- •. -::,-• THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. E. JUST, Howard University E. G. CONKLIN, Princeton University FRANK R. LlLLIE, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University EDMUND B. WILSON, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor VOLUME LXXIV FEBRUARY TO JUNE, 1938 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. 11 THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: \Yheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between June 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the re- mainder of the year. ICntercd October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. LANCASTER I'Kl.SS. INC.. LANCASTER. PA. CONTENTS No. 1. FEBRUARY, 1938 PAGE CLEVELAND, L. R. Longitudinal and Transverse Division in Two Closely Related Flagellates 1 CLEVELAND, L. R. Origin and Development of the Achromatic Figure 41 WAKSMAN, SELMAN A., AND UNTO VARTIOVAARA The Adsorption of Bacteria by Marine Bottom 56 COE, W. R. Primary Sexual Phases in the Oviparous Oyster (Ostrea vir- ginica) 64 SONNEBORN, T. M. The Delayed Occurrence and Total Omission of Endomixis in Selected Lines of Paramecium aurelia 76 REDFIELD, ALFRED C., AND ANCEL B. KEYS The Distribution of Ammonia in the Waters of the Gulf of Maine 83 BRACKET, JEAN The Oxygen Consumption of Artificially Activated and Fer- tilized Chaetopterus Eggs TYLER, ALBERT, AND N. H. HOROWITZ On the Energetics of Differentiation. VII 99 BARNES, T. CUNLIFFE Experiments on Ligia in Bermuda. V. Further effects of salts and of heavy sea water 108 SUMMERS, F. M. Some Aspects of Normal Development in the Colonial Ciliate Zoothamnium alternans 117 SUMMERS, F. M. Form Regulation in Zoothamnium alternans No. 2. APRIL, 1938 EARTH, L. G. Quantitative Studies of the Factors Governing the Rate of Regeneration in Tubularia iii jv CONTENTS PAGE KIDDER, GEORGE \Y., AND C. LLOYD CLAFF Cytological Investigations of Colpoda cucullus. 178 NOVIKOFF, ALEX B. Embryonic Determination in the Annelid, Sabellaria vulgaris. I 198 NOVIKOFF, ALEX B. Embryonic Determination in the Annelid, Sabellaria vulgaris. II 211 PIERSON, BERNICE F. The Relation of Mortality after Endomixis to the Prior Inter- endomictic Interval in Paramecium aurelia 235 GELBER, JULIUS The Effect of Shorter than Normal Interendomictic Intervals on Mortality after Endomixis in Paramecium aurelia. . . . 244 SPEICHER, KATHRYN G., AND B. R. SPEICHER Diploids from Unfertilized Eggs in Habrobracon 247 MAXWELL, JANE Inactivation of Sperm by X-Radiation in Habrobracon . . . 253 SCHMIEDER, RUDOLF G. The Sex Ratio in Melittobia chalybii Ashmead, Gameto- genesis and Cleavage in Females and in Haploid Males (Hy- menoptera: Chalcidoidea) 256 HAYES, FREDERICK RONALD The Relation of Fat Changes to the General Chemical Em- bryology of the Sea Urchin 267 ABRAMOWITZ, A. A., AND R. K. ABRAMOWITZ On the Specificity and Related Properties of the Crustacean Chromatophorotropic Hormone 278 MAST, S. ()., AND COLEEN 1 < > \VLER The Effect of Sodium, Potassium and Calcium Ions on Change- in Volume of Amoeba proteus. . . . 297 MAST, S. O. The Contractile Yacunlr in Amoeba proteus (Leidy) 306 I.II.I.KLAND, OLE Duration of Life without Food in Drosophila pseudoobscura . 314 YOUNG, K. T. The Life History of a Trem.it ode (Levinseniella cruzi?) from the Shore Birds (Limos.i fecloa and Catoptrophorus semi- palmatus inornatus) 319 GIESK, ARTHUR C. The Effects of I'ltra-Yiolet Radiation of X 2537A upon Cleav- age of Sea- Urchin Eggs 330 CONTENTS v PAGE OLSON, MAGNUS The Histology of the Retractor Muscles of Thyone briareus Lesueur 342 No. 3. JUNE, 1938 OWEN, BRADFORD B. A Simple Teleost Kidney in the Genus Cyclothone 349 WELSH, J. H., AND F. A. CHACE, JR. Eyes of Deep-Sea Crustaceans. II. Sergestidae 364 LEAVITT, BENJAMIN B. The Quantitative Vertical Distribution of Macrozooplankton in the Atlantic Ocean Basin 376 MORGAN, T. H. A Reconsideration of the Evidence Concerning a Dorso- ventral Pre-organization of the Egg of Chaetopterus 395 MORGAN, T. H., AND ALBERT TYLER The Relation between Entrance Point of the Spermatozoon and Bilaterality of the Egg of Chaetopterus 401 SUMNER, F. B., AND P. DOUDOROFF Some Experiments upon Temperature Acclimatization and Respiratory Metabolism in Fishes 403 STILES, KARL A. Intermediate-winged Aphids and the Time-of-Determination Theory 430 GALTSOFF, PAUL S. Physiology of Reproduction of Ostrea virginica, I 461 INDEX 487 Vol. LXXIV, No. 1 February, 1938 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY LONGITUDINAL AND TRANSVERSE DIVISION IN TWO CLOSELY RELATED FLAGELLATES L. R. CLEVELAND (From the Biological Laboratories, Harvard University, Cambridge, Massachusetts} The two flagellates considered in this paper are closely related morphologically, and are probably of common origin; yet one divides longitudinally in a typical flagellate manner, and the other trans- versely, as in ciliates. In the latter, the anterior daughter cell retains the parent extranuclear organelles, while the other daughter develops new organelles at the posterior end. This organism may represent the beginning stage in the development of a new group of flagellates. The chromosomes of both organisms are interesting because they are large, two in number (except in rare instances), clearly differenti- ated in size, persist through the interphase, show structural details plainly, divide in the telophase, and are unmistakably moved to the poles by the extranuclear chromosomal fibres. The achromatic figure is also interesting because the fibres com- posing it are as distinct as the flagella which the organisms possess, and it is plain that they have the same individuality as the flagella. They arise from the centrioles, which follow the flagellar bands, and are sometimes so intermingled with the flagella that they can be distinguished from them only by their function in nuclear division. Eleven species of Spirotrichonympha have been described from termites but, with the exception of 5. polygyra, the species described by Cupp (1930), the descriptions are so brief that it is impossible to determine how many species are valid. And the description of Cupp left unanswered many vital questions concerning the structure and behavior of extranuclear organelles and the processes of nuclear division. Further, an examination of the protozoa of Kalotermes (Paraneotermes Light) simplicicornis (Banks), the termite harboring the S. polygyra described by Cupp, shows that her description dealt with two distinct organisms, and is therefore not only inadequate but, in many respects, inaccurate. However, it is not surprising that Cupp overlooked the fact that she was dealing with two organisms, because 1 L. R. CLEVELAND they are so much alike in some respects — and yet so different in others. If the number of flagellar bands is not considered, their differences may be seen only when the processes of cell division are studied. These processes Cupp states she did not understand because of failure to find a sufficient number of successive stages. An examination of the interphase flagellates of this termite shows that the number of organisms with two spiral flagellar bands is about equal to the number with four spiral flagellar bands. This fact Cupp noted, but she supposed that those with two bands were recently divided four-banded forms which would soon develop two new bands. But the number of two-banded forms is just as great in preparations without dividing organisms as in those with them. This alone indicates that the two-banded forms are not young or developing four-banded forms. There are also forms with three, five, and six spiral flagellar bands, but these are very rare in comparison with the two- and four-banded forms. In all hypermastigote flagellates occasional irregularities occur in the flagellated areas, and therefore one should expect to find a few organisms, when countless thousands are studied, whose spiral flagellar bands likewise deviate from the normal numbers. These abnormalities are probably due, at least in part, as unpublished observations on B ar bid any mpha show, to exceptional behavior of the centriole, the organelle from which the flagellated areas arise; sometimes, one centriole fails to produce another from itself during cell division, in which case one daughter cell gets only one centriole while the other gets two; and at other times, one centriole produces two new centrioles instead of one from itself, thus giving one daughter three centrioles and the other two. Organisms with one flagellar band also occur, and in preparations containing many dividing individuals they are present in greater numbers than the forms with three, five, and six flagellar bands; some of these are probably the result of irregularities in the behavior of the centriole, but most of them, as will be made clear presently, are posterior daughters produced in the transverse division of the body of the forms with two spiral flagellar bands. Kven though Cupp considered the forms with two spiral flagellar to be identical with those having four bands, her species, • Polygyra, is valid for the forms with four bands because she con- sidered four to IK- the normal number of bands. It is only necessary to give a name to the form with two bands. This, however, is not an task, for its relationship to previously described species and genera must be considered from three angles: should it be placed in LONGITUDINAL AND TRANSVERSE DIVISION any of the three genera which possess two flagellar bands; should Spirotrichonympha, which possesses four flagellar bands, be amended so as to receive it; or should a new genus be erected for it. Except for the fact that it has two instead of four flagellar bands, it is indis- tinguishable from 5. polygyra when only interphase extranuclear organelles are considered, and its interphase nucleus does not differ greatly from that of S. polygyra; the four chromosomal coils are in two sheaths, while in 5". polygyra each of the four chromosomal coils is in a separate sheath. This difference results from the failure of the sheaths containing the daughter chromosomal coils to separate in the late telophase of the two-banded form. But when the processes of cell division, particularly the behavior of the centrioles, the direction of the achromatic figure, and the plane of cytoplasmic division, are considered the two-banded form is so different from S. polygyra that one wonders whether it is desirable to place it in the same genus with S. polygyra. On the basis of the chromosomes alone, no one familiar with hypermastigotes, especially those with spiral flagellar bands, would hesitate to separate S. polygyra generically from Macrospiro- nympha, Leptospironympha, and Spirotrichosoma, the previously described forms with two spiral flagellar bands, for S. polygyra has two rod-shaped chromosomes, while Macros pironympha, Leptospironympha, and Spirotrichosoma each has a fairly large number of V-shaped chromosomes. But the two-banded form under consideration also has two rod-shaped chromosomes which, like those of S. polygyra, are clearly differentiated morphologically into a short one and a long one; and it, too, should be separated from the previously described genera with two spiral flagellar bands. It is plainly a question, then, of whether one should separate S. polygyra and the two-banded form generically on a basis of four clear-cut differences; number of flagellar bands, behavior of centrioles, direction plane of the achromatic figure, and the plane of cytoplasmic division. Aside from the number of flagellar bands, the differences are largely physiological and the present tendency is not to place organisms in separate genera on this basis. This leaves for consideration only the number of flagellar bands. If the extranuclear organelles other than the spiral flagellar bands of S. polygyra and the two-banded form were at all different, the two organisms should be separated generically, but it has been impossible, after using various fixatives and stains, to detect a single difference. Hence, I believe the two organisms are more closely related than the number of flagellar bands indicates. In other words, the number of flagellar bands in this case does not represent a fundamental difference, and since it doesn't, it should not be used to separate the two organisms generically. LIBRARY teV ~ h 4 L. R. CLEVELAND In view of the facts presented in the discussion above, the hyper- mastigote flagellate with two spiral flagellar bands in Kalotermes (Paraneotermes Light) simplicicornis (Banks) is placed in the genus Spirotrichonympha Grassi and Foa, 1911 and the species bisfnra, the species being new. In order to conserve space and to bring out more clearly the similarities and differences between S. polygyra and S. bispira, the morphology and processes of cell division in the two species are described in one paper. I am indebted to Professor S. F. Light for a generous supply of the termites used in this study, to Miss Jane Collier for technical and research assistance, to Miss Dorothy G. Harris for making the drawings, and to the Penrose Fund of the American Philosophical Society for financial assistance. It was necessary to make and study eleven hundred permanent preparations before obtaining the infor- mation presented. SPIROTRICHONYMPHA POLYGYRA Morphology In ten individuals the body ranged in length from 63 to 112 microns with an average of 81; the width at the widest portion ranged from 25 to 60 microns with an average of 45; the distance of the greatest width from the anterior end averaged 56 microns; the distance from the anterior end to the nucleus averaged 21 microns; the transverse diameter of the nucleus averaged 11 microns; the distance from the posterior termination of the spiral flagellar bands to the posterior end of the body averaged 26 microns; the length of that portion of the body covered by the flagellar bands (the flagellated area) averaged 55 microns; the number of turns or spirals made by the flagellar bands averaged 45; the length of flagella from the surface of the body averaged 20 microns; the length of flagella from the basal granules to the surface of the body averaged 7 microns; the total length of flagella averaged 27 microns; the distance from the flagellar bands (straight line) to the surface of the body averaged 5 microns, and the width of a flagellar band is about 1 micron. In the anterior end the spirals made by the flagellar bands are small and close together and do not vary much for a distance of 5 to 7 microns. Then they gradually become larger and farther apart until the mid-region of the body is reached. From this point on to their termination there is little variation in their diameter and in tin- distance between them (Fig. 1). They turn sharply at their point of origin in the anterior end, and this gives them a ring-like appearance when viewed vertically. \Yhen viewed laterally at their point of LONGITUDINAL AND TKANSN ICRSE DIVISION 5 origin, they may be seen to originate in two groups, two in each group (Text-fig. A, 3). In other words, there are two points from which the bands grow out posteriorly, and two of the bands arising from each point, lie side by side for a short distance. Sometimes, just before, but usually shortly after, cytoplasmic division the process of duplicating the flagellar bands begins. It cannot begin earlier, as in JJHJflPS ^ r^rr. TEXT-FIG. .1. Details of portions of Spirotrichonympha polygyra. 1. Surface view of spiral flagellar bands. X 3000. 2. Flat surface of a band, showing rows of basal granules and the flagella that arise from the granules. X 4800. 3. Anterior end, showing how the flagellar bands begin to spiral. X 4800. some related genera, because the four bands of the parent cell must unwind first and distribute two bands to each daughter, a process that is not completed in this flagellate until just before cytoplasmic division (Figs. 28-37). If the two new or daughter bands were to grow out before the unwinding of the parent bands was completed or nearly so, they would become entangled with the parent bands and 6 L. K. CLEVELAND could never lake their places besides them to form the iiiterphasc organism with four spiral llagellar bands. When viewed laterally, depressions may be seen between the band- as shown in Text-fig. .-1 , 1. In other word-, the bands appear to be on ridges This impression is probably produced by the ertopla-mir layer, which extends Irom the surface of the body to the bands, pushing inward between the bands. The llai surface of the llagcllar bands does not lie parallel \\ith the antero-posterior axis of the body as do the bands in Macros f>iron yniplni, but lies at an angle, the po- terior margin of the band being nearer the surface of the body than the anterior, so that it is only when the bands are flattened that views of them as shown in Text-fig. .1. 2 may be obtained. The basal granules, from which the llagella arise, lie in rows antero-posteriorly directed across the flagellar bands. \\ it h two granule- in each row (Text-fig. A, 2). The two llagella that arise from the granules in each row adhere to each other from their point of origin until they reach the surface of the body big. 1, Text-fig. .1, 1. 2). After leaving tin- band, the llagella extend posteriorly almost to the next band, from which point they turn someuhat abruptly and extend to the surface of the body, slanting posteriorly as shown in Fig. 1. The large single axostyle is a hyaloid structure containing many fibres, extends from the anterior end through the central portion of the body, becomes wider as it approaches the nucleus which it sur- rounds, then smaller, enlarges again near the posterior end of the body, and after leaving the body tapers to a line point. In the portion anterior to the nucleus the axostylar fibres are grouped close together, so that a few ol them sometimes retain a small amount of stain in individuals suitably destained for nuclear details. This portion was mistaken for a centriole by ( upp (1930). By varying the detaining l inie, all degrees of des tain ing of t he axostyle max be obtained, ranging from a heavily stained structure to one \\ilh no stain at all. The so-called parabasals differ greatly in appearance from those of most genera of flagellates, but their reaction to fixatives and stains is the same, and I presume they should be regarded as parabasals. Instead of being long and lender, they are spherical and tollo\\ the spirals of the flagellar bands from a point near the posterior border of the nucleus to the termination ol the bands 'I'ig. 1). They are probably attached to the band- since they follou them -o regularly and since they retain their relationship to the band- \\hen the band- and the body .ire broken up by drastic treatment; but it has been impossible to see a connection between them and the bands-. There an- other strut line-, o\ei looked by ('upp and oilier- \\lio have studied Spirolrichonympfia, \\hich follou the llagellar bands just LONGITUDINAL AND TRANSVERSE DIVISION 7 as closely as the parabasals and which are clearly connected to the bands by fine threads, each structure being connected by a single thread. These structures are considerably smaller than the parabasals, stain deeply following fixatives which do not render the parabasals stainable, and follow the bands a considerable distance anteriorly beyond the termination of the parabasals (Fig. 1). In every indi- vidual, they terminate at practically the same point, namely halfway between the mid-portion of the nucleus and the anterior end; and they extend posteriorly to the termination of the bands. But, unlike the parabasals, they lie closer together along the bands in the region of the nucleus and anterior to the nucleus than elsewhere. They, like the parabasals, probably arise from the bands, since various stages in their growth may be noted in new bands which are increasing in length posteriorly. Mitosis The chromosomes are duplicated in the late telophase (Fig. 15), and as a result. of this each interphase nucleus contains four chromo- somes (Figs. 3-7). Each chromosome contains a distinct coil within a sheath and is anchored to the anterior margin of the nuclear mem- brane by an intranuclear chromosomal fibre. These ends of the chromosomes are referred to as the anterior ends or as the attachment ends, while the opposite ends, which are attached either to a chromatin nucleolus or to portions of it, are referred to as the posterior ends. The portions of the chromatin nucleolus attached to the chromosomes vary considerably in size and shape (Figs. 3, 5) ; sometimes a single piece may be attached to all the chromosomes (Fig. 4) ; at other times a piece may be attached to each pair of chromosomes (Fig. 8). The daughter chromosomes are usually twisted around each other, the degree of twisting varying considerably in different nuclei. When all the chromosomes are attached to one nucleolus, they are twisted little if any, and in such nuclei the direction of the chromosomal coils may be followed plainly from one end of the chromosomes to the other (Fig. 4). It is also possible in many of the other nuclei where the chromosomes are each attached to a portion of the chromatin nucleolus and hence are permitted to twist considerably, to follow the coils clearly from the anterior to the posterior end of the chromosomes and to see that the coil sometimes reverses its direction in an individual chromosome (Fig. 3) ; and wThen each pair of daughter chromosomes is attached to a portion of the chromatin nucleolus, as illustrated in Fig. 8, the direction of the coils may be determined at a glance because they are so plain. A coil may reverse its direction at any point in a chromosome, and the coils of daughter chromosomes may be directed 8 L. R. CLEVELAND either alike or differently. Models of chromosomal coils, made of wires, show that the shifting of direction of coiling in a chromosome does not interfere with the coil splitting lengthwise, as it does in the telophase, and separating to become two individual coils of daughter chromosomes. The first step in cell division is the separation of the spiral flagellar bands at their anterior ends, two of the bands moving in one direction and two in the other; and as they separate they unwind, a process that is not completed until shortly before cytoplasmic division (Figs. 28-37). Almost immediately after the bands begin to separate and unwind, the achromatic figure begins to form between the twro groups of separating bands (Figs. 28, 29). In most dividing organisms the fibres of the achromatic figure, like the flagella, appear to arise directly from the flagellar bands about midway between the nuclei and the anterior ends of the cells, but in a few individuals — especially those distorted by drastic treatment — a fine, darkly staining line may be seen following one band of each group (of two bands). This line is the elongate centriole from which the fibres of the achromatic figure arise (Fig. 31). It should be noted here that in other genera of the Spirotrichonymphida1 previously studied (Cleveland, Hall, Sanders, and Collier, 1934) two elongate centrioles follow the spiral flagellar bands from their point of origin at the anterior ends of the cells for a considerable distance posteriorly, but their posterior or distal ends, which produce the achromatic figure, are free of the bands. In Pseudotrichonympha, a hypermastigote without flagellar bands (Cleve- land, 1935), the unusually long centrioles adhere to the inner margin of the rostral portion of the flagellated area, although this, compara- tively speaking, is a short portion of the centrioles; in Barbulanympha, a hypermastigote with two, short flagellated areas at the anterior end instead of flagellar bands, the centrioles which are about 40 microns in length, adhere to the lamella underneath the flagellated areas for a distance of 6 to 10 microns; and in Trichonympha, another hyper- mastigote without flagellar bands, the posteriorly directed elongate centrioles are free except for their point of origin where they are fastened to the lamella underneath the flagellated area. Thus, there are all degrees of anchorage of the centrioles to another organelle, ranging from Trichonympha, Joenia, Joenopsis, and other genera, where the anchorage is slight, through Barbulanympha, Urinympha, Rhynchonympha, Eucomonympha, Teranympha, Staurojoenina, Macro- spironympha, Spirotrichosoma, and Leptospironympha to Spirotricho- nympha, where it is complete. Without this series, the situation in Spirotrichonympha would seem unusual and would be difficult to understand. LONGITUDINAL AND TRANSVERSE DIVISION The manner in which the achromatic figure arises from the centri- oles that follow two of the four flagellar bands is similar to that in other hypermastigotes: Astral rays arise from the distal end of each centriole, meet, join, overlap, and grow along one another to form the central spindle (Figs. 28-31). As more rays overlap, the central spindle becomes broader; and as the rays increase in length, it becomes longer. However, a point is reached shortly after nuclear division when the central spindle ceases to broaden, and it soon begins to narrow, because the daughter flagellar-band centriole complexes, to which its ends are anchored, move in opposite directions and thus pull apart the astral rays that formed the central spindle, the last-joined astral rays being the first ones to pull apart (Figs. 33, 35). Finally, the central spindle is pulled apart and disappears (Fig. 36). The early central spindle lies a considerable distance anterior to the nucleus (Figs. 28-30), but, as it increases in size and the anterior ends of the developing daughter cells move in opposite directions, it is gradually moved toward the nucleus, finally stretching directly across it (Figs. 31, 32). The direction of the central spindle in S. polygyra is crosswise to the long axis of the body, while in S. bispira, as will be described presently, it extends parallel with the long axis. Shortly before the central spindle comes in contact with the ever intact nuclear membrane as a result of its moving posteriorly, some of the astral rays become extranuclear chromosomal fibres by making contact with the intranuclear chromosomal fibres anchored in the nuclear membrane, and as the central spindle increases in length, due to the centrioles from which it arises moving in opposite directions, these extranuclear chromosomal fibres begin to move the chromosomes toward the poles, one short and one long daughter chromosome moving toward each pole (Figs. 9, 10). In order to conserve space and since the same situation exists in S. bispira, where it is illustrated, the extranuclear chromosomal fibres have not been drawn; nor has the relation of the central spindle to the nucleus and chromosomes been shown in most instances. However, the position of the central spindle and of the extranuclear chromosomal fibres can be determined in the drawings of the nuclei by the direction of the intranuclear chromosomal fibres; the extranuclear chromosomal fibres extend from the intranuclear chromosomal fibres in the nuclear membrane in the same direction as the intranuclear chromosomal fibres (Figs. 9-12); and the direction of the extranuclear chromosomal fibres indicates the position of the central spindle, since they and the fibres of the central spindle arise from the distal ends of the centrioles. The nuclei (Figs. 3-15) are all mounted so that their anterior surfaces point toward the top of the page. 10 L. R. CLEVELAND Just before the chromosomes begin their poleward movement, they become considerably shorter and the turns made by their spirals become broader. These stubby chromosomes persist until shortly after nuclear division (Figs. 9-12); then they gradually elongate and their coils and sheaths split lengthwise, thus producing one short and one long pair of chromosomes which persist through the interphase (Figs. 13-15). SPIROTRICHONYMPHA BISPIRA SP. NOV. Morphology In ten individuals the body ranged in length from 59 to 102 microns with an average of 81; the width at the widest portion ranged from 32 to 48 microns with an average of 40; the distance of the greatest width from the anterior end averaged 57 microns; the distance from the anterior end to the nucleus averaged 21 microns; the transverse diameter of the nucleus averaged 10 microns; the distance from the posterior termination of the spiral flagellar bands to the posterior end of the body averaged 27 microns; the length of that portion of the body covered by the flagellar bands (the flagellated area) averaged 54 microns; the number of turns or spirals made by the flagellar bands averaged 34; the length of the flagella from the surface of the body averaged 20 microns; the length of the flagella from the basal granules to the surface of the body averaged 7 microns; the total length of the flagella averaged 27 microns; the distance from the flagellar bands (straight line) to the surface of the body averaged 5 microns, and the width of a flagellar band is about 1 micron. It is clear from these measurements that S. bispira differs from S. polygyra neither in the size of the body nor in the size of the organ- elles. And the interphase organelles of the two organisms do not differ in appearance and number except for the number of bands and the enclosure of each pair of daughter chromosomal coils in a common sheath in S. bispira, differences already noted. There is, however, a notable difference in the number of turns made by the flagellar bands, those of S. polygyra making 45 and those of S. bispira making 34. Here is a fact which in itself indicates that the two-banded and four- banded forms are distinct organisms; for, if the forms with two bands making 34 turns developed two more bands and these new bands followed the old ones, as they would have to do, the number of turns .\ "iild IK- f)S. The bands of .S'. bispira are slightly farther apart, >ecially in the anterior end, than those of S. polygyra (Figs. 1,2). The description given for the axostyle, flagella, parabasals, and other • AtnimK -lear organelles of S. polygyra takes care of these organelles LONGITUDINAL AND TRANSVERSE DIVISION 1 1 in S. bispira, and we may proceed immediately with the description of mitosis. Mitosis The various processes connected in one way or another with mitosis and cell division in S. bispira are unusually interesting because they are so plain, because of their unique features, and because they differ decidedly from those of S. polygyra. There is one short and one long chromosomal sheath in the interphase nucleus and each sheath contains two chromosomal coils (Fig. 16). The division of the coils occurs in the late telophase (Fig. 27). Each coil is attached to the nuclear membrane by a fibre which is termed the intranuclear chromosomal fibre. Until the early anaphase, each coil is thus anchored to the anterior margin of the nuclear membrane (Figs. 16—20). (In all the drawings of nuclei the anterior margins are directed toward the top of the page.) The opposite or posterior ends of the chromosomes are all attached either to a single chromatin nucleolus or to portions of a chromatin nucleolus, as described in 5. polygyra. The first step in nuclear division is the division of the chromosomal sheath and the moving apart of the daughter chromosomal coils (Fig. 17). As the daughter chromosomes move apart, they frequently coil around each other (Figs. 17-19). This movement of the chromo- somes is not poleward movement, but merely movement within the nucleus, and the daughters are probably interconnected during this period — although interconnections cannot always be seen— for later when they contract and straighten, just before poleward movement begins, interconnections may be seen plainly (Fig. 20) ; and these strands connecting the chromosomes are not broken until considerable progress in poleward movement has been made (Figs. 21-23). It should be noted here that poleward movement of the chromosomes is not to the right and left, as in 5. polygyra, when the anterior end of the organism and the anterior surface of the nucleus are foremost, but is anterior and posterior (up and down). In other words, the chromosomes of 5. polygyra move at right angles to the long axis of the cell, while those of S. bispira move parallel with the long axis. The reason for this will be clear when the development and function of the achromatic figure is explained. So far as I know, there is no cell hitherto described where the behavior of the centrioles during the formation of the achromatic figure is the same as in S. bispira. There are two spiral flagellar bands in each interphase organism, but it has been impossible to determine whether a centriole follows one of these bands prior to the beginning 12 L. R. CLEVELAND of cell division, because during division a centriole usually follows one band so closely that the centriole and the band appear as one structure. In other words, the centriole cannot be distinguished from the band, and its presence during division, with the exception of a few instances, can be demonstrated only by its function. The elongate centriole which follows the band during division may persist as an elongate centriole during the interphase, just as the centrioles of Barbulanympha and many other genera persist; it may degenerate except for its anterior portion in the late telophase, just as the centrioles of Pseudotricho- nymplia do; or there may be one short and one long centriole in the interphase as in Trichonympha. At present, I see no way to determine which possibility is correct. But, irrespective of whether the posterior portion of one centriole which follows a band during division degener- ates or persists during the interphase, the manner in which it functions during cell division may be demonstrated clearly. In order to facilitate the explanation of the processes concerned with cell division, it seems desirable to name and label some of the structures involved in the processes. The two interphase flagellar bands which persist and do not change position during cell division are designated parent flagellar bands (p.f.b.). In most organisms a centriole follows one of these bands from its point of origin at the anterior end of the cell for a distance of 4 to 5 turns or spirals, and these two structures are designated centriole parent flagellar band (c.p.f.b.). Occasionally the centriole does not follow the band for more than two turns before it separates from it and continues inde- pendently; then it is labeled centriole (c.). The earliest stage in cell division is the growth of a new flagellar band from the point of origin of the parent flagellar bands posteriorly. This new flagellar band has a centriole which, in the early stages of development, follows it for its entire length, and where the (wo follow each other they are desig- nated centriole new flagellar band (c.n.f.b.). This new band soon breaks loose from its point of origin, becomes free, and migrates posteriorly; then the end which at one time was connected to the parent flagellar bands at their point of origin in the anterior end of the cell is designated the anterior end (a.e.); the other end which, as development progresses, elongates beyond the point where the centriole terminates is designated the elongating end (e.e.). As this new band elongates posteriorly from its point of origin, it develops basal granules from which flagella (/.) grow out. Astral rays (a.r.) arise from the distal portion of the centriole following (he parent flagellar band and soon meet those that arise from the same portion of the centriole that follows the new flagellar band; \\hen they meet, they join, overlap, LONGITUDINAL AND TRANSVERSE DIVISION 13 grow along one another and thus form the central spindle (c.s.}. The centriole that follows one of the parent flagellar bands may be referred to as a stationary centriole, and the one that follows the new or free flagellar band as a free centriole, although it is not free in a sense because it is attached to the band. Similarly, the astral rays arising from the two centrioles may be referred to (Figs. 52, 53) as astral rays from free centriole (a.r.f.c.) and astral rays from stationary centriole (a.r.s.c.}. The formation of a new flagellar band — first attached at its point of origin, then free and migrating posteriorly while an achromatic figure is formed between it and one of the (stationary) parent flagellar bands — presents an interesting picture in the mechanics of cell division and renewal of extranuclear organelles. This migrating flagellar band, extending in length as it migrates, evidently has considerable difficulty in finding its way to the posterior end of the cell, for it is sometimes entirely outside the parent flagellar bands and sometimes partly outside and partly inside. Its movement is further handicapped because it is joined to one of the parent flagellar bands by the central spindle arising from its centriole and that of the parent flagellar band. But in most instances it eventually makes its way to the posterior end of the cell and, while so doing, picks up and carries with it two chromo- somes, half of the nuclear membrane, and other nuclear materials. However, it sometimes fails to make this journey, and these instances are particularly interesting from the standpoint of the manner in which the achromatic figure is formed and its function in nuclear division. Two such instances are illustrated: in one (Fig. 52) the new band appears to be caught by the parent flagellar bands near its anterior end; in the other (Fig. 53) the new band appears to be lodged to the side and anterior to the parent flagellar bands. In the first instance, it is difficult to distinguish clearly astral rays arising from the centriole following the new band from the flagella that are arising from the new band, but the astral rays that are arising from the distal end of the centriole that follows one of the parent flagellar bands are long and plain and extend between the spirals of the parent bands to the nucleus (Fig. 52). In the second instance, astral rays may be seen clearly arising from the centriole following the new band; they extend inwardly and the flagella, for the most part, outwardly (Fig. 53). The astral rays that arise from the stationary centriole (a.r.s.c.) extend within the spirals to the nucleus. In neither instance is a central spindle formed, because the centrioles, from which the astral rays forming the central spindle arise, are so placed that the astral rays arising from one centriole cannot meet those arising from the 14 L. R. CLEVELAND other. In both instances the chromosomes have the appearance of telophase chromosomes, yet the pairs of daughter chromosomes have not moved apart because the proper formation and function of the achromatic figure was upset by the position of the free centriole which adheres to the improperly placed new flagellar band. The fact that the free centriole in this organism assumes almost every conceivable position during its posterior migration is responsible for the great variation in the appearance of the achromatic figure. For example, the central spindle of the achromatic figure may extend from the free centriole to the stationary centriole entirely within the parent flagellar bands (Figs. 49-51, 55); it may be entirely outside the parent bands (Fig. 47); one portion of it may be inside and the other outside (Figs. 43, 48) ; or, as is more often the case, it may be partly outside and partly inside (Figs. 42, 44, 45). The end of the central spindle adjacent to the stationary centriole is fairly constant in shape when the centriole follows the parent flagellar band closely, because it assumes the shape of that portion of the centriole-band complex from which the fibres arise, and hence is cylindrical (Figs. 50, 51); but when this centriole does not follow the band closely, as occasionally happens, there is nothing to direct the fibres into a cylindrical position and they merely extend posteriorly from the distal portion of the centriole (Figs. 41, 42, 45). The shape of the other end of the central spindle depends entirely on the position of the new flagellar band and the centriole following it. If the centriole and band lie in almost a straight line, the adjacent end of the central spindle is flat or nearly so (Figs. 47, 51); if they lie somewhat in a cylindrical position, the adjacent end of the central spindle is somewhat cylindrical (Figs. 44, 50) ; if they are zigzag or wavy, so is the adjacent end of the central spindle (Figs. 43, 45, 46, 49). The position of the free band and centriole may also be such that only a few of the astral rays arising from the stationary centriole join those arising from the free centriole to form a central spindle outside the parent flagellar bands, while the other astral rays arising from the stationary centriole extend between the parent bands to the nucleus (Fig. 46). Many variations in the numbers of fibres that join or fail to join occur. The r61e of those astral rays which, by becoming extranuclear chromosomal fibres, function in the movement of the chromosomes to the poles is plain. These rays are longer than those which in other hypermastigotes perform the same function. This is because the nuc-leus lies farther away from the centrioles, the point of origin of the rays, than in most cells. In other words, the astral rays become long bctnre they reach the nucleus. In some instances these extranuclear LONGITUDINAL AND TRANSVERSE DIVISION 15 chromosomal fibres may be followed all the way from their point of contact with the intranuclear chromosomal fibres in the nuclear membrane to the poles (Figs. 45, 55); in others they are lost among the fibres of the central spindle soon after they leave the nuclear membrane (Fig. 51). There can be no question regarding the fact that they carry one short and one long chromosome anteriorly and one short and one long one posteriorly, because, when the centrioles from which they arise are so situated in the cell that these fibres cannot reach the chromosomes, the chromosomes do not move apart and the nucleus does not divide. The chromosomes are anchored to the anterior surface of the nuclear membrane until the anaphase, and if there were no inter- chromosomal connections between the daughters of the two pairs, two of the chromosomes would not move at all until the nuclear membrane began to elongate prior to pulling in two. The situation would be very simple, as is occasionally the case when the connections between the daughter chromosomes are broken in the prophase or metaphase: two chromosomes (a long one and a short one) move to the posterior surface of the nuclear membrane, the nuclear membrane elongates, pulls in two, and daughter nuclei are developed. But the more usual thing is for the connections between the daughter chromo- somes to persist until the two chromosomes that are going to form a part of the posterior daughter nucleus are pulled to the posterior surface of the nuclear membrane (Figs. 21, 22). The connections between the posterior ends of the chromosomes are the last ones to break because the pull on the chromosomes is from their anterior ends. Incidentally, we have here an explanation of what produces the so-called equatorial plate stage in chromosomes; the pull on the daughter chromosomes from opposite directions (antero-posteriorly in this organism) causes them to take up a more or less central position within the nucleus. Variations of this stage, of which there are many in different organisms, result from the irregularity in the behavior of the connections between daughter chromosomes, those in some organisms breaking sooner than in others. Cell Division The type of cell division in 5. bispira is so unusual that it deserves special attention. Shortly after the division of the nucleus and the disappearance of the achromatic figure, the new flagellar band, which now lies in the posterior end of the cell (Fig. 56), begins to arrange itself in the form of a spiral, the first portion to so arrange itself being the anterior end (Fig. 57). This is the end which in the early stages 16 L. R. CLEVELAND of mitosis lies adjacent to the anterior ends of the two parent flagellar bands (Figs. 38, 39), the end which does not elongate and which is labeled a.e. in the illustrations of the formation and development of the achromatic figure (Figs. 39-55). It now becomes the anterior end of the daughter cell which is developing at the posterior end of the parent cell (Figs. 56-59). This new band soon becomes arranged in spirals from the anterior to the posterior end (Fig. 58). Meanwhile, both the posterior and anterior developing daughter cells begin to form new axostyles (Figs. 56-58). Transverse cytoplasmic division soon occurs, producing two independent daughter cells; one, the anterior daughter, obtaining a nucleus and all the parent extranuclear organelles intact except the axostyle; the other, the posterior daughter, obtaining a nucleus, the new flagellar band, the centriole adhering to the new band, a small amount of the parent cytoplasm, and the extranuclear organelles such as flagella, parabasals, axostyles, etc. developed before cytoplasmic division. In other words, all the extranuclear organelles of the parent cell except the axostyle, which is resorbed, are carried over into the anterior daughter, and the posterior daughter gets the new organelles. The axostyle surrounds the nucleus in such a manner that the nucleus probably could not divide if this organelle remained intact. So far as I know, the extranuclear organ- dies of flagellates are either resorbed and each daughter develops a new set of organelles or the parent organelles are distributed among each daughter. S. bispira shows a radical departure from either procedure, and one wonders if this is not the beginning of a new stage of evolution in flagellates. After cytoplasmic division, the posterior daughter, which is considerably smaller than the anterior one, continues the development of its extranuclear organelles. The axostyle increases in length and breadth and the flagellar band extends posteriori},', forming spirals as it does so, although the spirals for some time are not perfectly arranged (Fig. 59). They soon become arranged more or less perfectly (Fig. 60), and presently another flagellar band begins to develop from the point of origin of the existing or first new band. This hand is referred to as the second new band. It extends posteriorly and soon takes its place along the first new band, so that the posterior daughter cell now has two spiral flagellar bands (Fig. 61). These bands increase in length and finally become as long as the parent bands were prior to the beginning of the processes of cell division. As the bands increase in length, they form more flagella and parabasals, the axostyle becomes grown, and the posterior daughter is now indistinguishable from the parent cell from which it originated (Fig. 2). Meanwhile, the axostyle LONGITUDINAL AND TRANSVERSE DIVISION 17 of the anterior daughter completes its growth, and it, too, becomes indistinguishable from the parent cell from which it originated. REFERENCES CITED CLEVELAND, L. R., 1935. The centrioles of Pseudotrichonympha and their role in mitosis. Biol. Bull., 69: 46. CLEVELAND, L. R., S. R. HALL, E. P. SANDERS, AND J. COLLIER, 1934. The wood- feeding roach Cryptocercus, its protozoa, and the symbiosis between pro- tozoa and roach. Mem. Acad. Arts and Sci., 17: 185. CUPP, E. E., 1930. Spirotrichonympha polygyra sp. nov. from Neotermes simplici- cornis Banks. Univ. Calif. Publ. Zool., 33: 351. 18 L. R. CLEVELAND EXPLANATION OF PLATES The drawings were made with the aid of a camera lucida from material fixed either in Schaudinn's or Flemming's fluids and stained with haematoxylin. PLATE 1 FIG. 1. Spirotrichonympha polygyra. Entire organism. Note four spiral flagellar bands. The flagella leave the bands in groups of two, and only two groups are shown from each band. The two adhering flagella extend posteriorly from their point of origin almost to the next band; then they turn and continue somewhat posteriorly to the surface of the body where they separate. The spherical parabasals follow the bands from a point just posterior to the nucleus to their termination. Other bodies, smaller than the parabasals, follow the bands from midway between the anterior end and nucleus to their termination. The axostyle extends from the anterior end around the nucleus to the posterior end of the body, terminating in a fine point slightly beyond the body. X 1600. PLATE 2 FIG. 2. Spirotrichonympha bispira sp. nov. Entire organism. The extra- nuclear organelles differ from those of S. polygyra only in that there are two instead of four spiral flagellar bands. X 1600. PLATE 3 Spirotrichonympha polygyra The nuclei are mounted so that their anterior surfaces are directed toward the top of the plate. FIG. 3. Interphase nucleus. There are four chromosomes. Each chromosome has a single coil lying within a sheath and is anchored to the anterior surface of the nuclear membrane by an intranuclear chromosomal fibre. The posterior end of each chromosome is attached to a portion of the chromatin nucleolus. There are two long and two short daughter chromosomes. The daughters of each group are some- what coiled around each other. X 3800. FIG. 4. Interphase nucleus. Both pairs of daughter chromosomes are attached to the chromatin nucleolus at their posterior ends; anteriorly they are attached to the nuclear membrane as explained in Fig. 3. X 3800. 1- IG. 5. Interphase nucleus. Here the pairs of daughter chromosomes are con- siderably coiled around each other, the long pair lies over the short pair, and the posterior end of each chromosome is attached to a portion of the chromatin nucleolus. X 2400. FIG. 6. Interphase nucleus more heavily stained than the three previous nuclei, so that other material in the nucleus besides the chromosomes stains and thus obscures slightly the arrangement of the chromosomes and their connections with portions of the chromatin nucleolus. X 3800. FIG. 7. Interphase nucleus so heavily stained that chromosomal details cannot be seen. X 2400. FIG. 8. Prophase nucleus. The daughter chromosomes arc no longer coiled around each other. The coils within each chromosomal sheath are plain and their direction may be followed clearly. X 3800. FIG. 9. Rather heavily stained nucleus. Each pair of chromosomes is short and stubby, the turns made by the coils within the sheaths are broader and closer together, and the daughter chromosomes are beginning to move apart. The chromo- somal coils here and until the late telophase are not so distinct as in the interphase and prophase. The attachments between chromosomes and port ions of the chromatin nucleolus will be lost presently, not to be renewed until the very late telophase. X 3800. LONGITUDINAL AND TRANSVERSE DIVISION 19 FIG. 10. The greatly shortened chromosomes have moved apart and a large one and a small one are preparing to line up on opposite sides of the nucleus. X 3800. FIG. 11. Nucleus has elongated. One large and one small chromosome have been moved to the left side of the nucleus and the other two chromosomes are in the process of being lined up in the same manner. X 3800. FIG. 12. The nuclear membrane has pulled in two and each daughter nucleus has a pair of dimorphic chromosomes. X 2400. FIG. 13. Daughter telophase nucleus. Chromosomes are beginning to elongate and the coils within their sheaths are again distinct. X 2400. FIG. 14. Daughter telophase nucleus. The chromosomal coils are in the process of duplicating themselves by longitudinal division. X 2400. FIG. 15. Daughter telophase nucleus. The chromosomal coils have divided and the chromosomal sheaths have almost completed the process of longitudinal division. Connections are being made between chromosomes and portions of the chromatin nucleolus. X 2400. PLATE 4 Spirotrichonympha bispira The nuclei are mounted so that their anterior surfaces are directed toward the top of the plate. FIG. 16. Interphase nucleus. Note a short and a long chromosomal sheath, each containing two chromosomal coils attached to the anterior margin of the nuclear membrane by an intranuclear chromosomal fibre. The posterior end of each sheath is attached to a portion of the chromatin nucleolus. X 3000. FIG. 17. Prophase. The chromosomal sheaths are dividing longitudinally and each sheath contains a chromosomal coil attached anteriorly to the nuclear membrane and posteriorly to the chromatin nucleolus. The daughter chromosomes are coiled around each other. X 3000. FIG. 18. Prophase. Daughter chromosomes arranged somewhat differently from preceding stage, the anterior attachments of the long pair having moved pos- teriorly, and each pair is attached posteriorly to a portion of the chromatin nucleolus. X 3000. FIG. 19. Same as the two preceding stages except that each daughter chromo- some is in the process of becoming attached to a single portion of the chromatin nucleolus. X 3000. FIG. 20. Chromosomes have become short and stubby preparatory to poleward movement. X 4800. FIG. 21. Early anaphase. The anterior ends of the long pair of chromosomes have made some progress in their poleward movement, one chromosome moving anteriorly, the other posteriorly. The short pair, as is sometimes the case, has been moved almost to the posterior end of the nucleus, although the chromosomes have not begun to separate. This was brought about by an extranuclear chromosomal fibre from the posterior pole (centriole) becoming attached to the intranuclear chro- mosomal fibre of the chromosome destined to go to the posterior pole. This fibre carried both chromosomes with it because the two chromosomes were connected. X 4800. FIG. 22. A slightly later stage. The anterior ends of each pair of chromosomes have made considerable progress in their poleward movement, while the posterior ends are still connected and have not moved poleward. The short pair lies to the left and separation has progressed farther than in the pair which lies to the right. Note the evagination of the nuclear membrane at the four points where the intra- nuclear chromosomal fibres connect the chromosomes with the nuclear membrane. The evaginations are produced at these points by the pull of the extranuclear chro- mosomal fibres. X 3000. 20 L. R. CLEVELAND FIG. 23. The chromosomes have turned completely around, a short one and a long one being directed toward each pole (anteriorly and posteriorly). Practically- all of the connections between the daughter chromosomes have been broken. X 3000. FIG. 24. The chromosomes have moved farther apart and the nucleus has be- gun to elongate. X 4800. FIG. 25. The chromosomes have moved still farther apart and have begun to elongate. The nuclear membrane is constricting preparatory to being pulled in two. One intranuclear chromosomal fibre has been duplicated. X 3000. FIG. 26. A later stage. Nucleus greatly elongated and will soon be pulled in two. Chromosomal coils in anterior pair are in the process of duplicating themselves by longitudinal division. X 3000. FIG. 27. Posterior daughter telophase nucleus. The intranuclear chromosomal fibres and the chromosomal coils have been duplicated. The chromosomal sheaths, while slightly constricted, have not divided. X 4800. PLATE 5 Spirotrichonympha polygyra FIG. 28. Early stage in the separation of the four spiral flagellar bands into two groups. As the bands separate, they unwind. A centriole follows one band of each group from the anterior end to the posterior termination of the achromatic figure. Astral rays have arisen from the distal ends of each centriole; some of these rays have met and formed a portion of the central spindle; others, the posterior ones, are in the process of meeting to form more central spindle. X 1500. FIG. 29. About the same stage as the previous one, but a different view of the two groups of separating and unwinding bands; anteriorly, the astral rays have met to form a portion of the central spindle; posteriorly, they have not met. X 1500. FIG. 30. A slightly later stage in the separation of the bands and the formation of the central spindle portion of the achromatic figure. The astral rays have met both anteriorly and posteriorly to form the broad, flat central spindle. Note position of nucleus in this and in the two previous illustrations. X 1500. FIG. 31. Later stage. Central spindle is longer, nearer the nucleus, bands are farther apart, and the unwinding of the bands has progressed considerably. X 1500. FIG. 32. Still later stage. Central spindle is longer and lies over the nucleus. The bands which the centrioles follow are separated from the other bands. At the ends of the central spindle astral rays are arising from the centrioles and flagella are arising from the bands, but the flagella have been omitted and the astral rays have been omitted except those that joined to form the central spindle. X 1500. FIG. 33. A much later stage. Nucleus has divided and daughter nuclei have moved a considerable distance apart; many of the astral rays have pulled apart so that the central spindle is now long and narrow. Note the manner in which the bands are unwinding. X 750. FIG. 33 A. Detail of the anterior end of Fig. 33. Note the relation of the ends of the central spindle to the bands. X 1500. PLATE 6 Spirotricliotiympha polygyra Fu;. 34. Vertical view of a stage slightly later than Fig. 32. Note relation of central spindle to flagellar bands. The flagella are not drawn full length so as to •id confusing them with the fibres of the central spindle. The astral rays that do not join to form the central spindle cannot be differentiated from flagella when the LONGITUDINAL AND TRANSVERSE DIVISION 21 two arise from the same point on the centrioles and flagellar bands. This is a six- banded individual. X 1500. FIG. 35. An intermediate stage between Figs. 32 and 33 drawn to show that the ends of the central spindle are not flat but semicircular when the centrioles (and flagellar bands) from which the central spindle arises are not straight (as in Figs. 31, 32). In other words, the end of the central spindle has the same shape as the centriole (and band) from which the astral rays composing the central spindle arise. X1500. FIG. 36. A late stage in the unwinding of the flagellar bands. Achromatic figure has disappeared and daughter nuclei are far apart. New daughter axostyles are growing out and have extended a short distance beyond the nuclei posteriorly. X 1200. FIG. 37. Elongate organism about to divide longitudinally. The unwinding of the parent bands is complete and new daughter bands will grow out presently. X 1200. PLATE 7 Spirotrichonyinpha bispira The flagella arising from the two parent or old bands and those arising from the new or free band are not drawn as long as they are. They are stopped where they leave the body. Explanation of labels used in Plates 7-10. a.e. — Anterior end (of new or free flagellar band). a.r. — Astral rays. a.r.f.c. — Astral rays from free centriole. a.r.s.c. — Astral rays from stationary centriole. c. — Centriole. c.n.f.b. — Centriole of (or following) new flagellar band. c.p.f.b. — Centriole of (or following) parent flagellar band. c.s. — Central spindle. e.e. — Elongating end (of new or free flagellar band). /. — Flagella. p.f.b. — Parent flagellar band. FIG. 38. Anterior end of a cell in the earliest stage of division showing the twro parent flagellar bands (p.f.b.), the new flagellar band and its centriole (c.n.f.b.), and flagella (/.), arising from the parent bands and the new band. The free centriole, which follows the new band from its point of origin in the anterior tip of the cell, has not begun to produce astral rays; nor has the stationary centriole which follows one of the parent bands. X 1500. FIG. 39. A slightly later stage. The elongating end (e.e.) of the new band has begun to extend laterally, while the anterior end (a.e.) still remains at the anterior tip of the cell. X 1500. FIG. 40. Later stage. The new band has begun to migrate posteriorly, its anterior end (a.e.) lying to the left and its elongating end (e.e.) to the right. The free centriole following the new band has given off astral rays which have joined some of those given off by the stationary centriole following one of the parent bands to form the very small, early central spindle (c.s.). In this instance, the centriole (c.) fol- lowing one of the parent bands does not follow the band for its entire length, the distal portion being free. X 1500. FIG. 41. The situation here is the same as in Fig. 40 except the new flagellar band is longer, has migrated farther posteriorly, the central spindle (c.s.) is longer and wider, and the stationary centriole is free of its band for a greater distance an- teriorly. X 1500. L. R. CLEVELAND FIG. 42. Later stage. The new band, especially the elongating end (e.e.) has migrated farther posteriorly and now lies near the nucleus. The central spindle (c.s.) is longer. The distal portion of the stationary centriole (c.) is free of the band. X 1500. FIG. 43. A considerably later stage in the development of the achromatic figure and the posterior migration of the new flagellar band. The central spindle (c.s.) is in two portions, one inside the parent flagellar bands and one outside. The portion inside the bands extends over the nucleus. No portion of the stationary centriole is free of the band. The shape of the end of the central spindle arising from the centriole following one of the parent bands is the same as the band and centriole from which it arises, semicircular. The other end of the central spindle is irregular in shape due to the position of the centriole from which it arises. Numerous astral rays (a.r.) extend toward the nucleus from the centriole that follows the new band. X 1500. FIG. 44. Approximately the same stage as illustrated by Fig. 42, but different in appearance because of the position of the new band and its centriole. The end of the central spindle (c.s.) adjacent to the centriole of the new band is cylindrical or nearly so and no portion of the centriole of the parent flagellar band is free. X 1500.' FIG. 45. About the same stage as Fig. 44, but the achromatic figure presents a very different appearance because the new flagellar band and its centriole lie near the periphery of the cell and extend antero-posteriorly in almost a straight line, thus giving the posterior end of the central spindle a greatly flattened appearance. Note the length of the functioning portion of the centriole of the new band. X 1500. FIG. 46. The achromatic figure here presents a strikingly different appearance from that of any of the previous illustrations, because many of the astral rays arising from the stationary centriole have extended posteriorly between the spirals of the parent flagellar bands instead of making their way toward the periphery, as a few of them have done, to join the astral rays arising from the free centriole (the one following the new flagellar band). Note astral rays extending from the free centriole toward the nucleus. When these rays join those arising from the other centriole and extending posteriorly within the spirals, another portion of the central spindle will be formed. This portion, like that illustrated in Fig. 43, will lie mostly within the flagellar bands. X 1500. PLATE 8 Spirotrichonympha bispira The flagella arising from the two parents or old bands and those arising from the new or free band are not drawn as long as they are. They are stopped where they leave the body. See explanation of Plate 7 for meaning of labels employed. FIG. 47. The central spindle (c.s.) lies almost entirely outside the parent flagellar bands. Its posterior portion, which arises from the centriole of the new flagellar band, is broad and flat because the new band extends in almost a straight line, while its anterior portion, which arises from the centriole of one of the parent flagellar bands, is semicircular. X 1500. FIG. 48. Here there are two portions of the central spindle (c.s.), one inside and one outside the spirals of the flagellar bands. The centriole (c.) following one of the parent flagellar bands is free of the band for a short distance. X 1500. FIG. 40. The central spindle (r.s.) lies within the spirals of the flagellar bands except for the posterior portion which is posterior to the bands. Astral rays arise from both the stationary and free centriole and extend toward the nucleus. Some of these rays have joined the intranuclear chromosomal fibres in the nuclear membrane and by so doing have become extranuclear chromosomal fibres which are pulling the chromosomes toward the poles, the centriole of the parent flagellar band (c.p.f.b.) and the centriole of the new flagellar band (c.n.f.b.). X 1500. LONGITUDINAL AND TRANSVERSE DIVISION PLATE 9 Spirotrichonympha bispira The flagella arising from the two parent or old bands and those arising from the new or free band are not drawn as long as they are. They are stopped where they leave the body. See explanation of Plate 7 for meaning of labels employed. FIG. 50. Note the manner in which the astral rays composing the anterior half of the central spindle arise from the distal end of the centriole following one of the parent flagellar bands. The central spindle is almost cylindrical at this end because of the spiraling of the portion of the centriole from which it arises, while the opposite end is almost flat. There are probably astral rays arising from the centriole of the new band and extending toward the nucleus but, since they cannot be distinguished with certainty from the flagella which are numerous in this area, they are not drawn. X 1500. FIG. 51. This central spindle is almost cylindrical anteriorly, flat posteriorly, and twisted in the mid-region. The astral rays that have become extranuclear chromosomal fibres by attaching themselves to the intranuclear chromosomal fibres in the nuclear membrane, as well as the free astral rays, are intermingled with the fibres of the central spindle and can scarcely be distinguished from them, thus presenting a situation closely resembling that of many metazoan mitoses. X 1500. FIG. 52. In this cell the new flagellar band and its centriole failed to migrate posteriorly so that the astral rays arising from the free centriole (a.r.f.c.) could not meet and join those arising from the stationary centriole (a.r.s.c.) to form a central spindle. X 1500. PLATE 10 Spirotrichonympha bispira The flagella arising from the two parent or old bands and those arising from the new or free band, except in Fig. 54, are not drawn as long as they are. They are stopped where they leave the body. See explanation of Plate 7 for meaning of labels employed. FIG. 53. Another cell where the free band and its centriole failed to migrate posteriorly and the astral rays arising from the free centriole (a.r.f.c.) could not meet and join those arising from the stationary centriole (a.r.s.c.} to form a central spindle. This would be a telophase if the proper formation of the achromatic figure had not been prevented by the position of the centrioles, for the chromosomes are telophasic in structure. X 1500. FIG. 54. Many of the astral rays arising from the stationary centriole have failed to meet those arising from the free centriole and only a small central spindle connects the two centrioles. The movement of the chromosomes and the division of the nucleus have been upset. The flagella of the new band are drawn full length, but only one flagellum is shown beyond the point of bifurcation. X 1500. FIG. 55. Nucleus is being pulled in two and the elongate central spindle is pulling apart. X 1500. FIG. 56. Nucleus has divided, achromatic figure has disappeared, chromosomes have elongated, new axostyles are developing, and the new or free flagellar band has migrated to the extreme posterior end of the parent cell. X 1500. PLATE 11 Spirotrichonympha bispira The flagella arising from the two parent or old bands and those arising from the first and second new bands are not shown as long as they are. They are stopped where they leave the body. 24 L. R. CLEVELAND FIG. 57. The two parent flagellar bands, as in previous illustrations of cell division of 5. bispira, remain intact. The new flagellar band which, although it has assumed various shapes in earlier stages, has not formed spirals, has now begun to spiral at its anterior end, which lies at the posterior end of the cell. The shape of the posterior end of the cell has changed, too. X 1500. FIG. 58. Later stage. Cell is more elongate, posterior end is pointed more like the anterior end, and all of the new flagellar band has formed spirals. This stage is just before the transverse division of the cytoplasm to form two daughter cells; one, the anterior daughter, getting the two parent flagellar bands and their associated organelles; the other, the posterior daughter which is smaller, getting the new flagellar band and its associated organelles. Both daughters form new axostyles. X 1500. FIG. 59. A posterior daughter cell after cytoplasmic division. The spirals of what was the new flagellar band before division of the body are little if any better arranged now. This organism has just separated from the anterior daughter. X 1500. FIG. 60. A slightly later stage in the post -cytoplasmic development of a pos- terior daughter cell. The spirals of the flagellar band have perfected their arrange- ment. X 1500. FIG. 61. A considerably later stage. A second new flagellar band has now appeared and is co-extensive with the first new band, the two forming spirals in the same manner as those of the parent cell previous to cell division. When these bands increase in length posteriorly (forming spirals as they do), the axostyle extends pos- teriorly, and the body increases in size, this cell will be like the parent which pro- duced'it. X 1500. HK)I.(X,ICAL I'.I LLETIX PLATE 1 Onrolhv <». Harris del. I.. I! BIOLOGICAL I'.i LI.KTIX PLATE 2 Dorothy Ci. Harris drl. I I! Cli-\cl:in,l BIOLOGICAL HI LLKT1X PLATK 8 15 Dorothy G. Harris del. L. R. Cleveland BIOLOGICAL BULLETIN PI, ATK 4 25 26 Dorothy (j. Harris del. L. K. Cleveland BIOLOGICAL BULLETIN PLATE 5 26 30 29 31 33 Dorothy G. Harris del. I.. H. Cleveland BIOLOGICAL BULLETIN PLATE 6 34 35 37 36 Dorothy G. Harris del. L. R. Cleveland BIOLOGICAL BULLETIN PLATE 7 42 -V-c.rxfb 46 Of. Dorothy G. Harris del. I.. K. Cleveland BIOLOGICAL BULLETIN PLATE 8 — c.p.p.b. c.s. 49 Dorothy G. Harris del. L. K. Cleveland HIOLOGICAL BULLETIN PLATE 9 C.n.fb.-, Q.r. — /- -z. Dorolhy G. Harris del. L. R. Cleveland BIOLOGICAL BULLETIN PLATE 10 Dorothy G. Harris del. L. R. Cleveland BULLETIN PLATE 11 57 Bl 58 59 Dorothy O. Harris del. I.. K. Cleveland ORIGIN AND DEVELOPMENT OF THE ACHROMATIC FIGURE L. R. CLEVELAND (From Ike Biological Laboratories, Harvard University, Cambridge, Massachusetts) The question of the origin and nature of the achromatic figure has been debated for more than fifty years, and its existence has been seriously questioned many times, particularly during recent years. Several reasons have been advanced for regarding it as an artifact, the principal one being the inability to demonstrate clearly its origin, development, and function in living cells. I have studied the achromatic figure in twenty genera of hyper- mastigote and ten genera of polymastigote flagellates. These organ- isms furnish a diversity of favorable material. In some of them, there is nc question regarding the manner in which the achromatic figure is formed from the centrioles and the reality of the fibres composing it; in others, the development of the achromatic figure is not so plain, and can be determined with certainty only in view of their relation to those organisms where the process is unmistakably clear. There are gradations between the two groups. In several genera, the develop- ment of the achromatic figure has been followed in living cells from its very beginning to its disappearance in the late telophase; and the picture presented is exactly the same as in fixed cells. In one genus, Barbulanympha, considerable time has been devoted to the origin and development of the achromatic figure when more than two centrioles are concerned in the process, and when only one is concerned ; because the process is so very plain under these conditions that no one can question it. In a previous paper (Cleveland, Hall, Sanders, and Collier, 1934. Mem. Am. Acad. Arts and Sci., 17: 185) Barbulanympha was described, and an account was given of the manner in which the two elongate, interphase centrioles produce daughter or new centrioles like themselves, a process which need not be considered here; but it is desirable, before considering the achro- matic figure when an abnormal number of centrioles are present, to consider it when two are present. The centrioles vary in length in the four species of Barbulanympha from 15 to 30 microns; in the unstained living cell they are of a dense hyaline nature, and may be differentiated easily from other cellular contents; when fixed in Schaudinn's fluid and stained with Heiden- 41 42 L. R. CLEVELAND hain's haematoxylin, they stain and destain in about the same manner as chromatin ; they are joined at their anterior ends by a desmose; and their distal ends, which lie 20 to 30 microns apart, are free (Fig. 1). The distal end of each centriole is surrounded by a spherical cen- trosome, which is 4 to 6 microns in diameter, and which moves with the centriole whenever the latter is moved, either by the natural movement of the cell or by mechanical manipulation. When the nucleus and centrioles are suitably destained for study, the centro- comes retain little or no stain. In the interphase cell, there are no fibres extending from either sentriole (Fig. 1). In the prophase, astral rays begin to grow out from the distal end of each centriole (Fig. 2). At first, the rays arising from one centriole are a considerable distance from those arising from the other; but, as they increase in length, the two sets of astral rays soon meet (Fig. 3), and, as they meet, the individual rays or fibres join, grow along one another and overlap to form the early central spindle portion of the achromatic figure (Fig. 4). That the astral rays arise from the centriole instead of the centrosome may be shown by observation of living cells; the rays may be traced through the cen- trosome to the centriole. The same may be demonstrated also in fixed cells. And further proof that the centrosome plays no part in the production of the astral rays may be obtained from a study of several other genera with elongate centrioles similar to those of BarbulanympJia but with no centrosomes surrounding their distal ends; in these, the connection of the astral rays to the centrioles may be seen at a glance. Those genera with centrosomes surrounding the distal ends of their centrioles form cylindrical central spindles, while those with no centrosomes form flat ones; and thus the role of the centrosome in directing the astral rays that form the central spindle may be seen. As the astral rays continue to increase in length, more of those arising from one centriole meet those arising from the other, and the central spindle becomes larger; and, at the same time, the central spindle increases in length, because the centrioles from which it arises move in opposite directions (Fig. 5). Those astral rays which do not participate in the formation of the central spindle, because they do not meet and overlap, also increase in length; at the same time, more astral rays arise from each centriole, so that the individual rays vary greatly in length; and it is not long after the central spindle is formed before some of the astral rays which do not participate in its formation become extranuclear chromosomal fibres by joining the intranuclear chromo- somal fibres, the fibres which anchor the chromosomes to the ever intact nuclear membrane (Fig. 6). Thus, the achromatic figure ORIGIN OF ACHROMATIC FIGURE 43 consists of astral rays which are joined together to form the central spindle, astral rays which are connected to the chromosomes and are responsible for their movement to the poles, and astral rays which merely radiate in the cytoplasm and, so far as known, perform no function. All of these fibres appear alike structurally, and the only justification for naming them is their functional differences. When three centrioles are present in the interphase (Fig. 7), the type of achromatic figure produced by them during cell division varies considerably, depending on their position. If the distal ends of two of the centrioles lie fairly close together, as illustrated in Fig. 8, two long central spindles and one short one are produced. In this instance, the centriole which lies somewhat apart functions jointly with each of the other centrioles in the production of the two long central spindles; and the two centrioles which lie fairly close together function jointly in the production of a short central spindle between them. In other words, the astral rays arising from the centriole which lies more or less alone, and which may be termed the remote centriole, meet, join, and overlap those arising from the two centrioles which lie fairly close together, and thus the remote centriole functions as much again as each of the other centrioles in the formation of the two long central spindles, because it supplies as many astral rays for central spindle production as both of the other centrioles combined. If, however, as is frequently seen, two centrioles lie adjacent and one apart, the three centrioles function in the production of one central spindle, which has the same appearance as when only two centrioles are functioning in the production of a central spindle (for example, as in Fig. 5). The achromatic figure appears exactly as that of Fig. 8 would appear if the two centrioles on the right were pulled together so that one centrosome touched the other. If the distal ends of the three centrioles are more or less equidistant, i.e., at the apices of an equilateral triangle, each centriole functions in approximately the same manner in the production of the achromatic figure; three central spindles, more or less equal in size, are produced, and each centriole functions in conjunction with two other centrioles in their formation (Fig. 10). If, on the other hand, the three centrioles are not equidistant, but lie at the apices of a triangle which is wider on one side, only two central spindles are produced — at least at first; a third may be formed later, as the astral rays become longer, if the centrioles between which a central spindle failed to form earlier do not lie too far apart (Fig. 9). Three is the maximum number of central spindles that can be produced from three centrioles. Many cells with four centrioles have been studied. The four inter- 44 L. R. CLEVELAND phase centrioles may all he mature, i.e., fully grown, or some of them may be in the process of development, as shown in Fig. 11 ; but they do not function until growth is completed; and when they function, the type of achromatic figure produced by them depends on their position. If they lie side by side in groups of two, one central spindle is produced, although all four centrioles participate in its formation; if they lie at the four corners of a square, six central spindles are pro- duced; if they lie at the corners of a rectangle, four central spindles are usually produced; if they lie fairly close together, but not adjacent, in groups of two, two central spindles are usually produced in the early stages of cell division, and in the later stages, if the centrioles of each group move apart, a central spindle is formed between the centrioles of each group. Sometimes, the four centrioles lie in the same plane in groups of two in such a manner that a criss-crossing occurs in the astral rays forming the central spindles. Figure 13 illustrates an example of this. Here astral rays arising from the upper centriole on the right have joined astral rays arising from both of the centrioles on the left, so that this centriole is connected by a central spindle to each of the centrioles on the left; and the astral rays arising from the lower centriole on the right have joined astral rays arising from both of the centrioles on the left, so that this centriole is also connected by a central spindle to each of the centrioles on the left. In brief, each of the four centrioles is connected to two other centrioles by a central spindle. Would such a series of connections occur as a result of "lines of force"? Other instances have been seen where, for example, the lower centriole on the right was connected to each centriole on the left by a central spindle, while the upper centriole on the right was connected only to the upper centriole on the left by a central spindle. And other vari- ations have been seen in the criss-cross central spindle interconnections of centrioles. When five non-adjacent centrioles are present, eight central spindles could be formed by an overlapping of the astral rays arising from them, although six, as shown in Fig. 12, is the usual number. Here each of the two upper centrioles is functioning in the formation of three central spindles; the lower right and left centrioles are each functioning in the formation of two central spindles; and the lower middle centriole is functioning in the formation of four central spindles, most of its astral rays being used to form central spindles. Many examples of a still larger number of centrioles functioning in the production of multiple achromatic figures have been studied, but only one has been illustrated; this is Fig. 16 where fifteen centrioles are in order to avoid confusion only the centrosomes are drawn). ORIGIN OF ACHROMATIC FIGURE 45 Here there is great variation in the length and width of the central spindles, resulting from differences in the time of their formation and in the number of fibres composing them. There are also many in- stances of the criss-crossing of the fibres of the central spindles; instead of the astral rays arising from one centriole joining those from another, they join those from several centrioles. Few of the central spindles are straight, that is, extend directly from one centriole to another; and it is difficult, owing to their curvation and the manner in which they are crowded together, to determine the number present. When two centrioles are present but are so widely separated that the astral rays arising from one cannot meet those arising from the other, the central spindle portion of the achromatic figure is not formed (Fig. 14). Similarly, when only one centriole is present, the development of the achromatic figure is incomplete; many astral rays extend to and around the nucleus, but they do not form a central spindle (Fig. 15). In a few instances where several interphase centrioles are present, decidedly abnormal conditions have been seen; the formation of cen- trosomes surrounding the distal ends of the centrioles is either com- pletely or partially upset, and strands, varying in size from a little larger than astral rays to half the size of the centrioles, extend pos- teriorly from the centrioles (Fig. 17). Just what these strands are is not known. They can scarcely be regarded as astral rays, or even a premature aberrant attempt to produce an achromatic figure; but they may be secondary extensions from the centrioles. I am indebted to the Penrose Fund of the American Philosophical Society for financial assistance. BIOLOGICAL BULLETIN PLATE 1 2 ORIGIN OF ACHROMATIC FIGURE 47 EXPLANATION OF PLATES All figures were drawn with the aid of a camera lucida from material fixed in Schaudinn's fluid and stained with Heidenhain's haematoxylin. The organism is Barbulanymplia. PLATE 1 FIG. 1. The two elongate interphase centrioles with centrosomes surrounding their distal ends. X 1400. FIG. 2. Prophase centrioles with astral rays arising from their distal ends. X 1400. FIG. 3. The astral rays arising from one centriole are meeting those arising from the other, preparatory to the formation of the central spindle. X 1400. FIG. 4. The astral rays have met, joined, and grown along one another to form the early central spindle. X 1400. FIG. 5. More astral rays have joined and the formation of the central spindle is complete. X 1400. 48 L. R. CLEVELAND PLATE 2 FIG. 6. Longitudinal section of anterior end of cell showing the interrelation of flagellated areas, centrioles, achromatic figure, and chromosomes. Some of the astral rays which do not join and overlap in the formation of the central spindle have become extranuclear chromosomal fibres by connecting with the intranuclear chromo- somal fibres in the nuclear membrane. X 1400. FIG. 7. Three interphase centrioles and centrosomes. X 1400. FIG. 8. The type of achromatic figure produced by three centrioles when they lie in this position. X 1400. BIOLOGICAL BULLETIN PLATE 2 8 BIOLOGICAL BULLETIN PLATE 3 10 Dorothy G. Harris del. 1,. K. Cleveland ORIGIN OF ACHROMATIC FIGURE 51 PLATE 3 FIG. 9. Three centrioles in another position and the achromatic figure produced by them. X 1600. FIG. 10. Still another position of three centrioles and the achromatic figure produced by them. X 1600. 52 L. R. CLEVELAND PLATE 4 FIG. 11. Four interphase centrioles, one mature and three immature. X 1400. FIG. 12. Five centrioles and their achromatic figure. X 1400. FIG. 13. Four centrioles and their achromatic figure. Note the criss-crossing of the astral rays forming the central spindle. X 1400. FIG. 14. Two centrioles so far apart that their astral rays could not meet to form the central spindle portion of the achromatic figure. X 600. BIOLOGICAL BULLETIN PLAT I I \ 14 BIOLOGICAL BULLETIN PLATE 5 15 Dorolhy G. Harris del. L. K. Cleveland ORIGIN OF ACHROMATIC FIGURE PLATE 5 FIG. 15. The type of achromatic figure produced when only one centriole is present. Note the absence of a central spindle. X 1400. FIG. 16. Fifteen centiioles are functioning in the production of an achromatic figure. In order to avoid confusion, the centrioles are omitted; but the position of their distal ends is indicated by the position of the centrosomes. X 1400. FIG. 17. Interphase centrioles with probable secondary posterior extensions. X 1400. THE ADSORPTION OF BACTERIA BY MARINE BOTTOM ' SELMAN A. \VAKSMAN AND UNTO VARTIOVAARA (From the Woods Hole Oceanographic Institution and New Jersey Agricultural Experiment Station) In various recent studies (1, 4, 5) on the adsorption of bacteria in soil, the impression is left that the bacteria exert, in that state, only little effect upon the cycle of life in the soil. In his book on Agricul- tural Microbiology, Chudiakov (2) states emphatically that the prob- lem of the condition of the bacteria in the soil hampered considerably a better understanding of soil microbiological processes. The existence of a surface relationship between the soil and microorganisms is true not only of bacteria but also of Protozoa, as shown by Cutler (3). Different species of bacteria were found to behave differently in regard to their adsorption by the soil particles; their mobility and ability to form zooglea seemed to be of special significance in this connection. The chemical activities of the bacteria, as measured by the evolution of CO2, were considerably modified by the adsorption process (4). The finer soil constituents, namely the clay and silt fractions, were found to have a much greater adsorptive effect than the sand. Different soil types were found to adsorb bacteria to a different degree (5). Rubentschik (8) made a detailed study of the phenomenon of bacterial adsorption in salt basins; the bacteria isolated from the mud were found to show a higher degree of adsorption than the bacteria found in the water. The conclusion was reached that the bacterial benthos consist of easily adsorbable species whereas the bacterial plankton contains organisms possessing a low degree of adsorption by the bottom sediments. A bacterial exchange took place in the mud, the adsorption of some species being accompanied by the desorption of others from the mud. The relative concentration of the bacteria, their nature and the type of bottom material exerted an influence upon the adsorption process. The activities of different species of bacteria were variously modified in the adsorbed state: the metabolism of some was lowered and that of others was increased. ' ( Ontriliution No. 161 of the Woods Hole OceanoRraphic Institution and J<>urn;il Scries Paper of the New Jersey Agricultural Experiment Station, Department of Soil Microbiology. 56 ADSORPTION OF BACTERIA BY MARINE BOTTOM 57 According to Peele (6), the adsorption of bacteria is probably due to the attraction of unlike electric charges. The nature of the base in the soil complex influences considerably the process of adsorption, the monovalent cations showing the least adsorption. In most of these studies, pure cultures of bacteria were used and the results obtained interpreted in terms of processes carried out by a complex population of microorganisms inhabiting the soil or the sea bottom. Usually a very short period of contact was allowed between the soil and the bacterial culture grown upon an artificial medium; no attempt was made to determine what happened after the culture had adjusted itself to the environment. These studies have an important bearing upon the influence of the sea bottom upon the bacterial activities in the water. A number of questions may arise in this connection; 1. Is the specific occurrence of certain bacteria in the bottom due to their adsorption by the bottom material? 2. Is the relatively low number of bacteria in the sea water, as compared with that in the bottom, due to their removal from the water by the adsorption process? 3. Is this relationship responsible for the difference in the rate of bacterial processes taking place in shallow seas over a sand bottom such as Georges Bank, compared with the corresponding processes over a mud bottom, such as the Gulf of Maine (7, 11)? These and other questions deserve fundamental treatment. No attempt will be made to give a definite answer in the following experi- ments, which must be considered as preliminary in nature. EXPERIMENTAL In order to eliminate the interfering effect of the bacterial popula- tion normally found in fresh bottom material, the first experiments were carried out with marine mud and sand which had been kept in a dry state for 5 years and in which the bacteria had been reduced to very low numbers. In the first experiment, the adsorption of bacteria from a mixed population and from a pure culture of a marine bacterium by two differ- ent types of dry marine mud and sand was studied. A mixed culture was obtained by allowing fresh sea water to remain, in a glass container, in the laboratory, for a period of 48 hours. This resulted in an increase in the numbers of bacteria from a few hundred to 303,000 per 1 cc., as determined by the plate method. As a pure culture, a marine agar liquefying bacterium (No. 11) was used. This organism was selected because of the ease of recognizing the colonies produced on the plate. It had been isolated from sea water and kept in culture for a period 58 SELMAN A. WAKSMAX AXD UXTO VARTIOVAARA of 2 years. The organism was grown in a medium poor in nutrients (1 gram peptone, 1 gram glucose and 0.5 gram K2HPO4 in one liter of sea water), for 24 or 48 hours. The culture was then diluted ten times with sea water sterilized by heating 30 minutes at 80° C. The two muds, No. 1329 and No. 1331, contained 2.46 and 1.58 per cent organic carbon and 0.28 and 0.16 per cent nitrogen, respectively (9). The sand contained 0.58 per rnit carbon and 0.05 per cent nitrogen. Ten- gram portions of mud or sand were placed in 250-cc. flasks containing either 100 cc. cultured sea water or 100 cc. of the diluted 24-hour culture of the bacterium. The flasks were shaken by hand for 10 minutes, allowed to stand 10 minutes, and 1-cc. portions of the super- natant liquid plated out using a sea water agar medium (No. 1). The TABLE I of bacteria from mixed and pure cultures by dry marine mud and sand Nature of culture l Sacteria in 1 cc. water, thousand s Cultured u-dti-r Control Start 303 10 minutes 2 hours 21 hours Plus mud 1329 28 13 ' 3,700 Plus mud 1 531 •5 75 530 Plus sand 180 115 32,000 Ii :• •' s MEAN' / / 15- 20- 25- 30- 35- 40- 45- 50- 55- 60- 65- 70- 19 24 29 34 39 44 49 54 59 64 69 74 FIG. 1. Graph showing correlation between sexuality and size at first breeding season, based on an unselected sample of 751 yearlings from the culture frames at the \t-w Jersey laboratory on the Shore of Delaware Bay near Cape May. The size of the class is shown at the bottom of the graph and the number of individuals in each class at the top. The mean percentage of females in the entire sample is also indicated. ing of 751 yearlings from the culture frames at the laboratory on the shore of Delaware Bay, 358 had a shell length of 50 mm. or more on July 7, 1937, while 393 were smaller. Examination of the gonads showed a ratio of 63.4 females for each 100 males in the group of larger individuals as compared with 26.6 females for each 100 males the smaller ones. This correlation between sexuality and size tin >c 751 young oysters is shown graphically in Fig. 1. It will be noted lltal none of the females was less than 25 nun. in length and PRIMARY SEXUAL PHASES IN OVIPAROUS OYSTER 69 none of the males more than 69 mm. The largest number of indi- viduals, however, were between 40 and 55 mm. in length and in these groups there was little difference in the percentage of females. DIRECT FEMALE DEVELOPMENT The primary gonad may be either distinctly bisexual in appearance or it may become differentiated very early into either of the two sexual phases (Coe, 1932a). In the latter case there is a direct transforma- tion of the primary undifferentiated gonia into the functional sexual cells. The female phase is then attained without indication of protandry. In Long Island Sound the sexual phase of the first breeding season is sometimes distinguishable as early as October or November in ex- ceptionally well nourished young at the age of only three or four months after setting. More frequently, however, morphological sex differentiation is delayed until mid-winter or later. At Beaufort, North Carolina, direct female development appears to be the general rule, with a much higher ratio of young females than in more northern localities (Table I). Samples of well nourished individuals from Delaware Bay, known to have set July 9, 1936, like- wise indicated direct development, when examined at the end of December. No evidence of incipient protandry was found at that time, although in many cases the gonads were sexually differentiated. INDIRECT FEMALE DEVELOPMENT It has been shown previously (Coe, 1932&), that the primary gonads of some young individuals develop directly into ovaries or spermaries, as the case may be, while those of other individuals of the same age are distinctly bisexual, or, to use a less ambiguous term, ambisexual. The latter are characterized by a cortical layer of ovogonia and young ovocytes, with spermatogenic cells at one side or in the lumen. In many cases these ovocytes later degenerate or they may remain inactive during the primary functional male phase, to become activated and functional toward the end of that phase or some time thereafter. In this way the sexual phase may change from male to female either during the first breeding season or, much more frequently, during the following autumn. A subsequent change from female to male phase evidently results from the later activation of some of the descendents of the primary undifferentiated gonia into spermatogonia. The propagation of such undifferentiated gonia may be continued year after year or even through a long lifetime, since some residual gonia always remain after spawning. 70 w. R. COE It was also previously reported (Coe, 1932a) that the developing gonads exhibit all intergradations between those of the so-called true males in which no indications of ovogonia or ovocytes can be detected and those which develop directly into ovaries. In a few individuals both types of cells in the ambisexual gonad multiply harmoniously, leading to the functionally hermaphroditic condition at the breeding season . In northern localities some of the spermatogonia in the young ambisexual type of gonad may undergo transformations during the autumn comparable with those taking place in normal spermatogenesis (Coe, 1932a). These cells may later degenerate, followed by the activation of ovogonia, thus leading to the functional female phase at the first breeding season during the following summer. Such incipient protandry seems to be much less frequent than the direct development in which the initial dominant phase, either male or female, is retained during the winter and the following breeding season. SEXUAL CONDITIONS IN SOUTHERN LOCALITIES From North Carolina southward the initial sexual phase in the earliest set of the year becomes functional toward the end of the same season, when the young oyster is only three or four months of age, as Burkenroad (193 la) reported for the Louisiana oyster. Young of the later sets are similar to those of more northern localities in their sexual conditions, although protandry may be less frequent (Table I). Thus at Beaufort, North Carolina, and at Apalachicola, Florida, on the Gulf of Mexico, there may be two generations of this species in one year, individuals of the early set spawning some three to four months after setting while those of later broods do not become ripe until the follow- ing spring. Cultures in both these localities are usually contaminated with more or less numerous individuals of the larviparous species, O. equestris, which matures at a still younger age and has a breeding season covering most of the year in warm situations. It is difficult to distinguish the two species externally when young, but the relations of the epibranchial chamber differ and the gonads are very different, since 0. equestris has a sequence of overlapping male and female sexual phases and the spermatogenic cells are larger than in O. virginica and arranged in dense clusters. A small collection of the over-wintered young at Beaufort in 1933, kindly supplied by Dr. H. F. Prytherch, showed a ratio of nearly 49 ti males to 100 males, while the first set of the season in 1936, stated to have occurred about June 1, showed a female ratio of about 40 when PRIMARY SEXUAL PHASES TN OVIPAROUS OYSTER 71 examined in August, September and October. The largest individuals had a shell length of 23 mm. in August, 40 mm. in September, and 70 mm. in October. In August there were only a few males with ripe sperm and toward the end of October most individuals of the early set were spawned out. Collections of the early set taken September 20, 1937, about fifteen weeks after setting, were from 20 to 50 mm. in length and showed a female ratio of 37.09 (Table I). Those which exceeded 35 mm. in length had a female ratio of about 61 as compared with 16 for those of smaller size. An unselected sample containing 88 individuals which were one year old, originating from the autumnal set during the last two weeks in September, 1936, on the other hand, showed a ratio of 75 females to 100 functional males. These were doubtless in their second spawning season, having previously participated in the spring spawning during June, 1937, when seven to eight months of age. The ratio of females was thus about twice as great as was found in the young at the first spawning period. This would indicate that the proportion of young males changing to the female phase for their second spawning period must have been about 22 per cent greater than the number of indi- viduals, if any, which experienced a sex change in the opposite direc- tion during the same time. Small collections from Apalachicola, Florida, obtained through the courtesy of Mr. R. O. Smith, were examined three to five months after setting in 1936. These, combined with an additional collection kindly supplied by Dr. A. E. Hopkins from the same locality in 1937, showed a female ratio of 7.09 (Table I), corresponding closely to the ratios from northern localities under unfavorable conditions. DISCUSSION The foregoing evidence concerning the sexual conditions in the Virginia oyster indicates that the sex-differentiating mechanism must be in a very labile condition since it responds so generally to environ- mental influences. Because of the different sex ratios under different environmental conditions, it seems probable that the responses noted may depend upon the interaction of several associated factors, both genotypic and phenotypic. Among these it is evident that all these populations have at least the three following categories of hereditary and environ- mental influences. (1). There is an inherent tendency toward protandry, as is the case with so many other mollusks, both pelecypods and gastropods, 72 w. R. COE including the larviparous oysters (Orton, 1927; Coe, 19326), Teredo (Coe, 193632c. Mistological basis of sex changes in the American oyster (Ostrea virginica). Science, 76: 175. PRIMARY SEXUAL PHASES IN OVIPAROUS OYSTER 75 COE, W. R., 1934. Alternation of sexuality in oysters. Am. Nat., 68: 236. COE, W. R., 1936a. Environment and sex in the oviparous oyster, Ostrea virginica. Biol. Bull., 71: 353. COE, W. R., 19366. Sexual phases in Crepidula. Jour. Rxper. ZooL, 72: 455. COE, W. R., 1936c. Sex ratios and sex changes in mollusks. Mem. Mus. Hist. Nat. Belgique (2 sen), fasc. 3: 69. COE, W. R., 1936J. Sequence of functional sexual phases in Teredo. Biol. Bull. 71: 122. COE, W. R , 1938. Conditions influencing change of sex in mollusks of the genus Crepidula. Jour. Rxper. ZooL, 77: 401. GALSTOFF, P. S., 1937. Observations and experiments on sex change in the adult American oyster, "Ostrea virginica." Collecting Net, 12: 187. LOOSANOFF, VICTOR L., 1936. Sexual phases in the quohog. Science, 83: 287. LOOSANOFF, VICTOR L., 1937. Development of the primary gonad and sexual phases in Venus mercenaria Linnaeus. Biol. Bull., 72: 389. NEEDLER, ALFREDA B., 1932a. American Atlantic oysters change their sex. Prog. Kept. Atlantic Biol. Sta. and Fish. Exp. Sta., 5: 3. NEEDLER, ALFREDA B., 19326. Sex reversal in Ostrea virginica. Contr. Can. Biol. and Fish., 7: 285. ORTON, ]. H., 1927. Observations and experiments on sex-change in the European oyster (O., edulis). Jour. Mar. Biol. Ass'n. 14: 967. ORTON, ]. H., 1928. Observations on Patella vulgata. Part I. Sex-phenomena, breeding and shell-growth. Jour. Mar. Biol. Ass'n. 15: 851. ORTON, ]. H.f 1936. Observations and experiments on sex-change in the European oyster (Ostrea edulis). Pt. V, A simultaneous study of spawning in 1927 in two distinct geographical localities. Mem. Mus. Hist. Nat. Belgique (2 sen) fasc. 3:997. ORTON, ]. H. AND P. R. AWATI, 1926. Modification by habitat in the Portuguese oyster, Ostrea (Gryphaea) angulata. Jour. Mar. Biol. Assn., 14: 227. PELSENEER, PAUL, 1935. Essai d'ethologie d'apres 1'etude des Mollusques. Pub. Fond. Agathon De Potter. Acad. Roy. Belgique Cl. Sci., Bruxelles, 1-662. ROUGHLEY, T. C., 1933. The life history of the Australian oyster (Ostrea com- mercialis). Proc. Linnean Soc. New South Wales, 58: 279. THE DELAYED OCCURRENCE AND TOTAL OMISSION OF ENDOMIXIS IN SELECTED LINES OF PARAMECIUM AURELIA T. M. SONNEBORN (From tJie Department of Zoology, Johns Hopkins University) Experimental control of the process of endomixis in Paramecium aurelia has been only partly achieved. Jollos (1916), Sonneborn (19376) and others have developed methods of inducing endomixis, but no method of avoiding it completely is known. Temporary suppres- sion of endomixis was probably attained by Jollos (1916); but the method employed volumes of culture medium so great as to make it impossible to ascertain with reliability whether endomixis was occur- ring or not. The present paper sets forth a method that avoids this difficulty and provides not only some lines of descent with greatly extended interendomictic intervals, but also others that never go into endomixis. The method is based on the observation (Sonneborn, 1937a) that the interendomictic interval varies greatly in sister lines cultivated under the same conditions. An attempt was therefore made to select the lines with the longer intervals. This was done by discarding lines as they went into endomixis and replacing them by new lines begun with surplus individuals from sister lines that had not yet gone into endomixis. When this is done with a group of 12 to 24 daily isolation lines cultivated under the conditions employed by Sonneborn (1936), the following phenomena occur. At first, none of the lines go into endomixis; then, as the "normal" time for endomixis approaches, more and more of the lines go into endomixis and have to be replaced by their sister lines that have not yet gone into endomixis. During a period of several weeks, this high frequency of endomixis continues and necessitates equally frequent eliminations and replacements; but dur- ing the following few weeks the frequency of endomixis greatly de- creases and thereafter occurs but rarely. Selection has thus been effective in seeking out lines in which endomixis completely fails to occur. However, such lines cannot be continued indefinitely; after four or five months they die. This method has been successfully applied in this laboratory many timo and is now a routine technique. As illustrations of the results <•lii.iin.ible with it, two typical histories will be given in the following 76 AVOIDANCE OF ENDOMIXIS IN PAKAMKCIUM 77 vo -c M R." ._=> Q T..T ' £*f"i? ?..? -T.. R •^ ^ -r v. d | _ i; §• a . LI j-i ifv. C^ >±. ^ Q 0 t*-, TV-) "^ *O T T w-t tO >0 O L.I ^ tn V) if) \ J. T T T <-\j I. . J . .1 _T T. -TN-^ K1 W) -O^ T Q T T* T. -T T-.TT» o Q a o e, I T |._ 1 1 10*5 £JipiLr M T..T » "J T - ,0 ^^M~ UJ T I. -lii — .J I.' — l.l LJ.ll. I UJ'J> f'V-Si tO I .(-, ^o^n ^Dl — I — '- - O if, — --- "~|LUUJMJ T ^ - T T TT '•-• — ul I. I LJ I. .. UJ -r *-T-.TT H-l - .1 LU UJ T- ..J.T.I- :8l J.L u~ CO ^ — — UJ 1 . . J T- -' 1 °^ - -.1 0 .y TT T. J LJ '•& LO O 10 I loJ I — — — ••*> o — I Ivfi I..I Ld UJ LL) ~ ~ ^j ro »O uJ |-,l— »O I.-T T- ^" UJ ^ u-, in m T__| uJ m I-J T--T « — — — I T — le-o S.°S OD P- ^^ ° Ld C UJ UJ — -1 = u,^ ^>!Qs[?- I? 2 I[ oj — — _ IUIQ uj r..J~ — UJUJ *O *LduJ UJ. .2__',. ... 4 id j-i ^n ao ° •? 1- - -w '- u js e "Tj 3H g g O C3 CO *L> u c"0 o> o «*• *i r 3 o> • - c £ «J 0) _c •§ £ 1 ° O &f -_ a; Z o — oj ~U c -* S CJ -— OJ t« -4-J il J2 cu -r1 •— cj 3 # "° IS *J U5 0) if Pll 3 O "" T3 !« -f .Z 'Z & 0 u ; ^ S u C ? •" •^ « co ^ 9"S o-c ^ = '5 c 15 ,S5 In 1> •« «3 u .SH B rt 0) O co y 4_t CJ 'a o T) C O c 11 &5 ^.2 "o P t) O o> -a c c us (after 41 to 63 fissions) following the initial endomixis. TABLE I The frequency of endomixis and of death in relation to the time since the last preceding endomixis in races R and \V of Paramecium aurelia Days since preceding endomixis X umber of lines in endomixis per 100 line-days Number of deaths per 100 line-days 1-13 0 0.8 14-55 23.3 2.6 R 56-76 6.4 2.4 77-122 0.6 5.9 123-130 0 35.8 1-30 0 2.0 W 31-76 3.8 0.8 77-130 0.5 4.8 131-163 0 13.4 Each of these line's was replaced at the time of endomixis by a new line started with a surplus individual from one of the sister lines that had not yet gone into endomixis. By repeating this process of re- placement in all the descendant lim ^ \\liuirver endomixis occurred or might have occurred, some of the lines lived for as much as 130 days (350 fissions) without endomixis. As shown in Table I, the life of this group of lines fell into five periods differing markedly in the frequency ot endomixis and death. The first period, extending to the 13th day, was characterized by the complete absence of endomixis. In the second period, from the 14th to 55th days, many lines went into endomixis and had to be replaced. The third period, from the 56th to 76th day, was marked by a great reduction of the number of lines going into endomixis. In the fourth period, from the 77th to 122nd day, only three lines went into cndo- AVOIDANCE OF ENDOMIXIS IN PARAMECIUM 79 mixis, but the number that died or stopped multiplying increased. Finally, in the fifth period, from the 123rd to 130th day, no line went into endomixis, but the death rate was so high that the group com- pletely died out. In Table I, lines which stopped multiplying and were found, on staining, not to be in endomixis, are included among those which died. Experience showed that this was their usual fate. The main feature of the preceding account is that endomixis did not occur in certain lines of descent carried through 350 fissions during 130 days. This was demonstrated by cytological studies. Each line of descent was stained on every day that fission occurred; and if fission failed to occur in any line for two successive days the line was stained and so brought to an end. Since every line that showed any nuclear condition even remotely suggesting endomixis was discarded, it is certain that no endomixis was overlooked and that the lines living through till the end of the experiment had not experienced endomixis during the entire time. RACE W The method of obtaining non-endomictic lines of descent was also applied successfully to race W, though here the detailed results differ from those obtained with race R. A typical set of results on race W is illustrated by Fig. 2. The group there represented began with four individuals in endomixis on February 10, 1935. As shown in Table I, the life of this group is divisible into four distinct periods. In the first period, extending to the 30th day, there was no endomixis and few deaths occurred; in the second period, from the 31st to the 76th day, the frequency of endomixis was at its highest and the death rate was still low; in the third period, from the 77th to 130th day, there were few endomixes (two certainly, and possibly four), but the death rate showed a marked increase; in the fourth period, from the 131st to 163rd day, there were no further endomixes and the death rate reached its peak, resulting in the extermination of the group. As in race R, the lines of race W which were carried to the end of the experiment showed no endomixis during the entire period of observation. Their life without endomixis extended for 163 days and included 302 successive fissions. This long period is 5| times as long as those reported to be typical for this race by Woodruff and Erdmann (1914) and it includes about seven times as many fissions. Comparison of the data for the two races given in Table I shows that the group of race W lines had a longer initial period without en- domixis (30 days as compared with 13), and that it at no time at- tained so high a rate of endomixis or death as did the group of race R lines. Further, the maximum period without endomixis was longer in race W than in race R, but included fewer fissions. § J= 3 •- o fvl Q » -5 | ir t »o " r* ^j o a •4-* 5* *^ S S ^ ^ 2§ P4 O "2 U £' ££ T K C i C i > | > tf K I *~_> (*s 1 • j K- t j w o "O CM C 0 J 1 __ 'o ^ "« *J c r** *o <\J •S S fc 'So T w > C p G r> - 4 ' : ^ u h 1 f 1 ; t r-^ at c — o _O £ -5 r ^ to c ; ^ T 1 J ir> t^J to V : ! o 1 O «O CT> •— — •— 0 ^ c u rO «/» rx. (N, -J to f «T ' ..I u u r c 7 tC J CNJ i tn L •> > J i * u L (, 8 I-. .. •| c :"=f$5 ) T •-\.::: - P • * •' o Is ...Jo in T -IV 'J- - M 1 ••• J IT) <"g IT 1 1 a u (< \f > 1 j 1 • ^ - M •\ ru in . — CVJ ^ _a» n 3 , i •- — " "5 C\J «NJ ° 5 r-N** QJ _ _ — . • ~— t* O j *" • c -s -e m ._ CNJ IB. I ' (^ rJ UJ ill • ~ ^ 4) «> C d3 in -C J> OJ u CO ^H .N ^ 3 " ^ "O 3 — r^ J§ ° ? -2 S-5 A r*. to CO ^— ~ ^j '£* '^f ^ r^- UJ ^* *"* C/) o ~~ 1 ( ^* Ii^ ^ |' r^ in r^ = 2 8 °"l . -. "-- l L § UJ , >•"=: ' uj uj o ;p R ^^^» g = U .= L CM ± ; f 0 5 g-g i UJ L u ^ U 3 § ^ UJ _ . , L o c ^c S C/3 ^ (U . — CM .. re '£ « "§ '!• = "-">"« || J o 'S,"^.^^. § C 5 " '> 00 00 co <& . J uj "" » "2 *o aO ^O uj i rv oo °0 T " 1 o 3 .. ^^ T..." 00 UJ r*x "§ S "5 ^ c r J 1 ^ l fV>" ^" n « •.- o 55 g s tr C "~~" **'1 I e "^ c o -S t i 9 * p «•* *O L S 10 c y (« « ~ c/l l ***^ t/i flj 1 a» •- E a> "ai __ 3 J= .§ « C . w J= 0 ia o <» aJ H _£ >— ' C ' "e UH £ C C JS .2 o O t/J "^ 80 AVOIDANCE OK KXDOMIX1S IN PARAMECIUM Three other races (//, E, and S) have been subjected to the same method of selection with similar results. It seems reasonable to sup- pose that what has been accomplished in these five races could prob- ably be done with most or all races <>t P. aurelia under similar cultural conditions: by selection, lines can be isolated that do without endo- mixis throughout their lives — a period 5 to 7 times as long as the ordi- nary interval between successive endomixes. Variability of the Interendomictic Interval In the absence of selection, the extent of the period during \vhich no nuclear reorganization occurs was shown by Sonneborn (1937a) to vary greatly. The preceding sections show that still greater vari- ability appears when the selection technique is applied. This greater variability must be due to the existence of relatively rare lines which, without any special experimental treatment, have extremely long interendomictic intervals or even fail completely to undergo endomixis. Since such lines are very rare, they would ordinarily not be found in work where the investigator is able to observe only a small sample of a race. The method of selection serves simply to seek out these lines and multiply them. Without resorting to any change of cultural conditions, it provides a means of obtaining both lines that long omit endomixis and lines that omit it entirely. SUMMARY If daily isolation lines of P. aurelia are regularly discarded as soon as they go into endomixis, and if these are then replaced by sister lines that have not yet gone into endomixis, it is possible to maintain for long periods lines which have not been in endomixis since the start of this procedure. With this method there were obtained lines of race R which omitted endomixis for as long as 130 days and 350 fissions, and lines of race W which omitted endomixis for 163 days and 303 fissions. At the end of these long periods, all lines died. In the culture of groups of lines selected in this way, there is an initial period during which endomixis does not occur; this is followed by a period in which endomixis occurs in many of the lines (which have to be discarded) ; during the remaining history, endomixis occurs rarely and eventually not at all, but the death rate rises and results in the final extinction of the group. The details as to the duration of these periods and their characteristic death and endomixis rates differ in different races. The method of selection emphasizes the enormous variability of the interendomictic interval. The extremely long intervals are very rare, and the selection method is simply a device for finding these rare lines and multiplying them. T. M. SONNEBORN LITERATURE CITED JOLLOS, V., 1916. Die FortpHanzung der Infusorien und die potentielle Unster- blichkeit der Einzelligen. Biol. Centralbl., 36: 497. SONNEBORN, T. M., 1936. Factors determining conjugation in Paramecium aurelia. I. The cyclical factor: The recency of nuclear reorganization. Genetics, 21: 503. SONNEBOR.N, T. M., 1937a. The extent of the interendomictic interval in Parame- cium aurelia and some factors determining its variability. Jour. Exper. Zool., 75: 471. SONNEBORN, T. M., 1937/j. Induction of endomixis in Paramecium aurelia. Biol. Bull., 72: 196. XYooDRi'FF, L. L. AM> R. Ki\i)\i ANN, 1914. A normal periodic reorganization process without cell fusion in 1'aramecium. Jour. Exper. Zool., 17: 425. THE DISTRIBUTION OF AMMONIA IN THE WATERS OF THE GULF OF MAINE ALFRED C. REDFIELD AND ANCEL B. KEYS (From the Woods Hole Oceanographic Institution l and the Biological Laboratories, Harvard University] In the decomposition of organic residues in the sea ammonia is apparently the first simple inorganic compound of nitrogen to be set free. Since it does not commonly occur in large quantities, it is evidently oxidized rapidly to nitrate and nitrite. As an ephemeral intermediate in the cycle of regeneration of nitrate, it is of interest in indicating the places in which the decomposition of nitrogenous ma- terial is taking place. The present paper presents the results of a survey of the occurrence of ammonia in the waters of the Gulf of Maine based on observations made during cruises of the "Atlantis" in September, 1933 and May and June, 1934. The only previous measurements of ammonia in these waters of which we are aware are those made by Seiwell (1931) in Penobscot and Frenchman's Bays and those recorded by Rakestraw (1936) in a study of the occurrence of nitrite in the sea. Our results are correlated with the studies of nitrite made by Rakestraw of the phosphorus cycle by Redfield, Smith and Ketchum (1937), and with observations on the distribution of plankton, based on collections made during these cruises. METHODS Ammonia was determined by the Teorell titration of the vacuum distillate as described by Krogh (1934). The apparatus could be used practically on shipboard, using alcohol lamps burning pure alcohol for heating and a Cenco Hyvac pump as a source of low pressure. Am- monia-free water was prepared ashore. It was found necessary to steam out the apparatus before the start of each series of analyses and to make blank analyses between each pair of unknowns. To be reli- able, all determinations must be made in duplicate, since unforeseen contamination frequently occurred. The sea water samples were taken in Nansen reversing bottle-, from which they were immediately transferred, with double rinsing, to gfess-stoppered bottles. These were stored in the refrigerator and 1 Contribution No. 159. 83 84 ALFRED C. REDF1ELD AND AXCEL B. KEYS analyzed as quickly as possible. Experiments recorded in Table 1 showed that sea water stored for about eight hours underwent no appreciable change in ammonia content. With longer periods of storage there is a progressive decrease in the ammonia content, which is not prevented by the addition of 0.1 per cent mercuric chloride (Keys, Christcnsen and Krogh, 1935). The reason for this loss is unknown. RESULTS The positions of the stations at which reasonably complete and satisfactory observations on the distribution of ammonia were obtained TAB Lie I Stability of ammonia in stored sea water samples * 1 analysis Second analysis Third analysis Hours after collection NHj-N, mg./M' Hours after collection NHj-N. mg./M' Hours alter ' dlcction NH3-N. mg./M3 1 40 7 41 290 10 2.5 11 8.5 L9 290 8 3.0 35 7 35 290 15 2.5 28 8 JO 290 2 4 20 — — 90 13 4 62 17 51 90 45 6 -12 — — . 90 25 6 44 18 24 — — 6 57 18 16 90 37 5 4') 17 28 90 23 4 46 17 32 — — 5 42 18 30 90 20 tl.5 36 290 8 t2 25 290 4 t2.5 21 290 15 ts.s 31 290 16 * All values are averages from duplicate analyses. t I'gCl? added to sample at time of collection. in September, 1933 and May, 1934 are shown in Figs. 1 and 2. Ob- servations were also made at six stations in the (inll between June 25 and June 30 ; the results obtained al that lime did not differ from those of a month earlier except that on the \\hole smaller amounts oi am- monia were present. In presenting the results it is convenient to separate the observations made in deep water where the stability of the \\.iter column preserves its heterogeneity from those obtained in -hallow water and in the shallower channels across the banks where a degree of turbulence prevails. DISTRIBUTION OF AMMONIA IN GULF OF MAINE 85 OBSERVATIONS IN DEEP WATER Figure 3 represents the concentrations of ammonia at various deep stations in September and in May. The stations are arranged as far as possible in the order of the positions through which the water is thought to circulate on penetrating the Gulf. Stations occupied in May are placed under stations occupied in comparable positions in September. In September, Station 1781, 45 miles southeast of the Eastern Channel, showed the presence of some 10 mg. N per cubic meter in the upper 60 meters, diminishing to zero at 100 meters, below which level no ammonia was detected down to the depth of 2,000 meters. The ammonia in the water in the Eastern Channel was similar in distribu- tion, there being none observable below 125 meters. Within the FIGS. 1 and 2. Position of stations at which distribution of ammonia was de- termined in September, 1933 and May, 1934 respectively. Gulf, on the other hand, the concentration of ammonia was distributed rather uniformly throughout the water column. The concentrations of ammonia occurring at any station increased towards the western basin of the Gulf, where amounts exceeding 50 mg. N per cubic meter were found. In May the distribution of ammonia was very different. Amounts in excess of 10 mg. N per cubic meter were rarely observed in the upper 20 meters and at depths greater than 100 meters. Between these depths, at 30 to 60 meters, maximal concentrations occurred ranging in value from 10 mg. N per cubic meter in the Eastern Channel to 45 mg. N per cubic meter in the southern part of the western basin. In the following month similar conditions were observed save that the maximal concentrations did not exceed 25 mg. N per cubic meter. 86 ALFRED C. REDFIELD AND ANCEL B. KEYS z cc UJ • : s •- (/I 5 C $ r- 0 - 4-> rt — fc/. o cn o UJ 5 z * "> CO •. I h- .*••* •. •- * ;; cc 0 i •f z § " • ' a: o "> UJ •» g: i .-• H K. $^ • 2 ° < • •i UJ ) / * . . * ^_-^^M^I i o Z -J?l a: LLI. uj Z' yt 1- Z to * w «tS N. O < I **" ^ Ul O ' • . •• • § i ft*^^; ^ t- »1 Z UJ "\ UJ Q- **• 'R Z 00 ?> , *«» *o Z" • * % ^ V. o i^ — 0 2 8 S 8 S 10 2 11? cj p«i 8 ^^5 o u> O lO 10 9 s; -• •j 1- 0 "» «\! • . I . s . 5 j • "• j i . o O O O in o >n o CU N ro . o i - ^_ o O» 4, £ = -C 4J -S ^.= « tn o U AJ I & «"< j5 fc I S *-• I— 1 O y c cs ^ J!S « <= c 2 c •- c o £ 2 3 °^ ^ !-5 — 'C x E -b o c 10 d m .SUM t a £ JJ £• a r-* C X o . ... • -• NO I7S9 IX- 11-33 50 100 150 NO 2193 V - 28-34 NO 2165 V- 21 - 34 NO 2214 VI- 2- 34 NO 22OI V- 29 - 34 NORTH CHANNEL CAPE COD SOUTH CHANNEL GEORGES BANK FIG. 4. Vertical distribution of ammonia in shallow waters of the Gulf of Maine. Upper row, in September, 1933; lower row, comparable positions in May, 1934. Concentrations of ammonia expressed as milligrams ammonia nitrogen per cubic meter along the abscissa. Channel. In the shallow stations the concentration sometimes in- creased with depth, sometimes the reverse. DISCUSSION The distribution of ammonia appeared very irregular when the observations were first made. The magnitude of the concentrations of ammonia observed agreed well with the general range of observa- tions made by earlier workers in various regions (Robinson and Wirth, 1934). It appeared, however, that great differences might be observed ALFRED C. REDFIELD AND ANCEL B. KEYS in similar situations at any time and in the same situation at different times. In several ways our observations support the conclusions drawn from similar studies of ammonia in the English Channel made by Cooper (1933). He considers the formation of ammonia to be a "surface" phenomenon with lesser activity at the bottom. In the spring the surface concentrations decrease from utilization by phyto- plankton. Our May observations would appear to be adequately explained on these assumptions. He also observed a general increase in summer and autumn, as did we in September. As additional evidence has accumulated concerning the biology and chemistry of the Gulf, a number of correlations appeared which seem to give the findings additional significance. RELATION TO PHOSPHATE CYCLE The most striking difference between the occurrence of ammonia in September and in May is its relative abundance in the deep water at the end of summer compared with its absence there in the spring. This is precisely the relation found for another intermediate product of decomposition — dissolved, organic phosphorus compounds. These were found by Redfield, Smith and Ketchum (1937) to disappear from the water of the western basin in mid-winter and gradually to ac- cumulate during the summer, reaching a maximum observed concen- tration in November. Like the ammonia in September, the concentra- tion was quite uniform at all depths. In the spring these phosphorus compounds appear in the subsurface layers, as does the ammonia, before they may be detected in considerable quantity in the depths. It will be of interest to observe whether the ammonia does not almost entirely disappear from the water in mid-winter, as does the dissolved organic phosphorus. Like the organic phosphorus, ammonia appears to accumulate at all depths in late summer. This should not be thought of as the accumu- lation of an inert product, however, for von Brand, Rakestraw, and Rcnn (1937) have shown, at least in a laboratory experiment, that the ammonia of decomposing plankton is nitrified completely in 30 days. It is probable that the ammonia is formed by decomposition throughout the water column, for in September there are no well- marked gradients of ammonia concentration which are necessary for its transport by eddy conductivity. There is no evidence, except in a few of the shallow stations, of increasing concentrations near the bottom, such as Seiwell (1931) observed in shallow bays and as might if decomposition on the sea bottom was responsible for a con- part of the regeneration of nitrogen. DISTRIBUTION OF AMMONIA IN GULF OF MAINE 89 RELATION TO NITRATE If the view to which Rakestraw (1936) has recently lent support is correct — that nitrite arises in the water through the oxidation of ammonia — one would expect a close correlation in the distribution of these nitrogen compounds. This expectation is in part fulfilled. The distribution of ammonia in May closely parallels that of nitrite, as shown in Fig. 5. Both may be explained through decomposition taking place in the subsurface waters. Rakestraw explains the absence of nitrite from the water immediately below the surface as being due to 100 10 20 40 50 150- 200- AMMONIA FIG. 5. Vertical distribution of nitrite and ammonia at Station 2212 May 31, 1934. Depths measured in meters downward along the ordinate. Concentrations measured in milligrams nitrogen present as nitrite or as ammonia per cubic meter. the assimilation of this substance by plankton. Evidently ammonia is similarly consumed. It is interesting, though not surprising, that there is no evidence that soluble organic phosphorus compounds are assimilated in a similar way — since they were found in as high con- centration in the surface as at any depth. Rakestraw remarks that the maximum for nitrite occurs between June and September. However, the nitrite does not appear to ac- cumulate toward the end of summer to the extent to which ammonia and organic phosphorus compounds do. Nitrite is evidently a shorter- 00 ALFRED C. REDFIELD AND AXCEL B. KEYS lived link in the chain of nitrogen transformations than is ammonia for even in May the ammonia nitrogen concentration exceeds that of nitrite nitrogen by ten-fold. HIGHEST AMMONIA SEPT 33 HIGHEST AMMONIA VOLUME OF ZOO-PLANKTON SEPT '34 VOLUME OF ZOO- PLANKTON MAY-JUNE '34 FIG. 6. The highest concentration of ammonia in the upper 80 meters of water in various parts of the Gulf of Maine in September, 1933 and May, 1934 compared with the quantitative distribution of zociplankton. Ammonia concentrations ex- pressed in milligrams ammonia nitrogen per cubic meter. Zooplankton volumes expressed as cubic centimeters of dry plankton per square meter of surface area caught in a vertical haul from near the bottom. RELATION TO PLANKTON The quantity of ammonia observed during both cruises varied !^i rally from station to station. In September the greater quantities DISTRIBUTION OF AMMONIA IN GULF OF MAINE 91 were observed in the northern and western parts of the Gulf; in May in the southern and eastern regions. These facts did not assume significance until the plankton catches made at the time had been analyzed. During each cruise a standard vertical zooplankton haul was made at each station occupied (a number far exceeding those at which the ammonia was investigated). A 1.5-meter silk net was hauled from near the bottom to the surface. The total catch at each station was freed of excess moisture on a filter and its volume measured by dis- placement. The results of these measurements expressed as the vol- ume of "dry" plankton per square meter surface have been plotted in Fig. 6 to show the general distribution of animal plankton. For comparison, the highest value of ammonia observed at each station in the upper 80 meters is also shown. While the correlation is not per- fect, it is evident that the highest concentrations of ammonia at both periods of observation occur in regions in which the zooplankton are densely distributed. This correlation suggests that the animal plankton may be in some way responsible for the appearance of ammonia. While specific observations on plankton animals are lacking, ammonia appears to be the principal nitrogenous waste product of many invertebrate animals (Delauney, 1931). Harvey (1934) has pointed out that the phyto- plankton community is continuously being grazed down by the zo- oplankton. These considerations point to the probability that zooplankton, through their metabolic products and through their own decay, are an important intermediary in the liberation and distribution of ammonia in sea water. SUMMARY 1. In the deeper basins of the Gulf of Maine in May ammonia occurred in minimal concentrations at the surface and at all depths below sixty meters; maximal concentrations varying up to 45 mg. N per cubic meter occurred in a definite stratum between 30 and 60 meters. 2. In September the concentration of ammonia was rather uniform at all depths and increased as the distance from the open sea increased, concentrations exceeding 50 mg. N per cubic meter occurring in the western basin. 3. In the tideways of the North and South Channels, ammonia is distributed uniformly with depth in both May and September. 4. In shallow waters its occurrence showed no regularity. 5. The occurrence of ammonia may be correlated in part with the distribution of organic phosphorus compounds, of nitrite, and of 92 ALFRED C. REDFIELD AND ANCEL B. KEYS zooplankton, so as to support the view that its distribution marks the place and the intensity of organic decomposition. REFERENCES VON BRAND, T., N. \V. RAKESTRAW, AND C. E. RENN, 1937. Biol. Bull, 72: 165. COOPER, L. H. N., 1933. Jour. Mar. Biol. Ass'n., 18: 677. DELAUNEY, H., 1931. Biol. Rei'.t 6: 265. HARVEY, H. W., 1934. Jour. Mar. Biol. Ass'n., 19: 775. KEYS, A., E. H. CHRISTENSEN, AND A. KROGH, 1935. Jour. Mar. Biol. Ass'n., 20: 181. KROGH, A., 1934. Biol. Bull., 67: 126. RAKESTRAW, N. VV., 1936. Biol. Bull., 71 : 133. REDFIELD, A. C., H. P. SMITH AND B. KKTCHUM, 1937. Biol. Bull., 73: 421. ROBINSON, R. J., AND H. E. WIRTH, 1934. Jour. Conseil. perm. int. pour I'explor. de la mer, 9: 15. SEIWELL, H. R., 1931. Ecology, 12: 485. THE OXYGEN CONSUMPTION OF ARTIFICIALLY ACTI- VATED AND FERTILIZED CH^TOPTERUS EGGS JEAN BRACKET (From the Marine Biological Laboratory, Woods Hole, Mass., Biology Department, Princeton University, and the Faculty of Medicine, University of Brussels) INTRODUCTION It is well known that unfertilized Chsetopterus eggs, when they are treated with various physical or chemical agents, undergo a special type of development (F. R. Lillie's (1902, 1906) differentiation without cleavage). Lillie's original observations have been recently confirmed and completed by Pasteels (1934) and by myself (1937). These inves- tigations show that if the unfertilized eggs are treated with 5 per cent isotonic KC1 in sea water, they first complete their maturation; im- mediately after the expulsion of the second polar body, they start undergoing a series of monaster cycles. During each of the cycles, there is an increase in the number of the chromosomes and the appear- ance of lobulations simulating cleavage. But the furrows fade out quickly and the egg resumes its original shape. There is, therefore, a considerable increase in the size of the egg nucleus and in the number of the chromosomes during the first hours of differentiation without cleavage. Later on, the big single nucleus breaks down and, at the same time, occur processes comparable to gastrulation; in the best cases, unicellular ciliated eggs, resembling somewhat trochophores in their general shape, can be obtained. It seemed of interest to measure the oxygen consumption of eggs undergoing differentiation without cleavage and to compare it with the respiration of unfertilized eggs: it was hoped that such an inves- tigation would throw some light on the question whether energy is needed for cleavage and differentiation, a problem which has recently been investigated by Tyler (1933, 1936) on other eggs. A beginning was made last year during a stay in the Zoological Laboratory in Naples and a report has been published recently (1937). The results, however, were far from satisfactory because a small per- centage only of the eggs could be fertilized, although differentiation without cleavage was easily obtained. Lacking the necessary controls, we could only compare the data obtained for the activated eggs with the results published by Whitaker on the respiration of fertilized eggs in Woods Hole. Obviously, no definite conclusion can be drawn from 93 94 JEAN BRACKET such a comparison and it was therefore necessary to reinvestigate the problem. MATERIAL AND METHODS The experiments were carried out following the technical details pointed out by \Vhitaker (1933) : one or two ripe females were allowed to shed after the tip of their parapodia had been cut with scissors. The eggs were then washed repeatedly (5 or 6 times) in sea water in order to remove the mucus and a suspension of a concentration varying from 1 : 25 to 1 : 40 was prepared by slightly packing the eggs in a graduated centrifuge-tube and adding the required amount of sea water. Aliquot parts of the suspension (2 cc.) were pipetted into Warburg manometer conical cups (ca. 10 cc. capacity). Fertilization with one drop of diluted sperm or activation with 0.1 cc. of isotonic KC1 usually occurred just before attaching the vessels to the manometers. Readings were taken after a 15-minutes equilibration period. The water-bath was kept at 24.8° and the manometers were shaken at a speed of 50 round- trip oscillations per minute with an excursion of 4 cm. That these conditions may be considered as adequate is indicated by the fact that increase in the speed of shaking induced no increase in the oxygen consumption ; furthermore, the respiration of unfertilized eggs usually kept constant during 6 or 7 hours and they sho\ved at that time little, if any, cytolysis. RESULTS We kno\v since Whitaker's experiments that there is a strong drop in the oxygen consumption of Chxtopterus eggs at the time of fer- tilization. My previous experiments from Naples indicated that a similar drop occurs in KCl-activated eggs. In most of those experi- ments, activation was produced by tipping the KC1 contained in the side-arm into the main part of the vessel after the respiration had already been measured during 1 or 2 hours. The results at Naples could be easily confirmed at Woods Hole: while the metabolism of the fertilized eggs dropped to 53 per cent of the initial level (average from 14 experiments in which the percentage of fertilized eggs exceeded 90 per cent), the oxygen uptake of the activated eggs (from the same females) fell to 51 per cent. There is therefore no doubt that, like the increase in respiration which follows fertilization in sea-urchin eggs, the drop observed in the present case is linked to the cytoplasmic changes resulting from activation and not to any special influence exerted by the spermatozoon on the metabolism. The following graph represents the oxygen consumption of Chsctop- terus eggs during the 7 hours following activation and fertilization. OXYGEN CONSUMPTION CH^TOPTERUS EGGS 95 In this graph, the rate of oxygen consumption has been plotted against time, the initial respiration of the unfertilized eggs being considered as 100 per cent. The individual points on the curves represent average values obtained from 14 experiments. Rate of 00 consum-pttoa **•% 200r -150 -100 50 u, •i 204507 Hours. FIG. 1. Oxygen consumption of Chxtoptems eggs during the seven hours fol- lowing activation and fertilization. Ordinates: rate of oxygen consumption; ab- scissae, hours. The curve F represents the oxygen consumption of fertilized eggs; the curve KC1 represents that of eggs treated with KC1; and the circles, u, represent that for unfertilized eggs. It is easy to see that, as stated before, the oxygen consumption of the unfertilized eggs (indicated by large circles) remains constant throughout the whole experimental period. It is evident too that respiration increases after the drop in oxygen consumption occurring after fertilization or activation, but obviously this increase is much stronger and faster in the case of the fertilized eggs than it is for the activated eggs: the fertilized eggs reach the initial level after 3y£ hours, i.e. at an early blastula stage. The respiration rate increases steadily later on and is almost 80 per cent higher at the end of the experiment than the oxygen uptake of the unfertilized eggs. On the other hand, the activated eggs differentiating without cleavage reach the initial rate only after 6 hours instead of 3>^. The fact that the same dif- ference in the respiration of two sets of eggs was found regularly in all the experiments indicates beyond doubt that the observed fact is really significant. It was possible that the difference might be linked to a depressive effect on the metabolism of the added KC1 which was present during the whole experiment. In order to check this possibility, 4 series of 96 JEAN BRACKET control experiments were run in which the respiration of activated eggs was measured both in presence of KC1 in excess as before and after repeated washings (10 times). In other manometers, the oxygen consumption of normal fertilized eggs was compared with the metab- olism of similar eggs to which 5 per cent isotonic KC1 was added 10 minutes after fertilization. All these control experiments showed that the presence or the absence of KC1 in that concentration does not exert any significant effect on the curves. In both cases, the initial respiratory level was reached after 3^2 hours in the case of the fertilized eggs and after 6 hours for the activated ones. It seems, however, possible that KC1 slightly enhances the absolute Q% consumption of the eggs. Obviously, the differences observed in the metabolic rate of eggs differentiating without cleavage and of fertilized eggs is some- how linked to their different type of development and not to a de- pressive effect of KC1 on the metabolism. It \vas also of some interest to see ho\v the thymonucleic acid synthesis, as a chemical index of the mitotic activity, would compare in fertilized eggs and in eggs differentiating without cleavage. The eggs were therefore taken out of the manometer vessels and preserved in acetone until a sufficient amount of material (ab'out 2 grams) could be collected. The thymonucleic acid content of unfertilized, activated and fertilized eggs was determined by Dische's (1930) colorimetric micromethod. Three estimations could be made. These showed that while the unfertilized eggs contained, as was expected, only traces of thymonucleic acid, the activated eggs had a content in this substance of about 0.35 milligrams per gram wet \veight and the fertilized eggs of 1 milligram per gram wet weight. The thymonucleic acid synthesized in 8 hours during differentiation without cleavage amounts thus only to 30 per cent of the amount produced in the fertilized eggs during that period. DISCUSSION The curve we obtained for fertilized eggs closely resembles the one published by \Yhitaker (1933), although its slope is somewhat steeper. In Whitaker's experiments, the unfertilized egg rate is reached by the fertilized eggs in 4^ hours and the increase over that rate at the end of 7 hours of the experiment is only 30 per cent. This difference is very likely due to the higher temperature (24.8° instead of 22°) at which our experiments were carried out: ciliary activity had just begun after 6 hours in XYhitaker's case while we noticed swimming blastulae after 4*/£ hours. It seems therefore probable that conditions similar to those described by Tyler (1936a, b) in his work on temperature coefficients of developmental processes and cellular respiration prevail OXYGEN CONSUMPTION CH^TOPTERUS EGGS 97 also in Chxtopterus eggs. It is very probable that the reduced meta- bolic activity of the eggs differentiating without cleavage is, likewise, somehow linked with their slower development. \Ye have already seen that the control experiments rule out the possibility that the observed difference should result from an inhibition effect of KC1 on the eggs' respiration. There is also no doubt that the development of the eggs differentiating without cleavage is slowed down to a consider- able extent: for instance, at the end of the experiments the fertilized eggs had turned into actively swimming larvae while the activated eggs develop cilia only much later. Likewise, the increase of nuclear material, as indicated by the thymonucleic acid estimations, goes on at a much slower rate in the eggs differentiating without cleavage than in the fertilized ones. It is therefore likely that both the reduced oxygen uptake and the slower development are linked together; such a conclusion could support Tyler's opinion that part of the energy available in the egg is needed for the growth and differentiation processes taking place during development. Recent findings by Runnstrom (1933), J. Brachet (1935), Privolnev (1936), Stefanelli (1937) that the respiration undergoes cyclic changes during mitosis in eggs of different species are also in good agreement with our observa- tions. It must be pointed out, however, that the interpretation of the results is complicated by a special factor; namely, that the activated eggs remain unicellular while the fertilized ones cleave into many cells. It is, of course, by no means impossible that oxidation processes might occur at the surfaces between the different blastomeres. Against such an interpretation, however, two facts can be cited; namely, that there is no direct relationship between the number of the blastomeres and the oxygen uptake of the egg (cf. Needham, 1931) and that the respira- tion of small marine invertebrate eggs is over a wide range independent of the O2 tension in the surrounding medium. SUMMARY 1. The oxygen consumption of unfertilized Chaetopterus eggs drops in the same proportion whether they are fertilized or activated with KC1. 2. The increase in respiratory rate after that initial drop is much faster in fertilized eggs than in activated (differentiating without cleavage) eggs. 3. The lower respiratory activity of the activated eggs is linked to their special type of development. 98 JEAN BRACKET BIBLIOGRAPHY BRACKET, J., 1935. Etude du mctnbolisme de 1'oeuf de Grenouille (Rana fusca) au cours du developpement. II. La respiration de 1'oeuf pendant la feconda- tion et la mitose. Arch, de Biol. 46: 1. BRACKET, J., 1937. La differentiation sans clivage dans 1'oeuf de Chetoptere envisagce aux points de vue cytologique et metabolique. Arch, de Biol., 48: 561. DISCHE, Z., 1930. Cber einige neue charakteristische Farbreaktionen der Thymonu- kleinsaure und eine Mikromethode zur Bestimmung derselben in tierischen Organen mit Hilfe dieser Reaktionen. Mikrochemie, 8: 4. LILLIE, F. R., 1902. Differentiation without cleavage in the egg of the annelid Chaetopterus pergamentaceus. Arch.f. entw. Mech,, 14: 477. LILLIE, F. R., 1906. Observations and experiments concerning the elementary phenomena of embryonic development in Chaetopterus. Jour, exper. Zool., 3: 153. NEEDHAM, J., 1931. Chemical Embryology. Cambridge University Press. PASTEELS, J., 1934. Recherches sur la morphogencse et le determinisme des seg- mentations inegales chez les Spiralia. Arch. d'Anat. microsc., 30: 161. PRIVOLNEV, T. I., 1936. Ueber den Atmungsrythmus bei der Furchung von Fluss- prickeneier (Lampetra fluviatilis). C. R. Ac. Sc., U. R. S. S. 4: 433-436. RUNNSTROM, J., 1933. Stoffwechselvorgange wahrend der ersten Mitose des See- igeleies. Protoplasma, 20: 1. STEFANELLI, A., 1937. Prime osservazioni sulle assunzione di O^ della uova et dei premi stadi embrionale dei Bufonidi. Bull. Soc. ital. Biol., 12: 284. TYLER, A., 1933. On the energetics of differentiation. I. A comparison of the oxygen consumption of "half" and whole embryos of the sea urchin. Publ. Staz. zool. Napoli, 13: 155. TYLER, A., 1935. On the energetics of differentiation. II. A comparison of the rates of development of giant and of normal sea-urchin embryos. Biol. Bull., 68: 451. TYLER, A., 1936a. III. Comparison of the temperature coefficients for cleavage and later stages in the development of the eggs of some marine animals. Biol. Bull., 71 : 59. TYLER, A., 19366. IV. Comparison of the rates of oxygen consumption and of development at different temperatures of eggs of some marine animals. Biol. Bull,, 71 : 82. \YHITAKER, D. M., 1933. On the rate of oxygen consumption by fertilized and unfertilized eggs. IV. Chaetopterus and Arbacia punctulata. Jour. Gen. Physiol., 16: 475. ON THE ENERGETICS OF DIFFERENTIATION. VII COMPARISON OF THE RESPIRATORY RATES OF PARTHENOGENETIC AND FERTILIZED URECHIS EGGS ALBERT TYLER AND N. H. HOROWITZ (From the William G. Kerckhoff Laboratories of the Biological Sciences, California Institute of Technology, Pasadena, California) The results of these experiments show that the rise in respiratory rate that occurs during development is correlated with cleavage; eggs that fail to cleave after activation show a greatly retarded rise while those that cleave show a rise that is roughly commensurate with their division rate. Inhibiting cleavage with phenylurethane affects the respiratory rate similarly. THEORETICAL PART It has been long been known that the rate of respiration rises during development. This increase in rate is evidently not directly propor- tional to the increase in the number of cells (cf. Needham, 1931). It might, nevertheless, depend upon changes in the egg brought about by cell division, so that when cleavage fails to occur the rise in respiration would be inhibited. We have considered in previous work the dependence of the form-changes on the respiration, the rate of oxygen consumption being taken as a measure of the energy available for the various developmental processes. We consider now the possibility that the developmental changes determine in turn the rate of respira- tion. If, for example, early cleavage is inhibited in a manner that does not affect the absolute rate of respiration at the particular stage, then we may expect, on this basis, failure of the subsequent rise. In the early work of Warburg (1910) it has been shown that cleav- age could be suppressed in sea urchin eggs by means of phenylurethane without immediately affecting the respiratory rate. However, the question of whether or not the rate would rise later was not investigated. Also it has been shown that after parthenogenetic activation of sea urchin eggs the same increase in rate occurs that is obtained normally upon fertilization (Warburg, 1910; Loeb and Wasteneys, 1913). But here again it would be desirable to know what happens later, especially since the parthenogenetically activated eggs develop much more slowly in general and often stop in early cleavage or even fail to divide. 99 100 ALBERT TYLER AND N. H. HOROWITZ In Urechis artificial activation with a single agent may produce dividing or non-dividing eggs depending on the length of treatment. Diluted sea water (Tyler, 1931) or ammoniacal sea water (Hiraiwa and Kawamura, 1936; Tyler and Bauer, 1937) may be used for this purpose. EXPERIMENTAL PART The respiration measurements were made by means of the usual Warburg method using the cylindrical type of vessel previously described (Tyler, 1936) but of 18 to 20 cc. calibration volume. The quantity of eggs present in each vessel was determined at the end of the run from the Kjeldahl nitrogen, and the oxygen consumption is expressed on that basis. The quantity of eggs employed was such as to give readings of 25 to 50 mm. Brodie fluid per hour at the start of the experiment. The non-dividing and the dividing parthenogenetic eggs were produced simply by treatment with ammoniacal sea water, as pre- viously described. Treatments of 2 to 7 minutes with 0.01 N NH3 in sea water give 100 per cent activation with normal polar body extrusion but no cleavage. Such eggs go through a series of monaster cycles after polar body formation, but at a slower rate than would correspond to the normal nuclear changes. After about 10 to 15 hours unicellular swimmers may develop similar to the embryos differentiating without cleavage, first described by Lillie (1902). Treatments of 12 to 17 minutes with 0.01 N NH3 in sea water give 100 per cent activation with as much as 90 to 100 per cent cleavage. Polar body formation is usually interfered with in such eggs, as previously described (Tyler and Bauer, 1937). After 8 to 12 hours swimming embryos including some normal ones may appear. Non-Cleaving Parthenogenetic Eggs The results of two sets of experiments with non-cleaving activated eggs are given in Table I, along with the fertilized controls. In both the treatment was for six minutes with 0.01 N NH3, and insemination of the control was done at the start of the parthenogenetic treatment. Activation and fertilization were 100 per cent in both these experi- ments. As may be seen in Table I the oxygen consumption of the parthenogenetic eggs during the first hour of measurement is very nearly the same as that of the fertilized eggs. In both experiments the parthenogenetic eggs give slightly higher values, but the difference is no greater here than the difference between the duplicate vessels. During the subsequent hours the rate rises steadily in the case of the Irrtili/rd eggs. The parthenogenetic eggs, however, show first a RESPIRATORY RATE AND CLEAVAGE 101 slight drop (second to fourth hours) followed by an increasing rate. After 8 hours the parthenogenetic eggs attain a rate that is about double their initial respiration, but only about half of that of the fertilized eggs. The eggs from the vessels were examined at the end of the run. In both experiments the parthenogenetic eggs gave about 2 per cent of unicellular swimmers while the fertilized eggs gave 90 to 98 per cent normal top-swimming trochophores. Cleaving Parthenogenetic Eggs The respiration data of two sets of experiments with eggs that divide after ammonia treatment are given in Table II. In both TABLE I Oxygen consumption of non-cleaving parthenogenetic eggs. Parthenogenetic treatment = 6 minutes with 0.01 N NH3 in sea water. Measurements begun at 43 minutes after treatment or insemination in Experiment I and 60 minutes after in Experiment II. Values given as cu. mm. C>2 per hr. per mg. N. Temperature = 22.0° C. Experiment [ II Hour Parth. Parth. Fert. Pert. Parth. Pert. Pert. 1st 4.64 4.48 4.38 4.24 4.72 4.53 4.14 2nd ... . 4.40 4.30 4.53 4.38 3.88 5.05 4.49 3rd 4.02 3.99 5 60 5 55 3.83 5.45 5.35 4th 5th 4.52 5.22 4.34 5.07 7.58 10.01 7.53 9.93 4.32 5.56 7.01 10.10 6.65 10.20 6th 6.19 6.02 13.60 13.43 7.40 13.83 13.48 7th 6.86 7.29 16.38 16.31 9.66 18.26 17.45 8th 8.91 8.87 18.66 18.47 10.93 21.64 20.81 experiments it may be seen that the respiration of the parthenogenetic eggs at the start is slightly higher than that of the fertilized controls. A third set not presented in the table shows the same initial difference, and also agrees very well throughout the run. The difference here is somewhat greater than was manifested in the experiments with the non-cleaving parthenogenetic eggs. It may be pointed out that the ammonia treatment is considerably longer here (13 to 17 minutes) and this may account for the higher initial rate. The subsequent readings give values that are fairly constant for the parthenogenetic eggs up to the fifth hour, while the fertilized eggs rise as usual during this time. Following this the rate rises and at the eighth hour the respiration is about half of the control rate. It is very nearly the same at that time as is obtained with the non-cleaving 102 ALBERT TYLER AND N. H. HOROWITZ parthenogenetic eggs (Table I). However, differences appear later as will be shown below. The cleavage of these parthenogenetic eggs is very slow compared with the fertilized controls. In Experiment I, at 2 hours after treat- ment, there were 43 per cent cleaved of which 15 per cent were in two cells, 22 in three and 6 in four. The fertilized controls were 100 per cent in four at this time. An hour later there was 75 per cent cleavage. At the end of the run (11 hours after treatment) the eggs from the vessels were 90 per cent cleaved and 30 per cent were swimmers. In Experiment II, 60 per cent of the parthenogenetic eggs were divided at two hours after treatment, the distribution being 23 per cent TABLE II Oxygen consumption of cleaving parthenogenetic eggs. Parthenogenetic treatment = 13 minutes (Experiment I) and 17 minutes (Experiment II) with 0.01 N NH3 in sea water. Measurements began at 74 minutes after treatment or insemination in Experiment I and 53 minutes after in Experiment II. Values given as cu. mm. O* per hour per mg. N. Temperature = 22° C. Experiment ] [ I I Hour Parth. Parth. Pert. Pert. Parth. Parth. Pert. Unfcrt. 1st 5.88 5.81 4.48 4.64 5.69 5.70 4.51 3.69 2nd 5.45 5.69 4.92 4.82 5.40 5.48 4.60 3.90 3rd 5.20 5.25 5.63 5.80 5.51 5.34 5.59 3.37 A 4-U 51 £. 51 A 7A7 A OA 5QC c 70 f\ QJ. 4tn .oO .v>0 .\)o o.yo .80 o. / U 5th 5.63 5.25 10.28 10.12 5.91 5.48 9.07 3.29 6th 5.99 5.92 14.05 13.80 7.01 6.86 13.17 3.37 7th 7.55 7.25 17.46 17.83 8.04 8.02 16.27 8th 8.62 8.04 20.90 21.14 10.25 10.18 20.06 9th 9.71 9.49 24.50 24.45 10th 10 65 1095 2855 in two cells, 30 in three and 7 in four. The fertilized controls were all in four cells at the time. At the end of the run (10 hours) the par- thenogenetic eggs from the vessels were 95 per cent cleaved of which 5 per cent were swimmers, while the fertilized eggs gave 90 per cent swimmers. Since the parthenogenetic eggs do not divide at all synchronously and since many may stop after one or more cleavages, the distribution becomes quite complicated. It is clear, however, that cleavage and the development of cilia are retarded. Later Stages Measurements on the later stages were made by culturing the eggs in dishes and washing the embryos before transfer to the vessels. RESPIRATORY RATE AND CLEAVAGE 103 This is to avoid possible effects of bacterial growth or other changes produced in the vessels during a prolonged run. In Table III the results of one set of experiments are given. Cleaving and non-cleaving parthenogenetic eggs were prepared from the same batch along with the fertilized controls, and allowed to develop for 11 hours at room temperature (18.2° C.) before transfer to the vessels. In the fertilized lot there were more than 95 per cent top-swimming young trocho- phores, and only the top swimmers were transferred to the vessels. In the non-cleaving parthenogenetic lot (Parth. 5) there were no swimmers at that time. In the cleaving parthenogenetic lot (Parth. 17) 95 per cent had divided and practically all of them were bottom- swimmers. TABLE III Oxygen consumption of late stages of non-cleaving and cleaving parthenogenetic eggs. Parthenogenetic treatment = 5 minutes (Parth. 5) and 17 minutes (Parth. 17) with 0.01 N NH3 in sea water. Eggs cultured for 11 hours at 18.2° before measure- ments were begun. Values given as cu. mm. O2 per hour per mg. N. Temperature = 22°. Hour Fert. Pert. Parth. 5 Parth. 5 Parth. 17 Parth. 17 1st 21.80 22.09 12.31 12.17 11.64 11.69 2nd 25.45 25.91 12.54 12.42 19.03 18.45 3rd 29.38 30.48 15.11 14.90 18.74 18.96 4th 32.39 33.32 15.01 14.65 20.75 20.26 5th 36.17 36.18 14.44 14.15 22.60 22.35 6th 39.94 40.48 15.76 15.72 24.76 24.98 7th 43.31 45.70 18.98 18.75 27.70 27.95 8th 46.50 47.70 22.18 22.73 31.73 32.67 As may be seen from Table III the two types of parthenogenetic eggs respire at very nearly the same rate at the start of the experiment. The rate is a little more than half of the fertilized rate at this time. The rate rises with time, but more rapidly in the case of the dividing eggs. At the end of the eight-hour period they are respiring at almost one and a half times the rate of the non-cleaving eggs, but still only two-thirds that of the fertilized controls. Examination of the eggs from the vessels at the end of the run showed 20 to 30 per cent bottom- swimmers (unicellular) in the case of the non-cleaving eggs (Parth. 5); 100 per cent swimmers of which 30 to 40 per cent were top-swimmers in the case of the cleaving eggs (Parth. 17); and 100 per cent top- swimmers in the case of the fertilized controls. Fertilized Eggs in Phenylurethane With the proper concentration of phenylurethane, cleavage may be suppressed without the initial respiratory rate being affected, as 104 ALBERT TYLER AND N. H. HOROWITZ Warburg (1910) showed on sea urchin eggs. The experiments were repeated for the purpose of covering a longer period. The results of such an experiment with UrecMs eggs are presented in Table IV. The eggs were placed in the solution at 45 minutes after insemination, and the measurements begun 40 minutes later. As the figures in Table IV show, the initial rate of respiration is the same as in the untreated controls. The rate then rises in both, but more slowly in the treated eggs. At the end of the run their respiration is less than three-fourths of the control rate. It is evident here that the failure of cytoplasmic cleavage is accompanied by a slower rise in respiratory rate. The rate does, however, actually rise. It may be pointed out in this connection that nuclear division goes on in the treated eggs but at a retarded rate. TABLE IV Oxygen consumption of phenylurelhane-treated eggs. Eggs placed in 5 X 10~* N phenylurethane after appearance of second polar body (45 minutes after insemina- tion). Measurements begun at 85 minutes after insemination and 40 minutes after immersion in phenylurethane. Values given as cu. mm. Os per hour per mg. N. Temperature = 22.0° C. Hour Control Control Phenylur. Phenylur. 1st 4.61 4.64 4.86 5.04 2nd 4.94 4.92 4.69 4.83 3rd 5.66 5.48 5.40 6.71 4th 6.99 7.05 6.72 5.13 5th 9.84 10.02 9.16 10.84 6th 13.34 13.54 10.80 9.26 7th 16.30 16.40 12.30 14.01 8th 18.83 18.79 13.68 14.01 9th 21.61 21.33 15.43 15 98 Runnstrom (1928) investigated the action of phenyl- and ethyl- urethane on the respiration of sea urchin eggs. The concentrations that he used gave a depression of the initial rate; nevertheless a distinct rise is manifested during the three-hour period of the experi- ments. He notes too that nuclear division proceeds although the respiration during the first hour may be only 35 per cent of that of the control. DISCUSSION The curves of Fig. 1 summarize the results. It is readily seen that the parthenogenetic eggs, although starting out at about the same rate as the fertilized eggs, do not give as rapid a rise with time. Tin- t\\o types "t parthenogenetic eggs give approximately the same values (luring the early stages, but later the dividing eggs manifest a RESPIRATORY RATE AND CLEAVAGE 105 more rapid rise. If only the early period were considered this might be taken to mean that the rise in respiratory rate is not at all correlated with cleavage. However, as pointed out above, the cleavage parthe- nogenetic eggs divide at a retarded rate. Also, it is evident that in normal fertilized eggs no appreciable rise occurs during the early cleavage. It is therefore fairly safe to conclude that the rise in rate is linked with cleavage. But that it is not merely a matter of cyto- plasmic cleavage is clear from the fact that in the non-cleaving parthe- nogenetic eggs and the phenylurethane-treated eggs the respiratory rate does rise with time. Here, as we have seen, nuclear division goes on and it is probably with this factor that the rise is connected. 7 9 II 13 Hours after fertilization 15 17 19 FIG. 1. Rate of oxygen consumption of Urechis eggs. Curves A and A'; fertilized eggs. B and B'; non-cleaving parthenogenetic eggs. C and C'; cleaving parthenogenetic eggs. D; phenylurethane-treated eggs. Temperature 22° C. Values are averages from all of the data. A', B' and C' are not direct continuations of A, B, and C since the eggs had been cultured at 18.2° C. for 11 hours. Similar experiments have recently been independently performed by Brachet (1938) on Cliaetopterus eggs that " differentiate without cleavage" and he has kindly allowed us to examine his manuscript. The results with Chaetopterus agree very well with those presented here. In addition Brachet has investigated the nucleic acid content and finds it to be much lower in the parthenogenetic (non-cleaving) eggs than in the fertilized controls. In experiments of this type it is, of course, important to rule out possible direct effects of the chemical agent employed. It is reasonable to assume that there has been no direct effect when the initial rate is unaltered. Such is the case in the experiments reported here with 106 ALBERT TYLER AND N. H. HOROWITZ the exception of the cleaving parthenogenetic eggs (Table II) which showed a significantly high initial respiratory rate. This is very likely due to the ammonia treatment and from the subsequent values we might assume that the effect passes off. In any event, the differ- ence would have to be greater and in the opposite direction to seriously affect the conclusion. As was pointed out in the introduction, we might expect no rise in respiration if cleavage were blocked by some means that does not alter the absolute respiratory rate at the particular stage. But we are not dealing here with such ideal cases. In these experiments cytoplasmic division fails (non-cleaving parthenogenetic eggs) or is prevented (phenylurethane-treated eggs) while nuclear division pro- ceeds at a retarded rate; or both cytoplasmic and nuclear division proceed at a retarded rate (cleaving parthenogenetic eggs). It is not surprising, therefore, to find that the respiratory rate does increase in these cases. That the parthenogenetic show a slower rise than the fertilized eggs is consistent with their slower development, as is also the difference between the phenylurethane-treated and the control eggs. The difference between the two types of parthenogenetic eggs may likewise be interpreted as due to differences in the rate at which comparable stages of development are reached. SUMMARY 1. Artificially activated eggs of Urechis respire at the same initial rate as do normally fertilized eggs. 2. The rate of respiration rises with time in the artificially activated eggs, but at a much slower rate than in the fertilized eggs. 3. The increase in respiratory rate with time is greater with cleaving than with non-cleaving parthenogenetic eggs. 4. Fertilized eggs in which cytoplasmic cleavage is inhibited and nuclear division retarded by means of phenylurethane give a retarded rise in respiratory rate, although the initial rate is the same as the control rate. 5. It is concluded that the delayed rise in respiration is linked with the slower development in all these cases. LITERATURE CITED BRACHET, J., 1938. The oxygen consumption of artificially activated and fertilized Chaetopterus eggs. Biol. Bull., 74: 93. HIKAIWA, J. K., AND T. KAWAMURA, 1936. Relation between maturation division and cleavage in artificially activated eggs of Urechis unicinctus (von Drasche). Biol. Bull., 70: 344. LII.LIE, F. R., 1902. Differentiation without cleavage in the egg of tin- annelid Chaetopterus pergamentaceus. Arch. Entw.-mech., 14: 477. RESPIRATORY RATE AND CLEAVAGE 107 LOEB, JM AND H. WASTENEYS, 1913. The influence of hypertonic solution upon the rate of oxidations in fertilized and unfertilized eggs. Jour. Biol. Chem., 14: 469. NEEDHAM, J., 1931. Chemical Embryology. The MacMillan Company, New York. RUNNSTROM, J., 1928. Struktur und Atmung bei der Entwicklungserregung des Seeigeleies. Ada. Zool., 9: 445. TYLER, A., 1931. The production of normal embryos by artificial parthenogenesis in the echiuroid, Urechis. Biol. Bull., 60: 187. TYLER, A., 1936. On the energetics of differentiation. IV. Comparison of the rates of oxygen consumption and of development at different temperatures of eggs of some marine animals. Biol. Bull., 71: 82. TYLER, A., AND H. BAUER, 1937. Polar body extrusion and cleavage in artificially activated eggs of Urechis caupo. Biol. Bull., 73: 164. WARBURG, O., 1910. Uber die Oxydationen in lebenden Zellen nach Versuchen am Seeigelei. Zeitschr.f. physiol. Chem., 66: 305. i:\TKklMI.\Tso\ LIGIA 1\ BERMUDA \'. FURTHER EFFECTS OF SALTS AND OF HEAVY SEA WATER T. CUNLIFFE BARNES (From tJte Osborn Zoological Laboratory, Yale University, and tlie Bermuda Biological Station) Four previous papers of this series (Barnes, 1932, 1934, 1935, 1936) have described aspects of the behavior, salt requirements and thermal range of Ligia baudiniana, an interesting isopod which is invading the land through the intertidal zone. The present report deals with fur- ther experiments on the capillary mechanism conveying sea water to the gills, reactions to filter paper saturated with diluted sea water, the protective action of Ca in hypotonic sea water and the combined effect of high temperature and heavy water. HABITS It was reported previously that the release of young from the brood pouch was observed only in specimens kept submerged in sea water, but during the past summer this phenomenon occurred in a few cases in females in air over filter paper moistened with sea water. On the other hand, the molting process has occurred only in air. No adult specimens have been taken in the sea, but in rare cases an isopod will enter sea water in a terrarium to feed. Ligia is provided with a capillary mechanism for keeping the gills moist without entering the sea. The first paper of this series stated that the uropods and spines are lowered into the water or onto a water film and the sea water then rises by this capillary path to the gills. However, Mr. M. D. Burkenroad has drawn my attention to a more important capillary conduit by which the water rises between the sixth and seventh legs to the gills and is then propelled by the gills down the uropods and spines (Fig. 1, A). The sixth and seventh leg on one or both sides may be used. Frequently the animal slides one leg over the other alternately at the start. The drainage down the uropods was described in a former paper (Barnes, 1935) in experiments in which sea water was dropped on the animals. The flow was followed in the present experiments by the addition of stains, fine particles or bubbles to the sea water. If the uropods are lowered when the last pair of legs are not drawn together, the first movement of the water is ii] i I lie capillary conduit of the uropods as originally described (which 108 EFFECTS OF MODIFIED SEA WATER ON LIGIA 109 may be observed by the movement of particles in the sea water). When the isopod is totally immersed, the currents produced by the beating of the gills in the surrounding sea water (indicated by the move- ment of particles) follow the same direction as the capillary circuit involving the rise of water between the last pair of legs (Fig. 1, B). The sixth or seventh leg was removed in some specimens to modify the method of obtaining water. In the absence of the seventh leg a specimen standing on moist filter paper usually takes up the position FIG. 1. Methods of moistening the gills in Ligia. (In these sketches the an- tennae and body segmentation are omitted.) A. Specimen in air on filter paper saturated with distilled water. Water rises between the sixth and seventh leg by capillarity and is drained down between the uropods as indicated by arrows. B. Specimen immersed in sea water. The arrows indicate the currents main- tained by the beating of the gills. C. Specimen with seventh leg removed standing on filter paper saturated with sea water. Sixth leg and uropods are in position for capillary circuit. D. Side view of specimen with sixth leg removed standing on film of sea water (stippled). E. Side view of specimen with sixth leg removed with abdomen lowered in film of sea water. F. Sixth and seventh legs permanently fused along segment indicated by dotted line (found on left side of one specimen). These partially fused legs functioned normally as a capillary channel. for the capillary circuit via the legs (Fig. 1, C). It feels about franti- cally with the sixth leg for the missing seventh leg to complete the capillary channel. No attempt is made to use the fifth leg. Similarly if the sixth leg is removed, the conduit is impossible. Some specimens "squat" on the water film and the water rises by capillarity on the whole undersurface of the body and gills (Fig. I, D). Other specimens with the sixth leg missing lower the abdomen and uropods and thus 110 T. CUNLIFFE BARNES secure water by capillarity (Fig. 1, £). It is clear that there are several methods of wetting the gills. Indeed, many specimens are found which have lost the uropods and spines. These simply lower the abdomen on a damp substratum. An interesting specimen was found in which the sixth and seventh leg were permanently fixed in the capillarity posture owing to the fusion of the last segments (Fig. 1, F). While walking, the double leg moved alternately with the seventh leg on the opposite side, but sometimes dragged. The joined legs functioned normally as a channel above the point of fusion. REACTION TO FILTER PAPER SATURATED WITH DILUTED SEA WATER Large filter papers (diameter 25") were cut in two, one half satu- rated with distilled water, the other with diluted sea water and placed TABLE I Reaction of Ligia to Filter Paper Saturated with Diluted Sea Water (The animals were tested in groups of four) Total number of Treatment of each half of paper isopods found on Ratio each half Sea water vs. distilled water 69 : 123 1 : 1.78 75 per cent sea water vs. distilled water . 27 : 73 1 : 2.74 50 per cent sea water vs. distilled water . 93 : 96 1 : 1.03 25 per cent sea water vs. distilled water . 109 : 73 1 : 4.9 10 per cent sea water vs. distilled water . 27 : 21 1.28 : 1 in a covered flat dish in a photographic darkroom having a light directly above the center of the dish. Four isopods (previously kept on sea- weed moistened with sea water) were placed in the dish and their distribution on the two halves was observed at ten to fifteen-minute intervals over a period of one to two hours. The dish was rotated 90° after each reading to eliminate any unsuspected source of orienta- tion. The animals showed a distinct tendency to collect on filter paper saturated with distilled water when the other half was moistened with sea water (Table I, and Barnes, 1935). It was found that this aversion was also shown to 75 per cent sea water, but not to dilutions of 50 per cent and below (Table I). The tendency to avoid sea water was most pronounced during the first observations in a given experiment and after an hour or more, when the paper containing distilled water was becoming dry, most specimens collected on the sea water side. Thus the ratio of specimens on the sea water side to those on the distilled water side was 1 : 3.3 EFFECTS OF MODIFIED SEA WATER ON LIGIA 111 for all the first readings and 2.3 : 1 for the last observation made after an average time interval of one and one half hours. THE PROTECTIVE ACTION OF CALCIUM IN HYPOTONIC SEA WATER Ligia survives for only about seven hours in 100 cc. of 25 per cent sea water, but as with many other forms, the addition of calcium protects the organism from hypotonic media, apparently by decreasing permeability. As in all the experiments with solutions, individual specimens were tested in 100 cc. of solution in finger bowls. As will be seen from Table II, there is a threshold for the Ca effect (at about TABLE II Longevity of Ligia in 25 per cent Sea Water Containing Added cc. f M CaCh added to 1 liter Average Maximum Coefficients of Number of 25 per cent longevity longevity variation specimens sea water hours hours 0 7 ± 0.33 15 4.7 33 2 6.5 ±0.20 11 3.0 30 5 7.6 ± 0.37 17 4.8 58 12 24.6 ± 3.66 119 14.8 30 15 24.1 ±3.70 144 15.3 20 20 51.5 ± 7.98 120 15.5 20 25 37.8 ± 5.09 168 13.5 30 50 38.5 ± 3.85 180 10.0 40 12 cc. | M CaCl2 added to a liter of 25 per cent sea water). It is also apparent that the addition of Ca above a critical quantity (20 cc. f M CaCl2 added to a liter of 25 per cent sea water) has little additional protective action. It is interesting to note (see Fig. 2) that the average survival in calcified 25 per cent sea water approaches the aver- age longevity in 100 cc. of natural sea water. THE REVERSIBILITY OF THE CALCIUM EFFECT Experiments were carried out to see if the protective action of calcium involves an irreversible chemical process. If so, preliminary treatment in solutions of high calcium content should lengthen the survival of isopods subsequently immersed in hypotonic sea water. However, it was found that preliminary immersion in calcified sea water from half an hour to over a day had little if any effect on the subsequent longevity of the treated specimens in 25 per cent sea water (see Table III). Likewise, preliminary treatment with sea water having a high content of sodium has no significant effect on subsequent action of hypotonic sea water. In all these tests many specimens 112 T. CUXLIFFE BARXES died during the preliminary treatment so that those tested in the 25 per cent sea water represented a selected population capable of with- standing the submerged state. It must be remembered, however, that the survival of Li^ia in any liquid medium is limited by unknown factors so that these selected specimens were already weakened by submersion. That these two factors balanced each otht-r is indicated by the similarity of the average longevity in 25 per cent sea water of untreated and treated specimens. Thus the 152 isopods exposed to 50 40 30 20 10 10 20 30 40 50 Ki<;. 2. Length of life of Ligia in 100 cc. of 25 per cent sea water with increasing amounts of CaCl2. Onlinates, average life in hours. Abscissa1, cc. of jj M CaCI2 added to 1,000 cc. of 25 per cent sea water. The graph represents 228 tests (see Table II). Lower dotted line is the expected survival in 25 per cent sea water. Upper dotted line is the average survival in natural sea water. The vertical lines are proportional to the probable errors. The sea water average represents 265 specimens and the 25 per cent sea water average, 63 specimens. These are cumulative totals for five seasons' work. In all cases an individual specimen was tested in 100 cc. of solution. previous solutions survived 8.9 hours in 25 per cent sea water compared to the usual average of about seven hours in this medium for freshly collected individuals. In all the tests the specimens were washed by rapid immersion in distilled water after treatment in tin- calcium solutions, but this had no effect on the subsequent survival in 25 per cent sea water. Thus controls treated with ordinary sea water and EFFECTS OF MODIFIED SEA WATER ON LIGIA 113 then washed in distilled water lived (he usual interval of about 7 hours in 25 per cent sea water. Experiments in which the isopods were tested in distilled water after treatment with CaCl2 give similar results (see Table IV). After exposure to f M CaCl2 the average survival was three hours in distilled water compared to 2.8 hours for untreated specimens. THE COMBINED EFFECTS OF HIGH TEMPERATURE AND HEAVY WATER It was previously shown that Ligia succumbs rapidly in a small volume of sea water at 38° C. The following experiments were performed to ascertain if sea water containing 99 per cent heavy water would protect the organism against heat death or else potentiate with high temperature by reducing the survival time. In each case 5 cc. TABLE III Longevity of Ligia in 25 per cent Sea Water after Treatment in Calcified Sea Water Subsequent Average average Maxi- Coefficients Number Initial solution length of longevity mum of of treatment in 25 percent longevity variation specimens sea water hours hours hours Sea water . . 12.2 6.2 ±0.30 13.5 4.8 34 50 cc. f M CaCl2 plus 950 cc. sea water .... 31.5 5.5 ±0.32 6.5 5.8 13 100 cc. f M CaCl2 plus 900 cc. sea water .... 21.1 6.5 ±0.33 9 5.0 22 200 cc. f M CaCl2 plus 800 cc. sea water .... 17.6 9.2 ±0.38 17 4.1 49 500 cc. | M CaCl2 plus 500 cc. sea water .... 0.5 10.5±0.39 18 3.7 40 400 cc. | M NaCl plus 600 cc. sea water .... 25.3 9.4±0.79 25 8.4 28 of sea water were evaporated and the salts redissolved in ordinary distilled water or in 99 per cent heavy water. It was found that death ensued twice as rapidly in heavy sea water at 38° C. The survival times in 5-cc. samples were 23 minutes for H2O sea water and eleven minutes for D2O sea water (see Table V) . Owing to the scarcity of heavy water, several specimens were tested in the same sample, but no appreciable difference was observed in the survival times of the first and last isopods treated. DISCUSSION The tendency of the isopods to collect on filter paper treated with distilled water rather than on sea water paper is of interest in connec- tion with the animals' aversion for the sea. The reaction appears to 114 T. Ct'XLIFFE BARXES TAHLE IV Longevity of Ligia in Distilled Water after Treatment with Calcium Subsequent Average longevity Maxi- Coefficients Number Initial solution length of in mum of of treatment distilled longevity variation specimens water minutes hours hours Xone 0 2. 8 ±0.2 5 5.2 8.9 22 500 cc. § M CaCI2 plus 500 cc. sea water .... 33 4.5 ±0.15 6 3.3 20 | M CaCU 12 3±0.20 4 6.7 12 be determined by the salt content of sea water since dilution of 50 per cent destroys the effect. It is possible that the salts on the paper stimulate the isopods to greater movement which would cause them to collect on the salt-free side. Gunn (1937) has described a hygro- kinetic effect in Porcellio scaber whereby the greater activity of animals in dry air causes them to collect in moist locations, but this mechanism is probably not responsible for the reaction of Ligia described above. As was shown in a previous paper (Barnes, 1935), specimens will collect on dry filter paper when the other side of the dish contains filter paper moistened with distilled water. This refers to moist specimens. It is not known what part the flushing mechanism of the gills plays in TABLE V Longevity of Ligia in 5 cc. of Ordinary and Heavy Sea Water at High Temperature Medium Tempera- ture Average longevity Maximum longevity Coefficients of variation Number of specimens Sea water °C. 38 seconds 1383 ±45.1 seconds 2005 3.26 27 Sea water containing 99 per cent D2O 38 660±57.0 1140 8.63 11 these reactions. Bateman (1933) found that salts become slightly concentrated in the blood of specimens kept in air and it is possible that flushing the gills, especially with distilled water, enables the animal to get rid of salts concentrated by evaporation. On the other hand, specimens which have been immersed in distilled water also show t he- aversion for paper soaked in sea water (Barnes, 1935). The well-known action of ('a as a factor enabling organisms to withstand hypotonic solutions (for reference cf. Barnes, 1934) is strikingly illustrated by Ligia, which survives almost as long in 100 cc. of 25 per cent sea water of approximately 0.015 M CaClo content as in 100 per cent of natural sea water. To the list of favorable artificial EFFECTS OF MODIFIED SEA WATER ON LIGIA H5 sea water solutions for Ligia described in previous papers may be added this new hypotonic mixture which is equivalent to sea water in which the Na, K and Mg content has been reduced to one fourth and the Ca content raised slightly. However, the maximum longevity in this medium is far short of the 297 hours observed in natural sea water. The very long survival of occasional specimens in sea water suggests that other factors besides salt effects are involved, such as the amount of food in the gut, the nature of the previous environment, the oxygen content, or the necessity for molting (which apparently does not occur in sea water). The high summer temperature of approximately 27° C. must also be considered. The failure of preliminary immersion in sea water rich in calcium to protect the organism against subsequent exposure to hypotonic sea water indicates that Ca forms a loose, rapidly reversible combination with material in the plasma membrane. The existence of a threshold concentration and of a limited range in which the effect of Ca is pro- portional to the concentration suggests that a surface reaction is involved. The rapid lethal action of 5 cc. of sea water at 38°, in which the ordinary water has been replaced by heavy water, was an unexpected result. The lower energy content of heavy water, which under certain conditions might be expected to protect an organism from high temperatures, was overbalanced by the well known toxic action of high concentrations of deuterium. Barbour (1937) has recently dis- cussed the toxicology of heavy water. In larger volumes of sea water or even fresh water Ligia withstands high temperatures for a much longer period, probably on account of the greater quantity of oxygen present. The 5-cc. samples used in the tests were exposed to the air in a Petri dish so that the oxygen content of both heavy and ordinary sea water was probably the same. SUMMARY 1. When presented with a "choice" between filter paper moistened with sea water or with distilled water, freshly caught specimens of Ligia tend to collect on the latter. Dilution of the sea water destroys this effect. 2. The survival of Ligia immersed in 25 per cent sea water with added calcium approaches the longevity in natural sea water. 3. Preliminary exposure of solutions of high calcium content does not protect Ligia from subsequent immersion in hypotonic sea water. 4. The lethal action of 5 cc. of sea water at 38° C. is enhanced by 99 per cent heavy water. 116 T. CUNLIFFE BARNES LITERATURE CITED BARBOUR, H. G., 1937. The basis of the pharmocological action of heavy water in mammals. Yale Jour. Biol. and Med. 9: 551. BARNES, T. C., 1932. Salt requirements and space orientation of the littoral isopod Ligia in Bermuda. Biol. Bull., 63: 496. BARNES, T. C., 1934. Further observations on the salt requirements of Ligia in Bermuda. Biol. Bull., 66: 124. BARNES, T. C., 1935. Salt requirements and orientation of Ligia in Bermuda. III. Biol. Bull., 69: 259. BARNES, T. C., 1936. Experiments on Ligia in Bermuda. IV. The effects of heavy water and temperature. Biol. Bull., 70: 109. BATEMAN, J. B., 1933. Osmotic and ionic regulation in the shore crab, Carcinus maenas, with notes on the blood concentrations of Gammarus locusta and Ligia oceanica. Brit. Jour. Exper. Biol., 10: 355. GUNN, D. L., 1937. The humidity reactions of the wood-louse, Porcellio scaber. Brit. Jour. Exper. Biol., 14: 178. SOME ASPECTS OF NORMAL DEVELOPMENT IN THE COLONIAL CILIATE ZOOTHAMNIUM ALTERNANS F. M. SUMMERS (From the Marine Biological Laboratory, Woods Hole, and Bard College, Columbia University) INTRODUCTION This is the first of a series of inquiries into some of the problems of organization and regulation in a colonial protozoan of a rather special type, Zoothamnium alternans, whose cells collectively possess in some degree many of the attributes of an individual organism. In order to fix some standard by which to judge experimental results it was first necessary to review in detail the characteristic features of normal development in a large number of cases. Although Claparede and Lachmann (1858) formulated a scheme of notation for cell lineage in this species, Faure-Fremiet (1930) was the first to make a comprehensive study of its development. His primary concern was the demonstration of cytological mechanisms by which potentialities for growth and ciliospore formation were assorted to cells in characteristic positions on the colony. According to his conclusions this "cytological" factor determines the general pattern of develop- ment. His second objective, a study of time as a limiting growth factor, has no immediate bearing on the subject to be treated here. The quantitative aspects and time relations in development are to be incorporated in a subsequent report. The data presented here augment in considerable detail the previous studies of Zoothamnium alternans, particularly with regard to the features of development which have an important bearing on the problems of regulation now being investigated. The work was done at the Marine Biological Laboratory, Woods Hole, Massachusetts, during the summer months of 1935-6. It is a privilege to acknowledge assistance from those whose friendly interest has sustained the author's enthusiasm over many of the dull periods of routine work. MATERIAL The basal portions of mature hydroids (Pennaria and Tubularia} and Bryozoa (Bugula} from many localities in Woods Hole harbor great numbers of attached colonies of Zoothamnium alternans. The 117 118 F. M. SUMMERS hydroid and hryozoan colonies are clean and free from the ciliates until about June 15. From then on throughout the summer the proto- zoa are very abundant, although their condition varies with particular habitats and with the water temperature. Grave (1933) found that the hydroids of this region are particularly healthy and free from bacterial growths, debris, etc. for the early spring months but that in mid-summer the contaminating organisms affect them adversely. This applies equally to the associated protozoa, for those obtainable in late July and August are commonly infested with parasites of several classes. The most annoying and destructive organisms are bacteria, filamentous algae, and an intra-cellular parasite which has been identified as a suctorian (Acineta) by Faure-Fremiet (1930); others of less consequence are small Crustacea, diatoms, and other ciliates, e.g. Amphileptits and Cothurnia. METHODS The procedure for maintaining selected colonies in aquaria is an elaboration of the technique employed by Faure-Fremiet (1930) for his study of growth in Zoothamnium alternans. Hydroids bearing the protozoa were cleaned, washed, and packed into large glass vessels of running sea water. Ruled and numbered slides were placed face down on the infected hydroid material and left for about 12 hours. They were then removed and washed by passing under a stream of running water. The precautions taken at these stages to remove adhering scum, debris, algae, other protozoa, and minute Crustacea, determined to a great extent the age to which the colonies could be grown . Some of the mature asexual migratory cells, the ciliospores (Wesen- berg-Lund, 1925), liberated from the "wild" colonies during the 12- hour interval, attached themselves to the ruled surface of the numbered slides. Slides were examined in Petri dishes, the positions of the recently affixed ciliospores recorded, and then transferred from the Petri dish to correspondingly numbered slots (vertical) in wooden racks, immersed in a 15-gallon aquarium. Cultural conditions were very much improved by the use of filtered sea water. Two 2-liter aspirator bottles packed with sand and stoppered with glass wool proved to be adequate for the purpose. When cleaned daily their combined outputs averaged about 50 gallons per hour. Routine counts were made with a 4 mm. dry objective, \\hen optical sections of the barely immersed colonies were not clear, the water level in the Petri dish was adjusted by a slight turn of the dish on the slanted microscope stage until the surface film flattened the NORMAL DEVELOPMENT IN ZOOTHAMNIUM ALTERNANS 119 colony. This was done as a last resort, for contact with the surface film elicited rapid contractile responses, frequently resulting in the loss of important zooids. 7MJ9 FIG. 1. A relatively mature colony of Zoathamnium alternans showing the characteristic arrangement of the zooids and branches. Axial microzooids on branches E, G, L, and N are differentiating into ciliospores; one of them at E is in an advanced stage of metamorphosis, almost ready to break away from the parent colony. Approximately X 100. GENERAL FORM OF THE MATURE COLONY The tapering frond-like configuration of the older expanded colonies is shown in Fig. 1. Its branches project obliquely from the per- 120 F. M. SUMMERS pendicular axis. They lie in the same plane but alternate on right and left sides at successive nodes. The zooids of the branches occur singly for the most part, likewise alternating in position along the branch axis. While the branches of this species usually consist only of a single axis, in rare cases the zooids may divide one or more times to produce branches of the second order. The expanded colony is essentially asymmetrical, i.e. characterized by three heteropolar axes. The primary axis is firmly cemented to the substrate at the basal end and bears the actively dividing terminal macrozooid (TM.) at the apex. The lateral axis is heteropolar by virtue of the alternating branches, each one terminating in a single cell, the dominant branch zooid. "Anterior" and "posterior" sur- faces (Faure-Fremiet, 1930) are defined by the curvature of the branches as well as by the position of the zooids: the branches curve anteriorly while the zooids are inclined posteriorly (toward the convex surface). A sample count showed that the first branch occurred nine times on the left side to four on the right. Both dexio- and leiotropic colonies were reported in another species, Zoothamnium arbuscula, by Furssenko (1929). Beginning at a point about mid-way between the basal (attached) end of the stalk and the first branch, a heavy contractile central cordon, or "neuro-muscular" cord, within the stalk substance ramifies through- out the colony (Fig. 1). Its branches terminate distally at the basal ends of the zooids. It is thus continuous from branch to branch and from cell to cell. More will be said about the importance of this structure in a later section. NORMAL DEVELOPMENT The scheme used to designate the lineage of the zooids is a modified form of Faure-Fremiet's adaptation of the original rule formulated by Claparede and Lachmann (1858-60), who were the first to note devel- opment in this species according to a definitely determined pattern. The mode of propagation in Zoothamnium alternans is almost exclusively asexual. Axillary branch zooids at various levels in larger colonies are transformed into migratory ciliospores. These liberated motile individuals are comparatively large cells, metamorphosed zooids, whose sub-conical bodies flatten antero-posteriorly into cells having a thickened bi-convex lens-like appearance. Their rapid swimming or creeping movements are effected by means of a heavy equatorial girdlet of cilia. The oral disc and peristomal cilia are retracted and probably undergo reorganization during the motile phase. NORMAL DEVELOPMENT IN ZOOTHAMNIUM ALTERNANS 121 After a migratory existence of several hours duration the ciliospores become relatively quiescent, hovering about within a limited radius with aboral end in contact with the substrate, finally fixing themselves to the surface by means of a scopula. The secretion of the stalk begins ; the girdlet of cilia becomes inactive and disappears as the peduncle elongates. The lens-shaped ciliospore then begins to assume the appearance of a very large sub-conical zooid with retracted adoral zone. The first section of the peduncle secreted is a fairly heavy, non- contractile cylinder with thickened cuticle and homogeneous hyaline medullary substance. At some point between 200-30(V the secretory activity of the ciliospore is altered: the continuity of the secretory process is interrupted, the basal end of the cell increases in diameter, and the formation of the neuro-muscular cord begins. This node marks the lower limit of stalk contractility. The contractile cord is an integral part of the stalk structure from this point onward. The division of the initial individual occurs approximately 15 hours after attachment. The first division furrow defines the median plane of the future colony. Succeeding furrows bear a definite and constant relation to the first, in such a manner that the daughter cells occur first on one side of the mid-line then on the other. The two daughter cells resulting from the first division are always markedly unequal in size (Fig. 2). The large daughter remains apical (terminal) in position. FIG. 2. Developing colony at the 2-cell stage. Drawn from a living specimen in which the macronuclei were faintly visible. X 250. In the current terminology it is designated as the terminal macrozooid No. 1. The small lateral daughter, the first cell of the first branch, is characterized as the median microzooid A. Dichotomous branching of the stalk occurs when the secretory basal portion of the cell constricts during the fission process. With the onset of secretory activity (i.e. stalk formation) the sister cells gradually move away from the point of origin. Material secreted by the terminal macrozooid prolongs the axial stalk. The branch cell produces a somewhat smaller stalk extending laterally from the junction. 122 R M. SUMMERS The apical cell (TM. No. 1) divides again at the second node in approximately 8 hours, producing an apical cell of the second genera- tion (TM. No. 2) and a median microzooid B. This time, however, the branch cell lies on the opposite side of the axis from the previously produced microzooid A. Repeated divisions of the apical cell extend the principal axis of the colony, each time similarly producing a ter- minal macrozooid of the succeeding generation (3, 4, 5, . . .) and an alternating branch zooid of a higher level (C, D, E, . . .). Tin- position of successive branch microzooids is strictly alternative. Colonies with as many as 25 branches were frequently observed on the culture slides but it is doubtful that this represents maximum develop- ment. One colony produced 33 branches before succumbing to external parasites. Following the first division of the apical cell there is a gradual reduction in the size inequality of the resulting daughters until, after 8 to 10 generations of axial development, or in some cases even fewer, their volumes immediately after a mitotic period are approximately equal. For experimental purposes it is essential to distinguish between recently formed daughters of later generations along the main axis. To this end it is necessary to refer back to the previous generation. If the left-hand cell of that generation is differentiating into a branch cell, then the right-hand position of the undifferentiated cell at the new mode may be taken as assurance of its presumptive branch rela- tions. Likewise position is tin- chief diagnostic criterion for the identification of undifferentiated branch zooids. The generalized pattern of branch development given below con- cerns only the position of the cells produced, for as will be seen later, the descendants of the median microzooids are not always equivalent in their prospective values. The symbol X (or .v) is used to denote branch generations in general (Fig. 3). Lateral extension of the stalk bearing the median microzooid A" continues until the cell divides. Microzooid A' gives rise to two struc- turally similar cells, microzooids 1AT and l.v. Microzooid 1A" occupies a median position with respect to the main axis of the colony, whereas \x is the lateral member. Actually l.v immediately assumes the terminal position on the branch axis, thus constituting the growing point of the branch. Marked differential behavior now characterizes the two sisters: 1A' lays down a short segment of stalk and then remains quiescent for some time ; on the other hand, l.r actively extends the branch axis, then divides again within a few hours. Microzooid l.v then produces 2*1 (median) and 2x2 (lateral). Cell 2.v2 remains behind as a common nutritive branch zooid while 2.Y1 assumes the terminal NORMAL DEVELOPMENT IN ZOOTHAMNIUM ALTERNANS 123 position and continues as before, giving origin to 3.x:1 (lateral) and 3jc2 (median) ; 3.V1 to 4*1 and 4jt2, etc. Each time the alternate secondary daughter is left behind as a common zooid (Fig. 1). Further mitotic activity in the common branch zooids 2x2, 3xz, 4#2, etc. was not characteristic of the normal colonies whose histories were charted. About 2 per cent of the total number of colonies under surveillance showed a tendency toward precocious development. In these cases the repeated divisions of all branch zooids led to such complex formations that the lineage soon became impossible to follow. Conversely, microzooid \X on branches up to about level J com- monly divides once, forming \Xi (median) and IX2 (lateral). Its mitotic activity on more distal branches is infrequent. On several Presumptive Ciliospores FIG. 3. Schematic representation of branch development. branches of each colony the cell IX or its descendants metamorphose into the migratory ciliospores which break away from the parent colony, leaving behind a short, stubby peduncle. Occasionally IX produces a second or sometimes a third generation thereby augmenting the number of propagatory zooids. DIFFERENTIATION The distribution of the heteromorphic zooids to be described is a generalized account based upon a tabular evaluation of 200 cases of normal and experimentally cut colonies (proximal portions only), many of which were lost before complete development. Inclusion of the basal parts of operated colonies in the tabulations seems to be justified since statistical summaries of both normal and operated 124 [-. M. SUMMERS colonies indicate no consistent variation as regards either rate of development or position of the differentiating zooids. 1 hiring the life history of a colony six types of zooids may be formed although not all of them are apt to be present at any one time. 1 . The common nutritive microzooids constitute the majority of the colony cells. They are the trumpet-shaped individuals characteristic of the genus. Notwithstanding the fact that most of them never assume any other form, they may be regarded as generalized com- ponents of prime colonies because (a) under undisturbed laboratory conditions some of them at random positions along the branch axis metamorphose into microgamonts, and (b) experiments may be per- formed in which variously situated common zooids may be induced to differentiate into actively reproducing apical cells. Their capacity to differentiate into another type under imposed conditions gradually diminishes in time so that in older colonies the "vegetative" function becomes fixed. They can then be considered as somewhat specialized zooids. The shape of zooids that have persisted for relatively long periods of time without division are decidedly more slender and elongate than the "average" zooids. This appears to be a modifica- tion of form rather than an alteration of volume. The extension of the protoplasmic body as well as its peduncle also appears to be gradual and continuous subsequent to the post-mitotic growth period. 2. Each colony bears a single terminal macrozooid (TM.} at the apex of the principal axis (Figs. 1 and 2). In the earlier stages of colony formation three structural peculiarities clearly distinguish it from the common zooids. It is considerably larger than any of the common types for the first ten or more axial generations, the body is characteristically flexed, and numerous annular striations in the pellicle are very pronounced. After ten generations or so its form individuality progressively diminishes. At levels beyond the twentieth branch only its position and generative activity are distinctive. 3. The terminal branch zooids (Lv, 2.v', 3.r', . . .) are almost diminu- tive replicas of the terminal macrozooid. For the first few generations they are identified shortly after dividing by fairly marked pellicular annulations and a moderate anterior flexure of the tapering body. Near the close of branch development their patent features are obscure. Standard criteria of position and, to some extent, proclivity toward more rapid stalk formation and division must be re-lied upon to dis- tinguish them from their common neighbors, the secondary daughters. Finally, at the terminus of a fully formed branch the last pair of zooids are usually indistinguishable. This point represents about four generations on branch A, ten at intermediate levels (e.g. //), and progressively fewer after that (Fig. 1). NORMAL DEVELOPMENT IN ZOOTHAMNIUM ALTERNANS 125 4. The most striking of the heteromorphic zooids are the immature ciliospores, or asexual propagative cells which represent differentiated microzooids of the order \X or descendants. It is important for later discussion to note that the function of ciliospore formation is normally but not invariably restricted to these microzooids. They have been observed to arise from microzooids 5d2 and 6d~ in one instance, and K)//1 and 10//2 in another. Faure-Fremiet (1930) found that the terminal macrozooid was sometimes transformed into a migrating ciliospore but, unlike the usual ciliospore-forming cells, was unaccom- panied by an endomictic reorganization. Prior to ciliospore formation microzooid IX undergoes a profound growth in size. Very accentuated concentrically disposed pellicular furrows appear during the growth process. It is impossible to predict whether or not it will divide, for in a small percentage of the cases it is transformed directly into a motile ciliospore thereby terminating the lineage of that cell on the colony. Division of IX in from 20 to 70 hours after derivation from the median microzooid X is the rule, however. It is noteworthy, perhaps, that fission may occur after the onset of differentiation. The increased volume at this time is con- siderably in excess of the pre-divisional growth of the other zooids. The two descendants, IX1 (median) and l^"2 (lateral), are seldom equivalent in their size relations; the lateral daughter is usually the larger and the first to metamorphose. Complete differentiation into motile cells already characterized involves the appearance of a girdle of long, closely set cilia about the equatorial region, the accentuation of the annulations, introversion of the oral disc, and differential growth in the plane of the diameter, eventually flattening the zooids into disc-shaped bodies. It is believed that they are able to feed to some extent when the oral disc is intro- verted and the adoral zone folded. The gullet remains open and the undulating membrane within persists in its activity. The final metamorphosis of IX1 may be deferred until a second generation of potential ciliospores (IX11 and l^12) are produced. The origin of at least four ciliospores from some axial microzooids has been observed. Whether or not they represent three generations of alternate develop- ment or two generations formed dichotomously is still an open question. The production of ciliospores is restricted to a limited number of loci on any given colony. Commonly there are but three or four such loci on one colony, although one protocol shows as many as seven, generalized colony as visualized from tabulated data bears florescences at levels D-K-P-T-AA. They have been recorded for all nodes along 126 F. M. SUMMERS the axis between B to IT inclusive, hut never at consecutive nodes in any one case. It should also be stated that 1A" on many branches divides without giving rise to ciliospores. Its descendants persist as common microzooids. 5. Microgamonts (microgonidia, Wesenberg-Lund, 1925; micro- conjugants, Fursscnko, 1()29) develop from otherwise visibly undiffer- entiated microzooids at the terminal or sub-terminal position along branch axes. The observed time interval between the formation of the microzooids and their complete transformation into mature, liberated swimmers varies between the limits of six to thirty-six hours. The first sign of approaching metamorphosis is the appearance of a faint furrow about the equatorial region of the tapered body. The zooid is thus marked into a larger peristomal portion and a smaller basal portion. Cilia gradually grow out from the body at the furrow. At first very short, they soon develop into a broad girdle whose move- ments simulate those of a slowly waving undulating membrane (Fig. 4). Meanwhile the diameter of the cell increases. The oral half FIG. 4. Differentiating microgamont about 5 hours before the completion of metamorphosis. X 250. containing the retracted peristomal disc is transformed into the slightly flat, sub-conical oral end of the swimmer. The posterior portion is further reduced to a very slightly convex aboral surface at whose center the stalk is still attached. Its struggling movements gradually accelerate until at full maturity it breaks away from the parent colony. At full term the microgamonts resemble ciliospores in their eccentric body outline, ciliature, and general movements. They are, however, not perceptibly larger than the common zooids, show but faint indica- tions of annulations, and swim more rapidly. While cytological studies were not attempted, it is definitely es- tablished that these are the migratory members of the copulatory series. When liberated some have been observed to swim rapidly back and forth in an arc about the parent colony, occasionally coming to rest on some zooid, or crawling about on the branches and over the individuals in various positions, and eventually swimming away as t hough its relatives were not receptive of its attentions. The duration of the migratory phase was not determined. One was observed to settle upon terminal macrozooid No. 3 of a young colony possessing nine cells in all. Two and one-half hours later the aboral end of the NORMAL DEVELOPMENT IN ZOOTHAMNIUM ALTERNANS 127 migrant was firmly fused with the lateral surface of the apical cell. By that time the ciliary crown of the swimmer had been resorbed. As regards origin, it is frequently difficult to determine if micro- gamonts arise from prospective terminal or sub-terminal branch microzooids because of the facility with which the prospective values of very recently produced daughter cells may be altered. For example, if the terminal cell (e.g. 2x1} is accidentally lost, the sub-terminal cell (2.r2) immediately assumes the terminal functions. If loss occurs soon after fission, when both daughters are superficially alike, the stalk remnant left behind at the node occurs in the alternate position, whereas the cell which normally would have been the alternate (2^2) becomes the growing point. The tendency of the terminal branch cell to differentiate into a microgamont is conclusively demonstrated by the successive metamorphosis of sister cells. This means that the continued lateral growth of a branch subsequent to the production of a microgamont may represent each time an early (undetected) trans- formation of the secondary zooid into a new terminal cell. The con- fusion is apt to arise because the alternate position of the stubby stalk left by the lost cell or by the migrating microgamont cannot always be diagnosed accurately. Fortunately in many instances the terminal branch cell is well indicated either by position or by symptoms of an approaching division by the time the alternate cell shows signs of becoming a microgamont. Tabulations show that 85 per cent of the microgamonts observed on colonies whose complete developments were charted, were produced by the four basal branches A, B, C, and D, enumerated in decreasing order of importance. The rest arose from zooids on branches E, F, G, and K. They were formed either by the terminal (-xl) or alternate (-x2) cells as far out on the branches as the sixth generation on A, B, and C (including 5a21 and 5a22), the seventh generation on D, and the eighth generation on F. None were observed to come from the median microzooid X, the axial microzooid IX or its descendants. The majority of the colonies bore but one or two microgamonts; a few produced as many as nine but not more than three for any one branch. They occurred simultaneously on different branches but successively from the alternate zooids on any given branch. And finally, it is important that they differentiated only from branch zooids at or near the terminal position, never from the more axial (older) microzooids. 6. Macrogamonts were found only in the terminal macrozooid position on the primary axis. Attempts to discover some morphologi- cal or behavioral distinction between terminal macrozooids and pre- 128 R M. SUMMERS conjugant macrogamonts were negative, but this possibility is not precluded since of the three or four hundred colonies inspected, none may have happened to have been preparing for conjugation. The presence of an attached microgamont constituted the sole criterion for identification. Conjugants at all levels between the terminal macrozooid positions three to twenty-four inclusive were observed. The youngest colony possessed but nine cells in all, including the conjugant. Two succes- sive conjugants, five generations apart, were recorded for one colony. The incidence of conjugation reached a maximum during the month of July; before or after that the conjugants were quite rare. Even in July only about 5 per cent of the colonies bore conjugants although more of them were producing microgamonts. Many of the colonies produced no gamonts whatever during the period of development on the slides. Some of the effects of conjugation upon colonial development will be treated in a later section. REACTIVE CAPACITIES The motor responses of Zoothamnium alternans are developed to a relatively high degree. Very local reactions by individual zooids result in the retraction of the ciliated oral disc. More general responses evoked by altered salinity of the water, sudden contact stimuli, etc., are expressed by an extremely rapid contraction of the stalk and the zooids. Gradual relaxation or extension ensues within fifteen to thirty seconds unless the stimulus is sustained. This manifestation of irritability constitutes one of the major technical difficulties for experimental work. When the colony is contracted, it is practically impossible to distinguish a given zooid or to cut the stalk at a desired point. Fortunately for the experimental work the stalk alone may be touched gently without disturbing the colony. A series of attempts to inactivate the colonies by subjecting them to a variety of narcotics in various concentrations were quite unsuccessful. Freezing tempera- tures did not modify the contractions to any great extent. SUMMARY 1. The frond-like colonies of Zoothamnium alternans develop ac- cording to a well-determined pattern. This development is described with particular reference to the origin of the heteromorphic zooids. 2. An asexual propagative zooid (ciliospore) detaches from a mother colony, swims away, and settles down to form a new colony. When affixed to the substrate the ciliospore begins to secrete a peduncle or NORMAL DEVELOPMENT IN ZOOTHAMNIUM ALTERNANS 129 stalk. THe first portion of the peduncle is flexible but not contractile. From a point some 200 to 300/i above the hold-fast, the ciliospore suddenly begins to produce a contractile cord in the core of the hyaline stalk substance. Distal to this point the contractile cord is an integral part of the stalk; it is continuous from branch to branch and from cell to cell. 3. When the stalk is approximately 500/x long the ciliospore divides unequally. The larger daughter remains axial in position. It re- presents the first generation of the terminal macrozooid series (TM. No. 1). The smaller lateral cell is the initial zooid of the first branch, the median microzooid A . 4. Successive divisions of the terminal macrozooid produce each time a terminal macrozooid of the next generation and a median micro- zooid. The latter are strictly alternate in position: they alternate on right and left sides of the primary axis at successive nodes. As many as thirty-three generations of the terminal macrozooid have been observed. 5. The first division of the initial branch zooid gives rise to an axial microzooid and a lateral stem cell. The stem cell generates alternating zooids of the main branch strain whereas the axial microzooids on some of the branches represent the presumptive ciliospores. 6. At least four types of zooids comprise a colony: (1) a single terminal macrozooid at the apex of the primary axis, (2) many common microzooids scattered along each branch axis, (3) a terminal zooid at the tip of each branch, and (4) immature ciliospores which arise from axial microzooids on some of the branches. During epidemics of conjugation two more types may be formed: the terminal macrozooid may be transformed into a macrogamont, and common microzooids at random positions may metamorphose into migratory microgamonts. The spatial distribution of the heteromorphic zooids is described in some detail. LITERATURE CITED CLAPAREDE, E., AND J. LACHMANN, 1858-60. Etudes sur les Infusoires et les Rhizopods. Mem. de Vlnstiiut Genevois, 5-7. FAURE-FREMIET, E., 1930. Growth and differentiation of the colonies of Zootham- nium alternans (Clap, and Lachm.). Biol. Bull., 58: 28. FURSSENKO, A., 1929. Lebenscyclus und Morphologie von Zoothamnium arbuscula Ehrenberg. Arch. f. Protist., 67: 376. GRAVE, B. H., 1933. Rate of growth, age at sexual maturity, and duration of life of certain sessile organisms, at Woods Hole, Massachusetts. Biol. Bull., 65: 375. WESENBERG-LUND, C, 1925. Contributions to the biology of Zoothamnium geniculatum Aryton. D. Kgl. Danske. Vidensk, Selsk. Skr., nature . og. math. Afd., 8. Raekke, X: 1. FORM REGULATION IN ZOOTHAMNIUM ALTERNANS F. M. SUMMERS (From Bard College, Columbia University) INTRODUCTION Interest in morphogenetic studies has long centered about the part- whole relationships obtaining throughout the formative period in organic development. In so far as a system comprising several parts is concerned, the term organization implies the existence of integrating factors that condition to some extent the limits and direction of regional specialization. A remarkable number of investigations on metazoan organizers have already demonstrated the importance of extrinsic factors upon determination in specific parts. It was felt that additional information about these factors could be gained by applying operative techniques to an animal type in which, presumably, the interrelationships have not attained so great a degree of complexity. The principal endeavor of this work was to investigate some of the qualitative and quantitative aspects of growth and differentiation in Zoolhamnium colonies, regulating after the removal of actively growing (distal) parts, and to compare regulative behavior with the normal development already described by Faurc-Fremiet (1930) and Sum- mers (1938). The results have made it possible to offer a rough map of potencies and prospective values of individual cells at various positions in the colonial pattern and to indicate some of the changes in the expression of inherent potencies which may be induced by experi- mental means. It is a pleasure to thank Professor L. R. Cleveland of Harvard University for extending to the author the privileges of his laboratory during January, 1937. MATERIALS AND METHODS The materials and techniques used in this study are similar to those described in the previous paper and need not be repeated here. The only additional detail pertains to the method of shearing the stalk. Two fine scalpels made from No. 9 sewing needles were used for the purpose. One was brought to rest against the surface of the stalk and the other sheared against it in scissors fashion. The plan of attack is by no means new, but the type of organization 130 FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 131 dealt with seems to promise a fresh approach to the current problems of form determination. A colony of Zoothamnium alternans is ad- mirably adapted to work of this kind by virtue of the regularity and precision with which the characteristic colonial pattern develops. The alternating arrangement of the branches and cells makes it compara- tively easy to follow the history of any one cell throughout the course of its development for evidences of growth, division, or differentiation. The spatial relationship of the cells minimizes to a great extent some of the factors so difficult to evaluate for compact tissues. Crowding effects such as mutual contact, pressure, etc. (Peebles, 1931) are of no great concern here. Then, too, the separated cells are uniformly bathed by an almost constant medium, filtered sea water. Physio- logical relations between them are effected through a well-defined channel, the stalk with its neuro-muscular cord (Faure-Fremiet's "cordon central"), rather than through the general expanse of juxta- posed cell membranes. REGULATIVE DEVELOPMENT Standards of Judgment. — Colonies maintained for several days on slides are apt to be attacked by internal parasites or covered by plant growths of one kind or another, especially in the basal regions. When operations are made the axial growth of a colony is retarded for an average of 23.1 hours pending the formation of a new terminal macro- zooid. It is during this period of arrested development that adverse environmental conditions are liable to bring about an incapacitation or loss of important zooids before a decision relative to the success of the operation can be reached. A small proportion of the successful operations shown in Tables I and II do not appear in subsequent tables because they were destroyed or abandoned after indubitable signs of new terminal macrozooid differentiation had appeared but before descendants were produced. In the absence of a regenerate, the responses were recorded only when all of the structural character- istics of the new terminal macrozooid were established and, in addition, the "activated" branch developed an anterior flexure. In conse- quence of the stalk curvature the new terminal macrozooid assumes the apical position upon an anteriorly directed axial stalk. The point of curvature marks the node (Figs. 3 and 5) at which the stalk suddenly increases to a diameter approaching that of the original axis. Regenerative responses were arbitrarily called negative only when one of the following conditions were realized: (a) there was no activity for at least 48 hours; (b) in the event that mitotic activity continued for a generation or two, a minimum of 72 hours was allowed for signs 132 F. M. SUMMERS of regulatory activity; (c) when the terminal branch zooid in the line of succession metamorphosed into a migrating zooid of some kind. The thirty-seven negative cases shown in Tables I and II were main- tained for a mean time of 94.3 hours after the last division, with ex- tremes of 52 to 212 hours. TABLE I PARTS CUT AWAY TM+O TM.-H TM.+2 TM.+3 TM.«-4 TM.+5 TM.+6 TM+7 Total - 1 1 P (1) (2) (1) (4) O 2(1) 1 (1) (D (D 3 (4) N (D (D z 0 H M L (D (1) (1) 1 (2) (2) OPERA K J 1 1 2(1) 1 (1) 1 (O 1 (I) 1 (2) 2 (1) 3 (1) O H 1 (1) 0) 1 (2) LJ LJ G F 3 (1) 4 1 (1) 1 (1) (0 4 (4) 5 E 1 1 1 3 D 3 2 5 C 6 4 1 1 12 B 3 (1) 4 (1) 2 1 (D 10 (3) A 9(6) 6 (3) 6 2(2) 1 1 1 1 27 (1 1) Total ^7(13) 21 (12) 9 (3) 5(6) 3 (1) i 0) 1 0) 1 78(37) The distribution of regulative responses in 115 operated colonies summarized according to the number of apical branches cut away: T.M. + 0 = only the terminal macrozooid removed; T.M. -f- 1, T.M. + 2, etc. == terminal macrozooid plus one, two, or more branches removed. The letters in the left-hand column designate the branches from which the regenerates arose. The numbers of cases in which no response occurred are shown in parentheses. The first responses of the successively operated colonies are included. No attempt has been made to indicate the number of zooids on the branches dissected away; it may be estimated by referring to Fig. 1 in the previous paper. All colonies were abandoned when the important zooids appeared to be unhealthy. Adjustment of Descriptive Notation.— \\\ arbitrary departure from the standard notation (see Faure-Fremiet, 1930; or Summers, 1938) seems to be advisable in view of the complications arising from the FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 133 designation of generations produced by the secondary zooid near tin- point of origin of the new terminal macrozooid. To illustrate: if the new terminal macrozooid differentiates directly from 2al, then the secondary zooid 2a2 continues to generate zooids of the branch A, viz. 2#21, 2a22; 2a211, 2a212, etc., which complicates a system already difficult to summarize briefly. At this point it is proposed to adjust the terminology so that the new terminal macrozooid (actually 2&1) corresponds in position to the TABLE II ORIGIN OF NEW TERMINAL MACROZOOID IX I IX1 Ix 2x' 2x* 3x' 4X1 4x2 5x' 6x' Total Q 1 1 P (D (0 (0 (D (4) O 2 ' (D (2) (1) 3(4) N (D (0 z o M (D 1 (D 1 (2) 1- < L 0) (0 (2) cr Ul K 1 0) 0) 1 (2) Q. O J 2 (0 2(1) Li. 1 1 1 1 (O 3(0 o _J H 1 (D (D 1 (2) Id G 1 2 1 (2) 0) (I) 4(4) LJ _l F 2 2 1 5 E 1 1 1 3 D 1 3 1 5 C 8 3 1 12 B 5 (2) 2 1 1 1 (O 10(3) A 1 0) 0) 15 (7) 5 0) 5 (1) 1 27(1 1) Total 5 (1) 2 (1) 37 (1 1) 19 (6) 3 (2) 8 (6) 4 (2) (3) (4) (1) 78(37) Lateral distribution of operations. This table summarizes the data from Table I according to regulation by branch generations irrespective of the amount of colony cut away. The left-hand column represents the branches from which the regenerates developed; headings indicate the position and generation of the zooids which produced the new terminal macrozooid. The negative responses are shown in parentheses. original terminal macrozooid (T. M. No. 1), but is distinguished from it by a prime (') symbol. The lateral branch zooid 2a2 becomes A', the first zooid of a new branch at the original A level (Fig. 1). This adjustment is valid if subsequent products are to be treated descriptively as new primary branches rather than as branches of the second order on the first branch below the cut. This is nothing more than a manipulation of terms to facilitate comparison of normal and regulating colonies. F. M. SUMMERS Trauma. — For a number of reasons it appears probable that traumatic shock effects are not significant factors in post-operative adjustments of colonial form. Cells adjacent to the cut areas and elsewhere soon expand and feed as before. Processes of mitosis or REGENERATING PORTION ( TM.2' ) / <3a!; \_y / \ i TM.I' i * / •~^< ?2a') FIG. 1. Diagram of regulative development. The revised notation for desig- nating cell lineage in regulating portions is indicated within the circles. The symbols in parentheses under the circles illustrate how cumbersome the conventional termi- nology would soon become if applied to the regenerates. The diagram shows the lineage of a regenerate from branch .1. It the principal axis is severed between branches A and fi at a time when there are three cells on branch .1, the regenerate usually arises from the terminal branch zooid 2al as illustrated. The sub-terminal zooid (2a2) assumes the terminal branch position A' and continues to generate branch cells. The new branch B' is produced by the first division of the new terminal macrozooid T.M. \'. The parts produced after cutting are drawn with broken lines. differentiation, in progress at the time of operation, continue without perceptible interruption. Or these processes may begin at varying intervals after cutting, in any of the branch zooids, except that which becomes the new terminal macrozooid. Furthermore, when there is FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 135 no regulation, or when the median microzooid divides one or more times before the new terminal macrozooid is recognizable, the first branch below the cut continues to generate common zooids as before. Relative to division rates, the available data indicate that, exclusive of the one which bears the presumptive terminal macrozooid, branch growth is not perceptibly altered after cutting. As a rule there is a lag in the development of the activated branch pending the differentia- tion of a new growing point. The only effects of mechanical disturbance are evidenced for a very short time after cutting by a state of irritability during which the contraction of decapitated colonies is frequent, irregular, and some- times tetanic. But normal overt behavior is resumed within a few moments when the stalks are shorn cleanly at internodal points. Operations were considered acceptable only when the normal reactions were regained within a relatively short time. Cases where only the neuro-muscular cord of the stalk was injured are to be treated in another section. Distribution of Operations. — A general resume of the experimental results in terms of initial regulative responses in Zoothamnium alternans is given in Tables I and II. Of the 144 protocols at hand (acceptable operations), 78 yielded positive responses and 37 were negative; the remainder were inconclusive according to the standards chosen and are omitted in the digests. Table I summarizes the responses to various types of cuts made at the several levels along the principal axis irrespective of the number of generations on the regulating branches. In Table II the same protocols are tabulated according to regulative responses by the various branch generations without regard to the number of branches or zooids removed. In this table the negative cases show only the zooids which were expected to reconstitute the axial growing point; some of these, failing to regulate, continued to develop laterally with- out further differentiation. The symbols X and x are used to indicate generations on a general- ized branch (Summers, 1938). With reference to a specific branch, e.g. branch D, the symbol lx refers to the microzooid Id, and IX means axial microzooid \D. Simple Cut-offs. — In general when the terminal macrozooid was cut off, the terminal cell of the first branch below the cut differentiated into a new, well-defined terminal macrozooid whose first and subse- quent divisions proliferated the alternating median microzooids (initial branch cells) of the regenerate. Some operations were made at a time when a single cell, the median microzooid X, represented the 136 F. M. SUMMERS adjacent rudimentary branch. In all such instances recorded at least one division followed without perceptible delay, thus producing an axial microzooid 1A" and the presumptive terminal macrozooid l.v. More frequently the removal of a terminal macrozooid left two or three cells on the last branch (Fig. 1). Simple cut-offs leaving this branch with more than three cells were rarely possible for the reason that a division of the terminal macrozooid usually preceded the third division on the adjacent branch. There were a number of cases where, after an operation, the ap- pointed branch continued to develop at a normal rate for one or more A FIG. 2. A. Colony 37/3-5 drawn approximately 53 hours after the operation. The apical cell was destroyed at the two-cell stage of development. The remaining microzooid A divided twice before the new T.M. differentiated from 2a>. Its first division produced the initial cell of the branch B' '. The alternate daughter (2a2) extended the original branch, and the axial microzooid I A persisted without further change. X 250. B. Schematic representation of the resulting growth. generations before another terminal macrozooid differentiated (Fig. 2). In approximately 12 per cent of the regulating colonies two divisions followed the operation, within normal time limits, before indications of a new terminal macrozooid appeared. In one case there were three pre-differentiational divisions. Regulation from the terminal branch zooid l.v following simple decapitation of the growing point occurred in a high percentage of the cases (see Tables I and II). As far as the various levels along the primary axis were explored, the cells of the first few branch generations exhibited relatively frequent regulative responses. For tin- small number of operations made at high levels on old colonies, the transfer- FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 137 mation of l.r into a new terminal macrozooid occurred with about the same frequency and with as much dispatch as for earlier periods, i.e. lower levels on younger colonies. The metamorphosis of the acti- vated microzooid \x at levels above E required no more than the average time necessary for differentiation of apical cells produced at lower levels. This appears to correlate with the more or less uniform rate of normal axial growth (Faure-Fremiet, 1930). Likewise 2xl zooids on the various branches responded in the majority of trials. Zooids of the 3xl generation or later failed to regenerate above the mid- region of the experimental colonies. In order to test the responses of a branch cell of the third generation it was necessary to cut away the newer branches which had formed above it along the main axis. In tests of the fourth or later branch generations, a relatively large part of the colony had to be removed. Compound Cuts. — When the terminal macrozooid and the terminal cell of the last branch were removed, the terminal cell of the second preceding branch was frequently induced to differentiate into a new terminal macrozooid. This particular relationship obtained for a limited number of successive branches and even then was unpredicta- ble. For instance, the terminal macrozooid was sometimes produced by the sub-terminal (a secondary) microzooid of the newest branch despite the presence of a healthy terminal cell on the penultimate branch. In rare cases the latter assumed the regulative function in the presence of a complete uninjured branch between it and the cut-off. The number of regenerates obtained from the sub-terminal branch zooids is given in Table II. Microzooids of the order 2x2 responded up to the level of branch /, whereas sub-terminal zooids of the fourth generation (4^2) did not respond at all. It is also certain that some colonies did not regenerate from either the sub-terminal zooid of the first branch below the cut or from the terminal zooid of the next ad- jacent branch. These branches continued to develop in a normal fashion for one or more generations without attempting to produce a new terminal macrozooid. The reactions of the older segments of well-developed colonies were tested by means of extensive "cut-backs," colonies cut off at some more basal internode. The data obtained (Table II) suggest an inverse relation between the number of regulative responses and the age of the activated zooids, i.e. the frequency of responses diminishes as the number of lateral generations increases. The frequency of negative cases even in younger generations increased in the high levels, which is probably an expression of the fewer generations re- quired to bring the more distal branches to full development. The 138 F. M. SUMMERS same data arranged according to amount of colony removed (Table I) show that basal branches may differentiate a new terminal macrozooid after as many as seven branches plus the apical zooid are cut away. The cut-back experiments were successful only as regards the demonstration of initial reactivity to surgical alterations. The cases in Tables I and II where relatively large portions of the colonies were removed show only that the prospective values of certain zooids along branch axes may or may not be modified, depending upon the amount of colony dissected away. The capacity of the responding zooids for sustained growth is not known because all of the colonies cut back four or more branches had to be discarded before the regenerate attained full growth. Indeed, some were maintained under experi- mental conditions only long enough to produce a new apical cell. The chief difficulty is referable to the fact that the basal branches were the first to be attacked by vegetable growths propagating over the surface of the slide. The affected basal zooids were shed before the colonies reached maturity. For this reason the zooids of branch A on colonies with eight or more branches were usually unsuitable for testing. The age of the colony at the time of operation plus the additional time required for the differentiation of one of its zooids gave to the parasites an advantage that was too frequently fatal to the experiment. Regulation from the Axial Microzooid Series. — Although nearly 9 per cent of the regenerates sprang from the axial microzooid (\X) series, their regulative behavior was capricious and could not be in- duced at will. One of the most striking facts in this connection was the origin of new growths from LY or descendants on complete, uninjured branches (Fig. 3). One originated from the third branch below the cut. I Miberate attempts to activate a given axial zooid by eliminating all other zooids on the branch resulted in (a) no further developmental activity, or (b) regeneration from some zooid on the next lower branch. A three-cell colony which was trimmed down to a single cell, the axial microzooid I A, remained without further change for 165.5 hours; its contractile and feeding responses appeared to be normal for the entire period of observation. Similarly, the corre- sponding zooids on branches E, L, M, and P of other colonies were tested without avail. In each case the regenerate developed from zooids on the next lower branch and is so recorded. On the other hand, it has been demonstrated that the axial microzooids are capable of regenerating (see IX and \Xl in Table II). The protocols show that IX or some of its descendants carry termi- nal macrozooid potencies. In some of these cases the regenerates FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 139 formed directly from IX without leaving ciliospore-producing cells on the activated branch. The others regulated after \Xl and IX2 were formed; IX1 gave rise to the new terminal macrozooid while IX2 produced one or, after division, two typical ciliospores. Regulation after Successive Cuts.— The successive operations were of two general classes: (a) progressive, in which the second and third FIG. 3. A. Drawing of colony 3/J-2 approximately 72 hours after cutting. In this case the regenerate was produced by the axial microzooid on the intact branch /. Axial microzooid II divided before the new terminal macrozooid differentiated: I/1 regulated, leaving I/2 near the base of the new axis. An increase in diameter of the new axis is evident near its junction with the original colony axis. X 250. B. Condition of the apical end of the colony before cutting at the point indicated by the arrow. C. Schematic representation of the resulting growth at the time the drawing was made. cuts removed regenerated parts distal to the preceding operation (Fig. 4) ; and (6) regressive, where the entire regenerate plus additional parts of the original colony were cut off. In the first group the second and third regenerates were themselves products of regenerated seg- ments. Those of the second group developed from some more basal 140 F. M. SUMMERS cell of the original colony. In this respect they were similar to the extensive cut-back types and subject to the same technical limitations. The regulative activity of the zooids of the second and third order regenerates in progressively cut colonies was similar in most respects to that evoked after simple cut-offs. In each of the nine cases studied the same morphological pattern and, as far as the limited number of cases permit judgment, similar developmental rates obtained after the second and third operations. Incomplete Section of the Stalk. — In a few cases out of many trials a local injury to the neuro-muscular cord was effected without destroy- ing the continuity of the cortical hyaline stalk substance. A re- FIG. 4. A. Development after two successive operations (54 hours after the first cut). The original axis lies on the right. The severed peduncle of the regener- ated terminal macrozooid may be seen near the base of the second regenerate. Axial microzooid 1C was badly parasitized; it dropped away soon after the drawing was made. X 250. B. Schematic representation of A. The cuts are indicated by arrows. *Para- sitized zooid. quisite degree of compression between the needles caused the cord to break down into a series of irregular protoplasmic droplets, some of which appeared to be independent of any attenuated membranous connectives. Several interesting facts we're brought out by this type of opera- tion. The severed part of the neuro-muscular cord did not recover from the injury, i.e. the structural or functional cont inuity between the separated parts was not re-established. After the injury there was no subsequent degeneration of the cord in either proximal or distal parts. The functional unity of the whole was permanently impaired. The FORM REGULATION IN ZOOTHAMNIUM ALTERNANS proximal and distal parts contracted independently of each other and, as esteiblished in one case, the distal portion continued to grow and differentiate in the manner of an intact colony, whereas the proximal part regenerated a new primary axis from a zooid below the injury (Fig. 5). More substantial data are obviously required before definite conclusions can be drawn. Nevertheless these results do suggest interesting possibilities for further study. A thoroughgoing investi- gation of the contractile, transportative, and transmissive properties of the neuro-muscular cord may lead to a further elucidation of coordi- nating factors in colony formation. At least here is a clear indication that, whatever the physico-chemical nature of the integrative factors, they are probably mediated through the substance of the cord. ^m A. FIG. 5. A. Branch C of colony 9B4 56 hours after injury to the neuro-muscular cord (drawn from above). The original colony of six branches was pinched in the mid-region, isolating ABC from DEF plus the terminal macrozooid. The terminal cell on branch C (3cl) at the time of the operation differentiated into a new terminal macrozooid which produced two new branches as shown. Note the increased diameter of the stalk just lateral to the second zooid. This marks the position of the microzooid 3cl at the time of injury. X 250. B. Schematic representation of branch C as drawn. Axial microzooid 1C was accidentally cut away from the position marked (x). Several completely isolated fragments were followed for a time by transferring every few hours to fresh filtered sea water. Nothing of unusual interest occurred in their development. Growth, differ- entiation, and regulation in progress at the time of cutting continued as before for the few generations that were followed. They did not re-attach to the substrate but developed as free-swimming fragments. Their growth capacities or the minimum size necessary for survival were not investigated. Differentiation in Regenerated Parts.— The regenerates formed on decapitated colonies, after single or successive operations, are capable of producing any of the six types of heteromorphic zooids previously described (Summers, 1938). According to the data compiled from 142 F. M. SUMMERS 77 protocols the type, number, and distribution of zooids on the regenerated parts compare favorably with the control colonies. The regenerates consisted of a new terminal macrozooid, varying numbers of common microzooids, a terminal cell at the tip of each branch, and one or more potential ciliospores, depending upon the degree of develop- ment following an operation. Macro- and microgamonts likewise differentiated on regulating parts with about the same frequency and vertical distribution as for corresponding regions of normal gamont- producing colonies. Regenerates sometimes differed from the con- trols in respect to the branch generation involved in the production of ciliospores and microgamonts. In normal colonies the ciliospores developed not earlier than the \X generation and the microgamonts only from 2xl or succeeding generations. On the regenerates one or the other of these two types of migrating zooids frequently developed from the initial branch zooid (Fig. 6B). This tendency towards .. FIG. 6. A. Terminal macrozooid 13' (above branch M) differentiating into a ciliospore 204 hours after cutting. The colony was sectioned between branches E and F. X 250. B. Metamorphosis of the median microzooid S' into a microgamont. I he drawing was made 127 hours after regulation from the axial microzooid 1O. X 250. earlier differentiation obtained not only in the young regenerates on immature colonies but also in those derived from older (basal) zooids of nearly mature colonies. Another noticeable deviation from the established norm was tin- occasional metamorphosis of the terminal macrozooid into a migra- tory zooid, either ciliospore or microgamont (Fig. 6A), thus bringing to a close the growth along that particular axis. The terminal macro- zooid of the regenerate may differentiate directly into any of the three reproductive zooids: microgamont, macrogamont, or ciliospore. Faure-Fremiet (1930) reported the formation of the latter type in run- instances during the normal development of this species but the process was unaccompanied by the endomictic process which he described for normal ciliospore development. While studying conjugation in Zodthamnium arbuscula, Furssenko FORM REGULATION IN ZOOTHAMN1UM ALTERNANS 143 (1929) observed that a local injury to one of the main branches affected the zooids distal to the region of injury, inducing many of the micro- zooids to metamorphose into microgamonts ("microconjugants"). Similar effects on the whole colony were induced by unfavorable environmental conditions, e.g. inanition or lack of oxygen. He did not observe a regulatory response subsequent to the injury. Regulation after Conjugation. — From the following fragmentary account of the growth activity manifest in conjugating colonies it is at once clear that this aspect of development alone constitutes a lengthy research problem. Only ten of the colonies whose lineage was being followed happened to conjugate so that a detailed analysis FIG. 7. A colony several days after the onset of conjugation at the level of branch N. The ex-conjugant divided into a cluster of large zooids, one of which differentiated into a new T.M. whose further development prolonged the main axis. The exact lineage of these cells is not known. X 250. of the process is not immediately available. Much of that which follows is based upon conjugants observed among the adventitious growths on the culture slides whose histories are but imperfectly known. Conjugation is introduced here because it affords one clue to qualitatively different physiological relations between the apical zooid, the conjugant in this case, and a large area of the subordinate regions of the colony. So far the results obtained from regulation experiments bespeak a regular functional correlation between the single cell in the apical position and the zooids in sub-adjacent regions, such that the latter 144 F. M. SUMMERS are subservient to the former. Potentialities known to be present in zooids of a lower order are presumably held in abeyance by the apical influence. Cutting away the apical region evokes a response in some zooid in a subordinate but adjacent region. The response is a differ- entiation of another apical zooid whose relations with the whole are seemingly homologous with those of the original apical cell. There may be a time between decapitation and subsequent regulation when the apical cell influences are altogether absent, yet the interim is not sufficiently great to seriously modify the observable growth phe- nomena. It is noteworthy that subordinate branches attain about the same end-point of lateral growth in control and decapitated colonies. As stated in the previous work (Summers, 1938), the macrogamonts were observed only in the terminal macrozooid positions 3 to 24 along the primary axis. The fusion of gamonts invariably brought axial development temporarily to a close some 12 to 13 hours after the last mitosis. The conjugant remained quiescent for periods of about four days, then divided into two moderately large zooids. One of the two ex- conjugants assumed the form of a terminal macrozooid and resumed axial development after the four-day interruption. The fate of the sister ex-conjugant is a matter for conjecture at present; some disap- peared from the colonies between observational periods while others divided into clusters of from two to seven large ciliospore-like zooids at the base of the new axis (Fig. 7). Tin- histories of these are likewise unknown. Apparently they do not propagate additional axes while associated with the parent colony. The development of ciliospores from some of the ex-conjugants in Zoothamniiim arbuscula (Furssenko, 1929) is suggestive, however. The point to be made relates primarily to the behavior of the colony as a whole following the conjugation process. Prior to the completion of conjugation and continuing thereafter a new growth phenomenon appeared. The first three or four branches below the presiding conjugant developed out of all proportion to the average expectations (Fig. 8). The number of branch generations was in some instances greater than twice that of corresponding branches in con- trols. Moreover, many of the common lateral or secondary zooids were activated to divide one or more times, originating second order branches which, in turn, sometimes produced tertiary branches. In this way each of the first few branches below the conjugant level grew almost as individual colonies. The greatest lateral growth effect obtained nearest the conjugant and diminished basally as a gradient. The normal tendency toward a pyramidal colony pattern was thereby FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 145 reversed in the environs of the conjugant. More precise information regarding growth intensity and capacity factors in both the ex-con- jugant strain and the subordinate branches awaits further investigation. DISCUSSION One of the most important consequences of the work is the demon- stration of qualitatively different physiological relations between FIG. 8. A. Schematic representation of protocol \9oe to show the dispropor- tionate development of branches K, M, and N. The lineage of branch L could not be deciphered. One of the ex-conjugant zooids at 0 produced the apical growth illustrated in the diagram. B. The corresponding portion of the largest of the 70 control colonies. spatially separated cells. Under normal conditions a specific pattern unfolds. When an apical region of a colony is cut away some zooid of a lower order, one whose complete developmental possibilities are otherwise never expressed, assumes the dominant generative functions, and the characteristic pattern perseveres. So far these results are intelligible in terms of what Child (1929) calls physiological correla- tion: the relations of dominance or control and subordination between 146 F. M. SUMMERS parts. He concludes "that dominance and subordination depend primarily on quantitative, rather than specific differences in physio- logical condition and that they represent a certain aspect of a physio- logical gradient." In Zoothamnium alternans the transformation of a terminal macrozooid into an ex-conjugant initiates an entirely different developmental phase which gives another clue to the general nature of apical control. Four days after the fusion of gamonts the normal growth relations of a varying number of branches below the conjugant K'vel are upset in a rather remarkable way. Each of the three or four adjacent branches develops out of all proportion to the normal expecta- tions. The precocious mitotic activity produces secondary and even tertiary axes on the affected parts. This unusual phenomenon does not occur when the terminal macrozooid is present or when it is absent ; it is effected by some new quality in the coordinating mechanism arising in consequence of conjugation activities in one particular cell— the apical cell. The effects obtained after decapitation and conjugation certainly suggest that the single cell in the apical position is responsible in a large measure for quantitative and qualitative regulation of the part- whole relationship and that the control varies with respect to the local activities of the parts. There also seems to be a coordination between the cells on different branches. A new terminal macrozooid arising from one of the branch cells exerts its influence from what was formerly considered to be a branch position. Perhaps branch-to-branch coordination also ex- plains the stable activity of the variously placed cells in the interim before a new apical cell differentiates; only one of the several possible zooids regulates. The control of a terminal branch cell over the cells on its own branch can be interpreted in a similar way. As long as the terminal cell presides over a branch strain its immediate relatives remain quiescent. If it is destroyed, however, the sister at tin- sub- terminal position assumes the functional role of the lost cell. A rather wide variation in the degree of regional correlation is suggested by those instances where a new terminal macrozooid arises, not on the first branch as usual, but on the second or even third branch below the operated level. The axial microzooids on every branch do not differentiate into ciliospores. Loci of metamorphosing zooids occur on about every third or fourth branch. The prospective value of a single axial microzooid (LY) is predictable at a relatively early period by a marked growth in size. The growth may be taken as a criterion of at least a partial differentiation, to be completed a good many hours later FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 147 by a further increment in volume, modification of form, appearance of motile organelles, etc. A peculiar characteristic of these large cells is that they may differentiate directly into mature ciliospores or they may divide, giving origin to two zooids of unequal size, both larger than any of the common types. The larger of the two matures first; the smaller grows to the size of its predecessor before metamorphosing or dividing again to produce two mature ciliospores in succession. Regenerates sometimes arise from either those axial microzooids which are, to all appearances, not predestined to metamorphose, or from those already partly differentiated. In the latter a potency or potencies in the process of expression apparently can be altered or superseded by others whose "urgency" toward expression is greater. The directional change in the process is referable to stimuli arising from the altered colonial organization. This is but another bit of evidence to the effect that cellular organization is dynamic and labile at certain periods and that changes going on within the cell which lead to recognizable morphogenetic characteristics are not necessarily irreversibly determined in direction. Many cases are known among the Protozoa where extrinsic, or intrinsic factors lead to periods of reorganization, varying in considerable degree for the different groups and at different periods in the life cycle. In Zoothamnium alternans we have a case where the re-direction of morphogenetic processes can be traced to an extrinsic cause: cutting the colony in the near vicinity of the cell subsequently affected. Several significant problems arise in connection with the variable response of the axial microzooids. Why do regenerates sometimes arise from axial microzooids (IX or descendants) when as a rule the new growths are derived from the terminal branch cells? Attempts to induce regulation from \X cells by trimming away all other cells on the branch gave no positive or predictable results. Until the question is investigated further in Zoothamnium we can only interpret the variable behavior in terms of other work. Some of the merotomy experiments on other protozoa are suggestive (Calkins, 191 la, b; Peebles, 1912; Young, 1922; Dembowska, 1926; Taylor, 1928; and others). With respect to regenerative capacity these investigators were able to demonstrate progressive physiological changes in the cellular organization during the inter-mitotic period. Fragments cut at successive intervals after fission gave an increasingly high percentage of perfect regenerates. In Uronychia (Calkins, 191 la) even an emicro- nucleated fragment regenerated when cut immediately before the onset of fission. But in nearly all of the different forms studied the regenerative tendency disappeared sometime during the division process. 148 p. M. SUMMERS f Another line of investigation summarized by Calkins (1934) and Summers (1935) demonstrates the cytological changes in cell organelles coincident with the division process. The resorption of old and the reappearance of new motor organelles, macronuclear reorganization, etc. suggest a brief period of cellular de-differentiation. There are probably analogous processes of alternating differentiation and de- differentiation in the history of the individual zooids in Zoothamnium. Time may be one of the important factors in individual cell behavior in relation to the balance between extrinsic coordinating influences and the aggregate of intracellular activities. That is to say, the intracellular activities of a Zoothamnium cell may lead to the "fixa- tion" of specific potencies at some critical period after cell division or, conversely, a cell may be more susceptible to the coordinating influences during or immediately after a division process. The axial microzooid, for example, divides in from 20 to 70 hours after its deriva- tion from the initial branch cell, whereas the terminal cell on the branch generally divides at intervals of about 12 hours. If a decapita- tion is made at a moment when the axial microzooid is in some phase of divisional reorganization and the terminal cell in a more stable condi- tion, the former instead of the latter may be activated or excited to prolonged generative activity. The supposition should be tested by a series of accurately timed operations above some particular branch. An explanation of morphogenetic processes in Zoothamnium al- ternans in terms of embryonic segregation at the* time of division has already been attempted by Faure-Fremiet (1930). In order to out- line several points for discussion it is essential to review briefly his cytological analysis of normal development in this species. First, as regards the early axial divisions, the first three generations along the primary axis are unequal divisions. The inequality of the resulting daughters is reflected in the assortment of macronuclear material; each time the zooids which remain in the terminal position (T<\L 1, 2, 3) receive a larger portion of the niacronuclei than the smaller branch microzooids (A, B, C). On the supposition that the enlarged end of the macronucleus apportioned to the terminal macrozooids represents a kind of segregation of chromatin material, these three unequal divisions are described by Faure-Fremiet as qualitatively and quanti- tatively differential divisions. Beginning with the fourth division (division of TM. 3) the extremities of the dividing macronuclei in .ill later axial divisions are similar in size but in each instance a bit more of the finely striated mid-portion of the macronucleus is received by the zo<")id remaining in the terminal position. All of these later divi- sions are characterized as quantitatively differential only. With FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 149 respect to the branch generations, all divisions along branches A, B, and C are similar and almost equal. The initial zooids of subsequently produced branches (D, E, F, etc.) undergo qualitatively differential divisions: the axial microzooids (ID, IE, \F, etc.) receive a greater share of the macronuclei than their lateral sisters (Id, \e, I/, etc.) although the cytoplasm in each case is distributed equally The axial microzooids (\D, IE, IF, etc.) represent the ciliospore-producing members of the colony. They undergo marked growth in size ac- companied by a disintegration and reconstitution of the macronuclei which, although not described in detail, is characterized as an endo- mictic process. Fission in the lateral sisters (Id, \e, If, etc.) is of no further interest cytologically ; these zooids constitute the main branch strains. From the foregoing description it follows that differential division occurs at two points in the formation of all branches above C. To illustrate (Fig. 9) : D receives less cytoplasm and macronucleus than -''.•> --^- *• FIG. 9. TM. 4, but the macronucleus in both resulting zooids is qualitatively similar. The division is therefore quantitatively differential. When the initial branch cell D divides its cytoplasm is distributed equally but the macronucleus is assorted differentially because a thickened granu- lar part goes wholly to the axial microzooid \D. The division is there- fore qualitatively differential. According to Faure-Fremiet, "It appears clearly then that during the growth of a colony of Zoothamnium alternans the two cells resulting from a division of one initial cell are never equivalent as to their 'potentialities.'" Also, "The character of the differential divisions on the main strain seems to determine the individual's differentiation of the colony; this differentiation depends not only on the individual' size, but also upon its physiological potencies." The differential divisions also appear to determine the characteristic features of the 150 F. M. SUMMERS median individuals and of the microzooids. The latter have a limited power of growth and multiplication. Median microzooids on branches A, B, and C remain as common nutritive zooids, whereas the corre- sponding zooids or their descendants on branches above C may undergo considerable growth and metamorphose into ciliospores. Is it to be inferred from this analysis that quantitatively differ- ential divisions determine (restrict) the subsequent power of division in branch strains and, further, that qualitatively differential divisions effect a segregation of potencies for ciliospore formation? Regarding the first inference, all alternating median microzooids, initial branch cells, are the lesser products of quantitatively differential divisions; they can divide only so many times, according to the number of generations normally produced on the branches of which they repre- sent the beginning. \Yhile there is no doubt about the inequality of the cytoplasmic distribution between terminal macrozooids and initial branch cells for the first ten generations or more, the inequality diminishes beyond this point until the two daughters are no longer differential as regards volume of cytoplasm. An equality may be achieved as early as the tenth and not later than the twentieth axial generation. Then what of those colonies that developed eight to fifteen generations beyond the twentieth node with similar axial- lateral relations? Another point to be made pertains to the regulative capacity of the branch zooids. Those distributed in alternate posi- tions along a branch axis seldom divide further so long as the integrity of the whole colony is preserved. When an apical piece is cut away from the colony some one of the more lateral zooids on the remaining portion is capable of assuming the principal generative functions for relatively long periods of time. This behavior is difficult to interpret on the assumption that mitotic or "growth" potentialities are condi- tioned at either or both of the first two generations on the regulating branch. The second inference may be challenged upon the grounds that segregation of ciliospore potencies does not occur at the specified division. Faure-Fremiet adduces cytological evidence of segregation at the first three axial divisions and thereafter at the division of the initial branch zooids. The latter is the fission at which the ciliospore- forming zooids are separated from the main branch strain. In the light of newer findings, the restriction of ciliospore formation to \D, \E, IF, etc. can be questioned. Ciliospores were observed to develop from axial microzooids on the second and third branches, and also from both daughters of the fifth and tenth generations on branches I) and // respectively of control colonies. Moreover, ciliospores oc- FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 151 curred on the regenerate in nearly every case of regulation from branch zooids lateral to the supposed differential division provided, of course, that they were maintained for a sufficient length of lime. The final bit of evidence comes from the demonstrated regulative capacity of the axial microzooids (1^0 on some of the operated colonies. When activated these were able to regenerate comparatively large sections of colony upon which all classes of zooids except gamonts appeared. The spatial relationship of cells should not be minimized as an important determining factor in an organism whose cells are so charac- teristically placed. The importance of this factor in development cannot be properly valuated until the general physico-chemical nature of the integrating mechanism and the medium through which it operates, presumably the neuro-muscular cord, are more fully under- stood. It is fairly certain, however, that it is not a specific factor, for in normal uncut colonies ciliospores occasionally develop in odd positions, and the microgamonts are apt to differentiate from the common zooids at almost any position lateral to the axial microzooids. Of related interest is the work of Buchanan (1927) on the flatworm Phagocata. The region from which a piece is taken with reference to the mouth of the intact worm is of no significance in determining the location of the mouth in the regenerate. Seyd (1935), on the other hand, reported a definite degree of regional specificity in the regenera- tion of a new mouth in Spirostomum; mouths in abnormal positions in the cut organisms degenerated and new ones formed at the correct locations. The conjugation processes in Zoothamnium arbuscula (Furssenko, 1929) and Z. alternans do not differ in essential detail. In the former the zooids at the terminal position on each of the two primary axes (Ae and Bc) becomes the "macroconjugants." Until metagamic divisions occur, further growth along these axes is arrested. Two metagamic divisions of the conjugant result in a cluster of four large zooids, two of which (A^ and C%) metamorphose directly into macro- zooids (ciliospores). The remaining two (B2 and DI) divide again, each giving rise to another macrozooid and a stem cell. A stem cell produces an additional macrozooid and a new (ex-con jugant) axis. A single conjugant therefore produces two ex-conjugant axes and from four to six successively produced macrozooids. The behavior of small lateral branches from the main axis at nodes basal to the conjugant in Z. arbuscula compares favorably with the precocious development of subordinate branches on a conjugating colony of Z. alternans. One or two of the small lateral branches (Seitenatschen) below the conjugant develop to the proportions of 152 F. M. SUMMERS regular main branches. In this fashion normal colonies with seven main axes are transformed, after conjugation, into colonies with nine to eleven chief axes. Furssenko refers to the new growths as "com- pensation" branches (Ersatzzweigen). It is his belief that the enor- mous growth of the macrozooids occurs at the expense of food obtained from nearby microzooids and, similarly, the compensation branches are destined to supply the energy needs of the ex-conjugant deriva- tives, i.e. the five or six clustered macrozooids and the new ex-con- jugant axes. The chief support of Furssenko's hypothesis that ex-conjugant generations develop at the expense of adjacent regions comes from two observations: (a) the ex-conjugants themselves are not active feeders, and (6) the macrozooids in neighboring regions either fail to mature or they divide to form common zooids. The idea presupposes a mobiliza- tion and free transport of nourishment to regions active in development. The observations on Zoothamnium alternans are not in any way contrary to a possible role of the stalk in transportative phenomena. It is also quite likely, although yet to be proved, that nutrient materials are utilized by some zooids at the expense of adjacent ones; this may be a cause contributing to growth inhibition or differentiation in nearby cells. Nevertheless it does appear that the precocious develop- ment of subordinate branches in a conjugating colony of Z. alternans is not primarily directed toward nutritional ends. In the first place, the actively developing apices of branch strains have energy require- ments which, when taken as a whole, undoubtedly exceed the demands of the single ex-conjugant or its first few non-feeding descendants. The flux would therefore be directed away from the conjugant node. Secondly, in nearly every instance recorded the unusual development of the subordinate branches was well under way before the first post- conjugant division occurred. It is problematical whether or not the change in food requirements coincident with the transformation of a terminal macrozooid into an ex-conjugant is sufficiently great to account for the relatively far-reaching alteration of the growth pattern. SUMMARY 1. Zoothamnium alternans is a colonial protozoan of a rather special type whose constituent cells collectively possess in some degree many of the attributes of an integrated organism. Some ot tin- integrating factors can be described in general terms from the work undertaken on form regulation. 2. When the apical cell of the primary axis is dissected away from a developing colony, a cell on some inferior branch, usually the first FORM REGULATION IN ZOOTHAMNIUM ALTERNANS 153 below the cut, will differentiate into a new apical cell. The geo- graphical limits within which positive regulative responses occur are given in the text. 3. Development of a colony continues from the newly differen- tiated apical cell. The structural and developmental characteristics of the normal colony persevere in the regenerated portion. 4. Evidence is presented to the effect that zooids retain, for a time at least, greater developmental potentialities than are actually expressed when they comprise a part of the intact colony. 5. Under varying physiological conditions in the apical region of a colony, the coordinating influences exerted upon the mitotic activity of neighboring zooids may be inhibitory (as shown by the responses evoked after decapitation) or excitatory (when the terminal macro- zooid is transformed into an ex-conjugant). 6. In the light of observations presented, the idea of dichotomous segregation or sifting out of potencies at fission is inadequate as an explanation of localization in this species. The experimental data do not confirm Faure-Fremiet's cytological account of qualitatively differential divisions at specified division nodes on the branches. 7. There is cause to suspect that morphogenetic processes in particular zooids of Zoothamnium alternans (e.g. the presumptive ciliospores), once initiated and partly expressed in visible structure, can be conditioned or modified by cuts made in some neighboring region. 8. An hypothesis is offered to account for the origin of a regenerate from one or the other of several dissimilar cells of a branch strain. The explanation is based upon the factor of time in relation to the balance between extrinsic influences and the aggregate of intracellular metabolic activities by which potentialities are realized. The cells are thought to be more susceptible to external control during the re- organizational period of mitosis. There may be a critical time in cellular differentiation beyond which the intrinsic processes are not influenced by stimuli arising in some other part of the colony. LITERATURE CITED BUCHANAN, J. W., 1927. The spatial relations between developing structures. I. The position of the mouth in regenerating pieces of Phagocata gracilis (Leidy). Jour. Exper. Zool., 49: 69. CALKINS, G. N., 191 la. Regeneration and cell division in Uronychia. Jour. Exper. Zool., 10: 95. CALKINS, G. N., 19116. Effects produced by cutting Paramecium cells. Bull., 21: 36. CALKINS, G. N., 1934. Factors controlling longevity in protozoan protoplasm. Biol. Bull, 67: 410. 154 F. M. SUMMERS CHILD, C. M., 1929. Physiological dominance and physiological isolation in de- velopment and reconstitution. Roux' Archiv. f. entu'.-mech., 117: 21. DEMBOWSKA, \Y. S., 1926. Studies on the regeneration of Protozoa. II. Regenera- tion of the ciliary apparatus in some marine Hypotricha. Jour. Exper. Zoo!., 43: 485. FAURE-FREMIET, E., 1930. Growth and differentiation of the colonies of Zootham- nium alternans (Clap, and Lachm.). Biol. Bull., 58: 28. Ft" RSSENKO, A., 1929. Lebenscyclus und Morphologic von Zoothamnium arbuscula Ehrenberg. Arch. f. Protist., 67: 376. LILLIE, F. R., 1929. Embryonic segregation and its role in the life history. Roux' Archiv. f. entw.-mech., 118: 499. PEEBLES, F., 1912. Regeneration and regulation in Paramecium caudatum. Biol. Bull., 23: 154. PEEBLES, F., 1931. Some growth-regulating factors in Tubularia. Physiol. Zoo/., 4: 1. SEVD, E. L., 1935. Studies on the regulation of Spirostomum ambiguum Ehrbg. Arch.f. Protist., 86: 454. SUMMERS, F. M., 1935. The division and reorganization of the macronuclei of Aspidisca lynceus M tiller, Diophrys appendiculata Stein, and Stylonychia pustulata Ehrbg. Arch.f. Protist., 85: 173. SUMMERS, F. M., 1938. Some aspects of normal development in the colonial ciliate Zoothamnium alternans. Biol. Bull., 74: 116. TAYLOR, C. V., 1928. Protoplasmic reorganization in Uronychia uncinata, sp. nov., during binary fission and regeneration. Physiol. Zoo/., 1:1. WEISS, PAUL, 1935. The so-called organizer and the problem of organization in amphibian development. Physiol. Rev., 15: 639. WESENBERG-LUND, C., 1925. Contributions to the biology of Zoothamnium geniculatum Ayrton. D. Kgl. Danske Vidensk. Selsk. Skr., natiirv. og. math., Afd., 8. Raekke, X: 1. VOIM., I). B., 1922. A contribution to the morphology and physiology of the genus Uronychia. Jour. Exper. Zoo/., 36: 353. Vol. LXXIV, No. 2 April, 1938 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY QUANTITATIVE STUDIES OF THE FACTORS GOVERNING THE RATE OF REGENERATION IN TUBULARIA L. G. EARTH (From the Department, of Zoology, Columbia University and the Marine Biological Laboratory, Woods Hole, Mass.) The purpose of these investigations is to determine the reciprocal effect of two regenerating regions in Tubularia. In order properly to design the experiment it is necessary first to work out the rates of regeneration of different parts of the stem and to determine the effect of size of stem on rate of regeneration. Following this two regions of known rates are allowed to compete with one another and the effect on rate of regeneration is measured. Since it is found that the region exhibiting the lower rate is most affected, further experiments are de- signed to measure this effect. This inhibitory effect exercised by the higher rate over the lower rate has been termed physiological dominance by Child (1929) and the term "dominance" is used in this paper to describe inhibition. When the stem of Tubularia is sectioned the hydranth differentiates in situ from the tissue of the cut end and after 30-40 hours at 18° C. the length of the primordium can be measured by means of an ocular micrometer. At this time a constriction appears at the base of the developing hydranth separating it from the rest of the stem. The perisarc being rigid, the diameter of the hydranth can be measured and the volume calculated. The time for regeneration is measured at two stages of development. t\ is the time in hours from the time at which the stem is cut to the time at which the constriction appears between the primordium and the rest of the stem. /2 is the time for complete regeneration when the hydranth is pushed out of the opening of the perisarc. In some experiments where short pieces of stem are used it is necessary to use t\ as a measure of time since the hydranth, although completely differentiated, does not emerge from the end. In most cases, however, /2 is used for the calculation of rate of regeneration. Rate of regeneration can then be defined as the volume of the hydranth in cubic micra divided by the time in hours. R : • Trr2L/t. 155 156 L. G. EARTH Naturally when t\ is used the rate is somewhat higher than when /2 is the measure of time. The values obtained are in regeneration units which are termed R. V . in this paper. It is convenient to assign symbols to the ends and this has been done in Fig. 1 which shows the method of naming the hydranths. DI, D-t, Ds, etc. are used to designate the distal or oral hydranth of each piece of the stem and are numbered consecutively from distal to Ill n\r Tin p\r III MM Mil, pr i -TT7T D' r^7 Dr i 2 E B •^Tinr Dr Dr rrimnvr P^ PI DI p; Ff P — < thread ligature D! / P! HID! 22 i Mil hydranth regenerating FIG. 1. A shows how the stems are prepared; B, C, D how halves, thirds and sixths are prepared and designated; E shows how inherent rates are determined In- tying off a small part of the stem; finally, /«", G, II show how the iL-ngth of stem is varied while the level remains constant. l)\r,D'£ • • • relative rate of distal hydranths; P\T, P'i • relative rate of proximal hydranths; D\', D-2' • • • inherent rate of distal hydranths; P\\ Pj • inherent rate of proximal hydranths. D22 and P22 are ex- amples of absolute rates. RATE OF REGENERATION IN TUBULARIA 157 proximal levels. D\ is always the most distal hydranth of the stem. PI, Pi, Ps, etc. are used for the proximal or aboral hydranths in the same order. A glance at Fig. 1, A, B, C, will make this clear. The stems are kept in a rectangular dish which is cooled by a glass coil through which running sea water is circulated. The stems rest on a strip of cheesecloth wyhich is stretched over a glass frame and are covered by one centimeter of sea water. Loose-fitting glass covers are placed over the top of the dish. A glass rod, which is drawn out and bent in the form of a hook, is used for picking up and transferring long stems. For short stems a pipette bent at a 45° angle is satisfactory. Since the stems of Tubularia are very variable in the rate at which they regenerate, controls were prepared for every experiment, and experiments involving the use of different stems are never compared. That is to say, one may not compare the rates of 15-mm. stems from different experiments for they may be vastly different, as measurements show (Table II, Experiments 5, 6 and 7). The general procedure was to select from several colonies those colonies with straight stems and no side branches. The stems were cut off at the base and placed in a large dish of sea water and further selected for similarity in length, diameter and appearance. Some stems were translucent, while others were opaque, and these two differed in rate of regeneration. After this selection, the hydranths were cut off, and the stems of the same length were selected at random for experimentals and controls. In long stems, the additional precau- tion of cutting off 3 mm. of the stem along with the hydranth was ob- served, as the region adjacent to the hydranth often regenerates at a low rate, due possibly to the use of this region in the formation of the very large gonophores. The Inherent Rates of Regeneration at Various Regions of the Stem The object of these experiments is to isolate various levels of the stem so that the inherent rate of regeneration can be determined. Isolation of a region from the rest of the stem can be obtained by means of a thread ligature which is tied about the stem shutting off circulation and cutting through the tissue. This technique was used by Morgan (1902) and lately by Peebles (1931). The perisarc, which is tough, does not crack but the coenosarc is completely severed so that there are no cellular connections across the ligature. If this liga- ture be applied about 2-3 mm. from the end of the stem a small piece of tissue 2-3 mm. in length is isolated and will form a hydranth. this means the rate of regeneration of a small piece of tissue at any level 158 L. G. P.ARTH of the stem can be measured without being influenced by the stem as a whole. The situation is analogous to self-differentiation of an explant from an embryo. It might be thought that the same result could be brought about more simply by cutting off a piece 2-3 mm. in length. However, in this case one meets with the difficulty that two centers of regeneration arise at the two cut ends and bipolar forms may arise. It is impossible to get any measure of the rate at which these forms develop. TABLE I Inherent rates of regeneration of distal and proximal hydranths at various levels of the stem of Tubularia. Stems 30 mm. long, cut into 6 pieces and each piece ligatured in middle. L = length in n, d = diameter of stem in /*, /i = time in hours from cutting to the constriction of the regenerating hydranth. L d h L /i =?•- /?, 1136 504 31.6 36.0 71.6 /?, 1044 488 34.4 30.3 564 D3 928 456 36.4 26.2 43.3 D^ 880 432 39.6 22.2 33 1 D6 592 388 39.8 14.8 17.6 Ds 404 364 49.1 8.2 8.7 P, 924 504 39.1 23.6 47.0 P, 824 456 39.3 21.0 33.9 P3 844 432 40.9 20.6 30.3 P^ 704 408 47.5 14.8 194 Ps 416 372 44.5 9.2 96 P6 244 340 52.8 4.6 4.2 The method of determining the inherent rates of regeneration at various levels of the stem is shown in Fig. 1, E, where the superscript "i" is used to designate inherent rate. Stems 30 mm. in length are first cut into 6 pieces and then each 5-mm. piece is ligatured in the middle. Thus in each 5-mm. stem the distal half is completely iso- lated from the proximal half. Since the stem which regenerates is so short, 2.5 mm., the hydranths have difficult}' in emerging and so the time recorded is t\. This is the time from cutting to the formation of the primordium of the hydranth. In Table I and Fig. 2 the data are recorded and plot N •2, is only partially dominant over the proximal hydranth, P2. The rate of P2 = 4.0 R. U. Finally, in Fig. 4, we have plotted the rates of regeneration of the six possible hydranths, Dlt Z>2, D3 and PI, P2, P3 of Fig. 1, C for 10-mm. lengths of stems originally 30 mm. long. The graded differences in rate of regeneration in different regions of the stem are shown by the rates of the distal hydranths and also by the proximal hydranths in Table III, Experiments D, E, F. It will be noted, however, that at any level at which stem is cut the distal hydranth of that level re- generates much more rapidly than a proximal hydranth at the same 164 L. G. BAKTH level. This means that the distal regenerating hydranths in each third of the stem are inhibiting the proximals in that same third. The situation can be summarized by saying that D{ > Ds > D3r and /Y • P,r • /Y, but that Ds • Pir and D3r > Ps, showing that both D\ and D» are dominant. JY is 0, showing the dominance of D3. 90-1 80- 10 15 LENGTH MM. 20 25 30 Fu;. 4. Relative rates of regeneration of cut steins. Squares = stems 26 nun. long cut into halves; triangles = stems 15 mm. long cut into halves; circles = stems 30 mm. long cut into thirds. Df, D-f, P\r, etc. = relative rates of respective li\- clranths. Rate = M3/hrs-105. The dotted line serves to connect adjacent regions of a cut stem. RATE OF REGENERATION IN TUBULARIA 165 The plot of rates in three sets of stems in Fig. 4 shows quite clearly that in each case the gradient is steeper in the upper (distal) levels of the stem. The slopes of the gradient in the upper levels of the three sets of stems are approximately parallel, which seems to indicate that the drop in rate per unit length of stem is about the same in different stems. It should be pointed out that the gradient here will be steeper than the gradient in inherent rates (Fig. 1), since dominance lowers the rate of proximal hydranths. The Reciprocal Influence of Two Regenerating Regions In stems of Tubularia from 10 to 30 mm. in length, a hydranth usually regenerates at both ends, and it is the purpose of the following experiments to show how the rate of regeneration changes when either the distal or proximal hydranth is prevented from regenerating by means of a ligature allowing the hydranth at the opposite end to re- generate free from dominance. These rates have been termed absolute rates, Da, to distinguish them from relative rates, Dr. When only one end of a stem is allowed to regenerate, the rate is termed "absolute rate" in contrast to relative rate. This absolute rate of one end can be determined by simply tying off the opposite end of a stem, which prevents regeneration at this end. This rate tends to be the maximal rate of regeneration of a given region since the region has the entire length of the stem affecting its rate and is also independent of a second regenerating region (Fig. 1, //, D-). It will also be seen that the abso- lute rate is simply the inherent rate plus the increase due to the addition of the stem. The Influence of the Distal Regenerating Region on the Proximal End. -In general (Table III), the shorter the stem the greater the inhibiting effect of the distal regenerating region, and the following cases are arranged according to the length of the stem. In each experiment three lots of stems were used. The first lot regenerated without a ligature, giving relative rates Dr and Pr. The second lot regenerated with the distal end tied off, giving the absolute rate of the proximal end Pa, while in the third group the proximal end was ligatured, giving the absolute rate of the distal end, Da. Without exception (Table III), the absolute rate of any proximal hydranth is higher than its relative rate. Indeed, in some cases the relative rate may be 0, while the absolute rate is fairly high, 36.8 units (Experiment A}. This method of treating the proximal end furnishes a measure of dominance, for if \ve know the absolute rate and the relative rate we can calculate the percentage reduction due to the distal end, i.e. the inhibition exercised by a dominant region. In 166 L. G. EARTH Table III, last column, this percentage has been calculated, and dominance varies from 100 per cent in short pieces to 7 per cent in longer pieces. The Influence of the Proximal Regenerating Region on the Distal End. —In contrast to the proximal hydranth, the distal hydranth is affected little if at all by the ligature at the proximal end. This is true in the upper levels of the stem. Thus, in Table III, Experiment B, the rela- tive rate of the distal hydranth, Df — 40.5 R. U., while the absolute rate of the same hydranth, D\a --- 38.6 R. U. and also, Experiment L, Dir •• 105.5, Dia - 108.0. However, in lower levels of the stem, the absolute rate of the distal hydranth is usually greater than the relative rate. Experiment F, the relative rate of the distal hydranth of the TABLE IV The relation of the inherent rates of regeneration to the relative rates in pieces M3 of Tubularia of varying length. Rate = — • 10s. D' = inherent rate of distal hydranth; P1 = inherent rate of proximal hydranth; Dr = relative rate of distal hydranth; Pr = relative rate of proximal hydranth. Liga- Liga- No 1 i«a No I.1LM Ratios r>i PI ture ture Dr Pr DlIPl Dr\PT mm. 25 54.6 22.0 72.5 27.9 2.46 2.60 Chiefly partition, no dominance. 20 38.2 13.9 49.0 19.0 2.74 2.60 Chiefly partition, no dominance. 14 81.5 22.6 93.0 4.9 3.61 19.0 Dominance incomplete. 5* 71.6 47.0 78.0 0 1.53 00 Dominance complete. * /i used in place of /2- proximal third of the stem Dsr • 19.1, the absolute rate of the same hydranth, D3a ---- 27.0; Experiment II, Df •• 41.4, D2a ---- 53.6. This is extremely interesting, because it shows that the proximal hydranth does have an inhibiting influence over the distal hydranth in lower levels of the stem. In proximal regions of the stem thru there is a reciprocal influence of the two regenerating ends on each other, each tending to inhibit the other. The Competition between Tivo Regenerating Regions Having Different Rates In the second section of this paper it has been shown that adding parts of the stem to either the proximal or distal end increases the rate of regeneration. Therefore the stem as a whole contributes materials to the regenerating ends, and the way in which I his material is par- RATE OF REGENERATION IN TUBULARIA 167 titioned can be studied in stems of different: lengths which provide different amounts of the material. The following experiments were designed to determine how the distal and proximal hydranths would divide or partition the effect of adding a definite volume of the stem to the system. For this purpose the inherent rates of regeneration of the proximal and distal hydranth were found by tying 3-mm. ligatures at the ends of the stems as in Fig. 1, F and H, D\l and PI\ In the same stems the opposite ends Z>i22 and Pi22 give the rates of regeneration with 19 mm. of stem added to each end. Finally, as in Fig. 1, A, the two ends were allowed to regenerate in competition for the intervening 19 mm. of stem. It is seen from Table IV that in long stems (20-25 mm. lengths) the addition of the middle of the stem to the distal and proximal ends TABLE V Increase in rate of regeneration of distal and proximal ends when the same volume of stem is added to each. Total Length Stem 3 mm. Inherent Rate Volume Added M3. 107 Stem 22 mm. Absolute Rate Increase in Rate mm. 25 Distal 54.6 280 66.8 12.2 Proximal 22.0 266 34.0 12.0 Stem 17 mm. Stem 3 mm. Inherent Rate Absolute Rate 20 Distal 38.2 163 51.2 13.0 Proximal 13.9 176 28.9 15.0 increases the rate of regeneration of these ends in direct proportion to their inherent rates. That is, the relative rates are directly propor- tional to the inherent rates Dr/Pr = D{\P\ Therefore there is simply a partition of materials without any dominance. This result would be expected if no factor other than available foodstuffs controlled the rates of regeneration. However, in short stems (14 and 5 mm. lengths) there is no longer a partitioning of substances in proportion to the inherent rates of the distal and proximal hydranths, but rather a dominant effect of the distal end, so that it takes more than its share. Dr/Pr > D^P*. As a matter of fact, in all non-ligatured 5-mm. pieces examined in this experiment and many others, no regeneration occurred at the proximal end at all, in spite of the fact that the proximal end had an inherent rate almost as great as the distal Di/Pi - •• 1.53. This situation repre- sents complete dominance in which all of the material goes to the distal hydranth. 168 L. G. EARTH A Comparison of the Effect of the Distal Half and the Proximal Half on the Inherent Rate of Regeneration of the Ends It has been shown that the middle regions of the stem contribute materials for the regeneration of the two ends. The amount which the stem contributes can be measured by the increase in the rate of regeneration of the hydranths under the assumption that rate is pro- portional to amount of substances available. Thus, in Table Y, the inherent rate of regeneration of the distal end is 54.6 R. U., and when we add 19 mm. of stem, the rate increases to 66.8 R. U. or an increment of 12.2 R. U. Similarly, Pl increases from 22.0 to 34.0 R. U., or an increase of 12.0 R. U. The increase is the same for both proximal and distal ends. (See 20-mm. stems also.) It is clear that the materials of the stem can be used by either the proximal or distal end. Now if we ligature the stem in the middle, the effect of the materials in distal and proximal halves on the regeneration of the ends may be TABLE VI Comparison of the effect of distal and proximal halves of a stem on the rate of regeneration of the ends. Df is the inherent rate of regeneration of the distal end. Under distal half is the rate of the distal end, with middle ligature. Similarly P{ is the inherent rate of the proximal end. Proximal half = rate of proximal end with middle ligature. D* Distal Half Increase P» Proximal Half Increase 25 mm. 54.6 64.6 10 22 26 4 14 mm 820 98 0 16 22.6 24.2 1.6 studied (Table VI). It is found in the two experiments available that the increase in rate is much greater in the distal as compared \\iili the proximal half. This may be taken to mean that there are more materials available in ihc distal half than the proximal. The curves for the effect of volume on rate (Fig. 3) also indicate that the distal half is more effective in increasing rate than the proximal. As volume is added to the distal end, the rate goes up sharply and then falls off. However, as volume is added to the proximal end, the rate goes up slowly at first and then sharply. The evidence seems rather conclusive that the two regions differ in their effects on rate of regeneration. Isolation of Distal and Proximal Regions by Means of a Middle Ligature and its Effect on the Relative Rate of Regeneration of the Ends This method of studying dominance was used in the first experi- ments, and it is complicated by the fact that not only is isolation pro- RATE OF REGENERATION IN TUHULARIA duced by the ligature but the volume of the stem is also reduced. The effect on the proximal hydranth is a dual one. Isolation increases the rate of regeneration, while reduction in volume decreases the rate, and the effect is a summation of the positive and negative action. Since, as we have shown, the reduction of the volume of the stem adjacent to the proximal hydranth decreases the rate to a small extent only, the chief effect is to remove the dominant region from the sphere of action, and the net result is an increase in the rate of the proximal hydranth (Table VII). Length is an important factor, as in 25-mm. stems the change in rate with ligature is slight, while in 15-mm. stems it may be much greater, depending on the condition of the stems. This is in keeping with the fact that little dominance is exerted in long stems. TABLE VII Rate of regeneration of distal and proximal hydranths of Tubularia with and without a middle ligature. Dr and Pr = relative rate; Dl and Pl -- rate under conditions of middle ligature. Experiment Length No. Dr Dl Pr Pi mm. 1 25 10 67.8 70.5 20.4 25.4 2 25 10 72.5 64.6 27.9 26.0 3 15 10 46.7 43.2 18.8 21.0 4 15 10 63.0 64.0 17.5 23.0 5 15 10 40.6 36.4 1.9 15.0 6 15 10 47.6 38.6 0 15.1 7 14 20 93.0 98.0 4.9 24.2 The Mechanism of the Dominance Exerted by the Distal Regenerating Hydranth over the Proximal End of the Cut Stem In all previous experiments a ligature was used to block the dom- inance exerted by the distal regenerating end over the proximal end. With ligature of the stem the proximal end is allowed to regenerate independently of the distal and its rate of regeneration is greatly in- creased. This ligature, however, not only stops circulation between the distal and proximal end but also severs cellular connections. Thus it is not clear whether the factor responsible for dominance is some- thing present in the circulation or something transmitted or trans- ported through the cells. Therefore it is- necessary to determine the effect on the proximal end of stopping circulation from the distal end but leaving cellular* connections intact. Use of a Loose Ligature for Blocking Circulation. — It is not easy to 1/0 L. G. EARTH shut off circulation between parts of the stem of Tubidaria as the cells will rearrange themselves after compression of the stem so that the circulation breaks through once more. However, there are indications which can be seen from Table Mil where the effects of a loose ligature which was tied so as to just stop circulation is compared with a tight ligature cutting through the coenosarc and breaking all cellular con- nections. The control with no ligature shows that we are dealing with stems in which the distal end is almost completely dominant over the proximal end: i.e. distal end, 38.7 R. U.; proximal end, 2.9 R. U. A tight ligature completely isolating the two halves sends the rate of regeneration of the proximal end up to 23.0 R. U. or an 8-fold increase. The distal end shows a slight reduction to 34.0 R. U. as it is cut off from the proximal half of the stem. Now with a loose ligature where cellular connections are still intact the proximal end shows an increased rate of 23.0 R. U. over controls in spite of the fact that at the end of the experiment circulation was reestablished in a few cases through TABLE VIII Rate of regeneration of distal and proximal ends of Tubidaria under conditions of ligature in the middle of the stem. Rate == L/t* where L is length of primordium in micra and t% is time in hours from cutting of stem to emergence of hydranth. No. and Length No Ligature Tight Ligature Loose Ligature of Stems DP DP DP 10 15 mm. 38.7 2.9 34.0 _'.v() 36.4 23.0 the ligature. This sort of experiment, while it appears conclusive, is not entirely satisfactory as it is difficult to control the tying of a ligature so as to cut off circulation without breaking cellular connec- tion between the two halves. Injection of Oil to Block Circulation. — The stem of Tubidaria is about 0.5 mm. in diameter and it is relatively easy to insert a micro- pipette for injection. It is necessary merely to crack the rigid perisarc first with two pairs of sharp watchmaker's forceps after which a pipette can be inserted while observing under a binocular microscope. A small drop of paraffin oil (Nujol) is injected and after the pipette is withdrawn the rigid perisarc snaps back into place. The perisarc must not be re- moved because then regeneration will take place at the exposed surface. Controls for this type of experiment were- stems in which the perisarc was ruptured and the pipette inserted without injection. Some of the experiments where oil was injected also served as controls since the drop was sometimes too small to block circulation. RATE OF REGENERATION IN TUBULARIA 171 Table IX records results. There are 25 stems in each sample and of the 25 controls only 2 proximal hydranths developed, making the rate 4.65 R. U. and thus showing that the distal end exerted rather complete dominance. When circulation is blocked, however, the 25 injected stems regenerate 12 proximal hydranths, bringing the average rate to 22.8 R. U., or a 5-fold increase in rate. The distal hydranth in the injected group shows a small decrease in rate, as might be expected from previous results on the use of ligatures. A ligature in a short stem increases the rate of the inhibited proximal end but decreases the rate of the dominant distal end by shortening the stem. Only 12 out of 25 possible hydranths appear at the proximal end of injected stems and it is interesting to examine the 13 stems which did not form a hydranth proximally. Of these 13 negative cases 9 showed that the oil drop had moved from its original position at the middle of TABLE IX Injection of an oil drop into the gastrovascular cavity of Tubularia. Stems 6 mm. long. Twenty-five stems used in each sample. R = rate of regeneration = nT2L//2 where r = radius of stem in micra; L = length of primordium in micra and /o = time in hours required for emergence of the newly regenerated hydranth. Oil Injected No Injection Distal Proximal Distal Proximal L 1384 282 40.0 86.1 528 264 50.5 22.8 1536 284 40.9 95.0 104 280 55.0 4.65 r /„ R the stem into the distal end. In the 12 stems which form a proximal hydranth all but 2 showed the oil in the original position. It is clear that because the size of the oil drop varies the smaller drops do not completely block circulation and as a result they are carried by the circulation to the distal end. When the drop is larger it is held firmly in place by the endodermal lining of the gastrovascular cavity and in these cases proximal hydranths appear associated with a complete block to circulation. In other experiments where dominance is not so complete there is a quantitative effect on rate of regeneration of the proximal end, with injection of oil into the gastrovascular cavity. It must be remembered that this effect is not so great as would be expected since not all of the oil drops are large enough to block circulation. An example is shown in Table X, where the proximal ends of control stems regenerate at a 172 L. G. EARTH TABLE X The effect of blocking circulation in Tubularia with oil drops. Oil injected into distal region. Controls consist in injury similar to that of injection. Twenty-one stems, 10-11 mm. in length used for each sample. R = irr2L/t. Oil Injected Injury Control Distal Proximal Distal Proximal L. . . 1200 254 35.4 67.0 1032 236 47.0 38.4 1352 252 37.3 72.0 612 228 50.0 20.0 r /, R rate of 20.0 R. U., hut upon injection of oil increase to 38.4 R. U. The rate of the distal hydranth is reduced from 72.0 R. U. in untreated TABLE XI Effect of blocking circulation by means of oil in Tubularia. Stems 8-10 mm. long isolated from long thick stems. Twenty-eight stems in each sample. R = rate using t\; I -= length of primorditim in micra; r -= radius of stem in micra; /i == time in hours from cutting to formation of primordium. Rate = ^3/hrs. -10s. Oil Control Distal Proximal Distal Proximal L 1388 286 37.8 94.0 884 276 46.9 45.0 1320 286 37.3 91.0 852 268 46.0 41.7 r . . . t\ hours R stems to 67.0 in injected stems. This decrease is in part caused by injection of the drop into the distal region of the stem which isolates a small portion. The average length of stem from the oil to the distal end was 2.5 mm. at the termination of the experiment. Finally, in stems where there is very little dominance there is little effect of injection of oil in the middle of the stem. This result was obtained from some 8-10 mm. pieces cut from long, thick stems. Table XI shows that neither the proximal nor distal hydranths are affected to any extent by the injection of oil. This result is to be expected from the section dealing with the use of a middle ligature to isolate the two ends. In stems where the distal hydranth exerts little dominance it was found that little effect was produced on the rate of regeneration of the proximal hydranth by a ligature. The experiments on injection of oil into the gastrovasctilar cavity RATE OF REGENERATION IN TUBULARIA 173 of Tubularia shows that the oil isolates the proximal end of the stem from the distal end, producing about the same effect as a ligature. In the former case the circulation is blocked while in (he latter both circu- lation and cellular transmission are blocked. It becomes important then to see just what the oil drop does in the gastrovascular cavity. It has already been pointed out that when the drop is small it has little effect in blocking dominance. Also, when there is little dominance, little or no effect of injection of oil is found. Therefore it is safe to say that there are no toxic chemical or physical effects of the oil on the cells which come in contact with the oil. Examination of the region into which the oil is injected shows it to be firmly held in place by the endoderm which it partially displaces. When it is not so held in place it moves during the course of hours to the distal end of the stem. Although the endoderm is displaced and perhaps the cellular connections in this layer are broken, the ectoderm remains intact and the cells can be seen to be continuous over the sur- face of the oil drop. As we were not satisfied with this observation, the conductivity over the bridge of ectoderm was tested by using an elec- trical stimulus. In some previous unpublished work on electrical stimulation in Tubularia it was found that upon applying a stimulus at the proximal end the tentacles of the distal hydranth would respond. Three stems in which dominance was blocked by injection of oil were treated in this manner and in each case the tentacles of the distal hydranth responded to an electrical stimulus applied at the proximal end. From these observations there can be little doubt that the ecto- dermal connections over the oil drop are morphologically and physio- logically intact and that dominance is not transmitted over this layer. Discussion Child (1907) pointed out that in Tubularia mesembryanthemum both the length of the cut stem and the level at which the stem was cut were factors determining the time for regeneration and size of the primordium. Driesch (1899) before this had measured the primordium of halves of stems and found that the oral half (distal half) formed longer primordia than the aboral (proximal half). Driesch also showed that the hydranths emerged faster in the oral (distal) half as compared with the aboral (proximal) half. Thus the regional differences in re- generative capacity in the stem of Tubularia are by no means new. The new treatment of the facts by combining two variables, size and time, into a rate has not been suggested hitherto. By utilizing both variables it is possible to express the rate of change within the stem at any level and so compare rates under various conditions. It is hoped that the 174 L. G. BARTH rate as defined in this paper is a measure of the chemical changes in- volved in the differentiation of the hydranth from the stem after cut- ting. It is not sufficient to express these changes in terms of time only since two hydranths of different sizes may regenerate in the same time and certainly the larger hydranth must have utilized more material than the smaller. Therefore the rate of chemical change must have been higher in the region which formed the larger hydranth. Simi- larly two hydranths of the same size may require different times for regeneration. A second difference between these experiments and those of early in vestigators is the use of a ligature to isolate regions in order to test their rate of regeneration. Driesch (18W) and Child (1^07) cut stems into halves and thirds and allowed both ends to regenerate. The size and time for regeneration of distal (oral) hydranths is modified by the presence of a second region of regeneration. Both investigators showed that the size and time factors varied with the length of stem cut. In my experiments by the use of the ligature the second regenerating region is eliminated and the size kept constant. Thus the "inherent" rate is measured. It is proposed that the "inherent " rate of regenera- tion be used as a base so that the effect of variables such as length of cut stem and the rate of regeneration of a second hydranth can be studied by means of appropriate ligatures. Since the stem of Tubularia shows a gradient of "inherent" rates there must be graded differences in the concentration of some substance or substances, which account for these different rates. Further, be- cause the rates are higher in the younger (distal) regions of the stem it is reasonable to assume that the substances are of the nature of a synthetic factor which is able to convert available materials into a hydranth. We will let this inherent synthetic factor be represented by E and assume that the concentration of R is proportional to the inherent rate of regeneration as measured by isolation of parts of the stem. E is then present in highest concentration in the young cells at the distal end and in lowest concentration at the proximal end. Then Fig. 2, giving rates of regeneration, may be taken to indicate the relative con- centration of E at various regions of the stem since only internal factors are responsible for these differing rates at various levels <>t the stem. But E is not the only factor affecting rate. In Fig. 3 it was shown that increase in length of stem adjacent to the regenerating region will also increase rate of regeneration. It is evident that something from the middle of the stem travels to the ends and causes an increase in rate. Let us call this factor or substance S. S is transported through RATE OF REGENERATION IN TUBULARIA 175 the gastrovascular cavity in the circulation which is easily observed in Tubular ia. A cross-section of the stem shows four channels in the endoderm and in the intact stem particles can be seen travelling up one side of the stem and back down the other so that a fairly rapid circulation exists. Timing the flowing particles gave a velocity of 6 mm. /minute. This means that in a short stem 6 mm. long a com- plete circulation of the contents of the coelenteron should take place in 2 minutes. If one end was using up materials in rapid regeneration it is conceivable that the concentration of substance, S, in the circula- tion would be lowered considerably so that at the opposite end sub- stances might pass into the gastrovascular cavity from the cells and so inhibit regeneration by removal of available materials. The effect of S on rate of regeneration is difficult to measure but we may take Fig. 3, which shows increase of rate with increased volume of stem as a provisional measure of S. Obviously, however, this does not give us the effect of 5 in very low concentrations. That 5 is a very important factor is seen by comparing the inherent rate of the proximal end with the relative rate (Table IV, 14-mm. and 5-mm. stems). In these cases something is actually removed from the proximal region so that although the stem is much larger the rate of regeneration is lower when the distal hydranth is regenerating. The situation may be summarized as follows. Regeneration is essentially the transformation of stem into hydranth and this requires E I chemical changes : S +±H. Let us assume E to be a catalyst in the cell which transforms 5 into H, H being the substances necessary for hydranth differentiation. The reaction is reversible, as Child (1923) has shown that hydranths may dedifferentiate into stem. I have also observed a hydranth partially differentiate from ccenosarc and then return to coenosarc. E, as we have pointed out, is present in the cells and is in highest concentration at the distal end. 5 is in the cells and also the gastrovascular cavity and is present in greatest amount in the longest stems. In long stems, where 5 is high (Table IV, 25-mm. and 20-mm. stems) it appears that S is partitioned to the distal and proximal hydranths according to their inherent rates of regeneration, as repre- sented by E. One way of expressing it is that, with increase in 5 while Ep (concentration of E at the proximal end) and Ed (concentra- tion of E at the distal end) remain constant, II v (concentration of //at proximal end) and Hd (concentration of H at distal end) increase pro- portionally. This result is the expected one. 1/6 L. G. EARTH The difficulty comes when we consider short stems, where Ep (concentration of E at the proximal end as measured by the proximal inherent rate) is close to Ed (concentration of E at the distal end as measured by the distal inherent rate) yet apparently 5 is used by the distal hydranth and not by the proximal hydranth. All that can be assumed is that the factor Ed converts 5 into II as fast as it appears in the coelenteron, so that the concentration of 5" is always below the minimal value necessary for regeneration at the opposite end. Since Ed > E,, it can synthesize II at a lower concentration of S. By writing the reaction as a reversible one, we may even suggest that // — » S at the proximal end in the case of short stems. This whole thought de- pends on the assumption that the difference between Ed and Ep in short stems, although small, is great enough to lower the concentration of 5 below a minimal value, Sm for the proximal hydranth. As S in- creases in amount (in longer stems) the concentration in the coelenteron rises above the minimal value Sm for the proximal hydranth and re- generation starts at the proximal end though at a low rate at first. S thus becomes a limiting factor for regeneration in low concentrations. It will be seen that the formal explanation given above is readily applied to other results, such as those from middle ligatures and the results showing the difference between relative and absolute rates of regeneration. From the comparison of the absolute rate of the distal hydranth Da and the relative rate of the same hydranths Dr in Table III, it is ap- parent that there is a maximal rate of regeneration for the distal hydranth, or a maximal concentration of 5 above which synthesis of // is not increased perceptibly. This is not unlikely. The above discussion throws the entire effect of dominance on the transport of available substances (S) for regeneration. In Tnbularia the transport system is extremely simple, and it seems an ideal system for experimentation. If in short stems the circulation of 5 is blocked (as by an oil drop) both the distal and proximal ends should regenerate independently of each other and no dominance will be exerted by the distal end. The experiments in this paper indicate that such is tin- case and thus show that S is a factor which circulates in the fluid of the gastrovascular cavity. Summary \ . The rate of regeneration of Tubularia has been measured by the formula R • • Trr-L/t where r • - radius of cross-section, 1. -- length ot regenerate, / = = time in hours for formation of primordium or the time in hours for emergence of the fully formed hydranth. 2. The rate of regeneration of isolated parts of the stem decreases RATE OF REGENERATION IN TUBULARIA 177 rapidly from the distal to more proximal regions. An increase in the length of stem adjacent to a regenerating end increases the rate of its regeneration. 3. When two regenerating regions are competing they partition the effect of adding the middle region of the stem if the stem is long. If the Stem is short the distal end becomes dominant and inhibits the prox- imal end. 4. A method has been devised for measuring dominance by expres- sing it as the percentage reduction in rate of regeneration due to the presence of the distal hydranth. 5. The circulation within the stem of Tubularia has been blocked by means of injection of a drop of oil with the result that the dominance (inhibition) exercised by the distal regenerating end over the proximal end is blocked. 6. A generalized mechanism based on a synthetic factor E in the cells and a circulatory factor S is suggested as a formal explanation of the phenomenon of dominance. LITERATURE CITED CHILD, C. M., 1907. An analysis of form regulation in Tubularia. IV. Regional and polar differences in the time of hydranth formation as a special case of regulation in a complex system. Arch. Entw.-mech., 24: 1. CHILD, C. M., 1923. The axial gradients in Hydrozoa. Biol. Bull., 45: 181. CHILD, C. M., 1929. Physiological dominance and physiological isolation in de- velopment and reconstitution. Arch. Entw.-mech., 117: 21. DRIESCH, HANS, 1899. Studien iiber das Regulations-vermogen der Organismen. Arch. Entw.-mech., 9: 103. MORGAN, T. H., 1902. Further experiments on the regeneration of Tubularia. Arch. Entw.-mech., 13: 528. PEEBLES, F., 1931. Some growth-regulating factors in Tubularia. Physiol. Zool., 4: 1. CYTOLOGICAL INVESTIGATIONS OF COLPODA CUCULLUS GEORGK \V. KIDDKR AND C. LLOYD CLAFF (From the Arnold Biological Laboratory of Brown University and the Marine Biological Laboratory, Woods Hole, Massachusetts) Inasmuch as the various species of Colpoda have interested many previous investigators, especially with regard to factors of encystment and excystment, it was thought worth while to make a careful study of the best known species, Colpoda cucullus Muller. So many interest- ing and hitherto unreported phenomena have been observed that it \\as thought advisable to present only the cytological results in this first report. Subsequent reports will deal with the many and interest- ing observations on other phases of the problem which are at hand and these will be supplemented by further and more complete data. The life histories of various species of Colpoda have been investi- gated from time to time starting with the work of Stein in 1854. He reported the encystment of Colpoda cucullus, the subsequent division into two, four, eight and even sixteen smaller individuals and their ultimate escape from the ruptured cyst. Rhumbler (1888) made an extended study of the process of encystment and division in Colpoda cucullus and C. steini and distinguished between the division cysts, 'Theilungscyste," and those cysts within which division does not occur, the permanent cysts, " Dauercyste." He describes in some detail the appearance and activity of freshly encysted and dividing forms, noting that the cilia are retained throughout the division process but are lost during their stay in the permanent cysts. He observed quite accurately the gradual loss of the food inclusions, during perma- nent cyst formation, although his interpretation of the method of this loss is open to serious question. \\Vnyon (1926) has given a rather diagrammatic account of the division of Colpoda steini, agreeing with the accounts of Stein and Rhumbler but adding some details of the nuclei. Wenyon's observa- tions were made on fixed and stained material while those of Stein and Rhumbler were mostly obtained from the living material. Very recently Penn (1U37) has reported the occurrence ot binary and quadruple division without encystment in a strain of Colpoda cucullus. Kncystment before division occurred in his race only when the cultures were old or when the cultures became crowded. He was 178 CYTOLOGICAL INVESTIGATIONS OF COI.I'ODA CUCULLUS 1/9 able to induce the formation of cyst walls by placing "healthy in- dividuals" in infusions containing gelatinous masses of bacteria. In this report are given descriptions of the nuclear phenomena which take place during the divisional process, but details are lacking. There has been no cytological report of the nuclei of the "per- manent" or resistant cysts published to date, as far as we are aware, and the reports of the nuclei during division have all been fragmentary. It is not difficult to see why resistant cysts have eluded observation in the past when one considers the fact that most nuclear investigation has been done after the employment of the various haematoxylin stains. As pointed out by Goodey (1913) the cyst walls ("ectocyst and endo- cyst") stain intensely with both Heidenhain's and Delafield's haema- toxylins. We have found it impossible to study the contents of the resistant cysts after treatment with these standard stains and no doubt previous workers have experienced like difficulty. It is with the aid of the nucleal reaction developed by Feulgen and Rossenbeck (1924) that the details of the nuclear complex of resistant cysts have been rendered observable. Haematoxylin stains do not react on the division cyst wall as they do on the resistant cyst walls, mainly because of the comparative delicacy of the former. It is possible to obtain a rough idea of what happens to the nuclei during division by employing these stains. But because of the densely packed food inclusions in the cytoplasm during this period fine details are obscured. After the Feulgen reaction, however, the history of the nuclear components may be readily fol- lowed. It has been found that this history is a rather surprising one and one that may shed considerable light on the role of the macro- nuclear chromatin in ciliate metabolism. Therefore, because we have observed with considerable exactitude the nuclear phenomena both during division and during the stay within the resistant cyst, and because we feel that these observations will contribute to our under- standing of related phenomena among ciliates in general, we offer the following cytological report as the first one of a comprehensive nature on this common organism, Colpoda cucullus. MATERIAL AND METHODS Colpoda cucullus is a very common form and may be collected in a great variety of places. Our original material wras obtained from dry hay taken from the banks of a brackish stream near Stuart, Florida. The hay was placed in spring water from which enormous numbers of the ciliates were later collected and transferred to small dishes. From these dishes a number of motile specimens were selected with a micro- pipette and placed in individual isolation culture dishes in a drop of 180 GEORGE \V. KIDDER AXI) C. I.LOYD CLAFF twenty-four-hour-old diluted hay tea. After a number of encystments and segregations had taken place one motile ciliate \vas selected and all of the others discarded. All glassware was then scrupulously cleaned and boiled for a long period of time to remove any danger of contamination. This obvious precaution was to insure the presence of only one species with which to work. All of our material, therefore, has descended from a single organism. The standard culture medium used in this work, while not the only one successfully used, has given consistent results and the isolation lines grown in it have exhibited a surprising vitality. It consists of nothing more than ordinary hay tea, used after twenty-four hours at room temperatures. Into this medium single motile ciliates were placed and invariably at the end of twenty-four hours three to four divisions had occurred. All that was necessary to obtain resistant cysts was to allow multiplication to proceed for forty-eight hours or longer without adding fresh medium. Under those conditions the division rate decreased until finally all of the ciliates secreted the heavy wall characteristic of the resistance phase, accompanied by the other phenomena described below. Because of the high fission rate, the hardiness of the species and the predictable response to certain environ- mental conditions there was always an abundance of material in every phase of the life history under investigation. For the preparation of permanent slides for cytological study we employed special "recovery dishes"1 for the collection of material designed by one of us (C. L. C.). Small circular flat-bottomed Pyrex dishes with straight sides, measuring 23 mm. X 12 mm. (inside measurements), were prepared. These dishes will accommodate Y% inch circular cover glasses, allowing them to rest on the bottom with very little free space about the edges. Into one of the dishes was placed a cover glass, in some cases thinly coated with fresh egg albumin. The dish was then filled with culture medium and inoculated with a single motile Colpoda. By frequent microscopic examinations the stages desired could be accurately noted, the cover glass taken out and placed in the fixing fluid. In this way we were able to recover hundreds of the different stages on each cover glass. Another ad- vantage this method has is that the organisms do not tend to pile up but each adheres to the glass more or less separately. There is a minimum of debris which makes for clarity of the final preparations. The success of these "recovery" dishes caused us to give up entirely our earlier methods of concentrating by centrifuging and of selecting individuals under the dissecting microscope. 1 These dishes, known as the ClalT Recovery Dishes, are being put on the mar- k' t by Clay-Adams Company, Inc., New York City. CYTOLOGICAL INVESTIGATIONS OF COLPODA CUCULLUS 181 As mentioned in the beginning, our clear preparations resulted from the use of the Feulgen nucleal reaction. We made many prepara- tions with the various haematoxylins and carmines, and while these preparations were entirely satisfactory for the motile forms and proved very useful, they were mediocre for divisional phases and entirely use- less for resistant cysts. We found that certain modifications of tin- standard Feulgen procedure were advantageous. After much experi- mentation it was found that the following times gave excellent results and we highly recommend them for future work on Colpoda cucullus: acid hydrolysis — 15 minutes at 60° C.; fuchsin sulphurous solution— 4 to 5 hours; sodium bisulphite-hydrochloric acid wash — 15 minutes (at least three changes). It was found that the increased time of washing had a decided tendency to clear the cytoplasm of any trace of free fuchsin and rendered the preparations beautifully clear. Many types of fixing fluids were used but it was found that wher- ever cyst walls were present the more penetrating varieties gave, as would be expected, the best results. Therefore most of our material was fixed in Schaudinn's with 5 per cent acetic, corrosive sublimate in 95 per cent alcohol with 5 per cent acetic acid added, or the Gilson- Carnoy mixture. The latter fixing fluid gave the best results on the resistant cysts. Our Feulgen prepared material was very often counterstained with either fast green in 95 per cent alcohol, methylene blue in 70 per cent alcohol, or the acid component of the Borrel mixture (Calkins, 1930). The last-named stain was modified as to balance from the original, containing 1/3 indigo carmine to % picric acid. Counter stains reacted well in the cytoplasm of trophic and divisional forms, penetrating the thin cyst membrane of the latter with ease. The resistant cyst mem- branes appear to be entirely impermeable to fast green, methylene blue and the indigo carmine component of the Borrel mixture. The picric acid of the Borrel stain penetrates very readily, however, staining the cytoplasm an intense yellow. After this counterstain striking preparations are obtained if the material has been taken from a culture in which trophic forms, divisional cysts and newly formed resistant cysts are present. The cytoplasm of the trophic forms and the divi- sional cysts is a brilliant green while the cytoplasm of the resistant cysts is yellow. It should be mentioned that numerous preparations employing the silver nitrate method of Klein were made on the trophic forms as an aid to our positive determination of the species at the beginning of the investigation. The extreme variability of the size and form of these ciliates is so great that specific identification becomes difficult without 182 GEORGE \Y. KIDDKR AND C. LLOYD CLAFF careful study of the ciliary pattern. The question of variation of the ciliary pattern \ve hope to present in another report. It has been found possible to check very accurately the times when the various nuclear changes occur during each phase by starting with resistant cysts, inducing them to excyst and fixing material at frequent, timed intervals. In carrying out these timed observations a large number of " recovery dishes" were prepared at one time and their con- tained organisms fixed in order, first noting the condition of each in the living state. Due to this procedure we have a complete series of slides demonstrating the nuclear activity from excystment, through several reproductive divisions and through a second resistant encystment. As a result of these observations we are certain that the sequence of events to be described is the normal one and occurs in a regular fashion in this ciliate. All the cytological details observed from the timed material have been repeatedly checked from mass cultures. THE LIFE CYCLE — GENERAL Conjugation has never been observed by us in our strain of Colpoda cucullns and as far as we have been able to determine it has never been reported in the literature. Enriques (1908) mentions conjugation as having been noted by him in Colpoda steini but not in C. cucullns. Therefore we will use the term "life cycle" to denote the sequence of events taking place from one resistant cyst stage back to another. Colpoda cucullns appears to possess the ability to accommodate itself to a wide variety of environmental conditions. The great factor in this extreme accommodation seems to be its ability to secrete, in a minimum of time, protective or semiprotective cyst membranes. Under normal conditions all reorganizational phases accompanying reproduction and resistant encystment are carried out when the organ- ism is enclosed in some type of membrane or membranes. Feeding, only occurring in the trophic stage, is accomplished in a very efficient manner, for within a few minutes after resistant excystment the cyto- plasm will be found to be literally packed with spherical food vacuoles. \\ lien placed ill fresh culture medium the resistant cyst undergoes certain changes very rapidly. These changes lead to a rupturing of the heavy outer cyst wall (ectocyst of Goodey, 1913) and finally to liberating the swimming ciliate from the thin inner wall (endocyst of Goodey). One of us (C.L.C ) has been investigating the mechanism of excystment under normal and experimental conditions and will report these findings at an early date. The newly excysted ciliate is devoid of food vacuoles but very rapidly engulfs great quantities of bacteria and bacterial debris. This CYTOLOGICAL INVESTIGATIONS OF COLPODA CUCULLUS 183 food is invariably formed into spherical compact masses and distri- buted at random throughout the cytoplasm. One average sized ciliate (70 11) may have as many as two hundred such food vacuoles (Plate I, Figs. 1 and 2). Growth ensues, the ultimate size appearing to be dependent on the excellence of the cultural conditions but independent of the ability to reproduce. Within a few hours after excystment the trophic form begins to round up, without, however, losing the motility of the cilia. The cytostome becomes indistinct and a thin membrane is secreted outside the still moving cilia. This is the "Theilungscyste" of Rhumbler (1888) and undoubtedly corresponds to the "pseudocyst" of Tillina canalifera recently described by Turner (1937). In the vast majority of cases quadruple division occurs within this cyst resulting in the liberation of four small daughter ciliates. The cilia may be continually observed throughout the process. Occasionally binary division occurs resulting in the liberation of but two daughter ciliates in which case they are proportionately larger in size. We have never observed divisions resulting in more than four daughters although Stein (1854) has reported as many as sixteen daughter ciliates escaping from a division cyst. The report of Penn (1937), in which he maintains the normal condition in Colpoda cucullus to be quadruple division without encystment, was not confirmed in our study. Binary and quadruple division without encystments were encountered only occasionally among the great number of encysted forms irrespective of the bacterial condition of the medium. It is possible that lack of encystment is a peculiarity of the strain studied by Penn. If no additions are made to the culture medium for a period of forty-eight hours the rate of division cyst formation decreases and stops and the ciliates become very small and less active. They round up very quickly and secrete a heavy, wrinkled cyst wall about them- selves. This results in the resistant cyst which is able to withstand desiccation. The formation of resistant cysts may be induced by simply drawing off a part of the medium and replacing it with old medium from a culture in which resistant cyst formation has taken place. In the literature are numerous accounts of the necessity for drying before resistant cyst formation occurs (Barker and Taylor, 1931 ; Bodine, 1923; etc.). We have found that resistant cysts form readily even in an abundance of fluid, and may be kept for long periods without recourse to evaporation. As pointed out by Penn (1937), if conditions inducing resistant cyst formation occur rapidly enough (evaporation, according to Penn), some of the division cysts become resistant cysts by the simple expedient of forming a heavy wall about the outside of the division cyst wall. Thus resistant cysts may contain one, two 184 GEORGE \V. K I ODER AND C. LLOYD CLAFF or occasionally four cells, all motor organelles dedifferentiating until nothing can be seen but the nuclear apparatus and the granular cytoplasm. A B K TEXT FIG. 1. Diagrammatic representation of the "life cycle" of Colpoda cuculhis illustrating the sequence of events during reproduction and resistant cyst formation. A-J, normal reproductive activity repeated (/ to B) under favorable cultural conditions. K-O, resistant cyst with space between N and () representing the lapse of an indefinite amount of time while the arrow from A' to N represents a short space of time during which macronuclear reorganization and chromatin elimi- nation takes place. Arrow from O to A represents the return of favorable conditions for excystment. A young resistant cyst possesses many food inclusions but these are rapidly absorbed, leaving characteristic refringenl inclusions in their -if, id. U'ilhin a few hours even the refringent bodies have disappeared CYTOLOGICAL INVESTIGATIONS OF COLPODA CUCULLUS 185 and the cytoplasm becomes compact and evenly granular, containing the nuclear apparatus. The compactness of the cytoplasm is brought about by the actual shrinkage in the size of the organism even after the cyst wall has been laid down. Resistant cysts are always very much smaller than all but the smallest of the trophic forms or division cysts. The above account agrees essentially with the previously published descriptions of the "cycle" of Colpoda cucullus and will serve as a basis for the details of the nuclear activity described below. Text figure 1 illustrates the normal course of events during the "life cycle" of Colpoda cucullus which, with the accompanying legend, will serve to clarify the above description. THE NUCLEI OF COLPODA CUCULLUS Enriques (1908) gives as a specific characteristic of Colpoda cucullus the possession of a macronucleus with a lobed karyosome "cariosoma lobato." He was able to observe this after staining with carmine. From his figures it may be assumed that, except for the karyosome, the macronucleus is devoid of stainable material (chromatin). Wen- yon (1926) figures the macronucleus of Colpoda cucullus with irregular karyosome-like bodies toward the center. He does not state what stain was employed but it would be safe to assume the use of a hxma- toxylin. That both of these observations were due to the type of stain used is demonstrated by a comparative study of organisms stained in borax carmine, Heidenhain's haematoxylin and after the Feulgen reaction. In the first two cases the stainable material seems to be concentrated toward the center of the macronucleus surrounded by a faintly stained, irregular periphery. But after the Feulgen re- action the picture is reversed. The chromatin is seen as irregular plaques surrounding and extending into the colorless nucleoplasm (Plate I, Fig. 1). After any of the counter stains used with the Feulgen reaction the nucleoplasm is sharply contrasted to the chromatin and is seen to have the form described as the karyosome by previous workers. Penn (1937) illustrates this general condition in his figures from Feulgen prepared material. We wish to emphasize the chromatin configuration of the macronucleus of the trophic form, therefore, as being in the form of irregular plaques around the periphery and sur- rounding the non-chromatin nucleoplasm of the center, because of the prevalent use of this character in classification (see the description of Kahl, 1931). A single micronucleus is always present in the trophic stage. It lies near the macronucleus but may be on any side of it. The chro- matin is quite compact and in fixed preparations is usually seen to have 186 GEORGE W. KIDDER AND C. LLOYD CLAFF shrunken away from the nuclear membrane. The micronucleus is quite commonly elongated, flattened on one surface, and slightly pointed at the ends. \Ye have never found more than the single micronucleus in the trophic forms. Penn (1937) mentions the oc- casional occurrence of two or four micronuclei in his strain. \Ye believe that it is possible that he interpreted the balls of extrusion chromatin from the macronucleus as micronuclei inasmuch as he completely over- looked these interesting bodies (see our Plate I, Fig. 23). In our material the micronucleus behaves in an orthodox fashion during the divisional activities. After the division cyst membrane is laid down it enlarges and assumes a spherical shape. The chromatin becomes slightly less compact and is seen to be finely granular (Plate I, Fig. 3). It then becomes distinctly striated and the whole nucleus elongates and enlarges. The striations within the chromatin become more marked and finally irregular threads may be observed all oriented with the long axis of the nucleus (Plate I, Fig. 4). Contraction of the chromatin threads results in the formation of the metaphase plate (Plate I, Figs. 5, 6, 9). There appear to be a great many chromosomes and we were never able to make even an approximate count, as a glance at the figures will reveal. The anaphase is formed by a separa- tion of the chromosomes of the metaphase plate into two groups; the nature of this separation has not been determined. The two daughter chromosome groups move to opposite ends of the fully formed spindle leaving between them definite clear fibers which remain into tin- elongated telophase (Plate I, Figs. 7 and 8). The chromatin becomes compact again in the daughter telophase groups and these move farther apart (Plate I, Figs. 10 and 19), retaining for some time the connecting strand formed from the nuclear membrane. Finally this strand breaks and the two micronuclei round up and take their positions at opposite ends of the now elongated macronucleus (Plate I, Figs. 11 and 12). This sequence is repeated prior to every cell division without any appreciable variation. It is within the chromatin of the macronucleus that events occur that offer an interesting problem in ciliate cytology. Before the division cyst membrane is laid down the chromatin loses its plaque-like configuration and becomes flocculent, being roughly dispersed through- out the nucleoplasm. It is still slightly granular but the granules are exceedingly minute (Plate I, Fig. 2). By the time the division cyst membrane is formed the chromatin of the macronucleus has begun to take on a definite configuration (Plate I, Fig. 3), that of granular aggregates suspended in the clear nucleoplasm (Plate I, Figs. 9, 10 and 11). This configuration is retained until the daughter organisms are CYTOLOGICAL INVESTIGATIONS OF COLPODA CUCULLUS 187 ready to emerge from the cyst, in most cases through the two divisions. The chromatin aggregates stain uniformly and rather intensely but because of their scattered condition the macronucleus as a whole appears as a loosely knit body, lying among the numerous food inclu- sions. By the time the micronucleus has reached its telophase stage the macronucleus has begun to elongate. This elongation marks the future division plane of the cell, being always at right angles to it. Further elongation stretches the macronuclear membrane until a typical constriction appears separating the chromatin into two daugh- ter halves (Plate I, Fig. 12). The two daughter halves of the macro- nucleus quickly separate and round up and the fission plane forms, dividing the cell into two equal daughter cells, each containing a single micronucleus and macronucleus and numerous food inclusions (Plate I, Fig. 13). Shortly after cell division there becomes differentiated simultane- ously within each daughter macronucleus an irregular, granular mass of chromatin. This mass stains much more intensely than the general macronuclear aggregates and is first seen lying close to the nuclear membrane. It is rapidly separated from the chromatin aggregates and is pushed out from the surface of the macronucleus being sur- rounded by a pocket formed from the macronuclear membrane. This activity is nearly or exactly synchronous in each daughter macro- nucleus (Plate I, Fig. 14). This deeply staining chromatin mass is the "extrusion chromatin" or "residual chromatin" and is ultimately broken away from the macronucleus and cast into the cytoplasm (Plate I, Figs. 15 and 17). Within the cytoplasm it becomes com- pacted into an intensely staining ball which rapidly diminishes in size until it disappears from view. It is usual that complete absorption of the extrusion chromatin is accomplished before the start of the second fission. The second division is initiated, as in the first, by the activity of the micronucleus. All stages appear to be the same as in the preceding division (Plate I, Figs. 18 and 19) with the result that four diuighter ciliates are formed within the original division cyst membrane. Oc- casionally the cyst wall becomes soft and irregular and liberates the two daughters before the second division. Figure 17 represents a case of binary fission just before the liberation of the daughter cells. The extrusion chromatin has not been absorbed yet, and may not be until after the daughters become free-swimming organisms. Following the second cell division each macronucleus again under- goes a reorganizational process whereby more extrusion chromatin is formed and cast out into the cytoplasm (Plate I, Figs. 20 and 21). KS8 GEORGE W. KIDDER AX I) C. I.I.OVI) CLAFF After this process has taken place the four daughter ciliates become more active, the cyst wall becomes softer and more irregular and finally ruptures. In the majority of cases enough time elapses during the freeing process for the extrusion chromatin to become absorbed but sometimes the cyst membrane ruptures early and the young daughter organisms each carry with them the remains of the residual ball (Plate I, Fig. 23). This residual ball might easily be mistaken for a supernumerary micronucleus, as we feel sure has been the case in the report of Penn ( 1937) on the occurrence of more than one micronucleus. When, as occasionally happens, as mentioned above, binary or quadruple division takes place without the formation of a definite cyst membrane there occurs exactly the same nuclear activity as found in the normal division cyst. Chromatin extrusion follows each division in as regular and predictable a fashion as that just described. Figure 22 represents the end result of a quadruple division without cyst formation. These daughter ciliates are completely reorganized and all trace of the residual chromatin has disappeared. We have been able to find all stages representing the above process of chromatin elimina- tion in these divisions but because of the duplication of the conditions found during the normal process and because division without cyst formation was the exception in our material we felt that illustrations would be redundant. The above process of the alternation of a free-swimming, feeding organism with reproduction within the division cyst is repeated every eight hours, on the average, either in isolation or mass cultures so long as fresh medium is provided. Whether or not there will be found a waning vitality over extended periods of time we cannot tell at present. Experiments testing this point are being carried out and will be re- ported at a later date. When conditions within the culture change due to the accumulation of waste products resistant cysts are formed. This process is accom- plished in a very short period of time and involves the laying down of a thick, relatively impermeable wall, the absorption of the food bodies and the concentration of the cytoplasm. No activity on the part of the micronucleus is observed but the macronucleus proceeds to the most profound reorganization as yet reported for any holotrichous ciliate. Resistant cyst formation in a trophic organism proceeds with a diminution in size and a rounding up of the ciliate. As the heavy cyst wall is secreted the cilia disappear and the cyst becomes firmly attached to the substrate. The nuclear apparatus is much the same in appearance as in the trophic stage except for the tact l licit both CYTOLOGICAL INVESTIGATIONS OF COLPODA CUCULLUS 189 nuclei become smaller and slightly more dense (Plate II, Fig. 24). The food inclusions are rapidly absorbed and the cytoplasm then con- tains numerous minute refringent bodies (Plate II, Figs. 25 and 26). After the disappearance of the food balls but before the disappearance of the refringent bodies the macronucleus begins to elongate and to differentiate into two distinct regions. One end becomes more in- tensely staining than the other as if the chromatin was becoming com- EXPLANATION OF PLATES All figures are of Colpoda cucullus, and with the exception of Figs. 4-8 the magnification is X 1,000. Figs. 4-8 represent a magnification of X 2,000. All figures are from preparations treated with the Feulgen reagent. Fixatives used were Schaudinn's fluid with 5 per cent acetic acid for the preparations illustrated on Plate I and Gilson-Carnoy fluid for the illustrations on Plate II. The cilia, which are present in all stages represented on Plate I, have been omitted from the illustra- tions. Asterisks (*) represent residual chromatin. PLATE I Explanation of Figures FIG. 1. Trophic ciliate showing typical nuclear apparatus, food inclusions, contractile vacuole and cytoplasmic vacuoles. Note the plaque-like arrangement of the macronuclear chromatin. FIG. 2. Ciliate about to undergo encystment prior to reproduction. The macronuclear chromatin has become dispersed and flocculent. FIG. 3. Early division cyst. Micronucleus enlarged and the macronuclear chromatin beginning to collect in aggregates. FIGS. 4-8. Representative micronuclei during mitosis. FIG. 9. Micronucleus in metaphase and macronuclear chromatin in the form of irregular aggregates. FIG. 10. Later stage. Macronucleus elongating. FIG. 11. Daughter micronuclei at the poles of the elongated macronucleus. FIG. 12. Constriction of the macronucleus. FIG. 13. Plasmotomy completed and the two daughter macronuclei rounded up. No indication of chromatin differentiation for elimination as yet. FIG. 14. Budding off of extrusion chromatin (*). FIG. 15. Slightly later stage. FIG. 16. About the same condition as the previous stage and a timed prepara- tion. This cyst represents the first one formed after emerging from the resistant cyst. Note small size. FIG. 17. Probably a case of binary fission with the two daughter ciliates about to leave the wrinkled cyst membrane. Note extrusion chromatin (*). FIG. 18. Prophase of the second division. The extrusion chromatin has been absorbed in the cytoplasm. FIG. 19. Constriction of the macronuclei for the second division. FIG. 20. Second cell division completed. Extrusion chromatin being given off from each daughter macronucleus. FIG. 21. Four small ciliates about to emerge from the division cyst. Within the cytoplasm of each will be seen the residual ball of chromatin (*). FIG. 22. A case of quadruple division without encystment. This represents the last stage with the daughter ciliates completely reorganized and about to separate. FIG. 23. A free-swimming ciliate just after emerging from the division cyst. The residual ball of chromatin (*) has not been absorbed as yet. 190 GEORGE W. K1DDER AND C. LLOYD CLAFF * * > '* '* I : . * . >*H7 m? **- ^ •.» i '-*• " y • 8 CYTOLOGICAL INVESTIGATIONS OF COLPODA CUCULLUS 191 . . g V^-» •% . ife :••:-. <" 20 , — 17 ^ 19 t U 16 ®:.*>&: .!& 22 192 GEORGE \V. K1DDER AND C LLOYD CLA1-T '& I 3 :> 24 25 26 28 29 30 31 32 35 •* '37 » "38 > 39 CYTOLOGICAL IN\ HSTK iATIONS OF COLPODA CUCULLUS 193 parted. This compacted region varies in size from one-third to one- half the whole nucleus (Plate II, Fig. 27). Very rapidly the compacted area buds off from the rest of the macronucleus leaving its chromatin in an irregular granular condition (Plate II, Fig. 28). The compact bud moves farther away from the nucleus until the connection between the two severs and the deeply-staining chromatin rounds up in the cytoplasm (Plate II, Figs. 29, 30 and 31). The amount of chromatin extruded varies considerably in the different cysts. In extreme cases it forms a ball as large as the remaining reorganized macronucleus but usually it is somewhat smaller (Plate II, Figs. 31 and 32). The ulti- mate resorption of the cast-out chromatin takes place by a gradual diminution in size but no apparent diminution in staining capacity (Plate II, Figs. 33 and 34). Within a few hours after encystment the extruded chromatin has disappeared and there is no further activity within the cyst until excystment. The refringent bodies have gradu- ally disappeared during this process of nuclear reorganization and the resting cyst then possesses a very clear, slightly granular cytoplasm in which the micro- and macronucleus are embedded (Plate II, Fig. 35). PLATE II FIG. 24. Newly-formed resistant cyst characterized by the possession of food spheres. FIG. 25. Later stage showing the diminution of food bodies and the concomitant appearance of the small, refringent bodies. FIG. 26. The food bodies have entirely disappeared and the refringent bodies are numerous. FIG. 27. Early stage in the differentiation of the extrusion chromatin within the macronucleus. The extrusion chromatin stains more intensely than the chroma- tin which is destined to remain. FIGS. 28-30. The budding off of the ball of extrusion chromatin. FIG. 31. The connection between the extrusion bud and the macronucleus has broken. Note the globular condition of the large extrusion mass (*). FIG. 32. A large extrusion mass (*) somewhat later than the preceding stage. FIGS. 33-34. The extrusion chromatin mass (*) diminishes in size. Note also that the refringent bodies within the cytoplasm have disappeared. FIG. 35. A reorganized resistant cyst with compact clear cytoplasm and a smoothly granular, compact macronucleus. In this condition the resistant cyst will remain until the cultural conditions are altered sufficiently to induce its excystment. FIGS. 36-39. Chromatin elimination from the macronuclei of two-cell resistant cysts duplicating, in each cell, the conditions seen in the single-celled resistant cyst. FIG. 40. Two-celled resistant cyst in which cyst formation occurred before the extrusion chromatin of the divisional reorganization was completely absorbed. The divisional extrusion chromatin is represented by the small, deeply staining balls (*d) while the resistant cyst extrusion chromatin is represented by the large masses (V). FIG. 41. Four-cell resistant stage showing typical extrusion chromatin in each cell. This type of cyst is relatively rare. FIG. 42. Small, clear, cyst-like structures which appeared in a few old liquid cultures and which are thought to represent degenerate resistant cysts. 194 GEORGE W. KIDDER AXI) C. LLOYD CLAFF \\V have found a few cases where two buds of waste chromatin have been extruded from the reorganizing macronucleus but these are rare. When, as often happens, division cysts become resistant cysts there is a complete reorganization within each of the daughter ciliates identi- cal with that of the single-celled cyst described above. Usually there appears to be enough time for the divisional reorganization of the macronucleus to proceed to completion and the subsequent re- sorption of the extrusion chromatin to take place before resistant cyst macronuclear reorganization sets in (Plate II, Figs. 36, 37, 38, 39 and 41). Rarely are there found resistant cysts where both divisional extrusion chromatin and resistant cyst extrusion chromatin are present (Plate II, Fig. 40). Also resistant cysts containing two cells are much more frequent in occurrence than those containing four cells. DISCUSSION In our opinion the actuality of macronuclear reorganization involv- ing the elimination of residual chromatin may possibly be demon- strated universally among the holotrichous ciliates. The definite estab- lishment of the elimination of residual chromatin as a single ball during divisional reorganization has been made in the following forms: Kid- deria (Conchophthirius} my till (Kidder, 1933a), Ancistruma isseli (Kidder, 1933/;), Conchophthirius anodontx, C. curtus, C. magna ( Kidder, 1934), Myxophyllum (Conchophthirius} steenstrupii (Rossolimo and Jakimowitsch, 1929), Allosphxrium convexa (Kidder and Summers, 1935) and Urocentrum turbo (Kidder and Diller, 1934). A number of other species undoubtedly fall into this group if we can judge by the published reports (see Loxocephalus, Behrend, 1916 and Enpoterion pernix, MacLennan and Connell, 1931). Post-divisional chromatin elimination, i.e. the casting into the cytoplasm of residual chromatin from each of the daughter macronuclei after separation, is known to occur in IclUhyophthirius multifiliis (Haas, 1933), Colpidium colpoda, C. campylum and Glaucoma scintillans (Kidder and Diller, 1934), Chilodonella labiata and C. faurei (MacDougall, 1936). In all the cases cited above the chromatin to be eliminated is dif- ferentiated within the macronucleus prior to its division and thus ad- vertises itself. All species, therefore, in which no differentiation into regions occur prior to fission have been described as having a "clean" macronuclear division. As the vast majority of ciliates divide without encystment, and the two daughter cells separate immediately after fission, this phase of their cytology has been neglected. Ii serins entirely possible to us that if attention were paid to the young daughter CYTOLOGICAL INVESTIGATIONS OF COLPODA CUCULLUS 195 cells after fission the occurrence of macron uclear reorganizational processes would be discovered in a great many if not all species. An- other possibility that suggests itself is that the many cases of chromatin- like fragments in the cytoplasm so often reported in ciliate studies may be explained by some process of macronuclear elimination during reorganization. \Ye wish to emphasize the necessity for more thor- ough and critical cytological work with this problem in mind to deter- mine whether or not macronuclear reorganization with chromatin elimination will be found to be a universal principle applicable to till holotrichous ciliates. As to the exact significance of this regular though complicated process, we are still unable to say. It has been suggested (Kidder, 1933a, 19336, 1934; Kidder and Diller, 1934) that the eliminated chromatin represents worn-out material and the process might be a cleaning out of the macronucleus toward a perfect organization. It was further suggested (Kidder and Diller, 1934) that the profound reorganization which occurs at every division of Colpidium colpoda, C. campylum and Glaucoma scintillans might account for the fact that conjugation rarely occurs in these species, the reorganizational process serving to restore the cells to their fundamental condition and thereby decreasing the necessity for conjugation. The above suggestion seems to us to apply as well as any other to Colpoda cucullus. Concerning the drastic reorganization that occurs immediately after the resistant cyst is formed, it seems logical to suppose that this represents the ridding of the macronucleus of materials no longer needed in the state of decreased or suspended metabolism. Materials accumulated in the macronucleus through the very activity of encyst- ment may be detrimental to the resting cell or to the process of excyst- ment, to come at a later date. Unfortunately we have very few records of what goes on within the resistant cysts of the various species of ciliates with which to make comparisons. We know from a few sources (see Tittler, 1935) that a process of endomyxis sometimes accompanies encystment, whereby the old macronucleus is completely discarded and a new one formed from micronuclear material. These cases would seem to represent simply a different method for accom- plishing the same general result as occurs in Colpoda cucullus, the production of a purified macronucleus. That the reorganizational process occurring within the resistant cyst takes the place of the divisional reorganization of the macro- nucleus is denied by direct observation. In the very first division after emergence from the resistant cyst the normal reorganizational chro- matin elimination occurs. This was determined by timed preparations 196 GEORGE W. KIDDER AND C. LLOYD CLAFF and Figure 16 illustrates a first division cyst. The small size is usual as the ciliates emerging from the resistant cysts are very small and usually reproduce before full growth is attained, a condition noted in the case of Tillina ma°na by Gregory (1909). The time factor may play an important role here, however, as there is the possibility of the aging of the macronucleus during its long period within the resistant cyst, resulting in the necessity for reorganization immediately upon again taking up an active life. The actuality of a regular and predictable macronuclear reorganiza- tion with the elimination of quantities of chromatin during division and within the resistant cyst has been established for Colpoda cucidlus but a completely satisfactory explanation of its significance awaits further investigation. Experiments are now under way which, it is hoped, will throw some light on this question. SUMMARY 1. A complete description of the nuclear activity of Colpoda cucullus Muller is given for the first time. 2. In our strain the normal method of reproduction takes place within a thin cyst membrane. Usually two divisions result giving rise to four daughter organisms which break out of the cyst and repeat the process. Occasionally binary fission occurs within the cyst. Rarely quadruple division occurs without encystment, as described by Penn (1937). 3. Following each cell division there occurs a reorganizational pro- cess within the daughter macronuclei resulting in the elimination of a quantity of residual chromatin. The residual chromatin is cast into the cytoplasm where it is absorbed. Elimination of residual chromatin is regular and synchronous in each cell whether the division has occurred within a cyst or not. 4. When cultural conditions are poor resistant cysts are formed. 5. The resistant cysts are formed by the secretion ot a heavy cyst membrane, the absorption of the food inclusions and the concentration of the whole protoplasmic mass. 6. Immediately following the formation of the resistant cyst membrane the macronucleus undergoes a profound reorganization during which a variable, but always a considerable amount of chromatin is budded off and cast into the cytoplasm where it is absorbed. \<> micronuclear activity occurs at this time. 7. The question of chromatin elimination from the macronuclei of holoirichous ciliates is discussed and the opinion expressed that this phenomenon may represent a universal principle. CYTOLOGICAL INVESTIGATIONS OF COLPODA CUCULLUS LITERATURE CITED BARKER, H. A., AND C. V. TAYLOR, 1931. A study of the conditions of encj imrni of Colpoda cucullus. Physiol. Zool., 4: 620. BEHREND, K., 1916. Zur Conjugation von Loxoct-phalus. Arch. f. Protist., 37: 1. BODINE, J. H., 1923. Excystation of Colpoda cucullus. Jour. Expcr. Zool., 37: 115. CALKINS, G. N., 1930. Uroleptus halseyi Calkins. II. The origin and fate of the macronudear chromatin. Arch.f. Protist., 69: 151. ENRIQUES, P., 1908. Sulla morfologia e sistematica del genere Colpoda. Ar!>••. el< .| .HP-HI oi i]i ii al i ili.i near i In- c.'.ip in i lie pn »totroch; ; I i In- appearance of contractions of the larva, albeit these contractions are considerably weaker than those of the normal larva; (5) the formation of two pairs of bristles of the same notched structure- as described lor the normal. Twenty-nine-hour Larva. — Normal (Fig. 4^4). The larva is begin- ning to assume the appearance of the fully-developed trochophore. The membrane is closely applied to the surface of tin- larva. The apical tuft has disappeared, and only the non-motile cilia remain at the anterior end. The prototroch is raised slightly, forming the so- called "hood fold" i Wilson, 1(J29). The single posterior cilium per- sists. The small yellow chromatophores are taking on a green tinge. The stiff dorsal cilia are longer and more numerous than in 1 he previous DETERMINATION IN SABELLARIA 205 stage. The oesophagus, lined with many active cilia, leads into the stomach. The stomach is partially separated from the intestine by an incomplete shelf; the intestine is lined with cilia of the type found in the stomach. The number of bristles varies from three to four pairs. On stimulation (i.e., when the slide or the dish is tapped gently), the larva rapidly contracts and the bristles are spread into a fan-like arrangement. After several seconds in this position, the animal returns to its usual form. Exogastrula (Fig. 4.B). The general appearance of the larva is similar to that of the preceding stage. The three or four pairs of bristles project laterally and, as the animal moves, are rotated in the manner described above. On stimulation, they are spread out as in FIG. 5. Larvae, forty-eight hours after fertilization. Abbreviations as in Fig. 3. A. Normal larva. Dorsal view. To the left of the stomach is the eye spot (e). B. Exogastrula. Dorsal view. A large part of the post-trochal region is covered with short cilia. the normal larva, although to a lesser extent. There is no internal gut, and a large cavity is present inside the larva. The chsetae-sacs project into the cavity. As in the normal larva of this age, only the apical cilia persist at the anterior end. The dorsal cilia have elongated and increased in number. The chromatophores, limited to about the anterior half of the larva, are becoming greener in color. The cilia on the exterior of the post-trochal region are found in two regions. Of these only the ventral region has the cilia of the cesophageal type. Forty-eight-hour Larva. — Normal (Fig. 5A). This is a well-devel- oped trochophore. The post-trochal region has lengthened somewhat. The apical cilia are numerous and the number of dorsal cilia has also increased. There is a single posterior cilium. A second, posterior row 206 ALEX B. NOVIKOFF of short cilia has appeared on the hood fold, just posterior to the long prototrochal cilia. On either side of the mouth there has developed a fold covered externally with cilia, the lip fold. The chromatophores have a more pronounced green coloration. On the left side of the larva, near the stomach, there has appeared a single eye spot, com- posed of closely packed orange-red granules. The gut is clearly differentiated into oesophagus, stomach, and intestine. The stomach has increased in size and has become more spherical. Running posteriorly from the mouth, in a groove along the ventral surface, are the long cilia which constitute the neurotroch. A larva at this time ordinarily has four pairs of bristles, which project posteriorly. Exogastrula (Fig. SB). Changes similar to those occurring in the normal larva? have taken place between this and the preceding stage. The apical cilia have become more numerous. The prototroch is better developed and a row of short cilia has been added posterior to it. The dorsal cilia have increased in number and si/e, and the posterior cilium is single and dorsal. On the dorsal surface, anterior to the prototroch, may be seen the eye spot. Four pair of bristles project laterally as do the bristles of the earlier larva1. There is no internal gut. The cilia characteristic of the oesophagus and the intestine, together with the cilia of the neurotroch and the lip fold, cover the post-trochal region almost completely. Special mention must be made of the eye spot, since it was not seen in all the larvae examined. It could not be found in thirteen of the ninety-three larva? studied. Partial Exogastrulx There are always found in the culture dishes containing the mem- braneless larva1 a few larva- which are intermediate in structure between those described as complete exogastrulu* and the normal. These larvae are illuminating for an understanding of the structure of the more extreme type. Figure 6 shows two views of a well-developed trochophore of this type. Not all the structures are visible in the drawings, but the larva possesses every structure found in the normal animals, ah hough these structures have in some cases differentiated in abnormal positions. The most outstanding feature of these larva? is the gut structure. The stomach and intestine are clearly normal in form, position, ,md type of ciliation. The intestine opens posteriorly, as usual, through the anus. At the anterior end of the stomach there is a circular opening lined with cilia which move in a manner characteristic of the cilia of the o^sophageal-stomach opening in the normal larva. But this opening leads here not to the oesophagus, but to the outer surface of the larva. In fact, there is no internal DETERMINATION IN SABELLARIA 207 oesophagus at all. On the ventral surface of the larva is a long, conical outgrowth, completely covered with long actively-moving cilia which are characteristic of the oesophagus. DISCUSSION OF RESULTS There are two important points of difference between larvae which develop from eggs without membranes and normal larvae. The first is the complete absence of an internal gut and the second the striking deviation from the normal form of the larva. Instead of possessing the fairly spherical shape, they have elongate bodies in which the posterior structures, the posterior cilium and the paired bristles, are dorsally displaced. The distortion of the posterior part of the larva may be accounted for if we assume that the endoderm cells, which differentiate normally inside the ectoderm to form the gut, grow out so that they come to lie on the surface of the larva. B FIG. 6. Partial exogastrulae. Seventy-two hours after fertilization. A. Dorsal view. The opening to the stomach, the stomach, and the intestine are visible. B. Ventral view. The bristles have been omitted in the drawing; only their points of origin in the chsetae-sacs are indicated. The opening to the stomach may be seen at the base of the long conical projection covered with short cilia. That this posterior outgrowth of the endoderm cells has indeed occurred is shown by the following considerations: 1. The gut is com- pletely lacking in these larvae, and within the post-trochal region of the larva there is a large cavity. 2. The cells situated at the posterior end of these larvae are not pigmented and never develop chromato- phores. In the normal larva the chromatophores are found in the ectoderm; only the endoderm is colorless. 3. In the normal larva, motile cilia do not appear on the external surface of the post-trochal region until approximately thirty-five hours after fertilization. \\ hen they do develop, they are restricted to a narrow region, the neurotroch. But in the membraneless larva cilia begin to appear on the surface of the post-trochal region before the end of the first day of development. By twenty-four hours after fertilization they have covered two wide areas. That these cilia are not those of the neurotroch is clear from 208 ALEX B. NOYIKOFF the fact that they cover two extensive regions and that they develop twelve hours earlier than do the neurotrochal cilia. Their appearance coincides precisely with the development, normally, of cilia on the inner surface of the endoderm cells. It is also possible to detect differences in the cilia which correspond to the differences between the resophageal cilia and the cilia of the stomach and intestine. \Ye may therefore conclude that in these eggs the endoderm has been turned inside out. Whatever the mechanism involved, the end result of this process is comparable to the exogastrula1 that have been described in the sea-urchin (Herbst, 1893; Driesch, 1893) and in the amphibia (Holtfreter, 1933a, b). In the partial exogastrulae, the stomach and intestine develop normally, and only the o-sophagus is exogastrulated. And, as we would expect on the basis of our assumption, the general form of the larva is more like the normal than is that of the complete exogastrulu. The most significant feature of exogastrula development in Sabel- laria is the complete self-differentiation of both endoderm and ecto- derm. In the amphibian egg, Holtfreter found that normal differen- tiation of the ectoderm is dependent upon contact with the mesendo- derm. When the latter, instead of coming to lie beneath the ectoderm as it does normally, evaginates, it leaves the ectoderm a wrinkled, hollow sac in which no trace of differentiation into nervous tissue appears. On the other hand, the mesendoderm, although turned inside-out, undergoes self-differentiation; it produces gut, thyroid, pancreas, liver, notochord, musculature, kidney, and gonad. Thus, in the normal course of development, the endoderm differentiates independently of the ectoderm, but the ectoderm must be in contact with the endoderm, or gut roof, to differentiate normally. It is this gut roof which acts in the capacity of an organizer — which induces the formation of a nervous system in the ectoderm. In Sabellaria, on the other hand, there is no deviation from normal differentiation of either the ectoderm or endoderm in the exogastrula?. The endoderm cells retain their morphological polarity, so that they develop cilia on what is in these larvae their external surface. And although the tripartite nature of the gut is lost, the type of cilia devel- oped by the exogastrulated . Organ isierungsstuf en nach regionaler Kombination von Entomesoderm mil Ektoderm. Biol. Zentralbl., 53: 404. HORSTADIUS, SVEN, 1936. In\x'>-t igation> on determination in the early development of Cerebratulus (Abstracl I. Ki»l. Hull., 71: 406. HORSTADIUS, SVEN, 1937. Experiments on determination in the early development of Cerebratulus lacteus. Biol. Bull., 73: 317. NOVIKOFF, ALEX B., 1936. Transplantation of the polar lobe in Sabellaria vulgaris (Abstract). Anat. Rec., 67: (Supplement 1) 57. NOVIKOFF, ALEX B., 1937. Sabellaria vulgaris. Culture Methods for Invertebrate Animals, by (ialtsolT et al. Comstock Publishing Co. Inc., p. 187. TuNCi, Ti-Cnow, 1934. Recherches sur les potentialtes des blastomcres chez Ascidiella scabra. Arch. d'Anat. Micros., 30: 381. WILSON, I)OU<;LAS P., 1929. The larva; of the British sabellarians. Jour. Mar. Hiol. Ass'n., 16: 221. EMBRYONIC DETERMINATION IN THE ANNELID, SABELLARIA VULGARIS II. TRANSPLANTATION OF POLAR LOBES AND BI.ASK >MKKES AS A TEST OF THEIR INDUCING CAPACITIES ALEX B. NOVIKOFF (From the Department of Zoology, Columbia University and the Marine Biological Laboratory, Woods Hole, Massachusetts] INTRODUCTION E. B. Wilson (1904a) demonstrated that the egg of the annelid, Lanice, belonged to the group of so-called mosaic eggs, since isolated halves of the two-cell stage developed into partial embryos. The anterior cell produced an embryo which lacked the post-trochal region; in the embryo formed from the posterior cell there was present a nearly typical post-trochal region. Delage (1899) had previously described a dwarf embryo from an egg-fragment of the same species. That annelid eggs, generally, are of the mosaic type is shown by the experiments of Penners (1924, 1926) with Tubifex, of Tyler (1930) with Chxtopterus, and of Hatt (1932) with Sabellaria. Wilson (1929) summarizes the evidence which indicates that there is no fundamental distinction between the mosaic and regulative types of ova. Among regulative eggs, where correlative differentiation, or embryonic induction, is most prominent, mosaic features can be found, and in mosaic eggs, there are suggestions that embryonic induction may play a part in early development. Wilson suggests that the polar lobe of such eggs as Dentalium may function as an organizing region similar to the dorsal blastoporal lip of amphibia, since only when the lobe is present does the larva develop the apical tuft and the post- trochal region. However, in the absence of transplantation experi- ments, no final conclusions could be reached. Schleip (1929) describes a "natural experiment" in which a second polar lobe is added to the egg of Dentalium. Among giant eggs, some are found which appear to be fusions of two ova at their vegetal hemispheres. In these eggs, a single large polar lobe may be formed, which goes in its entirety into one of the cells. This leaves one c-u with no lobe, and the other with two. However, such eggs do not develop. Schleip then tried to transplant isolated polar lobes to blastomeres, but all attempts were unsuccessful. 211 212 ALEX B. NOVIKOFF Transplantation experiments can readily be performed in the egg of Sabellaria vulgaris. In this species, large polar lobes, similar to those found in Den tali urn, are formed in the course of the first three cleavages. Although the egg is only about sixty micra in diameter, it is not difficult to remove the polar lobe or to separate individual blastomeres. Both blastomeres and polar lobes can be fused together in desired combinations and the eggs reared through a well -developed trochophore stage. In all, 247 successful transplant operations have been studied, including 80 in which either the first or second polar lobe was transplanted. MATERIAL AND METHOD All experiments reported in this paper were performed at the Marine Biological Laboratory, Woods Hole, Mass , during the summers of 1935, 1936, and 1937. The animals used were dredged from Vineyard Sound and the eggs were obtained in the manner described elsewhere by the writer (1937). The egg of Sabellaria vulgaris possesses a tough, wrinkled vitelline membrane which resists cutting with glass needles. In addition, within the membrane, in the perivitelline space, there is a dense jelly which makes difficult the separation of individual blasto- meres and which completely prevents bringing blastomeres together. Thus, to perform transplantations, it is necessary to first remove both membrane and jelly from the egg. The vitelline membrane is removed from fertilized eggs of Sabellaria by treatment with an isotonic solution of NaCl, brought to pH 9.6 by the addition of Na2CO3 (Novikoff, 1938). In most cases, the eggs are thus treated within the ten minutes that elapse between the forma- tion of the first polar body and the second. After they have been washed once in sea water, the denuded eggs are placed into a Syracuse dish of freshly-filtered sea water. The dish is allowed to remain without disturbance on the stage of a dissecting microscope. Within two or three minutes, the eggs have settled and are adhering to the glass bottom of the dish. By means of fine glass needles, each egg is then lifted from the jelly which remains adherent to the bottom. This process may have to be repeated several times in order to remove the jelly completely. The jelly being invisible, its removal can best be ascertained by bringing together the individual eggs; they come into contact with each other only when the jelly has been removed. At this time the eggs are quite sticky and if allowed to adhere too long to the dish they flatten out. Since such flattened eggs do not develop normally, it is important that the eggs be lifted, at close intervals, from the bottom of the dish. DETERMINATION IN SABELLARIA 213 The eggs are cut, free-hand, under the dissecting microscope. In order to determine the orientations of cells when fused, small spots are marked on the eggs before they are cut. This is done by bringing into contact with the surface of the egg the open end of a fine capillary tube filled with agar, in which is dissolved a vital dye such as Nile Blue sulphate. In order to fill an exceedingly fine bore, the following pro- cedure is followed. A short piece of capillary tubing is partly filled by immersing one end in a warm solution of the agar containing the dye. When the agar has cooled, that part of the tube which has no agar is heated by a microflame and pulled out to the desired width of bore. The other end of the tube is then sealed off and the microflame applied to the part of the agar nearer the narrowed end. The agar melts and moves into the free end of the tube. On cooling, the agar, in many cases, remains at the opening of the tube. The tube is brought into contact with the egg by means of a Zeiss- Peterfi micro- manipulator. The length of time during which the agar must remain in contact with the egg varies with the concentration of dye used. In some cases, where it is possible to determine the polarity of the isolated cell without previously staining a particular region, the entire cell is stained before transplantation. This is usually done before the mem- brane is removed, and a dilute solution of Nile Blue sulphate in sea water is used. To effect the fusion of blastomeres it is only necessary to bring them into contact with each other, after the membrane and jelly have ' been removed. It is usually sufficient to press them together for several seconds, although this may, in some cases, have to be repeated several times before they finally stay together. Following the operation, the eggs generally develop into swimming larvae within five to six hours after fertilization. They can at that time no longer be left in the open dish for they soon swim to the surface, where they are quickly torn by the surface film. The procedure which leaves the least amount of surface exposed to the air and which, at the same time, is most convenient for the detailed study of the living larvae, is to allow the embryos to develop in small drops of sea water between a glass slide and coverslip. When an embryo begins to show signs of movement, it is transferred, by means of a mouth pipette (Horstadius, 193 7a), to a small drop of sea water. Evaporation from the drop is prevented by sealing the edges of the coverslip with a thin layer of vaseline. The results to be described are based on a total of 403 experimental larvae, 156 developing from isolated blastomeres and 247 from fusions of various blastomeres. The percentages of larvae surviving are: 73 214 ALEX B. XOVIKOFF per cent on the first day, and 51 per cent on the second day, from isolated cells; 91 per cent on the first day, and 74 per cent on the second day, from combinations of blastomeres. As described in the first paper of the present series (Novikoff, 1938), a great number of the larvae arising from denuded eggs develop into exogastrula?, in which there is no internal gut and the endoderm cells are turned inside out. Under the coverslips, all structures differentiate in the larva?, just as they would in larger volumes of sea water, with the exception of the orange- red eye spot, and, to some extent, the yellow-green chromatophores. These are frequently absent or irregular. That the failure of the eye spot and the chromatophores to develop normally is due to some general factor such as lowered oxygen tension or increased pressure, is indicated by the development of normal eggs (i.e., eggs from which the membranes have not been removed) under similar conditions. Al- though the cell arrangement and the tissues of the larvae from such eggs are manifestly similar to those of normal larva?, the eye spot does not form and the chromatophores show the same variable character as do those of the membraneless larvae. NORMAL DEVELOPMENT Changes in the egg of Sabellaria vulgaris at maturation and during fertilization are described elsewhere (Novikoff, in press). Both polar bodies are formed after insemination and cleavage begins approximately twenty minutes * after the extrusion of the second polar body. In the course of the first cleavage division, a large part of the vegetal hemisphere of the egg becomes constricted in the form of a spherical lobe. Figure 1, A is a photograph of an egg at the "trefoil" stage, when the polar lobe is at its maximum size. When viewed from the side, the lobe appears to be equal in size to either of the first two blastomeres. But when viewed from the vegetal pole, it is seen to be considerably smaller. The visible constituents of the lobe cytoplasm do not differ from that of either blastomere, except that there is no spindle area in the lobe. About fifteen minutes after it first appears, the polar lobe flows into one of the blastomeres. This blastomere is the CD cell and it is now much larger than the other, the AB blastomere (Fig. 1, B). Figure 1, C shows an egg during the second cleavage, when the second polar lobe is at its maximum size. This lobe forms in the CD blastomere only, and is smaller Mian the first lobe. When it flows back into one of the daughter cells at the completion of the division, the four quarter-blastomeres consist of two equal-sized cells, A and B (the products of the division of AB), a slightly larger cell, C, 1 All time intervals are for room temperatures, varying from 19° to 25° C. DETERMINATION IN SAI1ELLARIA 215 and a much larger cell, D (Fig. 1 , D). It is the D blastomere which has received the contents of the second polar lobe. During the next divi- sion, when the micromeres are produced, a third polar lobe, formed B FIGS. 1-10. All figures, except 1 and 5, are composed of camera lucida drawings of living larvae, magnified approximately 260 times. Abbreviations used are: p prototroch at apical tuft ac apical cilia ec cilia of exogastrulated endoderm cells pc posterior cilium dc dorsal cilium pb post-trochal bristle e eye m mouth FIG. 1. Photomicrographs of living eggs of Sabellaria vulgar is. A. Trefoil stage. Side view. The polar lobe is slightly out of focus. B. Two-cell stage. Blastomere CD is to the right of A B. C. Second cleavage. The second polar lobe and two of the quarter-blastomeres are in sharp focus; the other two cells are not in focus. D. Four -cell stage. Seen from the vegetal pole. The large D cell is to the right and the A cell is uppermost. E. Early trochophore larva. Sixteen hours after fertilization. Photographed with dark- field illumination; shows apical tuft and prototrochal cilia. F. Later trochophore larva. Forty-eight hours after ferti- lization; shows stomach, intestine, dark eye spot, prototrochal cilia, and post-trochal bristles. from the D cell, flows into the D macromere, ID. This lobe is smaller than the second lobe and is more variable than the preceding lobes; 216 ALEX B. NOVIKOFF in many cases this lobe does not become distinctly separated from the dividing D cell. The later cleavages of the egg have not been described. A detailed description of the development of the larva is presented in the first paper of this series (Novikoff, 1938). In the normal course of development, the ectoderm gives rise to the following structures: a prototroch of long active cilia; chromatophores; an apical tuft ; non- motile apical cilia; paired chaeta>sacs from which extend long, serrated bristles; a single posterior non-motile cilium; several long non-motile cilia on the dorsal surface; an eye spot; and a neurotroch of rapidly moving cilia. The endoderm differentiates into a tripartite gut con- sisting of oesophagus, stomach, and intestine. In many ot the mem- braneless larvie, the gut is exogastrulate and the three portions can not be distinguished. Of the ectodermal structures, those most easily observed in living embryos — and therefore the ones best suited for FIG. 2. Larva; from isolated blastomeres. A. E-PL1 larva, seventy-one hours; with internal jjut. B. E-PLl larva, fifty hours; with exo^astrulated endoderm. C. AB larva, fifty-one hours; with internal .nut. D. AH larva, fifty-six hours; \\ith exo.ya-trulated endoderm. the present study — are the apical tuft, the post-trochal bristles, the prototrochal cilia, and the apical cilia. The apical tuft forms at about six hours after fertilization, and persists for approximately twenty hours (Fig. 1, /'."). Before it disappears, there develop at the apical end, a number of stiff cilia; these apical cilia remain throughout larval development. The prototrorhal cilia appear at about the same time as the apical tuft and they remain throughout larval development. The post-trochal bristles make their appearance toward the end of the first day of development. They increase in length and number as development progresses (Fig. 1, F). ISOLATION EXPERIMENTS E-TLl . An egg from which the polar lobe is removed at the treloil stage is labelled K-l'/J. The first cleavage of E-PLl differs from i In- normal second cleavage in that no polar lobe is formed. As a result, DETERMINATION IN SABELLARIA 217 the quarter-blastomeres are equal in size. In addition, no lobe is formed at the next division. The larva? of such eggs differ from the normal larva in that: (1) no apical tuft forms; (2) no post-trochal region (including the bristles) appears; and (3) the prototrochal cilia, although of the same size as in normal larva?, are at the posterior end of the larval ectoderm (Fig. 2A, B). Although no apical tuft is present, at ap- proximately twenty-four hours after fertilization the typical non- motile cilia appear. AB. — The isolated AB cleaves without the formation of polar lobes and gives rise to a spherical larva similar to that of E-PL1. It lacks the apical tuft and post-trochal bristles. It develops normal proto- trochal and apical cilia, although the former are situated at the posterior end of the larval ectoderm (Fig. 2C, D). FIG. 3. Larvae from isolated blastomeres. A. CD larva, twenty-five hours. B. CD larva, eighty-five hours. C. CD-PL2 larva, fifty-six hours. D. E-PL2 larva, forty-five hours. (On the next day, the apical tuft was no longer present.) E. E-PL2 larva, eighteen hours. It is exceptional in that it possesses two apical tufts. CD. — During the first cleavage of CD there is formed a polar lobe of the same size as the normal second lobe. After the division, it flows into the D cell. At the second cleavage, another, smaller lobe forms from the D cell. This lobe has the variable character of the normal third lobe. The early CD larvae appear to be quite normal — the prototrochal cilia are in their usual position and the typical apical tufts are formed (Fig. 3, A}. However, since the cilia and tuft are of the normal size, they are, proportionately, too large for these reduced larva?. Later, the paired bristles are formed. However, no apical cilia develop, and in many instances (17 out of 29), the apical tuft does not disappear when it does in controls (Fig. 3, B). CD-PL2. — -If during the course of the first cleavage of the CD blastomere, the polar lobe is removed, the next cleavage occurs without the formation of a lobe. The CD-PL2 larva possesses a typical apical 218 ALKX B. XOYIKOl-F tuft but lacks the post-trochal structures as well as the apical cilia (Fig. 3, C). E-PL2. — It is with difficulty that the second polar lobe is removed from a whole egg. In each of the three larva? obtained, normal apical tufts and prototrochal cilia develop. (One larva possesses two apical tufts — Fig. 3, £.) The older larva? show no post-trochal structures, but do develop apical cilia (Fig. 3, D}. BC. — Before the cells have shifted their position after the second cleavage, it is possible to divide them so that BC is separated from AD. The BC combination forms no polar lobe and gives rise to a larva which possesses the typical apical tuft and prototrochal cilia, and a few apical cilia. It lacks completely the post-trochal region (Fig. 4, B). FIG. 4. Larva? from isolated blastomeres. A. Al> larva, forty-nine hours. 11. B C larva, fifty hours. C. ABC larva, t unit v-six hours. D. ABD larva, sixty- six hours. E. D larva, forty-seven hours. F. C larva, thirty-three hours. G. D-PL3 larva, forty-one hours. AD. — At the first cleavage of AD, a small lobe, which has the variable character of the normal third lobe, forms in the D cell. The larva developing from this combination of cells forms no apical tuft. It possesses normal prototrochal cilia and the typical post-trochal region from which extend the bristles. Apical cilia are also present (Fig. 4, A}. ABC. — By destroying the D cell at the completion of the second cleavage, ABC combinations are obtained. No polar lobe is formed at the first cleavage. The larva which develops possesses the typical apical tuft and apical cilia. However, no post-trochal bristles are produced (Fig. 4, C). ABD. — Destruction of the C cell leaves the ABD blastomeres. At the next cleavage, the D cell forms a small polar lobe. The ABD DETERMINATION IN SABELLARIA 219 larvae develop normal prototrochal cilia, but no apical tufts. Later in development, apical cilia and post-trochal bristles appear (Fig. 4, D). C. — No polar lobes are formed during the cleavage of the C blaslo- mere. The larva developing from the C cell develops typical proto- trochal cilia and apical tuft. But neither post-trochal structures nor apical cilia are formed (Fig. 4, F). D. — During the first cleavage of the D cell, the small variable lobe appears and passes into the macromere, \D. The D larva has typical prototrochal cilia and post-trochal bristles. But it develops no apical tuft and no apical cilia (Fig. 4, £). D-PL3. — In some cases, the lobe formed by the D cell constricts sufficiently so that it may be removed. The larva developing from TABLE I Summary of larvie obtained from isolation experiments. Number Operated Number Surviving Apical Tufts + ? Post-trochal Bristles + Prototrochal Cilia + - ? Apical Cilia + - ? E-PL1 30 21 0 21 0 11 21 0 14 0 AB 36 25 0 24 0 18 23 0 1 11 6 1 CD . 56 47 37 1 9 32 1 42 0 5 1 26 6 CD-PL2 8 5 5 0 0 4 5 0 0 3 1 E-PL2 3 3 3 0 0 2 3 0 2 0 BC 10 8 6 2 1 5 8 0 6 0 AD 6 3 0 3 3 0 3 0 2 0 1 ABC 10 8 7 0 1 0 5 8 0 4 1 ABD 7 5 0 5 4 0 5 0 4 0 C 17 11 7 3 1 0 7 11 0 0 7 D . . . 23 15 0 12 3 9 4 15 0 094 D-PL3 7 5 0 5 0 3 5 0 0 3 213 156 the D-PL3 lacks the post-trochal structures as well as the apical tuft and apical cilia (Fig. 4, G). Table I summarizes the isolation experiments. All combinations of cells develop prototrochal cilia. But only those combinations which include the substance of the C cell (earlier found in the first polar lobe) develop apical tufts; only those which have the ID cell materials (and previously found in the three polar lobes) form the post-trochal region; and the apical cilia develop only when the A or B cell is present. Three exceptions, a BC larva with bristles, a CD larva with apical cilia, and an E-PL2 larva with two apical tufts, will be discussed later. 220 ALEX B. NOYIKOFF Behavior of Isolated Polar Lobes The changes in form of isolated polar lobes have been described in Dentaliiim, by \Yilson (1904&) and in Ilyanassa, by Morgan (1933, 1935). In Dentalium, the isolated lobe usually constricts periodically to form lobe-like structures by a process which simulates the form.it ion of polar lobes in the- whole egg. There 'are three such constrictions, and they occur at approximately the same time as do the cleavages of the whole egg. One case is described in which the final constriction gives rise, not to a temporary lobe, but to a distinctly separated portion of cytoplasm. Wilson interprets the first of these changes as the formation of lobes within the isolated polar lobes, at the time when tin; whole egg normally forms lobes. The final stage he regards as a PLl PLz V PLl PL 2 Fl«. 5. Form changes in isolated polar loin--. PLl, first polar lobe. PL2, second polar lobe. Numerals 1 1 indicate the time of the first four cleavages of the whole egg. Explanation in text. permanent division of the isolated lobe into two at the time of the fourth cleavage of the whole egg, in which (he material of the polar lobe no longer forms a temporary polar lobe, but is permanently cut off by a cell division. Morgan argues against a literal interpreta- tion of the form changes in the isolated lobe as lobes. In Ilyanassa, he finds that : ( 1) The changes in the lobe are not strictly synchronous with the cleavage of the whole egg. (2) The constrictions in isolated lobes come and go at least three times, whereas in the whole egg, the lobe would appear only once more. (3) The change in shape does not, strictly speaking, give a reduced picture of the changes in the whole egg; and, (4) the form changes in the lobe resemble more the process of DETERMINATION IN SABELLARIA 221 micromere constriction rather than lobe formation. Therefore, Morgan interprets the later changes of the isolated lobe as related to the constriction of micromeres, and not to the formation of polar lobes. First Polar Lobe. — Isolated first polar lobes of Sabellaria were observed continuously for two to three hours. Although, as both Wilson and Morgan found, the behavior is somewhat variable, most lobes show a remarkable constancy in their changes. Of the 31 iso- lated lobes studied, 25 were of the type drawn in Fig. 5; the other six were more or less variable. No change in the shape of the isolated lobe occurs at the time of the first cleavage of E-PL1, i.e., at the time of the second cleavage of the whole egg (Fig. 5, b). About ten to fifteen minutes after the first cleavage of E-PL1, the lobe is deeply constricted to form a de- finite lobe-like structure (Fig. 5 , c) . At the time of the second cleavage, the lobe is spherical (Fig. 5, d), but another deep constriction appears about ten to fifteen minutes later (Fig. 5, e). As this second constric- tion disappears, a slight flattening of the lobe occurs (Fig. 5,/). This lasts for a short time, and when E-PL1 cleaves for the third time, the lobe is again spherical (Fig. 5, g). A third constriction forms after the third cleavage of E-PL1 (Fig. 5, h), but before it disappears completely, the entire cell becomes irregular in outline, and long irregular pseudo- podia are formed (Fig. 5, i). These are later withdrawn (Fig. 5, j) and the cell again becomes spherical (Fig. 5, k). However, it remains in this condition for only a short time, until the irregular pseudopodia are formed again (Fig. 5, /). The extension and retraction of the pseudopodia is not synchronous with cleavage, and continues until the lobe cytolyzes. In several cases, the process was still going on twenty-eight hours after the removal of the lobe; in one case, it was observed up to forty-eight hours after the separation of the lobe. Second Polar Lobe. — The second polar lobes were removed from five isolated CD cells, and their behavior followed continuously for two hours. Four of the five lobes produced fairly deep constrictions twice (Fig. 5). The other produced only the first of the two constrictions; at the time when the second constriction would form, the cell elongated somewhat without constricting. The behavior of the four lobes which formed the two constrictions was fairly uniform. When CD-PL^ divides for the first time, i.e., at the time of the third cleavage of the whole egg, the isolated lobe elongates slightly (Fig. 5, b'} and within a minute is rounded out (Fig. 5, c'). A similar elongation forms again, in three of the lobes, in about eight to ten minutes after the first (Fig. 5, d'}. At about seventeen to eighteen minutes after the first cleavage, the lobe develops a constriction which persists for two ALEX B. NOVIKOFF minutes (Fig. 5,/'). As the constriction disappears, the lobe flattens slightly (Fig. 5, g'}. \Yithin a few minutes, the second cleavage of CD-PL2 occurs. Following the second cleavage a second constriction forms (Fig. 5, i'), and as it disappears the cell becomes irregular and gives rise to pseudopodia (Fig. 5, /). The behavior of the pseudopodia is similar to those of isolated first lobes (Fig. 5, j'-ri}. The constrictions formed by the isolated first and second polar lobes resemble in appearance the polar lobes formed by the dividing ovum. Although these constrictions do not occur at the same time as the cleavages of the ovum, they must, in some way, be related to events taking place in the egg during division. This is brought out by a comparison of the time elapsing between successive cleavages of the lobeless egg and the constrictions of the isolated lobe, in the twenty- five cases where the first polar lobe was removed and the five in which the second lobe was removed. The average time between the first and second cleavages of E-PL1 is 20.6 minutes and that between the first and second constrictions of the isolated first lobe 18.6 minutes. The time between the second and third cleavage of E-PL1 is 24.6 minutes and that between the second and third constriction is 21.2 minutes. For the isolated second lobes, the average time between the two constrictions is 21.8 minutes, while the interval between the first and second division of CD-PL2 is 21.0 minutes. The changes in the form of the isolated lobes are apparently synchronous with the cleavages of the ovum, except that all events in the isolated lobes are pushed back by a delay in the appearance of the initial constriction. This delay is approximately ten to fifteen minutes for the isolated first lobe and about seventeen to eighteen minutes for the isolated second lobe. On the basis that the periodic constrictions formed by isolated polar lobes are correlated with cytoplasmic changes occurring in whole eggs at the time of the cleavages in which polar lobes are formed, we would expect the isolated second lobe to form one fewer constriction than the isolated first lobe. The isolated first lobe forms three con- strictions; the isolated second lobe two. And, as would be expected, the time elapsing between the first and second constrictions of the isolated second lobe (21.8 minutes) is almost identical with the time between the second and third constrictions of the isolated first lobe (21.2 minutes). However, certain differences between the behavior of the isolated lobes and that of polar lobes of the whole egg must be noted. The whole egg forms two polar lobes after (he formation ol the first lobe, and one after the second. The isolated first lobe, however, forms three constrictions and the isolated second lobe two. DETERMINATION IN SAP.ELLARIA 223 Although there is a decrease in size of successive lobes in the whole egg, the constrictions of the isolated lobe are all of approximately the same size. But the facts to be emphasized are: (1) that there reside in the isolated polar lobes materials which take part, independently of the nucleus or the mitotic apparatus, in reactions affecting the tension at the surface of the cell, and (2) that these reactions in isolated lobes occur synchronously (if we discount the initial delay) with events in the whole egg, or in the egg from which the polar lobe has been removed. TABLE II Differentiation of larvae after transplantation of polar lobes and blastomeres. Type of Operation Num- ber Oper- ated Num- ber Sur- vived Larva- with: Larva with: Larvae with: 0 Api- cal Tuft 1 Api- cal Tuft 2 Api- cal Tufts OSets of Bris- tles ISet of Bris- tles 2 Sets of Bris- tles 0 Groups of Apical Cilia 1 Group of Apical Cilia 2 Groups of Apical Cilia A. Polar lobe transplants Transplant Host PL1 Whole egg.. AB E-PL1.. PL2 Whole egg . . AB... . 26 35 15 G 6 23 32 14 5 G 7 32 14 0 6 10 0 0 5 0 0 0 0 0 0 8 20 10 2 3 14 0 0 o 0 0 0 0 0 0 7 5 3 1 0 15 14 7 3 3 0 0 0 0 0 B. Transplantation of blas- tomeres to whole egg Transplant CD 9 10 16 11 9 10 14 11 0 0 2 2 o 4 10 9 7 6 0 0 0 0 0 2 1 8 4 8 G 0 10 0 1 2 2 1 5 G 12 4 0 0 0 5 C D AB C. Transplantation of blas- tomeres to E-PL1 Transplant CD 16 20 15 4 12 3 16 18 15 3 11 3 2 6 15 1 11 1 11 12 0 2 0 2 0 0 0 0 0 0 2 14 0 2 9 0 12 0 13 0 0 3 0 0 0 0 0 0 5 5 4 0 0 0 9 8 9 2 3 1 0 0 0 0 G 2 C . D CD-PL? AB Whole egg . . D. Fusions of blastomeres Combination CD, CD 20 9 6 4 12 3 1 19 8 4 3 10 2 1 0 0 1 0 11) 2 1 11 7 2 2 0 0 0 8 0 1 0 0 0 0 2 1 0 2 8 1 1 2 1 2 0 0 0 0 11 5 0 0 0 0 0 15 7 2 2 1 0 Thret 0 0 0 0 2 0 groups o 0 0 0 0 f) 1 "cilia CD, D CD, C. . C C AB, AB 3 AB's. . 5 AB's TRANSPLANTATION EXPERIMENTS Transplantation of Polar Lobes Transplants of first and second polar lobes were made to whole eggs at the trefoil, two-cell, and four-cell stages, to AB blastomeres, and to E-PL1. The lobes were placed at the animal pole, at the vegetal pole, and at the equator of the dividing egg. Although 224 ALEX B. NOYIKOFF fused to the blastomeres, the lobes go through essentially the same form changes as do isolated lobes. When cleavage progresses, the blastomeres may grow over and completely enclose the lobe, or the lobe may remain at one end of the larva. The lobe is still part of the larva when the cilia appear (Fig. 6A, C). Prototrochal cilia form in all larvae, and the apical tufts develop only in those which include C cells (Table II, A and Fig. 6). At about fifteen hours after fertilixa- tion, the lobe, more or less completely cytolyzed, is extruded from the larva (Fig. 6, D}. This is generally followed by a cytolysis of a portion of the embryo, especially in those cases where the lobe was deeply embedded within the embryo. Thus, only 59 out of the 80 larva? FIG. 6. Differentiation of larvae after transplantation of polar lobes. A. AB plus PL1 larva, llj-jj hours. The lobe (stippled) is within the larva. B. AB plus PL1 larva, 24 hours. The larva has moved away from the cytolyzed spheres of the ejected lobe. C. E plus PL1 larva, 13 hours. The lobe (stippled) is within the larva. D. E plus PL1 larva, eighteen hours. Tin- lobe has been ejected. E. AB plus PL1 larva, 45 hours. The ejected lobe has broken up into small spheres. F. E plus PL1 larva, 26 hours. G. E-PL1 larva, 50 hours. The first polar lobe was in contact with the dividing egg for five hours, at the end of which time it was removed. survived beyond twenty-four hours. Where the larva is not great ly damaged by the loss of the lobe, differentiation progresses normally. The AB and E-PLl larva> form apical cilia but no post-trochal regions (Fig. 6, B). The larva? produced by entire eggs lose their apical tufts when apical cilia appear, and they develop post-trochal regions with typical bristles (Fig. 6, F). In many cases, the polar lobe was stained heavily with Nile Blue sulphate, before being transplanted. The dye diffused from the lobe into the adjacent cells, so that the larval tissues acquired a pronounced blue coloration. DETERMINATION IN SABELLAKIA 225 To avoid possible effects of the lobe cytolysis on the differentiation of the larva, the lobe was, in fifteen cases, allowed to remain fused to the dividing cells only until the embryo began to show signs of move- ment, when it was removed. The contact is long enough so that the blue color of the dye diffuses from the lobe into adjacent cells. The larvae which develop from such cells are in no essential way altered by contact with the lobe (Fig. 6, G}. The presence of the polar lobe, fused to the larva for five hours (when it is removed), or within the larva for eleven hours (at which time it is ejected), does not induce the formation of larval structures. Transplantation of Blastomeres to Whole Eggs Table II, B summarizes the types of transplants made to whole eggs, at the trefoil, two-cell, and four-cell stages. The transplants FIG. 7. Differentiation of larvae after transplantation of blastomeres to whole eggs. A. E plus CD larva, 27 hours. B. E plus C larva, 24 hours. C. E plus I) larva, 79 hours. D. E plus AB larva, 40 hours. are in some cases stained with Nile Blue sulphate before being fused to the host. It is possible to follow the stained region through the first day of development, but beyond this time the dye is not visible. The orientation is varied, but no correlation between any particular orientation and type of development is found. The differentiation of the transplanted AB, CD, C, or D blasto- meres does not bring about the development of additional structures in the host; those structures which are duplicated in the larva arise from the self-differentiation of the transplanted cells. Seven of the nine E plus CD larvae show two apical tufts; six (out of the seven surviving) show two sets of post-trochal bristles (Fig.- 7, A). In six of the ten E plus C larvae, two apical tufts are visible (Fig. 7, B). Ten of the fourteen E plus D larvae possess two sets of bristles (Fig. 7, C). The E plus AB larvae show no duplication of either the apical tuft or 226 ALEX B. NOVIKOFF bristles, but five of the ten possess two distinct groups of apical cilia Transplantation of Blastomeres to E-PLI The polar lobes are removed from eggs at the trefoil stage, and in their place are put half- or quarter-blastomeres, or whole eggs. In other experiments, transplants are placed at the animal pole, or to the side of E-PLI . As in previous operations, the transplanted cells are first stained, either in loto, or locally, to mark their polarity. The resultant larvae are summarized in Table II, C and a few of the various kinds are shown in Fig. 8. In all larva', the host cells give rise to FIG. 8. Differentiation of larvae after transplantation of blastomeres to E-PLI. A. (E-PLJ) plus CD lafvn, 43 hours. B. (E-PLl) plus C larva, 30 hours. C. (E-PLI) plus C larva, 48 hours. D. (E-PLI) plus D larva, 17 hours. E. Same larva as shown in D, 43 hours. F. (E-PLI) plus AB larva, 54 hours. G. (E-PLI) plus E larva, 25 hours. H. Same larva as shown in G, 91 hours. prototrochal cilia, and in at least 36 of 55, they form apical cilia. They form no apical tuft and no post-trochal bristles. The apical tufts, post-trochal bristles, or supernumerary prototrochal and apical cilia which are present in the larva* arise only through the self-differentia- tion of the transplant. Fusions of CD, C, and D Blastomeres Thirty-four operations involving combinations of CD cells with CD, D, and C blastomeres, and fusions of two Cecils were performed (Table II, /; and Fig. larva- with the Cl>, CD combination, 8 show two apical tufts; in 11, only one tufl is distinct Iv DETERMINATION IN SARELLARIA 227 visible. Of the 14 larva* surviving beyond (he first day, 2 fail to develop post-trochal bristles, 2 form one set of bristles, and 11 develop two distinct sets. There are 7 larva? with the CD, D combination. Each of these larvae produces one apical tuft, and 5 of the 7 show two sets of post-trochal bristles. ( )m> larva forms one set of bristles and another forms none. Four larvae of the CD, C constitution show the following: two apical tufts, 1 ; one apical tuft, 2; no apical tuft, 1. Two survive long enough to develop bristles; in each there is one set. Only two C plus C larva? survive; each possesses only one apical tuft and none forms post-trochal bristles except those which include D cells. FIG. 9. Differentiation of larvae arising from fusions of blastomeres. A. CD plus CD larva, 13 hours. B. Same larva as shown in A, 59 hours. C. C plus C larva, 28 hours. E. CD plus D larva, 55 hours. D. AB plus AB larva, 50 hours. Fusions of AB Blastomeres At the time of the first cleavage of isolated AB cells, various combinations are effected. Of the larva? surviving, ten come from fusions of two AB cells, two from three AB's, and one from five. Five of seven of the AB, AB larvae show two sets of apical cilia, and two show one set; in the single surviving larva from the combination of three AB's, two sets of apical cilia are seen ; in the larva from the fusion of the five AB's three distinct sets of apical cilia are visible. Although there is a superabundance of cellular material, none of the larvae develops an apical tuft and none develops a post-trochal bristle (Table II, D and Fig. 9, D). 228 ALEX B. NOVIKOFF Fusions of Whole Eggs Two eggs are fused at the trefoil stage or at the two-cell stage. The orientation of the two eggs with respect to each other is varied as follows: (1) the animal-vegetal axes of the ova remain parallel, but the eggs are rotated to different degrees; (2) one of the two eggs is inverted so that the fusion occurs at the animal poles or at the vegetal poles of the two eggs, with the rotation of the eggs varied as in (1). j FIG. 10. Differentiation of larvae arising from fusions of two eggs. A. Larva, 23 hours. B. Same larva as shown in A, 52 hours. C. Another larva, 27 hours. />. Another larva, 49 hours, with one internal gut. On the outer surface is an area of ciliated cells, devoid of chromatophores. E. Another larva, 55 hours, with two internal guts. Of the twenty-live fusions effected, 23 larvae are alive at the end of twenty hours of development, and \() at the end of thirty-five hours. Of the 23, 2 larva* show no apical tuft, 7 show one tuft, and 11 sliou two tufts. Among the 19 larva?, the post-trochal bristles fail to form in one case, 4 larva; have one set of bristles, and 14 have two sets; the apical cilia are not visible in 2 larvae, 5 show one group of cilia, and 12 show two distinct groups. In none of the larva? are there more than two apical tufts, or more than (wo sets of post-trochal bristles, or more DETERMINATION IN SABELLARIA 229 than two groups of apical cilia. Figure 10 includes three advanced larvae of different types. In one type (Fig. 10, B), the endoderm cells of both eggs have been exogastrulated. In another (Fig. 10, E), there are two distinct internal guts, of fairly normal structure. In a third type (Fig. 10, D), only one internal gut is found. This gut has the typical tripartite character and is no larger than the normal gut. On the outer surface of the larva, a wide, delimited, area is devoid of the chromatophores characteristic of the ectoderm, and the cells of this area are covered with the rapidly-moving cilia, characteristic of the gut cells. The constitution of this area is therefore interpreted as the endoderm cells of one of the eggs. Instead of giving rise to a gut, these cells have become part of the outer covering of the larva; but in spite of their new location, they continue to differentiate as they would normally. In two instances, three eggs are fused at the two-cell stage. The resultant larvae develop three sets of post-trochal bristles. In one case, four eggs are fused at the two-cell stage, and the larva develops four sets of bristles. DISCUSSION OF RESULTS By means of isolation experiments, the independent developmental capacities of early blastomeres of Sabellaria vulgaris were determined. These experiments included the usual separation of half- and quarter- blastomeres, and, in addition, the separation of other combinations of cells (AD, BC, ABD, and ABC). Also, the effect of the removal of three polar lobes was studied. Table I classifies the 156 surviving larvae, with respect to the presence or absence of the prototroch, apical tuft, post-trochal bristles, and apical cilia. Prototrochal cilia were present in all larvae. The apical tuft formed only in those larvae which included the first polar lobe and the C cell. The post-trochal bristles developed only when the three polar lobes and the \D cell were present. The differentiation of apical cilia occurred only when either the A or the B cell was present. Two of the larvae were exceptions to this conclusion: one CD larva, which developed apical cilia, and one BC larva, which developed post-trochal bristles. A third larva, of the E-PL2 type, formed two apical tufts instead of one. Having determined the fate of the blastomeres when isolated, the behavior of these same cells when placed in contact with each other atypically, or with isolated polar lobes, was investigated. The twenty- two types of transplantation experiments, including a total of 247 larvae, are summarized in Table II. Not a single case is found in which the transplant had induced the formation of any particular structure 230 ALEX H. NOYIKOFF in (he host. In any combination of blastomeres, apical tufts develop only when Cecils are present, post-trochal bristles are dependent upon the presence of D cells, and apical cilia form only when either the A or B cell is included. In Table III, the larva? are classified on a different basis. The number of C, D, and A or B cells included in I he makeup of tin- individuals are compared with the number of apical tufts, post-trochal structures, and apical cilia. In general, the number of apical tufts is TABLE III Classification of larva; developing from transplantation experiments. Number of C Cells Total No. of Larvae Number of Larva with 0 Apical Tufts 1 Apical Tul't 2 Apical Tutts 0 88 88 0 0 1 95 21 74 0 2 64 3 28 33 Number of D Cells * Total No. of Larvae Number of Larva* with 0 Sets of Bristles 1 Set of Bristles 2 Sets of Bristles 0 69 69 0 0 1 63 14 49 0 2 62 4 12 46 Number of A (or /:. Cells Total No. of Larvae Number of Larva: with 0 Apical Cilia 1 Group of Apical Cilia 2 Groups of Apical Cilia 0 26 26 0 0 1 128 35 93 0 2 49 4 15 30 * Two larva? had 3 D cells; both had 3 sets of bristles. One larva had 4 D cells; it had 4 sets of bri-i It •-. determined by the number of C cells present, the number of sets of post-trochal bristles by the number of D cells, and the number of groups of apical cilia by the number of A (or B} cells. The absence of a higher correlation, particularly in the cases of the apical tuft and apical cilia, may possibly be due to several reasons: (1) The larva? are actively swimming and, especially in small, healthy individuals, such minute structures may be overlooked. (The post-trochal bristles are much more readily seen. The correlation is higher with this structure than DETERMINATION IN SAIJKLLAUI A 231 with either of the other two.) (2) Some of the operated individuals may have been injured and the failure to form a particular structure may be a manifestation of their reduced vitality. (3) Actively swim- ming larv;e may have some of their cells torn away by the surface at the edge of the drop of water. It is in many cases not possible to deter- mine whether this has occurred. (4) The mechanical effects of the neighboring tissues may prevent the cells from giving rise to the partic- ular structure. If true, this effect might possibly be conceived as a kind of regulatory process. The most significant feature of the transplantation experiments is that in no instance is the number of apical tufts greater than the number of C cells, nor the number of post-trochal regions greater than the number of D cells, nor the number of groups of apical cilia greater than the number of A (or B] cells. Of special interest are several larvae which developed from a fusion of two eggs at the two-cell stage. In these larvae (Fig. 10, D), there is present a single gut, of normal size and typical tripartite structure. In one of the two eggs, the cells whose prospective value is gut endoderm have been incorporated into the outer covering of the larva. Although "ectodermal" in the sense of location, these cells continue to differentiate into endoderm, i.e., they do not develop chromatophores and they become ciliated on their outer surface. The self-differentiation of the endoderm in this position indicates, as does the development of exogastrulse (Novikoff, 1938), the complete independence of endodermal and ectodermal differentiation in Sabel- laria vulgaris. Since the polar lobe, as well as any of the quarter- or half-blasto- meres, does not affect the differentiation of any cell through contact with that cell, it is not possible to consider the polar lobe an "or- ganizer," in the sense of Spemann (cf. Wilson, 1929, pp. 202-205, and Huxley and deBeer, 1934, pp. 171-172). The experiments of Tyler (1930) have demonstrated that when the first cleavage of the Chxtopterus egg is made equal — either through the retraction of the polar lobe into the smaller, AB, blastomere, or through the cleavage furrow dividing both the egg and the polar lobe equally — then the two half-blastomeres are totipotent. In both types of equal cleaving eggs, the AB blastomere receives polar lobe material. If allowed to develop in toto, such eggs produce double monsters of the cruciata type. (At the second cleavage, two very small polar lobes may be formed.) When separated, each half-blastomere produces a fairly normal larva. There is in reality no "AB" cell; each cell behaves like a CD. But there is no evidence that the substance, whose altered distribution ALEX B. NOVIKOFF changes the prospective value of the cell from that of an AB to that of a CD cell, is located within the polar lobe. The fact that a double embryo is produced when the polar lobe goes in its entirety into the AB, i.e., that the CD can produce an embryo without the materials of the polar lobe, indicates the complexity of the situation. That the vegetal hemisphere of the molluscan or annelid egg possesses a parti- cular substance at the time of the first cleavage is well established by the work on Ilyanassa (Crampton, 1896), on Dentalium (Wilson, 19046), on Tubifex (Penners, 1924, 1926), and on SabeUaria (Hatt, 1932, and the present paper). In Tiibifex, if the first cleavage of the whole egg or the CD blastomere is made to take place equally instead of unequally (through heat or lack of oxygen) — and the pole-plasms are distributed equally to the two cells — double cruciata monsters are produced (Penners, 1924). A single case was observed in SabeUaria in which the first two blastomeres were of equal size. \Yhen isolated, each of the blastomeres produced polar lobes at the next two cleavages, and each gave rise to a larva possessing an apical tuft. The fact that each cell forms polar lobes and that the two cells are of equal size indicates that each cell probably received materials from the first polar lobe. It would, then, appear that some substance present in the first polar lobe does have the ability to change the course of development of a cell, but that this substance does not act by contact with a cell ; it must become a part of the cell. Normally, only the CD cell develops an apical tuft since the materials of the lobe flow only into that cell. The exceptional production of apical cilia by a CD cell and the appearance of post-trochal bristles in a BC combination, mentioned earlier, may be due to a deviation in either of the first two cleavage furrows, with a resultant unusual allocation of materials. The doubling of the apical tuft in the E-PL2 larva may also be due to an unusual pattern, in which the materials giving rise to apical tufts are separated into differ- ent cells. In the so-called regulative eggs, it is in many cases possible to alter the course of differentiation of a cell by transplanting the cell to a new position in the developing embryo. By varying the stage at which the operation is performed, the time of determination of a structure may be ascertained. Due to the scarcity of similar experiments on mosaic eggs, relatively little information is available concerning the effect of one part of an embryo on another during the course of develop- ment, or the existence of inducing, or organizing, regions in these eggs. Penners (1926, 1934) destroyed varying numbers of mesodermal and ectodermal teloblasts of Tubifex, at different stages, to test (1) the inter-dependence of ectoderm and mesoderm during development, and DETERMINATION IN SARELLARIA 233 (2) the inducing capacity of the teloblasts. He found that the ecto- derm and mesoderm show complete independent differentiation, except for a slight influence of the development of one upon the form and upon the rate of development of the other. The fact that following the destruction of the teloblasts the embryos continue to develop normally indicates that the teloblasts are not organizing centers. Horstadius (19376) combined various quartets of blastomeres of the sixteen-cell stage in Cerebratuhis; he found no effect of one layer upon the differentiation of the others. These results are in agreement with those of the present investigation, in which the polar lobe, half-, and quarter- blastomeres are shown to be ineffective in directing the development of Sabellaria. However, in another egg which was thought to be mosaic, the egg of the ascidian, effects of one cell upon the development of the other have been reported. Tung (1934) found that some factor outside the brain is responsible for the formation of the sense organ, in Ascidiella. Also, Tung found indications that the adhesive organ is induced. A recent paper by Rose (1937) reports that, similarly, in the egg of Styela, the eye spot is induced by the gray macromeres. SUMMARY 1. Isolation experiments on the egg of Sabellaria vulgaris demon- strate that the formation of the apical tuft in partial larvae is dependent upon the presence of the first polar lobe and the C cell ; that the post- trochal region develops only when the three polar lobes and the ID cell are present; and that apical cilia form only if the A or B cell is included. 2. Form changes in isolated first and second polar lobes are de- scribed. The early changes are synchronous with the cleavages of the ovum, except that all events in the isolated lobe are delayed. 3. The results of the following transplantation experiments are reported: (a) Transplantation of polar lobes, (b) Transplantation of blastomeres to the whole egg. (c) Transplantation of blastomeres to E-PL1. (d) Fusions of half- and quarter-blastomeres. (e) Fusions of two eggs. In all combinations, complete self-differen- tiation of individual blastomeres occurs. Apical tufts develop only when C cells are present, post-trochal bristles are dependent upon the presence of D cells, and apical cilia form only when either the A or B cell is included. 4. The results of this investigation are compared with those from experiments on other mosaic eggs. It is with pleasure that the writer expresses his gratitude to Pro- 234 ALEX B. NOVIKOFF fessor L. G. Earth for his untiring assistance and constant encourage- ment throughout the course of this investigation, and to Professor E. B. Wilson for his inspiring interest in the work. BIBLIOGRAPHY CRAMPTON, HENRY E., 1896. Experimental studies on gasteropod development. Arch. f. Entw.-mech., 3:1. DELAGK, YVKS, 1899. Etudes sur la merogonie. Arch, de Zool. exper. el gen. (Ser. Ill), 7: 383. HATT, PIERRE, 1932. Essais experimentaux sur les localizations germinales dans 1'oeuf d'un Annelide (Sabellaria alveolata L.). Arch. d'Anat. Micros., 28: 81. HORSTADIUS, SVEN, 1937a. Microdissection, Free-hand Manipulations. In Hand- book of Microscopical Technique, edited by C. E. McClung. Paul B. Hoeber, p. 43. HORSTADIUS, SVEN, 19376. Experiments on determination in the early development of Cerebratulus lacteus. Bid. Bull., 73: 317. HUXLEY, JULIAN S., AND G. R. UKBEER, 1934. The Elements of Experimental Embryology. Cambridge. University Press. MORGAN, THOMAS H., 1933. The formation of the antipolar lobe in Ilyanassa. Jour. Exper. Zool., 64: 433. MORGAN, THOMAS H., 1935. The rhythmic changes in form of the isolated anti- polar lobe of Ilyanassa. Kiol. Bull., 68: 296. NOVIKOFF, ALEX B., 1937. Sabellaria vulgaris. In Culture Methods for Inverte brate Animals, edited by Galtsoff et al. Comstock Publishing Co. Inc., p. 187. NOVIKOFF, ALEX B., 1938. Embryonic determination in the annelid, Sabellaria vulgaris. I. liiol. Bull., 74: 198. PENNERS, ANDREAS, 1924. Experimentelle Untersuchungen -/.urn Determinations- problem am Keim von Tubifex rivulorum Lam. I. Arch. Mikr. Anat. u. Entw.-mech., 102: 51. PENNERS, ANDREAS, 1926. Experimentelle Untersuchungen zum Determinations- problem am Keim von Tubifex rivulorum Lam. II. Zeitschr. f. Wiss. Zool., 127: 1. PENNERS, ANDREAS, 1934. Experimentelle Untersuchungen zum Determinations- problem am Keim von Tubifex rivulorum Lam. III. Zeitschr. f. Wiss. Zool., 145: 220. ROSE, MERYL S., 1937. The induction of pigment spots in Styela partita (Abstract). Anat. Rec., 70 (Supplement 1): 102. SCHLEIP, WALDEMAR, 1929. Die Determination der Primitiventwicklung. Leipzig, Akad. Verlags. TUNG, Ti-Cnow, 1934. Recherches sur les potentialitcs des blastonu-res chez Ascidiella scabra. Arch. d'Anat. Micros., 30: 381. TYLER, ALBERT, 1930. Experimental production of double embryos in annelids and mollusks. Jour. Expt-.r. Zool., 57: 347. WILSON, EDMUND B., 1904a. Mosaic development in the annelid egg. Science, 20: 748. WlLSON, EDMUND B., 19046. Experimental studies on germinal localization. I. The germ-regions in the egg of Dentalium. Jour. Exper. Zool., 1:1. WILSON, EDMUND B., 1929. The development of egg-fragments in annelids. Arch, f. Entw.-mech., 117: 179. THE RELATION OF MORTALITY AFTER ENDOMIXIS TO THE PRIOR INTERENDOMICTIC INTERVAL IN PARAMECIUM AURELIA l BERNICE FRANCES PIERSON (From the Department of Zoology, The Johns Hopkins University) Investigators of endomixis in Paramecium have frequently observed that animals often die during or soon after endomixis. According to Erdmann and Woodruff (1916), few Paramecium caudatum survived endomixis. In Paramecium aurelia, Caldwell (1933) found that death occurred 2.85 to 9.40 times more frequently at endomixis than in the middle of the period between endomixes. The present paper is a presentation of the results of a study of one of the factors determining such deaths, namely, the length of the preceding interendomictic interval. As will appear below, when the prior interendomictic inter- val is unusually long, endomixis results in a greater percentage of deaths than when the interval is of the ordinary duration. Moreover, the greater the interendomictic interval, the greater is the mortality resulting from endomixis, until, after very long intervals, endomixis invariably results in death. In order to investigate this question, it was essential to have avail- able, simultaneously, lines with normal interendomictic intervals and sister lines with unusually long intervals. This was accomplished by employing recently developed methods of inducing endomixis (Sonne- born, 1937) and of obtaining lines with long interendomictic intervals (Sonneborn, 1938). To induce endomixis, the surplus animals from daily isolation lines of cultivation were collected in a small amount of fresh culture medium and kept at 31° C. for a few days until endomixis occurred. To obtain lines with long interendomictic intervals, daily isolation lines which went into endomixis were replaced by sister lines which had not yet gone into endomixis. In this way, lines with long intervals are selected for study while the lines with shorter intervals are eliminated. Using these techniques, the following experiment was performed. The vegetative descendants of a single endomictic individual were cultivated as 24 daily isolation lines of descent for 165 days. During this time, all lines that went into endomixis were eliminated and re- 1 This work was suggested by Dr. T. M. Sonneborn, to whom I wish to express my sincere appreciation for his helpful advice and assistance. 235 236 BERNICE FRANCES PIERSON placed by surplus animals from sister lines; so that at all times this group (I, Fig. 1) consisted of 24 lines that had not been in endomixis since the start of the experiment. At five successive intervals of 21 to 31 days, surplus animals from this group were induced to go into Time from initiol endomixis, in days 0 3 1 5 6 6 1 1C 2 12 b 165 T A' H ie>< '- 1C . ID' ^- IE u IE> i i i— DC r- ID r- IE i >v II Oe : 12 Jar - K .6 Jar 25, 31 feb 21 ^ 21 Mor 25 •. 15 Apr. 24 Intervals between successive cndomixes, m days FIG. 1. Plan of Experiment The horizontal lines (solid) represent groups of isolation culture lines. Long, solid line " I " represents a group of 24 isolation lines from which all endomictics were eliminated. Long, solid line " II " represents a group of 24 isolation lines, interrupted periodically (at points marked e) for the induction of endomixis in small mass cultures. The short, solid, horizontal lines represent groups of isolation lines carried for 15 generations. The length of the short lines corresponds to the average number of days that the 15 generations lived in all the groups. The vertical lines (broken) connect the source groups (represented by long, solid, horizontal lines) with their derived groups (represented by short, solid, horizontal lines). In each case, the derived group began with animals in endomixis taken from mass cultures of animals from the source group. A, B, C, D, and E are the designations given to the derived groups from both I and II. The four derived pairs of groups of isolation lines between which comparisons are made art.- groups In, Iln; Ic, lie; ID, I1|>; IK, UK. Group II was derived from Group I.\ by continuing, without endomixis, 24 of the lines of this group. e stands for climax of endomixis. This symbol is put in on the day on which endomictics were isolated from the mass cultures. T indicates the time of termination of the experiment. Time in days, from the initial endomixis in Group I, is shown on the horizontal axis at the top of the diagram. The number of days between successive endomixes is shown at the bottom of the diagram. endomixis and the resulting mortality in them was determined. In these groups (IA, In, Ic, ID, and IK, Fig. 1), therefore, tin- mortality after interendomictic intervals of 31, 56, SI, 102, and 125 days, re- spectively, was ascertained. At the same time that each of these MORTALITY AND ENDOMIXIS IN PARAMECIUM 237 groups (except the first) was being studied, there was examined a control group of lines (hat went through endomixis at the same time, but had instead an interendomictic interval of normal extent (21 to 25 days). The four control groups (I IB, He, I In, and HE) were ob- tained by inducing four successive endomixes at the proper time in descendants of the group I.\. Comparisons of mortality were thus made between the following pairs of groups: IB (previous interval 56 days) with HB (previous interval 25 days); Ic (previous interval 81 days) with He (interval 25 days); ID (interval 102 days) with IID (interval 21 days); and IE (interval 125 days) with II E (interval 23 days). In this experiment, the animals employed were all descended vegetatively from one which was isolated from a Johns Hopkins stock TABLE I Number of fissions from climax of endomixis until death in sixty non-viable lines among one hundred and twenty exendomictic lines followed through until death or the next induction of endomixis. (In these lines, the maximum interendomictic interval was 71 fissions ) Number of Fissions until Death Frequency 1 2 2 3 3 4 4 6 5 13 6 7 7 7 8 6 9 6 10 1 14 1 27 . . . . 1 29 1 30 1 37. 1 mass culture of Woodruff's long-lived Yale race. The organisms were cultivated throughout according to the methods described by Sonne- born (-1936). Samples of all lines, except those not carried through till the next endomixis, were stained daily to determine the nuclear con- dition. Relationship of the Interendomictic Interval to the Percentage of Mortality after Endomixis The percentage mortality was computed as follows: In each group, each endomictic animal was cultivated as a single daily isolation line until the fifteenth fission after the climax of endomixis. The per- BERNICE FRANCES PIERSOX centage of lines that died during this period was the percentage mor- tality for the group. The period of 15 fissions was decided upon as a conservative standard in view of the facts that the process of endomixis itself lasts not more than 9 or 10 fissions and that experience of previous investigators (e.g. Caldwell, 1933) as well as our own showed that deaths rarely occur in the interval from the tenth generation after the climax of one endomixis until the following endomixis. Table I gives i In- frequency of death at various stages of the interendomictic interval in the material here investigated. As appears from the table, deaths occurred but rarely after the tenth fission. The results of the experiment are summarized in Tables II to IV. Table II gives a general view of the relation of mortality after endo- mixis to the length of the previous interendomictic interval. As appears in the table, the percentage mortality increases as the previous interendomictic interval increases. It rose from 32.2 per cent after TABLE II Relation of mortality after endomixis to length of previous interendomictic interval. Previous Interendoiuii i ic Interval in Days 21 Percentage of Mortality after Endomixis 32.2 Number of Endomictic Lines Observed 90 23 62.5 48 25 .... 59.3 . . . 91 25 57.0 79 31 66 6 .... 48 56 . . . 71.4 . 49 81 79.2 53 102 89 3 . . . . 84 125. .100.0. . .29 an interval of 21 days to 100 per cent after an interval of 125 days. The first five intervals, 21, 23, 25, 25, and 31 days are normal inter- endomictic intervals for Woodruff's stock of Paramecium anrelia. The last four intervals of 56, 81, 102, and 125 days are abnormally long intervals for this stock. It will be noted that even in normal inter- endomictic intervals the percentage of mortality was high. It will be observed, from an examination of Table II, that there may be great differences in mortality after endomixis even when the inter- vals between endomixes are practically the same. Such a difference appears between the percentages of mortality after the 21- and 23-day intervals when the percentages of mortality were 32.2 per cent and 62.5 per cent, respectively. Obviously there are factors oilier than inter- endomictic interval involved in the production of mortality. Knviron- mental factors probably play an important role here. Therefore, it MORTALITY AND ENDOMIXIS IN PARAMECIUM 239 was essential, for the purposes of this study, to make simultaneous comparisons of groups kept under identical cultural conditions. The results of such simultaneous comparisons are given in Table III. There the percentages of mortality from endomixis after four abnor- mally long interendomictic intervals are compared with the percentages of mortality from concurrent endomixes after four normal intervals. Thus, when the interendomictic interval was 56 days, there was 71.4 per cent mortality, as compared with 59.3 per cent in the concurrent group with a normal interval of 25 days. When the interval was increased to 81 days, the mortality was between 79.2 and 84.0 per cent (the exact figure depending upon how many of 20 animals that died without dividing — and hence without a determination of whether they had been in endomixis — were in endomixis), as compared with TABLE III Relation between length of previous interendomictic interval and percentage of mortality after endomixis. Groups with Normal Groups with Long Interendomictic Intervals Interendomictic Intervals Previous In- Percentage Percentage Previous In- Name of terendomictic Mortality Mortality terendomictic Name of Group Interval in after after Interval in Group Days Endomixis Endomixis Days HB 25 59.3 71.4 56 IB He 25 57.0 79.2 81 Ic IID 21 32.2 89.3 102 ID HE 23 62.5 100.0 125 IE 57.0 per cent in the concurrent group with a normal interval of 25 days. When the interval was still greater, 102 days, the mortality likewise increased to 89.3 per cent as compared with 32.2 per cent in the controls with a normal interval of 21 days. Finally, when the interval reached 125 days, the mortality was 100 per cent, as compared with 62.5 percent in the controls with a normal interval of 23 days. The effects of still greater intervals could not be studied because efforts to induce endo- mixis at intervals of 148, 158, and 161 days all failed; all but a few of the animals in the induction cultures died within 48 hours. The results of the experiment, as summarized in Tables II and III, show clearly that as the time between two successive endomixes in- creases, the percentage of lines that are unable to survive the second endomixis also increases until eventually no animals can survive. 240 BERXICE FRANCES PIERSOX Relationship of the Interendomictic Interval to the Number of Generations between Endomixis and Death The relationship between the length of the previous interendomictic interval and the number of generations which the non-viable exendo- mictic lines lived after the climax of endomixis is shown by the data in Table IV. When the previous interval was of normal extent, i.e. about 25 days, as was the case in the control groups, the mean number of generations which the lines lived after the climax of endomixis re- mained fairly constant, ranging only from 5.0 to 5.9 generations. TABLE IV Length of life in number of generations from climax of endomixis until death, in the non-viable exendomictics, in relation to prior interendomictic interval. (The number dying without fission wa^ not determined in experiments 1 and 2.) Ex- peri- ment i troup Inter- endo- mictic Inter- val in Days Number of Generation-; from Climax of Endomixis until Death To- tal Non- viable Exen- domic- tics Mean 0 1 2 3 4 5 6 7 8 0 10 11 12 13 14 15 1 Us 25 ? 1 1 2 1 s 5 15 7 6 7 8 8 2 4 1 54 5.9 IB 56 ? 7 '> 5 3 1 35 4.8 2 He 25 ? 1 1 5 6 13 4 2 1 1 2 1 1 45 5.8 Ic 81 ? 7 3 3 6 9 8 4 1 1 42 4.4 3 HD 21 3 1 1 3 3 3 7 8 2 4 1 1 2 1 29 5.0 ID 102 6 1 5 3 L3 18 l > 1 4 2 2 75 4.4 4 HE 23 4 4 3 5 1 2 2 1 1 1 1 30 5.5 IE 125 21 1 2 2 2 1 29 0.9 Among the experimental groups ihr mean number of generations decreased from 4.8 general inns when the interval was 56 days to 0.(J generation when the interval was 125 days. In contrast to the steady increase in mortality rate with increasing interendomictic intervals, the mean survival period shows no such steady change. When the interendomictic interval was twice as great as normal the survival period decreased but little. Kven when the interendomictic interval was 3 or 4 times as great as normal, the sur- vival period was decreased, on the average, by only about one fission. MORTALITY AND ENDOMIXIS IN PAKAM 1'XTUM 241 Differences of this magnitude are of doubtful significance because the method of determining when the climax of endomixis occurred involved an uncertainty. No direct observation could be made on the endo- mictic animals removed from the induction cultures to begin the experimental and control groups; but the next day one or more products of their fission were stained and the nuclear condition recorded. On the basis of the number of fragments of the old macro- nucleus, their size and the intensity of stain, and the size, form, and intensity of stain of the new macronucleus or its anlage, an estimate could be made of how many fissions had occurred since the climax of endomixis; but such an estimate may often be in error by one or two fissions. For this reason, little significance is attached to the slight differences among the preceding groups in the mean number of genera- tions that the non-viable lines survived. In the group with the longest interendomictic interval, however, the difference is so great as to be unquestionably significant. After an interendomictic interval of 125 days, the non-viable exendomictics went through only one-sixth as many fissions as the non-viable exendo- mictics with normal prior interendomictic intervals. Indeed, 72.4 per cent of them failed to divide at all after the climax of endomixis. On the third day without fission they were all stained and found to contain fragments of the old macronucleus, but no new anlage or macronucleus. It might be suggested that about three-fourths of the parent lines had lost their micronuclei before endomixis was induced. In such animals only the destructive phases of endomixis can take place as there is no reserve micronucleus from which a new macronucleus can be formed. Loss of the micronucleus after long omission of endomixis has in fact been observed by others in this laboratory (unpublished). Discussion 1. In view of the complex and superficially paradoxical relations between endomixis and mortality, it has sometimes been held that endomixis is neither a definite phenomenon nor a normal one, but that it is a pathological response of the organism to adverse conditions, the type of response and its consequences varying with the degree of unfavorableness of the environment. In the present work, effects due to differences of environment were avoided by systematically exchang- ing culture medium between the various groups compared and by restricting comparisons to groups examined at the same time under the same cultural conditions. Thus, the differences in mortality after endomixis were not consequences of environmental differences. Nor were they due to the cumulative action of unfavorable conditions, for : 242 BERXICE FRANCES PIERSON each case the groups compared had been subjected equally long to the same conditions. The differences in mortality after endomixis must therefore have been due to the difference in the prior interendomictic interval. The fact that mortality is increased after longer intervals shows that the frequent occurrence of endomixis is an advantage to the stock. In view of this, the interpretation of endomixis as pathological is untenable. Similar conclusions were reached by Kimball (1937) from a study of the precise ratios in which sex segregates after endo- mixis in this species. 2. The vit-w that Protozoa are potentially immortal and that natu- ral death does not exist among them became popular during the last quarter of the nineteenth century. It has long been known, however, that while a race as a whole may be potentially immortal, certain mem- bers of the race are doomed, from internal causes, to die. Thus, Jennings (1913) and others showed that conjugation often resulted in unavoidable death. Erdmann and Woodruff (1916), Jennings, Raffel, Lynch, and Sonneborn (1932), Raffel (1932),,Caldwell (1933), Sonne- born and Lynch (1937), and others have shown that endomixis likewise often results in death. \Yoodruft (1917), Sonneborn (1935), and Jennings and Sonneborn (1936) have shown that long omission of endo- mixis ultimately results in death. To these intrinsic causes of death in Paramecium the present paper adds another: The mortality at endo- mixis is directly proportional to the preceding interval without endomixis. SUMMARY 1. Using an interendomictic interval of 20 to 30 days as a standard, it was shown that intervals approximately two, three, four, and five times this long resulted in progressive increases in mortality after endomixis until 100 per cent mortality occurred. 2. At intervals greater than this, the animals died before endomixis could be induced under conditions favorable for its induction. 3. The mean number of generations which non-viable lines sur- vived the climax of endomixis was 5.0 to 5.9 generations when the pre- vious interval was of normal extent. After an interval of 125 days, survival dropped greatly to a mean of 0.9 generation. 4. The results are shown to disagree with current interpretations of mortality at endomixis based on the concept of endomixis as a patho- logical process. 5. The results show that unusually long interendomictic intervals are, like other previously known conditions, a cause of " natural death " in Protozoa. MORTALITY AND ENDOMIXIS IN PARAMECIUM 243 LITERATURE CITED CALDWKLL, L., 1933. The production of inherited diversities at enclomixis in Paramecium aurelia. Jour. Exper. Zool., 66: 371. ERDMANN, R., AND L. L. WOODRUFF, 1916. The periodic reorganization process in Parannpcium caudatum. Jour. Expcr. Zool., 20: 59. JENNINGS, H. S., 1913. The effect of conjugation in Paramecium. Jour. Exper. Zool., 14: 279. JENNINGS, H. S., DANIEL RAFFEL, RUTH STOCKING LYNCH, AND T. M. SONNEBORN, 1932. The diverse biotypes produced by conjugation within a clone of Paramecium aurelia. Jour. Exper. Zool., 62: 363. JENNINGS, H. S., AND T. M. SONNEBORN, 1936. Relation of endomixis to vitality in Paramecium aurelia. Comptes rendus du XIIC Congres International de Zoologie (Lisbonne 1935), pp. 416-420. KIMBALL, R. F., 1937. The inheritance of sex at endomixis in Paramecium aurelia. Proc. Nat. Acad. Sri., 23: 469. RAFFEL, DANIEL, 1932. The occurrence of gene mutations in Paramecium aurelia. Jour. Exper. Zool., 63: 371. SONNEBORN, T. M., 1935. The relation of endomixis to vitality in Paramecium aurelia. Anal. Rcc., 64: (Supplement No. 1) 103. SONNEBORN, T. M., 1936. Factors determining conjugation in Paramecium aurelia, I. The cyclical factor: the recency of nuclear reorganization. Genetics, 21: 503. SONNEBORX, T. M., 1937. Induction of endomixis in Paramecium aurelia. Biol. Bull., 72: 196. SONNEBORN, T. M., 1938. The delayed occurrence and total omission of endomixis in selected lines of Paramecium aurelia. Biol. Bull., 74: 76. SONNEBORN, T. M., AND R. S. LYNCH, 1937. Factors determining conjugation in Paramecium aurelia. III. A genetic factor: the origin at endomixis of genetic diversities. Genetics, 22: 284. WOODRUFF, L. L., 1917. The influence of general environmental conditions on the periodicity of endomixis in Paramecium aurelia. Biol. Bull., 33: 437. THE EFFP:CT OF SHORTFR THAN NORMAL IXTER- ENDOMICTIC INTERVALS ON MORTALITY AFTER ENDOMIXIS IX PARAMECIUM AURELIA JULIUS GELBER1 (From the Department of Zoology, The Johns Hopkins University) In the preceding paper, Pierson (1938) has shown for Parameciiim aurclia thai mortality after endomixis is directly proportional to the length of the interendomictic interval, when the latter is longer than normal. The present paper examines the same question when the interendomictic intervals are shorter than normal. The same race of Paramecium aurelia is examined in both studies. All individuals in the present study were descended without endomixis from one taken from a stock mass culture of this race on February 22, 1936. The methods of culture and of induction of endomixis were similar to those employed by Pierson. The experiment was performed as indicated in Fig. 1. From a single endomictic individual, 24 daily isolation lines of cultivation were followed for 28 days (Group I, Fig. 1). During this time, samples of each line stained daily showed that no endomixis occurred. On the eighth and sixteenth days after the initial endomixis, the surplus animals from the isolation lines were collected in a mass culture and placed at 31° C. In each case, a high percentage of individuals went into endomixis after 48 hours in these conditions. From each of these, a group of endomictic individuals was isolated and cultivated (Groups II and III, Fig. 1). On the twenty-sixth day after the initial endo- mi\K iliree mass cultures were set up, each consisting of the surplus animals from one of the three groups under cultivation. These cul- turo were placed at 31° C. and all contained numerous endomictic individuals two days later. From each of the three cultures a group of endomictic individuals was isolated and cultivated. These three groups of animals \\ere thus all in endomixis at the same time, but they dittncd in the interval since the last preceding endomixis: in one, the interval was 10 days; in the second, 18 days; and in the third, 28 days. The three groups will be designated the 10, 18, and 28-day groups, in reference to their prior interendomictic intervals. The 10-day 1 The author is grateful to Dr. T. M. Sonneborn who suggested the problem and offered his advice throughout the experiment, and to Bernice Pierson and Nathaniel Finkelstein for their kind assistance. 244 MORTALITY AND ENDOMIXIS IN PARAMECIUM 245 group was begun with 192 individuals from the culture in which endo- mixis had been induced; but only 51 of these were in endomixis, as determined by staining products of their fissions on the next day. From each of these 51 individuals, a single daily isolation line was cultivated until it died or until 15 successive fissions had taken place. Of the 51 lines, 48 lived through the 15 fissions and three died after 5 to 12 fissions, giving a mortality rate of 5.9 per cent. The 18-day group was begun with 144 individuals from the induc- tion culture; but only 94 of these were in endomixis, as subsequently determined. Of these 94 endomictic individuals, 74 lived through the following 15 fissions, and 20 died after 1 to 7 fissions, giving a mortality rate of 21.3 per cent. Time in Days Percentage Mortality 41.8 5.9 21.3 i v i w ** Y FIG. 1. Plan of Experiment The three horizontal lines, I, II, and III, represent groups of isolation culture lines carried without endomixis for 28, 18, and 10 days, respectively. The vertical lines connect the source group (horizontal line I) with the two derived groups (hori- zontal lines II and III). In each case the derived group began with animals in endo- mixis taken from mass culture of animals from the source group. e stands for endomixis. The upper line shows the time in days since the initial endomixis in Group I. The percentage mortality following endomixis in each group is shown to the right of the line representing that group. The 28-day group was begun with 144 individuals from the induc- tion culture. Of these, 110 were shown by subsequent staining of their descendants to have been in endomixis. These 110 endomictic individuals were cultivated in the same way as were those of the other two groups. In this group, 64 lived through the 15 fission period of observation and 46 died after 0 to 13 fissions, giving a mortality rate of 41.8 per cent. Thus, the group with a normal interendomictic interval of 28 days suffered a mortality rate of 41.8 per cent after endomixis, as compared with 21.3 per cent and 5.9 per cent mortality following abnormally short interendomictic intervals of 18 and 10 days, respectively. The results therefore extend those of Pierson to include abnormally 0 10 18 28 I ...,,.... 1 ....... . «_l (28 day [Group (lO day (Group fl8 day \G r o u p • n m. 0 1 C 246 JULIUS GELBER short as well as abnormally long interendomictic intervals. Through- out the entire range of intervals investigated, the mortality after endomixis is directly proportional to the extent of the preceding inter- endomictic interval. LITERATURE CITED PIERSON, BERNICE F., 1938. The relation of mortality after endomixis to the prior interendomictic interval in Paramecium aurelia. Biol. Bull., 74: 235. DIPLOIDS FROM UNFERTILIZED EGGS IN HABROBRACON KATHRVN G. SPEICHER AND H. R. SPEICHER1 (From the Department of Zoology, University of Maine, and the Marine Biological Laboratory, Woods Hole, Mass.) FEMALES The occurrence of impaternate females (females from unfertilized eggs) in the parasitic wasp Ilabrobracon juglandis (Ashmead) has previously been reported (Speicher, 1934). Such females occur sporadically from various virgins, and regularly constitute about 1 per cent of the F2 population from Fi virgins produced by outcrossing females from tapering or reverted tapering stocks. The hypothesis was offered that these diploid impaternate females might be produced by the failure of the second maturation division in the unfertilized egg and, carrying two chromatids from a single tetrad, would provide material for studying the mechanics of crossing-over. When FI virgins were heterozygous for recessive factors, ¥2 im- paternate females occurred in the ratio of one homozygous dominant, to two heterozygotes, to one homozygous recessive for each locus involved. This 1:2:1 ratio was at variance with results obtained in other organisms where more than one strand may be recovered from a single tetrad, notably Drosophila (Anderson, 1925) and Neurospora (Lindegren, 1933). Here the first maturation division is reductional at the spindle fiber and in dyads from that division the homozygosis of any locus depends upon the amount of crossing-over between it and the fiber, and hence is a function of its location along the chromosome. In Ilabrobracon, except for the locus of the sex-linked factor fused (Whiting and Speicher, 1935), the amount of homozygosis was con- sistently 50 per cent for all loci tested, even including those of two recessives known to be linked and separated by a distance of ten units. It was realized from the beginning of the work that other hypoth- eses could be advanced to explain the formation of impaternate females. And it \vas apparent that the 1:2:1 ratio obtained would be expected for all loci if the two homologous chromosome strands of an impaternate female came from two tetrads, independent and com- 1 The authors are indebted to Professor P. W. Whiting of the University of Pennsylvania for the use of special microscopic equipment which was supplied in part by a grant to him from the Elizabeth Thompson Science Fund. 247 248 KATIIKVN G. SPEICHER AND B. R. SPEICHER pletely reduced, rather than from a single tetrad which underwent only partial reduction. As an investigation of oogenesis in Ilabrobracon had already been begun (Speicher, 1936), it seemed best to suspend further genetic studies until the method of formation of impaternate females could be determined cytologically. Those findings are here reported for the first time. Since impaternate females had been shown to be genetically diploid, whereas their sibs are haploid, a cytological analysis of their formation seemed practical. According to past work on other forms at least three basic hypotheses could be considered. First, one of the two maturation divisions of an unfertilized egg could be suppressed, as was observed by Silvestri (1908) in the wasp Prospalta. This would leave a cleavage nucleus containing the diploid number of chromatids, which would restore the diploid number of chromosomes if the homo- logues separated. Second, fusion of two reduced egg nuclei present TABLE I Progeny from No. 25/reta virgin females. Fi virgin 9 9 i 2* 3 4 5 6 7* 8 9 10 11 Sons . ... 83 0 01 26 OS Q7 s 67 9Q 11 * 6* Daughters 1 2 0 0 5 0 0 0 o 0 7 Eggs collected 4Q 3 50 S6 46 61 11 7Q 48 S6 S] * Died before completion of experiment. in a binucleate egg also would restore the diploid condition. Third, the egg might originate as a tetraploid, undergo reduction and thus become diploid. A fourth hypothesis, the doubling of chromosomes in the haploid egg during cleavage, is eliminated because tests have shown that impaternate females may be genetically heterozygous. The first two theories were tested together. Over 300 eggs were collected from virgin females produced by crossing reverted tapering and stock 25. They were fixed at first cleavage prophase and stained by the Feulgen whole-mount method. The same females produced collectively 15 impaternate daughters among 724 sons, over 2 per cent of the total. Among the eggs studied approximately the same pro- portion would be expected to show cytological differences if either theory were correct. Suppression of a maturation division would re- sult in a decrease in the normal number of polar nuclei formed at the egg margin. A binucleate egg would be expected to show two groups DIPLO1DS, UNKKKTIUZED EGGS OK HABROBRACON 249 of polar nuclei, totaling twice the normal number. An examination of all eggs revealed none showing either of these two differences. Treatment of the third hypothesis requires chromosome counts during maturation; accordingly it was necessary to obtain a new lot of eggs fixed at an earlier stage than the above. It had been noticed previously that some virgin females produced impaternate daughters in small groups while others in the same experiment produced no daughters whatever. Inclusion in the data of the offspring from the latter virgins lowers the percentage of impaternate females among total offspring. It was therefore possible to raise the expected percentage of exceptional eggs by selecting eggs only from virgins known to be producing impaternate daughters. This was accom- plished as follows. Eleven females from a cross of reverted tapering by stock 25 were placed with host caterpillars. Eggs laid upon the caterpillars during seven consecutive days were fixed at first anaphase and temporarily stored in alcohol. Eggs laid at night over the same period were allowed to develop, in order to indicate which of the eleven females were thelytokous. Results are shown in Table I. ;• Kig. 1. Fig. 2. FIG. 1. First anaphase of normal egg. X 3,000. FIG. 2. First anaphase of tetraploid egg. X 3,000. Seven females, producing a total of 428 sons, had no daughters, while the remaining four produced 241 sons and 15 daughters. The 149 eggs collected from the latter four females were then stained by the Feulgen whole-mount technique. One hundred and ten of these were in condition to study; the remainder were either lost in handling or were collapsed. Ninety-eight eggs were unquestionably diploid, seven were unquestionably tetraploid and five more were questionably tetraploid. The clear cases of tetraploidy showed twenty chromo- somes, presumably bivalents although individual chromatids have never been observed in Habrobracon oogenesis due probably to their small size, moving to one or to each pole. Normal diploid eggs show only ten chromosomes going to each pole, Figs. 1 and 2. These cytological data, and the fact that impaternate females come in groups from certain mothers suggest the probability that production of tetraploid eggs, as developed from patches of tetraploid 250 KATHRYX G. S PHI CHICK AND B. R. S PIC 1C HICK ovarian tissue, is responsible for the appearance of cliploid impaternate females among haploid brothers. M.M.KS 1 >iploid males of biparenial origin have been reported repeatedly in Ilabrobracon. Since diploid females are produced by virgins as a result of tetraploidy in egg cells, it seems plausible that diploid im- paternate males may be produced in the same way. According to Whiting's scheme of sex-determination in Ilabrobracon (AYhiting, 1933) those eggs which were diploid after reduction and contained sex chromosomes A" and Y would produce impaternate females, while those which contained chromosomes XX or YY would produce diploid impaternate males. If distribution of chromatids is random the number of diploid females and males produced should be equal. TABLE II I >.itu from tests for diploid impaternate males. K, Q Q I- li.tpl'nd c?d\ I "•j impaternate 9 ? .1 or (i .1,1 Aa aa o'/od 62" 6 10 5 Le/le 6070 1 7 6 St/st 4408 4 2 0 Three experiments were set up in an effort to produce diploid impaternate males that could be distinguished genetically from their haploid brothers. Since the highest percentages of impaternate FZ diploids had previously resulted from outcrosses of tapering and reverted tapering females, tapering was again used as the maternal stock in one experiment. In the other two it was necessary to intro- duce recessive factors into the cross through the females. In order to insure the occurrence of thelytoky in these cases the recessives were repeatedly bred up to reverted tapering, and a stock related to it but having the desired genetic constitution was then derived and supplied the females for the parental crosses. In each experiment FI virgins were produced carrying both of two allels that give rise to a distinct phenol \ pe when they are heterozygous. Results are given in Table II where the first column shows the pairs of allels u-e