, 1901, to July 14, 1903, during which time 860 generations were attained. Culture B was started December 10, 1902, and died out through an acci- dent November 22, 1903. Pleurotricha culture A was isolated November 10, 1902, and became extinct July 3, 1903. Culture B was carried continuously from November 25, 1902, to March 13, 1904, when it was lost by an accident. ‘The culture of Gastrostyla was started May 28, 1904, and died out December 5, 1904. The life-history of each culture is represented graphically by a curve which is plotted by averaging the number of divisions per day of the four lines constituting each culture, and then averaging this for five- or ten-day periods. 3. All the cultures give incontestable proof that the species studied pass through periods of greater and less general vitality The Life-History of Hypotrichous Infusorta. 627 as measured by the rate of division. ‘This cyclical change is most prominent in the Oxytricha A-culture. The periods of depres- sion lead to death if the culture is subjected continuously to the same environment. 4. Minor fluctuations occur in the division-rate which I have termed “rhythms”? and which are to be clearly distinguished from cycles. The rhythms are probably indicative of a rhythmical change in the metabolism of the organism, though they are influenced somewhat by almost imperceptible changes in the environment. 5. he results of the experiments seem to indicate that “rhythms” and “cycles” should be defined as follows: Arhythm is a minor periodic rise and fall of the fission-rate, due to some unknown factor in cell-metabolism, from which recovery is autonomous. A cycle is a periodic rise and fall of the fission-rate, extending over a varying number of rhythms, and ending in the extinction of the race unless it is “rejuvenated” by conjugation or by changed environment. 6. Changes in the environment will revive the lagging func- tions during the descending cycle, as is shown conclusively by the sudden recuperation of Oxytricha A during July, 1902. There is every reason to believe that this “reyuvenescence”” was produced by treatment with extract of beef. 7. Seasonal and temperature changes have no apparent influence on the cyclical fluctuations of vitality. Variation in temperature, however, undoubtedly affects somewhat the daily rate of division, if not directly, at least through the food supply. 8. The number of generations which constitute a cycle is not at all constant; and there is no evidence to show that duration in time is of any significance in the forms studied. g. Periods of extreme depression of vitality are manifested on the physiological side chiefly by a greatly decreased division- rate, and by the comparative frequency of pathological divisions. Morphological changes are apparent chiefly in (1) an increased vacuolization of the cytoplasm; (2) distortion and fragmentation of the macronuclei; (3) numerical increase of the micronuclei; and finally (4) ina reduction of the ciliary apparatus. 10. Variation in the size of the infusorians during the life- cycle is marked; the organisms being very small during periods 628 Lorande Loss Woodruff. of high reproductive activity and progressively increasing in size as “degeneration” advances. In the last couple of generations before death ensues the size is secondarily reduced by a shrinking of the cytoplasm. 11. A disappearance of one of the micronuclei occurred at certain periods of high reproductive activity. 12. [hese cultures strongly suggest that it is customary to regard the structure most frequently observed in “wild” Infusoria as too constant in character, and to overlook the fact that, under varying conditions, modifications may occur which are in no way abnormal. 13. Throughout the entire period of the cultures no tendency to conjugate was shown in any of the series, and experiments for endogamous and exogamous syzygies failed to produce a single case. 14. Experiments with KH,PO, K,HPO, KCl, KBr, K,SO,, MgSO, and NaCl gave evidence of the extreme sensitiveness of Protozoa to solutions of electrolytes. Initial stimulation with KH,PO,, K,SO,, and KBr in ;2,; solutions caused in each case a slight acceleration of the division-rate; while initial stimulation with ;", K,HPO,, KCl, NaCl, and MgSO, caused a slowing of the rate. Daily stimulation with the same solutions of each of these salts invariably caused a marked inhibition of the fission- rate. Initial stimulation with KH,PO, 3.3, showed no change in the rate while K,HPO, 7.3, produced a marked increase. K,SO, 33, accelerated division; and KCland NaCl each in 3, solutions, retarded it; while KBr -3, accelerated the fission-rate. Comparison of the effects of the two solutions of each salt shows that, almost without exception, the more dilute solution produced the greater variation in the rate from the control. 15. Stimulation with K,HPO, 7.3, gave different results at various periods of the life-cycle, which indicates that the state of the general vitality of the culture, and also the rhythms, are factors which must be taken into account in experimental work of this nature. 16. Light has little or no direct effect on the division-rate of Oxytricha fallax. Zodlogical Laboratory, Columbia University, New York, 1905. T he Life-History of H ypotrichous Infusorta. 629 LITERATURE. Butscuu, O., ’76.—Studien tiber die ersten Entwickelungsvorgange der Eizelle, der Zelltheilung und der Konjugation der Infusorien. Abh. d. Senckenb. nat. Gesellsch. Frankfurt a. M., x. ’83.—Protozoa. Bronn’s Klassen und Ordnungen der Thierreichs. Catxins, Gary N., ’01.—The Protozoa. Columbia University Press. ’02, 1.—Studies on the Life-history of Protozoa. I. The Life-Cycle of Parameecium caudatum. Archiy fiir Entwickelungsmechanik der Organismen, xy, I. ’o2, 2. (With C. C. Lieb.)—Studies on the Life-history of Protozoa. II. The Effect of Stimuli on the Life-Cycle of Paramcecium caudatum. Archiv fiir Protistenkunde, 1, I. ’02, 3.—Studies on the Life-history of Protozoa. III. The Six Hundred and Twentieth Generation of Paramcecium caudatum. Biol. Bull., iu, 5. ’o4.—Studies on the Life-history of Protozoa. IV. Death of the A- Series of Paramcecium caudatum. Conclusions. Journal of Experimental Zodlogy, i, 3. Dujarpin, F., ’41.—Histoire naturelle des zodphytes Infusoires, comprenant la physiologie et la classification de ces animaux, etc. ENGELMANN, T. W., ’76.—Ueber Entwickelung und Fortpflanzung von Infusorien. Morphologisches Jahrbuch, 1. GeppEs, P., anD THomson, J. A., ’00.—The Evolution of Sex. Second ed. Scribners. Gree Ley, A. W., ’04.—Experiments on the Physical Structure of the Protoplasm of Paramcecium and its Relation to the Reactions of the Organism to Thermal, Chemical and Electrical Stimuli. Biol. Bull., vii, 1. Hertwic, R., ’03.—Ueber Korrelation von Zell- und Kerngrésse und ihre Bedeu- tung fur die geschlechtliche Differenzierung und die Teilung der Zelle. Biol. Centralblatt, Bd. xxiii. Jounson, H. P., ’93.—A Contribution to the Morphology and Biology of the Stentors. Jour. of Morphol., viii, 3. Jouxowsky, D., ’98.—Beitrage zur Frage nach den Bedingungen der Vermehrung und des Eintrittes der Konjugation bei den Ciliaten. Verh. Nat. Med. Ver. Heidelberg, xxvi. Lyon, E. P., ’04.—Rhythms of Susceptibility and of Carbon Dioxide Production in Cleavage. Amer. Journal of Physiology, xi, I. ’02.—Effects of Potassium Cyanide and of Lack of Oxygen upon the Fertilized Eggs and the Embryos of the Sea Urchin (Arbacia punctulata). Amer. Journal of Physiology, vii, I. 630 Lorande Loss Woodruff. Mauras, E., ’88.—Recherches expérimentales sur la multiplication des Infusoires ciliés. Arch. d. Zool. exper. et gén., 2me sér., Vi. ’89.—Le rejeunissement karyogamique chez les Cilies. Arch. d. zodl. exper. et gén., 2me sé€r., Vil. Peters, A. W., ’04.—Metabolism and Division in Protozoa. Proc. Amer. Acad. Arts and Sci., xxxix, 20. Rywoscn, D., ’00.—Ueber die Bedeutung der Salz fiir das Leben der Organismen. Biol. Centralblatt, xx, 12. Scott, J. W., ’°03.—Periods of Susceptibility in the Differentiation of Unfertilized Eggs of Amphitrite. Biol. Bull., v, 1. Srmpson, J. Y., ’o1, 1.—Observations on Binary Fission in the Life-history of Ciliata. Proc. Roy. Soc. Edinb., vol. xxiii. ‘or, 2..-Studies in Protozoa. I. Notes on the Intimate Structure of Protozoa, as Exhibited by Intra-vital Staining. Proc. Scot. Microscopical Soc., ili, 2. STEIN, F., ’83.—Der Organismus der Infusionsthiere. Leipzig. STErRKI, V.,’78.—Beitrage zur Morphologie der Oxytrichinen. Z. w. Z., xxxi. Tow e, ErizapetH W., ’04.—A Study of the Effects of Certain Stimuli, Single and Combined, upon Paramcecium. Amer. Jour. of Physiol., xiy ee Verworn, M., ’97.—Allegemenine Physiologie. WALLENGREN, H., ’01.—Inanitionserscheinungen der Zelle. Zeit. f. allg. Physiolo- mie, 4, T- WEIsMANN, A., ’84.—Ueber Leben und Tod. Witson, E. B., ’00.—The Cell in Development and Inheritance. Second ed. Columbia University Press. T he Life-History of Hypotrichous Infusorta. 631 EXPLANATION OF PLATES. The photographs were taken by Dr. Edward Leaming, of Columbia University, from permanent preparations stained with picrocarmin. The magnification is the same in every case and the relative sizes, therefore, represent absolute differences. The figures, unless otherwise specified, are of Oxytricha fallax, Culture A. Prate I. Figs. 1 and 2. Two individuals in the 230th generation, period 16, April 2,1902. (Cf. Diagram I.) The cytoplasm is vacuolated and the macronuclei are vacuolated and displaced in the cell. A charac- teristic ‘‘halo” is visible about the macronuclei. The individual shown in Fig. 2 has three micronuclei. Figs. 3 and 4. Individuals in the 239th and 241st generation respectively. Period 24, June 1902. The two macronuclei in each are fused and their structure appears somewhat more homogenous than is the case in those illustrated in Figs. 1 and 2. Fig. 5. Specimen in the 243d generation, period 25, June 24, 1902, showing an extreme case of cytoplasmic vacuolization. The nuclei are exceptionally normal for this period of the cycle. Fig. 6. Specimen in the 246th generation, period 25, July 1, 1902. Fig. 7. Individual in the 246th generation (A-2), period 25, July 2, 1902. The macronuclei are surrounded by a ‘‘halo” (cf. Fig. 1). é Fig. 8. Individual in the 247th generation (A-1), period 25, July 2, 1902. Note the condition of the cytoplasm. Prarte II. Fig. 9. Specimen in the 250th generation (A-1), period 26, July 6, 1902. The cell is shrunken and the cytoplasm considerably vacuolated. Note the somewhat reduced size and irregular contour of the cell. This is the last of the line A-1 before it was “rejuvenated.” Figs. 10 and 11. Specimens in the 255th generation, period 27, July 21, 1902. These individuals are from line A-2 which remained dividing, at this time, at the slow rate. The specimen photographed in Fig. 10 has ingested a Trachelomonas volvocina. Fig. 12. Specimen in the 256th generation (A-1), period 26, July 8, 1902. This line had divided six times within the past forty-eight hours. Note the normal condition of cytoplasm and nuclei as compared with the preceding specimens. Fig. 13. Specimen in the 287th generation (A-1), period 27, July 20, 1902. Size is reduced. Compare with Fig. 12. Fig. 14. Individual in the 331st generation (A-1), period 29, August 7, 1902. Size is reduced. Nuclei are proportionately large. Fig. 15. Specimen in the 4ogth generation, period 32, September 1, 1902. Apparently a ‘‘normal” individual in every respect. Fig. 16. Specimen in the 542d generation, period 36, October 17, 1902. Cytoplasmic vacuoliza- tion begins to appear. Fig. 17. Individual in the 829th generation, period 56, April 29, 1903. Nuclear fragmentation has begun. Fig. 18. Specimen in the 853d generation, period 62, July 2, 1903. Nuclear and cytoplasmic degeneration is far advanced. The size of the cell is greatly increased. 632 Lorande Loss Woodruff. a : Fig. 19. Same as Fig. 18, Plate.