THE BIOLOGY OF AND OF SOME OTHER COELENTERATES; EDITED BY: HOWARD M. LENHOFF V. FARNSWORTH LOOMIS -gSu^ THE BIOLOGY OF HYDRA AND OF SOME OTHER COELENTERATES: 1961 EDITED BY: HOWARD M. LENHOFF W. FARNSWORTH LOOMIS THE BIOLOGY OF HYDRA AND OF SOME OTHER COELENTERATES : 1961 Edited by Howard M. Lenhoff and W. Farnsworth Loomis UNIVERSITY OF MIAMI PRESS CORAL GABLES, FLORIDA Copyright 1961 l)y The Loomis Institute tor Seientifie Keseareli, Iiie. Library of Congress Card Number 61-18157 Printed in the United States of America bj- Rose Printing Company Tallahassee, Miami, Jacksonville, St. Augustine Forevv^ord "Further, I discovered a little animal whose body was at times long, at times drawn up short, and to the middle of whose body .... a still lesser animalcule of the same make seemed to be fixed fast by its hinder end .... [At that time the little animalcule] had only four very short little horns, yet after the lapse of sixteen hours I saw that its body and its horns had increased in bigness, and four hours later still I saw it had forsaken its mother.''^ In this remarkable letter written on Christmas Day, 1702, An- tony van Leeuwenhoek amazed members of the Royal Society by annonncing a discovery of dual significance. While reporting the initial description of the organism which we now call hydra," he also described the first instance of asexual reproduction ever ob- served in animals. Thus, from their very discovery hydra have served to reveal new biological phenomena. More startling findings with hydra followed when, in 1744, Abraham Trembley published in his superb Memoires an exposition of: the first controlled experiments on regeneration; the first suc- cessful animal grafting experiments; the first investigations of photo- taxis in lower invertebrates; the first vital staining of tissues; and thorough proof of asexual reproduction by budding. Two centuries have passed since Trembley made these revolutionary discoveries, an interim during which research on hydra was sporadic, and hydra were relegated to a subsidiary role in classroom instruction. In the 1 Letter 149, December 25, 1702. Quoted in Antony van Leeuwenhoek and his "Little Animals" by Clifford Dobell, Dover Pubis., N. Y., 1960, pp. 280-281. - In this volume we have adopted, whenever possible, the following usages for purposes of unifonnity and clarity: (a) Hydra, when referring to one or more speci- mens of this genus if the species has already been clearly indicated; (b) hydra, when referring to one or more specimens of the Hydridae in general, and when the species is not indicated; (c) hydras, when referring to a number of genera of the Hydridae. last decade, however, a renaissance in the use of hydra as a labora- tory animal has been in progress. The return of hydra to their original status as laboratory ani- mals is marked by the pul^lication of Tlie Biolog,y of Hydra : 1961. This is the first book since Trembley's Memoires devoted to original research reports dealing for the most part with hydra. The present volume is a record of a symposium on the Physiology and Ultra- structure of Hydra and of some other Coelenterates held March 29- 31, 1961, at the Fairchild Tropical Gardens, Coral Gables, Florida. In this symposium. North American workers representing many different fields of biology described their current work. They started with a discussion of the fine structure of hydra cells and mesoglea. Following a session devoted to the development, chemistry, and function of nematocysts, they considered the sul^jects of feeding and nutrition. Next, research on tissue culture, symbiosis, and cal- cification were discussed. A session concerning the various forces responsible for the patterns of colonial hydroids led, in turn, to a consideration of cellular differentiation and then of aging in both mortal and immortal coelenterate types. Appropriately, attention turned at last to regeneration and to new birth as seen in budding. In organizing this symposium, the editors desired to bring about an integration of knowledge from a large variety of disciplines. Electron microscopists, naturalists, biochemists, and developmental biologists ordinarily do not read or publish in the same journals. The aim of the symposium was, therefore, to effect an interdisci- plinary synthesis which might otherwise take years by normal chan- nels. Accordingly, the discussions that followed each talk are included because they point out some of the many unsolved prob- lems and therefore should prove of value in stimulating further investigations. Much of the work presented at this symposium is in an early stage. At times we have thought that perhaps these results are too preliminary and should only be compiled after more data have been accumulated. The situation is analogous to constructing a new building. At times we might feel that all such work should proceed behind walls marked "Work in Progress. No Admittance." At other times we are intrigued with the very smell of sawdust and of wet paint. It is in this latter spirit that the volume was compiled, for these efforts, given time, may well show that hydra are particularly favorable material for the investigation of cellular and intercellular problems. History at least supports this view, because it was in hydroid material that asexual reproduction and regeneration were first discovered over two hundred years ago. '7 cut off the heads of the one that had seven, and after a few days I saw in it a prodigy scarcehj inferior to the fahidous Hydre of Lernaea. It acquired seven new heads ... .But here is something more than the legend dared to invent: the seven heads that I cut off from this Hydre, after being fed, became perfect animals. . . ."^ W. Farnsworth Loomis, M.D. Howard M. Lenhoff, Ph.D. Greenwich, Connecticut Miami, Florida September 21. 1961 ^ Abraham Trembley, 1744. Memoircs, pour servir « Vhistoire d'un genre de polypes d'eau douce, a bras en forme de comes. Leide (Verbeek), p. 246. Quoted in Abraham Trembley of Geneva, John R. Baker, Arnold & Co., London, 1952, p. 34. (A complete translation, to be published, S. G. Lenhoff and H. M. Lenhoff, University of Miami Press. ) - 'S :5 J ^ ^ I ^ ^ r d J ffi ;i; £ d^ < "-n ffi J ■ 5] . bC a; ^ J W g ^^ U ^ ^ ^ m . 2 &h' • ^ K ^^3 •„ bO C H S r>^ ^ §^Q -^ _Q F-ii - -^ ^ t: ^ ^ S^-^'^ C/2 ^1°^ >. . J S W bio . &: Pd 2 U > ^ W a; ^ « ^ .S hJ Jg"^ ■^ ^- ^ I w -5? = -^ I O U ^ -^ ^ o t,- ^^ 3 ^ ^^d| . C ffi W • ^ ^ a; 2 -"^ >:s ti. d i^ w -^ c c/: Participants Barbara O. Alving Eric Alving Reza Bashey Robert J. Boucek B. Bourne John Bovaird Patricia Broberg Robert R. Bryden Stanley Burg Allison Burnett F. Gray Butcher Edward L. Chambers George B. Chapman David L. Claybrook Sears Crowell Robert E. Eakin Don W. Favv^cett Hector Fernandez Bernard Fritzie Chandler Fulton Geraldine F. Gauthier Lauren C. Gilman Thomas Goreau C. T. Grabom^ski Cadet Hand Arthur Hess Ray M. Iverson Edward Kline A. R. Krall Charles E. Lane Edward Larson W. Henry Leigh Howard M. Lenhoff Yu-YiNG Fu Li Alfred L. Loomis W. Farnsworth Loomis Philip Lunger Charles F. Lytle G. O. Mackie A. G. Matoltsy J. Marsh Edgar J. Martin N. Mason Leonard Muscatine Nancy L. Noble Edward E. Palincsar Helen D. Park L. M. Passano John H. Phillips Earl R. Rich Gordon C. Ring D. M. Ross K. Savard Harriet Schapiro David B. Slautterback Daryl Stafford Bernard L. Strehler W. J. VAN Wagtendonk Stephen A. Wainwright Peter Wangersky Eleanor D. Wangersky John H. Welsh Richard L. Wood Edmund Zaharowic:z Acknowledgements It is a pleasure to aeknowledge all those who helped in this venture: Mr. Nixon Smiley, Director, and Mr. S. Kiem, Superin- tendent, of the Fairchild Tropical Gardens, for their .special care in providing facilities conducive to di.scussion; Miss H. Schapiro, Mr. B. Fritzie, Mr. H. Fernandez, Mr. D. Stafford, and Mr. E. Zaharowicz, graduate students of the University of Miami, for their help throughout the symposium; Mr. R. Conklin of the Miami Sea- quarium, for being a generous host to the participants and their families; Mrs. N. Jaffe, Mrs. E. Hirshhorn, Mr. J. Bovaird, and Mr. H. Reasor, who, in addition to their regular responsibilities, con- tributed in the preparation of the recorded discus.sions and of the index of this volume; Dr. B. Strehler, for his most helpful sugges- tions in expediting publication; Drs. E. Muscatine and E. L. Cham- bers, for their many suggestions; and Dr. R. }. Boucek of the Howard Hughes Medical Institute, for his encouragement and for providing facilities for arranging the symposium. Finally, we wish to express our deep thanks to every participant of the symposium for his cheerful cooperation in responding to our seemingly endless requests for corrected manuscripts, di.scussions, and galleys. Contents PAGE The Fine Structltre of Cells in Hydra ... 1 Arthur Hess The Fine Structure of Intercellular and Mesogleal Attachments of Epithelial Cells in Hydra , . 51 Richard L. Wood Is there a Nervous System in Hydra? .... 69 Floor Discussion Nematocyst Development ...... 77 David B. Slautterback The Fine Structure of the Stenoteles of Hydra . 131 George B. Chapman Chemistry of Nematocyst Capsule and Toxin of Hydra littoralis .......... 153 Edward S. Kline Physalia Nematocysts and Their Toxin . . . 169 Charles E. Lane Compounds of Pharmacological Interest in Coelen- terates .......... 179 John H. Welsh Page Present State of Nematocyst Research: Types, Struc- ture, AND Function ........ 187 Cadet Hand Activation of the Feeding Reflex in Hydra Uftoralis . 203 Howard M. Lenhoff The Nutrition of Hydra ...... 233 David L. Claybrook Isolation and Maintenance in Tissue Culture of Coe- LENTERATE CeLL LiNES ....... 245 John H. Phillips Symbiosis in Marine and Fresh Water Coelenterates . 255 Leonard Muscatine On the Relation of Calcification to Primary Produc- tivity IN Reef Building Organisms ..... 269 T. F. Goreau The Development of Cordylophora .... 287 Chandler Fulton Developmental Problems in Campanidoria . . . 297 Sears Crowell Patterns of Budding in the Freshwater Hydroid Cras- pcdacusta .......... 317 Charles F. Lytle Page Feedback Factors Affecting Sexual Differentiation in Hydra littoralis ......... 337 W. F. LooMis Apparent Rhythmicity in Sexual Differentiation of Hydra littoralis 363 Helen D. Park Aging in Coelenterates ....... 373 Bernard L. Strehler Studies on Chemical Inhibition of Regeneration in Hydra 399 Robert E. Eakin A Study of Normal and Abnormal Regeneration of Hydra 413 Dorothy B. Spangenberg Growth Factors in the Tissues of Hydra . . . 425 Allison L. Burnett Nucleic Acid and Protein Changes in Budding Hydra littoralis 441 Yu-Ying Fu Li and Howard M. Lenhoff Index 449 The Fine Structure of Cells in Hydra Arthur Hess^ Department of Anatomy, Washington Univers^ity School of Medicine, St. Louis, Missouri. Hydra can be considered to have the following anatomical regions: tentacles, hypostome or mouth region, column or stomach, peduncle and basal disk. Sections of the colvmm will serve most frequently to introduce the general histology of hydra. Then variations of the different body regions will be presented. Hydra has in general two cellular layers, ectoderm or epidermis and endoderm or gastrodermis separated by a layer called mesoglea. The ectoderm is composed basically of epithelio-muscular cells and contains dispersed cnidoblasts or nematocyst-bearing cells and interstitial cells or undifferentiated cells. Gland cells occur in spe- cialized regions. The endoderm contains gland cells, digestive cells and interstitial cells in its generalized areas. Cnidoblasts occur only rarely in the endoderm. The mesoglea is acellular. Whole Hydra oligactis were fixed in an extended state in Dalton's fluid, a solution containing 1% osmium tetroxide, 1% potassium dichromate at a pH of 7.2 to 7.6 and 0.85% sodium chlo- ride, in an ice bath for 15-45 minutes. Sometimes, the Hydra was divided into its various body regions; at other times, the animals were treated as a whole. They were then dehydrated in alcohol and embedded in methacrylate or araldite. Some sections were stained with lead acetate or potassium permanganate. They were photo- graphed in the electron microscope. iThe author wishes to acknowledge the participation of Dr. A. I. Cohen and Mrs. Dorothy Sanderson in tliis study. The author's present address is Department of Physiology, University of Utah College of Medicine, Salt Lake City, Utah. THE BIOLOGY OF HYDRA : 1961 GENERAL HISTOLOGY THE ECTODERM Epithelio-muscular cells (Figs. 1, 3). Vacuolated cells are seen in the ectoderm. Their nuclei are large, of even granular texture, and contain prominent nucleoli. These cells have a few double mem- branes and many mitochondria in their cytoplasm. They frequently contain dense inclusions, which we have not as yet identified. Small vacuoles, in addition to the large ones, are present. Within these cells and accumulated at their base, closely packed bundles of fine fibrils arranged in parallel and running longitudinally with respect to the column axis are seen. The muscle system will be considered separately later. These cells, frequently but not always, are the surface cells of Hydra. The surface of Hydra is covered by a granular material resting on two membranes (Fig. 6). One membrane obviously belongs to the surface cells, usually epithelio-muscular, but can be, at times, the cnidoblast. The other membrane apparently does not belong to a cell. A short \arying distance separates the outer mem- brane of the surface cell and the membrane on which the granular material rests. Hydra, therefore, appears to be covered over most of its surface by this thin cuticular material. Interstitial cells (Figs. 4, 5). Groups of small, rounded cells occur in the ectoderm. They are numerous in some areas and absent from others. These appear to be interstitial cells. They are charac- terized by having a very finely granular particulate cytoplasm with no double membranes. Mitochondria and a Golgi apparatus are present. Their nuclei are evenh^ granular with one or more promi- nent nucleoli. The cells are frequently very intimately related to each other and at times, the limiting membranes between two adja- cent cells appear to be lacking and the cells appear to be syncytial (Fig. 5). Since these cells give rise to cnidoblasts, some interstitial cells can be seen with a few double membranes in their cytoplasm suggesting that they are beginning their differentiation. These cells can be seen at times adjacent to the muscle layer on the mesoglea. Cnidobkists (Figs. 3, 7-10). Cnidoblasts frequently occur in groups and can be found near the mesoglea or sometimes forming ARTHUR HESS 3 the surface cell of Hydra. These cells have mitochondria and a Golgi apparatus. However, it is the presence of the double mem- branes or endoplasmic reticulum which renders these cells distinc- tive. The cnidoblasts apparently are deri\ed from interstitial cells. The cnidoblasts bearing \ery immature nematocysts have a series of vesicles (Fig. 7). As the nematocyst matures, these vesicles in- crease in amount and extent and apparently coalesce until the sys- tem of double membranes within the cell becomes quite elaborate and striking ( Fig. 8 ) . The nematocyst increases in size and dis- places the nucleus. In cnidoblasts with well-developed nemato- cysts, the double membranes begin to decrease in amount ( Fig. 9 ) . In cnidocytes having what appear to be mature nematocysts, the double membrane system appears to have regressed and the cyto- plasm of the cells is again granular with only a few strands of double membrane remaining ( Fig. 10 ) . These cells are also apparently in syncytial relation to each other and frequently, the cell membranes between adjacent cnidoblasts can be seen to be lacking. Apparently the syncytium is no longer present after the nematocysts are mature and the cnidoblasts have completed their differentiation and are called cnidocytes. Each mature cnidocyte appears surrounded by a complete cell membrane in the tentacle, as will be shown later. THE MESOGLEA (Figs. 1, 15, 16, 19-21). The mesoglea presents a vary- ing appearance in electron micrographs. It may appear fibrous or granular. Some of this \ariability may be due to the state of con- traction or extension of the Hydra during fixation. No cells are present. Pieces of cytoplasm seen in the mesoglea can be seen to be connected to ectoderm or endoderm cells which are pushing into the mesoglea. These pieces of cytoplasm are surrounded by a cell membrane and thus strictly are outside the mesoglea. The mesoglea is apparently not surrounded by its own limiting membrane. It penetrates between the cells of the ectoderm and endoderm (see especially Figs. 15 and 19), and granules, similar to those seen in the mesoglea, can be found in extracellular spaces between ectodennal and endodermal cells (Figs. 15, 19, 23, 25). 4 THE BIOLOGY OF HYDRA : 1961 Thus, the constituent cells of Hydra can be considered as embedded in mesoglea and the mesoglea forms a supporting substance for the cells. THE ENDODERM Gland cells (Figs. 11, 12). The gland cell pours its secretion into the lumen to break down the food and make its products avail- able for digestion. Essentially only one kind of gland cell has been found. This cell contains a series of large interlacing vacuoles, which most frequently appear light, but sometimes dark. Toward the base of the cell, the vacuoles frequently are smaller than in the portion of cell near the lumen. The cell appears to be under- going a process of manufacture of the vacuoles starting toward the base. Thus, various vacuolar arrangements can be seen, but they are believed to be stages in the appearance of a single kind of gland cell. Between the vacuoles, some mitochondria and double mem- branes appear. Toward the base of the cell, the vacuoles are not present and the cytoplasm is filled with mitochondria and double membranes. It is probably here where the manufacture of the vacuolar contents, which will be secreted into the lumen, begins. The nucleus of the cell is toward its base. This cell apparently does not rest upon the mesoglea. The digestive cell (Figs. 13-16). The digestive cell absorbs the food products after action of the gland cell. The digestive cell also undergoes cyclical changes according to the feeding activities of Hydra and also contains various inclusions depending on the kind of food and time of feeding. The cell can appear columnar and rather well organized or can contain huge vacuoles. It has a light cytoplasm with mitochondria and a Golgi apparatus. The surface of the digestive cell usually has a series of small cytoplasmic projec- tions or villi extending into the lumen (Fig. 14). The digestiv^e cells contain the endodermal muscle filaments at their base and rest upon the mesoglea (Figs. 15, 16). Flagella (Fig. 18). Apparently both gland cell and digestive cell have flagella. It is difficult to determine exactly how many project from each cell. Two to four flagella are commonly seen. The flagella present the nine peripheral and two central longitudinal ARTHUR HESS 5 filaments characteristic of motile flagella in other animals. These flagella differ slightly from those of other organisms in that they possess a relatively thick membrane surrounding the filaments which frequently becomes separated from the filaments so that its rela- tion to the filaments does not appear as intimate as the relatively thin membrane enclosing flagella elsewhere. THE MUSCULAR SYSTEM (Figs. 1, 15, 16, 17, 20, 21). The ectodermal muscle layer runs essentially longitudinally, while the endodermal layer is predom- inantly transversely oriented. The muscle filaments contained as a cell organelle in the base of the epithelio-muscular and diges- tive cells run parallel to each other, appear to be essentially of one kind, present no cross striation, and hence can be considered as smooth muscle filaments. The muscle fibers run along the meso- glea. They appear to be anchored to the mesoglea by small cyto- plasmic extensions of the cells containing them (see especially Figs. 15 and 16 ) . These extensions are frequently more numerous and robust on the ectodermal side and sometimes muscle filaments extend into these cytoplasmic attachment roots. The ectodermal muscle filaments in the cytoplasmic extensions of the base of one epithelio-muscular cell are very intimately related to the muscle filaments of an adjacent epithelio-muscular cell. The extensions of the cells can dovetail with each other in finger-like extensions or can overlap each other. However, the filaments do not pass from one cell to another. The filaments sometimes appear to insert on the cell membrane and when this happens in adjacent cells, an apparent thickening of the adjacent cell membranes occurs and a desmosome-like effect is produced ( Fig. 17 ) . The digestive cells usually do not undergo such an intimate arrangement and adjacent digestive cells are related to each other by relatively smooth membranes. There are points along which the mesoglea appears very thin or interrupted and where the ectoderm and endodermal muscle filaments, or at least the membranes of the cells containing them, are practically in contiguity (Figs. 20, 21). Probably some very thin mesogleal substance intervenes between them since, as men- 6 THE BIOLOGY OF HYDRA : 1961 tioned above, all the eells are probably embedded in mesoglea. These points of contact between the muscle layers are fairly fre- quent and occur in all areas investigated. THE RELATIONS OF CELLS TO EACH OTHER The special relationships of muscle cells and the fact that all cells appear embedded in mesoglea have already been discussed. However, there are other peculiar relations of cells that should be mentioned. By no means are the limiting membranes of the cells smooth. At times, a button or snap fastener arrangement can be seen where one cell evaginates a piece of cytoplasm to rest in an indentation of an adjacent cell ( Fig. 2 ) . This causes the frequent appearance of circular areas of cytoplasm located between cells. In addition, terminal bars are seen between some cells lining the lumen (Fig. 14) and between other cells near the mesoglea (Fig. 19). NERVE CELLS AND FIBERS No cell was foimd which could be called a nerve cell. As explained above, the small circles located between cells, which may sometimes form clusters and appear like bundles of nerve fibers (Figs. 15, 18, 21, 26), probably result from the peculiar formations of the cell borders. There is the possibility that nerve tissue of Hydra may appear different in the electron micro- scope from that of other organisms, and we are thus unable to identify nerve cells or fibers in our electron micrographs. However, if the absence of nerve tissue in Hydra can be accepted, one may perhaps go further and wonder if, indeed. Hydra needs any nerves. The epithelio-muscular cells containing the muscle filaments are on the surface of the animal and there can act as receptor cells to lead the impulse to its muscle filament organelles. The impulse of one ectodermal muscle fiber could easily be transmitted to muscle filaments in adjacent epithelio-muscular cells. Endodermal muscle could conduct an impulse from one cell to another in a similar manner. Lastly, the interaction of ectodermal and endodermal muscle could well be achieved through the points where mesoglea AHTUL'R HESS is practically absent and the two muscle layers are essentially in contact. The ordinary slow moxement which Hydra performs could well be subserved bv muscle to muscle transmission. REGIONAL HISTOLOGY llic Jiypvstoinc (Figs. 26, 27). The hypostome has a relatively low-lying ectoderm (Fig. 26). The endoderm is extremely well- developed ( Fig. 27 ) . The very large cells and dense accumulations of gland cells sometimes practically obliterate the lumen. The peduncle. The endoderm of the peduncle is reduced in extent and gland cells are absent. The digestixe cells consist of very vacuolated thin strands of cytoplasm. The ectoderm is similar to that already described. The epithelio-muscular cells of this region are characterized b\- ha\"ing granules near their surface (see Fig. 30). T]u' based disk (Fig. 29). The endoderm of this area is like that of the peduncle. The ectoderm is characterized b\- the presence of a type of gland cell which consists mostl)- of double membranes and has large granules, similar to those seen in the epithelio-mus- cular cells of the peduncle, l)ut much larger. These granules of the gland cells of the ectoderm in this area are apparently the substance produced to cement Hydra to the substratum. The ectodermal cells of the pedal disk have small extensions of cytoplasm or villi on their surface. No granular and cuticular material is present on the surface of Hydra at this level. Hie tentacle (Figs. 22, 24, 28). The tentacle arises at the level of the Inpostome. Sections through this region reveal a gradual change of the cells with the tentacle compared to the hypostome having a reduced endoderm and ectoderm, reduced number of gland cells, increased vacuolation of digestive cells, increased num- ber of cnidocytes, and perhaps better de\ elopment of the muscle filaments in the epithelio-muscular cells. The endoderm-mesoglea interface at this level exhibits a characteristic scalloped appearance (Fig. 28). The endoderm of the tentacle is severely reduced (Fig. 22). It consists of \'ery thin wisps of cytoplasm of digestive cells enclos- 8 THE BIOLOGY OF HYDRA : 1961 ing huge vacuoles. The ectoderm is also very thin and has a series of bulges or ridges. At the height of each ridge are present the cnidocytes containing apparently mature nematocysts (Fig. 22). The cnidocytes can be surrounded by the cytoplasm of epithelio- muscular cells. Sometimes the cnidocytes rest on the muscle layer, in which case, epithelio-muscular cell cytoplasm is on three sides of them (Figs. 22, 24). At other times, the cnidocyte is the surface cell (Fig. 24). However, as far as we can determine, each cni- docyte is enclosed by a complete cell membrane, even when one cnidocyte abuts against another (Fig. 24). Hence, the syncytial relationship of cnidoblasts with immature nematocysts seen in the column has broken down during the maturation of these cells seen as cnidocytes in the tentacle. Frequently, muscle filaments are present in the cnidocyte (Fig. 24). At what stage of cnidoblast development these muscle filaments make their appearance is unknown. The bud (Fig. 30). We have not studied the bud in detail, but we have noticed that the ectodermal cells of mother and bud fuse insensibly with the cuticular layer of the mother continuous over the surface of the bud. The portion of bud attaching to the mother Hydra looks essentially like the peduncle of the mother Hydra with very vacuolated endodermal digestive cells and ectodermal epithe- lio-muscular cells containing granules near their surface. Pertinent literature is cited in: Hess, A., A. I. Cohen, and E. A. Robson. 1957. Observations on the structure of hydra as seen with the electron and hght microscopes. Quart. }. Microscop. Sci. 98: 315-326. EXPLANATION OF PLATES All photographs are electron micrographs of Hydra. The line on the photographs indicates 1/j,. PLATE I. Fig. 1. Cross section of the ectoderm showing the epithelio- muscular cells with their nuclei (N), inclusion bodies (I) frequently found in these cells, the muscle filaments at the base of the epithelio-muscular cell forming the ectodermal muscular layer (E), and the vacuoles (V) in the cells. M is the mesoglea. Fig. 2. The arrows point to the "snap-fastener" relationship between two cells in the ectoderm where a portion of cytoplasm of one cell evaginates and indents an adjacent cell. H.:M^' ■/"^^^ 1 4 lit wl/^\"^^?^ ";\ '^^ fcr-^ .//;■ >t v.- :^ PLATE II. Fig. 3. Cross section of the ectoderm (surface of animal to to the left) showing epithelio-muscular cells (E), interstitial cells (I) and cnidoblasts (C), the latter in apparent syncytial relationship and having nematocysts and prominent double membranes. 10 11 PLATE III. Fig. 4. An interstitial cell with its nucleus, granular cyto- plasm, small mitochondria and a Golgi apparatus (G). Fig. 5. The line, indicating the magnification, passes through an appar- ent cytoplasmic bridge between two adjacent interstitial cells. Fig. 6. The surface of Hydra showing the cuticular substance resting on a membrane (Arrow # 1). Arrow # 2 shows another membrane, probably the limiting membrane of the surface cells. 12 -^iim^m. 13 PLATE IV. Fig. 7. Cnidoblasts in syncytial relationship during the be- ginning of nematocyst development. Double membranes are present in the cytoplasm. See figures 8 to 10. Fig. 8. Cnidoblasts in syncytial relationship at the height of nematocyst development. The double membranes in the cytoplasm have increased in amount and are very conspicuous in the cell. 14 •*-* 15 PLATE V. Fig. 9. Cnidoblast with a well-developed nematocyst. The double membranes are present, but are regressing in amount. See figure 8. Fig. 10. Cnidocyte (mature cnidoblast) which contains a fully-developed nematocyst (not seen in figure). The double membranes in the cytoplasm are severely reduced in amount and only a few strands are left. G is a Golgi apparatus. 16 •"' •» , i .i " 1' / -:Sf'-'' « e * *) « "G. '' \ V ! 10 " #v i7 PLATE VI. Fig. 11. A gland cell with light vacuoles and concentrations of double membranes toward the base of the cell around its nucleus. The lumen is to the right. Fig. 12. Dark vacuolar contents in another gland cell. Gland cells with dark vacuolar contents are seen only rarely. 18 '^ }}.ri'I^ 19 PLATE VII. Fig. 13. Digestive cells with relatively small vacuoles and large dark inclusion bodies, probably lipid. The nuclei of digestive cells and fairly light cytoplasm with mitochondria and some double membranes can also be seen. Fig. 14. The surfaces of two digestive cells lining the lumen. The cells contain inclusion bodies. The cell membranes at the junction of the two cells near the lumen are rather dense for a short distance and resemble a terminal bar. Short tortuous process of cytoplasm extend into the lumen. 20 Plate VIII. Fig. 15. Sections showing the relations of ectoderm (EC), endoderm (EN) and mesoglea (M). The endoderm cells have muscle filaments in their base and send processes into the mesoglea. If the origin of the processes is missed in section, the processes appear to be lying as organelles or inclusions in the mesoglea. The mesoglea has no membrane around it. If the cell membranes of adjacent cells are traced for a short distance from the mesoglea, collections of granules, similar to those in the mesoglea, can be found in the extracellular space (arrow). 22 L EC M V J ^ i, I \ % w- EN r i / JOI' / li f' W ^ / i V /. -^ .#*•" I /' \ i 23 PLATE IX. Fig. 16. The ectoderm is at the top, the endoderm at the bottom and separated from each other by mesoglea. A rather robust process passes from an endodermal cell into the mesoglea. Fig. 17. Longitudinal section of the junction of two muscle fibers in the ectoderm. The alternating light and dark densities on adjacent cell mem- branes yield a desmosome-like effect. Muscle filaments do not pass from one cell to the other. Fig. 18. Flagella in the lumen of the hypostome surrounded by a mem- brane and exhibiting the characteristic pattern of filaments. 24 *v T": 1 -^ /' -. .- / \ '■•■ J ■^q' ^ ■ ■»_^,. fcrf-^assfcsof '^» ^>Cjr'" ■ ,^.h.a ' 0- ^^•'V- ,Wi_ .^ 16 »£ 17 i EN y c 29 PLATE XII. Fig. 22. A ridge or low elevation in the ectoderm of the tentacle containing cnidocytes (C) apparently embedded in the cytoplasm of an epithelio-muscular cell (EP). A thin mesogleal layer (arrow) separates the ectodermal cells from the attenuated, highly vacuolated endoderm (EN). Fig. 23. An extracellular space containing granules and enclosed by the cell membranes of four endoderm cells, yielding a star-shaped effect. 30 ti EP' 'A •■\ ■♦. ; .. i« 31 PLATE XIII. Fig. 24. Cnidocytes in the tentacle embedded in the cyto- plasm of epithelio-muscular cells (EP). The cnidocytes are not syncytial, but are enclosed individually in their limiting membranes. One of the cnidocytes has muscle filaments (M). S is the surface of the animal. Near S, a cnidocil is seen. Fig. 25. An extracellular space containing granules and enclosed by the membranes of endoderm cells and yielding a star-shaped effect. 32 5*iiS?^«*^1 f -M •#»%. EP .-' A ..■^ 24 1 v. Z5 PLATE XIV. Fig. 26. Section through the hypostome showing relations of ectoderm, endoderm and mesoglea (M). The epithelio-muscular cell (EP) contains vacuoles, organelles and inclusions as described previously and muscle filaments in its base. The digestive cells of the endoderm (EN) con- tain vacuoles, organelles and inclusion bodies and have muscle filaments in their base. S is the surface of the animal. 34 .■..•'A'Sf, «^' ^• M ». ■* :r^^:^-^^ Iv.^x ;^«^ 26 f -• PLATE XV. Fig. 27. Section through the endoderm of the hypostome. Gland cells (G), with their vacuoles and digestive cells (D) containing inclu- sion bodies, probably absorbed food, line the lumen (L). The flagella and cytoplasmic processes extending from these cells are seen in the lumen. 36 PLATE XVI. Fig. 28. Section through the junction of the hypostome and the tentacle. The arrows indicate the approximate plane of attachment of the tentacle (on the right of the arrows) to the hypostome. The interface of the endodermal cells (EN) and the mesoglea (M) has a characteristic scal- loped appearance. Muscle filaments on the epithlio-muscular cells are per- haps better developd in the tentacle than in the hypostome. 3H # ^ I h M '/A EN %>^: i '"' k- %\ ^ / i / I ' / / / / ^ X -■^a*C ; 3Q§ IIhI • '^ ^ '< J s ^ 39 PLATE XVII. Fig. 29. The gland cells of the ectoderm of the pedal disk. There is no cuticular substance on the surface (S). Small cytoplasmic proc- esses extend from the surface of the cells. The granules (G) are probably the secretion manufactured by these cells, especially those with more double membranes and fev/er and smaller granules farther from the surface, to cement Hydra to the substratum. 40 29 I i m :*«? ^-V \..^. J ^ •#v 1J •:>>' n .>%^ ", 4i PLATE XVIII. Fig. 30. Place of origin of bud from mother. The arrow indicates approximately the point of attachment of the bud to the mother. Epithelio-muscular cells of the mother are on the bottom of the photograph. The more highly vacuolated epithelio-muscular cells of the bud are similar to the epithelio-muscular cells seen at the level of the peduncle. Similarly, the granules seen near the surface of the epithelio-muscular cells of both mother and bud are characteristic of epithelio-muscular cells of the peduncle. The cuticular layer is continuous over the surfaces of bud and mother. 42 'S'«?.^'*:%3**^ ^r "t-^SN ~V^->^ \ ( i^ r ^--v, ic •i-' ■ *! !%i.tr.5^.: - XN -■^ ^3..^.: 43 44 THE BIOLOGY OF HYDRA : 1961 DISCUSSION LOOMIS: At what point do interstitial cells start differentiating the four types of nematocysts? HESS: I'm sure someone else later in the program could answer that. I have not actually worked on the structure of the nematocyst per se, just the cnidoblast. SLAUTTERBACK: Dr. Hess, I'm afraid your fine micrographs have stolen the thunder from the rest of the electron microscopists here. I did not want to raise the issue of what you have called the gland cell. I believe that on the basis of location, staining proper- ties and appearance of the secretory granules in electron micro- graphs, your term includes two distinct cell types as has been sug- gested in the classical literature. We have been calling the cell which is more prominent in the hypostome region and resembles the goblet cell of the vertebrate digestive system, a mucous cell. The other type, which is more prominent below the hypos'come and resembles the pancreatic acinar cell, we have called the zymogenic cell. HESS: It seemed to me that these different appearances might be cyclical changes. Most of the cells have light vacuoles and only rarely do some of them stain darker in the electron microscope. I haven't done histochemical staining, and you might be right that two different cell types occur because many people speak of these two kinds of cells. BURNETT: I would like to mention some histochemical results we have obtained on regenerating hydra. If the hypostome of the hydra is excised, we find that mucous cells begin to appear in abundance in the gastrodermis at the point of excision about 12-18 hours after cutting. The secretory material in these cells is PAS positive, stains with alcian blue, is metachromatic after toluidine blue or methylene blue staining, and is removable by hyaluronidase digestion. This material is most certainly an acid mucopolysac- charide. Gland cells appear six hours after excision. The secretory droplets in these cells are several times larger than those found in the mucous cells. Moreover, gland cells do not stain with alcian ARTHUR HESS 45 blue and are not metachromatic, but positive to Millon's reaction for proteins. These two types of cells are, therefore, quite different from one another both histochemically and morphologically. I have a question. Were the cnidoblasts in the same cluster forming the same type of nematocysts? HESS: I didn't notice the type of nematocyst, but all those within a cluster seem to be in the same stage of development. FAWCETT: I would like to comment on that point. It has been our experience that within any single cluster of cnidoblasts, they are all forming nematocysts of the same kind. They are also pre- cisely synchronized in their development. I would comment further, if I may, on the syncytial relationship that was mentioned. I noticed in Dr. Hess' pictures two distinct kinds of syncytial relationships. In a number of instances, the connections between cells appeared simply as small discontinuities of varying lengths in the pairs of membranes constituting the boundaries between cells. We have seen such apparent communications, but although our technique was seemingly good enough to make it unlikely that these were artifactitious breaks in the continuity of the cell membranes, this has nevertheless always been a disturbing possibility. There is an- other kind of syncytial relationship between cnidoblasts which is clearly not artifactual, and is of considerable interest in relation to the mechanism of cell division and the control of differentiation. It is this kind of intercellular bridge, found in both interstitial cells ( Fig. 1 ) and cnidoblasts ( Fig. 2 ) , that I would like to illustrate in order to emphasize the special nature and probable significance of the syncytial relationship between cnidoblasts. Groups of eight or sixteen cells arising by proliferation from a single interstitial cell remain connected by bridges a micron or so in diameter, en- closed by a specialized, thickened area of membrane that has a characteristic contour. There is no possibility that this localized thickening of the plasmalemma and special configuration of the sin-face could arise as an artifact of specimen preparation. Notice the heartshaped outline of the intercellular space and the definite ridge that encircles the waist of the intercellular connection. Dr. Slautterback and I believe that such bridges arise during division of the interstitial cells when the constricting cleavage furrow en- .-/ji Fig. 1. Intercellular bridge of interstitial cells. counters the spindle remnant, and is arrested by it for a time. This occurs very commonly in mitotic divisions in many kinds of genninal and somatic cells and gives rise to a transient structure called a spindle bridge. Usually, however, such connections between the daughter cells endure only for several minutes and then when the spindle remnants have resorbed the cleavage is completed. Evi- dently in the case being described here, cleavage does not resume and absorption of the spindle filaments leaves the daughter cells in open communication through short cylindrical bridges large enough to permit mitochondria and other formed elements of the cytoplasm to pass from one cell body to another. As a consequence of the matter in which they are formed, there is never more than one Fig. 2. Intercellular bridge of cnidoblasts. 46 I Nematoc^t p^. / nterceltular Bridge Nematocyst 47 48 THE BIOLOGY OF HYDRA : 1961 such bridge between any two cells in the cnidoblast cluster. The bridges persist throughout the period of differentiation of the nema- tocysts. If the nematocysts are eventually to migrate as individual cells, the bridges connecting them must be severed at some time late in their differentiation, but this process has not yet been ob- served. We believe that the syncytial relationship of the cnidoblasts is probably the morphological basis for the synchrony of their differ- entiation. It is interesting that the same kind of synchrony is seen in the groups of developing germ cells in the testes and these are also connected by intercellular bridges that form in the same way. CLAYBROOK: Do either of you find cytoplasmic bridges between different cell types, or are they only between two of the same kind? HESS: I've only seen them between the same cell type. How about you? FAWCETT: Bridges of the kind I have been describing occur only between cells of the same type. HESS: I've seen a break in the cell membranes of the spermatids, like the first type of interconnection of which you spoke. We thought that it was an artifact until we saw cytoplasm and mito- chondria in the intercellular bridge running between the two syncy- tial cells. GAUTHIER: May we return to the subject of gland cells? If the two cell types represent only a cyclical change in one cell type, would you expect that starvation might produce a levelling off so that only one type would be present? HESS: Well, I thought that the different appearances of gland cells indicated cyclical changes of one cell type, but others here appar- ently disagree. GAUTHIER: In preliminary experiments with starved hydra, I have found that two distinct types of gland cells persist for as long as twelve days, GOREAU: I am interested in the so-called microvilli you have shown. We have seen microvilli in gorgonian and scleractinian material which have a much more regular and permanent appear- ARTHUR HESS 49 ance than anything you ha\'e shown. The processes in your sections of Hydra epidermis look to me hke temporary cytoplasmic pseudo- podia. They certainly don't have the same well organized distribu- tion that is seen, for example, in the epidermal cells of corals where the microvilli are arranged in a regular ring around the base of the flagella (Goreau and Philpott, 1956. Exptl. Cell Research 10:552). I'm also interested to see that the epidermal cells of Hydra are not flagellated. HESS: All the flagella of Hydra arise from endodermal cells and extend into the lumen. GOREAU: We've never found more than one flagellum per cell, whereas you seem to think there are more than one. HESS: Yes, in the gastrodermis, each cell apparently has from two to four flagella. The Fine Structure of Intercellular and Mesogleal Attachments of Epithelial Cells in Hydra Richard L. Wood Department of Anatotny, University of WasJiington Scliool of Mediciiie, Seattle. Cellular interactions in multicellular organisms have been ex- amined both from the physiological and the morphological points of view. As a result of these studies it has become clear that there are certain general features of epithelia which are related to special kinds of adhesive properties. It is further realized that these special adhesive properties are not distributed uniformly over the cell surface. The epithelial layers of hydra share these general properties of epithelia, although the details of intercellular attach- ment sites seem not to have been studied extensively in the past. Hydra consists essentially of a bicellular leaflet of epithelia and, therefore, is well suited for studies of epithelial cell interactions. The epithelia of hydra are perfectly good epithelia, but at the same time the individual cells serve several functions, many of which are not usually associated with functions of epithelium in a single layer in higher organisms. The presence of well developed terminal bar type attachment areas between these epithelial cells of hydra is certainly to be expected from our knowledge of higher organisms. Such areas do occur and the detailed structure differs from previ- ously described intercellular attachments. Basal processes of many epithelial cells in hydra contain muscle fibers. Special relationships between adjacent muscle fibers and ^This research was aided in part by Grant No. H-2698 from the National Institutes of Health, Public Health Service. 51 52 THE BIOLOGY OF HYDRA : 1961 between muscle elements and connective tissue, or mesoglea, would also be expected, and indeed they also occur. The purpose of this presentation is to review some of these relationships as I have observed them using light and electron microscopy. These observa- tions pose a great number of additional questions which will require some new approaches for further elucidation. In this presentation I will refer to the intercellular attachments as desmosomes. I prefer desmosome as a general descriptive term for intercellular attachments because the term was originally pro- posed with a recognition of the functional relationship and basic similarity of the various forms of intercellular attachment (9). The concept of desmosome (literally "bonding body") seems to be well substantiated by micromanipulation experiments with various kinds of epithelium from different organisms. The present observations were made on specimens of Chloro- hydra viridissima and Pelmafohijdra oligactis. Material was fixed in osmium tetroxide buffered in acetate-veronal (6) or s-collidine (1) at pH 7.4. The tissue was dehydrated in ethyl alcohol and embedded in a mixture of n-butyl and methyl methacrylates or in either Araldite or Epon epoxy resin (see Luft, ref. 5). Light micrographs were made from one micron sections cut from epoxy embedded blocks and stained according to the method of Rich- ardson, et al. (7). The electron microscopy was done on an RCA- 2C with an improved power supply and with a Siemens Elmiskop I. The epithelial layers of hydra mostly consist of single layers of cuboidal to columnar epithelial cells. In the epidermis interstitial cells occur between the epithelial cells near their bases and nema- tocytes occur between epithelial cells at the outer surface of the animal. The gastroderm contains two easily identifiable cell types, nutrient cells and glandular cells. A thin lamella of mesoglea sepa- rates the two epithelial layers. This general configuration is dem- onstrated in the first illustration. Figure la is a light micrograph of a transverse section through the region of the hypostome in Chlorohydra. Glandular cells and basally located intracellular symbiotic Zoochlorellu may be identified in the gastrodermis. Light areas near the mesoglea at the base of the epidermal cells represent cross sections of muscle fibers. Figure \h shows a trans- verse section through the column of Pelmatohydra. The larger "•P* ^Sk 7^* "* " * r -/"Ifc « ^ _. Fig. 1. Light microscope pictures of a, Chlorohydra and b, Pelmatohydra. In a the epidermis is at the top and the gastrodermis at the bottom. The two layers are separated by the mesolamella along which muscle fibers may be seen. In b the epidermis is at the right and the gastrodermis at the left with the mesolamella between. Note the obvious muscle fibers at the base of the epidermis and the connection between epithelia at the circle. Desmosomes appear at the arrows. The black circular objects in the gastrodermis of a are Zoochlorella; in b similar structures are food particles. C, cnidoblasts; N, nucleus, o— 2200X. fc— 2200X. 53 54 THE BIOLOGY OF HYDRA : 1961 size and lack of Zoochlorella make Pelmotohi/dra easier to ex- amine. At the free outer surface, adjacent epidermal cells are bound together by terminal bar type desmosomes ( arrows ) . These desmosomes were not described in earlier light microscope studies of hydra or in more recent electron microscope studies by other workers (2, 4, 11). Interstitial cells, nematoblasts and gastroderm- al nutrient cells are seen clearly in Figure lb. These general features of hydra epithelia are shown to even better advantage in low magnification electron micrographs. Figure 2 is an electron micrograph of a section through the gastric region of Chlorohijdra. The prominent dense bodies in the gastro- dermal cells are Zoochlorella. Other identifiable features include nuclei, microvilli, other cellular inclusions and muscle processes. The two epithelial layers are separated by the thin mesolamella which appears dense in this picture. Desmosomes appear as areas of increased density between adjacent cell surfaces, especially near the outer surface of epidermal cells and the lumenal surface of gastrodermal cells (arrows). Similar densities occur between adjacent membranes of interdigitated muscle processes (Fig. 2, circle ) . In both species of hydra examined the desmosomes which are present near the free surfaces of epidermal and gastrodermal cells display a very complex morphology when viewed at higher magni- fication. The two apposed plasma membranes each exhibit the dual profile of the "unit membrane" of Robertson (8), the two peaks of density being about 70 Angstrom units apart. The increase in density noted by light microscopy and in lower magnification electron micrographs is seen to be due to a condensation of intracel- lular material and to the presence of a specially oriented intercel- lular matrix. These features are shown in Figure 3, a and b, an example of an epidermal desmosome in a specimen prepared in the usual way and then stained with phosphotungstic acid prior to embedding. In this preparation the junction of at least three differ- ent epidermal cells is represented. The condensation of intracellular material appears somewhat vague at this magnification but the organization of intercellular material is well demonstrated. The two apposed cell surfaces are connected directly by a series of parallel densities oriented perpendicular to the plane of the plasma RICHARD L. WOOD 55 membranes. The intercellular space is thereby divided into a series of compartments. From examination of oblique or longitudinal sections of these desmosomes it is clear that the intercellular connections are not \ 33t^ O \ y9i Fig. 2. Low magnification electron micrograph of Chlorohydra. The epi- dermis is at the top and the gastrodermis at the bottom. Zoochlorella appear in the gastrodermal cells. Cross sections of muscle fibers lie adjacent to the mesolamella in the epidermis. Desmosomes are apparent in both the epi- dermis and the gastrodermis (arrows). Specialized muscle-to-muscle attach- ment is indicated by increased densities such as at the small circle. Note the large intracellular vacuoles in both epithelial layers. V, intracellular vacuole. 2700X. (Originally published in J. Biophysic. and Biochem. Cytol., 6: 343-352, 1959). Plasma membrane 0.1// Fig. 3. An epidermal septate desmosome of Pelamatohydra. The junction of three cells in a shows the reflection of ceil surfaces into the attachment region (arrow) and the prominent cross connections. At b the lower central portion of a (framed) is shown at higher magnification. At the double arrow the outer dense component of the lower plasma membrane appears to be continuous with the dense lines of the intercellular septa. The diagram at c illustrates the arrangement of septate desmosomes as visualized from these observations. See text, o— 53,000X. 6— 130,000X. (These illustrations origi- nally appeared in J. Biophysic. and Biocbem. Cytol., 6: 343-352, 1959). 56 Fig. 4. a. End-to-end apposition of muscle fibers in the epidermis of Pelmatohydra. Note the irregular line of contact and the increased density associated with the two cell surfaces (arrows). The myofilaments appear as small streaks oriented towards the attachment zone. The subjacent mesoglea exhibits very fine filaments more or less randomly arranged, b. End-to-end apposition of gastrodermal muscle fibers at high magnification. The fila- ments in gastrodermal muscle appear less conspicuous than those in the epi- dermis. M, mitochondrion; ME, mesoglea. a — 17,000X. b — 80,000X. 57 58 THE BIOLOGY OF HYDRA : 1961 simple bars but actually form lamellar partitions, or septa. The exact nature of the septa is not yet clear but there is some indication that they may be continuous with the outer dense components of the two apposed "unit" membranes (Fig. Sb, arrow). A diagram- matic representation of this type of desmosome is shown in Figure 3c. The two plasma membranes are joined by septa which may possibly have direct connections to the outer components of the apposed plasma membranes. Lack of continuity, as illustrated at B, is more commonly seen than continuity shown at A, so it is uncertain which configuration is more accurate. Perhaps both conditions occur along the course of the same septum. Another type of intercellular attachment occurs in hydra where muscle processes are apposed end to end. Myofilaments appear to insert into regions of increased density and the two cell surfaces are maintained always in close approximation. This relationship, shown in Figure 4 (a, h) resembles the intercalated disc of verte- brate cardiac muscle. The intercalated disc is now recognized as a kind of desmosome (see Sjostrand and Andersson, ref. 10). This type of attachment is particularly clear in longitudinal sections of the epidermis (Fig. 4«). In cross section they appear at the base of the gastrodermis and may be distinguished as irregu- lar, dark streaks in light micrographs (Fig. \h). In the basal region of the tentacles, and in the upper part of the column, there is a special type of relationship of the muscle pro- cesses to the mesoglea. This type of attachment may also be identi- fied by light microscopy in favorable preparations. Figure 5a is a light micrograph of a longitudinal section of a tentacle near its junction with the hypostome. Near the mesoglea an area of in- creased density is quite apparent, but the details of its structure are not obvious. A similar region viewed in the electron microscope (Fig. 5i>) shows that the density is caused by a specialized muscle insertion on mesoglea. The attachment is accomplished by means of a narrow finger of epitheliomuscular cell cytoplasm which becomes intimately associated with an area of increased density in the adjacent mesoglea. The cytoplasmic finger contains a condensa- tion of material which appears to be organized into a series of small tubular elements arranged at right angles to the plane of the plasma membrane (Fig. 6). The disposition of these tubules sug- Fig. 5. Attachment of muscle to mesoglea in Pelmatohydra; o is a light micrograph of a longitudinal section of a tentacle. The epidermis is at the left and its scalloped surface indicates partial contraction of the tentacle. The dense line at the base of the epidermis (arrow) indicates a specialized form of attachment of muscle to mesoglea. A similar region viewed in cross section with the electron microscope is shown at b. Epidermal muscle fibers lie adjacent to the mesoglea. An extension of one muscle fiber becomes as- sociated with a projection of mesoglea. See text. L, lumen; N, nucleus; M, muscle; ME, mesoglea. o— 2,000X. 6— 20,500X. 59 60 THE BIOLOGY OF HYDRA : 1981 gests a supporting function such as might be required in areas where there is increased mechanical stress. The final example of a possible attachment mechanism in hydra which I will present is another arrangement of epithelial cell surfaces at the level of the mesoglea. In my preparations, both Fig. 6. High magnification of muscle attachment to mesoglea at a ten- tacle base in Pelmatobydra. The mesoglea is to the right. The cytoplasmic finger of the muscle fiber extends vertically through the center of the pic- ture. Note the transversely oriented tubular structures (arrow) and the pat- terns of increased density. See text. ME, mesoglea. 120,000X. RICHARD L. WOOD 61 Fig. 7. Mesogleal relationship of epidermal muscle processes (top). The muscle fibers extend irregular processes into the mesoglea, some of which traverse the mesoglea completely (center). Those which traverse the mes- oglea may abut against similar processes from the gastrodermal cells (bot- tom). C, cnidoblast; M, muscle. 7,000X. for light microscopy and for electron microscopy, the mesogleal sm'face of epitheliomuscular cells is plicated and irregular. Fre- quently the mesoglea is completely traversed by narrow cytoplas- mic processes (Fig. lb, circle; Fig. 7). These connections were seen and illustrated by Hadzi in 1909 (3) but have not captured the attention of morphologists again until rather recently. They extend from epithelial cells situated in both layers. Within the mesoglea, or at one epithelial surface, the processes may abut against the oppo- site epithelium either along a fairly broad surface or in a very limited area (Figs. 7,8). So far as has been observed, the processes extending across mesoglea represent regions of contact between the two epithelial layers but not cytoplasmic continuity. Two dis- tinct plasma membranes have always been seen although a reduc- tion of the spacing between the apposed membranes is often evident. In fact, the typical 150-200 A separation may essentially disappear, as is illustrated in Figure 8. 62 THE BIOLOGY OF HYDRA : 1961 The irregular profile of epithelial cell surfaces being presented to the mesoglea could possibly reflect a mode of insertion into the extracellular matrix, as suggested by Hess, Cohen and Robson (4). Cell contacts across the mesoglea could be related to an attachment function but could also be related to the transfer of nutrients from gastrodermis to epidermis or to some mechanism of direct integra- tion between the two muscle layers. In this paper I have attempted to present a brief account of some of the various types of attachment that occur between epithe- lial cells and between the epithelial cells and mesoglea in hydra. The conclusion that all of these specializations represent kinds of cellular attachment is based on comparison with other organisms and on attempts to correlate structure with function. These attempts take into consideration special physiological problems related to the fresh water environment and the mode of feeding of these organisms. A permeability barrier for the organism seems essential and attempts to find a structural basis for this barrier have been unsuccessful in the past. I have postulated that the septate form of desmosome could be important in preventing the influx of excessive fluid to the inter- cellular spaces (12). There is no direct evidence, however, that septate desmosomes are any more effective in this respect than are ordinary terminal bars found in ductile epithelium or gut of higher forms. In fact, I am not sure that one can say positively that term- inal bars function to preserve the intercellular milieu in any situa- tion but evidence seems to favor such an interpretation. At end-to-end and lateral contacts of interdigitating muscle fibers a strong adhesion is something which would appear essential for the efficient transmission of force during contraction of the muscle fibers. By the same token, the special kinds of attachments of muscle to mesoglea might be expected in areas of particular stress, such as presumably occurs at the bases of the tentacles. All these forms of attachment must also be interpreted as having importance for preserving relative cell positions during active movements of the animal. The chemical or molecular organization of the cell surfaces is certainly not yet known in suflicient detail to permit conclusions about the actual mechanism of attachment either between adjacent cells or between cells and mesoglea. I believe, however, that additional information will be obtained through further studies using techniques for dissociating cells and by using RICHARD L. WOOD 63 specific enzyme digestion. Analysis of appropriately treated material with high resolution electron microscopy may provide further information not only on the mechanism of intercellular attachment but also on the molecular structure of cell membranes themselves. Fig. 8. Interepithelial connection across mesoglea of Pelmatohydra. The epidermal process (top) abuts against a gastrodermal muscle fiber. Note the collapse of the normal intercellular separation at the region of contact (arrow). ME, mesoglea. 13,000X. 64 THE BIOLOGY OF HYDRA : 1961 REFERENCES 1. Bennett, H. S., and J. H. Luft. 1959. s-Collidine as a basis for buffering fixa- tives. ]. Biophys^ic. and Biochem. Cijtol. 6: 113-114. 2. Chapman, G., and L. Tilney. 1959. Cytological studies of the nematocysts of Hydra. I. Desmonemes, isorhizas, cnidocils and supporting structures. II. The stenoteles. /. Biophtjsic. and Biochem. Cijtol. 5: 69-84. 3. Hadzi, J. 1909. Ueber das Nervensystem von Hydra. Arb. zool. Inst. Wien. 17: 225-268. 4. Hess, A., A. Cohen and E. Robson. 1957. Observations on the structure of hydra as seen with the electron and Hght microscopes. Quart. J. Micr. Sc. 98: 31.5-326. 5. Luft, J. 1961. Improvements in epoxy resin embedding methods. ]. Biopliysic. arid Biochem. Cytol. 9: 409-414. 6. Palade, G. 1952. A study of fixation for electron microscopy. /. £.v/;. Med. 95: 285-297. 7. Richardson, K. C, L. Jarett and E. H. Finke. 1960. Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technology 35: 313-323. 8. Robertson, J. D. 1959. New observations on the ultrastructure of the membranes of frog peripheral nerve fibers. /. Biophysic. and Biochem. Cytol. 3: 1043-1047. 9. ScHAFFER, J. 1920. Vorlesungen iiher Histologic iind Histogenese. W. Engle- mann, Leipzig, pp. 69-100. 10. Sjostrand, F., and E. Andersson. 1954. Electron microscopy of the intercalated discs of cardiac muscle tissue. Experientia 10: 369-370. 11. Slautterback, D., and D. W. Fawcett. 1959. The development of cnidoblasts of Hydra. An electron microscope study of differentiation. /. Biophtjsic. and Biochem. Cytol. 5: 441-452. 12. Wood, R. 1959. Intercellular attachment in the epithelium of Hydra as revealed by electron microscopy. /. Biophysic. and Biochem. Cytol. 6: 343-352. DISCUSSION WAINWRIGHT: Do you have any ideas concerning the site of synthesis of the mesoglea? WOOD: In Hyman and other textbooks it is claimed that the mesoglea comes from both epithelial layers. I really have little more to add. It is always strictly extracellular and has no hmiting mem- brane, as Dr. Hess has already pointed out. It corresponds to the connective tissue of higher forms. I don't know exactly how it arises. RICHARD L. WOOD 65 HESS: I've seen mesoglea in very young buds almost immediately after their formation. FAWCETT: I have no reason to regard the mesoglea as different from any other epithelial basement membrane except for its greater thickness. Where one has two epithelial or endothelial layers arranged base-to-base in higher forms, one finds a layer of amor- phous, PAS positive material which looks very much like a thin mesoglea. I've always found this a very attractive prospect in hydra research. Perhaps here is the best place to study the structure and properties of basement membranes, and we might gain informa- tion from the mesoglea that could be carried over to the basement membranes which are such physiologically important structures in higher forms. I would ask you a question on terminology. I wonder why you choose not to call these specialized zones of attachment "terminal bars"? I certainly agree with you that "desmosomes" is preferable from every point of view to the term "attachment plaques," but isn't there an adequate functional reason for making a distinction between desmosomes and those devices that occur next to the free surface extending for the full length of the cell boundary and which may very well have the function of preventing access of material to the intercellular space. Isn't it desirable to distinguish these elongated structures from the desmosomes which are circular plaques that occur at many points along the confronted surfaces of the epithelial cells and seem to be solely for attachment? WOOD: I agree, Dr. Fawcett, but in my own terminology I regard the term desmosomes as a more general term. I then say this is a "terminal bar" type of desmosome. I'm sorry I didn't make it clearer in my presentation. This concept of the generality of the term desmosome comes from Schalfer's original description. I think "Schussleisten," which, of course, was the terminal bar, is an earlier term. Schalfer regarded the terminal bar as possibly arising from fusion of a series of small plaques. I've used the term desmosome in this general sense. I don't feel rigid about it, however. HESS: We all try to get hydra fixed in an extended state. Some of the things we see in the mesoglea might be very different, I think, depending on the state of contraction of the hydra. 66 THE BIOLOGY OF HYDRA : 1961 LUNGER: I have electron micrographs of Campamilaria endo- derm showing "terminal bar" desmosomes similar to those demon- strated for hydra by Dr. Wood. WOOD: This has been observed in several other forms. They appear in planaria, and one is described briefly in Grimstone, Home, Pantin and Robson's publication on Metridium. SLAUTTERBACK: I'm willing to call these things "terminal bars," but we must not put a functional significance upon this name because we don't have any way of knowing that these structures are excluding something from the epithelium and preventing it from reaching the mesoglea. As far as I know the only obvious function is attachment, is that right? To put it another way, we should not apply the name terminal bars to the desmosomes of hydra because the function of terminal bars, namely the impeding or preventing the flow of water, electrolytes or other substances between cells, has not yet been proved to exist in any organism or tissue other than mammalian kidney. WOOD: I think that the concept of terminal bar involves more than just this concept of separating the lumen from the intercellular space. It is a type of attachment which surrounds the entire surface of the cell. In longitudinal sections it has a bar-like structure which appears dense with certain types of stain. I agree that there is no direct evidence that these specialized desmosomes of hydra func- tion to prevent passage of water or other material intercellularly, but I think that the idea is certainly reasonable because hydra is a fresh water invertebrate and must osmoregulate somehow. There is no kidney to do this and a reduction of exposed cell surface would be one way to improve the situation. FAWCETT: There is a piece of evidence not found in hydra which indicates that terminal bars do have the function that has long been attributed to them. In recent work on the proximal con- voluted tubule of the mammalian nephron, Miller found that when he administered hemoglol)in solution to mice, the hemoglobin that filtered through the glomerulus and accumulated in the lumen of the tubule is electron dense and served as a good contrast medium. In electron micrographs one can follow the electron density of the hemoglobin between the cells of the proximal tubule as far as the RICHARD L. WOOD 67 terminal bar but no farther. Thus, at least for higher forms, large molecules do not penetrate between cells and the traditional inter- pretation of the terminal bar as a device for sealing the intercellular spaces now has some experimental substantiation. HESS: Some substances might even use those cross striations as the steps of a ladder to climb into the hydra. WOOD: I've thought of these cross connections as a system of baffle plates that might slow down penetration between the cells. Discussion on : Is there a Nervous System in Hydra? HESS: Electron microscopists say that they can't see a nervous system in hydra. And some of them say that a nervous system is not needed to account for the movements of hydra because the muscle cells in both ectoderm and endoderm contact each other allowing muscle to muscle transmission to take place. Dr. George Mackie from the Department of Zoology, Univer- sity of Alberta, has a few slides showing some silver stains of the nervous system. He might have something more convincing to convey about the presence of nerve tissue in hydra. MACKIE: This is a brief report on the results of a recent at- tempt to stain ner\'es in the body wall of hydra and Cordylophora using the classical Holmes silver technique. This work is still in its initial stages. I will begin with CorcJi/lophora. General topography of the nerve net. There is only one neuron- system in Cordylophora, unlike Velella which has two histologi- cally distinct plexuses. Neurons are abundant in the ectoderm of tentacles (Fig. 2) and hydranth (Fig. 1). We have the following figures for relative abundance of three tissue elements in a hydranth preparation where all showed well: Epithelio-muscular cells Neurons Cnidoblasts 231 94 64 Neurons also run in the ectoderm of the stem. In the hydranth the neurons lie external to the muscle fiber sheet, running in the spaces between the stems of the epithelio-muscular cells. They do not follow the cell outlines, seen in surface view. 6.9 70 THE BIOLOGY OF HYDRA : 1961 Neuron types. Structurally there seems to be little difference between neuro-sensory elements and purely nervous elements. Ap- proximately one in eight neurons has a process running up to the surface with a hair projecting externally (Fig. 3, 4, 8, 9), but of those which are entirely sub-epithelial the majority have what seems to be a reduced or rudimentary sense hair projecting into the sur- rounding tissue space (Figs. 4, 7). It is possible that such cells are modified sensory elements that have become or are becoming transformed into neurons in the strict sense. However, this does not exclude the possibility that they retain a sensory function, serv- ing for instance to record deep touch or to give position sense. The fibrous processes or neurites are similar in all these elements, whether the cell has a hair or not. Interconnections. The neurites associate freely, running side by side for long or short stretches, but there is nothing to suggest that they regularly form continuous connections. This nervous system is quite unlike the closed system of Velella which shows every sign of being a syncytium. It is much more like the non-syncytial open system of Velella. The only evidence for continuous connections is that here and there one finds binucleate neurons and in some places there are suspicious-looking pairs of neurons which could be the two halves of a binucleate pulled apart, but still in primary connection. This gives me the opportunity to insert a remark about the retention of primary connections between cells which was discussed earlier, following the paper by Hess. Such connec- tions have long been known in a variety of coelenterate cell types including young cnidoblasts, interstitial cells and epithelio-muscular Ectodermal nervous system of Cordylophora (Figs. 1-9) and hydra (Figs. 10-12) as seen in silver-stained whole mounts. Scale indicates 10 m/^. Fig. 1. area of hydranth wall showing parts of five neurons; Fig. 2. neurons in a tentacle; Fig. 3. neuro-sensory cell; Fig. 4. the three types of neurons; Fig. 5. nerve fibers in contact with young cnidoblast; Fig. 6. bipolar ganglion cell; Fig. 7. well-extended neurons in expanded epithelium; Fig. 8. neuro- sensory cell showing root of hair in cytoplasm; Fig. 9. neuro-sensory ceil; Fig. 10. neurosensory cell in contact with cnidoblast; Fig. 11. nerve fiber tract: only two out of four fibers are in focus; Fig. 12. bipolar ganglion cell. Abbreviations: en. cnidoblast; g. ganglion cell; hs. subepithelial hair; n. nerve fiber; p. process of neuro-sensory cell running to surface carrying external hair. 2 ' "' L 3 cn 4^, 7i 72 THE BIOLOGY OF HYDRA : 1961 cells. In cases where the intercellular bridge is long and slender and still contains the relic of the mitotic spindle apparatus (Hirschler's fusome) the structure may bear a strong resemblance to a nerve fiber, especially in silver preparations where the fibers take the stain like nerve fibers. I suspect that such fibers may have been mistaken for nerves by certain workers. "Innervation" of cnidohlasts. Given the abundance of neurites and cnidoblasts it is not surprising to find frequent instances where the two are in contact (Fig. 5). A rough estimate suggests that about one in five cnidoblasts are in contact with part of a neuron or neurosensory cell. No cases have been found where a neurite termi- nates directly upon a cnidoblast such as Spangenberg and Ham describe in H. litforalis. The contacts are apparently quite casual and undifferentiated. Perhaps we should not speak of innervation until we can show that these associations have functional significance. Comparison of Cordijlophora and Hydra. Hydra has proved harder to examine than Cordijlophora because the tissue is histologi- cally denser and more elaborate. However, the silver preparations do quite clearly show nerve elements. All that can be said at this stage in the work is that the system appears generally similar to that of Cordijlophora. Conventional neuro-sensory cells (Fig. 10) such as Hadzi describes have been seen as well as subepithelial ganglion cells (Fig. 12), some of which have a rudimentary hair such as occurs in Cordijlophora. If there is a noteworthy difference between the two forms it would seem to be the greater tendency in hydra for neurites to run in bundles. This has been seen near the hypostome, where bundles of up to four or five neurites (Fig. 11) have been followed for short distances, running around the ani- mal in a circular direction. As to the connections, which many claim to be continuous, I have nothing to say at the moment, except that I have not seen any junctions which I would confidently interpret as being continuous. HESS: Does anyone else have any comments? CLAY BROOK: 1 am very sorry that Dr. Spangenberg of the Texas group was not able to attend this meeting to present her studies of the nervous system in H. littoralis. I am afraid 1 cannot do a very good job of describing her methods and conclusions. DISCUSSION ON NERVOUS SYSTEM 73 Dr. Spangenberg used a methylene blue \'ital staining procedure, with a neutral red counterstain, to demonstrate the nerve cells in intact Hydra. I refer you to her recent publication (Spangenberg and Ham, 1960, /. Exp. Zool. 14S, 195-202) for detailed descrip- tions. I obser\ ed many of Dr. Spangenberg's methylene blue prepara- tions under phase contrast, and can report that they compare very closely to Dr. Mackie's silver preparations. Nerve cells with from one to seven fibrous processes were observed with interconnecting fibers between many cells. While a complete nerve net could not be stained all at once in any one animal, ner\'e networks in all regions of the bod)- were seen in various specimens. As Dr. Mackie reported, cnidoblasts are often found in close contact with ner\'e cells. This doesn't indicate necessarily that there is innervation of the cnidoblast, but the frequency of coincidence is suggestive of that. Dr. Spangenberg also identified multi-polar cells with the dis- tinct morphology of neurons in Hydra preparations dissociated into single cells with Hertwig-Schneider fluid. 1 think there is little doubt that nerve cells and a nerve net do exist in Hydra. HESS: If one wanted to be skeptical, it might be said that the "nerves" that the Texas group shows associated with the cnidoblasts are the discharged tubes of nematocysts. BURNETT: I have recently received some photographs from Semal Van-Gansen at the University of Brussels. She has dissected out nerve elements from hydra with a fine needle. In the epider- mis she finds the typical nerve net described by Hadzi (Fig. 1). In the gastrodermis she does not find a net. Instead she finds a more sparse distribution of nerve cells which do not resemble the small Fig. 1. Isolated epidermal nerve cell (Semal Van-Gansen) 74 THE BIOLOGY OF HYDRA : 1961 bi-polar and tri-polar neurons of the epidermis. Those in the gastro- dermis possess extremely long proeesses which branch profusely (Fig. 2). She has suggested to me that perhaps the epidemial net serves to coordinate the fast contraction of the longitudinal fibers, and the neurons in the gastrodermis control the slower contracting circular muscle fibers. She has been able to find sensory cells both in the epidermis and gastrodermis. I have been able to consistently demonstrate an epidermal nerve net by simply fixing a whole hydra for Yi hour in 100% alcohol and then staining for a few minutes in 0.1"? methylene blue. The nerve set is especially clear in the transparent areas of the tentacles and peduncle. If this interlacing network of bi-polar and tri-polar cells is not a nervous system then morphologically it is a unique system in the animal kingdom and one that must be reckoned with. Personally, I feel certain it is a nerve net. Fig. 2. Isolated gastrodermal nerve cell (Semal Van-Gansen). HESS: Couldn't these "nerve cells and fibers" be cell membranes radiating out from the intercellular spaces? Do the intercellular spaces stain? This is a dissection, is it not? BURNETT: Yes, this is a dissection. HESS: Well, the cell membranes could be left intact radiating from intercellular spaces filled with extensions of mesogleal sub- stance. Impregnation of these elements could yield a picture appear- ing like nerve cells and fibers. DISCUSSION ON NERVOUS SYSTEM 75 SLAUTTERBACK: Before the argument is lost Iw default Td like to inject a little bit of skepticism. I have no way of proving that the nervous system does not exist, in fact, I am not sure that I really doubt it. ( I was expecting Dr. Fawcett to stand up ahead of me and say this.) But I would like to say that most of us who have hunted for nerve cells with the electron microscope have been un- able to find any. It is at least possible that this is because the morphology of invertebrate nerves or hydra nerves is not readily recognizable. But tliis is disturbing in view of the fact that there are clear morphological criteria for the identification of nerves in vertebrate tissues; they are readily recognizable with the elec- tron microscope. In fact, I'd say more easily identified than in the light microscope. Then too, it seems to me that the musculo-epitheli- al cells are so beautifully organized for conduction in hydra, that we don't really have to postulate the existence of a nervous system which we can't see in order to account for the behavior pattern. I recognize that it will probably take arguments more cogent than these to refute a concept which has delighted liiologists for at least 70 years. I have only to say that we can't see a nervous system. We'd like things a little more sure. HESS: Muscle to muscle connections, of course, are present e\'en in mammalian smooth muscle. It wouldn't be an impossible situa- tion for hydra to use muscle to muscle transmission to execute its movements. PASSANO: I doubt that this answers our discussion, l)ut it might be of interest to tell this group of our success in recording action potentials from hydra. A few years ago C. B. McCullough and I tried to find out whether or not hydra showed non-decre- mental through conduction by looking for ner\e action potentials. We attempted to pick up actix ity of indixidual neurons, but what we got, probably, were near-simultaneous action potentials from several contiguous cells. We had results with two types of preparations. The tentacle- hypostome preparation (we cut off and discarded the column just below the tentacular base) was threaded on a silver rod through the mouth. In addition to serving to immobilize the animal the rod served as a neutral electrode. While observing with a water 76 THE BIOLOGY OF HYDRA : 1961 immersion objective we brought the tip of a conventional capillary microelectrode close to the cell body of one of the bipolar cells miderlying the epidermis between the tentackilar bases. Occasional- ly we picked up fairly strong, slow spikes, lasting 20 to 50 millisec- onds and somewhat various in shape. They were always associated with strong tentackilar contractions and always clearly came before any movement was discernable in the area under ob- servation. The tentacular reaction to glutathione did not elicit action potentials, however. The other successful preparation also gave action potentials associated with strong muscle contractions. Here we used an intact hydra suspended from the surface film and surrounded by a wire ring to immobilize the animal and to be the indifferent elec- trode. The microelectrode picked up action potentials after pene- trating the basal disk, when the gastrodermal longitudinal muscles contracted. We believe that these electrical changes associated with either tentacular or column "quick withdrawal" responses were nerve action potentials, not muscle action potentials, since they came well prior to muscle contraction, only with the quick, coordinated contractions of all the muscle fibers, and since we only picked them up sporadically. HESS: From a nerve cell? Can you get your electrode inside a nerve cell of hydra? PASSANO: We think that they are from nerve cells, since we attempted to place our recording electrode in the small bipolar cells that underlie the epidermis. Since we did not have direct coupled amplifiers available, we are not able to say whether or not we ever penetrated nerve cells. Frankly, I doubt it. Nematocyst Development' David B. Slautterback Depai'tmcnt of Anafonuj, The University of Wisconsin, Madison, Wisconsin. To a cytologist one of the most intriguing aspects of the nema- tocyst is that it is a secretory product hke many another, but unhke those commonly studied, it possesses a very high order of structural detail. To my knowledge, there are few rivals in this respect, among them being the protozoan trichocyst which serves to remind us that the coelenterates are not the only group with such highly organized secretory products. Though understanding it not at all, we are accustomed to the extremely intricate structures which cells, in an enviable demonstration of community effort, can con- struct in the extracellular space, such as hair and teeth. Still more commonplace, and seemingly more intelligible, are intracellular deposits of crystalline material. It does not stretch our imagination seriously to conceive of the mechanism which brings about this level of organization, impressive though it may be; for we can produce this same or similar structure in the laboratory without the inter\ention of cells. But comprehension of the mechanisms involved in the intracellular elaboration of such a highly organized body as the nematocyst challenges the best of our imaginative ca- pacities. Speaking for the cytologist, the rewards are well worth whatever effort is required for we can reasonably anticipate even more than elucidation of this one mechanism common to a single group of animals. Certainly new and better understanding of the organelles with which all cells must work will ensue. This after- noon we will hear several approaches to the understanding of nematocysts, their production, structure and functions. For my part I shall make a rather free interpretation of my assigned ^The work reported here was done during the tenure of U.S. Pul^lic Health Researcli Grants RG5651 and RG6934. 77 78 THE BIOLOGY OF HYDRA : 1961 topic, devoting most of my time to one of the lines of differentiation available to interstitial cells— the cnidoblast. Since Dr. Hess has shown you excellent low power electron micrographs for orientation I shall not include them in my presentation. The small, relatively undifferentiated interstitial cell is found in the gastroderm where it gives rise (at least) to the zymogenic and mucous cells, and in the ectoderm where it may differentiate into cells of the gonads, cnidoblasts and possibly some others. Figure 1 is an electron micrograph of a pair of interstitial cells in the ectoderm of hydra. The nucleus is large and the nucleolus very dense, but undoubtedly the most impressive feature of these cells is the large number of cytoplasmic granules which are molecules of ribonucleoprotein (RNP). In the cytoplasm of these cells, aside from the ribonucleoprotein granules, or ribosomes, as they are known to biochemists, there are no elements of the endoplasmic reticulum, or at least they are very sparse. The Golgi complex is represented, but only by a very few vesicles, showing a low degree of organization. Another pair of interstitial cells is seen in Figures 2 and 3. They illustrate the fact that the nuclear membrane of these cells has a specialization common to many other cell types, as at "Po" in the figure. These small circles which appear in a tangen- tial view of the nuclear membrane, and in longitudinal sections as indicated by the arrows, represent what have been called nu- clear pores. Whether or not they are physiologically "pores" or "holes" in the membrane, I think remains unproved. But in any case, it is likely that they represent specialized areas for transmis- sion of materials from nucleus to cytoplasm. This is exactly the kind of thing one would like to see in a cell which is about to differ- entiate, or for that matter, in a cell which is undergoing rapid mitotic division. However, the great desire to believe in such things, does not really substantiate their functional significance. So, while they may represent the lines of communication along which the nucleus tells the cytoplasm "now it's time to divide," or "now it's time to differentiate," this is largely speculative. These pores may be seen to better advantage in Figure 4, where a rather large piece of nuclear membrane has been cut in tangential section. The abundance of these structures in the nuclear envelope can be seen clearly. DAVID B. SLAUTTERBACK 79 Another pair of interstitial cells is illustrated in Figure 5. These show the same complex; the absence of endoplasmic reticulum membranes and now an intercellular bridge (mentioned earlier today) which shows a distinct confluence of cytoplasm between the conjoined cells. And as usual, there is an accumulation in the extracellular space of small dense particles. They measure about 250 to 300 Angstroms and in all respects resemble the particulate glycogen described by Fawcett and Selby in the atrial muscle of turtle heart and by now in numerous other cell types. I should point out, however, that it is not very common to find glycogen particles extracellularly except here in the ectoderm of hydra. And in these cells, glycogen, in my experience, as particulate glycogen, has never been demonstrated intracellularly. Never within the interstitial cells nor developing cnidoblast; only extracellularly. This would fit well with the suggestion that glycogen is broken down at the cell membrane. Returning to the intercellular bridges, your attention is direct- ed to its thickened membrane which seems to impart enough rigidity to the structure to resist deformation by the frequent shape changes of the animal as a whole. The plasmalemma is continuous from one cell to the other through the tubular bridge, although it is sharply reflected upon itself twice, and bears a peculiar annular expansion midway along the length of the bridge. Figure 6 is a striking demonstration of this form and the continuity of cytoplasm between the two cells. The vesicles in the center of the bridge could hardly be said to belong to either one cell or the other. Prob- ably the most important function of the intracellular bridge is to synchronize differentiation and thus provide large numbers of cnidoblasts in the same stage— reaching maturity at the same time. But, also in the early stages of cnidoblast development, when the cell is primarily concerned with proliferation, these intercellular bridges undoubtedly serve to synchronize the mitoses. It is possible, with some speculative stretch of the mind to suppose that the sub- stance which synchronizes these mitoses must therefore be a soluble substance, readily and rapidly transmitted from one cell to the other. And, so we have here some evidence for the fact that the nucleus when telling cytoplasm to begin a mitotic division, transmits this information by some relatively small molecule, or 80 THE BIOLOGY OF HYDRA : 1961 at least a rapidly diffusible one which quickly can reach an equili- brium level within the group of developing cnidoblasts. In my ex- perience these are usually 14 to 18 cells joined in a cluster, from which it is evident that a rapid diffusion rate is necessary to keep them all very closely synchronized. This synchrony is illustrated by the pair of interstitial cells in Figure 7. The dense clumps of granules are the chromatin material, and only remnants of the ruptured nuclear membrane persist. These are not two daughter cells in anaphase, they are in late prophase, so the mitoses are quite closely synchronized. When these cells divide for the last time, the diplosome remains near the plasmalemma ( Fig. 8 ) . The remnants of the achromatic figure, the spindle fibers can be seen clearly (S). They appear to be thin tubular structures on the order of 200 Angstrom units in diame- ter. Whether or not these spindle fibers have any progeny, or any remnant left in the fully differentiated cnidoblast, cannot yet be said. The possibility exists, and I shall point out at a later time what I believe to be their fate. You will see at "G" in the figure, a large number of vesicles belonging to the Golgi complex. Most of them do not have ribo- nucleo-protein granules upon their surfaces; but some do and still others have granules on one side and none on the other which may be interpreted as supporting the arguments for the continuity be- tween the endoplasm reticulum and the Golgi complex. Dr. Fawcett pointed out earlier today that some groups of cells are not joined together by intercellular bridges of the very intricate structure that you have just seen, but rather by simple dis- continuities of the membranes, an example of which appears in Figure 9. It is difficult indeed to argue that these are not artifacts of preparation techniques. But one can only say that they are fre- quently seen, and they appear in cells which otherwise seem very well preserved. However, two of the cells in the micrograph are bound together by an intercellular bridge of the specific type, and, it is not at all uncommon to see both types of continuit}^ within the same cluster. In fact, joining the same two cells together. Now when the endoplasmic reticulum begins to appear, we see coincidentally the first appearance of the nematocyst. The reticulum first appears as scattered vesicles in the cytoplasm, DAVID B. SLAUTTERBACK 81 rather spherical in appearance (Fig. 10); they have a very low density content. You can see at the arrows, for instance, a small amount of material within those vesicles. The nematocysts are indi- cated by "Ne"; one in the upper left hand corner and one in the lower right hand corner. It is quite difficult to detemiine exactly which name belongs with which nematocyst. But I would like to say by way of record here, that within one cluster all of the nemato- cysts we have seen are definitely of the same type and they continue to be the same type throughout the stages of differentiation. The relatively homogeneous area is the capsule of the nematocyst, and the granular area will become the tube. Around the open end, where the operculum will finally appear, there is a very dense ag- gregation of smooth vesicles which clearly belong to the Golgi Complex "GC." Notice again, the presence of glycogen granules between cells. In Figure 11 there is a cluster of cnidoblasts, early in their differentiation, and you see several sections of nematocysts and the nuclei of these cells. The intercellular bridges are quite conspicu- ous and now the endoplasmic reticulum has become considerably more prominent. The latter is seen mostly as sections of tubular structures, but there is some tendency to form flattened cysternae, typical of such secretory cells as the pancreatic acinar cell, for example. This section, however, has missed the Golgi zones. This particular illustration serves p r i m a r i 1 y to point out the remaining cytoplasmic bridges, and the progressive increase in vesicles of the endoplasmic reticulum. Figure 12 is a higher magni- fication view of cells at a slightly more advanced stage to emphasize the persistence of the intercellular bridges and the continuity of organelles, not simply continuity of cytoplasmic matrix, but or- ganelles seem to be shared between the cells. As the nematocyst develops, it acquires the appearance in Fig- ure 13. The Golgi complex is becoming very much more abundant. The centrioles, which are really a diplosome, remain at the open end of the forming capsule. The capsule is the lighter amorphous or faintly fibrous part, and the darker granular material is the forming tube protruding from the open opercular end of the cap- sule. Notice that the Golgi complex forms a close-fitting cap over the growing end of the tube. There is a continuous membrane sur- 82 THE BIOLOGY OF HYDRA : 1961 rounding this forming nematocyst which is agranular, and in all respects resembles that of the Golgi complex. You will see that the Golgi complex is formed as usual in vertebrates of flattened vesicles, expanded vesicles, and small spherical ones. It has been said that such appearances are not common in invertebrates and represent more of a vertebrate type of Golgi complex, so then hydra cnido- blasts have a vertebrate type of Golgi complex, if that's the case. The large body here at the top of the figure is lipid droplet, and in our experience lipid droplets are a ubiquitous finding in all secretory cells. Of course, lipid droplets are found in virtually all cell types, but a relatively sudden accumulation of lipid seems to go hand in hand with the differentiation of these secretory cells. In another section of the opercular end of a developing nemato- cyst ( Fig. 14 ) the Golgi membranes surrounding the growing tubule can be seen more clearly. In the Golgi zone, the three types of vesicles are evident and especially prominent in this micrograph, is this large expanded one (indicated by an arrow) whose contents appear every similar to those of the nematocyst capsule. The only appreciable difference seems to be a slightly greater den- sity of the material in the nematocyst than in the Golgi vesicle. One can often see areas where these Golgi vesicles seem to increase gradually in size getting larger and larger, and finally one of the vesicles seems to join by fusion of its membrane with that of the membrane surrounding the nematocyst tube or rod. This process is illustrated in Figure 15. It bears a remarkable resemblance to the mode of release of secretory granules in other cells in which the Gol- gi membrane surrounding the granule fuses with the plasmalemma and the membrane is broken at the point of fusion releasing the secretory product and adding the Golgi membrane to the plasma- lemma. As you can see the endoplasmic reticulum is continuing to develop. We are not yet past the peak of protein synthesis in this cell. That similar configurations are present in the isorhizas is evident from Figure 16. Here a large Golgi vesicle is being added to the nematocyst tube. Though some degree of uncertainty re- mains as to the identity of these developing nematocysts, those which you have seen before were probably desmosomes, but this one is an isorhiza, though whether it should be regarded as holo- trichous or atrichous, I cannot say. But again, you see the cen- DAVID B. SLAUTTERBACK 83 trioles at the opercular end, and the Golgi complex aggregated around the open end of the nematocyst. Figure 17 illustrates a very recent observation in our laboratory. The micrograph shows a cross section of the neck region of a devel- oping nematocyst. The accumulation of vesicles of the Golgi complex indicates the forming tul:>e has not yet extended very far out of the capsule. Immediately surrounding the Golgi membrane, which encases the nematocyst, is a row of very small tubules. They are about 180 A in diameter with a lumen about 75-80 A in diameter and a wall thickness of al)Out 50 A. In the upper right quadrant of the figure they are seen in perfect cross section. The function of these elements is not yet clear, but some of their structural relation- ships may be significant. In the interstitial cells they are found in groups scattered through the cytoplasm. Within the groups tubules are arranged at right angles to each other. They are evidently contin- uous with the tubules which have been interpreted as spindle fibers in Figure 8. In intermediate stages they are as figured here and in later stages (as Fig. 27) they continue to surround the nematocyst, oriented parallel to its long axis and are continuous at one end with the rootlets of the stiff rods ( described later in this paper ) and at the other with dense coils of tubules in the nuclear zone and seen as fibrous bodies in Figure 27. The only suggestion of function is seen in the relationship at the arrow in Figure 17. Here one tubule appears to be in direct communication with one of the small spherical vesicles of the Golgi complex. Whether this indicates a separate mechanism for the production of nematocyt capsule is not yet clear. Now to return \ er\' briefly to the endoplasmic reticulum. Figure 18 shows a fairly earh' group of cells with small tubular elements of the reticulum. In Figure 19 you will see a fairly late stage in the development of the cnidoblast. The cell has about reached the peak of its synthetic activity, and the endoplasmic reticulum now assumes a more packed formation and you see many flattened sacs which are disposed in a concentric array around the nucleus. The wider spaces (also marked with a star in Fig. 20) are areas where the reticulum has been cut obliquely and are not in reality such wide diameter structures. And finally the condition illustrated in Figure 20 is reached when the reticulum fills most of the cell. During the fomia- 84 THE BIOLOGY OF HYDRA : 1961 tion of the nematocyst, the Golgi complex is at all times in close proximity to the tip of the forming tube and that tube is formed out in the cytoplasm. It may become very very long and coiled around through the cytoplasm, but the Golgi complex caps the growing tip. In Figure 21 is a cnidoblast which has passed its peak of syn- thetic activity. We considered for sometime that the expansion of these endoplasmic reticulum vesicles was a fixation artifact due to osmotic differences in the fixative as compared to those within the cell. But by using a very wide variety of osmotic strengths and hydrogen ion concentrations, we have convinced ourselves that this is exactly what happens to the reticulum after it has passed the peak of synthetic activity. It begins to swell up, perhaps with an acute hydration of its contents. I wouldn't like to extend myself on that point, but in any case they do become vesicular again. In Figure 22 you will see a nematocyst, which shows how this forming tubule continues around through the cytoplasm. The ab- sence of Golgi vesicles from this section clearly indicates that there are still more coils of tubule elsewhere in this cell for if the tip were here we would see the Golgi membranes surrounding it. The cell in Figure 23 shows a still more advanced condition and this one is an early stenotele. The Golgi membrane sur- rounding the nematocyst is clearly discernable, and now we begin to see a concentration or aggregation of dense granules which were once randomly distributed. It is in this zone that the spines and thorns of the nematocyst tube will be formed. In this micrograph there are four sections through the coiled tube which is still outside of the nematocyst capsule. The darker bodies are mitochondria, and the endoplasmic reticulum is clearly vesicular and considerably de- creased in amount indicating the end of the synthetic phase. Though not present in this illustration, the Golgi complex is still active evi- dently collecting and concentrating material synthesized earlier in the now regressing reticulum. A more advanced stenotele cut longitudinaly is seen in Figure 24. The tubule has been withdrawn and the open end of the capsule is closed by the operculum. The laminated structure of the operculum can be seen in Figure 31. The arrows point to the ele- ments which were originally distributed at random throughout the DAVID B. SLAUTTERBACK 85 substance of the forming tube, and have now just begun to form the tubular wall and the spines and thorns. So this, I am sorry to say, is the stage soon after the retraction of the tube, which was as you saw before, wound throughout the cytoplasm. And I presume that this retraction is a very rapid process because we have never seen (or recognized) it in progress. On the other hand it may be that some of the tubes which we see lying coiled out in the cyto- plasm having a cross section somewhat thicker than usual, are these tubes undergoing withdrawal. In any case it is clear that there is no visible structure in the tube before it has been withdrawn into the capsule and that all of the intricate structures which appear later on are formed without immediate contact with cytoplasmic organelles and the mechanism of this astonishing feat remains an enigma. Figure 25 illustrates some of the elaborate detail of the structure of a stenotele and points out that the endoplasmic reticulum, which has reached a vesicular stage, is now disappearing and that the phospholipids of that membrane have gone some place else. It might be interesting to follow the displacement of these phospho- lipids with histochemical procedures, but we have not as yet tried such things. The isorhiza in Figure 26 illustrates a similar course of events in that type of nematocyst: the endoplasmic reticulum has become vesicular and vanished to a very large degree. The coiled tube is in- dicated at "T," and I presume that this is a holotrichous isorhiza be- cause, in some areas (arrows), we see what appear to be develop- ing thorns. At "Cn" in the upper right of the figure is the region where before we saw the diplosome and now we see what Hyman has referred to as the stiff rods which surround the operculum, a part of the cnidocil appartus. A similar degree of differentiation is seen in Figure 27, but the section has passed through the operculum and the cnidocil. One of the centroiles of the basal granule is at the base of the cilium, which, I believe, is just in the process of forming, and is quite broad in diameter. And just outside it, you can see one of the stiff rods. Now this is not the dense part, which you saw in the section im- mediately preceding, but this is the part which corresponds to the body of the cilium itself. The endoplasmic reticulum is much dimin- ished. The Golgi complex has retreated to the basal area of the cell 86 THE BIOLOGY OF HYDRA : 1961 and has often been described here, by silver stains, as a dense body in the basal part of the cnidoblast, but I believe it is simply an inac- tive Golgi complex. Immediately below it are very fine filaments which by newer techniques appear to be fine tubules (see Fig. 17). Figure 28 is a fortuitous section through a stenotele which is fully developed. The parts of the nematocyst are readily recog- nizable including the operculum (O), two of the three spines, and the faintly striated tubule. The membrane surrounding this structure is quite obvious. The cilium with its basal granule and one of the "stiff rods" are also prominent. Now I think it's obvious that this so-called stiff rod is very similar to the cilium in structure, but you can see faint longitudinal striations in the cilium which are absent in the stiff rod. Another structure which appears often in this zone is the multivesicular ( M V ) body which most closely resembles the lysosomes of DeDuve. I should also like to point out that there are very fine filaments visible in this micrograph which are attached to the cilium and to the stiff rods; in more favorable sections they also appear to be attached to the circumference of the operculum, and may serve in the mechanism of firing the nematocyst. In Figure 29 is a cross section of a stenotele. In the center of the micrograph the three heavy spines of the base of the tube can be seen; a dense material is gradually accumulating in them from the periphery inward. The peculiarly convoluted material aromid the spines is the base of the tubule itself and the conspicuous cross striation of it has a repeat period of about 150 A; that is, each light line measures about 75 A wide as does each dark line. The fact that this same periodicity is seen in longitudinal section (Fig. 30) suggests that the tubule is composed of a crystalline array of rod shaped molecules. ( I am not able to explain the difference between my measurements and those of Dr. Chapman though it is not impos- sible that they vary with degree of development or dehydration.) I am not going to deal extensively with the cnidocil structure at this time, but I would like to make a few additional observa- tions. In Figure 31 you will see that the stiff rods of the cnidocil appear first as a straight row of dense bodies connected by a fine dense line. And at one end of that row of bodies, there is a basal granule of an unmodified cilium (not visible in this section). This cilium can be seen in figure 32 where the stiff rods, now quite well DAVID B. SLAUTTERBACK 87 developed have begun to form a circle around the operculum. Notice that it is surrounded by fine filamentous structures which show a repeat period somewhat larger than 300 A. This section is slightly oblique to the plane of the rods so that in the upper left it has passed through the modified ciliary part and in the center and to the right has passed through the cross-striated rootlet. There are 21 of these plus the true cilium. In the next illustration you can see the relationship between the rootlet-like structure and the ciliary-like structure of the stiff rod. If you follow the membrane around the ciliary x)art you see that it passes below and peripheral to the upper end of the clearly cross-striated rootlet. It is remarkable that the rootlet which in the ordinary cilium is supposed to lend structural and function- al stability, should be offset in this way. Though there is little evi- dence to support the notion at this time, such an arrangement might function as a hinge with the ciliary part bending outward and the rootlet remaining fixed.'' Figure 34 shows a slightly oblique section through the complet- ed apparatus. Notice the 21 rods and the eccentrically placed cilium. Again the fine filamentous material which interconnects all parts of the apparatus and the operculum. The last two micrographs (Figs. 35 and 36) are taken from a section of a very different animal, but I want to use them to illus- trate an important consideration about the functioning of the endoplasmic reticulum in the cnidoblast. It is not evident from the developing cnidoblast that the ril^onucleoprotein granules must be or even can be attached to a membrane of the endoplasmic reticulum in order to function in the synthesis of new protein. It has been argued for some time that only the free granules of ribonu- cleoprotein are active and that after synthesis is completed the free granules move with their product to the endoplasmic reticulum where the product is separated and added to the contents of the lumen of the vesicle. We have seen in the proliferating inter- stitial cells that free granules arc active in the production of protein "for domestic consumption," i.e. new protoplasm. In the case at hand we have a secretory cell in the gut of a small earthworm. •"'It should be pointed out that the tubules illustrated in Figures 8 and 17 appear to be continuous with the rootlets of the stiff rods. 88 THE BIOLOCY OF HYDRA : 1961 Enchytraeus fmgmcntosus (Fig. 35). This cell is a very active protein secretor and this is the peak of its synthetic activity. It has become completely filled with endoplasmic reticulum plus a few secretory droplets and a very few mitochondria. In Figure 36 I think I can convince you that there are no free ribonucleoprotein granules in this cell; thus, attached RNP granules induce synthesis of protein for export from the cell; whether or not free ones do, I cannot say. (I cannot distribute the responsibility for the interpretations presented here, but I would like to acknowledge the important con- tribution of Prof. Don W. Fawcett to this work. ) Figures 18, 23 and 15 are reprinted here by courtesy of the Journal of Biophysical and Biochemical Cytology. They appeared in volume 3, page 441 of that Journal. The following abbreviations have been used in the accompanying illustrations: Centriole — Ce Nucleus — N Cnidocil Apparatus — Cn Nuclear Envelope — Np Endoplasmic Reticulum — ER Nuclear Pores — Po Golgi Complex — GC Nucleolus — No Intercellular Bridge — B Operculum — Lipid Droplet — L Particulate Glycogen — Gy Mitochondrion — M Plasmalemma — P Multivesicular Body — MV Ribonucleoprotein Granules — RNP Nematocyst — Ne Spindle Fibers — S Nematocyst Capsule — C Zymogen Droplet — Z Nematocyst Tube — T Fig. 1. A pair of interstitial cells showing granular cytoplasm. 8,900X. Fig. 2. Nuclear pores in an interstitial ceil. 10,000X. RNP 'mm Np r \ji ■> .^-Wiix.. m GC ^ Fig. 3. Same cell as Figure 2 somewhat enlarged. 13,000X. Fig. 4. Tangential section of the nuclear envelope in an interstitial cell showing the distribution of "nuclear pores." 32,000X. 90 --% 91 Fig. 5. A pair of interstitial ceils bound together by an intercellular bridge. 12,500X. Fig. 6. Enlargement of intercellular bridge similar to Figure 5. 22,000X. 92 us Fig. 7. Two interstitial cells from a single cluster, both in late prophase. 9,500X. Fig. 8. Diplosome of an interstitial cell with attached spindle fibers which are in fact tubules. 29,000X. 94 D5 Fig. 9. Two types of protoplasmic continuity in a cluster of interstitial cells. 12,000X. Fig. 10. Early cnidoblasts showing beginning development of endoplasmic reticulum, Golgi complex and nematocyst coincidentally. 17,000X. 9& B ' > ' w-/^ i <-' SSI- ' ,.r ** '! ER Ne -^ •c.. ^ ''1 1 j^^tm&^S^^ fe'" 1% • 1 ; Ne 10 Fig. 11. A cluster of cnidoblasts slightly more advanced than those in Figure 10. 17,000X. 98 ■K ^^1 I r yy -^ / Fig. 12. An intercellular bridge in a pair of cnidoblasts slightly more advanced than those in Figure 11. 22,000X. Fig. 13. This longitudinal section of a nematocyst shows the cop-like arrangement of the Golgi complex over the growing tip of the tubule. 15,000X. 100 .'W'* \ >3- 101 Fig. 14. Similar to Figure 13, but somewhat enlarged. 19,000X. Fig. 15. The membrane of a large Golgi vesicle has just fused to the mem- brane surrounding the nematocyst and in this way added its content to the previously synthesized tube material. 32,000X. 102 GC 14 '^MOKaBkSxmsk ^i lu^ Fig. 16. A cnidoblast containing a developing isorhiza. The same process lustrated in Figure 15 is seen here. 23,000X. Fig. 17. Cross section of the neck of a developing nematocyst. 46,000X. 104 m~ t-' ,/■ ■^1 ,v.' "^ '^'^«»«ta™SMiii fjjjUh-T. .f 17 105 Fig. 18. A cluster of early cnidobiasts undergoing synchronous differen- tiation. 13,500X. Fig. 19. A cnidoblast approaching the peak of synthetic activity. 20,000X. 106 / ^' a 18 J 101 Fig. 20. In this cnidoblast protein synthesis is going on at the maximum rate. 22,000X. 108 m'^^w<'mi i 20 lOU Fig. 21. This cell has passed the peak of synthetic activity and the reticu- lum has begun to vesiculate. 8,900X. Fig. 22. A considerable part of the coiled external tube has been cut in longitudinal section. 15,000X. 110 y' .'t ,? ER 21 / \. ^., 22 111 Fig. 23. Further regression of the endoplasmic reticulum is evident in this cnidoblast, as are several sections of the coiled tube. 8,900X. Fig. 24. The tube has been withdrawn into the capsule and the open end closed by an operculum. Fine structure of the tube has begun to form. 21,000X. 112 2 3 4. . • %; 2 4 ¥*^ 00^' 113 Fig. 25. A nearly mature stenotele. Rupture of the capsule is artifactual. n,500X. Fig. 26. A longitudinal section of an isorhiza showing regression of the endoplasmic reticulum and development of the cnidocil apparatus. 18,500X. 114 115 Fig. 27. This figure is similar to Figure 26 but the nematocyst is a sten- otele. 10,500X. Fig. 28. A stage simiior to that in Figure 27 but somewhat enlarged. 16,500X. 116 • 27 i 117 Fig. 29. Cross section of a stenotele tubule. 42,000X. Fig. 30. Longitudinal section of a stenotele tubule. 30,000X. 118 30, .^ "., 119 Fig. 31. This is the same cell as in Figure 24. Notice lamination of oper- culum and straight row of dense bodies of early cnidocil. 44,000X. Fig. 32. An oblique section through the opercular end of a stenotele showing partial encirclement by the developing cnidocil. 29,000X. 120 ' vm-ifes-.\ '\*V N>^ ""' IV^ ''^-' ^^!l*^ *._ ' ■ M * \ --y' 121 Fig. 33. Longitudinal section of a "stiff rod." 49,000X. Fig. 34. Oblique section of a fully differentiated cnidocil apparatus. 32,000X. 122 -t ^y-^ ■jr :^ 33 ...•fit ■' ...JXOt • .- -«^ ■ '•% • t e.^«& ^.: V^iil^ 1► SUCCINIC DEHYDROGENASE- -^CYT. c, ->► CYT. c > Fig. 3. A pathway for the oxidation of succinate. Not shown are the fat-soluble factors implicated in succinate oxidation, e.g., coenzyme Q, tocopherol and vitamin K. by the use of specific assays to locate in this chain the general and perhaps the exact site at which an inhibitor acts. The quantitative effect of the purified inhibitor on succinoxidase is presented in Figure 4. These data show that the inhibition is linear to about the 50% level. The maximum level of inhibition is less than 100%, although, in other experiments the 100% level has been reached. The primary portion of the succinoxidase system, succinic dehydro- genase, does not appear to be the area in which the Hydra inhibi- tor operates ( Table 6 ) . Here, with 48 times the amount of material needed to inhibit succinoxidase 50%, there is less than 20% inhibition. The terminal portion of the succinoxidase chain, cytochrome oxi- dase, is not inhibited at all by 22 times the 50% inhibitory level for succinoxidase (Table 7). The next subsystem that we studied was succinate-cytochrome-c reductase. This system probably includes a EDWARD S. KLINE 161 .04 .08 .12 .16 .20 .24 .28 ML. OF PURIFIED INHIBITOR 32 .36 Fig. 4. The inhibitory effect of the purified inhibitor on the succinoxidose activity of mouse liver homogenate. In both Curve A and Curve B, the flasks contain 0.1 ml. of liver homogenate (5 mg. of wet tissue per flask) and the purified inhibitor (2.9 /;g. of protein nitrogen per ml. of inhibitor). Total volume in all flasks: 3.1 ml. Each curve represents separate experiments (6). TABLE 6 Effect of the purified inhibitor on succinic dedydrogenase Material Inhibitor to Succinic dehydrogenase Inhibition tissue ratio activity Hg.: mg. Q02 (wet weight) % Mouse liver homogenate Mouse liver homogenate plus purified inhibitor ( 100 ^g. protein N ) 2.5 : 1 4.0 3.3 17.5 Each vessel contained 40 mg. of aqueous liver homogenate in final volume of 2.9 ml. 162 THE BIOLOGY OF HYDRA : 1961 TABLE 7 Action of the purified inhibitor on cytochrome oxidase Material Inhibitor to Cytochrome Inhibition tissue ratio oxidase activity fig.: nig. Qoo (wet weight) % Mouse Hver homogenate ■ 96 Mouse liver homogenate plus purified inhibitor (1.16 /ig. protein N) 1.16 : I 110 0.0 Each vessel contained 1.0 mg. of aqueous liver homogenate in final volume of 2.9 ml. flavin moiety and components of the intermediary portions of succinoxidase, terminating at cytochrome c. The exact nature of this system in the enzyme preparation we used is not thoroughly understood. If the purified inhibitor is preincubated with the mouse hver homogenate, we find significant inhibition of succinate-cyto- chrome-c reductase. The inhibition, although less than with succin- oxidase, is pronounced and occurs with the concentration of inhib- itor which produces 50% reduction of succinoxidase (Table 8). Our overall results in this area of the investigation lead us to postulate that the inhibition is specific [based on criteria of Keilin and Hartee ( 4 ) and Slater ( 22 ) ] and that the reduction of cyto- chrome c is blocked, that is, the inhibition occurs on the substrate side of cytochrome c. Fmthermore, since there is no evidence for TABLE 8 Effect of the purified inhibitor on succinate-cytochrome-c reductase Material Inhibitor to Reductase activity Inhibition tissue ratio ( cyt. c reduction at 550 m/x ) fig.: mg. fi moles in 10 min. % Mouse liver homogenate 0.0237 ■ Mouse liver homogenate plus purified inhibitor (0.0116 /ig. protein N) 0.058 : 1 0.0209 12 Mouse liver homogenate plus purified inliibitor ( 0.0309 /ig. protein N) 0.174 : 1 0.0094 60 Aqueous liver homogenate preincubated with purified inhibitor for 30 minutes at room temperature. Each cuvette in assay had 0.2 mg. of homogenate. Volume in each cuvette = 3.0 ml. EDWARD S. KLINE 163 direct inhibition of cytochrome b or any component preceding cyto- chrome b, we postulate that the inhibition occurs between cyto- chrome b and cytochrome c, ( based on scheme in Fig. 3 ) . This part of the system is not thoroughly characterized and I will not attempt to expand further on the above conclusions, except to add that these conclusions are similar in many ways to those drawn by Slater in his work with BAL (21, 22, 23), by Potter and Reif with antimycin A (18, 19, 20), and by Lightbown and Jackson with 2-heptyl-4- hydroxy-quinoline N oxide ( 12 ) . The purified inhibitor from Hydra has toxic effects on the mouse and the fiddler crab (5). When injected with the inhibitor (10 micrograms per gram body weight) most or all of the fiddler crabs became sluggish and about one-half of them lost their ability to right themselves, when placed on their backs. Eventually some of the animals died but most recovered. These effects were opposed to those of the boiled inhibitor, with which little or no adverse affects were noticed. We have done only a small number of experi- ments of this kind, with only 5 to 6 animals in each group, thus, we cannot really say much about the toxicity except that it occurs. Based on the amount of inhibitor required to elicit a discernible response in both the mouse and crab it does not appear that this material can account for more than a portion of the potent effect of toxic material present in Hydra. Welsh and Frock (25) have found tetramethylammonium in Hydra littoralis. If this compound is present in the nematocysts I expect that it may account for a large measure of the toxins potency. Hydroxijindoleamines — Both hydroxyindoles and hydroxyin- doleamines have been demonstrated in various coelenterates ( 14-17, 24, 26). One of these reports contained studies on H. oligactis (26). In it, Welsh showed the presence of significant amounts of 5-hydroxytryptamine (serotonin) in homogenates of this animal. Dr. Weissbach and I have found high concentrations of a 5-hy- droxytryptamine in Hydra littoralis. We induced nematocyst dis- charge by electric shock and compared the amounts of 5-hydroxyin- doleamine in the Hydra medium with that present in the whole animal (Table 9). This experiment showed that the discharged hydroxyindoleamine was present in more than 10 times the con- centration than was found in the whole animals. Several attempts 164 THE BIOLOGY OF HYDRA : 1961 TABLE 9 Hydroxyindoleamine distribution in fractions from Hydra littoralis Preparation Hydroxyindoleamine concentration fig. per ml. of preparation ^g. per gram dry tissue Whole Hydra homogenate 2.56 52 Hi/dra culture medium— before shock 0.011 57 Hydra culture medium— after shock 0.27 534 Conditions for the shock experiment similar to those used for experiment shown in Table 4. were made to determine if the amine was serotonin. We have dem- onstrated a 5-hydroxyindoleamine by paper chromotography but as yet have not been able to obtain a sufficiently clean extract to determine precisely the identity of the compound. We do have indirect evidence that the compound is serotonin rather than bufo- tenine. This evidence consists of partition coefficients between ether and an alkaline aqueous phase. A 5-hydroxytryptophan decar- boxylase is present in Hydra but this only shows that the animal can synthesize serotonin, not that serotonin is there. We are inclined to believe that serotonin is present in Hydra littoralis, but direct proof is still lacking. CONCLUDING STATEMENTS This study represents but a start toward the elucidation of the chemical composition of the nematocyst. It will be of interest not only to further characterize such preparations as the ones we have studied, but also to separate and compare the various types of nematocysts present in hydra, as well as the components of each nematocyst type. This information, coupled with the excellent morphology studies that are being carried out in various labora- tories, could form the basis for an understanding of the manner in which these intriguing organoids develop and function. EDWARD S. KLINE 165 REFERENCES 1. Glaser, O. C, and C. M. Sparrow. 1909. The physiology of neniatocvsts. /. Exp. Zool. 6: 361-382. 2. Hyman, L. H. 1940. The Invertehraies: Protozoa through Ctcnophora. McGraw- Hill Book Co., Inc., New York, p. 382. 3. Johnson, F. B., and H. M. Lenhoff. 1958. Histochemical study of purified Hydra nematocysts. /. Hist. Cytochem. 6: 394. 4. Keilin, D., and E. F. Hartree. 1949. Activity of the succinic dehydrogenase — cytochrome system in different tissue preparations. Biochem. J. 44: 205-218. 5. Kline, E. S., and V. S. Waravdekar. 1959. Toxic effects of a material isolated from Hydra littoralis. Amer. Soc. Pharmacol. Exp. Therap. 1: 62. 6. Kline, E. S., and V. S. Waravdekar. 1960. Inhibitor of succinoxidase activitv from Htjdra littoralis: J. Biol. Chem. 235: 1803-1808. 7. Kline, E. S., and V. S. Waravdekar. 1960a. On the site of action of a suc- cinoxidase inhibitor from Hydra. Fed. Proc. 19: .35. 8. Lane, C. E., and E. Dodge. 1958. The toxicity of Physalia nematocysts. Biol. Bull. 115: 219-226. 9. Lenhoff, H. M., E. S. Kline, and R. Hurley. 1957. A hydroxyproline-rich, intracellular, collagen-like protein of Hydra nematocysts. Biochem. Biophys. Acta. 26: 204-205. 10. Lenhoff, H. M., and E. S. Kline. 1958. The high imino acid content of the capsule from Hydra nematocysts. Anat. Rec. 130: 425. 11. Lenhoff, H. M., and J Bovaird. 1961. A quantitative chemical approach to problems of nematocvst distribution and replacement in Hydra. Develop. Biol. 3: 227-240. 12. Lightbown, J. W., and F. L. Jackson. 1956. Inhibition of cytochrome systems of heart muscle and certain bacteria by the antagonists of diliydrostrepto- mycin: 2-alkyl-4-hydroxyquinoline N-oxides. Biochem. J. 63: 130-137. 13. LooMis, W. F., and H. M. Lenhoff. 1956. Growth and sexual differentiation of hydra in mass culture. /. Exp. Zool. 132: 555-568. 14. Mathxas, a. p., D. M. Ross, and M. Schachter. 1957. Identification and dis- tribution of 5-hydroxytryptamine in a sea anemone. Nature 180: 658-659. 15. Mathias, a. p., D. M. Ross, and M. Schachter. 1960. The distribution of 5- hydroxytryptamine, tetramethylammonium, homarine, and other substances in sea anemones. /. Physiol. 151: 296-311. 16. Phillips, J. H. 1956. Isolation of active nematocysts of Metridium senile and their chemical composition. Nature 178: 932. 17. Phillips, J. H., and D. P. Abbott. 1957. Isolation and assay of die nematocyst toxin of Metridium senile fimhriatum. Biol. Bull. 113: 296-301. 18. Potter, van R., and A. F. Reif. 1952. Inhibition of an electron transport com- ponent by antimycin A. /. Biol. Chem. 194: 287-297. 19. Reif, A. F., and van R. Potter. 1953. Studies on succinoxidase inhibition: 1. Pseudoreversible inhibition by a napthoquinone and by antimycin A. ;. Biol. Chem. 205: 279-290. 20. Reif, A. F., and van R. Potter. 1954. Oxidative pathways insensitive to anti- mycin A. Arch. Biochem. 48: 1-6. 21. Slater, E. C. 1948. A factor in heart muscle required for the reduction of cytochrome c by cytochrome h. Nature 161: 405-406. 22. Slater, E. C. 1949. The action of inhibitors on the system of enzymes which catalyze the aerobic oxidation of succinate. Biochem. J. 45: 8-13. 166 THE BIOLOGY OF HYDRA : 1961 23. Slater, E. C. 1949. A respiratory catalyst required for the reduction of cyto- chrome c by cytochrome b. Biochem. }. 45: 14-30. 24. Welsh, J. H. 1955. On the nature and action of coelenterate toxins. Deep Sea Research, Suppl. 3: 287-297. 25. Welsh, J. H., and P. B. Prock. 1958. Quaternary ammonium bases in the coelen- terates. Biol. Bull. 115: 551-561. 26. Welsh, J. H. 1960. 5-Hydro.xytrytamine in coelenterates. Nature 186: 811-812. DISCUSSION WELSH: Were these serotonin values on a dry weight or wet weight basis? KLINE: Dry weight basis. WELSH: What do you say about the heat and pH stabihty of the succinoxidase inhibitor? KLINE: It is stable at pH 5.8 and 8, and since one step in the purification of the inhibitor is a pH 4 precipitation it has appre- ciable stability even at this pH. Heat stability is an interesting point. We felt that the inhibitor from Hydra littoralis could have been a phospholipase A. Phospho- lipase As are heat stable and the succinoxidase inhibitor in snake venom is believed to be this enzyme. We heated separately some Crotalus adamanteus venom and our inhibitor at pH 5.8 in a boil- ing water bath for 15 minutes. The venom lost none of its effec- tiveness against succinoxidase while the purified inhibitor from Hydra lost about 75% of its activity. EAKIN: What is its behavior on dialysis? KLINE: Essentially all of the activity is non-dialyzable. LENHOFF: Does the inhibitor do anything to mitochondria? KLINE: We have done one or two preliminary studies and there seems to be some effect on the mitochondria, but as yet we have done too little to make any definite statements. MARTIN: Have you ever tried to extract active substances from the nematocyst-poor parts of the Hydra? And if so, did they show any similarity with the nematocyst content? EDWARD S. KLINE 167 KLINE: Which component? Serotonin? MARTIN: Serotonin or the enzyme inhibitor. KLINE: We have not done that. I beheve that the best proof for the locahzation of these compounds will come when we can quantitatively isolate pure, undischarged nematocysts from the animal. ROSS: I'm very interested and pleased to see your results with serotonin. But I'd like to hear your comments on some observations that Mathias, Schachter, and I made in London on the distribution of serotonin in sea anemones, because our results would indicate that we cannot extend this conception generally over a whole group from one species unless one looks at the distribution very care- fully. We found extracts from tentacles separated from the column, or both separated from the tissues lining the coelenteron did not contain much serotonin in 3 of the 4 species of sea anemones that we used, viz., Metridiiim senile, Actinia equina and Anemonia sulcata. The only place where we found a significant amount of serotonin was in the "coelenteric tissue" of CalUactis parasitica, and there it was present in large quantities, 500-600 mg. per gram of freeze-dried matter. This was about 60 times the concentration found in the tentacles. Thus there seemed to be no correlation between the distribution of serotonin and nematocysts, or be- tween different species. I wonder if you have any comment to make on that? KLINE: I am aware of your work and it might appear that the findings in various laboratories are contradictory. But as you have said, we cannot necessarily extend results from one animal to an- other. I believe Phillips thought his hydroxyindoleamine was bufotenin rather than serotonin and that it was not localized in the nematocysts. Your group found serotonin in certain anemones and not in others and you feel that it is not concentrated in the nemato- cysts. Is this correct? ROSS: Well, it's in a part of the animal where there are fewer nematocysts. KLINE: Dr. Welsh's study with anemones points to it being serotonin and in the nematocysts. For the most part we all have 163 THE BIOLOGY OF HYDRA : 1961 been studying different animals with different approaches. As time passes I become more impressed by the variabiHty between closely related animals. PHILLIPS: Have you detected any hexosamines or uronic acids in the capsule? KLINE: We have not looked for them. STREHLER: What percentage of the total weight did you calcu- late would be collagen on the basis of this hydroxyproline content? KLINE: Based on 20% hydroxyproline, the collagen-like protein represents about 10% of the total protein of H. littoralis. Physalia Nematocysts and their Toxin Charles E. Lane Institute of Marine Sciences, University of Miami, Miami, Florida Nematocysts in Physalia are widely distributed through the epithehum clothing most of the members of the colony. These organelles are formed in cnidoblasts by so far imdescribed cyto- genetic processes. The upper surface of the float and the proximal portions of the gastrozooids and of the fishing tentacles are relatively deficient in mature nematocysts. Over the surface of the fishing tentacle cnidoblasts are concentrated in the epithelium clothing the batteries. These are permanent structures distributed in bead- like fashion along the length of one edge of the tentacle, and they are illustrated in Figure 1, which shows a three-dimensional re- construction of a segment of the fishing tentacle of Physalia. The batteries appear as discrete saccular enlargements along one edge. A longitudinal section through a portion of the tentacle, including a single battery cut equatorially appears in Figure 2. The battery is lined by gastrodermis continuous with that lining the gastrovascular extension in the tentacle. The mesoglea is a thick band of fibrous connective tissue external to the gastro- dermis. The epidermal layer bearing cnidoblasts clothes the entire structure. At the equator of the battery the epidermis thickens abruptly where the external hemisphere acquires its population of mature cnidoblasts. Perhaps the most outstanding histological char- acteristic of this epithelium is the regular distribution through it of nematocysts belonging to two different size groups. The total thickness of the epithelium is just sufficient to clothe the large nematocysts, which range from 25 to 30 microns in diameter. 169 170 THE BIOLOGY OF HYDRA : 1961 Fig. 1. Reconstruction in wax of a segment of the fishing tentacle of Physalia. The extension of the gastro-vascular cavity into the tentacle and the relationship between the cavities of the batteries and of the tentacle are clearly shown. Regularly spaced between cnidoblasts bearing the large nemato- cysts occur small cnidoblasts whose capsules range from 7 to 15 microns in diameter. In favorable preparations each of the cni- doblasts may be seen to be provided with a cnidocil, projecting through the cuticular layer of the epithelium into the ambient water. The light microscope reveals a perinuclear basketwork of elastic fibers within the cnidoblast, which appears to surround the nema- tocyst. Other than this perinuclear network and the nematocyst capsule, the cytoplasm of the definitive cnidoblast appears to pre- sent very little structural specialization. Cnidoblasts are regularly distributed throughout the epidermis of the external battery hemisphere, and there is also a repeating pattern of internal structure in adjacent nematocysts. The internal coiled thread, characteristic of the nematocysts of all Cnidaria, in each of the nematocysts originates at about the same point in the CHARLES E. LANE 171 l4» Fig. 2. Frontal Section through the fishing tentacle of Physalia X 200. The gastrovascular cavity of the tentacle communicates in the center of the field with the cavity of a single "battery." Hypertrophy of the gastrodermis begins at the equator of the battery. capsule and coils in the same clock-wise direction in approximately the same plane. If the surviving tentacle be stimulated by gradually increasing the concentration of solutes in the surrounding water, the nema- tocysts may be made to discharge. This is a dramatic, explosive process, the nematocyst threads being hurled from the capsule with sufficient force to penetrate the surface film. This observation explains our early experience of being severely stung even through a surgical glove. In our laboratory we isolate surviving nematocysts by con- trolled autolysis at 4", followed by screening, sieving, washing, and settling. The washing process is continued until the wash water is no longer toxic when injected into the hemocoele of the fiddler crab. This point may acquire some significance when one attempts to compare the activity and biochemistry of the toxin 172 THE BIOLOGY OF HYDRA : 1961 prepared in our laboratory with reports in the hterature describing the activity of other Physalia toxin preparations. Earher investiga- tors, ahnost uniformly, have homogenized and extracted entire tentacle material. It will later appear that there are active extra- nematocyst substances present in the tentacle; however, the bio- chemistry and pharmacology of these materials has not been studied in our laboratory. Isolated nematocysts may be concentrated by settling and de- cantation of the supernatant water. They do not survive centri- fugation without discharge so it is necessary to permit them to settle by gravity alone. The putty-like concentrate resulting from our procedure is virtually free of tissue fragments and contains very few (less than 1%) discharged nematocysts. The concen- trated nematocysts are frozen and stored in the deepfreeze where they retain their reactivity for periods of at least four years. Surviving nematocysts are homogenized in an all-glass homog- enizer in a minimum volume of distilled water for about twenty minutes, or until an aliquot shows no more than 10% unbroken capsules. The resulting brei is centrifuged at 4° in a refrigerated centrifuge at 15,000 > gravity; the residue is resuspended in a TABLE 1 Amino acids in an acid hydrolysate of the crude toxin Amino ^xA//Sample Relative Acid Concentration Alanine 0.37 5 Arginine 0.12 1 Aspartic Acid 0.32 7 Glutamic Acid 0.85 20 Glycine 0.75 9 Histidine 0.05 1 Isoleucine 0.19 4 Leucine 0.25 5 Lysine 0.25 6 Phenylalanine 0.15 4 Proline 0.38 7 Serine 0.23 4 Threonine 0.07 1 Tyrosine Valine 0.21 4 CHARLES E. LANE 173 minimum of water and recentrifuged. The supernatant solutions from these two centrifugations are combined and lyophiUzed. The lyophihzed "crude" toxin has regularly assayed between 15 and 16'i nitrogen by micro-Kjeldahl. All tests for polysaccharide have been negative. A sample of crude toxin was hydrolyzed in 6N HCl, and the hydrolysate was analyzed on the Beckman Spinco amino acid analyzer with the results shown in Table I. The lyophihzed toxin is lethal to mice at dosage levels of 1.7 mg. kilogram. When crude toxin was chromatographed one-dimensionally with 80^ »-propanol as the solvent system, a series of nine spots ap- peared when the paper was developed with ninhydrin. Each of the spots was separately eluted and assayed for total activity in the fiddler crab, Uco piigilafor. Four of the spots accounted for 95% of the total biological activity of the crude toxin. The active regions on the chromatographic papers were eluted, hydrolyzed, and rechromatographed. Each was shown to contain more than one amino acid. Since this chromatography had been accomplished in the pre- sence of a solvent and at room temperature it was felt that consid- erable loss of activity may have occurred. Such an attenuation might be sufficient to mask activity in other fractions. Accordingly, the crude toxin was next fractionated on the Beckman refrigerated paper curtain electrophoresis apparatus, using phthallate buffer pH 5.8 at 2". Four fractions were separated; after dialysis and lyophili- zation they were carefully diluted to their relative concentration in the original toxin and bioassayed on Uca pugilator. The results are shown in Figure 3. One peptide nearly equals the activity of the original whole toxin, although representing less than 10% of its weight, and it therefore appears that some inert masking protein materials may have been removed by electrophoresis. Physalio toxin, therefore, appears to he a relatively simple protein consisting of only a few toxic peptides. Our future studies will seek to describe the precise molecular configuration of these peptides and to relate biochemical structure to pharmacologic activity. I may be permitted to speculate briefly about the origin and synthesis of Physalia toxin. The gastrodermis lining the battery 174 THE BIOLOGY OF HYDRA : 1961 Toxin Fraction Fig. 3. Results of bioassay on groups of 10 ilea pugilator of: (0) whole crude toxin, and various fractions (1, 2, 3, 4) separated by electrophoresis from the crude toxin. Fractions were injected at levels approximating their separate concentration (by weight) in the crude toxin. undergoes characteristic hyperplasia beneath that portion of the epidermis containing mature nematocysts. This histological change involves structural polarization, extensive vacuolation of the cyto- plasm, and a change in the staining characteristics and chromatic density of the nucleus (see Fig. 2). The mesoglea separating the hyperplastic gastroderm cells from the nematocyst-containing epi- dermis is also modified. In preparations stained with Mallory's trichrome, the mesoglea shows discrete circular patches which stain differently from the rest of the mesoglea. These patches are always located between hypertrophied gastroderm cells and cnidoblasts in the surface epithelium. We have shown the gastrovascular cavity of Physalia to contain and circulate a protein fluid. There is open communication between the gastrovascular cavity and tlie cavity of the battery. It is tempting to speculate that the modified gastrodenn cells basal to the cnido- CHARLES E. LANE 175 blasts absorb precursor materials from the circulating gastrovascular fluid and from these synthesize the toxin which they subsequently secrete through the mesoglea and into the cnidoblasts. Slautterback and Fawcett have shown that the nematocyst thread in hydra originates from outside the nematocyst capsule and is subsequently introduced into the cavity of the nematocyst. If a structural component of the nematocyst may be formed external to the nematocyst capsule and subsequently introduced into it, it should not stretch our credulity too far to accept the suggestion that a soluble protein toxin may be synthesized outside the nema- tocyst and later may pass through it. One disturbing observation is that this would suggest or almost require that the nematocyst capsule be permeable to the toxin. We have repeatedly observed that the toxin does not leach from surviving purified nematocysts. Presumably, therefore, there is one stage in the morphogenetic history of the nematocyst when the capsule wall may be permeable to toxin but in the mature nematocyst these permeability relation- ships may be completely changed. I suggest that Phijsalia toxin is synthesized by gastrodermal cells, passes through the mesoglea, and into the nematocyst during the morphogenesis of this structure. DISCUSSION CROWELL: Where are the nematocysts manufactured in Phijsa- lia? Where is the differentiation of the cnidoblasts taking place? LANE: I can't answer because I don't know yet. I can tell you a few of the things we do know. In adult animals the float is generally free of cnidoblasts. The basal ends of gastrozoids are deficient in cnidoblasts. Cnidoblasts appear to be reasonably uni- formly distributed throughout the length of the fishing tentacle. I have seen no clear histological evidence of interstitial cells such as we have heard about in hydra. Obviously they must be there, but I haven't seen them. CROWELL: Is it possible that the whole tentacle is continuously growing so that it's always young basally and degenerating api- 176 THE BIOLOGY OF HYDRA : 1961 cally? If so, there is no need to replace the nematocysts along the length of the tentacle. LANE: I think this is entirely possible. CROW ELL : Two other possibilities are that nematocysts are made all along the tentacles, and that they are built back in headquarters and are transported to the tentacles by unknown means. LENHOFF: In the chromatograms you showed, were you run- ning the entire fluid or a hydrolyzate of the fluid? LANE: This was the entire material. We took the entire gastro- vascular fluid without any treatment. I suspect we have amino acids and peptides. We are now analyzing this fluid using paper electrophoresis. LENHOFF : What was the solvent? LANE : n-propanol. LENHOFF: In H. littoralis we find that the gastrodermis takes up mostly particles, and leaves behind the free amino acids in the gut. Your chromatogram looks somewhat like a normal pattern of free amino acids. Do you think Physalia does the same thing? LANE: We'll know more about this very soon. We find that toxin peptides distribute very much like this and we have eluted, hydro- lyzed, and rechromatographed them separately. We know they are peptides. So without actually having done it on this gastrovas- cular material, I feel fairly certain that these are peptides also. WOOD: Do you have any real evidence that the gastrodermis extrudes materials into the mesogleal extracellular space, which are then picked up by the epidermal cells? I question this because it seems to me that it would be more efficient to transfer such materials directly. This bears on whether your specialized area in the mesoglea is cellular or is purely connective tissue? LANE: That was the way we had interpreted it, but this is purely tentative and subject to change. Having seen the way in which both endodermal and ectodermal processes interdigitate and weave their way through the mesoglea in hydra, we could easily CHARLES E. LANE 177 expect the same thing to take place here. It may be that these tremendously hypertrophied endodermal cells penetrate through the mesoglea in these regions. MARTIN: I want to mention an experiment which supports Dr. Lane's hypothesis. We didn't work with Phy.salia or hydra, but with Anthopleura elcgantissima. We separated the tentacles from the column and took the mesenteries out. Then we ground up the column and tentacles separately, made extracts and mea- sured their toxicity by injecting them into mice and we found that the extracts of the nematocyst-poor column was as toxic as a crown, which is nematocyst-rich. HAND: Did you remove the mesenterial filaments? MARTIN: Yes. HAND: Fine. MARTIN: By the way, the mesenteric filaments were less toxic than the other two fractions. HAND: That's quite contrary to what I would have expected, since they have the bulk of the internally located nematocysts. SLAUTTERBACK: There is a mesogleal formation in hydra somewhat similar to the specialized areas you described. A great accumulation of mesoglea is sometimes seen under the pedal disk secretory cells. The predominant component here is amphorous. Whether it is the same material commonly found in the mesoglea has not been determined. The fine filaments and glycogen granules are not increased as much. Also, is the greatly enlarged part of the hypertrophied gastro- derm cells an enlarged "central vacuole "? LANE: Yes. SLAUTTERBACK: Were you ever able to fix anything in that vacuole? LANE : We've tried a wide spectrum of fixatives on these vacuoles but they've always been clear. BURNETT: Have you ever found nematocysts in the gastroder- 178 THE BIOLOGY OF HYDRA : 1961 mal cells suggesting that there may be a migration through the gastrovascular cavity? LANE: Yes, I've found them in gastrodermal cells, but they have been in cells which have incorporated this material from prey organisms. BURNETT: Are they in a state of digestion? LANE: That's right. Normally the nematocysts end at the lip of the gastrozoid. The lip is always identifiable by having cnidoblasts in its ectoderm, but none in its gastroderm. Actually, we've had a great deal of trouble in keeping these animals in captivity. Probably the reason for this is that they have no protection against dragging their tentacles on the bottom. When- ever this happens the fragile surface epithelium is destroyed so that the next time the tentacle contracts, it squirts out some of this gastrovascular fluid. It's interesting that within an hour of placing a mature PhysoUa in an ordinary aquarium tank, the surrounding water becomes ninhydrin-positive. He loses much fluid. This is one reason why we have been unable to keep these animals in captivity long enough to feed them, and then study the distribution of digested food materials to the gastroderm. LARSON What can you tell us about the pharmacological action of the toxin? LANE: We haven't enough information on the pharmacology of the toxin to justify any statements. GOREAU: Can rabbits be immunized against Physalia toxin? LANE: Yes. The material is sufficiently antigenic to develop good titers. It is difficult, however, to difl^erentiate between a lethal and an immunizing dose. GORDEAU: That's the problem of anaphylactic shock which was discovered with Physalia toxin by Richet. If you could immunize an animal against the toxin and label the antibodies with suitable fluorescent groups it might be possible to find out whether there is transfer of toxin from the gastroderm through the mesoglea into the epidermal nematocyst batteries. LANE : Yes, that would be an interesting experiment. Compounds of Pharmacological Interest in Coelenterates John H. Welsh Biological Laboratories, Harvard University, Cambridge, Massachusetts The nematocysts of coelenterates appear to serve two principal functions: one, a means of protection, the other, a role in feeding. A person once badly stung by Physalia, Cijanca or certain of the cubomedusae avoids contact with one of these a second time. It may be assumed that an animal that is stung and survives also may avoid future contact with a coelenterate if it is capable of learning. More important, perhaps, to the coelenterates, is the paralyzing or relaxing action of the contents of the nematocysts when injected into their prey. Since the very early years of this century, efforts have been made to identify the substances in coelenterates that are responsible for the symptoms that result from their sting. In most of the earlier work, extracts of whole coelenterates or of nematocyst-bearing parts (tentacles and acontia) have been used. Therefore, it has not been possible to attribute an observed action to nematocyst con- tents. The recently developed methods of isolating clean nemato- cysts will obviate this difficulty if it can be shown that they lose none of their contents during the isolation procedure. A condensed and incomplete summary of substances or frac- tions obtained from various coelenterates follows. Some of these derive from nematocysts; others, almost certainly, do not. I. Early attempts to isolate toxic components of coelenterates by Richet and Portier (19, 20, 21) yielded three active extracts: "thallasin," "congestin" and hypnotoxin." None of these was chem- ically identified and each was doubtless a mixture of substances 179 180 THE BIOLOGY OF HYDRA : 1961 (see refs. 13, 22, 23, 25, for summaries of this and other earHer work ) . II. Quarternary ammonimii compomids: Several nitrogenous bases have been isolated from various coe- lenterates, including the following: „ . ° ° References tetramethyl ammonium hydroxide or "tetramine" 4, 11, 17, 27 N-methylpyridinium hydroxide 5 homarine 2, 10, 17, 27 trigonelline 2, 27 y-butyrobetaine 1, 27 zoo-anemonin 3, 6, 27 Of these bases, the only one that has marked paralyzing action is tetramine (4, 27). It is the only one fomid thus far in a fresh- water coelenterate (27). It is a known toxic component of certain molluscan tissues (7, 9). With the exception of zoo-anemonin, the other bases listed above are widely distributed among marine invertebrates where they may play a role in osmoregulation ( 10, 27 ) . III. 5-Hydroxytryptamine (serotonin, 5-HT) : This very potent pain-producer and histamine releaser has now been identified in a variety of coelenterates (17, 25, 26). It is pres- ent in the coelenteric tissues of Calliactis parasitica in very large amounts ( 17 ) but in other coelenterates it is most abundant in regions (tentacles and acontia) where nematocysts are concentrated (25,26). IV. Histamine and histamine releasers : Histamine has been found in some coelenterates but not in others (17, 24). Potent histamine releasers have been extracted from a sea anemone ( 12) and Cyanea (24) . V. Active proteins: Much evidence indicates that the paralyzing and edema-pro- ducing actions of coelenterate toxins are due, in large measure, to a protein component(s) (8, 13, 14, 15, 16, 21, 22, 23). There is some evidence that this component acts on cholinergic neurons JOHN H. WELSH 181 in such a manner as to block conduction and or transmission ( 17 ) . The neutrahzing action of certain acetylchohne blockers such as tetraethylammonium (TEA) on the paralyzing action of ten- tacle extracts supports this view (25). Certain of the symptoms that follow a coelenterate sting such as pain, burning, itching, localized edema and hemorrhaging could result from injected 5-HT (a potent pain producer and his- tamine releaser), from histamine itself, and from other histamine releasers. These sul^stances, however, cannot be responsible for the paralyzing action of the nematocyst contents. Many quaternary ammonium compounds do have a paralyzing action as junctional blocking agents. Of those listed above, only tetramine can qualify as a candidate for the paralyzing action. In the first place it is the only one that has been identified in hydra extracts, while most of the others are widely distributed among the marine invertebrates. In the second place, tetramine is an effective poison and is the toxic component of the salivary glands of certain marine gastropods (7, 9), while the others are surprisingly non-toxic (cf. 27). Fur- thermore, the earlier observed antagonism of coelenterate extracts by tetraethylammonium chloride or Banthine (25) strongly sug- gests that a methylated quaternary nitrogen compound is, in some way, involved in the paralysis resulting from a coelenterate sting. However, calculations may be made that indicate that there is not enough tetramine, in the extracts that we have used, to account for their paralyzing action, at least on arthropods. Evidence has been accumulating over the years that the paralyz- ing factor in coelenterate toxins is a protein or group of proteins. Several recent studies show that toxicity remains after dialysis but is destroyed by boiling and by treatment with certain proteolytic enzymes (14, 15, 16, 18). The exact mode of action of the toxic protein ( s ) is not yet clear. RfiSUME OF SOME EXPERIMENTS THAT ARE CURRENTLY IN PROGRESS We are, at present, comparing the actions of homogenates of Metridium acontia and whole Hydra, and of material discharged 182 THE BIOLOGY OF HYDRA : 1961 from their nematocysts by electrical stimulation, on Carcinus mae- nas, Uca ptigilator and several species of cockroaches. A brief resume of some of the experiments and tentative results follows: 1) The minimum lethal dose of a homogenate of Hydra ameri- cana, in terms of the number of Hydra injected is between 5 and 10 Hydra for Carcinus weighing 20-30 gms.; 2-3 Hydra for Uca weigh- ing 4-5 gms.; and about 5 Hydra for female Bryostria sp. (cock- roach) weighing 4-5 gms. These are doses that usually kill in from 1 to 24 hours. The average dry weight of Hydra amcricaua, reared in the laboratory, is about 35 [xg. If the paralyzing factor constitutes something like 0.1% of the total dry weight, it appears that 0.2-0.4 fig. of toxic substance is lethal for a 20-30 gm. green crab. 2) Heating a Hydra homogenate for 30 min. at 100 "^ results in complete or nearly complete loss of paralyzing action. 3) Electrical stimulation of numbers of Hydra (200-300) in a minimum volume of distilled water discharges many of the nema- tocysts. Injection of a small volume (0.05 ml.) of the fluid sur- rounding the Hydra into Uca, produces symptoms that are quali- tatively like those seen when whole Hydra homogenate is injected. 4) Hydra, and Mctridiiim acontia, have been homogenized in 1.0% tetraethylammonium chloride (TEA). When volumes are in- jected known to contain minimum lethal doses of Hydra or acon- tia, none of the characteristic symptoms develop and most test animals survive indefinitely. This agrees with earlier observations on the autotomy reflex in brachyurans when it was found that TEA very effectively antagonized the effects of coelenterate ex- tracts (25). If the TEA is blocking the action of a toxic protein component, and not tetramine only, this may provide a clue to the mode of action of the toxin. REFERENCES 1. AcKERMAN, D., 1927. tjber die IdentiUit des Atkinins mit dcm 7-Butyrobetaine. Zeitschr. f. Biologie, 86: 199-202. 2. AcKERMAN, D., 1953. Uber das Vorkommeu von Homarin, Trisonellin und einer neuen Base Anemonin in der Anthozoa Ancmonio sulcata. Zeitschr. f. physiol. Chemie, 295: 1-9. 3. AcKERMAN, D., 1954. Richtigstellung: "Zoo-Anemonin" statt Anemonin. Zeitschr. f. physiol. Chemie, 296: 286. JOHN H. WELSH 183 4. AcKERMAN, D., F. HoLTZ, aiid H. Reinwein, 1923. Reindarstellung und Kon- stitutionsermittelung des Tetramines, eines Giftes aus Aktinia equina. Zeitschr.f. Biologie, 79: 113-120. 5. AcKEKNiAN, D., F. HoLTZ, and H. Reinwein, 1924: tjber die ExtraktstofFe von Aktinia equiyui. Zeitschr. f. Biologie, 80: 131-136. 6. AcKERMAN, D., and P. H. List, 1960. Zur Konstitution des Zooanemonins und des Herbipolins. Zeitschr. f. physiol. Chemie, 318: 281. 7. AsANO, M., and M. Itoh, 1960. Salivary poison of a marine gastropod, Neptunea arthritica Bemhardi, and the seasonal variation of its toxicity. Ann. N.Y. Acad. Sci., 90: 674-688. 8. Cantacuzene, J., and A. Damboviceanu, 1934. Caracteres physico-chimiques du poison des acconties d'Adamsia palliata. C. R. Soc. Biol., Paris, 117: 138-140. 9. Fange, R. 1960. The salivary gland of Neptunea antiqua. Ann. N. Y. Acad. Sci., 90: 689-694. 10. Gasteiger, E. L., p. S. Haake, and J- A. Gergen, 1960. An investigation of the di.stribution and function of homarine ( N-methyl picolinic acid). Ann. N. Y. Acad. Sci., 90: 622-636. 11. Haurowitz, F., and H. Waelsch, 1926. Uber die chemische Zusammensetzung der Qualle Velella spirans. Zeitschr. f. physiol. Chemie, 161: 330-317. 12. Jacques, R., and M. Schachter, 1954. A sea anenome extract (thalassine) which liberates histamine and a slow contracting substance. Brit. J. Pharmacol., 9: 49-52. 13. Kaiser, E., and H. Michl, 1958. Die Biochemie der tierischcn Gifte. Franz Deuticke, Vienna. 14. Lane, C. E. 1960. The toxin of Physalia nematocysts. Ann. N.Y. Acad. Sci. 90: 742-750. 15. Lane, C. E., and E. Dodge, 1958. The to.xicity of Physalia nematocysts. Biol. Bull, 115: 219. 16. Martrm, E. J., 1960. Observations on the toxic sea anenome, Rhodactis howesii ( Coelenterata ) . Pacific Science, 14: 403-407. 17. Mathias, a. p., D. M. Ross and M. Schachter, 1960. The distribution of 5-hydroxytryptamine, tetramethylammonium, homarine, and other sub- stances in sea anenomes. /. Physiol., 151: 296-311. 18. Phillips, J. H. Jr., and D. P. Abbott, 1957. Isolation and assay of the nematocyst toxin of Metridium senile fimhriatum. Biol. Bull., 113: 296-301. 19. RiCHET, C. 1902. Du poison pruritogene et urticant contenu dans les tentacules d'Actinies. C. R. Soc. Biol, Paris, 54: 1438. 20. RiCHET, C., 1903. Des poisons contenus dans les tentacules des Actinies, con- gestine et thalassine. C. R. Soc. Biol, Paris, 55: 246. 21. RiCHET, C., and P. Portier, 1936. Recherches sur la toxine des coelenteres et les phenomenes d' anaphylaxie. Resultats des campagnes scientifiques, Monaco 95: 3-24. 22. SoNDERHOFF, R., 1936. tJber das Gift der Seeanenionen. I. Ein Beitrag zur Kenntnis der Nesselgifte. Liehig's Ann., 525: 138-150. 23. Thiel, M. E. 1935. Uber die Wirkung des Nesselgiftes der Quallen auf den Menschen. Ergebnisse u. Fortschr. der Zoologie, 8: 1-35. 24. UvNAS, B. 1960. Mechanism of action of a histamine-liberating principle in jellyfish [Cyanea capillata). Ann. N.Y. Acad. Sci., 90: 751-759. 25. Welsh, J. H., 1956. On the nature and action of coelenterate toxins. Deep Sea Research, 3(suppl): 287-297. 26. Welsh, J. H., 1960. 5-Hydroxytryptamine in coelenterates. Nature, 186: 811. 27. Welsh, J. H., and P. B. Prock, 1958. Quaternary ammonium bases in the coelenterates. Biol Bull, 115: 551-561. 184 THE BIOLOGY OF HYDRA : 1961 DISCUSSION HAND: The extra serotonin that you find in the acontia seems reasonable in view of some very simple observations that one can make on Metridium and other acontiate anemones. They commonly eat small worms, copepods, and things of this nature. If you get a small transparent anemone, you can see that after the food is swal- lowed the prey is still kicking, wriggling and squirming. It gets into the coelenteron and the acontium coils around the animal, presum- ably the nematocysts of the acontium discharge, and this very quickly subdues it. It quivers a couple of times, and then stops. The acontia, of course, are rich in nematocysts, ROSS: Do you think that the amounts of serotonin that you find in the acontia, ca. 1 Mg./g., is significant? Compared with the amounts that we found in other parts of anemones, they seem so small that we would have dismissed them. WELSH: Well, I think if I may say so, it was unfortunate that you looked at Calliactis first. I think if you had looked at other anemones you would have viewed this situation differently. ROSS: Not at all. WELSH: Let me put it this way. If you go out and catch a vicious stinging wasp, you can get out of its venom a perfectly tremendous amount of serotonin. You measure it as 6 to 20 milligrams per gram of venom. Now if you do its nervous system, you get a few tenths of a microgram. I believe that the serotonin in the nervous system is just as important in the life of the wasp as the serotonin in its venom. The most we have in any part of our nervous system is 0.4 micrograms per gram of hypothalmus. And if the tranquilizer reserpine is doing what they say its doing, releasing serotonin, then this brings this down to a 10th of that, and here we're working in the lOths and hundredths of micro- grams per gram range. This is less than the concentration range of serotonin that one finds in acontia. ROSS: We found 600 times as much in the lining of the coelen- teron in Calliactis, so this made us think it couldn't possibly be associated with nematocyst poisons. JOHN H. WELSH 185 WELSH: But when you looked at other parts, you found that the tentacles were richer than the body wall? ROSS: A bit, but on the borderline. LENHOFF: Couldn't we view the tetramethylammonium com- pounds not as toxins, but as part of the normal nervous system trans- mitters of coelenterates since tetramine is present in all of the tis- sues assayed? I ask this question because when glutathione activates the feeding response in H. littoralis, some of the few substances that enhance the response are certain tetramethylammonium com- pounds. Possibly the transmission of the glutathione stimulus goes through a tetramine-mediated pathway rather than through an acetylchloline-mediated pathway? WELSH: I think it is entirely possible. We have no evidence on the tetramine one way or the other. However, tetramine does occur in a number of venoms; it occurs in the salivary glands of some marine gastropods in large amounts. And, of course, other choline esters, and other quaternary ammonium compounds occur in certain molluscs. But that tetramine may be taking the place of acetylcho- line in the coelenterate nervous system is a good possibility. HESS: Do these animals have choline esterase or acetylcholine? WELSH: There is choline esterase. PASSANO: I suspect that the acetylcholine esterase system is not significant in the functioning of the scyphozoan nervous sys- tem, and we know that 5-hydroxytryptamine also fails to have any effect. Could it be that the use of these substances, toxic to other animals as nematocyst toxins, might be valuable to the coelenterates because they would avoid the danger of self-inHicted paralysis? Is this even why their neuropharmacology is different from that of other animals? WELSH: Venomous animals are generally successful in keeping their venoms away from themselves. PASSANO: Well I would like to ask then, in other people^s experience in studying the feeding responses of nematocyst-bear- ing animals, are the nematocysts always prevented from penetrating 186 THE BIOLOGY OF HYDRA : 1961 the animal that possesses them? The independent effector is quite different from the effector at the end of a wasp; it is not so neatly controlled. The tentacle of a coelenterate coils around its prey. There is a great chance for nematocysts to be discharged into a tentacle. This would obviously create difficulties if the tentacle was paralyzed by its own poison. BURNETT: It is common for hydra to pierce its own tissues with nematocysts during feeding. PHILLIPS: I think the experiments of Dr. Ross, and of Dr. Martin and the ones that I did on Metridium suggest that caution should be employed in the interpretation of work using whole tissue extracts. Sometime ago, when I was working on the toxin, I detected a 5-hydroxyindole compound, which at that time I thought corresponded more closely to bufotenin. On purification I noticed that the level of 5-hydroxyindole compounds decreased steadily. In fact, pure suspensions of nematocysts contained no detectable 5-hydroxyindole compounds at all, yet the nematocysts were still capable of discharging and still possessed toxicity. WELSH: In that connection, I would be interested to know if the 5-hydroxytryptamine washed out of the nematocyst. It's a small, soluble molecule that diffuses readily through some cell surfaces. PHILLIPS: This is a possibility. But the nematocyst suspensions after purification still should show toxicity. WELSH: I don't think that the serotonin is really toxic. You can put a large amount of serotonin into a crab and it gets very nervous and jittery. An hour later it is normal. PHILLIPS: Diffusion from the nematocyst during purification, of course, is always a possibility. At the same time, nematocysts are still susceptible to osmotic discharge, so that gross permeability changes do not seem to have occurred. LOOMIS: How do you keep your nematocysts from discharging while you separate them? PHILLIPS: With high concentrations of sucrose. Present State of Nematocyst Research: Types, Structure and Function Cadet Hand Department of Zoology, University of California, Berkeley, CalifomiOi. I want to start by quoting an admirable passage from the Intro- duction of the recent paper by Burnett and colleagues (1). On page 247 they state "One of the most structurally complex and cer- tainly one of the most enigmatic organelles in the animal kingdom is the nematocyst of coelenterates. For nearly a century hosts of scientists, too numerous to mention, have concentrated their at- tentions on the mode of formation, the migration pathways, the mechanism of discharge, and the chemical nature of these unusual structures. . . ." These same authors go on to make the statement that ". . . none of these subjects of investigation has been resolved to any degree of satisfaction." In many ways this statement is accu- rate and acceptable, but I think in many ways I would disagree with the generality. A good deal is known about each of the subjects they cite and I for one have found considerable satisfaction in the numerous papers on nematocysts that I have examined. I also want to acknowledge that some of my satisfaction has come from reading the papers of Burnett and his co-workers. I want to talk today about types, structure and function of nema- tocysts. I also want to make it clear that I do not work on nemato- cysts, I work with them. My interests in them are twofold. First, nematocysts are a truly \'aluable systematic tool and many coelen- terates can be positively identified by their nematocysts alone. Not only this, but nematocysts are useful in relating higher taxa such 187 188 THE BIOLOGY OF HYDRA : 1961 as genera, families or even orders, and in the broad view even classes. Second, as a student of coelenterates I am interested in the biology of these animals, and the nematocysts are intimately in- volved in numerous aspects of the lives of coelenterates. There have been several attempts to classify nematocysts and some of the results of these have come down to us in the form of such useful and descriptive names as penetrants and glutinants. However, it was not until the elaborate system of Weill (10) was published that any real uniformity of nomenclature of nematocysts was arrived at. With the introduction of Weill's terminology some people complained that the system was too clumsy and the names too long to be useful. For example the commonest penetrant of many anthozoans could be called a hoplotelic microbasic masti- gophoric rhabdoidic heteronemic stomocnidic nematocyst, or a stenotele could be called a stenotelic rhopaloidic heteronemic stomocnidic nematocyst. In common practice, and as Weill's termi- nology is being applied, the names microbasic mastigophore (or just mastigophore) and stenotele suffice. Weill's system is only for- bidding when one first meets it, but it is a defined system which makes possible far greater accuracy in communication than any other so far devised. To use the full nomenclature, as in the exam- ples I cited, is just as absurd as to start the name of some species with the phylum name, add in the names of the class, order and family and finally tack on the specific binomial. Weill's system recognizes two categories of cnidae, spirocysts and proper nematocysts. Spirocysts are restricted to the zoantharian anthozoans while all coelenterates have nematocysts. The struc- ture and function of spirocysts are obscure. Weill (10) believes that spirocysts have but a single layered wall and it is extremely rare to see a spirocyst which has everted its thread. Cutress (5) has ar- gued rather convincingly that spirocysts are nematocysts and from his comments one could conclude that they represent a form of holotrichous nematocyst. The test of this conclusion will undoubt- edly come when a study of these cnidae is carried out with an electron microscope. The nematocysts proper have two major subdivisions, astomo- cnidae whose tubes are closed and stomocnidae with tubes open at the tip. The astomocnidae are divided in turn into two categories. CADET HAND 189 the familiar desmonemcs or volvents, and the much less familiar acrophores and anacwphorcs of the Siphonophora, which are col- lectively called rhopalonemcs and have a sac-like tube rather than the coiled or corkscrew tube of the desmonemcs. The stomocnidae show much more variety in form. They can be divided into the haploncmcs, whose tube has no enlarged basal portion or butt, and the hcteroncmcs which have a butt. Among the haplonemes we find the familiar armed holotrichs and unarmed atrichs, as well as partially armed forms we call basitrichs. These haplonemes have a thread or tube of constant diameter and are technically isorhizic. A second type of haploneme has an aniso- diametric tube which may taper or be slightly swollen near the base. These are the anisorhizic nematocysts of various siphonophores and Tubularia. The heteronemes, which you recall have a butt, can be divided into the rhahdoidcs whose butts are isodiametric and the rhopa- loides whose butts are anisodiametric. The rhabdoides can in turn be subdivided into masiifiophores with a terminal thread and amastigophores which have no terminal thread, while the rhopa- loides may be subdivided into eurijtclcs whose butts are dilated at their distal ends and stenoteles whose butts are dilated at their bases. Further subdivisions of a number of the nematocyst categories mentioned abo\e were proposed by Weill (10) but it is not neces- sary to review them further here. Weill's system described a total of eighteen different nematocyst categories, and in fact made it possi- ble by applying the terms hoplotelic for armed threads and anaplo- telic for unarmed threads to distinguish two sub-types within most of the subdivisions of the heteronemes. Working from Weill's system still other kinds of nematocysts have been described. Carlgren (2) divided mastigophores into b-mastigophores and p-mastigophores, based on the appearance of the end of the inverted butt. The p-mastigophore was the type Weill (10) had described and the b-mastigophore was a new category which in its unexploded condition looked like a basitrich but when exploded looked like a mastigophore. Another worker, Cutress (5), using the light microscope de- scribed two further categories of nematocysts, q-mastigophores and macrobasic p-mastigophorcs, and proposed the elimination of 190 THE BIOLOGY OF HYDRA : 1961 amastigophores (microbasic and macrobasic amastigophores of Weill). Ciitress also made a number of claims about nematocysts, some of which are wrong and others certainly are questionable. Unfortunately we have not yet progressed far enough in our study to analyze critically all of the structural details of all nematocysts, and until electron microscope studies have been extended to many more types of nematocysts, a number of suggestions Cutress has made cannot be proven or disproven. One of Cutress' suggestions is that the shaft or butt of mastigo- phores is folded within itself as well as being inverted before ex- plosion. This would bring the point of the butt to the tip of the cap- sule, would keep the point in the lead as the basal half of the butt everts, and he claims the thread is attached to this leading tip of the shaft. The tread would evert after the shaft has completely emerged. Miss Jane Westfall of the Department of Zoology at the University of California at Berkeley has been examining a number of nemato- cyst types with the electron microscope and has been particularly interested in mastigophores. Her studies have not yet progressed to a point where publication seems warranted, but we can comment on Cutress' suggestion. Both cross and longitudinal thin sections have been examined as well as whole exploded nematocysts. The material has been primarily the nematocysts of the acontia of our West Coast Metridiiim senile fimbriatum. Cross sections of micro- basic amastigophores, microbasic b-mastigophores and basitrichs (sensu Weill and Carlgren, refs. 2, 10) show clearly that the shaft is not folded on itself and contains only the spines. The spines are blades, as was shown so clearly by Robson (8), and are oriented with their tips toward the open end of the capsule. Longitudinal sections of amastigophores also show that the notch seen in the light microscope at the distal end of the shaft of amastigophores and p-mastigophores is the result of this being the end of the armored region of the shaft. Moreover, there is no thread within the shaft as Cutress has proposed. From these observations we conclude that Cutress is wrong, as were certain earlier workers who proposed folded as well as inverted shafts. It also should be noted that from the work of Picken ( 7 ) and Robson ( 8 ) that Cutress' claim that the holotrichs of Corynactis, which he calls macrobasic p-mastigo- phores, has an inverted and folded shaft is wrong. CADET HAND 191 The proposal of Cutress (5) to eliminate the categories micro- basic and macrobasic amastigophores also is not acceptable. It is true that there frequently is a short thread on many amastigophores, but this thread is apparently sometimes entirely absent. In our elec- tron microscope studies we have failed to find more than a wisp of a thread at the end of the shaft of these nematocysts in Metri- dium and in thin sections we have not been able to verify, as Cutress suggested, that this thread is attached to the inner capsular wall near the end of the shaft. Studies such as Cutress', which were based on the light microscope alone, cannot resolve problems such as this and we must await definitive electron microscope studies. The new category of nematocysts, microbasic q-mastigophores, which Cutress described may indeed be a valid type although this too is open to question. I and other workers have noticed dart- like structures which characterize q-mastigophores lying among exploded nematocysts. Weill (10) reports a number of such oc- currences and reviews some older accounts. These darts, which Cut- ress says are unattached discrete structures, occur within the shafts of certain microbasic mastigophores of acontiate anemones. Cutress reports them from the genera Metridiwn and Aiptasia, to which I can add Diadumene. It was my conclusion that the darts in Diadiimenc franciscana were nothing more than the mass of spines which should have armed the shaft. These spines are tightly curled within the shaft as we have seen in electron micrographs (unpub- lished) and are commonly sloughed off soon after eversion of the shaft as many workers have noted. Little would be required for this mass of spines to stick together, lose their contact with the shaft and form the dart. Whether this happens accidentally or as a nonnal process is not known. In Diadumene franciseana the darts could usually be found lying near a mastigophore with no spination on the shaft. Cutress, however, figures darts emerging from mastigo- phores with spined shafts and associated with nematocysts with spined shafts. If these are accurate observations, the recognition of a special nematocyst, the microbasic q-mastigophore, certainly is called for. It is unfortunate that Cutress did not choose some other name than dart for the organized structure contained in his q-masti- gophores. This name, dart, had already been used by Picken (7) to describe the tip of the packed spines as they emerge from the 192 THE BIOLOGY OF HYDRA : 1961 everting thread. Both structures would appear to be for penetration, and both may be the same if Cutress' interpretation is wrong. If, however, Cutress is correct two things would seem apparent. First with such a large structure as Cutress' dart seems to be, the spines of his mastigophore cannot be as large as those figured by Robson (8) nor as seen in Miss Westfall's micrographs because there would not be space for both. In Miss Westfall's unpublished electron micrographs the spines completely fill the shaft, and Cutress figures spines which would appear to be normal, at least as we see them in the light microscope (see ref. 5, p. 132, Fig. 7b and c). Second, it will continue to be confusing if two dissimilar parts of nematocysts have the same name and Picken's use of the word dart has priority. Another difficult point in Cutress' work concerns basitrichs. It is his contention that the category of nematocysts Weill ( 10 ) identi- fied and defined as basitrichs are in fact for the most part better assigned to the category microbasic b-mastigophore. Cutress is correct when he notes the difficulty in solving the problem with the light microscope because the basic problem here is to determine whether one is dealing with isodiametric isorhizas or with hete- ronemes with a butt. The magnitude of the difference between butt and thread may be as little as 0.1 microns Cutress notes, and this is not a readily resolvable difference with a light microscope. Cut- ress solves the problem by arbitrarily deciding that when one sees a straight inverted shaft, as in Weill's basitrich, this means the tube of this portion is differentiated as a shaft, is greater in diameter than the thread and that the tube itself is stilfer than the thread. The fact that this portion of the tube, the straight part carrying the armature, may be stiffened and not coiled only because it is packed with spines seems not to have occurred to Cutress. Cutress suggests we restrict basitrichs to certain nematocysts which so far are known only from anthozoans and have no stiffened or straight part in the inverted tube. These nematocysts, as he shows in his Figure 3, are basitrichs in every sense. In our electron microscope work we have examined uneverted basitrichs. The wall of the spined portion is not thicker than the wall of the thread. We cannot comment on diametric relationships since it would be the everted, not uneverted picture which should be examined and we have not done this. These CADET HAND 193 basitrichs look structurally very much like the much larger micro- basic b-mastigophores we have looked at in the same tissue, the acontia of Metridium. Cutress may be correct in writing "It may be presumptuous to state that the man who defined almost the entire system of cnidae classification failed to recognize his own categories" (ref. 5, page 126), but it seems "presumptuous" to me for Cutress to have done this on what appears to be spurious logic which assumes a shaft, rather than on factual evidence such as the electron micro- scope could have produced. At any rate, the evidence is not in yet and whether most basitrichs, as we have known them from the liter- ature, are in fact b-mastigophores remains to be seen. If Cutress is correct the identification of microbasic b-mastigophores will be much easier than it is today. My last comments on Cutress concern his new category of macro- basic p-mastigophores. By definition this category is said to have the undischarged shaft inverted and folded back on itself. This certainly is not so as I noted earlier, nor do I believe that this cate- gory includes the holotrichs of Corynactis as Cutress states. In our Corynactis caJifornica the holotrichs are good isorhizas, that is the thread is isodiametric. The category Cutress proposes would in- clude the former macrobasic amastigophores, and again I would say that the shortness of the thread, if it exists at all, is good reason for keeping the amastigophore separate from the p-mastigophore. It also seems reasonable that macrobasic p-mastigophores do exist, but they differ strikingly in their appearance from the microbasic p-mastigophore which has the obvious long coiled terminal thread within the capsule. The comments I have made so far concern both structure and types of nematocysts and I do not intend to review the details of fine structure which are so well known to so many and which we are adding to almost daily as new electron micrographs are exam- ined. The work of Chapman and Tilney (3, 4) stands out as the best work to date on the fine structure of fully formed nemato- cysts, and the work of Slautterback and Fawcett (9) on the de- velopment of nematocysts is clearly the best on this subject to this date. That this elegant work is being done on hydra is little wonder when one considers how easy this beast is to handle in the labora- tory, primarily as a result of Loomis' studies. What are needed are 194 THE BIOLOGY OF HYDRA : 1961 studies of many different coelenteiates so that all of the types may be fully explored rather than merely the limited cnidom of hydra. I would like to briefly explore one other aspect of nematocysts, namely their function and functioning. We have not yet arrived at a point where any single explanation can be had as to how a nema- tocyst discharges nor do we understand the meaning of diversity in nematocysts. Diversity in some microscopic structures such as lepi- dopteran scales and perhaps some sponge spicules seems not to be adaptive. This is, they all perform the same function and as long as a given size or distribution is maintained, variation in shape and ornamentation apparently can occur without selective forces coming into action. In nematocysts we do know that some of the diversity is adap- tive. There is little doubt as to the role of stenoteles and desmonemes in hydra and the recent work of Burnett, Lentz and Warren ( 1 ) has shown that the desmonemes respond before the stenoteles, trap the prey and hold it till the stenoteles discharge and kill it. Also, it appears clear from the work of Ewer ( 6 ) that the atrichs discharge against smooth surfaces and presumably are sticky, or are gluti- nants. Ewer also showed that foodstuffs or extracts from food in- hibited the atrichs while enhancing the discharge of stenoteles. Any- one who has worked with nematocysts has soon discovered that not all types respond to all stimuli, and some types like atrichs and spiro- cysts are very difficult to discharge under most conditions. How- ever, with all the work that has gone on we still can identify only three functions for nematocysts as far as the biology of the animal is concerned, namely adhering, entangling and penetrating, al- though Ewer ( 6 ) has suggested that the holotrichs of hydra may be purely defensive. The penetrating types are all assumed to deliver toxins and poison to the prey or foe but this has not been proven. We have no described or specific function for most nematocyst types and in fact our knowledge is limited in that what is known about function comes entirely from hydra. The work on the nema- tocysts of other types of coelenterates has concerned itself with biochemical problems, with studies of discharge mechanisms, the toxins and the makeup of the capsule rather than the function of the many varied types. Thus we are left with about twenty types of nematocyst of which we known the function of three. It would ap- CADET HAND 195 pear that all the heteronemic stomocnidae are penetrants, but the functions of most haplonemic stomocnidae are not known though we may assume they are adhesive. Among the astomocnidae we find the entangling desmonemes, but what of the rhopalonemes? As well as being in doubt of the function of most nematocysts we are again faced with diversity for which it is not easy to see adap- tive values. Cleverly contrived experiments may be able to answer many of these questions, but the possibility exists that nematocysts may be another example of variation without functional significance. At the moment it is difficult for me to imagine what functional differences one could ascribe to a series of mastigophores with no threads, short threads or long threads. Such variations exist, how- ever, and in discrete places and patterns, that is one species may have one type in one tissue, another in some other tissue, while a second species will show only one type in one place. Certainly types deserves attention. The problem of how nematocysts discharge is a complicated one and one to which many authors have addressed themselves. The cnidocil, which is so characteristic of at least some nemato- cysts of hydra, is not known to be associated with most nemato- cyst types, and in fact has been reported only in hydrozoans. When and if a final relationship between cnidocil and discharge in stenoteles and desmonemes is worked out we still will have to re- solve the problem of how other nematocysts are related to what is found here. We have seen no signs of cnidocils in Metriditim acontia. What the operculum is, or even if it exists in most nematocysts is a difficult problem. There seems to be little doubt that some sort of a plug or structure exists at the point on a capsule where the thread or tube starts everting. In stenoteles the operculum is a real structure as demonstrated by the electron microscope studies of Chapman and Tilney (4). In Miss Westfall's studies of nemato- cysts no operculum has yet been seen, although the material has not made optimum observation on this point possible to date. The mechanism of discharge has been analyzed by many people and I do not feel a detailed summary is called for here. The recent summary of Chapman and Tilney (3) cites the conclusions of the various authors and I would single out the reports of Picken (7) and Robson (8) as those which are most significant. New 196 THE BIOLOGY OF HYDRA : 1961 information will be added as we gather more information on fine structure and as further chemical and biochemical studies are car- ried out. One could suggest from observations on available elec- tron micrographs such as those of Yanagita and Wada (11), Chap- man and Tilney (4) and unpublished ones of Miss Westfall's that the shaft of some heteronemes is folded accordion style. If to this we add the fact that the capsule contracts on explosion, we could imagine that the shaft of these nematocysts unfolds as it everts. This may account for the full eversion of heavily armed shafts and only later would uptake of water play a role in eversion. This sug- gestion can be at least partly tested by critical analysis of the length of the sculptured or folded outline of uneverted shafts as com- pared with the full length of everted ones. It is a rare field of biology where one can say the last word has been said and one wonders if such a field exists, but the study of nematocysts seems clearly to be in its infancy and there is little chance of running out of problems ( or words ) . I do feel, however, that with the renewed interest in these intriguing and complicated structures which has appeared in recent years there is high hope that many of our problems will be solved. I look forward with ex- citement to the time in the future when we have enough knowl- edge to talk about the types, structure and function of nematocysts rather than what is not known. REFERENCES 1. Burnett, A. L., T. Lentz and M. Warren, 1960. The nematocysts of hydra ( Part I ) . The question of control of nematocyst discharge reaction by fully fed hydra. Ann. Soc. Royal Zool. Belgique 90: 247-267. 2. Carlgren, O. 1940. A contribution to the knowledge of structure and distribu- tion of cnidae in the Anthozoa. Kungl. Ftjmig. Sdllskapets Handl. N. F. 51 : 1-62. 3. Chapman, G. B. and L. G. Tilney, 1959. Cytological studies of the nemato- cysts of Hydra. I. Desmonemes, isorhizas, cnidocils, and supporting struc- tures. /. Biophijsic. and Biochem. Cijtol. .5: 69-78. 4. Chapman, G. B. and L. G. Tilney, 1959. Cytological studies of the nemato- cysts of Hydra. II. The Stenoteles. /. Biophysic. and Biochem. Cytol. 5: 79-84. 5. Cutress, C. 1955. An interpretation of the structure and distribution of cnidae in the Anthozoa. Systematic Zoology 4: 120-137. CADET HAND 197 6. Ewer, R. F. 1947. On the functions and mode of action of the nematocysts of Hydra. Proc. Zool. Soc. London 117: 365-376. 7. PiCKEN, L. E. R. 1953. A note on the nematocysts of Conjnactis viridis. Quart. Jour. Micros. Sci. 94: 203-227. 8. RoBSON, E. A. 1953. Nematocysts of Corynactis: The Activity of the filament during discharge. Quart. Jour. Micros. Sci. 94: 229-235. 9. Slautterback, D. L. and D. W. Fawcett, 1959. The development of the cnido- blasts of Hydra. An electron microscope study of cell differentiation. /. Biophysic. and Biochem. Cytol. 5: 441-452. 10. Weill, R. 1934. Contributions a I'etude des Cnidaires et de leur Nematocystes. Trav. Stat. Zool. d. Wimereux Tome 10, 11. Paris. 11. Yanagita, T. M. and T. Wada, 1959. Physiological mechanism of nematocyst responses in sea-anenome VI. A note on the microscopical structure of acontium, with special reference to the situation of cnidae within its sur- face. Cytologia 24: 81-97. DISCUSSION GOREAU: To those of us who swim in reefs and sometimes come into painful contact with Millepora complanata and similar stinging species, it would be of interest to know what nematocysts produce the burning sensation and the erythema. HAND: Four categories have been described: atrichs, basitrichs, macrobasic mastigophores and stenoteles. One could guess that the stenoteles and macrobasic mastigophores give you the kick. MUSCATINE: Has anyone observed the extrusion of substances from the end of nematocyst threads? HAND: Yes, I think there is a lot of information about material being extruded, and one of the places this is most readily visible is in the big holotrichs that corallimorpharian anemones and some corals have. First, there is an uptake of methylene blue. Then there is eversion of the thread as Picken and Robson have explained so beautifully. And then real droplets of the material leave the terminal end of the thread. One can see this happening in a fresh prepara- tion. Whether or not this is the toxin, and what relation this has to the total picture, is not at aU clear. But certainly there is something leaving the capsule. And the total volume of the everted system is in general greater than the uneverted system. In order to create this, something has had to move into the system or expand within it. 198 THE BIOLOGY OF HYDRA : 1961 ROSS:^ I would like to report some work which is partly on the point of Dr. Hand's talk. By chance, a few months ago, I stumbled on a phenomenon that I think has some bearing on the specialized function that certain nematocysts can perform. The sea anemone, Calliocfis parasitica, which I mentioned earlier today, lives on shells of hermit-crabs in British and Mediterranean waters. About 2 years ago I found that the anemone gets on the shell by a rather interesting behavior pattern ( Fig. 1 ) . It will transfer from another surface to the shell by a maneuver which begins with the adhesion of the tentacles to the shell; subsequently the animal detaches the pedal disc which then Fig. 1. Calliactis parasitica adhering to shell by tentacles and (a) detach- ing pedal disc from plastic plate and (b) swinging detached pedal disc over towards shell for eventual settling. 4 min. between (a) and (b). (From Ross, D.M. 1960. Proc. Zoo/. Soc. London, 134: 43-57. Reprinted by the courtesy of the Society) swings over and settles on the shell (Ross, D.M. (1960). Pwc. zool. Soc. Lond. 134: 43-57). But the important point to which I wish to draw attention is this initial response of the tentacles when iDr. Donald Ross, Department of Zoology, University of Alberta, Edmonton, Alberta, Canada. CADET HAND 199 they adhere to the shell. A few months ago, working at Banyuls on the French Mediterranean Coast, D. Davenport, L. Sutton and I looked at this phenomenon and satisfied ourselves that this initial sticking of the tentacles was due to the discharge of nematocysts (Davenport, D., D. M. Ross, and L. Sutton. 1961. Vie et Milieu, in the press). I don't know what kind of nematocyst was involved so I can't add anything about particular nematocyst types and their functions, but certainly it was a nematocyst response to the shell. Now that raises a puzzling point, because these tentacles of Calli- actis stick very readily to shells when the anemone itself is not on a shell; but if you pass a shell over the tentacles of a Calliactis that is already on a shell, its tentacles do not stick. In other words, these nematocysts seem, at any rate from this first observation, to be affected b\' whether the anemone's foot is on the shell or not. We did some experiments to extend this observation. We had 20 Calliactis; 10 of them were settled on shells and 10 were lying unattached on the floor of a tank. By taking a test shell and touch- ing it to single tentacles around the disc, one can get a score of the number of tentacles that stick. When the anemones are on the shell, one gets a score of the order of 5 or less "tentacle-sticks" in 100 shell- tentacle contacts. With the animals lying prone in the tank, one gets a score of the order of 50 or more "tentacle-sticks" using the same shell. In our experiments we transferred these same animals, al- lowing those that had been unattached to settle on shells and strip- ping off those that were attached so that the experiment could be done in reverse. And then we got a good reversal of the scores; the animals which were now on shells, which when unattached had given scores of the order of 50, had now dropped to 5 or less, and the other group, which when attached had given scores of 5 or less, had now climbed up to about 50 "tentacle-sticks" per 100. To my mind this phenomenon raises a crucial point as to whether nemato- cysts are independent effectors or not. I say this because the only change made in the experiment is that in one case the anemone has its pedal disc attached to the shell, and in the other case the pedal disc is free and unattached. So this observation forces one to con- clude that the threshold for this kind of nematocyst discharge could be affected by some form of remote control which in this case seemed to originate in the pedal disc. 200 THE BIOLOGY OF HYDRA : 1961 HAND: I'd like first just to applaud this work and say this is exactly what I was asking for, except that you must find out what these nematocysts are! CROWELL: About how long did you wait before you retested? ROSS: A few hours. The anenomes, when you strip them off, take at least an hour to open up and relax. The other anemones will also take about an hour to settle securely on the shells. One has to wait until all are open and all are settled. So several hours always elapsed between the two sets of observations in our reversal experi- ments. But we did several such experiments, and each time ob- tained clear evidence of big differences in the threshold of nema- tocyst discharge as measured by "stickiness" of tentacles to shell. GOREAU: There is a matter which may be important in connec- tion with what Ross just said. Not too long ago we observed at a depth of about 70 feet a large anemone, probably Bartholomea annuhta, which has living amongst its tentacles a small red fish and several shrimp of the genus Periclimcnses. This shrimp moves freely amongst the tentacles, climbs around on them or hovers just in front of them, waiting for small fish to come along. As soon as a fish is in position, the shrimp climbs aboard and proceeds to remove ectoparasites from the head and mouth. Once finished with the job, the shrimp returns to its host anemone. Neither the shrimp nor the commensal fish living among these tentacles excite any sort of feeding reflex on the part of the anenome. The questions I'd like to ask are these: "What protects these commensals against the nematocysts of the host anemone? Do the nematocysts fail to dis- charge into the animals at all, or are they immune to the action of the nematocysts? " The observations made by Ross seem to indicate that there is complete failure to discharge any menatocysts. In other words, commensal animals living among the tentacles of anemones can probably do so because they somehow inhibit nematocysts discharge and do not trigger off any sort of feeding reflex. That's the thing I don't understand, because I know that such anemones react instantly to any bits of meat dropped on the tentacles. This immediately sets off a feeding reaction resulting in flexion of the tentacles and opening of the stomodeum. CADET HAND 201 ROSS: From my experience, I think this is a failure of the nemato- cysts to discharge. Anemones are usually very active when they are responding to chemical stimuli and discharging their nemato- cysts. If nematocysts were being discharged, in this case, one would expect signs of this in the anemone's behavior. HAND: Davenport and Norris (Biol Bull. 115, 1958) working with the anenome Stoichactis and the fish Amphiprion, which I be- lieve were Philippine in origin, concluded that the nematocysts were not discharging. When a single scale was removed, however, then the fish gets it fast. As soon as the mucous layer is broken the nematocysts respond and the animal is in trouble. MARTIN: In the experiment which Dr. Ross described, I won- der if you are sure if the reaction of stickiness is a virtue of the nematocysts or of the epithelium of the tentacles? ROSS: We managed to induce a nematocyst discharge by offering small pieces of shell to tentacles, and observing under the binocu- lars that a large number of nematocyst threads were attached to the piece of shell. We also found that nematocysts were discharged into "Cutex " impregnated with tiny shell fragments, but not into "Cutex" alone used as a control. We were satisfied that it was a nematocyst discharge when we witnessed the following pheneome- non: You can present a shell to a Calliactis by bringing it up very carefully to a single tentacle. If it sticks, that tentacle adheres so strongly at the tip that you can lift up the whole animal by lifting the shell. It is impressive to see one of these large anemones hang- ing from the shell and attached only by the tip of a single tentacle. I cannot conceive of anything other than a powerful local nemato- cyst discharge that could produce this particular effect. MUSCATINE: Have you ruled out a mucous adhesive? ROSS: We satisfied ourselves by direct observation that they were not adhering by mucous strands, but that the tentacle was sticking directly to the shell at definite points of contact and not over the whole surface. BURNETT: Perhaps your animals which did not adhere to the shell were still discharging nematocysts? We have found that satiat- 202 THE BIOLOGY OF HYDRA : 1961 ed hydra still discharge nematocysts when an Artemia strikes the tentacle, but the nematocyst is quickly released from the tissues of the hydra and the AHemia falls to the bottom of the culture dish. If satiated with food, the hydra makes no effort to hold its prey. Perhaps in your experiments, the nematocysts discharged but were not retained by the cnidoblasts. ROSS: I only refer to the original observation which was that when an animal is on a shell you can brush another shell across it and there is not the faintest sign of a response. The tentacles are just brushed aside; they don't stick to it in any way. Yet you have this phenomenal behavior which is elicited when the animal is off a shell; it practically pounces on the test shell. Starting from that observation, we went on and did this other experiment. I wouldn't say that this is the complete answer, but I think it raises the whole question of nematocyst control \ ery sharply, even though more in- vestigation is required to clear it up. You certainly have a very different type of behavior depending whether the animal's foot is on a shell or not. It seems to us that this bcha\ ior, when it occurs, begins with nematocyst discharge. LOOM IS: Do you think it might be a matter of the shell trans- mitting calcium to the tentacle and making it sticky? ROSS: I've tried a good many models of shells and also shells boiled in alkali to remove organic material. The anemone does not respond to these; cleaning the shells destroys the activity. If you present CaUiactis with a perfect plaster of Paris model of the shell, it shows no interest. The rest of the story (I haven't time for the evidence here ) is that some substance in the mollusc shell, and not derived from the crab but from the mollusc, triggers the nematocyst discharge and the subsequent behavior pattern. It is not responding to the calcium of the shell, or to any other inorganic constituent, or the characteristically sculptured surface of the shell. In fact, the anemone gets on the shell occupied by the hermit-crab by respond- ing to the ghost of the long-dead mollusc that used to live there. It has nothing to do with the crab as such. SLAUTTERRACK: Does anyone care to go into metaphysics further? If not, I declare this meeting adjourned. Activation of the Feeding Reflex in Hydra lift oralis HowAiiD M. Lenhoff Laboidtories- uf Biochemist nj, Howard Hiiiilies Medical Institute, and 7.oology Department, University of Miami, Mianii, Florida Throughout tliis talk, I will often speak of experimenting on Hydra as if these animals were systems of purified enzymes. I speak in these terms more eonfidently toda\' than I could have a few- years ago when I first tried to adapt my former training in enzym- ology to experimentation with live Hydra. In enzymology I was able to treat a relatively simple experimental system in a limited number of ways, and the results were usualh' clear and unambig- uous. I soon found, however, that Hydra could be treated in vir- tually an unlimited number of ways and that the measurable responses of the animal were more difficult to interpret correctly. During a rewarding apprenticeship with Dr. Loomis, I was introduced to his method for rearing Hydra in the laboratory in solutions of known composition (16), a development that has en- abled inxestigators to experiment with hydra using the same rigor- ous controlled conditions which are applicable to simpler systems. These first discoveries of Loomis opened the door wide to contem- porary hydra research. His selection of Hydra for use in quantitative studies of cellular problems was a happy one because of at least three intrinsic prop- erties of the animal. First, genotypic constancy is practically guar- anteed by using animals descended from a single individual by budding. Second, their small size and lack of skeleton lend them to many of the quantitative techniques (7, 14) applicable to simpler systems. But perhaps the feature of hydra which makes 20.3 204 THE BIOLOGY OF HYDRA : 1961 them so remarkably adaptable to quantitative study is their lack of a definite self-regulated internal extra-cellular fluid. In place of this fluid is their culture solution, a solution regulated by the investigator. Once the environment is controlled, individual variation between hydra is minimized and thus the results are rendered less ambiguous. Working on the assumption that the intact Hydra can be treated with the same controlled conditions that we normally employ with an enzyme in solution, we find that in order to get reliable results with the glutathione-//j/f/rfl system, we must control precisely and within restricted limits the following factors, some of which I will report on today: pH, nature of the buffer, ionic strength, the nature of both the cations and anions, temperature, presence of trace metals, amount of aeration, concentration of glutathione or related compounds, presence or absence of proteases or glutamic acid, and length of time since previous exposure to glutathione or since last feeding. Undoubtedly, there are other factors that are as yet unknown. Of course, when studying developmental phenomena, more com- plex problems are met with. At present such phenomena as regen- eration, budding, and cell migration have none of the convenient environmental chemical "handles" (comparable to glutathione and pCOo) which have so often provided the means of attacking a problem. Yet certainly many of the environmental factors aflFecting the feeding reflex also influence developmental phenomena. For example. Hydra grown in a culture solution low in sodium have smooth short tentacles and few nematocysts. At even lower sodium concentrations the ectoderm thickens, developmental abnormalities occur, and often cellular areas begin to disintegrate. These abnor- malities never occur in a medium of the proper sodium content ( 11 ). Research with a whole animal challenges the quantitative biol- ogist. When he treats hydra with the same precision that he treats an in vitro system, he will find that much of the mystery surround- ing the animal disappears and that the excitement of a new under- standing beckons. Now let us consider the activation of the feeding reflex in Hydra littoralis by the tripeptide reduced glutathione. We owe HOWARD M. LENHOFF 205 the discovery of this phenomenon to two independent studies : one, by Helen Park, who, while studying the effects of radiation on Hydra, observed that the anti-radiation compound reduced gluta- thione caused the Hydras mouth to open (20) ; the other by Loomis, who, in a systematic search, identified reduced glutathione as the substance present in crustacean extracts that activated the feeding reflex in Hydra ( 17 ) . The significant aspects of this discovery are many. From an evolutionary viewpoint, data on the distribution of the glutathione- activated response has been used to deduce the sequence in geo- logical time that the feeding mechanisms of some coelenterates evolved (6, 15, 17). On the whole animal level, the feeding re- sponse is an example of an elaborate behavioral pattern controlled by a single environmental compound. At the cellular level, the glu- tathione-activated feeding reflex is a clear example of chemorecep- tion specific for only one molecule. This morning I would like to dwell on a fundamental subcellular aspect: the mechanism by which glutathione combines with and activates the glutathione-receptor. DESCRIPTION OF THE NORMAL FEEDING REFLEX All measurements are based on Hydra's characteristic feed- ing movements, described earlier by Ewer ( 4 ) and Loomis ( 17 ) . The drawings in Figure 1 illustrate each of these steps. A Hydra in the absence of the glutathione has its mouth closed, and its tenta- cles outstretched and relatively motionless. After the addition of glutathione, the tentacles begin to writhe and sweep inwards to- ward the longitudinal axis of the animal ( Fig. 1 A ) . Next, the tenta- cles bend toward the mouth, and the mouth opens (Fig. IB). Shown in this composite drawing (Fig. IB) are the various positions that a tentacle takes before contracting. These movements, culminating in mouth opening, usually all take place within half a minute. Figure IC shows how a Hydra looks during the greater portion of the feed- ing reflex, its mouth open wide and the tentacles in various phases of contraction. Frequently, the tips of the tentacles are observed within the Hydra's mouth, as shown in Figures IB and IC. 206 THE BIOLOGY OF HYDRA : 1961 A B C Fig. 1. Stages of the feeding reflex (see text) (From Ref. 8). A QUANTITATIVE ASSAY Requisite for quantitative studies of any biological phenomenon are accurate and reliable measurements. Therefore, special empha- sis is placed on the assay procedure which has as its basis the visual measurement of the mechanical process of mouth opening. Meaningful measurements of the feeding reflex require Hydra that respond to glutathione in a quantitatively reproducible manner. Large numbers of such experimental animals were obtained by HOWARD M. LENHOFF 207 starving for one or two days mass cultures of Hydra Uttoralis (18) that had been reproducing asexually in a sokition consisting of 10~^ M CaCL and 10~^ M NaHCO;; in deionized water. Special care was taken to remove most of the organic waste products from the cul- tures twice daily (18). The animals in each tray were not allowed to reach a density of over two or three thousand hydranths per 1500 ml. of culture solution. The assay procedure used in most of these experiments was as follows: Five starved Hydra obtained from the mass cultures were rinsed three times in 30 ml. portions of a solution lacking gluta- thione and consisting of 10"-^ M CaCl,, 10"^ A/ NaCl, and 10~^ M histidine chloride buffer, pH 6.2. The fixe Hydra were then transferred in one drop of the solution into 2 ml. of the same solu- tion containing glutathione (Sigma, St. Louis, Mo.). (Reduced glutathione is not readily oxidizable at pH 6.2.) This glutathione solution was in the spherical concavity (36 mm. diam. x 5 mm. deep) of a Maximov tissue culture slide. The Hydra were immediately observed through a binocular dissecting microscope set at a magni- fication of 19.5. The time intervals between the moment the Hydra were placed in the glutathione solution and the initial and final (ti and tf) times that the mouth of each animal was open were recorded. The magnitude of the feeding reflex is expressed as the average time (tf-ti) during which the mouths of the Hydra remained open in response to glutathione. In Table 1 are shown the results of four different experiments (a-d) which were carried out in excess glutathione. In these experi- ments each Hydra opened its mouth within 0.4 to 1.0 minutes (ti) after being placed in the glutathione solution. Under optimal con- ditions, the variations observed in the opening time ti were small when compared to tf, and did not significantly alter the o\'erall time during which the mouth was open (tf-ti). The closing time (tf) for the individual Hydra in each experi- ment (Table 1, expts. a through d) was about 35 minutes. Because the standard deviations were small in comparison to the total length of the response, they were not routinely calculated. At sub-optimal concentrations of glutathione (Table 1, expt. e), or in the presence of a compound known to compete with gluta- 208 THE BIOLOGY OF HYDRA : 1961 thione for the glutathione-receptor (13) (Table 1, expt. f), some Hydra took as long as 6 niiniites to open their mouths, while others did not carry out the feeding reflex at all. In these cases, the stand- ard deviation is large relative to tr-ti. Data of this type are similarly expressed as the average time (tf-ti) during which the mouths of the five Hydra tested remained open regardless of the number that responded positively. TABLE 1 Method of expressing the duration of the feeding reflex Glutathione t^-tj ( min. ) Expt. concentration tj(min. ) tf(min.) Mean rt S.D. (a) 10--^ M 0.43, 0.46, 0.60, 0.78, 1..33 33.00, 35.36, 38.08, 39.71, 41.00 36.71 ± 2.95 (b) 10 = M 0.50, 0.53, 0.71, 0.88, 0.91 32.80,-33.16,36.25, 36.43, .38. 11 34.64 ± 2.10 (c) 7.5 X 10-« M 0.43, 0.46, 0.58, 0.68, 0.96 28.21, 36.50, 36.50, 38.30, 43.45 .35.97 ± 5..30 (d) 5 X 10-* M 0.48,0.50,0.61, 0.78, 1.05 26.88, 35.25, 37.41, 42.00, 43.41 36.31 ± 6.36 (e) 5 X 10-7 M 0.68, 2.33, 2.63, 4.75, 00 5.08, 16.25, 22.50, 25.60,-. 11.81 ± 9.29 (f) 2 X 10-6 M 2.45, 3.75, 6.00, 6.41, 13.45, 22.91, 6.11 ±7.23 (+1C |-^ M glutamine) 00, 00 — , — . Hydra starved for two days were used in all experiments. Data from reference 8. The values for tr-ti at excess glutathione concentrations are usually within the range of 25 or 35 minutes, depending upon whether the Hydra were starved one or two days preceding the experiment. This fact should be borne in mind when comparing data from different sets of experiments. ( It now also is known that small changes in temperature influence tf-ti significantly as shown in Table 7. ) These data, in addition to providing a basis for the assay, give insight into central problems concerning the mechanism by which glutathione elicits the feeding reflex. The values given for ti must include the time required for at least two major processes to occur: ( a ) the union of glutathione with its receptor, which in experiments a-d is probably rapid (i.e., within a few seconds), and (b) all HOWARD M. LENHOFF 209 of the subsequent events leading to mouth opening. The values for ti (0.4 to 1.0 minutes) may represent, for the most part, the latter events. Large values of ti (those greater than 2.0 minutes) indicate that the experimental conditions for the feeding reflex are not optimal. For example, it takes longer for the mouth to open at low gluta- thione concentrations (Table 1, expt. e) or in the presence of a competitive inhibitor (Table 1, expt. f) than at excess glutathione concentrations under optimum conditions. Similarly, cellular poi- sons, such as N-ethyl maleimide or heavy metals, also cause an increase in ti (9). Further, Hydra in distilled water take longer to respond than do Hydra in distilled water containing added calcium (10). In the cases mentioned here, it would appear that the large values of t. result from the interference with the activation of a sufficient number of functional receptor sites needed to elicit an optimally rapid response. Another cause of a delay in mouth opening might stem from interference with some of the cellular events initiated by the combination of glutathione with its receptor. At sub-optimal concentrations of glutathione (Table 1, expt. e, and Fig. 3 at concentrations less than 5 ■ 10"*' M glutathione) the tf-ti values were small in comparison to those obtained at higher glutathione concentrations. These results show that graded responses can occur when conditions are not optimum. In addition, it is gen- erally observed that the larger the value of ti, the smaller the value of tf-ti. EFFECT OF GLUTAMIC ACID AND GLUTATHIONE ANALOGS Using glutathione analogs, we have undertaken a study of the size and configuration of the glutathione molecule necessary for activation of the response. The results, summarized in Table 2, show that the glutathione-receptor has a most unusual specificity compared to proteins which react with glutathione ( 13 ) . The recep- tor (a) is not dependent upon the sulfhydryl moiety of glutathione for activation, (b) has a high order of specificity for the structure of the tripeptide "backbone" of glutathione, and (c) is inhibited by glutamic acid. 210 THE BIOLOGY OF HYDRA : 1961 TABLE 2 Activators and inhibitors of the feeding reflex R CH. O.C-CH-CH.-CH.-CO- NH-CH-CO- NH-CH.-CO-2 A lutamyl - B alanyl C - jjlycine Activators Inhibitors ( tripeptide ) ( others ) R = -H R = -CH. R = -SH R = -S-CH ., R = -SO.H R = -SO:;H R = -S-COC R = -S(N-et R = _S-SG H:, lylsuccininiido) -CH.CO- glutamic acid glutamine cysteinylglycine R = SH and A =-0,C-CH 1 1 +NHo In confirmation of Cliffe and Waley ( 3 ) , we were able to extend their striking resnlts demonstrating that the sulfhydryl group is not necessary for the action of gkitathione on Hydra and that this group can be altered within certain limits. These workers obtained a posi- tive response in Ilijdm exposed to the lens tripeptide ophthalmic acid (y-glutamyl-«-amino-/]-butyryl-glycine). This tripeptide, as well as nor-ophthalmic acid, activates the feeding reflex in Hydra. Both peptides are identical to glutathione except that they have respectively a methyl or a hydrogen atom instead of the sulfhydryl. We further show that the S-methyl analog of glutathione also acti- vates the feeding response ( 13 ) . On the other hand, substitution of large groups for the sulf- hydryl moiety leads to analogs which do not have the right configuration to activate a response. Rather, such analogs (y-L-glutamyl-L-sulphi-alanylglycine, y-L-glutamyl-L-sulpho-alanyl- HOWARD M. LENHOFF 211 glycine, S-acetyl glutathione, S-succinimido glutathione, and oxidized glutathione ) when at high concentrations inhibit the action of glutathione (3, 13). These inhibitions are overcome by increasing the glutathione concentration. Thus, those analogs which retain the tripeptide backbone of glutathione act as competitive inhibitors in the activation of the feeding reflex. Another tripeptide which acts as a competitive inhibitor is asparthione (/?-aspartylcysteinylglycine) (13). This compound is nearly identical to reduced glutathione except that it lacks one methylene group, having aspartic acid substituted for glutamic acid. Loomis first showed that asparthione fails to activate the feed- ing reflex ( 17 ) . These characteristics of asparthione point out the importance of the y-glutamyl moiety of the tripeptide for the acti- vation process, as well as providing additional proof that the pres- ence of a sulfhydryl group on a tripeptide similar to glutathione is not sufficient for activity. Contrastingly, glyoxylase, another glu- tathione-requiring system, functions with asparthione (1). Further evidence that the y-glutamyl moiety is of special impor- tance in the active structure of glutathione is the action of both glutamic acid and glutamine as competitive inhibitors of gluta- thione, while neither cysteine nor glycine have this efi^ect ( 13 ) . The importance of the a-amino group of the glutamyl moiety is emphasized by the failure of a-keto glutaric acid and of glutaric acid to inhibit. Also, as might l)e anticipated, neither aspartic acid nor asparagine were inhibitory (13). These data indicate that the receptor has a high affinity for the y-glutamyl group, that the sulfhydryl group is important only in that it conforms to certain size limitations, and that the glycine is needed to complete the fit of the tripeptide into the receptor (Loomis has shown that y-glutamylcysteine does not activate a response. Ref. 17). As more analogs become available, we hope to determine the exact structural requirements for the stimulatory activity of glutathione. In addition, it should be possible by com- paring the Ki's of the different inhibitors to determine the relative affinities of the receptor for the different parts of the glutathione molecule. No other system known to require glutathione has such exacting requirements for the peptide backbone of glutathione. Regardless 212 THE BIOLOGY OF HYDRA : 1961 of this remarkable specificity and of the ample glutathione in the fluids emitted from Hydra s captured prey in nature, the possibility remained that some unknown trace factors present in these fluids are the natural activators of the feeding reflex. To examine this possibility, the following series of experiments were carried out ( Table 3 ) : a diluted aqueous extract ( 30;iig. ) from homogenized Arfemia elicited a 37 minute feeding response in Hydra (expt. 1). A similar extract containing 10~^ M added glutamic acid gave only a 7 minute response (expt. 2). The glu- tamic acid was presumably competitively inhibiting the glutathione in the Artcmia extract because, if in addition to 10~^ M glutamic acid, 10~"' M glutathione, was also included, then the inhibition was reversed ( expt. 3 ) . The inhibition was also reversed by increas- ing the amount of shrimp extract (expt. 4). Further evidence that the glutamic acid was acting competitively and was not irreversibly TABLE 3 The inhibition by glutamic acid of the feeding reflex induced by Artemia extracts, and the reversal of that inhibition by glutathione Expt. Test Solution tj-t; ( mill. ) 1. 30 p,g. Artemia extract. 37.3 2. 30 fig. Artemia extract and 10^ M glutamic acid. 7.4 3. 30 /ig. Artemia extract, 10~* M glutamic acid, and 10"^ M reduced glutathione. 40.6 4. 140 ng. Artemia extract and 10-' M glutamic acid. 26.5 5. Hydra from expt. 2 in 10"" M reduced glutathione. 29.1 6. 10-" M reduced glutathione 42.1 Hydra starved for two days were used in all experiments. HOWARD M. LENHOFF 213 poisoning the Hydra was shown in experiment 5 where the inhibited animals from experiment 2 were washed and then immediately placed in a fresh glutathione solution; these inhibited Hydra responded again for an additional 29 minutes. These experiments leave little doubt that reduced glutathione emitted from the prey is the major natural substance activating the feeding reflex in Hydra littoralis. EFFECT OF ENVIRONMENTAL CATIONS AND ANIONS As emphasized earlier, the Hydra's external aqueous environ- ment takes on a special importance when studying the feeding reflex. This fluid serves a dual role: it conveys glutathione from the prey to Hydra, and it bathes both the receptor and the ecto- dermal effector cells which are involved in part of the contractile processes of the feeding reflex. Therefore, before directly investigat- ing the mechanism by which glutathione activates the response, it is necessary to consider the influence on the feeding reflex of the ions in the media bathing the animals. Knowledge of the effects of these ions might be useful in gaining insight into the mechanisms involved TABLE 4 Ionic requirements for the activation of the feeding reflex Cations Anions Expt. Text Solution tf-ti Expt. Test Solution tf-tj (min.) 10-3 M (min.) 1. 10-^ M EDTA l.a. CaCL. 28.5 2. 10-^ M EDTA and 10-=' M CaCLo 28.5 2.a. CaBr. 23.0 3. 10-^ M EDTA and 3.a. CaL. 9.2 10-=' M SrCL 6.8 4.a. Ca(NO.). 8.3 Hydra starved for one day used in all experiments. All ions were dissolved in a solution of 10-* M NaHCOs Data from reference 10. 214 THE BIOLOGY OF HYDRA : 1961 in the feeding reflex in addition to defining the experimental limits within which the ionic composition can be varied. Summarized in Table 4 are data concerning the ionic require- ments for the feeding reflex (10). Hydra placed in a 10~^ M solution of the chelating agent ethylenediamine tetraacetic acid (EDTA) lost their ability to respond to glutathione (expt. 1). Since EDTA is known to chelate calcium ion, one of the two environmen- tal cations required for the growth of Hydra, 10~'' M CaCL was added to this same solution, and the Hydra responded normally (expt. 2). No other cation would replace calcium to any degree in reversing the inhibitory action of EDTA except strontium ( expt. 3 ) . Since this metal behaves chemically like calcium, experiment 3 strengthens the evidence that calcium is required to effect the feeding reflex. Further evidence for the calcium requirement was obtained 0.4 r- --0.2 - [CO-H+] Molarity x 10 ^) Fig. 2. The inhibition of the feeding reflex by magnesium ions, and its reversal by calcium ions. HOWARD M. LENHOFF 215 using magnesium ions, an ion known to compete with calcium in many biological systems. To show the competitive nature of the magnesium inhibition, our data is expressed in a fashion analogous to the Lineweaver-Burke plot. Here (Fig. 2) we plot the reciprocal of the duration of the feeding reflex against the reciprocal of the calcium concentration. These experiments were carried out in the absence of magnesium, or in 10~^ M MgCL, or in 10~^ M MgCL. The data show that the higher the concentration of magnesium, the greater is the inhibition. Furthermore, as the calcium concen- tration is increased, the magnesium inhibition is completely re- versed. These experiments leave little doubt that magnesium is interfering with the normal function of calcium ions in the feeding reflex. Sodium ions also exhibit similar competitive inhibitory efl^ects (9). However, for a comparable inhibition higher concentrations of sodium ions are necessary. At present there is little evidence as to whether the calcium functions at the glutathione-receptor, or in the efl^ector system. (It appears that the trypsin activation of the feeding reflex, which will be described later, also requires environmental calcium, thus favor- ing the involvement of calcium in the effector system.) Anions were also found to influence the feeding reflex (Table 4 ) . Holding the calcium content constant, the order of effectiveness of the anions in increasing the duration of the feeding reflex was: CI > Br > I = NO:, (10). The relationship of this order to the lyo- tropic series suggests that these ions influence the state of some proteins involved in the feeding reflex. From a practical viewpoint, these results point out the necessity of controlling precisely the ionic environment for the quantitative study of the feeding complex. COMBINATION OF GLUTATHIONE WITH THE RECEPTOR Most of the data just described concerns environmental chem- istry. Now we have to ask questions about the first physiological event of the feeding reflex, the combination of glutathione with the receptor. 216 THE BIOLOGY OF HYDRA : 1961 TABLE 5 Time required for mouths to close on removal of glutathione Time in Glutathione Time to Close ( min. ) ( min. ) 5.0 0.94 10.0 0.92 12.5 0.99 15.0 0.76 20.0 0.72 22.5 0.44 25.0 0.41 All experiments were carried out using Hydra starved for one day, and at excess glutathione ( 10"^ M). Data from reference 8. The following simple experiment demonstrated that glutathione did not act as a "trigger," if a triggered response is defined as one that continues after the initial stimulus is removed. Groups of Hydra were incubated in 10"'' M glutathione for periods varying from 5 to 25 minutes (Table 5). At the end of each incubation period, the animals in one drop of glutathione solution, were placed in 30 ml. of a solution of the same composition but lacking glutathione. In all cases the mouths closed in less than one minute (Table 5). The results indicate that glutathione had to be constantly present during the total time of the feeding reflex in order for the response to con- tinue. In addition, since the mouths close repidly on removal of glutathione it is concluded that the equilibrium between glutathione and the receptor is rapidly attained. This observation that the continued presence of glutathione is required for the activation of the feeding reflex allows us to formu- late a hypothesis on how glutathione activates the receptor. We visualize the receptor as an inactive protein on the siuface of certain Hydra cells. When that protein combines with glutathione, its tertiary structure is altered, rendering the receptor protein physi- ologically active. The active protein is then capable of either initi- ating, or allowing to go on to completion, the events involved in the receptor-eflFector system. These data also indicate that the longer Hydra were exposed to the glutathione, the sooner the mouths closed when glutathione HOWARD M. LENHOFF 217 was removed. The time that it took for the mouths to close prob- ably represent the time required both for the dissociation of the glutathione and for the cessation of the cellular events involved in the receptor-effector system. The observations that mouth closure was more rapid the longer Hydra were exposed to glutathione may imply that bound metabolic intermediates or cofactors, postulated to be released by and to take part in this system (12), become depleted as the feeding reHex nears completion. When considering the quantitative aspects of the union of gluta- thione with the receptor, we found that the data were more mean- ingful if they were treated according to concepts borrowed from enzymology. Therefore, we investigated the effect of glutathione concentration on the "activity" of the receptor-effector system, the "activity" in this case being expressed as the duration of the feeding reflex (Fig. 3). For each concentration of glutathione, duplicate groups of five animals were used. In experiments employing Hydra starved for two days (solid curve), a maximum response was ob- served at concentrations of 5 X 10 ""^ M and greater. No further increase in the duration of the feeding reflex occurred at higher glutathione concentrations. At lower glutathione concentrations, the duration of the feeding reflex increased in nearly direct propor- tion to the amount of glutathione added. However, at these smaller values there was greater variation in the response of the individual Hydra, some not responding at all, as indicated by the symbols used in Figure 3. The similarity of this plot (Fig. 3) to the Lang- muir adsorption isotherm, and to a curve illustrating the saturation of an enzyme by its substrate is apparent. Accordingly, the results (Fig. 3) are interpreted as indicating that at glutathione concen- trations greater than 5 X 10"** M all of the glutathione-receptors are saturated. In these experiments we have not been able to demon- strate that the glutathione is metabolized in a manner analogous to the metabolism of a substrate by its enzyme. But rather it appears as if the glutathione continues to activate all of the receptor- effector systems until the response ceases. At subsaturation levels of glutathione, the animal does not respond to its fullest capacity (see also Table 1, expt. e). Another useful concept, analogous to the Michaelis constant, or K^r, used in enzymology, is the concentration of glutathione eliciting 218 THE BIOLOGY OF HYDRA : 1961 a half-maximum response. For the g\utathione-H ydm system this value, ca. 10"^ M, probably closely approximates a true dissociation constant because of the apparent absence of glutathione metabolic products. A rough mass law treatment using the method of Scatch- ard (21) indicates that this constant can be measured within a factor of 2. The significance of this constant is threefold: First, its smallness indicates that the receptor has a high affinity for gluta- thione. Second the value of 10~^ M is within the physiologically active range expected to occur under natural conditions of feeding. And, third, this number provides a means of characterizing the receptor; that is, the glutathione receptor can be said to have a constant of 10~^ M. This constant has been found to be a charac- teristic of the receptor and to be nearly the same no matter what the nutritional state of the Hydra. For example, Figure 3 demon- strates that Hydra starved for two days respond to higher concen- trations of glutathione for a greater period of time than do Hydra starved for one day (lower curve). Nonetheless, the concentration of glutathione eliciting a half-maximal response on both sets of Hydra was 10~^ M. The difference in the maximum response observed in Hydra starved one or two days ( Fig. 3 ) become understandable if another comparison is made with enzyme systems. Just as the maximum activity of an enzyme reaction is dependent on the quantity of enzyme present and is not a specific property of the enzyme, in a similar manner the duration of the reflex at concentrations eliciting the maximum response is dependent upon the quantity of com- pleted^ receptor-effector systems of the Hydra. The maximum response is not an intrinsic property of the receptor or of Hydra as is the Km. Thus, Hydra starved for one day are interpreted to have fewer completed receptor-effector systems than Hydra starved for two days. As emphasized in the above comparison, just as the total enzyme activity at saturating concentrations of substrate is proportional to the amount of enzyme, so the total maximum response of Hydra to lA completed receptor-effector system is defined as one containing all of the com- ponents necessary for it to function when in combination with glutathione. When all the receptor-effector systems are completed, the Hydra is capable of carrying out a maximum response. HOWARD M. LENHOFF 219 HYDRA STARVED FOR TWO DAYS GLUTATHIONE CONCENTRATION (Molorify x 10") Fig. 3. Effect of glutathione concentration on the duration of the feeding reflex. Each point represents the mean for five Hydra. The type of symbol used indicates the number of Hydra in the group of five responding to glutathione: i.e. o, five; •, four; [H, three; A/ fwo; and A/ one. (From reference 8). excess glutathione is proportional to the number of active receptor- effector systems in each Hydra. Thus, in order to get comparable results, it is imperative in experiments using excess glutathione concentrations (10^' M) that each Hydra possess approximately the same number of such systems. Since it is impossible to know beforehand the number of completed receptor-effector systems per Hydra, the only criterion for obtaining quantitatively reproducible results is to select Hydra reared under nearly identical laboratory conditions. We repeatedly find that the standard deviation of the response of Hydra to excess glutathione is lov^^ if these animals come from the same mass culture (Table 1, expts. a-d). Therefore, one should not compare experiments employing Hydra taken from different mass cultures. Variation might result either from differ- 220 THE BIOLOGY OF HYDRA : 1961 ences in the following factors: the time elapsed since the previous exposure to glutathione, the ratios of environmental cations or anions, the temperature of the experiment, or to some presently unidentified factors. EVIDENCE FOR AN INTRINSIC LIMIT TO THE RESPONSE The data in Table 1 and Figure 3 show that the feeding reflex is limited to 25-35 minutes, depending upon the conditions of the experiments. In order to determine whether this mouth closure resulted from some intrinsic change within the Hydra, or from the oxidation or alteration of glutathione in the culture solution the following experiment was performed: A group of 5 Hydra were exposed to 2 ml. of 10~^ M glutathione until their mouths closed (Table 6). The same glutathione solution was then transferred to Response of TABLE 6 different groups of Hydra exposed to the some solution of excitatory compound used three times DURATION OF FEEDING tf-tj (mill.) REFLEX Group of Hydra Glutathione Ophthalmic Acid 10-= M lO-fi M S-Methyl Glutatliione 10-= M 1 2 3 27.1 27.6 19.8 23.1 28.1 24.0 29.5 26.2 21.7 AU Hydra were starved for one day. Data from reference 8. another group of 5 Hydra; this latter group of Hydra opened their mouths for 27 minutes, indicating that sufficient glutathione re- mained to elicit a near-maximum response. This transfer process was repeated, and again the Hijdra responded positively, although for a somewhat shorter time. Using the p-mercuribenzoate proce- HOWARD M. LENHOFF 221 dure of Boyer (2), parallel experiments were run in which the respective solutions were assayed for sulfhydryl groups after the Hydra had closed their mouths. No perceptible decrease in the sulfhydryl content of the solution occurred. Similar experiments were carried out using the glutathione ana- logs ophthalmic acid and S-methyl glutathione, compounds that activate the feeding reflex and are not auto-oxidizable. As shown in Table 6, these analogs like glutathione, retain much of their activ- ity after several exposures to Hydra. It can be concluded from all the experiments summarized in Table 6 that the feeding reflex nor- mally ends as a result of some other cause than the oxidation, disap- pearance, or alteration of the glutathione molecule; also it does not end because of the accumulation of inhibitors in the culture solution. Further examination of Table 6, however, does indicate some shortening of tf after using the same glutathione solution on three successive groups of Hydra. Thus, it appears that either the gluta- thione concentration was in some manner slightly lowered, or that some inhibitory factor gradually accumulated in the environment. It is not necessary to assume that the glutathione or glutathione analogs are altered or destroyed when combining with the receptor. There are known instances in which a biological response is initi- ated by a molecule (non-coenzymic in function) combining with a specific site without being metabolized. For example, thiogalacto- side induces the adaptive fonnation of the enzyme /?-galactosidase without being hydrolyzed ( 19 ) . Thus, from the data in Table 6, we might postulate as one result of glutathione activation, the consumption of some substance in the receptor-effector system, the concentration of which limits the duration of the feeding reflex to 25-35 minutes. If this postulate is true, then one might expect that after the Hydra have carried out a maximum response, there will be a period during which they give no further response to a fresh solution of glutathione. Secondly, there will be another period in which they regain their ability to respond maximally. This proved to be the case as shown in Figure 4. In this experiment large numbers of Hydra were exposed to gluta- thione for forty minutes. The animals were then washed with and placed into the glutathione-free culture solution, and, at intervals, exposed to a fresh solution of glutathione. The results show that 222 THE BIOLOGY OF HYDRA : 1961 40 I- 10 20 30 40 HOURS AFTER INITIAL EXPOSURE TO 6SH Fig. 4. Time for recovery of the ability to respond to glutathione (see text). during the first hour, the animals give Httle if any response. By the tenth hour, however, the Hydra had regained their abihty to respond for about 15 minutes, and after one day, responded maximally. Extending the interval between exposures to o\'er 70 hours did not result in any further increase in the length of their response to fresh glutathione. This lag and gradual resumption in the ability of Hydra to respond to a fresh stimulus of glutathione is interpreted to signify the period for the resynthesis of some substance, called "X," which we postulate to be limiting in the receptor-effector system. This view places a greater emphasis on the state of the receptor-effector system than on the physiology of the whole animal. EFFECTS OF TEMPERATURE The effects of temperature on the feeding response were studied primarily to provide more evidence concerning limiting substance X HOWARD M. LENHOFF 223 and information concerning its role in the execution of the response. These experiments are still in the notebook stage and will be sum- marized here only in order to show you some of the directions our research is taking. If the reactions of the feeding reflex (Fig. 5) are depicted as in- x'olving the conversion of limiting substance X to Y, then one might expect two major results of lowering the temperature. First, a small lowering in temperature should lower the rate of all the thermo- chemical reactions. However, by slowing down the reaction con- verting the X to Y we should therefore slow down the rate at which the supply of X is depleted, and thus increase the length of time that the mouth remains opened. This proved to be the case as shown in Table 7 where, as the temperature approaches 15 \ the Hi/dm respond for nearly 100 minutes. TABLE 7 Effect of temperature on the duration of the feeding reflex Temperature tf-ti Temperature tf-ti ( min. ) (min.) 6.2° 5.9 18.1° 86.9 8.9° 36.4 18.6° 55.0 10.3° 59.8 19.7° 59.2 12.5° 70.7 20.6° .55.0 14.5° 60.0 21.9° 35.1 15.4° 88.6 24.1° 29.4 16.3° 99.7 25.3° 21.5 27.7° 19.7 All Hydra were starved for two da) s. As a second efl^ect of lowering temperature, the limiting reaction may go so slowly that the optimum (threshold) conditions neces- sary for the feeding reflex are not maintained. Thus, when the Hydra are held below 15° they are observed to open their mouths for a few minutes, then close, open, etc. until they finally stop responding. The total duration of the responses below 15 becomes progressively less until the Hydra barely respond (Table 7). In fact, when the temperature is lowered from 20" to 5", the mouth takes 224 THE BIOLOGY OF HYDRA : 1961 longer to open (Table 8). These results (Table 8) are interpreted to mean that as the temperature is lowered, it takes longer for the completion of all the reactions (including the limiting one) leading to mouth opening.- TABLE 8 Effect of temperature on time of mouth opening Temperature 1 /t, (min. ) 5.2° 0.10 7.9° 0.35 9.5° 0.62 13.0° 1.11 16.3° 1.72 18.5° 1.56 20.1° 2.56 All Hydra were starved for two days. Each value is the mean for 25 animals. ACTIVATION BY PROTEOLYTIC ENZYMES Recently we have been carrying out some experiments in acti- vating the feeding response in the absence of added glutathione by using certain proteolytic enzymes. Although still in the preliminary stage, these experiments may help illucidate the sequence of events taking place in the receptor-effector system, and thus are of sufficient interest for some of them to be reported here. We have previously shown that papain, ficin, and trypsin acti- vate a feeding response in Hydra (12). This proteolytic activation was shown not to be the result of any toxic action of the enzymes for the Hydra were intact and alive after one day's exposure to the proteases. Dialyzed ficin, like papain (Table 9), did not activate a response unless cysteine was added to render the enzyme active. The boiled enzymes could not be activated by cysteine. The action -At temperatures below 13° Hydra vary greatly in their response, some animals not responding at all. Therefore, the data are expressed as 1/tj rather than as tj because in cases where there was no response, the t; values would range to infinity. HOWARD M. LENHOFF 225 of trypsin was inhibited by trypsin inhibitor (Table 9). Thus, the response seems to be a result of the proteolytic activities of these enzymes. Of twenty other purified proteins, only chymotrypsin gave a significant (8 min.) response (12). It does not seem likely that the proteases are acting by releasing reduced glutathione from Hydra because y-glutamyl linkages are rare in proteins, and because furthermore glutamic acid, a specific inhibitor for glutathione (13), does not inhibit the action of trypsin (9) . The possible effects of proteases on a whole animal are so numer- ous that it would be difficult at this time to single out any one action that would explain their effect on Hydra. Nevertheless, the important fact remains that proteases do activate a response, and thus a study of their effects might help in arriving at an understand- ing of the actual mechanism. For example, trypsin can activate only an 18-minute response; if glutathione, however, is added to the same Hydra, they respond an additional 17-18 minutes. In con- trast, after a 35 minute response initiated by glutathione, the addi- tion of trypsin has no effect. A mixture of excess glutathione and excess trypsin, interestingly enough, elicits a response equal only to that initiated by glutathione alone. Thus, these preliminary exper- iments indicate that the protease probably activates a series of events common to those activated by glutathione and involving the consumption of limiting substance X (9). Therefore, in Figure 5, the arrow indicating the site of action of the protease is drawn TABLE 9 Activation of feeding reflex by proteases Expt. Test Solution t^-tj (min.) La. 20 /ig./ml. papain 0.1 b. 20 fig./ml. papain -|- 10'3 M cysteine 19.8 c. 20 /ig./ml. papain -|- 10'3 M cysteine, boiled 5 min. at 100° 2.a. 0.1 mg./ml. trypsin 17.8 b. 0.1 mg./ml. trypsin -(- 0.1 mg./ml. trypsin inhibitor All Hydra were starved for two days. Data from reference 12. 226 THE BIOLOGY OF HYDRA : 1961 somewhere between the receptor and before the reaction involving the consumption of X. The activation by proteases has also been useful in determining the relative site at which calcium functions. Since the presence of environmental calcium ion is required for the activation of both glutathione (10) and proteases (9), we feel that calcium plays a role in the effector system rather than in the combination of gluta- thione with the receptor. A recent development which places added importance upon the activation of the feeding reflex by f)roteases is the discovery by Fulton (6) that proteases also activate the feeding reflex in Cordy- lophora. His results are striking in that he has also shown that Cordylophora do not carry out the feeding reflex in response to the peptide reduced glutathione, but rather to the single imino acid pro- line (5). Thus, although Hydra and Cordylophora have different specific excitatory compounds, the feeding reflex in both animals can be activated by proteases. In addition, Physalia gastroozoids, which normally respond to glutathione ( 15 ) , also are activated by i)ro- teases ( 12 ) . All of these results suggest that the protease is acting on some step which is common to the feeding reflex of all these organisms irrespective of the excitatory compound involved. SUMMARY AND CONCLUSIONS With the aid of the simplified scheme shown in Figure 5, I would like to summarize the present state of knowledge concerning the mechanism by which glutathione combines with and activates the glutathione-receptor of Hydra to elicit the feeding reflex. The activity of the glutathione resides in the size and configuration of the 7-glutamylalanylglycine backbone of the tripeptide, and not in the reducing properties of the molecule ( 3, 13, 17 ) . Concentrations of glutathione greater than 5 X 10 ~"' M activate all of the receptor- effector systems (Fig. 3), which are probably localized in the area immediately around the mouth and on the tentacles (8). The con- centration of glutathione eliciting a half maximum response is 10~*^ M (Fig. 3). In order for a response to take place, the glutathione must be constantly present at the receptor site (Table 5). The associ- HOWARD M. LENHOFF 227 ation of glutathione with the receptor is rapidly attained ( Table 5 ) ; the affinity of the receptor for glutathione is high (Fig. 3). After glutathione combines with the receptor, it takes about 0.5 minutes for all the events necessary for mouth opening to occur ( Table 1 ) . Once the reflex begins, it will continue for 25 to 35 minutes (Tables 1, 3, 4 and 6; Figs. 3 and 4). The response does not stop because of any alteration in the glutathione molecule (Table 6), but rather because of some inherent property of Hydra. The duration of the response is probably directly related to the conversion of some lim- iting substance X to its product Y (Tables 1, 3, 4 and 6). Lowered temperatures increase the duration of the feeding reflex, probably by decreasing the rate at which the supply of X is exhausted ( Table 7 ) . It takes about 24 hours for X to be resynthesized either from Y or anew (Fig. 4). The response can be stimulated in the absence of glutathione by certain proteases (12). The protease probably acts before the step involving the consumption of X. Furthermore, the presence of small amounts of calcium ion in the medium sur- rounding Hydra are required in order that a response may occur (10). The calcium appears to be involved in steps occurring between the site of activation by proteases and the effector system. X Y V f ^^Feeding GSH4-Rec^[GSH-Rec] — >E,-^E„-^=^E, »E., <^j > I ^k- Reflex Protease Fig. 5. Schematic outline of the glutathione receptor-effector system. Rec represents the receptor; E,„ E,,, E,, and E,„ enzymes; X, the limiting substance; and Y, its metabolic product. These results are concerned with a single biological system in which a specific excitatory compound combines with its receptor to activate a coordinated response. Activations by an excitatory com- pound comprise the common step in many basic biological phe- nomena such as chemoreception and hormone action. Some of the 228 THE BIOLOGY OF HYDRA : 1961 results described here on the interaction of gkitathione with the Hydra receptor may bear a relation to the functioning of some of these other systems. ACKNOWLEDGEMENTS It is a pleasure to acknowledge the superb assistance that Mr. John Bovaird has provided throughout this study. The criticisms of this manuscript by Drs. J. F. Woessner, Jr., A. Phillips, E. L. Chambers, and W. D. Dandliker are greatly appreciated. REFERENCES 1. Behrens, O. 1941. Coenzymes for glyoxylase. }. Biol. Chein. 141: 503-508. 2. BoYER, P. D. 1954. Spectophotometric study of the reaction of protein sulfhydr\'l groups with organic mercurials. /. Am. Cheiri. Soc. 76: 4331-4337. 3. Cliffe, E.. E., and S. G. Waley. 1958. Effect of analogues of glutathione on the feeding reaction of hydra. Nature 182: 804-805. 4. Ewer, R. F. 1947. On the function and mode of action of the nematocysts of Hydra. Proc. Zool. Soc. London 117: 365-376. 5. Fulton, C. 1960. The biology of a colonial hydroid. Ph.D. Thesis. The Rocke- feller Institute. 6. Fulton, C. (in press). The growth and feeding of Cordylophora and other hydroids. In The Lower Metozoa: Comparative Biology and Phtjlogeny, edited by M. B. Allen. Academic Press, Inc., New York. 7. Lenhoff, H. M., 1961. Digestion of protein in Hydra as studied using radio- autography and fractionation by differential solubilities. Exptl. Cell Research 23: 335-353. 8. Lenhoff, H. M. (in press). Activation of the feeding reHex in Hydra littoralis. I. Role played by reduced glutathione, and quantitative assay of the feeding reflex. /. Gen. Physiol. 9. Lenhoff, H. M. Unpublished observations. 10. Lenhoff, H. M. and J. Boviard. 1959. Requirement of bound calcium for the action of surface chemoreceptors. Science 130: 1474-1476. 11. Lenhoff, H. M. and J. Bo\aird. 1960. The requirement of trace amounts of environmental sodium for the growth and development of Hydra. Exptl. Cell Research 20: 384-394. 12. Lenhoff, H. M. and J. Bovaird. 1960. Enzymatic activation of a hormone-like response in Hydra by proteases. Nature 187: 671-673. 13. Lenhoff, H. M. and J. Bovaird. 1961. Action of glutamic acid and glutatliione analogues on the Hydra glutathione-receptor. Nature 189: 486-487. 14. Lenhoff, H. M. and W. F. Loomis. 1957. Environmental factors controlling respiration in hydra. /. Exp. Zool. 134: 171-182. 15. Lenhoff, H. M. and H. A. Schneiderman. 1959. The chemical control of feeding in the Portuguese man-of-war, Physalia Physalis L. and its bearing on the evolution of the Cnidaria. Biol. Bull. 116: 452-460. HOWARD M. LENHOFF 229 16. LooMis, W. F. 1954. Environmental factors controlling growth in livdra. ]. Exp. Zool. 126: 223-234. 17. LooMis, W. F. 1955. Glutathione control of the specific feeding reactions of h\dra. Ann. N. Y. Acad. Sci. 62: 209-228. 18. LooMis, W. F. and H. M. Lenhoff. 1956. Growth and sexual differentiation of hydra in mass culture. /. Exp. Zool. 132: 555-574. 19. MoNOD, J. 1956. Remarks on the mechanism of enzyme induction. In Enzymes: Units of Biological Structure and Function, edited by O. H. Gaebler. Academic Press, Inc., New York, pp. 7-28. 20. Park, H. D. 1953. In \V. F. Loomis, reference 17, p. 211. 21. ScATCHARD, G. 1949. The attraction of proteins for small molecules and ions. Ann. N. Y. Acad. Sci. 51: 660-672. DISCUSSION LANE: Would you care to speculate about the nature of the gluta- thione-receptors, and their location? LENHOFF: I can only guess that the receptor is a very specific protein, probably a lipoprotein on the cell membrane. The evidence is not too good concerning the location of the receptors on Hydra. Experiments using isolated parts of Hydra show that some are lo- cated on the tentacles, and others on the hypostome. We tried to localize the receptor by radioautography using glutathione. But the glutathione washes readily off. SLAUTTERBACK: Aren't you inhibiting the oxidative enzymes severely when \'0u get down to 6 degrees and thus reduce the gen- eral motility of the animal? LENHOFF: No doubt we are slowing down many reactions by lowering the temperature, but the limiting reaction is the one that we think causes this delay in mouth opening. SLAUTTERBACK: Are these animals still moving around actively? LENHOFF: Yes. In assaying the feeding reflex, we observe mouth opening, tentacle waving and contraction. All these movements seem normal as does contraction after a mechanical stimulus. BURNETT: Maybe you could explain this preliminary experiment. We placed hydra in a 10~^' M solution of glutathione and waited until all of them had closed their mouths and discontinued their 230 THE BIOLOGY OF HYDRA : 1961 feeding response. At this time we offered the hydra several hun- dred brine shrimp. The hydra readily captured, killed, and ingested the shrimp. LENHOFF: I can give some explanation. When a Hydra punc- tures a shrimp, all sorts of new and unknown substances present in the body fluids of the shrimp flow into the media. There is a possibility that these emitted fluids contain substances which en- hance the feeding reflex. In fact, we have some preliminary indica- tions that phospholipids present in serum do just this. Since I think that the initial activation takes place on the cell membrane, it is possible that the phospholipids act there. BURNETT: I suppose it is enhancing something already present. LENHOFF: Yes. The point I want to emphasize is that it is very hard to know what is happening since you do not know what is present in the fluids coming out of the shrimp. So many factors affect the feeding reflex, as I have shown you already. Chandler Fulton also shows that Cordylophora respond some- what to shrimp after they no longer respond to proline. This may be a similar phenomenon. GOREAU: Have you tried amino acids? I ask because we recently noted that small amounts of methionine caused corals to extrude mesenterial filaments. The entire colonies become covered with tangled white masses of filaments that stayed out as long as the methionine (2 /xglO ml) was in the medium. Extrusion of masen- terial filaments is a typical response of some corals when feeding in the presence of thick plankton swarms, but I have never seen such a strong sustained reaction with other stimulants, including clam juice, as with methionine. LENHOFF: I haven't tried methionine, although I doubt whether it would cause Hydra to respond. I fully agree that other com- pounds may work on other organisms. Fulton, for example, has shown that proline activates the feeding reflex of Cordylophora. GOREAU: Glutathione seems to have little effect on those corals on which it was tried. Zooplankton swarms probably secrete detec- HOWARD M. LENHOFF 231 able amounts of all kinds of organic substances which activate chemoreceptors to trigger the corals' feeding posture. Corals feed any time there is plankton around. The classic story that reef corals expand only at night is untrue. In fact, we have frequently seen corals feeding on swarms of zooplankton in the middle of the day irrespective of light intensity, using tentacles and extruded mesen- terial filaments to catch and entangle their prey. LENHOFF: It would he nice to see whether methionine analogs will inhibit this response in corals elicited by clam juice. This would provide strong evidence that methionine is the actixe compound in the clam juice. STREHLER: Langdon found that the reduced chain of insulin is a competitive inhil:)itor of glutathione-TPN reductase. Have you tried reduced insulin? LENHOFF: We have not tried insulin or reduced insulin yet. But Langdon's finding places this experiment high on our list. Another point I find exciting is that Langdon calls insulin a "prohormone." That is, he suggests that insulin will not work unless it is first split, although here it is split by reduction. Thus, insulin may represent a case of an excitatory compound being activated by the unmasking of an essential group. We think that unmasking phenomena (possibly proteolytic) may operate in control systems generally. BURNETT: Did you repeat the experiments of Balke and Steiner showing that lactic and ascorbic acids elicited a feeding reflex? LENHOFF: Yes. I found neither lactic nor ascorbic acid to work. However, I still wouldn't be surprised if under certain conditions other compounds also activate. For example, they may act, like the proteases, along the chain of reactions involved in the feeding reflex. Perhaps lactic acid, under their conditions affected some step of the response. And there still remains the possibility that Hydra has more than one receptor. All I can say is that in Hydra littoralis all the factors that I mentioned in my talk influence the response, and that there is no question that reduced glutathione is a natural activator. 232 THE BIOLOGY OF HYDRA : 1961 BURNETT: We once found that dilute concentrations of bovine testes hyaluronidase stimulated the feeding response. At that time we assumed that our enzyme preparation was contaminated with glutathione. A more recent preparation consisting of crystals quite different from our original preparation was not effective. LENHOFF: These are factors that you have to consider. First you must dialyze to remove endogenous glutathione. For example we found some of our enzyme preparations elicted a response before dialysis but not afterwards. KLINE: Some compounds may cause the mouth to open without producing the true feeding reflex. LENHOFF: Definitely. You can get mouth opening, but not the true feeding reflex from many compounds, usually toxic ones. As Dr. Loomis pointed out in his original paper, the best proof that a compound can initiate the feeding reflex is to give the Hydra some inert material impregnated with the compound you are testing. If the Hydra ingests the inert material, then a true feeding reflex was elicited by that compound. The Nutrition of Hydra David L. Claybrook Dcpt. of Phijsiolo^ij 6- Fliarmacologij , Wayne State Universitij College of Medi- cine, Detroit, Michigan. The study of hydra nutrition is in its infancy. In fact, we are not aware of any investigation of specific nutrient requirements prior to our own. I suspect that the absence of such studies has been due to the complexity of the prol^lem rather than to a lack of appreciation for its importance. The laboratory culture of hydra was more of an art than a science until Dr. Loomis' fundamental research defining environmental conditions for optimal growth ( 12 ) . The numlier of environmental variables was then greatly reduced to the point that the rate of growth coidd be directly con- trolled by limiting the food supply. The hydra's apparent refusal to ingest non-living food made it essentially impossible to feed a formulated diet. When the gluta- thione control of the feeding reaction was revealed (3), it offered a means for feeding to the animals particulate preparations of the experimenter's choice. With these possibilities in mind, we began a study of hydra nutrition. We undertook this investigation for two main reasons. First, we wanted to know to what extent the hydra's requirements were similar to and different from those of other animals. With the exceptions of the protozoa and insects, very little work has been devoted to the nutrition of invertebrates. Information on coelen- terate nutrition would contribute significantly to our knowledge of comparative biochemistry. The second, and primary purpose of the project was to in- crease the usefulness of the hydra as a biological material for the iln part from a dissertation submitted to the Graduate School, The University of Texas, in partial fulfillment of the requirements for the degree of Doctor of Phi- losophy, August, 1960. 233 234 THE BIOLOGY OF HYDRA : 1961 chemical study of development and differentiation processes — a field to be discussed by Dr. Eakin. Since the nutritional state of an animal affects all of its physiological processes to some extent, it is desirable to be able to control the nutritional state during the study of other physiological phenomena. The development of chem- ically defined nutrient preparations in which cultures of hydra or hydra cells could be grown aseptically would give the investigator complete biochemical control over the organisms. Our ultimate goal was the propagation of hydra cells in a chemically defined medium. With this sytem, we should be able to determine the role of each tissue layer in processes such as regeneration and sexual differentiation. However, we chose to begin our experimentation with whole animals for two reasons: in general, the requirements for cell propagation are much more critical than those for growth of the intact organism. In the intact animal, speci- fic trace nutrients may be supplied by specialized cells. There is also a more rapid loss of essential nutrilites to the external solution from the isolated cell. We hoped to discover approximate require- ments before proceeding to precise studies at the cellular level. Secondly, techniques for quantitative study had already been de- veloped for whole hydra but not for dissociated cells. Thus the nutritional value of an experimental diet could be determined by its effect on an observable physiological process such as asexual growth. Our stock Hydra clone was obtained from a locally-isolated strain of Hydra littoralis, and was grown according to the methods of Ham and Eakin ( 1 ) , When fed daily with an excess of freshly- hatched Artemia larvae, the Hydra grows at a maximum logarith- mic rate. Presumably the animal receives an excess of all exogenous requirements, and the limitation of growth rate is due to necessary metabolic conversions within the cells. If the exogenous supply of a growth factor is reduced below the maximum utilizable by the animal, a reduction in the observed rate of growth should result. In the search for a non-living diet, it was found that heat- killed Artemia would support asexual growth of Hydra for at least six months, but at a rate significantly below maximum. Al- though the killed Artemia contained adequate amounts of reduced glutathione to stimulate the feeding reaction, the solution had to be DAVID L. CLAYBROOK 235 stirred gently to bring them into contact with the Hydras tentacles. The effect of the period of heating on the subsequent growth rate is depicted in Table 1. The reduction in the growth rate is seen to be progressive with time of heating. This indicated to us that some substance was being inactivated by the heat treatment so that its availability to the Hydra became limiting to the growth process. TABLE 1 Relation of growth rate of Hydra littoralis. Ham strain, to period of heating of Artemia (70°) Growth Rate 1.9 3 3.3 7 3.5 15 4.5 30 5.2 U- = hi 2 T Heating time Doul:)hng time Growth Rate Constant (min.) (days) (/c)** .36 .21 .20 .15 .13 On the assumption that replacement of the growth-limiting fac- tors to a nutritionally deficient diet would increase the rate of ]:)udding, we assayed numerous biochemical and biological sub- stances for their capacities to stimulate budding when added as supplements to a diet of heated shrimp. Artemia heated for 7 minutes at 70 were fed to Hydra cultures for at least a week before the Hydra were used for bioassays. This period served to deplete the animals of any reserve of the growth factor. The heated Artemia diet was first supplemented with defined and complex substances dissolved in the salt solution, bathing the Hydra. Natural extracts, vitamins, amino acids, and other possible growth factors, alone and in various combinations, were tested in this system. No stimulatory effect on the growth rate was observed in any expriment. Since the lack of growth response to external supplements could have been due to relative impermeability of the ectodermal cells to dissolved materials, it was necessary to devise a technique for introducing the test materials directly into the coelenteron 236 THE BIOLOGY OF HYDRA : 1961 where normal absorption could take place. A diagrammatic pic- ture of the apparatus which was designed to inject a measured volume into the individual organism is shown in Figure 1. The apparatus features a micrometer-driven micro-liter syringe for de- livering quantities in the micro-liter range. ^) ^ fe to foot switch for motor to foot pedal for release bar Fig. 1. Micro-injector for feeding Hyc/ro. In our standard injection test, adult Hydra without buds were selected from the cultures maintained on heated Arteniia. The animals were placed in 9-depression spot plates in large Petri dishes, and each one was force-fed 0.2 [xl from a glass capillary containing semi-solid agar in which the experimental diet was dissolved or suspended. Twenty-four hours after injection, the ani- mals were examined under a dissecting microscope, and the num- ber of new buds in each dish of nine Hydra was recorded and compared with that of the unfed control dish. The relation of growth response to the quantity of material injected is shown in Figure 2. The response was proportional to the DAVID L. CLAYBROOK 237 B "^^h 1.0 0.8 0.6 BUDDING RESPONSE TO INJECTED FRACTIONS • Woter- insoluble Fraction of Liver A Water-soluble, non- diolyzoble Fraction of Liver Water- insoluble Fraction of Artemio ^ - '^''^ ^ ^ '^--'^ y ,^ y u. * y ,' / ' 1 1 1 1 1 1 1 1 1 1 1 1 hydro (Average of 9 Replicates) 0.4 0.2 I 2 4 6 8 10 20 30 ^^/hydro Fig. 2. Budding response to injected fractions. logarithm of the dosage in some cases, while in the other experi- ments log responses were not observed. The relative potencies of some natural materials showing ac- tivity in this system are listed in Table 2. The potencies on a dry TABLE 2 Relative activities of natural supplements for promoting budding in Hydra Potency Material (dry weight) Bovine liver acetone powder" 10 Mouse liver 10 Mouse kidney 8 Mouse heart 10 Chick embryo extract 10 Escherichia coll 6 Dried yeast 10 Chlorella ellipsoiclea 4 Bovine liver extract, 10 non-dialyzable fraction 'Used as standard and assigned arbitrary value of 10. 238 THE BIOLOGY OF HYDRA : 1961 weight basis are expressed relative to an arbitrary standard, bovine liver acetone powder. Activity was found in micro-organisms as well as in crude mammalian tissues. Substances with no demon- strable activity when fed internally included vitamins, amino acids, protein digests, nucleic acids, carbohydrates, and microbiological media. An active soluble preparation, bovine liver extract, was sub- jected to a number of physical and chemical tests in an effort to characterize the active constituents. The results of such tests are shown now in Table 3. The activity was found to be non-dialyzable, TABLE 3 Potencies of modified non-dialyzable soluble extract Fraction of Treatment Total Solids Potency Unmodified non-dialyzable extract 1.00 10 Ashing 0.02 Heating ( 2 hours, 70° ) Soluble fraction 0.20 Precipitate 0.80 12 Trypsin digestion 1.00 1-2 and was destroyed by ashing, characteristic of an organic macro- molecule. Heating in solution precipitated but did not destroy the active material. Incubation of the extract with trypsin or chymo- trypsin resulted in the disappearance of nearly all biological ac- tivity. The ultra-violet absorption spectrum of the soluble extract TABLE 4 Potencies of ammonium sulfate fractions of non-dialyzable soluble extract Ammonium Sulfate* Fraction of Fraction Total Solids Potency 1.00 10 0-33% 0.28 10 33-66% 0.50 10 66-100% 0.13 1 Soluble at 100% 0.09 'Fraction precipitated between the indicated points of saturation. DAVID L. CLAYBROOK 239 showed a peak absorption near 280 m/x, and the optical density per milhgram of extract indicated a high percentage of protein. All evidence, then, indicated that the active species were included in the protein fraction. Fractionation of the active extract with ammonium sulfate (Ta- ble 4) revealed that all activity was salted out, but was distributed Buds 0.6 - 0.4 - 0.2 - 0.0 Incubation Time in Hours at 37°C Fig. 3. Effect of period of incubation with chymotrypsin on growth pro- moting activity of non-dialyzable soluble extract. among several fractions. While supporting the conclusion that the active components were proteins, this data showed that the activity was apparently common to several classes of protein. The rate of inactivation of the extract by chymotrypsin is shown in Figure 3. From this curve it would appear that intact protein molecules, or relatively large fragments of them, are essential to activitv in this extract. 240 THE BIOLOGY OF HYDRA : 1961 It is interesting to note that all purified proteins which have been assayed were inactive in this system. These include casein, bovine albumin, insulin, hemoglobin, and six bovine plasma frac- tions. The wide distribution of activity in crude protein fractions, contrasted with the absence of detectable activity in highly purified proteins, suggests that the growth-stimulating factors could be small molecules bound firmly to crude protein, but removable by repeated purification. The evidence at hand has not enabled us to identify the Hydra growth-promoting principle with any specific previously recognized growth factors for other organisms. While the micro-injection technique has been a very useful method in the initial investigation of nutrition, it is still a tedious procedure because of the individual attention required for each Hydra. The mass culture of intact animals on a defined diet would obviously require different methods. It appears from consideration of other tissues cultured in vitro that the absolute biochemical re- quirements can be determined only by study at the cellular level. With the current progress toward maintaining coelenterate cells in vitro, the time may be near when hydra cells may be used in nutritional research. I think the significance of our own experiments lies not in the determination of specific nutrient requirements, but in the demon- stration that Hydra can live and grow on a non-living diet, and that nutrition of Hydra can be studied quantitatively by its effects on a measurable physiological process — namely the asexual growth process. Although we have only made a start toward understand- ing the nutrition of Hydra, we are encouraged to believe that it is a step toward developing the full potential of hydra as an experi- mental system. REFERENCES 1. Ham, R. G., and R. E. Eakin. 1958. Time sequence of certain physiological events during regeneration in hydra. /. Exp. Zool. 139: 33-54. 2. LooMis, W. F. 1954. Environmental factors controlling growth in hydra. /. Exp. Zool. 126: 223-234. 3. LooMis, W. F. 1955. Glutathione control of the specific feeding reactions of hydra. Ann. N.Y. Acad. Sci. 62- 209-228. DAVID L. CLAYBROOK 241 DISCUSSION STREHLER: Do }'Ou need to include any particles along with these soluble proteni fractions that were capable of supporting growth? CLAYRROOK: Well, our solutions were centrifuged for six hours at 33,000 g, which means that any sur\iving particles must have been rather small. LENHOFF: I think what Dr. Strehler is getting at is that perhaps the protein is being coagulated in the gut and is being engulfed as particles. We have some evidence that H. littoralis gastrodermis takes up mostly particles and leaves free amino acids behind in the gut (Lenhoft, H. 1961. Exptl. Cell Research, 23: 335-353). Thus, maybe the proteolytic enzymes destroy the growth-promoting prop- erties of the heat-labile protein by reducing it to a non-particulate solution of free amino acids that cannot be taken up by the gastro- dermis. CLAYRROOK: We don't know what happens after it gets inside the gut. GOREAU: What is Hydras digestive juice made out of? CLAYRROOK: I have no information on this. Do others? LENHOFF: We have fed about a million H. littoralis with shrimp, until we knew, by other measurements, that the food was mostly taken up by the gastroderm. Then we forced the Hydra to re- gurgitate, took the extract, and precipitated it with 80% ammonium sulfate. We found that there was proteolytic activity at pH 2.5 and 7. These proteolytic enzymes probably aid in degrading the cells into particles. But I doubt that the extracellular enzymes degrade the particles all the way to free amino acids, because the particles, when small enough, are rapidly phagocytized by the gastroderm. GOREAU: The reason I ask is that Claybrook's very lovely method allows one to withdraw things as well as introduce them, and I was wondering if one could do microchemical analyses on contents of the gut of the animal during various stages of digestion? CLAYRROOK: I haven't tried this at all. I don't know. 242 THE BIOLOGY OF HYDRA : 1961 KLINE: When you maintain the Hydra on heat killed Artemia, does the growth rate remain constant, even if reduced? CLAYBROOK'. Fairly constant. It varies slightly with the various lots of shrimp. KLINE: Then you didn't totally destroy something that is needed. Perhaps you reduced its concentration. How did you interpret the results? CLAYBROOK: The growth factor is not completely destroyed, but becomes limiting to growth. KLINE: In one experiment you had heat precipitated material on which the Hydra were able to grow quite well. CLAYBROOK: Right. This is the liver extract. We have not fractionated shrimp because the relative supply of liver and shrimp are not the same. LENHOFF: Is it possible that the more you heat the shrimp, the more the shrimp's cellular integiity is destroyed? And when you put these damaged shrimp in water, essential factors leak out? A few years ago Dr. Loomis and I were able to grow Hydra on frozen shrimp, but had no success with boiled shrimp. We thought then that boiling either destroyed a heat labile factor or allowed essential heat-stable factors to leak out. CLAYBROOK: It is possible, but in a few experiments we found no activity in the supernatant that the shrimp were boiled in. I wouldn't say this was conclusive. LENHOFF: Was this supernatant solution either ninhydrin or protein-positive? CLAYBROOK: We didn't check at this stage but I'm sure that there were ninhydrin-positive components there. GOREAU: What I am speculating on now assumes a nervous system! Living AHemia may be required because the struggle with the prey could set up a reflex which causes hydra to secrete enzymes or produce preabsorptive changes in the gastroderm, which would allow digestion to proceed in a much more complete DAVID L. CLAYBROOK 243 manner. The point is this. Perhaps the animal needs to struggle with its prey? This is, of course, a complete speculation but we may be dealing here with a phenomenon on the physiological rather than biochemical level. CLAYBROOK: I haven't tried any experiment which would answer your question. LENHOFF: Didn't Hijdra grow well on frozen shrimp? CLAYBROOK: They grow at a reduced rate. LENHOFF: At a very reduced rate? CLAYBROOK: Not very reduced. But below that found with live shrimp. The answer to this may also be leakage from the shrimp. LENHOFF: But they do grow on frozen shrimp. I would think that this would answer Dr. Goreau's speculation l^y showing that the struggling of live prey is not required. LOOMIS: It is interesting that apparently no carbohydrate is necessary. In other words, pure protein is enough. CLAYBROOK: Let's say carbohydrate is not limiting at this state. LOOMIS: But you feed them solely on the 0-66% ammonium sulfate fraction of liver protein? CLAYBROOK: This alone will not support continued growth. This is only a specific assay for the heat labile factor. EAKIN: I think that some of you were not able to hear Dr. Claybrook clearly when he described his method for demonstrating the requirements of Hydra for the heat-labile factor. The Hydra which he used as test organisms had been cultured at a sub-optimal level of nutrition by feeding them on heated brine shrimp. The response we studied was that of boosting them from this bare maintenance level to that which we get when they are fed live brine shrimp. LOOMIS: It is a specific assay for the heat-labile factor? EAKIN: That's right. Even the poorly growing controls are get- ting a highly complicated diet in the heated Artemia. 244 THE BIOLOGY OF HYDRA : 1961 LOOMIS: I have often tried to get micro-pipettes into the mouths of H. littoralis and out again without having them then regurgitate what I put in their stomachs. Perhaps you open their mouths with gkitathione? CLAYBROOK: No, I force it open. LOOMIS : Maybe that's the secret! CLAYBROOK: Yes. Then I wait until he closes his mouth on the pipette before injecting the material. LOOMIS : And then it is water-tight as you pull out? CLAYBROOK: The semi-solid consistency of the medium is es- sential here. You can't use a liquid. LOOMIS : How much agar do you use? CLAYBROOK: I use 0A%, which is relatively thick. The Hydra closes its mouth when the pipette is withdrawn and the viscous solution remains in the gut. LOOMIS : Will it flow down a microcapillary? CLAYBROOK : Yes, if under pressure. Isolation and Maintenance in Tissue Culture of Coelenterate Cell Lines John H. Phillips Department of Baeteriolofiij, University of California, Berkeley, California The in vitro cultivation of coelenterate tissues has been reported before (1, 2). However, attempts at the maintenance of such cul- tures for prolonged periods of time and serial transfer of cultured material have not apparently met with success. In addition, the evi- dence in support of true multiplication of cells has not been entirely convincing. The methods that will be discussed have led to the establishment of cell cultures from the anemone Anfhopleura ele- gantissima. These cultures have been transferred twenty to thirty times and have been under in vitro cultivation for more than a year. In addition, eight of the cell lines have been through one single cell cloning. The resulting clones of eight to thirty-two cells have given rise to cultures containing 10*^ to 10" cells. First will be described the procedures which have been used in isolation, cultivation and cloning of the cells, and this methodol- ogy will be followed by a description of the cells and some of their properties. A somewhat more detailed account of methods will soon be published (5). All glassware and rubber stoppers were cleaned by autoclaving in 0.1% NaoCO... (4). Glassware was wrapped in aluminum foil and sterilized by dry heat. Rubber stoppers were autoclaved in large Petri dishes. All nutrient solutions were steri- hzed by filtration, using Millipore filters. The nutrient medium that has been found to be most useful consists of 0.7% Edamine\ 1 Sheffield Chemical Company, Inc., Norwich, N. Y. 245 246 THE BIOLOGY OF HYDRA : 1961 an enzymatic digest of lactalbumin, in 90 /f artificial sea water (3) containing 500 units of penicillin and 0.5 mg. streptomycin per ml. Growth can be obtained over a range of Edamine concentrations from 0.4 to 1.5% and artificial sea water concentrations of 40% to 100%. Yeasts and molds which are not inhibited by the antibiotic mixture have at times presented difficulties. Mycostatin has been used at a cencentration of 50 units/ml. to free cultures of these contaminants. This antibiotic has not been included routinely be- cause it appears to be somewhat toxic to the anemone cells. The cell suspension used for the initial isolations is prepared by mincing the animal in a beaker with a pair of scissors. In some cases, lysozyme ( 1.5 mg./ml. of animal) was added to degrade the mucus that is secreted. Approximately five volumes of arti- ficial sea water is added per volume of minced tissue, and the mixture is stirred briefly. It is allowed to stand in an ice bath for approximately five minutes and filtered through two layers of cheese cloth. The filtered suspension is freed of large tissue frag- ments by centrifugation at 5 for 30 seconds at approximately 1000 g. The cell suspension containing very few tissue fragments is then centrifuged as above for 10 minutes. The cells are resuspended in sterile artificial sea water and the differential centrifugation is repeated. The cells are finally washed three more times with sterile artificial sea water. The last washing employs artificial sea water to which has been added antibiotics at the above-mentioned concentration. The cell suspension is diluted to a concentration of close to 3 X 10' cells/ml. This corresponds to 0.100 O.D. at 660 m/x in the Beckman Spectrophotometer Model DU and is about equal to 100 [ig. of cell protein/ml. or 5 X 10~^ ml. of packed cells/ml. One-tenth ml. of such a suspension is used as the culture inoculum. Figure 1 shows the appearance of such a suspension. There is great heterogeneity of cell type, and the outstanding con- taminant appears to be fragments of fibrous material from the mesoglea. The two comet-shaped objects in the center of Figure 1 are this material. The cells show a size range from 8—2 /x. The inoculum is placed in either a 60 mm Petri dish or into a test tube of 15 X 130 mm. containing a piece of coverslip 10 X 20 mm. Five ml. of nutrient solution is added and mixed with the inoculum. The test tubes are slanted to allow the cells to settle JOHN H. PHILLIPS 247 <» Fig. 1. Suspension of cells obtained from A. elegantlssima. Stained with periodic acid Schiff's. Magnification 900x. on the piece of coveislip. Either kind of preparation is incubated at 15°. The cultures are examined microscopically at a magni- fication of lOOx. Figure 2 shows well-developed clones growing on the side of a tube culture. The piece of coverslip may be removed from such cultures and used for more detailed examination. Grow- ing cultures can be maintained in tubes for prolonged periods of time, provided that fresh nutrient solution is added at weekly intervals. A suspension of cells can be obtained for transfer to new cultures by simply scraping some of the growth from the glass surface with a sterile spatula or through the use of lysozyme 1.5 mg./ml. in 0.3 M ethylene diamine tetracetic acid adjusted to pH 8.3 with NaOH. In either case, the final dispersal of the clumps of cells requires agitation. Generally, the suspension is drawn back and forth through a pipette. Cell suspensions may be standardized as indicated above; however, complete dispersal is generally not attained. The isolation of clones developing from 248 THE BIOLOGY OF HYDRA : 1961 .♦ .•> Fig. 2. Clones of A. elegantissima growing on the side of a test tube. Magnification 129x. single cells is generally made difficult by the slight movement of cells over the surface of the glass. Therefore, cloning procedure of Puck (7) is generally used. Cells are mixed w^ith 10 ml. of nutrient medium containing 0.2 /y agar, and the mixture is placed in a Petri dish containing a layer of 1% agar in artificial sea water. Developing clones are observed as clusters of cells that are generally separated from one another by a distance equal to the cell diameter. Development from a four through a thirty-two cell stage can be observed. The generation time is somewhat in excess of twenty- four hours. The clone can be removed with a capillary pipette and transferred to fresh nutrient solution. Because of the dis- tinctive appearance of a developing clone, there is no difficulty in avoiding clumps of cells which were present in the inoculum. The appearance of the cells growing in vitro shows certain JOHN H. PHILLIPS 249 m Fig. 3. Twelve-day-old clone of cells from A. elegantissima. Stained with periodic acid Schiff's. Magnification 900x. peculiarities, some of which it is hoped may be corrected through the use of a better nutrient medium. Suspensions of single cells obtained either directly from animals or from cultures do not show reaggregation. On the contrary, a developing clone generally shows outgrowth and separation of cells from the growing center. The separated cells occasionally move a short distance before becoming new centers of growth. Figure 3 shows a twelve-day-old clone developing on a cover- slip. The preparation was fixed in methanol and stained with periodic acid Schiff's stain (5). The cells are filled with a granular material that makes observation of the nucleus very difficult. These granules, when observed in living cells by phase contrast micros- copy appear as barred objects resembling mitochondria. Similar intracellular structures can be observed in suspensions of cells obtained directly from the animal, but such cells do not show as high a concentration of these objects. When these cultured cells are removed to artificial sea water containing ethylenediamine tetraacetic acid (EDTA), they rapidly change their appearance to that shown in Figure 4. The addition of sodium acetate to 0.1% Edamine medium appears to produce a similar effect which is under investigation at the time of this writing. Until the concentra- 250 THE BIOLOGY OF HYDRA : 1961 Fig. 4. Cultured cells washed with artificial sea water. Stained with periodic acid Schiff's. Magnification 900x. tion of these particles can be controlled, observation of mitosis in developing clones is impossible. The cells, particularly those toward the center of the clone in Figure 3, are surrounded by a red staining, carbohydrate-con- taining material which apparently acts as an intercellular cement. It can be weakened by both lysozyme and EDTA, but these agents even in combination do not result in complete separation of the cells. Since lysozyme functions as a i8(l-^4) N-acetyl hexosamini- dase (8), the presence of this carbohydrate derivative in the material appears likely. The material is not susceptible to the chitinase of Helix pomatia, hyaluronidase, nor trypsin. The action of EDTA suggests either the presence of bridges formed by diva- lent ions or possibly the activation of the lysozyme-like enzyme that has been detected in the secretions of these animals (6). Pollak's trichrome stain (9) has also been used in studies of this material. It is again stained red. This staining reaction is given by elastic fibers. Mucus assumes a green coloration by this stain- ing procedure. It appears likely that the material in question is other than mucus. Until the nature of this material is better under- stood, the methods for its degradation are available, quantitative JOHN H. PHILLIPS 251 work— for example, the accurate determination of generation time and cloning efficiency— is made difficult. The ease with which cell lines from this anemone can be established and maintained in the laboratory is encouraging. It will be of interest to determine if the cells of other coelenterates behave in a similar manner. These studies were supported l:)y grants from the National Science Foundation and the United States PubHc Health Service. REFERENCES 1. Gary, L. R. 1931. Report on invertebrate tissue culture. Carnegie Inst. Wash. Yr. Bk. 30: 379-381. 2. Lewis, M. R. 1915-1916. Sea water as a medium for tissue cultures. Anat. Rec. 10: 287-299. 3. MacLeod, R. A., E. Onofrey and M. E. Norris. 1954. Nutrition and metabo- lism of marine bacteria. 1. Survey of nutritional requirements. /. Bad. 68: 680-686. 4. Madin, S. H., p. C. Andriese and N. B. Darby. 1957. The in vitro cultivation of tissues of domestic and laboratory animals. Amer. J. of Vet. Res. 69: 932-941. 5. Phillips, J. H. In vitro maintenance and cultivation of cells from marine invertebrates. Methods in Medical Research (in press). 6. Phillips, J. H. Immune mechanisms in the Phylum Coelenterata, Second Annual Symposium on Comparative Biology. The Lower Metazoa: Comparative Biology and Phylogenij. To be published by Academic Press, N. Y. 7. Puck, T. T., P. I. Marcus and S. J. Cieciura. 1956. Clonal growth of mam- malian cells in vitro. ]. Exp. Med. 103: 273-284. 8. Salton, M. J. R. and J. M. Ghuysen. 1959. The structure of di and tetra sac- charides released from cell walls by lysozyme and streptomyces Fi enzyme and the y8(1^4) N-acetyl hexosaminidase activity of these enzymes. Biochim. et Biophys. Acta 36: 552-554. 9. Sano, M. E. 1949. Trichrome stain for tissue section, culture, or smear. Amer. J. Clin. Path. 19: 898. DISCUSSION MUSCATINE: Was the animal kept in artificial sea water? PHILLIPS: Our artificial sea water preparation is capable of maintaining the intact animal for a long period of time, but they are normally kept in real sea water. MUSCATINE: Is there any particular criterion that you use for the well-being of the animal? 252 THE BIOLOGY OF HYDRA : 1961 PHILLIPS: No. It just continues to look healthy. GOREAU: That is a beautiful piece of work. Do you know what cell types your cultures actually come from? Have you tried adding zooxanthellae? PHILLIPS: I haven't tried adding zooxanthellae. Some people at Stanford are interested in this problem. I am planning to give them my cultures to do this. With respect to the cell that I have growing in culture, this becomes an extremely difficult question to answer. For one thing, the appearance of the cells growing in culture may be markedly different from the cells that one sees in the intact animal as all the cells tend to round up on being freed from the tissue mass. This makes it impossible, on the basis of cell shape, to decide whether it is endoderm, mesoglea, or ectoderm. GOREAU: Perhaps you could start your cultures with scrapings from specific areas rather than the whole animal. PHILLIPS: This is something we want to try. I have not devoted a great deal of work to these culture lines although I've had them in the laboratory for sometime. GOREAU: A very important matter to anyone who has ever tried to dissect living coelenterates is the horrible problem of being flooded with mucus. Are you actually cutting this down with lysozyme? PHILLIPS: Definitely. There is one trick to that. The lysozyme should not be added to sea water. High electrolyte concentration is quite inhibitory to the action of lysozyme. It decreases its activity by almost 50%. That's the reason I add it directly to the animal before mincing the tissue. There is another thing I should mention, namely, the use of fluorescent antibody techniques for identification of materials with- in tissue. I have carried out work of this sort with these cells using rabbit anti-anemone serum and fluorescently labeled dog anti- rabbit globulin serum. This leads to a nice fluorescent uptake by the cells growing in culture, and it also results in an uptake of fluorescence by whole cell suspensions. But, I would not care to put this forth as anything but supporting evidence for these cells JOHN H. PHILLIPS 253 being from the anemone. I think this proof must come from repeated isolations, such as we carried out, and from a comparative study of the morphology of the cells. Also a consideration of the cloning efficiency assists in discarding the possibility that the cultured cell is some parasite present in small numbers within the animal. WAINWRIGHT: Have you tried collagenase on the intercellular material? PHILLIPS: No. Those are the only enzymes I have tried so far. It is resistant to trypsin and hyaluronidase but degraded by lyso- zyme. WOOD: I was not quite clear about your statements concerning the mitochondria. Have you tried a specific mitochondrial staining technique or do you have other criteria for identification? PHILLIPS: No. I simply said that they resembled mitochondria in that they were markedly bar shaped. That's all. STREHLER: Is there only one morphological type of cell? PHILLIPS: One sees a variety of cell types in developing cul- tures. For example, the ratio between nuclear and cytoplasmic size varies as well as the distribution of the granules within the cells. At the same time clone cultures derived from a single cell also shows this variation. SLAUTTERBACK: If the anemone is anything like hydra, you can determine whether or not they are gastroderm cells by expos- ing the animal to a thorotrast solution for a short time. Thorium dioxide serves as an excellent tag because only gastroderm cells pinocytize it. WOOD: Could you be certain that free cells derived from ecto- derm would not pinocytize or phagocytize some thorotrast? PHILLIPS: These cells do show a rapid uptake of such ma- terials as bovine and human serum albumin. If one labels such proteins with azo dyes within 15 minutes you get cells with brightly stained inclusions and the cells remain colored for long periods of time. In fact, it was in connection with immunological studies that I first became interested in cultivating these cells. 254 THE BIOLOGY OF HYDRA : 1961 SLAUTTERBACK: In response to Dr. Wood's comment, my suggestion was that the animal be exposed to the colloidal thorium dioxide before it was cut up. In that case there would be no thorium in the ectoderm cells. PHILLIPS: True, if the label remains. PASSANO: What is the chromosomal integrity in your clones over a period of time? PHILLIPS: I don't know. Until I can get rid of these granules and control their formation I do not want to even attempt to ob- serve mitotic figures. STREHLER: Does colchicine block their mitosis? PHILLIPS: I have not tried it yet. Symbiosis in Marine and Fresh Water Coelenterates Leonard Muscatine Laboratories of Biochcmistrij, Howard Hughes Medical Instittttc, Miami, Florida In studying the significance of symbiotic algae for the nutrition and growth of their invertebrate hosts, we have been guided by two objectives : a ) to estabhsh the existence of a nutritional relationship between algae and host, and b) to characterize the chemical basis of this relationship. Direct evidence for the contribution of carbon compounds from symbiotic algae to the tissues of the host has been demonstrated in a sea anemone ( 9 ) , a coral ( 3 ) , and in green hydra ( 5 ) . In this paper, we demonstrate a direct relationship between algal symbionts and changes in mass or growth of a marine and a fresh- water coelenterate. Our data show that retarded weight loss, en- hanced growth, and prolonged survival of the animals studied could be attributed to the presence of symbiotic algae. STUDIES ON SEA ANEMONES' - Experiments demonstrating retardation of weight loss were con- ducted on Anthopleura elegantissima (Brandt, 1835), an intertidal anemone which contains zooxanthellae within its gastrodermal cells (Fig. 1). Specimens without algae, found beneath fish canneries ^Part of a tliesis submitted in partial fulfillment of the requirements for the degree of doctor of Philosophy, Department of Zoology, University of California, Berkeley. -This investigation was supported by a fellowship (EF-9653) from the National Insti- tutes of Allergy and Infectious Diseases, Public Health Service. 255 256 THE BIOLOGY OF HYDRA : 1961 ^^' -1^'- Fig. 1. Electron micrograph of a transverse section through a musculo- epitheiial cell of an anemone showing the intracellular location of an algal cell, a) animal cell, b) algal cell, c) chromatophore. d) pyrenoid. (Prepared with the assistance of Miss Jane Westfall) at Pacific Grove, California, served as controls and are referred to as albinos. In order to evaluate quantitatively the effect of the algae on the nutrition of the host, we measured changes in weight of normal and albino anemones starved in light and darkness for 11 weeks. Reduced weight, i.e., weight under water, was used to measure weight changes. This method eliminates error from surplus fluid and, in contrast to dry weight, allows repeated measurements on living individuals.^ ^There were no major changes in the specific gravity of the animals themselves during the course of the experiment, showing that all of the observed changes were true weight changes. LEONARD MUSCATINE 257 Fig. 2. Arrangement of apparatus for rapid measurement of the reduced weight of a sea anemone in sea water of known temperature and density. The animal is suspended by a thin constantan wire hooked into its actino- phorynx. Two groups of five normal anemones were placed into aerated containers of twice-filtered sea water at 14.0^ ± 1.5^. One group was continually illuminated by 200 ft. c. of fluorescent illumination ( Champion— Warm White) while the other was kept continually in darkness. Both groups were allowed to starve.^ The reduced weight of each individual was measured (Fig. 2) at intervals of four days or more and the sea water in all containers was renewed weekly. Individuals in darkness were weighed in dim light. As additional controls, two groups of five albino anemones were treated in a manner identical to the normal svmbiotized anemones. Details ^Fed anemones were unsatisfactory experimental animals. Erratic behavior ( e.g. pre- mature egestion, failure to feed) interfered with attempts to control feeding. 5.58 THE BIOLOGY OF HYDRA : 1961 o NORMAL, LIGHT • NORMAL, DARK K. - • A ALBINO, LIGHT 3: - t A ALBINO, DARK ^ 5 - 8 • • -J • 5 oo i^ _ t o G 1 10 — ^ ^ Ci i- A i • • n * § 5 15 - A 1 • u c^ - ▲ • \ o _ A ▲ 8 o o ;:£ 20 — ▲ A .A • o - ^ ▲ \ 8 ^ _ ▲ • p 5 25 : A t • • t • • • K - A • • Uj 30 - A ▲ • • ^ - A t t • Ci: - A ki ~ a. - 35 - 40 - 1 1 1 1 1 A 1 1 1 1 1 1 23 456789 10 II 12 TIME IN WEEKS Fig. 3. Chonge in reduced weight of normal and albino anemones starved in light and darkness. The ordinate is the percent change from initial weight and denotes weight loss. of other methods in these experiments are given elsewhere (8). Reduced weight changes of symbiotized and albino anemones starved in light and darkness are depicted in Figure 3, and expressed as percent change from initial weight vs. time in weeks. The results show that durmg starvation all anemones lost weight at a near con- stant rate, and that symbiotized anemones lost weight at about half the rate of albinos. The possiliility tliat light could have directly LEONARD MUSCATINE 259 affected weight loss was tentatively ruled out since albinos in light and darkness lost weight at the same rate. We therefore conclude that the lower rate of weight loss by symbiotized anemones is re- lated to the presence of algae. These observations, along with evidence from tracer studies (9), suggest that during starvation carbon contributed by the algae, together with host excretory nitrogen, is used for the synthesis of organic compounds necessary for maintenance of weight. This view emphasizes the possible secondary role of the algae in reclaiming waste nitrogen. STUDIES ON HYDRA With the introduction of techniques for the mass culture of Hydra (7) an opportunity presented itself for quantitative studies on plant-animal symbiosis in the laboratory. Chlorohydm viridis- sima/' a hydra containing zoochlorellae within its gastrodermal cells ( 14 ) , may now be grown under controlled environmental conditions and in a fluid of known ionic constitution. Growth may be measured in terms of protein or logarithmic increase in number of hydranths (6). In addition, problems encountered with sea anemones, such as removal of algae and erratic feeding behavior, are easily resolved. C. viridissima was routinely grown in our laboratory in the fol- lowing culture medium ("M" solution): 10 "^M Tris (hydroxy) methylaminomethane buffer, pH 7.6, lO'-^M CaCL, 10 'M NaHCOs, 10-W KCl, and IQ-^M MgCL, in de-ionized water. Algae-free C. viridissima were obtained by growing green indi- viduals in M solution plus 0.5% glycerin (v/v) for 7-10 days, following the original technique of Whitney (13). These albinos then grew normally in M solution and did not regain an algal flora. Growth studies. All growth experiments were conducted at 21°-23° using the method of Loomis (7). Ten hydranths (five uni- form hydra, see ref. 4) from mass cultures were put in 30 ml. of M solution in shallow Petri dishes placed four inches from a single 40-watt fluorescent light ( Sylvania-Cool White). Daily, the ^Tentative identification. 260 THE BIOLOGY OF HYDRA : 1961 FED DAILY GREEN ALBINO 10 TIME IN DAYS Fig. 4. Semi-log plot of growth rates of duplicate cultures of green and albino C. yiridissima fed daily in the light. number of hydranths was counted and then each hydranth was fed on a dense suspension of Artemia nauphi. One hour after feeding and again, six hours later, the culture medium was renewed. This routine was followed for 5-7 days. Figure 4 shows that green and albino C. viridissima, when fed daily, have nearly identical logarithmic growth rates. These results imply that the algae do not contribute anything to the host that cannot be acquired from an exogenous food supply. Under optimal conditions, nutritional benefit would not be expected to manifest itself in terms of growth of the host because the maximum growth rate (kmax), a property intrinsic to the species, cannot be exceeded, regardless of the magnitude of the algal contribution. Therefore, we conducted growth experiments in which the amount of food was limited, reasoning that this would then allow benefit from the algae to express itself. Figure 5 demonstrates that, when fed every second LEONARD MUSCATINE 261 10 FED EVERY SECOND DAY GREEN t ALBINO TIME IN DAYS Fig. 5. Same as Figure 4 but fed every second day. Arrow denotes time of feeding. day, green liydra deviated only slightly from normal logarithmic growth. But under the same conditions, albino hydra showed not only a more pronounced deviation, but also required more time to regain normal growth after feeding was resumed. More striking dif- ferences appeared when these two groups were fed every third day. Figure 6 illustrates the sharp decline in rate of budding by both groups during the interval without food. But after feeding, green hydra immediately resumed a normal maximum growth rate. In contrast, growth of albinos lagged and did not return to normal. The effect of complete elimination of food is shown in Figure 7. Ten hydranths of each kind were removed from mass cultures and starved in the light in 30 ml. of M solution in Petri dishes (4" diam.). The culture medium was renewed once daily. Under these conditions, green hydra continued to produce buds for 7-9 days and 262 THE BIOLOGY OF HYDRA : 1961 ^ 80 10 FED EVERY THIRD DAY GREEN ^ ALBINO t J I \ \ \ L TIME IN DAYS Fig. 6. Same as Figure 5 but fed every third day. survived an additional 7-10 days until gradual diminution in size resulted in death. In contrast, albino C. viridissima, under these conditions, stopped budding after 1-3 days; within the next six days, all had disintegrated. These results show that the algae are essential for prolonged survival under starvation conditions. Early disintegration and death of albinos was unusual since a characteristic of most species of hydra, including green C. viridis- sima, is to gradually "waste away" when starved ( 1 ) . One explana- tion of this death gives us a clue to a possible nutritional role of the algae. Dixon (2) has stated that tissue death results from inability to synthesize coenzymes. By removing algae from C. viridissima we may have removed a source of coenzymes, or coenzyme precursors, normally available from algae during starvation or from food during normal feeding conditions. This idea fits well with results of limited food experiments, where green hydra show maximum growth imme- LEOXARD MUSCATINE 263 co20 ^ 15 o 10 I 5 - STARVED • • — • o • o • • o o o 1 1 1 1 8 10 TIME IN DAYS Fig. 7. Bud production and survival of green (closed circles) and albino (open circles) C. virid'issima starved in light. Each point represents the mean number of hydranths in duplicate culture vessels. diately after feeding (Figs. 5, 6) suggesting that they are primed with cofactors necessar\- for the effieient con\'ersion of crustacean protein into coelenterate protein. In contrast, albinos showed a lag after feeding. This lag may represent the time during which a cof actor from food is mobilized. These data take on special interest when compared to the results of field studies on the nutrition of corals. Corals contain- ing zooxanthellae grow optimally in spite of a low exogenous food supply (10, 11, 12). Our results with C. viridissima suggest that 264 THE BIOLOGY OF HYDRA : 1961 symbiotic algae can account for this by promoting efficient utiliza- tion of available food. ACKNOWLEDGEMENTS It is a pleasure to acknowledge the counsel of Drs. C. Hand and R. I. Smith at the University of California, Berkeley. Studies on hydra were initiated under the guidance of Dr. H. M. Lenhoff, to whom I am indebted for help in all phases of this investigation. REFERENCES 1. Brien, p. 1961. The fresh-water hydra. Amcr. Sci. 48: 461-475. 2. Dixon, M. 1941. Multi-enzyme systems. Cambridge, 100 pp. 3. GoREAU, T. F. and N. I. Goreau. 1960. Distribution of labeled carbon in rccf- building corals with and without zooxanthellae. Science 131: 668-669. 4. Lenhoff, H. M. and J. Bovaird. 1961. A quantitative chemical approach to problems of nematocyst distribution and replacement in Hydra. Detyelop. Biol. 3: 227-240. 5. Lenhoff, H. M. and K. F. Zimmerman. 1959. Biochemical studies of symbiosis in Chlorohydra viridissima. Anat. Rec. 134: 559. 6. LooMis, W. F. 1954. Environmental factors controlling growth in hydra. /. Exp. Zool. 126: 223-234. 7. LooMis, W. F. and H. M. Lenhoff. 1956. Growth and sexual differentiation of hydra in mass culture. /. Exp. Zool. 132: 555-574. 8. Muscatine, L. 1961. Some aspects of the relationship between a sea anemone and its symbiotic algae. Ph.D. Thesis, University of California, Berkeley, 100 pp. 9. Muscatine, L. and C. Hand. 1958. Direct evidence for transfer of materials from symbiotic algae to the tissues of a coelcnterate. Proc. Nat. Acad. Sci. 44: 1259-1263. 10. Odum, H. T. and E. P. Odum. 1955. Trophic structure and productivity of a windward coral reef community on Eniwetok atoll. Ecol. Monogr. 25: 291-320. 11. Sargent, M. and T. S. Austin. 1949. Organic productivity of an atoll. Traits. Amer. Geophys. Union 30: 245-249. 12. Sargent, M. and T. S. Austin. 1954. Biologic economy of coral reefs. U. S. Geol. Survey Prof. Pap. 260E, pp. 299-300. 13. Whitney, D. D. 1907. Artificial removal of the green bodies of Hydra viridis. Biol. Bull. 13: 291-299. 14. Wood, R. L. 1959. Intercellular attachment in the epithelium of Hydra as revealed by electron microscopy. /. Biophysic. Biochem. Cytol. 6: 343-352. See also p. 55, this volume. LEONARD MUSCATINE 265 DISCUSSION WAINWRIGHT: Is the number of hydranths proportional to the total amount of protein? MUSCATINE: Yes, the relationship is linear up to about 75 hy- dranths. WAINWRIGHT: Is this so in the starved experiment? MUSCATINE: Preliminary experiments show that after 5 days of starvation albinos consist of less protein per hydranth than the greens. EAKIN: Are you carrying along the colorless algae as a parasite in your albino organism? MUSCATINE: Microscopic examination after treatment with gly- cerin indicates that algae are no longer present. We use only those albinos which do not regain an algal flora when placed back in the glycerin-free culture solution. EAKIN: I will be reporting on an organism we developed by cul- turing Chlorohijcha in the dark, one which we call "brown ChJoro- hydra," and which undoubtedly corresponds to your "albino." We have not been able to detect the presence of any colorless algae in them. BURNETT: One of my students, Peter Wernik, finds that albinos take in more glycogen and protein reserve droplets than do green hydra. MUSCATINE: Do you feel that the green hydra use food more efficiently than the albinos? BURNETT: I don't know. Possibly the greens aren't requiring as much; a hydra always takes in just about what he needs. Did I understand you to say that your animals budded during 8 days of starvation? MUSCATINE: Yes, budding by green hydra persists for a week. They double their number in this time. GOREAU: What is the ratio of plant to animal biomass in Chloro- hijdra? 266 THE BIOLOCY OF HYDRA : 1961 MUSCATINE: I have no information on the algae in Chlorohydra yet. But I have good data for Anthoplcura clcgantissima where one can determine the biomass of the alga flora by using quantative pigment techniques. GOREAU: You mean chlorophyll? MUSCATINE: Yes, and the various carotenoids. Using the method of Richards with Thompson, and using cell counts and dry weight data from pure suspensions of zooxanthellae, we find that in Antlwpk'ura, the ratio of animal to algae on a dry weight basis is about 332 to 1. GOREAU: Such data is very important in relation to turnover studies. CLAYBROOK: How critical is the magnesium requirement for Chlorohydra? Is this essential or does it merely increase the growth rate? MUSCATINE: The maximum doubling time of green or albino hydra is about 1.2 days. Without magnesium it is only 1.9 to 2.8 days. LENHOFF: I think it's important to add that they reqidrc mag- nesium in order to grow. When we first received our Chlorohydra, we could not grow them on any of our other culture solutions. The last cation that we tried was magnesium. Then they doubled nearly every day. CLAYBROOK: In our experiments with Chlorohydra, we don't add any magnesium to the solution. MUSCATINE: Well, there is a possibility that they get enough in their food, or perhaps you have a different strain of animals? CLAYBROOK: It could be. EAKIN: Although we have maintained our Chlorohydra in syn- thetic solutions to which we have added no magnesium (solutions which give optimal maintenance conditions for Hydra littoralis), we find that the addition of Mg+ + decreases the doubling time and on occasions has caused clones showing signs of depression to return to normal. LEONARD MUSCATINE 267 LOOMIS: We grow them happily in 5% artificial seawater. We make up an MBL artificial seawater with deionized water, not dis- tilled. That's the main thing, no copper. In fact, we have nearly a dozen hydroids growing in artificial seawater. Cordylophora grows nicely in 10% MBL water while Chlorohydra grows in 55^ MBL. Of course, that has magnesium in it. I would like to make another point. Some day, somebody ought to study how glycerine makes the endothelial cells spit out their contained Chlorella. It would he interesting to study this incredible reaction, as well as to try and reinfect albino green hydra with free Chlorella. I don't think this has ever been done with green hydra, or lichens either. In other words, you can separate; but no one has yet recombined the two symbiotic forms that I know of. SLAUTTERBACK: Regarding reinfection, I have taken eggs, which as you know are white, from Chlorohydra and hatched them separately. If the resulting albinos are returned to a culture of green hydra, they remain white for about 2 weeks. FULTON: I think some German workers have succeeded in re- infecting white Chlorohydra, but not other species of hydra. Inci- dentally, we found that the antibiotic chloramphenical cures green hydra of the algae in a couple of days, much faster than glycerine. MUSCATINE: What concentration was used? FULTON: I'm not certain, but I think it was 200 /.ig. per ml. AlUSCATINE: I tried various algicides with the anemones, and neither the commercial product Algaedyne, which is a colloidal sil- ver solution, nor a high concentration of Streptomycin, nor starva- tion in darkness succeeded in totally ridding the animal of its algae. When starved in darkness the animal becomes very small but still retains algal cells which can be shown to increase in pigment con- tent. This might be regarded as evidence for heterotrophic activity in these zooxanthellae. FULTON: Did you try chloroamphenicol? MUSCATINE: No. BURNETT: Have you e\'er noticed differences in the distribution of the algae along the column? 26S THE BIOLOGY OF HYDRA : 1961 MUSCATINE: Yes, very often one sees regional differences. How- ever, I am not sure of the significance of this. BURNETT: I mention this because I had a chance to observe a very interesting green hydra in Brien's laboratory. This animal ( H. viridis) underwent what seems to be a somatic mutation. The peduncle on this form resembles a stolon and is several times larger than the gastric region. The whole animal may be one and a half inches long, surprising dimensions for a green hydra. The peduncle, unlike that of normal H. viridis, contains more algal bodies per cell than the gastric region. Also food materials pass into this region in greater amounts than into a normal peduncle. What is most inter- esting is that this mutant form not only reproduces asexually by budding but also by pinching off the distal portion of its peduncle. This detached portion then regenerates into a complete organism. It is marvelous! EAKIN: Did you try increasing the oxygen tension while growing the albino hydras? MUSCATINE: Well, we're just getting into gas analysis. We have conducted preliminary experiments growing green and albino hydra in air plus 0.4% COo, but the results were not definitive. Eventually we will control pCOo and pOo. EAKIN: It will be interesting to see if the high oxygen tension can reverse some of the effects observed in the absence of the algal chlorophyll. MUSCATINE: Yes, that's a good way to attack this, going through the algae. I would also like to see if green hydra show an action spectrum for growth rates which can be related to the absorption spectrum of chlorophyll. On the Relation of Calcification to Primary Productivity in Reef Building Organisms T. F. GOREAU Physiology Department, University College of the West Indies, Mona, Jamaieu, W. I., and Department of Marine Biochemistry and Ecology, New York Zoological Society. Coral reefs are tropical shallow water communities built up by calcareous organisms attached to the sea bottom. Such ecosystems may be regarded as biochemical factories which catalyse a large scale transfer of dissolved calcium and carbonate ions from sea water into the sediments as insoluble calcium carbonate. The result- ing reef limestones are deposited in typical formations which may in time become several thousand feet thick, as for example in some of the Pacific atolls ( 9 ) . A unique characteristic of coral reefs, found in no other deposi- tional system in the biosphere, is that maximum biological accre- tion of calcareous matter takes place only in the turbulent surface waters where the forces of mechanical and chemical erosion are also at a maximum. Corals and algae which build reefs do so by secreting hard calcareous masses that become aggregated into an organised coherent structure adapted for maximum attenuation of mechanical stresses set up by the constant battering of the seas, yet so shaped as to expose a maximum surface area for efficient matter-energy exchange with the environment. The papers of Tracey ef al. (16) and Emery ct al. (1) should be consulted for further aspects of this problem. In the West Indies, the interlocking reef framework is built up by the larger Scleractinia and Milleporidae, their separate colon- 269 270 THE BIOLOGY OF HYDRA : 1961 ies becoming cemented into a single nnit by lithothamnioid algae. The finer, more \'oIuminous, lagoon and forward slope sediments are produced chiefly by calcareous green algae, with Scleractinia, Gorgonia, Foraminifera, sponges, mollusks, arthropods and ech- inoderms contributing in xarious proportions depending on local factors. Owing to its stability and exposure to the seas, the frame- work is probably the site where most of the calcium carbonate pro- duction of the reef occurs. Only a fraction of this is ultimately de- posited in situ since the greater part of the calcareous material pro- duced here is washed out by waves and redeposited in the calmer water of the lagoon or the seaward slope. A large proportion of the total biomass of coral reefs is due to algae which grow in great abundance in all zones, ranging from the shallowest parts of the rampart to depths exceeding two hundred feet on the forward slope. The algal population of reefs can be divid- ed into two categories: the free-li\'ing fleshy, filamentous, and cal- careous algae; and the symbiotic unicellular zooxanthellae living in coelenterates. All reef-building Scleractinian corals without exception contain zooxanthellae. So do most Hydrocorals, Actinaria, Zoanthidea, Alcyonaria and Gorgonia living in reefs. According to the existing nomenclature, those calcareous coelenterates which have zooxan- thellae are said to be hermatypic, or reef-building; whereas those species lacking zooxanthellae are said to be ahermatypic or non- reef building. The former are limited in their vertical distribution to the upper parts of the euphotic zone and never grow in dark places. The ahermatypes are usually found in deep water be- low the euphotic zone although some species occur in shallow wa- ter where they tend to favour dark crevices. The basic difference be- tween hermatypic and ahermatypic coelenterates is that the former grow much faster to much larger sizes than the latter. Never- theless, some ahermatypic corals are known under certain condi- tions to form deep-sea banks which bear a superficial resemblance to shallow^ water reefs ( 14) . Although there is an absolute correlation between the pres- ence of zooxanthellae in calcareous coelenterates and their ability to build reefs, the relationship of the algae to their hosts and to the bio-economy of the reef as a whole is not yet clearly under- T. F. GOREAU 271 stood. The so-called "zooxanthella problem has been the sub- ject of much controversy because some investigators ha\e failed to recognise the multiplicity of host-symbiont relationships in the different groups of coelenterates: ranging from total nutri- tional dependence on zooxanthellae in some xeniid Alcyonacea (2) to nutritional independence in the Scleractinia which are spe- cialised carni^'ores (21). There can be little doubt that zooxanthella- coelenterate symbioses have exolved independenth' in many unre- lated groups at different times, thus accounting for the haphazard variety of the association in the \ arious classes and orders of the phylum. For further details and references regarding the zooxan- thella problem, the papers of Yonge (19, 20), Vaughan and Wells (17), Odum and Odum (10), and Goreau (4) should he consulted. CALCIUM DEPOSITION AND PHOTOSYNTHESIS IN REEF CORALS Growth in corals is achie\ ed b\' an increase in mass of the cal- careous skeleton and a concommittant proliferation of the overly- ing tissues. Our recent underwater studies on reef corals suggest that e\'en within any given species there may be no constant re- lationship between these two kinds of growth and that colony shape is to a certain extent controlled by ^'ariations in the ratio of new skeleton to new tissue. To study the factors which regulate calci- fication in corals and other calcareous organisms, we ha\e dexeloped new methods for the fast quantitative assay of growth by the use of radioactive tracers. Calcification is determined from the rate with which Ca^"* ions added to the sea water medium is deposited into the skeleton as Ca^' CO.,. under various conditions, e.g. light and dark. The procedure, which has been described elsewhere (3, 6), requires only a few hours; the experimental runs can be carried out in the field, and growth gradients are determined by sampling different parts of experimental colonies. Our observations demonstrate that calcification in reef-build- ing corals is dependent on the ambient light intensity to the ex- tent that growth in foiuteen species tested is on the a\'erage ten times faster in sunlight than in darkness (6). Calcification is re- 272 THE BIOLOGY OF HYDRA : 1961 duced by approximately fifty per cent on a cloudy day under other- wise similar conditions. By contrast, the calcification rates of some shallow water ahermatypic corals lacking zooxanthellae do not re- spond significantly to changes in light intensity. The stimulant effect of light on reef coral calcification disappears when the zoo- xanthellae are removed by culturing corals in darkness for about three months. Inherent species specific factors, independent of the zooxanthel- lae, also exert an important influence on calcium deposition. One example of this is the growth gradient of ramose corals such as Acropora cerviconiis where the large pale apical polyps that con- tain relatively few zooxanthellae calcify several times faster than the much smaller adjacent lateral corallities the tissues of which are packed with large masses of zooxanthellae. The enzyme car- bonic anhydrase also appears to play an important role in coral calcification. We have found carbonic anhydrase activity in repre- sentative species of all major groups of Coelenterata. The occur- rence of the enzyme has no relationship to the calcareous habit, or to the presence of zooxanthellae, which themselves do not contain significant amounts of carbonic anhydrase. The treatment of reef corals with a specific carbonic anhydrase inhibitor ( Diamox, Lederle) results in an average fifty per cent reduction of the calci- fication rate in the light, and a seventy five per cent reduction in darkness. The effect of carbonic anhydrase inhibition on the calci- fication rate is partially reversed in the light when the zooxanthel- lae are photosynthesizing. It therefore appears that carbonic anhy- drase and the zooxanthellae act in synergy to potentiate calcium deposition in corals ( 3 ) . The mechanisms responsible for the stimulation of skeleto- genesis in corals by photosynthesis of zooxanthellae are not clear- ly understood. If the two reactions are linked through some common pathway, the coupling must be of a facultative type since cal- cification can proceed in the absence of photosynthesis, although at a much reduced rate. We have observed that calcification is speeded up very quickly following the exposure of the corals to adequate light intensities. The short time constant of the potentiation makes it unlikely that the stimulation is due to produc- tion of nutrients by the zooxanthellae, but rather to prompt changes T. F. GOREAU 273 in concentration of some substrate common to photosynthesis and calcification. In previous papers (3, 4) we advanced the working hypothesis that acceleration of CaCOg deposition would occur if algal photosynthesis were to remo\e COo from the system and cause the equilibrium reaction T Ca(HCO,)., ^ CaCO, + H,CO, i to go to the right. Although the evidence for this is fairly pursua- sive, other mechanisms may also be involved. Some of these will be discussed below. In principle, the rate of CaCO.; production could be stimulat- ed in at least two ways : directly through control of the steady state bicarbonate concentration in the tissues as shown above, or indi- rectly by augmenting the supply of free energy available for active calcium transport through an increase in the rate and efficiency of cellular metabolism. In the discussion below, we will consider some of the possible indirect mechanisms. The onset of photosynthesis by the zooxanthellae immediately produces a rise in the intracel- lular oxygen concentration which may result in some increase in the rate and efficiency of metabolism in the coral. Thiel (15) and Yonge (19), among others, have already emphasized the probable importance of in situ production to the coral, but no spe- cific mechanisms were proposed. There is at present no information on the relation between the pO^. of the medium and the rate of coral growth. Nearly all hermatypic corals are net oxygen producers during the day, and the water circulating in the growing parts of the reef is as a rule supersaturated with oxygen (8, 10, 11, 12, 13) so that the dependence of calcification on oxygen would be diffi- cult to measure in these organisms. In two ahermatypic corals lacking zooxanthellae {Tuhastrca and Asirangia) we observed no significant changes in calcium deposition rates under conditions where the oxygen saturation of the medium varied between fifty and one hundred and twenty two per cent, suggesting that calcifica- tion rates in these corals are relatively independent of oxygen con- centration within the limits tested. 274 THE BIOLOGY OF HYDRA : 1961 Given an adequate supply of oxygen in the medium, far reach- ing effects on the rate and efficiency of metaboHc reactions can be brought about by increasing the rate with which soluble waste products are removed from the coral cells (20). This is a far more potent metabolic stimulant than increasing the oxygen concentra- tion. It has long been known that velocities of metabolic reac- tions are strictly limited by the rates with which the end products are removed from the immediate environment. In higher animals, this is accomplished by specialised circulatory and excre- tory systems which are lacking in the coelenterates. In the ab- sence of zooxanthellae, or in darkness, corals are forced to rely on diffusion alone to get rid of the soluble inorganic waste products of cell metabolism. This is a slow process, especially when the sur- face area for exchange is reduced by retraction of the polyps into the calyces. This situation is radically altered in the presence of zooxanthellae which require for photosynthesis and j)riniary pro- duction those very substances that the coral host must get rid of, e.g. COo, phosphates, nitrates, sulphates, ammonia, etc. Yonge and Nicholls (21) showed for some corals that zooxanthellae are capable of sufficiently high rates of photosynthesis to utilise not only all the soluble inorganic phosphate produced by coral colonies, but that additional phosphate is absorbed from the surrounding sea water. Under conditions of adequate illumination, the zooxanthellae are to be regarded as combined intracellular lungs and kidneys. The observed speeding up of calcification in reef corals exposed to bright light may in part be due to an increase of the rate and effici- ency with which metabolism can supply free energy to the car- rier mechanism concerned with active calcium transport. The ques- tion whether the calcification rate is indeed related to the metabolic rate, and whether this is in turn influenced by the level of algal photosynthesis in the manner indicated above is now under investigation in our laboratory. CARBONATE DEPOSITION, GROWTH AND PRODUCTIVITY Elsewhere, we advocated the view that Ca~^ and HCO^g ions dissolved in the ambient medium are the source of the mineral de- T. F. GOREAU 275 posited ill the skeleton as CaCO^, and that these are brought to the calcification site by separate pathways ( 3 ) . In order to test this directly we developed a technique in which the uptakes of Ca"*' and C^^ carbonate were measured simultaneously in a variety of calcareous coelenterates and algae, under natural conditions in the reef. As before, light and dark runs were carried out simultane- ously, the experiments lasting lietween five and six hours. After washing and drying the specimens, activities due to Ca^' and C^^ deposited in the skeleton were quantitatively isolated, and sep- arated from the C^^ activity fixed in the coenosarc as organic mat- ter by photosynthesis of the zooxanthellae. A detailed description of this technique will be published later. The data in Table I summarises results of field experiments carried out for the purpose of measuring simultaneously calcium and carbonate transfer rates from the medium into the test organ- isms. The plants and animals used in these investigations, and list- ed in Table I, belong to three different ecological categories: Group 1 consists of shallow water ahermatypic coelenterates which contri- bute only insignificant amounts of calcareous matter to the reef; Group 2 contains three hermatypic coelenterates which are chiefly reef framework builders; Group 3 has three hermatypic algae, the remains of which form the bulk of the fine calcareous lagoon and slope sediments. All these species are found in the actively grow- ing part of the reef rampart at Maiden Cay, Jamaica, where these experiments were carried out. The first two columns of Table I give the transfer rates of Ca"*^"^"*" and HC^^O^g into the mineral skeleton, the third column gives the rate of photosynthetic fixation of C^'' into organic matter, e.g. the primary producti\'ity. In the ahermatypic coelenterates lacking zooxanthellae, there are no significant light-dark differences in the calcium deposition rates, but in the hermatypic coelenterates con- taining zooxanthellae and in the hermatypic algae, these differ- ences are extremely pronounced. An exception was the red alga Ampliiroa where the calcification rate in darkness was much higher than in light. Not unexpectedly, the organic carbon fixation val- ues observed in ahermatypic species were extremely low, and were probably due to heterotrophic exchange, or photosynthesis of bor- ing algae in the skeleton. 276 THE BIOLOGY OF HYDRA : 1961 The primary carbon fixation observed in hermatypic coelenter- ates was due to photosynthesis l3y zooxanthellae. The boring algae were present in only very small amoimts in our samples and it is assumed that their contribution to the total productivity was also very small. Owing to uncertainty of the proportion of the plant biomass in corals, the data are given in terms of total nitrogen, e. g. animal plus plant. The highest calcification and productivity rates were observed in the hermatypic algae. The two species of HaUineda behaved like the hermatypic corals in that calcification was much faster in light than in darkness, but in Amphiroa there was a nega- tive correlation between photosynthesis and skeletogensis. We be- lieve that light inhibition of calcification in this species is pro- duced by a shortage of available carbonate due to competition for CO2 as a common substrate by extremely high levels of photosyn- thesis. This problem is now being investigated in our laboratory. There is a positive correlation between the calcium deposition rate and the photosynthetic rate as measured by the specific pri- mary productivity, e. g. the amount of organic matter produced in TABLE 1 Specific calcification and productivity rates of hermatypic and ahermatypic organisms. Light Category Species or ^g.Ca/mg.N/hr jixg.carbonate- ^g.organic- Dark C/mg.N/hr. C/mg.N/hr. Ahennatypic S. roseus light 12.0 3.30 1.250 Coelenterata dark 13.2 2.46 0.489 without A. solitaria light 8.7 1.33 0.547 Zooxanthellae dark 8.6 0.77 0.400 T. tcnuilumellosa light 5.5 0.56 0.217 dark 5.6 0.85 0.161 Hermatypic A. cervicornis light 126.3 17.93 12.090 Coelenterata (apical cm.) dark 35.1 4.09 0.861 with M. complanata light 59.6 10.19 19.680 Zooxanthellae dark 25.0 6.44 1.640 P. fiircata light 26.7 8.14 13.800 dark 5.6 0.63 0.532 Hermatypic H. tuna light 178.0 23.21 26.390 Algae dark 77.9 9.36 0.905 H. opuntia light 256.1 38.46 50.520 dark 72.6 11.82 0.899 A. fragilissima light 68.3 43.33 56.320 dark 792.6 87.24 2.180 T. F. GOREAV 277 fig carbon fixed per milligram nitrogen per hour. The highest calci- fication and productivity values are observed in the calcareous algae. In the light, the calcification rates in the two Halimedas are about 1.5 to 10 times faster than in the hermatypic corals, and about 20 to 40 times faster than in the ahermatypes. The carbon fixation rates in the Halimedas are only from 2.5 to 4 times greater than those in the hermatypic corals, the productivity values for the ahermatypes being neglected as they have no zooxanthellae. The approximate diurnal calcification and carbon fixation rates of the various species tested are shown in Table II. The daily cal- cium deposition was calculated on the simplifying assumption of twelve hours darkness and twelve hours sunshine equal in inten- sity to the average isolation between 10 a.m. and 4 p.m. during a late winter day in Jamaica. The daily productivity values were cal- culated on a twelve hourly liasis since no photosynthesis occurs at night. These figures are uncorrected for respiration. Tables I and II show that the differences in the calcification rates be- tween groups are far greater than the corresponding differences in the carbon fixation rates, but more data are needed to establish whether a quantitative correlation exists here. Obviously such com- parisons can have meaning only on a broad ecological level since we do not yet know if the physiological mechanisms of calcification in the various groups of organisms used for these experiments are equivalent. Nevertheless, the overall correlation is prob- ably not due to chance; it emphasizes the fundamental role TABLE 2 Daily calcification and carbon fixation rates of hermatypic and ahermatypic organisms. Category Species Calcium deposition in j(i,g./mg.N/dav Carbon fixation in ;Ug/mg.N/day Ahermatypic Coelenterata without zooxanthellae S. roseus A. solitarid T. tenuilamellosa 292.4 207.6 133.2 Hermatypic Coelenterata with zooxanthellae A. cervicornis (apical cm.) M. complanata P. furcata 1936.8 1015.2 387.6 145.08 236.16 165.60 Hermatypic algae H. tuna H. opuntia A. fragilissinia 3070.8 3944.4 10330.8 316.70 606.24 675.84 278 THE BIOLOGY OF HYDRA : 1961 played by photosynthesis in facihtating the deposition of calcare- ous matter in a wide variety of hermatypic organisms, irrespective of the possibility that the mechanisms concerned may be very different. Comparison of the results summarised in the first two columns of Table I shows that skeletogenesis rates calculated from Ca^' uptake are much higher than those calculated from the simul- taneous C^^ carbonate uptake. In CaCO... the stoichiometric mass ratio of calcium to carbon is 40 12 or about 3.335. This ratio should apply to the mineral constituent of the coelenterate and algal skel- etons which is mostly CaCO:., though some of the algae may con- tain traces of dolomite in addition to calcite and aragonite (18). However, the ratios calculated from our data are nearly all higher than the theoretical value, and they vary over a wide range. This either indicates that the organisms are secreting a skeletal mineral greatly enriched in calcium, or that the specific activities of the C^^ and Ca^*" labelled percursors change with respect to the external medium, and to each other, during the process of deposition. As there is no experimental evidence for calcium enrichment we are inclined to explain the apparent carbonate deficit shown in our data on the basis of the second alternative. The transfer rates given in Table I were calculated on the as- sumption that during CaCO;, deposition the specific activities of the Ca*'' and C^^ labelled percursors do not change with respect to the sea water or to each other, a condition that would occur only if the system were in isotopic equilibrium. However, this was not the case in our experiments which were run over sufficiently short periods of time that it was impossible for the test colonies to achieve isotopic equilibrium. Therefore it is to be expected that the specific activities in the newly formed skeletal CaCO;. would be less than in the dissolved Ca+^ and HCO^ of the medium if the labelled ex- ogenous atoms were to exchange with intracellular stores of un- labelled atoms to final deposition into the skeleton. Given that the molar fluxes of calcium and carl:>onate are equal and linked by some common pathway, and using the specific activi- ties of the precursors dissolved in the sea water as a reference base, the calculated deposition rates will be the higher for that component which has suffered the least isotopic dilution, e. g. cal- T. F. GOREAU 279 cium, and the lower for that constituent which was diluted the most during its passage through the cells, e. g. carbonate. This sug- gests that the reservoir of intracellular carbonate available for ex- change with absorbed exogenous carbonate is much greater than the internal pool of freely exchangeable calcium, and that the tissue calcium turnover rates must therefore be much higher than those of carbonate. In previous experiments, we have demonstrated that the exchangeable calcium in corals is indeed maintained at a low level in corals (5, 7 ) . The simultaneous introduction of isotopically labelled calcium and carbon makes it possible to assess the relative sizes of the pools of exchangeable endogenous calcium and carbon by the principle of dilution volumes in a situation where no isotopic equilibration has occurred. Under these conditions, our calculated transfer rates indicate that the internal pool of carbon available for exchange with exogenous carbonate being deposited into the skel- eton is about two to fifteen times greater than the amount of ex- changeable calcium. SUMMARY AND CONCLUSIONS 1. Coral reefs are tropical shallow water communities where intensive biological calcification occurs, resulting in net accumula- tion of limestone into the sediments. Photosynthesis appears to be in some way essential to reef formation. The most important reef-building organisms are calcareous algae and coelenterates, cor- als included. All reef-building coelenterates without exception con- tain symbiotic zooxanthellae. Corals without zooxanthellae grow slowly and never play a significant role in the building of reefs. 2. The zooxanthellae do not themselves calcify, but their presence results in a very powerful enhancement of calcification in the coral host as soon as photosynthesis begins. We have shown that stimulation of growth by light requires zooxanthellae since this efi^ect does not occur in reef corals from which zooxanthellae are removed, nor does it occur in ahermatypic corals which never have algal symbionts. Of three calcareous algae tested, two calcified much faster in light than in darkness, and in one the efl^ect was re- versed. 280 THE BIOLOGY OF HYDRA : 1961 3. There is a rough correlation between calcification rate and specific photosynthetic rate as measured by the organic productivity. The highest calcification and productivity rates were noted in the calcareous algae, but in one of these we observed a very strong reduction of CaCO.j deposition in the light in the presence of a very high rate of photosynthesis. Calcification and primary pro- ductivity rates in three hermatypic coelenterates with zooxanthel- lae are on the average about sixty per cent lower than in the calcareous algae. Their slowest calcification rates were observed in the ahermatypic corals that have no zooxanthellae. 4. Under the conditions of our experiments, it was found that labelled calcium was deposited up to seventeen times faster than labelled carbonate. This discrepancy may be the result of very large diflFerences in the amount of exchangeable endogenous car- bon in relation to the amount of calcium available for exchange, the former being very much larger than the latter so that intracel- lular dilution of the absorbed C^^ was much greater than that of Ca"* '. 5. Several mechanisms linking photosynthesis and calcifica- tion are discussed. CaCOy production may be enhanced: (1) through removal of CO2 from the calcification site by photosyn- thesis and/or carbonic anhydrase; (2) from stimulation of coral metabolism by photosynthesis of the zooxanthellae, which in turn increases the amount of energy available for active calcium and carbonate transport through the tissues into the skeleton. There is no evidence that metabolic efficiency in reef corals is increased by augmenting the oxygen supply over and above that already available from the environment. The zooxanthellae probably exert their effect by speeding up the rate with which metabolic waste products are removed from the vicinity of the host's cells since the algae require as raw material for photosynthesis those very inor- ganic substances that the coral must get rid of. Rapid removal of these from the host cells must set up strong local concentration gradients resulting in a large increase of metabolic efficiency, thus making more free energy available for a CaCO.^ secretion. 6. Photosynthesis plays a double role vis a vis the reef: it in- creases the free energy of the community through primary produc- tion and it produces in corals and algae the optimum physiological conditions necessary for rapid and efficient secretion of calcium car- T. F. GOREAU 281 bonate. In corals, the coupling of the calcification reaction to photosynthesis, though facultatixe, is almost certainly due to a direct link \'ia a common metabolic pathway, rather than to synthesis and diffusion of nutrients from the zooxanthellae to the host. There can be no question that the great increase in rate and efficiency of limestone secretion associated with photosynthesis must, on a community level, be of decisive importance to the for- mation, growth and maintenance of tropical coral reef ecosystems. REFERENCES 1. Emery, K. O., J. 1. Tracey and H. S. Ladd. 1954. Geology of Bikini and nearby atolls. U.S. Gcol. Survey Prof. Pap. 260-A 265pp. 2. GoHAR, H. A. F. 1940. Stndies on the .Xeniidae of the Red Sea. Mar. Biol. Sta. Ghardaqa, Egypt, Pub. 2: 25-118. .3. GoREAU, T. F. 1959. The physiology of skeleton formation in corals. 1. A method for measnring the rate of caleiimi deposition by corals under different conditions. Biol. Bull. 116: 59-75. 4. GoREAU, T. F. 1961. Problems of growth and calcium deposition in reef corals. Endeavour 20: 32-39. 5. GoREAU, T. F. and \'. T. Bowen. 1955. Calcium uptake by a coral. Science 122: 1188-1189. 6. Goreau, T. F. and N. I. Goreau. 1959. The physiology of skeleton formation in corals. 11. Calcium deposition by hermatypic corals under various con- ditions in the reef. Biol. Bull. 117: 239-250. 7. Gore.\u, T. F. and N. I. Goreau. 1960. The physiology of skeleton foniiation in corals. IV. On isotopic equilibrium exchange of calcium between coral- lum and environment in living and dead reef-building corals. Biol. Bull. 119: 416-427. 8. KoHN, A. J. and P. Helfrich. 1957. Primary organic productivity of a Hawa- iian coral reef. Limn, and Oceanogr. 2: 241-251. 9. Ladd, H. S., E. Ingebson, R. C. Townsend, R. C. Russell and H. K. Stephen- son. 1953. Drilling on Euiwetok Atoll, Marshall Islands. Am. Assoc. Petr. Geol. Bull. 37: 2257. 10. Odum, H. T. and E. P. Odum. 1955. Trophic structure and productivity of a windward coral reef community on Eniwetok Atoll. Ecol. Monogr. 25: 291-320. 11. Odum, H. T., P. R. Burkholder, and J. A. Rivero. 1959. Measurements of producti\ity of turtle grass flats, reefs, and the Bahia Fosforescente of Southern Puerto Rico. Inst. Mar. Sci. (Texas). Publ. 6: 159. 12. Sargent, M. C. and T. S. Austin. 1949. Organic productivity of an atoll. Amer. Geophys. Union Trans. 30: 245-249. 13. Sargent, M. C. and T. S. Austin. 1954. Biologic economy of coral reefs. U. S. Gcol. Surv. Prof. Pap. 260-E: 293-300. 282 THE BIOLOGY OF HYDRA : 1961 14. Teichert, C. 1958. Cold and deep water coral banks. Am. Assoc. Petr. Gcol. Bull. 42: 1064. 15. Thiel, M. E. 1929. Zur Fra,^e der Ernahning der Steinkorallen und der Bredeu- tung Ihrer Zooxanthellen. Zoo/. Anz. 81: 295. 16. Tracey, J. I., H. S. Lauu and J. E. Hokfmeister. 1948. Reefs of Bikini, Mar- shall Islands. Geol. Soc Am. Bull. 59: 861-878. 17. Vaughan, T. W. and J. W. Wells. 1943. Revision of the suborders, families and genera of the Scleractinia. Geol. Soc. Amer. Spec. Pap. 44: 363 pp. 18. Vinogradov, A. P. 1953. The elementary chemical composition of marine organisms. Sears Found. Mar. Res. Mem. II: 647 pp. 19. Yonge, C. M. 1940. The biology of reef building corals. Gt. Barrier Reef Expecl. Set. Rep. 7(13): 353-391.' 20. Yonge, C. M. 1957. Symbiosis. Geol. Soc. Amer. Mem. 67 (l): 429-442. 21. Yonge, C. M. and A. G. Nicholls. 1931. Studies on the Physiology of corals. V. The effect of starvation in light and darkness on the relationship between corals and zooxanthellae. Gt. Barrier Reef Expecl. Sci. Rep. 1(7): 177-211. DISCUSSION WAINWRIGHT: First I'd like to wave a small flag because you who have trays of hydra in your laboratory and even you ocean- ographers with laboratories in a ship don't have any idea under what difficulties Dr. Goreau is working and what he has done in taking his laboratory down onto the reef. Think of diving to 100 feet with 200 pounds of machinery on your back and then doing a critical experiment using glassware, radioisotopes, and lixing animals. Now I want to ask a question. Do you know what the limiting factors in calcification are? GOREAU: No, not yet, if we exclude light for the moment. Con- trary to what I said earlier, it may be possible to culture some species of corals in vitro. We must never assume, however, that the growth or accretion rates we measure under those conditions are equal to those occurring on the reef. Nevertheless, laboratory studies are use- ful because we can rigidly control the environment, the concentra- tion of such substances as HCO^ and Ca++ and the additions of inhibitors or stimulants, etc. We are planning such studies, but haven't gotten around to them yet, so I cannot really answer your question. MUSCATINE: Do you feel that calcification in corals is augment- ed by removal of COo by zooxanthellae? T. F. GOREAU 283 GOREAU: Yes. If we assume the hypothetical scheme of calcifica- tion which I published some years ago, then the removal of CO- from the system would tend to drive the equilibrium to the right and increase the rate of CaCO.; formation. MUSCATINE: This differs from the scheme of Wilbur and Jod- rey who found that calcification in their oyster mantle preparations was increased about five fold if a source of COo such as oxaloace- tate was added to the external medium. GOREAU: Oxaloacetate is an intermediate in the Krebs cycle. Any increase in the rate of this cycle may have rather non-specific effects, and changes in calcification rates would tell us little. Nevertheless, it's a very interesting possibility and we are planning work along similar lines. Unfortunately, as Wainwright mentioned, there are certain small difficulties in running such experiments. MARTIN: In mammals, the accretion of bone substance is not a one-way affair, but as accretion goes on, elimination and dissolu- tion of bone material also goes on. I wonder if these views contri- bute any insight into the problem. GOREAU: Yes. This is a very important point. Bone and coral differ in at least one fundamental way. Bone is mesodermal and remains at all times part of the internal medium of the body. At least 2(y/c of the bone mineral is exchangeable with calcium and phosphate dissolved in the body fluids. In addition, mammalian bone is vascularized and full of cells. The corallum, on the other hand, is an ectodermal mineral deposit which lies outside the body of the coral polyp. We have evidence that once the CaCO;^ is deposited there, it undergoes little or no further exchange with the environ- ment or with the coral; that is, it seems to be essentially isolated as long as it is covered by a layer of living tissue. LOO MIS: Dr. Goreau has shown that the rate of calcification at the end of a coral branch is something like tenfold what it is at a shoulder. I find this position effect fascinating since the two en- vironments appear identical at first glance. Another point is that CO^. plays a double role: (a) it is part of the calcium carbonate which is part of the corallum, and (b) it 284 THE BIOLOGY OF HYDRA : 1961 exerts a pH effect. Now wlien the light is shining on the algae, free CO2 is rapidly photosynthesized and the pH goes up to maybe 11 or 12. GOREAU: Corals have alkaline phosphatases with optima at about pH 11.0 (Goreau, 1953, P.N.A.S. 39: 1291). We thought at first that these enzymes were concerned with calcifica- tion, but results of our histochemical studies (Goreau, 1956. Na- ture 177: 1029) make this appear unlikely. LOOMIS: Under illuminated conditions you get precipitation of calcium carbonate through increase of pH. Therefore, COo has two roles in calcification: one as the carbonate ion, and one as free CO2. GOREAU: I am not sure that I agree with you. I wish we could measure CO2 and pH in living calcifying corals. Let me comment on the first part of your question regarding differential growth at tips and sides of branches in Acropora cewicornis. Actually conditions are almost certainly not identical at the tips and sides of branches. This species has an inborn factor which controls the rate and pat- tern of calcification in the colony — and thus determines colony shape. It is a function inherent in the coral not the zooxanthellae, and within some limits seems to have little relationship to photo- synthetic carbon fixation as I mentioned in my talk. PHILLIPS: How long a period of photosynthesis do you allow in these experiments? GOREAU: Approximately 6 hours, PHILLIPS: Bean and Hassid (Assimilation of C^^Oo by a Photo- synthesizing Red Alga, Iridophycus flaccidum. Bean, R. C. and W. Z. Hassid. 1955. /. Biol Chem. 2i2;411-425) found in their studies an assimilation of C^Oo in Iridophycus flaccidum, a ma- rine red algae, that 90 odd percent of the C^^ was in an alcohol- soluble phase. Alcohol extraction might be a possible way of getting around your wet ashing. What is the method you use? GOREAU : It is a modification of a technic published by Folch and Van Slyke. Instead of using a mixture of concentrated sul- T. F. GOREAU 285 phuric and phosphoric acids as the primary ashing agent, we use mixtures of 70% perchloric and concentrated nitric acids with a bit of potassium iodate added. We cannot use sulphate in any form because we wish to avoid converting the calcium to the sulfate and phosphate salts. PHILLIPS: The 80% ethanol might be worth trying since it would avoid the use of this rather explosive reaction mixture. GOREAU: We have had no trouble with it because we are using only 300 mg. samples in which there is less than 20 mgs. of organic matter present. HAND: Would you comment on the number of algae in the growing tip as compared with the number farther away. GOREAU: Histological sections show fewer zooxanthellae in the growing tip of A. cervicornis. The mg. N/mg. chlorophyll a ratio is also much higher in the axial polyps than in the lateral polyps- indicating a lower specific photosynthetic rate in the growing tip. HAND: This suggests that where there is less algae, there is more calcification. GOREAU: Yes, at least in A. palmofa and A. cervicornis. The Development of Cordylophora^ Chandler Fulton Department of Biolop.ij, Brandeis University, Waltham 54, Massachusetts One of the challenging problems of development is the manner in which a multi-cellular organism acquires and regulates its shape, pattern, or proportion. Colonial hydroids offer especially favorable material for study of this problem because their colonies are com- posed of a repeating pattern of hydranths arranged on tubular stems and stolons ( Fig. 1 ) . Hydroid colonies grow asexually by the elaboration of stolons attached to a substratum; at regular intervals the stolons send up uprights which bear hydranths, grow, and branch. The primary concern of this paper is the manner in which colonies develop this regular, repeating pattern. I chose to work with the brackish-water hydroid, Cordijlophoia lacustris, because it is exceptionally hardy and has a simple colony pattern. For study of the development of colonies, it is advantageous to have a refined and reproducible method of laboratory cultiva- tion similar to that de\eloped by Loomis for Hydra Uttoralis. One can grow Cordijlophoia colonies on glass microscope slides slanted in beakers of culture solution, with no flow of water or other spe- cial treatment ( 1 ) . The defined culture solution contains ionic so- dium, potassium, calcium, magnesium, chloride, and bicarbonate. All of these ions, with the exception of bicarbonate, are essential for growth at a maximum rate, and the proportion of the ions is critical. The cultures are fed Artemia larvae once daily, and the cul- ture solution changed after feeding and again later in the day. Be- tween feedings, the beakers are kept in the dark at 22 , though lA much abridged form of the paper presented at the meeting. Relevant Hterature citations and supporting data will be presented in papers to be published elsewhere, and may be found in reference ( 2 ) . 287 288 THE BIOLOGY OF HYDRA : 1961 neither light nor shght variations in temperature are critical. These standard conditions ( 1 ) have been used for all the experiments dis- cussed here, since variation of the conditions leads to alterations in colony pattern. The number of hydranths in a Cordylophora colony increases exponentially with time in the beaker-slide cultures, as do the hy- dranths of Ilydia in Loomis cidtures. It is thus possible to compute the growth rate of this colonial organism, using standard equa- tions for exponential growth. This growth rate has been used to eval- uate the growth conditions described above. Cordylophora colonies double about every three days, or more slowly than Hydra littoralis, which doubles in less than two days. The fact that Cordylophora colonies grow exponentially even though they are colonial is ol interest and we shall return to it later. stolon lip Figure 1. Diagram illustrating the basic pattern and macroscopic features of a Cordylophora colony. Sketched from a photograph of a laboratory colony. This culture method provided uniform Cordylophora colonies with which I could begin to study colony fomiation. Time-lapse movies taken to study the growth of colonies revealed a markedly organized system of peristaltic waves, which probably act to circu- late nutrients through the colonies.- These waves are proximally oriented, beginning at the tip of each hydranth and passing down -A movie demonstrating the features of peristalsis in Cordylophora was shown at tlie meeting. The apparent synchrony of peristalsis is still being studied. CHANDLER FULTON 289 through the tissue of the colony to the tips of the stolons. The waves are rhythmic, though very slow, occurring about two or three times an hour in a resting colony. The rate of peristalsis jumps threefold on feeding, to a frequency of about eight times an hour, and then declines back to the resting rate. The most striking feature of this peristalsis is that it is sychro- nized throughout a colony, in that the waves begin at the tip of each hydranth simultaneously. Further, if one ties a ligature on any of the uprights in a colony, the hydranth at the apex of that upright will, in time, begin to beat out of synchrony with the rest of the colony. In other words, disrupting the integrity of the colony (both tissue and coelenteron fluid ) eliminates the synchrony. Even if one accepts the conclusion that Cordylophora has nerve cells (Mackie, this symposium), I find it difficult to envision how a stimulus is trans- ferred through a colony in such a manner that each hydranth begins a perstaltic wave at the same time. I would suggest, however, that the synchrony indicates an order of integration in these colonial organisms which we have not hitherto suspected. I suspect also that understanding of colony development will involve further consider- ation of the orientation, rhythmicity and synchronization of the peristalisis. On superficial examination, a Cordylophora colony looks like a forest of little trees. I have attempted to distinguish the component events which produce this forest, and in so doing have found it pos- sible to describe in simple, quantitative terms how the forest de- velops. Careful observation of colonies reveals that they are entirely composed of a series of interconnected pipes, each consisting of a cylinder of tissue surrounded by a tubular perisarc." These tubes are of essentially uniform diameter. Thus one can conceive of a Cordylophora colony as a plumbing system with 0.2 mm. pipelines; the description of a colony can be reduced to a description of the kinds of tubes which comprise it, the relative positions of these tubes with respect to one another, and the way in which they are formed and grow. Stolon tubes, as they grow along the substratum, can give rise ^This approach to tlie colonies excludes the hydranths from consideration. Interesting observations on factors influencing the shape of hydranths, as well as entire colonies, have been presented by Kinne ( 3 ) . 290 THE BIOLOGY OF HYDRA : 1961 to two types of tubes: secondary stolons and uprights. Secondary stolons leave their parent stolons at right angles along the sub- stratum, while uprights leave at right angles away from the sub- stratum. Uprights, in contrast to stolons, are hydranth-bearing tubes, and give rise to one additional hydranth-bearing tube, the side branch. Side branches leave upright tubes at about 45 degree angles away from the substratum. Thus one can classify three types of tubes: stolon, upright, and side branch. Other differences further distinguish these tubes. Hydranth- bearing tubes develop only directly behind growing tips; they never develop in any other part of the colony. They are spaced at regular intervals along their tube of origin; upright tubes in particular occur at about three mm. distances along the stolon. In contrast, stolon tubes never develop at growing tips, but always come out of some old part of the colony, as at the base of a well-developed up- right. Further, stolon tubes are not spaced regularly; rather secon- dary stolons develop erratically with respect to any other part of the colony. How do these tubes grow? Since they are of uniform diameter, one can detemiine the growth rate of individual tubes by measuring increase in length with time. This has been done by photographing a colony over the course of a few days or a week in a growth cham- ber in front of a time-lapse camera. The movie is then used to plot the extension of the tube as a function of time. Such plots, for both stolons and uprights (side branch growth has not been measured), demonstrate that these tubes increase in length linearly with time. Stolons grow at a rate of about 0.1 mm. per hour, and uprights at a rate of 0.05 mm. per hour. You will recall that a colony as a whole grows exponentially in terms of hydranth number. The colony also grows exponentially in temis of dry weight, so that hydranth number is a measure of the mass of a colony. The observation of linear growth of tubes poses a dilemma : if the tubes which comprise a colony grow at a constant rate how does the colony as a whole grow exponentially? This ques- tion was first examined by model-building. One can diagram a col- ony in the form of a geometric progression, such that linear growth of tubes with regular branching at constant inten'als gives rise to exponential growth of the whole. Such a model does not look CHANDLER FULTON 291 like a Cordyloplwra colony in that 1 ) there is more branching than in an actual colony, and 2) the uprights are too tall relative to their parent stolons. The geometric progression model was redrawn in terms of the appearance of colonies growing under standard conditions, as shown in Figure 2. The growth during any unit of time is indicated by a pattern: black, stippled, etc. The stolon is visualized as grow- r ^ ?? i\\\\\\\t Time units Figure 2. A model illustrating the growth of a hypothetical colony over a period of six time units. See text for explanation. ing one unit per unit time (i.e., linearly), and producing uprights at a rate of one per unit time. During the same time unit, an upright grows only one-half unit, and a side branch only one-quarter unit. However, uprights and side branches continue to produce new tubes at the same distances as uprights are produced by stolons (i.e., one unit), and thus produce new' tubes at rates of 0.5 and 0.25 tubes per unit time respectively. Such a model takes into account the linear growth of tubes and normal branching pattern, and gives rise to a two-dimensional colony which bears a striking resemblance to laboratory colonies (cf. Figs. 1 and 2), If one computes the increase in hydranth num- ber of such a hypothetical colony with time, however, one finds that it continually falls away from exponential. This is in contrast to act- 292 THE BIOLOGY OF HYDRA : 1961 iial colonies, which do approach exponential increase in hydranth number. One can escape this new dilemma by doing what the colon- ies do, namely by introducing secondary stolons at intervals. If one adds such secondary stolons at appropriate times, one can make the growth of the model colony closely approach exponential. I do not know as yet whether or not this is the way colonies maintain ex- ponential growth. Laboratory colonies appear to develop in accord with the mod- el. This has been determined by measuring every relevent variable of the pattern of individual colonies, a task much facilitated by the use of a marking technique. If colonies are dipped into trypan blue, the perisarc is stained a deep blue while the tissue is unstained and unaflFected. When such a colony is grown in the absence of trypan blue, all new growth is colorless while that part of the colony present as perisarc at the time of marking remains blue. Thus new growth can be precisely measured as separate from old. The meas- urements support the picture of colony formation just described, except that branch tubes appear to grow more slowly, or at about one-eighth the rate of stolon tubes. But upright tubes grow at al- most exactly one-half the rate of stolon tubes. In conclusion, it has been possible to reduce the development of a Cordylophora colony to the growth and branching of a series of tubes: stolons, uprights, and side branches. The parameters of col- ony shape may be summarized in tabular form: Source Angle and position Spacing Relative growth rate Tube Colonies Model Stolon Upright** Branch** Stolon Stolon Upright 90°, along substratum 90°, away from subst. 45°, away from subst. erratic r—^ 3 mm. — 3 mm. 1 1 *These tubes also differ from stolons in that they bear hydranths and only develop at growing tips. A model has been developed integrating many of these aspects of asexual colony development, and the development of indi- vidual colonies studied in relation to the model. CHANDLER FULTON 293 From my point of view, the major result of this study is that, by reducing the development of a colony to a series of constituent events, it becomes possible to analyze the individual events which give rise to the shape of a colony. Many questions immediately pose themselves. For example, why do upright tubes grow at half the rate of stolon tubes? Why do hydranth-bearing tubes develop only be- hind growing tips, while stolon tubes develop away from these tips? What produces the regular spacing of upright tubes? What de- termines the angle at which each tube leaves its parent tube? As yet, none of these questions has even a preliminary answer, but I hope that at least I have provided you with a more dynamic pic- ture of these hydranth-bearing pipelines. REFERENCES 1. Fulton, C. 1960. Culture of a colonial hydroid under controlled conditions. Science 132: 473-474. 2. Fulton, C. 1960. The Biology of a Colonial Hydroid. Ph.D. Thesis, The Rocke- feller Institute, New York. 3. KiNNE, O. 1958. Adaptation to salinity variations: some facts and problems. In Physiological Adaptation ( C. L. Prosser, ed. ) . Washington, American Physiological Society, pp. 92-106. DISCUSSION MACKIE: Before the discussion turns to the main topics of Dr. Fulton's paper I'd like to comment on the colonial rhythm shown by Cordylophora— the synchronized waves of peristalsis in the hy- dranths. We have also seen this in Dr. Strehler's film of Fennaria. This sort of activity demands a specialized conduction system. Re- cently, R. K. Josephson has recorded action potentials from the stems of Cordylophora and Tuhularia. I cannot give the full details but in Tuhularia there are two rhythmically occurring patterns of activity and one of these patterns has distinct motor effects. I'd also like to reiterate that neurons have been identified his- tologically throughout stems and hydranths in Cordylophora, so there's no need for scepticism about the existence of a ner\ous system in these colonial forms. 294 THE BIOLOGY OF HYDRA : 1961 CROWELL: This frequency (of three times an hour or so for hy- dranth movement) surprises me, because, in the stolon anyway, if one watches the movement back and forth of the fluid, one gets a periodicity in the order of 3 to 5 minutes in all the hydroids I've looked at. FULTON: Have you looked at Cordylophora? CROWELL: Yes. FULTON: In the Cordylophora stolons I've followed, there are a pair of filling and emptying cycles about every twenty minutes, which corresponds to the frequency at the hydranths. CROWELL: I don't doubt that. What you see in the hydranths, I think, is different from the typical back and forth flow in the stolons. FULTON: I don't think so, but we're still in the process of finding out. CHAPMAN: I wonder if you have any information about tlie relationship between culture conditions, such as tonicity, tempera- ture, and pH, on the spacing of these uprights? FULTON: I have voluminous information. Actually not much affects interupright spacing, but many things affect the general pattern of colonies. Kinne has made a thorough study of the ef- fects of different dilutions of seawater and of different tempera- tures on colony pattern. I think that all it would be wise to say right now is that the pattern which I get is the pattern one gets in stand- ard culture solutions at 22° with one feeding a day and all the ritual. One can get almost any colony shape one wants simply by varying one parameter or another. So this is quite a labile system. LOOMIS: What strikes me in your nice growth records is a sort of feeling that the stolon is trying to escape from itself. In other words, it is trapped in its own one dimensional line and starts grow- ing a shoot upwards. Then growth has to escape from this shoot and does so first to the right and then to the left. New growth largely takes place in a new axis at right angles to old growth, which is another way of saying that growth can take place only at an open CHANDLER FULTON 295 and advancing tip. This growth inhibition along an estabhshed stolon may be related to the fact that Cordtjlophora, unlike hydra, will not grow on the bottom of a Petri dish but needs to be suspend- ed on a microscope slide in a beaker of water. The reason for this "Fulton effect" as I call it, seems to be the greater sensitivity of Cordylophora to pC02, for we have found that a pC02 as low as 1.5% atm. inhibits its growth while Hydra can stand up to 10% atm. Thus, Cordylophora on the bottom of a Petri dish sits in its own "halo zone" of high pCOo and inhibits itself, whereas Cordylo- phora on a slide is continually bathed by the thermal currents that exist within a beaker and can easily be shown with methylene blue. Perhaps this apical inhibition of stolon growth by pCOi. may partially explain the growth pattern of Cordylophora. Developmental Problems in Cam^anularid Sears Crowell" Departtiient of Zoology, Indiana University, Blooiningtun, Indiana This report reviews a limited number of experimental studies on the thecate or calyptoblastic colonial hydroid Campanularia flexuosa (Hincks). The selection of topics has been biased by the fact that most studies of developmental problems in hydroids have employed either hydra or athecate (gymnoblastic) species (e.g. Cordylophora, Corymorpha, Hydractinia, Tuhularia). This report 1 think, can be most useful if emphasis is placed upon the peculiar features of thecate forms and on differences between the two groups. No attempt has been made to cover comprehensively the morpho- genesis of thecate hydroids or related work of other investiga- tors. I have tried to point out a few of the interesting unsolved problems The principal topics are: 1. Patterns of colonial growth 2. Alterations of the pattern of growth 3. Aging 4. Regression and replacement of hydranths 5. Reconstitution studies 6. Hvdranth differentition iThe research lias been supported by a research grant (H-1948) from the National Heart Institute, U.S.P.H.S., and by a grant-in-aid from the American Cancer Society. ^Department of Zoology, Indiana University, Bloomington, Indiana and the Marine Biological Laboratory, Woods Hole, Mass. Contribution No. 707 from the Depart- ment of Zoology, Indiana University. 297 298 THE BIOLOGY OF HYDRA : 1961 PATTERNS OF COLONIAL GROWTH This brief report cannot cover the extensive hterature on pat- terns of growth. By 1914 Kiihn (10) had provided a comprehen- sive review and his figures have been used and recopied ever since. Recently Berrill has clarified many points concerning hydroid morphogenesis, and his recent book (2) provides us with both an excellent survey and a bibliography. The pattern of colonial growth of a typical athecate hydroid is shown in Figure 1 A. The oldest hydranth, terminal in position, is designated as 1, the next oldest, 2, etc. There is a zone of growth just proximal to each hydranth. Each such zone contributes to furth- er increase in the size of the colony in two ways: it lengthens the pedicel or stem in which it lies, and it gives off laterally at regular intervals a new hydranth bud with its own distinct growth zone. A newly produced hydranth initially has few tentacles and is small. Tentacles are gradually added as the hydranth grows in size. It is easy to determine the relative ages of the hydranths of a col- ony with this growth pattern, on the basis of both the position and 7 /^6 H 2 3 Fig. 1. Diagrams to show the growth pattern of colonial hydroids. A. The pattern typical of most colonial athecate species. B. The pattern of many thecate species, e.g. Campanularia, Obelia. The black regions are zones of growth. The numbers show relative age of hydranths. From Kiihn (10). SEARS CROWELL 299 the size of hydranths. This is clearly illustrated in the photographs of Cordijlophora (Fig. 2) and Pcnnaria (Fig. 3); and both corre- spond almost perfectly with the idealized pattern of Figure 1 A. Fig. 2. Pattern of a small colony of Cordylophora. graph by Charles Wyttenbach. From a color photo- Figure 1 B illustrates the typical growth pattern found in many thecate species. Growth zones give rise to the stems (pedicels) of new hydranths but do not add to the length of the stem itself. Hence the order of the age of the hydranths (in a young colony) is from the base upward, 1, 2, etc., in Fig. 1 B. The youngest hy- dranth is terminal— the opposite of the pattern in athecate species. In thecate species the pedicel of a new hydranth is completely formed before the hydranth itself is produced. After the pedicel at- tains its full length it enlarges at its tip to make a hydranth bud, which then quickly differentiates into a hydranth. By the time the hydranth emerges from its enclosing hydrotheca, it is fully func- tional, and has its full set of tentacles and its full size. It grows 300 THE BIOLOGY OF HYDRA : 1961 no more. The photograph of Campamdaria (Fig. 4) shows that all hydranths are of the same size. The bud of the hydranth which will be produced next is at the top, and proximal to this is the begin- ning of the outgrowth of the next pedicel. Some species of both thecate and athecate hydroids are solitary, and there are other species in which all hydranths arise only from the attaching stolon (e.g. Htjclr actinia). Yet another pattern of colonial growth, in which the growth zones are apical, is seen in sertularians and plumularians— presumably the most advanced of the thecate hydroids. These too provide challenging problems for experimental morphologists but cannot be considered here. Fig. 3. Pattern of a small portion of a colony of Pennaria. From a color photo- graph by Charles Wyttenbach. Fig. 4. (right). Pattern of growth for Campanularia flexuosa. From a color photo- graph by Charles Wyttenbach. SEARS CROWELL 301 The precise patterns of growth in hydroids tempt one to con- struct mathematical models such as those which Fulton has devel- oped and presented so well in this symposium. I am confident that similar models could be constructed for Campamdaria. The pre- ciseness of patterns also invites us to attempt to alter them. ALTERATIONS OF THE PATTERN OF GROWTH The basis for our first studies on Campanularia was the belief that procedures which would alter the pattern of growth would give some insight into the underlying controlling conditions. Young colonies grown at different temperatures gave colonies or similar form, but their growth schedule was strikingly altered. At higher temperatures the apical growth of each new pedicel and hydranth was accelerated, l)ut at cooler temperatures the initiation of the growth of each new pedicel occurred so much sooner that these colonies as a whole grew just as rapidly (7). This experiment showed that the factors which control the initi- ation of new growth are different from those which control rate of growth in an already established growing region. In a more elaborate experiment all growth zones and prospec- tive growth zones were compared in colonies kept at different nu- tritional levels. Figure 5 C shows diagramatically all of these zones. It could be predicted that with sub-optimal feeding there must be either a general uniform slowing down of all activities or a favoring of some at the expense of others. The latter proved to be the case. In general, lowered nutrition did not greatly affect the rate of growth in an already established part, but it did delay or stop the initiation of new growth. For example, the main stolon grew almost as well in nearly starved specimens as in well fed ones, and it produced new uprights. However, the initiation of subterminal growth by the uprights was delayed. As a consequence of these two effects the whole pattern of partly starved colonies was strikingly different from that of well fed ones. The two were about equally extensive along the substrate, but the height was conspicuously different. 302 THE BIOLOGY OF HYDRA : 1961 It is easy to conjecture that this difference is adaptive in na- ture: It is better for a colony at an unsatisfactory feeding site to move along than to add more feeding units where it is. Fig. 5. Campanularia. A. Technique of subculturing by placing an up- right beneath a thread which has been tied around a slide. The new growth is suggested by the dotted line. B. Pattern of a colony of the age used in the nutrition experiment discussed in the text. The numbers designating age of the upright correspond with those in Fig. 6. C. The zones of growth and prospective growth in Campanularia are indicated: W. to Z. (With permission; Fig. 1 of ref. 4). AGING The most striking observation, by serendipity, in the experi- ment just discussed was that the increase in height of the older uprights (stems with their hydranths) was much more adversely aflFected by reduced nutrition than the comparable growth of young- er uprights in the same colony. The growth in length of the up- rights depends on recurrent initiation of each new node— it is SEARS CROWELL 303 intermittent, not continuous. Figure 6 summarizes the experimental results. The groups are arranged in the order of decreasing food supply, and in each group the oldest upright is No. 1 at the left. In the well fed groups, glut and 4 2, old and young uprights had grown at the same rate. In all the others the younger grew faster (4). The effect of age of stem in slowing or limiting terminal growth was studied further (8). In one test the more basal levels of an up- right were removed every few days so that it consisted of only the 4 to 8 youngest hydranths. The terminal growth, in these cases, did not stop; the total length of stem produced was more than three 12 3 4 S 6 7 8 9 10 u uuuuuuuuu I 13 4 5 6 7 B 9 10 II uuuuuuuuuuu GLUT 2- — <- I. dD nn nD r~ini~ir~ir-ir-ir-n— inrnf— 1 JD. oD 2 5/ JIL jnEL ^n lO dddD r-ii-ir-n-ir-ir-n-ii-ir-n-n-i 34 56 78 9 10 II Fig. 6. Terminal growth related to nutritive level and to age (height) of upright. The subfigures are arranged in decreasing order of nutritive level; within each the oldest upright is at the left. (With permission; Fig. 3 of ref. 4). 304 THE BIOLOGY OF HYDRA : 1961 times greater than that observed in normal specimens or than that reported as the maximum height for this species in nature. Oth- er experiments, but not all, showed evidence of an aging factor inhibitory to growth. These studies are being continued. REGRESSION AND REPLACEMENT OF HYDRANTHS In all thecate hydroids which have been examined hydranths are short-lived; they regress and are resorbed after about one week (3). In Figure 4, for example, it may be noticed that there is only a pedicel at the location, lowest left, where the oldest hydranth "ought" to be; it had regressed. In this symposium. Dr. Strehler is presenting much of our information ( 12 ) concerning this regres- sion-replacement cycle and its implications for the understanding of aging. When regression occurs, the materials of the hydranth go back into the colony and are available as nutrition for further growth, a point which has been proved by Berrill ( 1 ) and Nathanson (11). [See comment by Crowell in the discussion of the paper by Streh- ler in this symposium (p. 396).] In contrast with thecate species athecate hydroids do not re- gress, so far as we know, except under adverse conditions. We have, for example, records of Cordylophora hydranths which lived for more than three months even when food was limited and growth was almost at a standstill ( not previously reported ) . RECONSTITUTION EXPERLMENTS Hydroid tissues can be dissociated mechanically giving tiny clumps of cells, which can be pushed together into a loose mass. In both thecate and athecate species these clusters reorganize them- selves into a double-layered hollow sphere with epidermal cells on the outside, endodermal cells inside. Up to this point thecate and athecate tissues are similar in behavior. The subsequent events differ strikingly and emphasize in a different way the contrast be- tween the two groups in the manner in which a hydranth develops. SEARS CROWELL 303 In a day or two in Coidijloplwra and other athecate forms a small bud (sometimes several) appears on the upper side of the cellular ball and quickly develops four or so tentacles. If fed, it will grow. In Campaniihha, a growth zone appears on the ball and produces either a stolon or a pedicel. This grows out for several days using the materials in the ball. Finally, after about a week, in exactly the same sequence as in ordinary hydranth development, a new small but complete hydranth is produced. Figure 7 shows sketches of this for Campanularia. Fig. 7. Sketches of the production of hydranths from dissociated tissues of Campanularia. At the top are clumps of cells which have been pushed together. Within a few hours these rearrange themselves into a hollow ball. This ball may produce a hydranth in either of two ways: at the left, by the production of a stolon from which a pedicel and then a hydranth develops; at the right, by the production of a pedicel at the top of which a hydranth develops (From Hartman, ref. 9). 306 THE BIOLOGY OF HYDRA : 1961 Here again we must raise the question: What is it that is being moved from the old jDart to make the new? Are cells moving? Are old cells breaking down to give substances that are moved and reutilized? We do not yet know the answers. At Indiana, Mr. Hartman (9) undertook to find differences among the tissues taken from different parts of a colony in re- spect to their capacities when dissociated. No differences were found among tissues from stem, stolon, or early hydranth buds of Campamdaria. Tissue taken from adult hydranths, however, did not reconstitute. This led, naturally, to tests of different stages of hy- dranth development. When a late stage of hydranth development was used, but one in which there was not yet any visible differentia- tion, Hartman found that the tissues reaggregated and within a few hours produced differentiated hydranth parts with an ir- regular organization. Two examples showing patches of tentacles, and in one case a hypostome, are illustrated in Figure 8. Evidently Fig. 8. Two examples of the irregular structures which differentiated when tissues from a late hydranth bud were dissociated and allowed to re- aggregate. There are patches of well developed tentacles, and in the example at the right there is a hypostome (From Hartman, ref. 9). each region of the scrambled tissues was already set in the course of its differentiation. A further test of the distal tissues of buds at this age showed that they were like the whole in making irregular structures at once. But the tissues taken from the proximal halves of such buds reconstituted according to the same pattern as stem, stolon, or early bud tissue. SEARS CROWELL 307 HYDRANTH DIFFERENTIATION^' The manner of development of thecate hydranths, their failure to grow, and the fact that they regress after only about a week sug- gest that they have little regenerative or regulative ability. We have cut tentacles from young hydranths and find that they do not regenerate appreciably. If the hypostome is cut off regression en- sues within a few hours. To carry this matter further back into stages of hydranth development we undertook several types of sim- ple operations on hydranth buds. Athecate hydranths which have had parts removed replace them. The three sketches of Figure 9, for example, illustrate the -T-rrrfi-^-TTr Fig. 9. Rapid restoration of tentacles and hypostome in Cordylophora following the removal of the hypostome and most of the tentacles. ^The experiments described in this section have not been presented elsewhere except in abstracts (5, 6 ) . 308 THE BIOLOGY OF HYDRA : 1961 quick regeneration which followed removal of the hypostome and most of the tentacles in a small young hydranth of Cordylophora. An analagous operation, illustrated in Figure 10, was performed several times on hydranth buds of Campanularia. Both the excised piece and the part which remained proceeded to differentiate just as they would have if no operation had been made. The isolated lit- tle pieces consisted of little more than tentacles and a hypostome. Such little creatures captured Artcmia larvae and passed them into the hypostome. They lived unchanged for about four days— a nor- mal life span for an unnourished hydranth. Similarly the "half hy- dranths" still on the colony showed normal activity but no restitu- tion of the missing tentacles. fh hr5 Fig. 10. The left half of the upper portion of a late hydranth bud of Campanularia is cut off. Both parts differentiate just what they would have produced normally, and there is no later restoration of missing parts. In another series of experiments we cut off and isolated very young hydranth buds of Campanularia, as shown in Figure 11. These were of such small mass that it would be impossible for them to develop a normal hydranth. Had these been athecate hydranth buds one would have predicted that they would produce either nothing, because of the small size, or at best a tiny hydranth. These isolated SEARS CROWELL 309 Fig. 11. Profile sketches of the morphogenesis of an isolated early hy- dranth bud of Campanularia. The finally differentiated disk consists of little more than a hypostome surrounded by a full circle of tentacles. The outer line represents the secreted perisarc; the tissue is stippled. buds of Campamdaria, however, showed an extraordinary ability to continue to perform the activities ordinarily performed by the distal-most tissues of a normally developing hydranth. They gradu- ally spread themselves laterally, laying down externally the hydro- thecal perisarc, and they continued to do so until a hydrotheca of ordinary size was produced. By this time the tissue itself was only a thin disk at the position where hypostome and tentacles would differentiate in a whole bud. Then the disk differentiated into just these distal-most parts. The whole process just described proceeded much more slow- ly than is the case in normal development. If one were dealing only or mainly with cell migration it would be expected that the events could occur at nearly normal speed. The slowness suggests that new cells are being produced, as is believed to be the case in ordinary hydranth development, and old ones are being de- stroyed and utilized. Regardless of the validity of this sugges- 310 THE BIOLOGY OF HYDRA : 1961 tion, it is clear that distal-most tissues have held rigidly to the se- quence of events characteristic of these tissues in normal development. CONCLUSIONS It is clear that the pattern of colonial growth can be altered in Campanularia by changes both in temperature and in nutritive level. The alterations are largely due to the sensitivity of zones of prospective growth. Differences in hydranth morphogenesis are striking when one compares the processes in thecate and athecate species. In the thecate form, Campanularia, a hydranth of full size is produced by a series of events which are not easily altered; they show little abil- ity to regulate. Once produced thecate hydranths do not grow, they do not regenerate parts which have been removed, and they regress and are resorbed after living for only a few days. In all these respects the reverse is true for athecate species. We know that old parts are utilized for new growth, but we do not know in what form materials are moved: as tissues? cells? fragments? chemical substances? This needs study. More attention also should be given to the initiation of new growth by zones of prospective growth. For analysis of these particular problems the- cate species, such as Campanularia, are probably better than athe- cate forms. ACKNOWLEDGEMENTS The author must acknowledge the assistance of Malcolm Rusk and Ruth Curtiss Telfer who were with him at the beginning of the studies of Campanularia; of Charles Wyttenbach who has made many contributions of ideas and time and whose photographs have been copied here; of Fred Wilt, Richard Manassa, Annelle Gibbon, Jean Lowiy, Maurice Hartman, and Pat Clapp all of whom have had some part in the work summarized here. The paper ought to be dedicated to the memory of Frederick S. Hammett who long ago proclaimed the special virtues of Cam- panularia for studies of growth. SEARS CROWELL 311 REFERENCES 1. Berrill, N. J. 1949. The polymorphic transformations of Ohelia. Quart. J. Micr. Set. 90: 235-264. 2. Berrill, N. J. 1961. Growth, Development, and Pattern. W. H. Freeman and Company, San Francisco. 555 pp. 3. Crowell, S. 1953. The regression-replacement cycle of hydranths of Obelia and Campanularia. Physiol. Zool. 26: 319-327. 4. Crowell, S. 1957. Differential responses of growth zones to nutritive level, age, and temperature in the colonial hydroid Campanularia. J. Exp. Zool. 134: 63-90. 5. Crowell, S. 1960. Non-regulative differentiation in the thecate hydroid Cam- panularia. Anat. Rec. 138: 341-342. 6. Crowell, S., and M. Hartman. 1960. Reorganization capacities of dissociated tissues of Campanularia flexuosa. Anat. Rec. 138: 342. 7. Crowell, S., and M. Rusk. 1950. Growth of Campanukiria colonies. Biol. Bull. 99: 357. 8. Crowell, S., and C. Wyttenbach. 1957. Factors affecting terminal growth in the hydroid Campanularia. Biol. Bull. 113: 233-244. 9. Hartman, M. E. 1960. A study of tlie reorganization capacities of dissociated tissues of Campanularia flexuosa. M.A. Thesis, Indiana University. 10. KuHN, A. 1914. Entwicklungsgeschichte und Verwandtschaftsbeziehungen der Hydrozoen. I Teil: Die Hydroiden. Ergeb. Forschr. Zool. 4: 1-284. 11. Nathanson, D. L. 1955. The relationship of regenerative ability to the regres- sion of hydranths of Campanularia. Biol. Bull. 109: 350. 12. Strehler, B. L., and S. Crowell. 1961. Studies on comparative physiology of aging. I. Function vs. age of Campanularia flexuosa. Gerontologia 5: 1-8. DISCUSSION FULTON: I am much impressed with the similarity of the growth pattern of Campanularia and Cordijlophora. For example, if you starve Cordt/lophora, the stolon is the least affected part. CROWELL: We didn't say anything about longevity. Cordijlop- Jiora hydranths don't die after a week or so as do calyptoblast hy- dranths. FULTON: As far as I know Cordijlophora hydranths never die. STREHLER: I would like to speak on that point. We have studied Bouganvillia hydranths for as long as 25 days and haven't seen a single individual die. They continued to increase in size as they got older. On the other hand the oldest Campanularia that we've 312 THE BIOLOGY OF HYDRA : 1961 ever found is eleven days of age. That's at about 17°. You can find older ones if you lower the temperature. Clijtia, by contrast to Campanularia adjusts in size to the amount of food that is avail- able. In Campanularia you get essentially the same size hydranths regardless of how well or poorly one feeds the colony. If it starts to make a hydranth it makes one of the standard size. Although Clytia hydranths do vary in size they don't grow after they're fully formed. You can get very tiny hydranths if the colony is starved and some hydranths as large as Campanularia if they are well fed. If Clytia is growing on Artemia and for some reason they don't catch their food on a regular basis, they very soon get to a size where they can't ingest Artemia because none of the hydranths are large enough. CROWELL: There is some variation in Campanularia. If one uses tissue masses of different sizes, one finds that there is a lower limit where one gets no hydranths. Above that, one gets specimens somewhat smaller than normal and with a smaller tentacle number. Then if one uses still larger masses one gets correspondingly larger hydranths. It's not very striking though. STREHLER: There is one implication in a word that you used. You said that there was a zone of "proliferation" down near the developing bud and I just wonder how you would explain certain experiments we did last summer which consisted of giving a colony 100,000 r of X-rays, enough so that the slides on which they were growing became deep amber in color. Still, after ten days, a few new hydranths were formed in the radiated colony. Just a few, it's true. CROWELL : Subterminal hydranths? STREHLER: These were replacements, I believe, i.e. subterm- inal. The point is, that it's hard for me to see how cell division could occur after that amount of radiation. I would propose alter- natively, that there are cells which have aheady divided and which probably lie in the stolon. At the proper signal these cells migrate into the region of what one might call growth, but which I think may better be considered as regions of differentiation and morphogenetic movement where no cell division is taking place. CROWELL: I think what you suggest is x^erfectly possible. The SEARS CROWELL 313 evidence for mitosis in these growing tips is most unsatisfactory. Berrill says mitosis occurs in growing hydranths but he never pre- sents any illustration of this mitotic activity. This is one reason why Mr. Lunger is now trying to look at these growth zones using the electron microscope. We hope to understand these processes at the cellular level. We certainly cannot right now. STREHLER: At the end of this afternoon's session I hope to show some time-lapse movies of an irradiated colony. I call this movie "On the Beach." FULTON: Can I interject something? I have been trying very hard to find out where cell division occurs in Cordylophora. I don't know whether it's me or the animal, but I cannot see any chromosomes. If anybody knows how to see mitosis in adult hy- droids I would be very happy to hear of it. CROWELL : Send me a copy of the letter. LYTLE: The only place we have been able to find mitotic figures in Cordylopliora is in early embryos. FULTON: This is easy. LYTLE: Not as easy as one might expect. We had to look at a lot of sections to find any mitotic figures. FULTON: Adult tissues must divide for they grow about one- tenth of a millimeter an hour. There must be cell division some- where. SLAUTTERBACK: In reference to the transected bud, I was quite interested in your "rob Peter to pay Paul" expression. I take this to mean that any one cell possesses not just a single pat- tern of differentiation, but all the possible patterns necessary for the production of a whole hydranth. And in this case, a cell may car- ry out each of these patterns sequentially until it has gone through all the steps normally carried out by many different cells. Do I understand correctly, or is there some mitosis going on and it is the daughter cells which make tentacles where the parent cell has made perisarc or stem or something else? CROWELL : I don't think we know. 314 THE BIOLOGY OF HYDRA : 1961 SLAUTTERBACK: This intrigues me very much because we've come upon dedifferentiation and redifferentiation in the pedal disc. If one amputates the pedal disc, the secretory cells are soon re- placed but not from the undifferentiated interstitial cell as might be expected, but by partial dedifferentiation of cnidoblasts fol- lowed by differentiation into secretory cells. This observation is possible because the nematocysts persists in these cells throughout the process. In fact, the mature secretory cells often contain a part- ly disintegrated nematocyst. Furthermore, even the organelle devel- opment characteristic of the cnidoblast persists for a time after the secretory cell, with its very different organelles, has begun to func- tion. I think this is one of the rarer demonstrations of a partial de- differentiation and then redifferentiation of the same cell into an entirely different cell line. I wonder if that is what is going on in your situation, or whether you have mitosis intervening, or what? CROW ELL: If one starts with a little colony consisting of a stolon and a few hydranths, and does not feed it, one often finds that there is new growth of the stolon and then production of new hydranths from the new stolon. I have seen this in Campanularia and Cor- dylophora; Berrill has described it. Of course, as new stolon and hydranths are growing at one end, old hydranths and stolon are regressing at the other end. One does get regression of hydranths of Cordylophora in this situation; however, there is no regression in well fed colonies. Of course, such a system gradually gets small- er—as long as it lasts it produces new parts at the expenses of the old. SLAUTTERBACK: I wonder whether there is a degradation of cells followed by reuse of the degraded material to make new cells, or whether there is a dedifferentiation, migration and redifferentia- tion of the original cells from the old hydranths. CROW ELL: That is just the point that is not understood. SLAUTTERBACK: In the pedal disk it is the old cells that are reused, i.e. redifferentiated. FULTON: This must also be the case in Cordylophora because the stolons of a starving colony will keep extending over the slide for months; the hydranths and stolon tissue behind the advancing SEARS CROWELL 315 tips being resorbed and regenerating continuously. Since there is no other source of nutrients, old cells must be reused. In line with this, I wanted to emphasize that there is normally no regression of hydranths in Cordylophora. Kinne ( 1956, Zool. Jahrb., Abt. Phy- siol. 66: 565 ) followed individual hydranths for about 140 days, and I have observed them for several months with no indications of regression. STREHLER: Does the hydranth continue to get larger during all that time? FULTON: They may grow very very slowly. They reach adult size I'd say in about a week of growth. CROWELL: You haven't said whether new cells are being pro- duced by using substances derived from old cells, or whether the same cells are producing the new parts by migrating. FULTON: I don't know. All I'm saying is that they can't be using up too much because they will go on for months. CROWELL: It will go on a long time. LYTLE: Or it's a very efficient system for recycling materials. STREHLER: It would be very interesting to know whether the same cells stay in the fully formed hydranth if it's not growing. That is, is there a cycle of cell replacement? One should be able to find out by seeing the effect of large doses of X-radiation on the longevity of CorchjJophora. Will it kill them in the same doses which double the longevity of Campanularia? SLAUTTERBACK: Every attempt we have made to demonstrate an increased mitotic rate following amputation of hydra heads has been unsuccessful. The formation of a new head with its tentacles appears to be strictly a matter of migration of cells from the column. There is no change in the level of differentiation of these cells nor is there any visible increase in mitotic activity. CROWELL: How successful are you in finding mitosis down in the lower region? SLAUTTERBACK: We can see them fairly commonly in the in- 316 THE BIOLOGY OF HYDRA : 1961 terstitial cells with the electron microscope but not with the light microscope. LENHOFF: We measure changes in the number of nematocysts in H. littoralis using a specific test for hydroxyproline, the imino acid that makes up much of the nematocyst capsule. We find that de- capitated Hydra which regenerate complete sets of tentacles show no net increase in hydroxyproline although starved Hydra are able to synthesize this unessential imino acid. Thus, it appears that re- generating animals use the nematocysts that they already have in their body tubes, and no new increase in the number of cnido- blasts occurs by cell division. BURNETT: We easily demonstrate mitosis in whole hydra by staining them in methylene blue at pH 7 after first digesting them with ribonuclease (1 mg. ml. for 3-5 hours). The enzyme re- moves all cytoplasmic RNA and makes the hydra more transpar- ent. By simply scanning the surface of the whole animal, one can see nests of interstitial cells in synchronous division. MACKIE: I have often seen mitosis in the cell-body part of epit- heliomuscular cells. The fiber part is not affected. It's rather inter- esting in silver preparations because the achromatic figure is chro- matic and the chromatic figures is achromatic. WOOD: Couldn't one use radioautography to trace the formation of DNA? This might give an indication of the mitotic rate or turn- over of cells. FULTON : If you can figure out how to get labeled thymidine into the animals, I'll be happy to do it. I've tried and seen nothing. Patterns of Budding in the Freshwater Hydroid Craspedacusta Charles F. Lytle- Dcpartment of Zoolofiy, Indiana University, Bloominp.t()n, Indiana Craspedacusta sowerbii Lankester is a freshwater hydrozoan observed sporadically in many lakes, ponds, quarries, and im- poundments of North America. It is best known for its conspicuous medusa stage (Fig. 1), although the life cycle also includes a nearly transparent polyp stage, which is microscopic and devoid of tentacles. These polyps are attached permanently to various sub- merged objects and grow as single hydranths or more commonly as small colonies of two to seven simple hydranths joined at their base (Fig. 2). There is no investing perisarc on the hydranths of C. sowerbii though a loose case of detritus can usually be seen around the basal portion of the hydranths and the base of the colony. This detritus is held by a mucous secretion of the epidermal cells. An individual hydranth is typically flask-shaped and measures approximately 0.3-0.5 mm. in length, while a colony composed of several hydranths may reach an overall diameter of two to three millimeters. The hydranths may be divided roughly into four regions: 1) a distal capitulum bearing several dozen nematocysts; 2) a constricted neck region; 3) expanded budding region; and ^This paper is contribution No. 706 from tlie Department of Zoology, Indiana Uni- versity and is based on a portion of a thesis submitted to the faculty of Indiana University for the Ph.D. degree. This investigation was supported in part by a pre- doctoral fellowship CF-8674 from the National Cancer Institute, United States Public Health Service. 2 Present address: Department of Zoology, Tulane University, New Orleans 18, Louisiana. The author wishes to express his appreciation for the guidance and support of Drs. Sears Crowell and Robert Briggs. 317 318 THE BIOLOGY OF HYDRA : 1961 Fig. 1. High-speed photograph of a swimming medusa. Magnification approximately 3X. 4) a basal region by which it attaches to the substrate and/or to neighboring hydranths. These hydranths carry on asexual reproduction by producing three types of buds ( Fig. 3 ) : 1 ) hydranth buds which remain attached to the parent to form small colonies; 2) frustule or planu- loid buds which separate from the parent and creep a short dis- tance before developing into new polyps; and 3) medusoid buds which are released as free-swimming medusae. Under optimal conditions all three types of buds are formed laterally as outgrowths of the body wall near the middle or budding region of the hy- dranth {vide ref. 21). Differential growth in the case of hydranth buds results in the subsequent basal attachment of adjacent hy- dranths. Several previous workers have observed the budding processes of Craspedacusta polyps (3, 6-10, 12, 14-22), but only Reisinger (20, 21) and McClary (14) have studied specific factors which influence the production of buds under laboratory conditions. Reisinger (20, 21) found that a sudden elevation of tempera- ture from 20" to 25-27° could initiate medusa budding. Mc- Clary (14) studied the growth and reproduction of polyps at CHARLES F. LYTLE 319 Fig. 2. Macrophotograph of a polyp colony with three hydranths. The neck and capitulum of the lower hydranth are reflexed. Magnification ap- proximately 40X. HYDRANTH FRUSTULE MEDUSA Fig. 3. Diagram illustrating the three types of buds produced by Cras- pedacusta polyps. 320 THE BIOLOGY OF HYDRA : 1961 four different temperatures and demonstrated that a temperature shift was not necessary for the initiation of medusa budding. He also observed that the three budding processes exhibited different tem- perature optima. In his experiments, frustule production was maxi- mal at 25°, hydranth budding was maximal at 12° and 20°, and medusa buds were produced only at 28°. The work discussed in the present report is concerned with the sequence of budding in developing colonies, some effects of temper- ature and nutrition on the growth and reproduction of polyp col- onies, and certain physiological interactions between the different budding processes. METHODS AND MATERIALS Polyps of C. sowerbii were collected on glass microscope slides submerged in a limestone quarry pool near Bloomington, Indiana, where populations of the medusae were known to occur regularly (13). Laboratory stocks were established; and for these experi- ments frustules were removed from stock cultures, isolated in Syra- cuse watch glasses, and incubated in an 18.5° (± 1.5°) constant temperature room. Approximately two days later the culture dishes were transferred to shallow glass trays through which charcoal- filtered tap water was continuously passed. In most experiments the shallow glass trays were partially immersed in constant-temper- ature baths. Culture water was provided from a charcoal filtration system manufactured by the Illinois Water Treatment Company, Rockford, Illinois (Model No. CC-24). Polyps were fed counted numbers of oligochaete worms {Aeo- losoma hemprichi Ehrenberg) by hand on alternate days or at spe- cified intervals. The worms were cultured on rice-agar plates con- taining a mixture of protozoa and l^acteria as described by Brand- wein (2). PATTERNS OF BUDDING The basic pattern of development and reproduction of a polyp colony is illustrated in Figure 4. Fifteen frustules were isolated at the CHARLES F. LYTLE 321 211 2 C 15 10 5 .7 5 .50 .25 Hydranths O o o- Frustules - o o- -O O O o o o -o o^,^^ °-, Medusae 6 8 10 12 14 16 WEEKS Fig. 4. Budding pattern of 15 polyps reared at 27°(d=2 ). Values on the abscissa represent the number of buds of each type produced per colony. start of this experiment and cultured at 27° (± 2°). All frustules had differentiated into polyps and produced an average of two hydranths each by the end of the first week. Hydranth budding declined during subsequent weeks but increased to a second peak during the 13th week. Frustule production began during the sixth week and declined to a minimum during the tenth week before ris- ing to a new high during the 15th week. Medusa buds appeared during the seventh week ( immediately following the decline of frus- tule production) and were produced continuously through the 16th week. The basic sequence of events exhibited by these colonies was an initial phase of rapid hydranth production, a phase of rapid frus- tule production, and a phase of medusa budding. Secondary in- creases in hydranth budding and frustule budding were also ob- served during the latter portion of the phase of medusa budding. 322 THE BIOLOGY OF HYDRA : 1961 A similar sequence of events was also observed in colonies reared at 20° and at 19-23 : At 20° {± 1°) (Fig. 5) the frequency of all three types of budding was reduced though the same basic bud- ding pattern was observed: an obvious initial peak of hydranth production, a phase of rapid frustule production followed by a slight decline, and a phase of medusa budding. At 19-23' (Fig. 6) all three types of budding were increased and the three phases of asexual reproduction were again clearly demonstrated. Colonies grown at different temperatures clearly demonstrate that under the relatively constant laboratory culture conditions three expressions of morphogenesis occur in a sequence of distinct phases. These activities are not mutually exclusive, but seem to ex- hibit a clear separation between the different phases. The com- mon basic pattern was observed at all temperatures, though certain specific variations were noted in the duration of each phase as well 201 IC 1.0- 0.5 4.5 3.0 1.5 0.4 0.2 _i 1- -O O' Hydranths Fr ustules Medusae 3 4 10 II 12 Fig. 5. Budding pattern of ten polyps reared at 20 (±1). Values on the abscissa represent the number of buds of each type produced per colony. CHARLES F. LYTLE 323 9-23 C H y d r a n t h s O O- o- Fr u St ul es o o o o- Medusae 10 12 14 16 18 WEEKS Fig. 6. Budding pattern of 13 polyps reared on a water table with the temperature rising slowly from an initial 19 to a maximum of 23 and re- turning to 19 at the end of 16 weeks. Values on the abscissa represent the number of buds of each type produced per colony. as in the absolute and relative numbers of buds of each type pro- duced at the various temperatures. My temperature experiments also indicate the existence of cer- tain interactions between the three budding processes. Figure 7 il- lustrates the relationship between hydranth budding and medusa budding. At all three temperatures there was an initial rapid rise in the number of hydranths per colony, medusa buds appearing only after the production of hydranths ceased or greatly declined. Me- dusa buds were produced earliest at 20' when the total number of hydranths produced was the smallest. Medusa buds were produced latest in the 19-23 colonies when the total number of hydranths was the greatest (Table 1). When subjected to statistical analysis ( analysis of variance ) , differences in the time of appearance of the 324 THE BIOLOGY OF HYDRA : 1961 o o ^^° 14 • I / I9-23°C 12 - 1 o 10 - o- o- ° o / 8 V / / c--"— — e 6 4 / «/ 27t2°C 2 - /^•^•" ^K — • 2 01I°C ..^ 8 10 12 14 WEEKS Fig. 7. Relationship of colonial growth and the initiation of medusa bud- ding at different temperatures. Values on the abscissa represent the cumu- lative number of hydranths per colony. Arrows indicate the appearance of the first medusa buds. first medusa buds were found to be significant at the 95% level. The relationship between the production of frustules and the appearance of medusa buds is illustrated in Figure 8. Colonies at all three temperatures exhibited an early rise in the production of frustules and a later decline. In each case frustule budding began after a decline in the initial rapid production of hydranth buds, and medusa buds appeared immediately following the decline in production of frustules. These experiments clearly suggest that hydranth budding may limit medusa budding, since medusa buds always appeared after hydranth budding had declined or stopped. Furthermore, the short- ened phase of hydranth budding at 20° is associated with the earliest formation of medusa buds, while the extended period of CHARLES F. LYTLE 325 TABLE 1 Age of polyp coionies at the appearance of the first medusa bud at various temperatures. 20° ( ±1°) 27° (±2°) 19-23° 45 d ays 45 davs 72 days 47 50 72 47 51 73 51 56 72 56 54 55 56 58 62 65 57 57 60 62 75 80 78 67 68 74 Mean: 48.6 days 56.3 days 73.1 days S.D. 2.7 2.8 1.6 hydranth budding at 19-23'^ is associated with a significant delay in the appearance of medusa buds. It also appears that medusa bud- ding may in turn limit frustule production since in all cases the appearance of medusa buds is preceded by a decline in the produc- tion of frustules. Further evidence for this interaction between medusa budding and frustule budding has been provided by McClary ( 14 ) . He ob- served no medusa budding in colonies reared at 12°, 20°, and 25°. In each of these groups there was an irregular but progressive in- crease in the rate of frustule budding for 102 days. His 28 colonies exhibited a rise and subsequent decline in the production of frustules, with the decline corresponding to a maximum in medusa budding. To study further interactions between hydranth budding, frus- tule budding, and medusa budding, we have investigated the effect of increased and decreased nutrition on polyp colonies in several ways. In the previous experiments described, frustules for the estab- lishment of experimental colonies were taken from stock cultures 326 THE BIOLOGY OF HYDRA : 1961 15 ■5 12 I 9 - 2 3 C Fig. 8. Relationship of frustule budding and medusa budding at different temperatures. Values on the abscissa represent the number of frustules produced per colony per week. Arrows indicate the appearance of the first medusa buds. maintained at 23" or below. Frustules from such cultures were gen- erally opaque as a result of large reserves of food material contained in the gastrodermal cells ( 16, 17 ) . These food reserves occurred in distinct cytoplasmic granules or droplets and appear to be simi- lar to the "protein reserve droplets ' or "spherules de reserves" con- tained in the gastrodermis of Hydra oligactis (4), Hydra attenuata (23), and in the polyp stage of the African freshwater medusa Limnocnida (1). Histochernical tests have indicated that these granules or "reserve bodies" may contain RNA, DNA, protein, car- bohydrate, and fats in varying proportions. Frustules produced by Craspedacusta colonies cultured at temperatures higher than 23° are appreciably less opaque, indi- cating smaller amounts of reserve food materials. Colonies reared from 27° frustules demonstrate a strikingly different developmental pattern from those rean^d from frustules produced at low temperatures. CHARLES F. LYTLE 327 Figure 9 illustrates the de\'elopment and budding of two groups of animals reared from 27 frustules at two different feeding rates. The animals represented by the open circles were fed on alternate days as in the previous experiments. They exhibited an initial phase of rapid hydranth production followed by the initiation and rap- id increase in frustule production— but no phase of medusa budding and no decline in the rate of frustule production. Therefore, in the absence of medusa budding the available food material went preferentially into the production of frustules. The second peak of hydranth budding during the 13th week does not appear related to the absence of medusa budding in these animals, since a similar secondary peak is observed at the same time in parallel groups of animals grown at this temperature which do produce medusa buds. 26t2°C i=8= H y d r a n t hs „^:c^^° \.,^_.«=e^^ 20 15 10 5 0.5 Frustules o „ / ^~^o o o o Medusae 8 ID 12 14 WE EKS Fig. 9. Budding pattern of colonies reared from 27 frustules at two different feeding rates. Colonies represented by the open circles were fed on alternate days and those represented by the filled circles were fed every third day. 328 THE BIOLOGY OF HYDRA : 1961 The animals represented by the filled chcles were fed every third day. These animals demonstrate that the rate of hydranth bud- ding is not significantly decreased by the lowered nutritional level but that frustule production is differentially affected. Therefore, this experiment provides direct evidence of a physiological interac- tion between medusa budding and frustule budding and further indicates that this interaction is at least partially nutritional. Another experiment with different nutritional levels further illus- trates the interactions between these three morphogenetic processes. Colonies were reared at 23° from frustules taken from low tempera- ture stocks until several hydranths had been formed. These colonies were starved for approximately four weeks to deplete their nutri- tional reserves and were divided into three groups fed at different rates. As indicated in Figure 10, the production of hydranths showed 10 n/2 Daily I Worm/ 2 Days 4 5 WEEKS Fig. 10. Colonial growth at three different feeding rates (n = the number of hydranths per colony at the time of feeding). Cultures maintained at 23° (±1°). CHARLES F. LYTLE 329 30 - 25 20 I 5 10 ^ Daily Fig. 11. Production of frustules at three different feeding rates (n = the number of hydranths per colony at the time of feeding). Cultures maintained at 23 (±:1 ). a direct and proportional increase with increased rates of feeding. The production of frustules, however, was affected differentially (Fig. 11). The animals at the two lower feeding rates produced only a few frustules while the animals at the highest rate showed a large increase in the number of frustules produced. Much of the additional food went preferentially into the production of frustules. The effect of these different feeding rates is summarized in Fig- ure 12. At the lowest feeding rate there were few buds of each type produced. At the intermediate feeding rate there was a 240% increase in the production of hydranths over those produced at the lowest rate of feeding and a 60^ increase in the production of me- dusa buds. Only a 5.1% increase was observed in the production of frustules. At the highest rate of feeding there was a further in- crease (211.8%) in the production of hydranth buds over the inter- 330 THE BIOLOGY OF HYDRA : 1961 ■ 2 4 0% Interr nediate 60% ■ i 5.1 % :^.jr--r--^^ 4 7l.7y« 1 1 1 High ■ 2 11.8% ■ m 1 2.5% — Hydranths Frustules Medusae Fig. 12. Differential utilization of food materials by the three different budding processes at three different feeding rates. The number of buds at each rate is expressed as a percentage of those produced at the next lower rate. mediate rate of feeding, but a much smaller increase in the produc- tion of medusa buds (12.5%). Frustule budding was tremendously increased (471.7%). Thus at the different nutritional levels food material was utilized differentially by the three different budding processes as observed also in the previous experiment. Hydranth budding appears to limit medusa budding since me- CHARLES F. LYTLE S31 dusa buds always appear after growth has decHned or stopped, and because the abbreviated phases of growth at 20° and 27" are associated with the early formation of medusa buds. The longer growth phase at 19-23 is associated with a delay in the forma- tion of medusa buds. Also hydranth budding is the least affected of the three types of budding by lowered nutritional level. A similar inverse relationship between growth and medusa budding was found in Hydr actinia by Hauenschild and in Obelia by Grell (11). Crowell (5) also found a definite order of priority in the utiliza- tion of nutritive substances among the several growth zones of Campanularia when overall growth was experimentally limited. Significantly, he observed that the formation of gonangia ap- peared to require a high nutritive level. Medusa budding appears to limit frustule production since in all cases the appearance of medusa buds is preceded by a decline in frustule budding. The production of frustules always reached an initial peak after the completion of the initial growth phase and de- clined prior to the appearance of medusa buds. This decline in frustule production was most marked in the 19-23 colonies which produced the greatest number of medusa buds, and least pronounced in the 20" colonies where the fewest medusae were produced. In cultures of high temperature frustules which produced no medusa buds, there was no subsequent decline in the rate of frustule pro- duction after the initial maximum was reached. The relationship between hydranth budding and frustule budding was less clearly demonstrated, but there were some indications of a similar inter- action between these two processes also. These experiments clearly demonstrate that temperature between 20° and 27° is not a limiting factor in the production of medusa buds by isolated colonies in culture if sufficient food is provided, contrary to the observations of Reisinger (20, 21) and McClary (14). Studies on nutrition have shown that lowering of the feeding rate within this temperature may diminish and/ or greatly delay the production of medusa buds. Experiments on the effect of various nutritional levels on the budding processes of isolated colonies demonstrate that the three budding processes are affected differentially by increased feeding rates. At very low feeding rates, medusa budding may be reduced 332 THE BIOLOGY OF HYDRA : 1961 or eliminated, few hydranths are formed, and few friistules are pro- duced. At intermediate rates a large proportion of the food materials are utilized in hydranth budding and in medusa budding. Frustule production is still low. At high feeding rates the largest portion of the food materials is utilized in the formation of frustules and pro- portionally less goes into the production of new hydranths and medusa buds. Therefore these experiments provide some physio- logical basis for the observed interactions between these three bud- ding processes and suggest that these three morphogenetic proc- esses are, at least in a sense, antagonistic, involving alternate pathways for the utilization of metabolic su1:>strates. My present hypothesis is that hydranth budding, frustule budding, and medusa budding represent alternate morphogenetic pathways, and that the control of budding in this system may depend upon physiological competition for specific metabolic substrates. REFERENCES 1. Bouillon, J. 1958. Etude monographique du genre Liiiinocnida ( Limnome- dusae). Ami. Soc. Roij. Zool. Belg. for 1956-1957. 87: 254-500. 2. Brandwein, p. 1937. The culture of some miscellaneous small invertebrates. In Culture Methods for Invertebrate Animals. Ed. P. S. Galtsoff et al. Com- stock Publishing Company, Ithaca, pp. 143-144. 3. Browne, E. T. 1906. On the freshwater medusa liberated by Microhijdra ryderi Potts, and a comparison with Linuiocodiuni. Quart. J. Microscop. Sci. 50 ( N.S. ) : 635-645. 4. Burnett, A. L. 1959. Histophysiology of growth in hydra. /. Exp. Zool. 140: 281-342. 5. Crowell, S. 1957. Differential responses of growth zones to nutritive level, age, and temperature in the colonial hydroid Campamdaria. }. Exp. Zool. 134: 63-90. 6. Dejdar, E. 1934. Die Siisswassenneduse Craspedacusta sowerbii Lankester in monographischer Darstellung. Z. Morph. Okol. Here 28: 595-691. 7. Dunham, D. W. 1941. Studies on tlie ecology and physiology of the freshwater jellyfish, Craspedacusta sowerbii. Ph.D. Thesis, Ohio State University, Columbus. 8. Fowler, G. H. 1890. Notes on the hydroid phase of Lintnocodium sowerbyi. Quart. J. Microscop. Sci. 80: 507-513. 9. GoETTE, A. 1909. Microhydra ryderi in Deutschland. Zool. Anz. 34: 89-90. 10. GoETTE, A. 1920. Uber die ungeschlechtliche Fortpflanzung von Microhydra ryderi. Zool. Anz. 51: 71-77. CHARLES F. LYILE 333 11. Hauenschild, C. 1954. Genetische und entwicklungsphysiologische Untersuch- ungen an Kulturen von Hijdractinia echinata Flemm. zur Frage der Se.xu- alitiit und Stockdifferenziemng. Zool. Jahrb., Aht. allg. Zoo/. Physiol. 64: 1-13. 12. KuHL, G. 1947. Zeitiafferfilm-untersuchungen iiber den Polypen von Craspedu- custa sowerhii ( Ungeschlechtliche Fortpflanzung, Okologie, und Regen- eration). Ahhandl. Senckenbeigischen NatuiiorscJwnden Ges. 473: 1-72. 13. Lytle, C. F. 1959. The records of freshwater medusae in Indiana. Pwc. Indiana Acad. Sci. 67: 304-308. 14. McClary, a. 1959. The effect of temperature on growth and reproduction in Craspedacusta sowcrJni. Ecology 40: 158-162. 15. MosER, J. 1930. Micwhydra E. Potts. Sitsber. Gcs. natuii. Freiindc, BerHn. pp. 283-303. 16. Payne, F. 1924. A study of the freshwater medusa, Craspedacusta ryderi. J. Morph. 38: 387-430. 17. Persch, H. 1933. Untersuchungen iiber Microhydra gcrnianica Roch. Z. wiss. Zool. 144: 163-210. 18. Potts, E. 1897. A North American freshwater jelly-fish. Amer. Nat. 31: 1032- 10.35. 19. Potts, E. 1906. On the medusa of Microhydra ryderi and on the forms of medii- sae inhabiting fresh water. Quart. J. Microscop. Sci. 50( N.S.): 623-633. 20. Reisinger, E. 1934. Die Siisswassermeduse Craspedacusta sowerbii Lankester und ihr Vorkommen in Flussgebiet von Rhein und Maas. Nafiir am Nie- derrhein 10: 33-43. 21. Reisinger, E. 1957. Zur Entwicklungsgeschichte und Entwicklungsmechanik von Craspedacusia ( Hydrozoa, Limnotrachylina ) . Z. Morph. Okol. Tiere 45: 656-698. 22. Ryder, T. A. 1885. The development and structure of Microlujdra ryderi. Amer. Nat. 29: 1232-12.36. 23. Semal-van Gansen, P. 1955. L'histophysiologie de rendodernie dc I'hydra d'eau douce. Ann. Sac. Roy. Zool. Belg. for 1954. 85: 217-278. DISCUSSION FULTON: I noticed that the patterns were the same, but the absohite numbers were very different when you grew them at 20' versus 19 to 23 '. Was one of these in the hght and the otlier in the dark, or anything like that? LYTLE: No. The animals in these experiments were all grown in an aquarium room with several large windows. No attempt was made to alter the normal photoperiod of light and darkness. FULTON: So far as you know the 19" to 23^^ versus the 20° are under otherwise identical conditions, but just the temperature varied? 334 THE BIOLOGY OF HYDRA : 1961 LYTLE: No. Unfortunately the conditions in these two experi- ments were not precisely the same, but I do think we can say that temperature is the most important variable here. The 20° cultures were maintained in running water in a constant temperature bath controlled ±1°. The 19-23° cultures, however, were maintained in running water on a water table at the temperature of the incoming water. During the course of this experiment the temperature rose gradually from an initial 19 to 23 and slowly returned to 19° at the end of 16 weeks. There was also a small diurnal variation in the temperature, in the order of about 1 ". Furthermore, because of a technical difficulty there was some difference in the rate of flow be- tween the 20° experiment and the 19-23° experiment, but I doubt that this had any great influence on our results. I believe that the gradual rise and decline of temperature was probably more impor- tant than the small difference in rate of flow or the actual difference in mean temperature between the two experiments, but this has to be investigated further. FULTON: I see that your absolute numbers were a lot bigger there. LYTLE: Definitely. The large colony with 22 hydranths which I showed at the beginning of my talk was grown on the water table with the rise and fall of temperature (19 -23° -19"). I have never gotten colonies this large in cultures grown under more closely con- trolled temperatures within this range. HAND: If I understood your summary, it sounded to me as if you were saying something backwards. You showed that when hy- dranth production falls off, frustule production comes on; and when frustule production falls off, medusa production comes on. It sound- ed as if you were saying that there was a backward action, that the second phenomenon was somehow affecting the first one. What were you thinking about? LYTLE: As I stated in my talk, there appears to be a definite hierarchy among the three budding processes. Hydranth budding has first priority, and it is only after hydranth production slows down that frustule production begins. Medusa budding does not be- gin for some time after hydranth budding has ceased or greatly CHARLES F. LYTLE 335 slowed down. In the interim there is a maximinn in tlie production of frustules. It appears that whenever metaboHc reserves are not being siphoned off by hydranth or medusa budding, they become avail- able for the production of frustules. Possibly the reason for the de- cline in frustule production two or three weeks prior to the appear- ance of medusa buds is that some of the reserve materials are al- ready going into the pathway leading to the production of medusa buds before the actual morphological appearance of buds. In other words, the biochemical machinery is being set in motion. Sim- ilar phenomena have been demonstrated in several other develop- mental systems where biochemical differentiation precedes morpho- logical differentiation. HAND: That's fine. But as I visualized what you were thinking about, it seemed to me that you were saying that there was a feed- back, and there can t have been in time; I think time doesn't run backwards. LYTLE: Not very well, but there is another experiment we have done which further illustrates this point. A group of animals was reared from frustules in the normal way to obtain colonies; then tlie feeding rate was suddenly doubled. In this case there was no significant effect on hydranth and medusa budding, but the produc- tion of frustules doubled. When the feeding rate was again dou- bled suddenly, frustule production once more doul:)led, while hy- dranth budding and medusa budding remained unaffected. There- fore, the additional food went only into the production of frustules. LOOMIS: We have been growing Cijanca artica in known solu- tion for about eight months and have observed a very similar situ- ation to the one you have described in Craspcdacusta. Thus, we find that they will bud indefinitely if fed every day with brine shrimp and then placed in clean water. They give no hint of form- ing medusae under these conditions. I left one culture in the ice box for a month, however, and then found that it had strobilized and was now giving off medusae. The thing that is pertinent to Dr. Hand's question, I believe, is that the new routine of starvation and stagnation without water change stops l^udding and induces 336 THE BIOLOGY OF HYDRA : 1961 medusa formation, probably by a feedback action by inducing partial anaerobiosis in the culture water. This problem is related to the sexual differentiation of H. littoralis which also appears on stag- nation, for in both animals the partial anaerobiosis of stagnation in- duces a second pattern of differentiation to be expressed, much as the butterfly pattern in the caterpillar becomes expressed during metamorphosis. LYTLE: We have done a similar experiment with the scyphisto- mae of Amelia, although our experiments took a lot longer than yours. We placed scyphistomae in a 5" cold room and left them there for about six months with only an occasional feeding. Shortly after we brought them back up into the laboratory (at 18.5 ), they strobilized. This was the only time we have obtained strobilae in the laboratory, although admittedly we haven't tried too serious- ly. We did try different rates of feeding without any success, but when we left them in the cold room they got dirty and eventually strobilized. CROWELL: Something similar happened with specimens of Aurelia which we gave to students at Bellarmine College. They tried, without success, to induce strobilization. Then, by accident, one of the students who had quit working but had a few polyps stored in a refrigerator, got medusae. So we have three explanations. Starvation is important, cold is important, and neglect is important. LOOM IS: Calculated neglect. CROWELL: Not even calculated neglect. Feedback Factors Affecting Sexual Differentiation in Hydra littoralis W. F. LooMis The Looniis Laboratory, Greenwich, Connecticut We have been trying to induce sexual differentiation in Hydra for some years now, because this instance of celhilar differentiation is controlled externally by the water in which these little animals live. This circumstance allows the investigator to analyze samples from cultures that have turned sexual, and then try his hand at recreating such water artificially. In this way, an approach to understanding the biochemical variables that control cellular differ- entiation becomes experimentally possible. We have found Hydra to be nearly ideal for such a study. Thus, any desired level of population density within a culture may be maintained indefinitely by simply removing all the baby Hydra that are produced daily by budding, baby Hydra being distin- guished from their parents by the fact that they do not yet possess buds of their own. Secondly, Hydra may be kept in simple saline 99% of the time, for they can feed on enough brine shrimp in fifteen minutes to supply their nutritional needs for the ensuing twenty-four hours. All the tedious routines of sterile tissue culture, therefore, become unnecessary when this instance of cellular differentiation is selected for study. Thirdly, the end result of cellular differentia- tion in this system is unusually clear-cut, for even an inexperienced observer can identify functional testes (or ova) on a Hydra if a dissecting microscope is available. Finally, since the differentia- tion of interstitial cells into gonadal tissue is an accessory path- 337 338 THE BIOLOGY OF HYDRA : 1961 way over and above their usual differentiation into nematocysts, the phenomenon is reversible and sexual Hydra may be obtained from asexual and vice versa. These various factors combined have made the following study experimentally feasible. Since several years' work will be reviewed in the next half hour, permit me to use an analogy to illustrate some otherwise confusing relationships. The analogy concerns a man who wears a little woolly sweater. Inside his skin we know the temperature to be 98.6 F. while the temperature of the room is perhaps 50 °F. Now the question is: What is the temperature to which his skin is exposed? Clearly the sweater markedly affects the answer, so that the air in contact with his skin is nearer 98.6 the thicker, and more impermeable the sweater. How does this relate to Hydra? Figure 1 is a photograph of some Hydra in a Petri dish in which a pH sensitive dye (brom cresol purple) is present as well as 0.5% agar. This is a small amount of agar, enough to increase the vis- cosity of the culture solution^ without making it actually gel. Ob- serve that each Hydra is surrounded by a halo of its own making, an area of increased acidity due to the increased pCOo adjacent to its body surface. Each Hydra, in other words, is inside a little woolly sweater, where the partial pressure or pCO:.. of carbon diox- ide is neither as high as it is in his tissues proper, nor as low as it is in the general macroenvironment of the Petri dish. This "halo zone" corresponds then to the area inside the man's sweater. It is the zone of partial anaerobiosis where the pCOo and pNHs are in- creased and the pO^ and pH are decreased in a microenvironment that is chemically quite different from that of the macroenvironment of the Petri dish proper. Note that the halo zone around each individual Hydra varies with the size of the Hydra, so that larger and older Hydra are exposed to greater degrees of anaerobiosis than smaller and younger ones. In addition, group effects are present around Hydra that happen to lie close together so that their halo zones overlap and mutually reinforce each other. This group effect is clearly visible iBVC solution composed of 100 mg./l. NaHCOa, 50 mg./l. disodiimi ediylenedia- mine tetraacetate ('"Versene") and 100 mg./l. CaCL., dissolved in deionized water from a Barnstead Bantam Demineralizer equipped with a red-cap Mixed Resin cartridge. W. F. LOOMIS 339 in Figure 1. It corresponds in our temperature analogy to the warmth generated by a group of baby birds that huddle together in the nest so that they create a microenvironment far warmer than the surrounding air. Figure 2 represents Rachevsky's formulation of such a halo zone ( 25 ) . He postulated that if a spherical cell of radius r should give off any metabolite such as CO^ at a rate q, then the concentra- Fig. 1. Halo zones of partial anaerobiosis around single Hydra. These vary in size with the size of the Hydra as well as with the closeness of ad- jacent Hydra. See text for details. 340 THE BIOLOGY OF HYDRA : 1961 tioii of this metabolite at the center of the cell would be the sum of four factors. At the bottom would be the macroenvironmental background, which in the case of pCO^ is 0.03^i atmosphere (0.22 mm. Hg) in all samples of aerated water but 5.3% atm. in mammalian blood. H,pNH3, pC02,etc. Fig. 2. Rachevsky's graph of the four zones that together determine the final degree of anaerobiosis to which the DNA in the nucleus of a cell will be exposed. This same analysis holds for a multicellular mass of cells such as a slime mold pseudoplasmodium. Hydra, or developing frog egg. See Rachevsky (25) for mathematical equation that determines the profile of this graph Both of these backgrounds remain constant because the percentage COo in the air (0.03%) is extremely constant while the pCO^ of the blood is homeostatically regulated by the medullary center of the brain. Above the background zone in Rachevsky's graph is seen the halo zone referred to above. This is the external gradient that forms around any respiring cell under stagnant conditions. It reflects both stagnation and crowding for the group effects mentioned above also increase as population density increases. W. F. LOOMIS 341 The third addition represents the cell membrane barrier, an addition that is very small in the case of COo and NH3 as the lipid cell membrane is known to be highly permeable to both these dis- solved gases, (it is almost impermeable to the HCO.,^ and NH4+ ions that are fat insoluble ) ( 10,27 ) . For present purposes, this third or membrane effect may be neglected. The fourth and final addition represents the intracellular pC02 gradient that varies both with q, the respiratory rate, and r the radius of the cell. Since cell division mechanisms keep r reason- ably constant, we can experimentally control this fourth factor by controlling q with a thermostat, for it has been shown that the respiratory rate of Hydra varies logarithmically with the tempera- ture, as well as also varying somewhat with the level of nutri- tion (11). The main factors that control the pCO^ in the center of a cell according to Rachevsky are then: 1) the external macroenviron- mental background; 2) the "halo zone" microenvironment; 3) the barrier effect that is small if only a cell membrane is involved but can be very large if it involves an impermeable chitinous perisarc; and 4 ) the internal gradient. This then was the thinking behind the various experiments re- ported below, experiments in which we studied the effects of tem- perature, rate of feeding, population density, stagnation, degree of aeration etc. on the sexual maturation of Hydra. It was our assumption that DNA in the nucleus of the interstitial cells in the hypostome can produce RNA and specific proteins such that gon- adal tissues form whenever their "programming" is correct in respect to such feedback variables as pH, pOo, pNHo and pCOo etc. When- ever the programming is not of this variety, then these same inter- stitial cells differentiate into nematocysts due to the intrinsic pro- gramming that, in this case, takes place wholly within the tissues of Hydra. Only in the case of sexual differentiation does the external halo, group, and background effects determine the outcome of the experiment. Only this case, therefore, can be experimentally manip- ulated by varying the external cultural conditions. Let us examine the results of the experiment in Table 1 from this point of view. This experiment was originally performed in 1957 (16) but has been repeated eight times since then with entirely 342 THE BIOLOGY OF HYDRA : 1961 consistent results, an exceptional record it might be said in a field where over a score of operational factors have been shown to affect the results. In this experiment, ten male Hydra littoralis were grown in 15 ml. beakers in BVC solution^ that had been aerated with oxygen in duplicate vessels 1 and 2, while vessels 3 to 8 re- ceived increasing amounts of BVC solution that had been equili- brated with oxygen containing 10'/ CO. gas. In all cases the Hydra were fed daily with an excess of brine shrimp and then rinsed and placed in clean BVC solution half an hour later when they had fed to repletion. In addition, each vessel was rinsed a second time about five hours later to remove any excreted material present at that time, the pCO^ being readjusted each time the water was changed. TABLE 1 Control of sexual differentiation in Hydra by pCO^ (From ref. 16) Vessel 1 2 3 4 5 6 7 8 Culture water shaken with 100 per cent O2 (ml.) 15 14 10 Culture water shaken with 10 percent CO2 and 90 per cent O2 ( ml. ) 1 5 10 Initial pCOo 0.0% 0.6% 2.8% 5.6% Day Percentage \ of sexual forms 1 2 3 4 5 6 7 8 9 10 30 70 70 70 60 10 60 50 100 100 100 100 11 70 60 100 100 100 100 12 100 60 100 100 100 100 13 100 70 100 100 100 100 1 See p. 338. W. F. LOOMIS 343 All vessels were kept at 25 and all newly-detached buds were removed daily with a medicine dropper so as to maintain a con- stant population density within the culture. Population density, temperature, nutrition, stagnation, depth of water, surface/volume ratio, calcium, sodium and versene concentrations, sex and species were thus held constant. This experiment demonstrates that under these exact conditions pCOo is a controlling factor in the sexual differentiation of these animals. The unusually high degree of repeatability of this experi- ment makes it significant, therefore, that a totally different result occurred when this experiment was repeated on a shaking machine that shook similar but closed vessels for a few seconds every twenty minutes day and night (Fig. 3). Under these shaken con- ditions, the same experiment failed to yield any sexually differ- entiated Hydra. In retrospect, this inhibitory effect of shaking is due to the breaking up of the halo zone by mixing the micro- environment with the macroenvironment every twenty minutes around the clock. Fig. 3. Automatic shaker that is turned on for 5 seconds every twenty minutes to destroy the halo zone by mixing the microenvironment of the Hydra with the background macroenvironment. 344 THE BIOLOGY OF HYDRA : 1961 Since the pCO- of the macroenvironment had been artificially increased in this negative experiment, it was concluded that high pCO^ was not the sole factor needed to induce sexual differentiation in Hydra littoralis ( 19 ) . The nature of the postulated second factor is still unknown; it does not appear to be simply lowered pOo or pH, or simply increased pNHg either, or all four factors combined, at least in any combination yet tried. Perhaps a fifth feedback factor exists, or even a sixth, but certainly some combination of known circumstances should be able to be brought together in the macro- environment such that even shaken Hydra are exposed to condi- tions equivalent to that found within the halo zone of stagnant Hydra. Before proceeding further, it is perhaps instructive to mention that a powerful group effect exists within this 1957 experiment, i.e. no sexual forms appear if one, rather than ten, Hydra are placed in each vessel. Here is an example of the crowding-effect referred to above in which several halo zones overlap to mutually reinforce one another." In contrast, single Hydra mature sexually when 0.1% agar is added to the BVC solution in which they are grown as in Figure 1, the viscosity being thus raised sufficiently to stop all thermal cur- rents and hence allow extra large halo zones to form around even single Hydra. Perhaps it is for this very reason that Puck's sludgy- agar method enables single cells to grow in tissue culture when otherwise groups of fifty to one hundred cells are needed as inocula to obtain growth ( 24 ) . With the realization that feedback factors associated with halo zone anaerobiosis were active in this system, it became important to develop quantitative means of measuring them. Rapid micro- methods were consequently devised for pOo, pNHg and pCOo, a Beckman micro glass electrode (Beckman 290-31 or 290-80) being already available for determining the pH of unaerated 0.5 ml. samples of water. All four methods are carried out in hypodermic -Heisenberg's principle that the act of observing something alters the thing observed enters here, for high levels of pCOs were first tried on ten Hydra so as to be statistic- ally significant. Only later did it become clear that tlie ten Hydra affected each other in a positive group effect so as to turn sexual when a lone Hydra would not, even though he was exposed to as high pCOo as were the ten. \V. F. LOOMIS 345 syringes to protect the samples from equilibrating with the gases in the atmosphere. Furthermore, the tip of the needle of a syringe may be placed at the exact point from which it is desired to take the sample. Our method of determining pOo has been described in detail elsewhere (13, 15). Basically, it consists of drawing 0.5 ml. of leuco indigo cannine reagent into a tuberculin syringe followed by 0.5 ml. of the water sample to be tested. After mixing, the red color that develops is measured at 586 m/x by placing the intact syringe within the light path of a Beckman spectrophotometer, thus avoiding all contamination of the reagent with atmospheric oxygen. NH:i is determined in our laboratory by mixing 0.5 ml. of water sample with 0.5 ml. of Nessler solution that has been diluted one to live. The resulting color in the syringe is measured at 480 uifi by the method described above for oxygen. pCOo is measured directly by a method that has been published elsewhere ( 17 ) . As originally described, this method required modi- fication of a Henderson-Haldane apparatus, but this has since been found unnecessary, the standard apparatus (New York Laboratory Supply Co. 44250) being found sufficiently accurate for all prac- tical purposes. One analysis takes about three minutes. It consists of 1 ) filling a 20 ml. syringe with 10 ml. of water sample and 10 ml. of air; 2) shaking the half-filled syringe for thirty seconds so as to enrich the gas phase with the CO2 dissolved in the water phase; and 3) measuring the percentage CO2 in tlie aii* phase volumetric- ally in the Henderson-Haldane apparatus by measuring its percentage shrinkage after exposure to NaOH. Using these four methods, pH, p02, and pNHg, pCOo may be determined in any given culture in less than ten minutes. Three of the methods require only 0.5 ml. water samples while even the fourth (the pCOo analysis) may be scaled down to 0.5 ml. if a Scholander burette is used in place of a Henderson-Haldane appa- ratus.'^ Alternatively, halo zone water may be prepared in large amounts by growing many Hydra in a closed vessel that is placed on the shaking machine described above so that the micro and macroenvironments are mixed every twenty minutes. Sexual dif- ^Personal communication from Dr. Leonard Muscatine. 346 THE BIOLOGY OF HYDRA : 1961 ferentiation appears in such shaken cultures when the population density is around one Hydra per ml., all Hydra being fed and cleaned once per day. This is our present approach to this fascinating problem. When completed it should be possible to place Hydra in a running stream of chemically treated water and have them turn sexual even though all feedback between them and their culture water has been eliminated. Figure 4 shows our apparatus for conducting such an experiment. It was used to show definitively that increased levels of pC02 alone were not sufficient to induce sexual differentiation in Hydra littoralis, the conclusion being that other feedback factors Fig. 4. Set of six syphons that allow 1-5 Hydra (in small beakers at lower end of syphons) to be maintained in a flowing stream of chemically known water with all feedback effects removed. The rate of water flow is varied by the size of the hypodermic needle used as well as by the level of water within the large bottles. A liter of BVC culture solution is added to each bottle daily and the air space flushed out for five minutes with whatever COo-Oi-N . mixture one desires. W. F. LOOMIS 347 were also necessary (19). When these other factors have been identified, and their appropriate dosage determined, it should be possible to add the necessary components to the reservoirs of Figure 4 and have the constantly-washed Hydra in the syphoned- beakers differentiate sexually because they "think" they are crowded, i.e. their ectoderm being exposed to the same conditions found within the halo zones of a crowded culture. Since the present multi-factor approach has evolved gradually over several years, it may be worthwhile to review briefly the route by which this investigation has progressed since this provides a framework within which to discuss various important observations. 1) po, Looking back, even our earliest observations suggested that sexual difl^erentiation occurred under conditions of partial anaero- biosis (11). Thus, we found that 1) a score of Hydra turned sexual in a stagnant aquarium tank full of living Daphnia; 2) they reverted to the asexual state a few days after the aerator of the aquarium was turned on; 3) the shape of the container, and its surface/volume ratio, strongly influenced the reaction as seen in Table 2; 4) crowding Hydra almost automatically induced them to turn sexual in BVC while 5) stagnation constituted a reciprocal TABLE 2 Percentage of sexual forms and oxygen tension in cultures of differing surface/volume ratio. Percentage Oxygen of sexual Depth tension fomis ( mm. ) (mg./l.) after 10 days 30 7.3 100 10 8.4 100 5 8.6 48 2.5 8.7 Each culture consisted of 25 Hydra in 25 ml. BVC solution contained in a 50 ml beaker and three sizes of Petri dishes. 348 THE BIOLOGY OF HYDRA : 1961 variable in that stagnant-but-not-ciowded cultures would turn sexual just as would crowded-but-not-stagnant ones. Indeed this last obser- vation explained why cultures of Hydra placed in an ice-box for several weeks sometimes turned sexual, a method often advocated by earlier workers who believed that they were mimicking the nat- ural drop in temperature found in ponds in the fall of the year when Hydra often spontaneously turn sexual. Our observations suggested that it was the stagnation rather than the lowered temperature that induced sexuality, for we observed other experimental cultures turn sexual at 20°, 25° and 30° (14). Analysis of over thirty spontaneously sexual cultures showed that the pO^ was uniformly reduced from the 21% atm. of fully aerated water to about 15% atm. (70% saturation with air). When Hydra were grown in BVC solution whose pOo had been artificially lowered this amount (by aeration with a 15% O^— 85% N^. gas mix- ture), no sexual differentiation took place. Closer analysis (Table 2) revealed that reduced pO^ and sexual differentiation occurred simul- taneously but not proportionately and it was concluded that lowered pOo was not the sole inductive stimulus if indeed it was not merely an unimportant accompanying factor (16). 2) pCO, Shortly after finding that lowered oxygen tension could not substitute for partial anaerobiosis, we began to investigate the possibility that an increase in the partial pressure — or pCOo — of carbon dioxide gas dissolved in the culture solution might be the inductive stimulus. This possibility was difficult to investigate at first because no accurate means of measuring pCO^. existed. As with oxygen tension therefore, it was first necessaiy to develop an accu- rate and convenient determination, and as soon as this was available ( 17 ) it was found that spontaneously sexual cultures routinely showed an increase in pCO^ from the 0.03% atm. of fully aerated water to around 0.60% atm. Indeed pCOo levels as high as 1.2% atm. were found in crowded cultures exposed to 100% oxygen rather than air, and a record level of 1.43% atm. was found to occur naturally in the hypolinmion of a neighboring fresh water pond in April ( 23 ) . W. F. LOOMIS 349 The next step was to expose Hydra to BVC solutions whose PCO2 had been raised artificially. Table 1 records the result of this 1957 experiment, an experiment that has been found to be highly repeatable as described above. Taking both the group and stag- nation factors of this particular experiment into account, it is clear that pCOo strongly affects the reaction. Just as clear, however, is the fact that an increase in pCO^ is not sufficient, for Hydra main- tained in a flowing stream of BVC (Fig. 4) whose pCO^ varied from 0.03% to 10% failed to differentiate sexually. Similar exposure of Hydra to conditions of both high pCO^. and low pO^ failed to induce sexual differentiation, and it was concluded that a third factor must be operative in the system ( 19 ) . 3) pNH, Evidence that a third factor existed induced us to assay samples of "crowded water" for such metabolic gases as carbon monoxide, methane, ethylene, H^S, SOo etc. Analysis by infra-red spectography, mass spectography and gas-liquid partition chromatography failed to show such gases to be present, only CO^. and NH,-, being detect- able beyond the gases found in normal air. Analysis for NH3 by the 1 ml. syringe method showed that sexual cultures usually con- tained about 1 mg.T. NH... and that Hydra secreted large amounts of ammonia after being fed with brine shrimp. Since the toxic level of NH3 varied with the pH, it was concluded that the active species was the NH3 molecule rather than the NH4+ ion as only the fonner could penetrate the lipid cell membrane which is largely imper- meable to polar solutes such as NH4+ (27). Exposure of Hydra to increased levels of pNHo was accom- plished by adding different amounts of NH4OH to buffered culture solutions, and it was found that this variable alone and in various combinations with increased pC02 and decreased pO^ was unable to induce sexuality in Hydra at least under the conditions tried to date. One insight came from these experiments, however: it became clear that Hydra release the salt ammonium bicarbonate into their halo zone and that this buffer is equivalent to NaHCOg which, as we will see below, strongly affects the system. 350 THE BIOLOGY OF HYDRA : 1961 4) pH Generally speaking, Hydra differentiate sexually above pH 7, the optimum being about pH 8. Since a pCO^, of 0.5%— 1% atm. is also required, the original pH of the unused culture solution must either be about pH 9 in weakly buffered solutions such as BVC, or else about pH 8 when strongly buffered with sodium bicarbonate, tris (hydroxmethyl) aminomethane, or Versene, which is a buffer as well as a chelating agent since it is an organic amine. In addition, we have seen that Hydra produce their own buffer — NH4HCO;i — in sufficient amounts to be very important. For example, the water from a dense culture of Hydra may contain as much as 5 mg./l. NH,, (i.e. 10 mg./l. NH4OH). At pH 8, this would be almost entirely in the form of ammonium bicarbonate, this concentration of ammo- nia having served to neutralize COo that otherwise would have created a pCOo of 0.80% atm. Since this newly formed ammonium bicarbonate now serves as so much extra bicarbonate, it is clear that the liberation of NH-; during digestion affects the pH, the bicarbonate concentration and both the pNH:j and the pCO:-. Since all determinations of pCO.. from pH depend on knowing the bicar- bonate concentration ( Henderson-Hasselbalch equation), it follows that all such measurements are suspect in crowded cultures since these can spontaneously increase their bicarbonate concentration through this mechanism. The direct method of measuring pCOo described above, of course, is not subject to this error. The powerful effect of buffer concentration is seen in the fact that for an entire year we failed to produce any sexual Hydra when they were grown in 70 mg. T. CaCL.; 350 mg/1. NaCl; and 10 mg./l. NaHC03 (12). When the NaHCO., was increased tenfold to 100 mg. 1. (14), almost every culture in the laboratory turned sexual ( 21 ) . In this connection, it is interesting that Dr. Park never observed any sexual Hydra over a period of six years while using an unbuf- fered culture solution composed of 0.4 mg. 1. KCl; 10 mg./l. NaCl; and 4.8 mg. T. CaCL. We have confirmed her observations and also found that Hydra rapidly turn sexual when 100 mg./l. NaHCOa is added to her solution. I would now like to describe in some detail a convenient method of growing Hydra (and other hydroids) under constant conditions W. F. LOOMIS 351 of pH, pOo, pNH;5 and pCO^. In essence, the method consists of setting these variables in the water of uncrowded cultures twice a day. These twice-daily water changes are usually carried out thirty minutes and five hours after the cultures have been given their daily feeding of brine shrimp. In addition, the closed culture vessels are left on a shaking machine that shakes them strongly every twenty minutes (Fig. 3). Continuous mixing mechanisms, such as tissue culture roller tubes have been found to be either damaging to Hydra or not strong enough to break up the halo zone. Figure 5 illustrates the method used in setting the pCOo of a series of Hydra cultures. Between one and ten Hydra are placed in a 30 ml. Pyrex weighing bottle with an interchangeable ground glass stopper within which are placed 25 ml. of culture solution which leaves 5 ml. of air space. Into such vessels are injected 1-10 ml. BVC culture solution whose pCO^. has been set at 10% atm. by bubbling it for ten minutes with the gas from a Matheson tank of compressed air containing 10% COo. This bubbling is carried out about once a week, for the CO^-enriched BVC solution can be stored in 100 ml. syringes as shown in Figure 5 since CO^ does not escape from solutions stored in this fashion. Figure 5 also demonstrates the method of filling such 100 ml. syringes from the bottom of a 500 ml. graduate in which the bubbling is carried out. The daily routine, therefore, consists of filling a 30 ml. dispens- ing syringe from the 100 ml. syringe-reservoir and injecting 0, 1, 2, 3, 4 ml. etc. of this solution through a long needle into the bottom of half filled culture vessels and then bringing their total fluid content to 25 ml. When these vessels are shaken, the gas and water phases equilibrate and then remain constant. The actual level of pCOo in the various cultures is determined the next morning by the direct method described above. The great solubility of NH-.. makes pNH;, easy to adjust for all that is necessary is to add varying amounts of NH4OH to aliquots of the culture solution whose pH has been set with a bufl^er system. Culture solutions containing various concentrations of NH4OH are thus prepared before the start of an experiment and then stored in capped gallon jugs. Whatever buffer is desired is also included in such solutions, provision being made for the change of pH that 352 THE BIOLOGY OF HYDRA : 1961 comes from the later injection of culture solution that has been enriched with dissolved COo gas. The final step in the twice daily setting of pH, pOs, pNHg and pCO^ involves setting the oxygen tension. Since oxygen is only *>v s^^« Fig. 5. The pCOo of a culture is adjusted by injecting a calculated amount of culture solution that has previously been bubbled for ten minutes with gas from a tank of compressed air containing 10% CO.. (see tank behind technician). This COj-rich water is stored in the large 100 ml. syringes on the bench for later use. Such syringes are filled with a long glass tube from the bottom of the tall cylinder as illustrated. slightly soluble in water, pOo is set by adjusting the air phase and then allowing the shaking machine to keep the air and water phase in equilibrium. In practice, a 100 ml. syringe is filled with No from a tank and then partially emptied and refilled to 100 ml. witli room air, the syringe thus containing whatever dilution of air in nitrogen that one desires. A needle is attached to the syringe and the W. F. LOOMIS 353 tip of the needle slipped under the ground glass stopper of the cul- ture vessel. Since the air space within the vessel is only 5 ml., it is clear that emptying the 100 ml. syringe into this space flushes it with twenty times its own volume and so sets the air phase to whatever percentage of oxygen one desires. Summarizing then, buffer and ammonia concentration are arranged at the start of an experiment by preparing sufficient amounts of appropriate culture solution to last for the duration of the experiment. Then pCOo-enriched water is injected at each water change from a syringe and the air space blown out with No-diluted air. The vessels are then left on a shaking machine that agitates them every twenty minutes. Proof that pH, pO., pNH., and pCO. do not change is obtained by analysis. For this, the pH and pCOo are first determined and the results plotted on semi-log paper as in Figure 6. After this it is an easy matter to follow the cultures with a daily pH which yields their pCOo as taken from the logarithmic calibration curve. The above system has been gradually evolved over several years. With its aid, we can now analyze and control most of the seemingly- magic variables that affect this system. How can the mere change of one Pyrex vessel for another completely alter the results (Table 2)? A second rinse? Leaving the cultures over the weekend? Chang- ing the bicarbonate concentration? Aerating the cultures? These and many other operational variables are reflected in the changed values found in our four feedback variables. Some of them are listed in Table 3 where they are correlated with Rachevsky's four zones as well as with their equivalents in the analogy of the man in the little woolly sweater. Perhaps further work will show that sexual differentiation can be chemically induced by appropriate levels of pH, pOo, pNH^ and pCO^, i.e. that it is just a matter of finding correct dosage levels. Alternatively, it may be that further feedback xariables are involved that will need identification, analytic quantitation and artificial application before Hydra will differentiate sexually in a no-feedback system. To date we have run experiments that appar- ently eliminate: carbon monoxide, ethylene, carbonic anhydrase, biotin, folic acid, lactic acid, as well as the possibility that a diurnal 354 THE BIOLOGY OF HYDRA : 1961 pCOa in % Atm on Log Scale 5.00 4.00 3.00 2.00 1.00 .50 .40 .30 .20 .10 .05 7.0 v^ "^ \ X K -Solution A \^ ^V K \, ^ N ?fv Solut on B \ ■\ ^. ^ K 7.2 7.4 7.6 7.8 8.0 82 PH Fig. 6. Logarithmic calibration curves relating pH and pCO., in two solu- tions (25 ): (1) 2 10 • M NaHCO,; 5 10"^ M CaCL„"and (2) BVT solution prepared as described in Loomis and Lenhoff (21). Values of pC02 obtained by method of Loomis (17). cycle of alternating high and low levels of pCOo is required for differentiation to occur. Before concluding this presentation, I would like to broaden the discussion by suggesting that pH, pO., pNH,. and pCO. affect many biological systems other than Hydra. Some of these systems have been mentioned in previous publications (14, 16, 18), but preliminary experimental work in this laboratory suggests that the following phenomena are controlled by one or more of these four feedback variables : 1. Tentacle number and rate of bud growth in Hydra (20). W. F. LOOMIS 355 2. Inhibition of Tubiilaria regeneration: Even an amputated hydranth can inhibit regeneration in adjacent Tubularia stems (5, 7, 28 ) . This may result from the low intracellu- lar pH that results from high background pCOo produced by bacterial decomposition (and possibly intrinsic respi- ration) of the amputated hydranth. 3. The Fulton effect in Cordylophora: Fulton observed that Cordylophora "may be grown at- tached to microscope slides slanted in 100 ml. beakers. Such cultures may be grown to considerable density, TABLE 3 Rachevsky zone ( see Fig. 2 ) Analogous zone Factors affecting sexual differ- entiation in hydra Internal gradient Body temperature Factors that affect the metabolic rate q: temperature, nutrition, etc. see (11) Factors affecting the radius r: size of individual hydra, which varies with age, also species of hydra Barrier zone (Cell membrane, perisarc, etc. ) Permanent insulation such as fur Genetic differences between strains, species, and genera of hydra and hydroids. May vary in thickness of perisarc, etc. External gradient or Halo zone Variable insulation such as a woolly sweater Effect of stagnation shaking second water rinse crowding (popula- tion density) agar (viscosity) Fulton effect Background zone Room temperature Effect of aeration shape of vessel surface/volmne absolute depth other respiring Ufe degree of stagnation bicarbonate concen- tration Versene concentra- tion pH, pNHs, PO2 pCO. of culture water 356 THE BIOLOGY OF HYDRA : 1961 whereas cultures grown in the bottoms of dishes quickly become necrotic." (6). This striking "position effect" in CordylopJwra contrasts with Hydra which can grow equally well on the bottoms or sides of dishes. Preliminary results suggest that growth of Cordylophora is especially sensitive to the self-induced acidity present in the halo zone and that such zones are largely prevented from forming on slide-grown cultures by thennal currents ( these can be made visible with methylene blue or other dyes ) . Experimental elevation of the pCOa in shaken cul- tures gradually inhibits Cordylophora growth whenever it is sufficient to lower their pH below about 6.7, the actual pCO:- varying with the buffering capacity of the solution employed. 4. Strobilization in Cyanca arctica. This organism buds indefinitely when fed and placed in clean water daily. One culture strobilized and produced many medusae after being left untouched for a month in an ice-box at 12'. 5. Spiral persons in Hydractinia and Podocoryne. Braverman has shown that spiral zooids of Podocoryne appear on the rim of hermit crab shells only if the shell is inhabited by a living hermit crab (4). He noted that spiral zooids never form on colonies grown on glass slides in the laboratory, an observation that we have confirmed on Hydractinia. Spiral zooids appear all over slide- grown Hydractinia cultures exposed to a pCO^ of 2% atm.: a result that suggests that CO2 coming from the respiration of the hermit crab is the stimulus that creates spiral zooids on lips of hermit crab shells. 6. Parthenogenetic reproduction in Daphnia. Both Daphnia longispina and Daphnia magna fail to re- produce parthenogenetically in aerated water while doing so in water whose pCO:.. is 1% atm. and whose pOo is 5% atm. Daphnia are thus neither aerobic nor anaerobic organisms, but like microaerophilic ("little air") bacteria, they require partial anaerobiosis to live. This fact ex- W. F. LOOMIS 357 plains their usual habitat which is the partially anaerobic environment of the hypolimnton (2) as well as their demand for the microaerophilic environment of a soil- manure culture ( 1 ) . Amoeboid motion. Amoeba proteus and Chaos chaos are both far larger than the usual metazoan cell. Their central protoplasm would become extremely anaerobic if it did not liquify and then flow peripherally in long pseudopods with a high surface/ volume ratio. The possibility that amoeboid motion results from this automatic gel-to-sol transformation under the high pCOo (and consequent low pH) existing in the center of these animals is supported by the experimental finding that a pCOo of 20% atm. "melts" their pseudopods back into their bodies so that they become spherical in form. W