THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board JOHN M. ANDERSON, Cornell University DAVID W. BISHOP, Carnegie Institution of Washington JAMES CASE, State University of Iowa JOHN W. GOWEN, Iowa State College LIBBIE H. HYMAN, American Museum of Natural History J. LOGAN IRVIN, University of North Carolina DONALD P. COSTELLO, University of North Carolina Managing Editor JOHN H. LOCHHEAD, University of Vermont V. L. LOOSANOFF, U. S. Fish and Wildlife Service L. H. KLEINHOLZ, Reed College BERTA SCHARRER, Albert Einstein College of Medicine FRANZ SCHRADER, Duke University . RANDOLPH TAYLOR, University of Michigan VOLUME 122 FEBRUARY TO JUNE, 1962 Printed and Issued by LANCASTER PRESS, Inc. PRINCE 8C LEMON STS. LANCASTER, PA. 11 THE BIOLOGICAL BULLETIN is issued six times a year at the Lancaster Press, Inc., Prince and Lemon Streets, Lancaster, Penn- sylvania. Subscriptions and similar matter should be addressed to The Biological Bulletin, Marine Biological Laboratory, Woods Hole, Massachusetts. Agent for Great Britain: Wheldon and Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W. C. 2. Single numbers $2.50. Subscription per volume (three issues), $6.00. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Massachusetts, between June 1 and September 1, and to Dr. Donald P. Costello, P.O. Box 429, Chapel Hill, North Carolina, during the remainder of the year. Second-class postage paid at Lancaster, Pa. LANCASTER PRESS, INC., LANCASTER, PA. CONTENTS No. 1. FEBRUARY, 1962 PAGE BECK, STANLEY D. Photoperiodic induction of diapause in an insect . : 1 BONNER, JOHN TYLER, AND MARYA R. DODD Aggregation territories in the cellular slime molds 13 BOWEN, SARANE THOMPSON The genetics of Artemia salina. I. The reproductive cycle 25 DAVIS, HARRY C., AND ALAN D. ANSELL Survival and growth of larvae of the European oyster, O. edulis, at lowered salinities 33 GOODBODY, IVAN The biology of Ascidia nigra (Savigny). I. Survival and mortality in an adult population 40 GORDON, MALCOLM S., BEN H. AMDUR AND P. F. SCHOLANDER Freezing resistance in some northern fishes 52 JENNINGS, J. B. A histochemical study of digestion and digestive enzymes in the rhyn- chocoelan Lineus ruber (O. F. Miiller) 63 KLEINHOLZ, L. H., P. R. BURGESS, D. B. CARLISLE AND O. PFLUEGER NCurosecretion and crustacean retinal pigment hormone: distribution of the light-adapting hormone 73 LOOSANOFF, V. L. Gametogenesis and spawning of the European oyster, O. edulis, in waters of Maine 86 M CLEAN, JAMES H. Sublittoral ecology of kelp beds of the open coast area near Carmel, California 95 NASH, DONALD J., AND JOHN W. GOWEN Effects of x-irradiation upon postnatal growth in the mouse 115 STUNKARD, HORACE W. Taeniocotyle nom. nov. for Macraspis Olsson, 1869, preoccupied, and systematic position of the Aspidobothrea 137 WELLS, HARRY W., AND MARY JANE WELLS The polychaete Ceratonereis tridentata as a pest of the scallop Aequi- pecten gibbus 149 YONGE, C. M. On the biology of the mesogastropocl Trichotropis cancellata Hinds, a benthic indicator species 160 80230 iv CONTENTS No. 2. APRIL, 1962 BOOLOOTIAN, RICHARD A., A. FARMANFARMAIAN AND A. C. GIESE On the reproductive cycle and breeding habits of two western species of Haliotis 183 BRANHAM, JOSEPH M., AND CHARLES B. METZ Inhibition of fertilizin agglutination and fertilization in Arbacia by Fucus extracts 194 DEHNEL, PAUL A. Aspects of osmoregulation in two species of intertidal crabs 208 KAMEMOTO, F. I., ALICE E. SPALDING AND SHARON M. KEISTER Ionic balance in blood and coelomic fluid of earthworms 228 NASS, SYLVAN Localization and properties of phosphoprotein phosphatase in the frog egg and embryo 232 REDMOND, JAMES R. Oxygen-hemocyanin relationships in the land crab, Cardisoma guanhumi 252 RINGLE, DAVID A., AND PAUL R. GROSS Organization and composition of the amphibian yolk platelet. I. Investi- gations on the organization of the platelet 263 RINGLE, DAVID A., AND PAUL R. GROSS Organization and composition of the amphibian yolk platelet. II. In- vestigations on yolk proteins 281 SLATER, JOHN V., AND JOHN W. TREMOR Radioactive phosphorus accumulation and distribution in Tetrahymena . 298 STEINBACH, H. BURR Ionic and water balance of planarians 310 No. 3. JUNE, 1962 ANDERSON, JOHN MAXWELL Studies on visceral regeneration in sea-stars. I. Regeneration of pyloric caeca in Henricia leviuscula (Stimpson) 321 FORREST, HELEN Lack of dependence of the feeding reaction in Hydra on reduced gluta- thione 343 GAREY, WALTER F. Cardiac responses of fishes in asphyxic environments 362 1 1 1 NTER, W. RUSSELL, AND DAVID C. GRANT Mechanics of the ligament in the bivalve Spisula solidissima in relation to mode of life 369 JAXDKK, RUDOLF The swimming plane of the crustacean Mysidium gracile (Dana) 380 RULON, OLIN The extension of fertilizability and a hypothesis on sperm entrance in sand dollar ej^s. 391 CONTENTS v SCOTT, SISTER FLORENCE MARIE Tissue affinity in Amaroecium. II. Reaggregation of three partial zooids into functioning Siamese twins 396 SEGAL, EARL, AND PAUL A. DEHNEL Osmotic behavior in an intertidal limpet, Acmaea limatula 417 WALL, BETTY J., AND C. L. RALPH Responses of specific neurosecretory cells of the cockroach, Blaberus giganteus, to dehydration 431 Vol. 122, No. 1 February, 1962 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY PHOTOPERIODIC INDUCTION OF DIAPAUSE IN AN INSECT 1 STANLEY D. BECK Department of Entomology, University of Wisconsin, Madison 6, Wisconsin Diapause is defined as a state of arrested development in which the arrest is enforced by a physiological mechanism rather than by concurrently unfavorable environmental conditions. Although not maintained directly by environmental factors, diapause is apparently induced, and in many species also terminated, in response to environmental stimuli. Most insect species are capable of diapause at some stage in the life cycle ; embryonic, larval, pupal, and adult diapauses have been reported. Since the classical study of Kogure (1933) on the role of photo- period in the induction of embryonic diapause in the silkworm, Boinby.v mori L., daylength has been reported as a major inducing factor in the diapause of many other species. The recent profusion of reviews and symposia on the subjects of diapause and photoperiodism makes a detailed review of literature undesirable here ; see Lees (1955, 1959, 1960), Running (1960), and Harker (1960b, 1961). The present study was undertaken to demonstrate some of the characteristics of photoperiods effectively inducing diapause in the European corn borer, Ostrinia nubilalis (Hbn.). Larval diapause in this species has been shown to be a response to short days and low temperatures (Mutchmor and Beckel. 1959; Beck and Hamec, 1960). These earlier studies involved measurement of the effects of ecologically possible photoperiods, i.e., a 24-hour total cycle. Under such limited experimental conditions, it was not possible to determine the relative importance of the light and dark phases of the over-all photoperiod. because one phase could not be varied except at the expense of the other. Neither was it possible to determine whether or not the photoperiodic response involved an endogenous circadian rhythm within the organism, as postulated by Bunning (1960) or a non-circadian "interval timer," as postulated by Lees (1960a, 1960b). The experimental work reported below was designed to contribute to the clarification of these problems. MATERIALS AND METHODS The European corn borers used in this study were from a restricted natural population occurring near Madison, Wisconsin. The use of a defined population 1 Approved for publication by the Director of the Wisconsin Agricultural Experiment Station. This study was supported in part by a research grant (RG-7557) from the National Institutes of Health. 1 Copyright 1962, by the Marine Biological Laboratory STANLEY D. BECK was necessary because of the demonstration of significant differences in photo- periodic responses among different geographical populations of this species (Beck and Apple, 1961). Overwintering borers were collected from the field in the fall of the year, and were stored at 4 C. As needed, groups of the stored borers were placed at 30 C. for pupation and emergence. The progeny of these insects were tised in the experiments described below. 100 90 80 70 2 60 u o cr UJ a 50 z 040 a- 30 20 10 0% DIAPAUSE 68% DIAPAUSE #-*-* X K I I 10 15 20 AGE IN DAYS 25 FIGURE 1. Typical pupation records of experimental populations of the European corn borer, illustrating the method of determining incidence of diapause. The experimental borers were reared aseptically on purified diets, according to the rearing techniques described by Beck and Smissman (1960). Photoperiod experiments were run at 30 C., and the larvae were exposed to the experimental photoperiods continuously from an age of less than 24 hours. Each experiment involved the use of 250 newly hatched larvae. The larvae were divided into 5 groups of 50 each; one group was reared in continuous dark as the experimental control ; the other 4 groups were exposed to the experimental conditions and were treated statistically as replicates. Because of mortality and microbial contamina- tion, the final number of insects in each replicate and the control varied from 40 PHOTOPERIOD AND DIAPAUSE to 46. As the larvae reached maturity, daily pupation records were taken. Diapause incidence was measured as the per cent of mature larvae that failed to pupate. This was determined on the basis of the sigmoid configuration of the pupation record (Fig. 1); when the curve had remained unchanged for several days, pupation was considered to have been completed. The remaining mature larvae were considered to be in diapause. The experiments were carried out in B.O.D. constant temperature incubators that had been modified to incorporate a thermistor temperature control system (Thermistemp Temperature Controller Model 71, Yellow Springs Instrument Company, Yellow Springs, Ohio). Control of photoperiod was effected through the use of 7-day cycle programmers wired to two 14-watt fluorescent lights in- 100 r 90 80 2 70 LJ o 60 a ? 50 LJ tn a. a 30 40 20 10 5 10 15 30 24 PHOTOPHASE (hr light/24 hr) FIGURE 2. Effect of photophase duration on diapause incidence among European corn borer larvae reared under 24-hour photoperiods. stalled in the incubator. In experiments involving temperature changes, a 24-hour temperature cycle was effected by using a clock motor to drive the thermistor temperature control unit through a prescribed cycle. The performance of the cycling apparatus was verified by the use of a recording thermograph. In the discussion of the results, which follows, the term photoperiod is restricted to refer to the total cycle composed of a period of light and a period of dark. The period of light within the photoperiod is referred to as the photophase; conversely, the dark portion of the photoperiod is termed the scotophasc. This terminology is recommended as a means of avoiding the usual ambiguous use of the term photoperiod. It has been commonly used to refer to ( 1 ) the total light/dark cycle, and (2) only the light portion of the total cycle. Since the structure of the word implies a periodicity involving light, and since the only periodic alternative to light is dark, the term should be used only in the sense of the total light/dark cycle. STANLEY D. BECK RESULTS At an incubation temperature of 30 C. and a photoperiod of 24 hours, the effect of photophase duration on the incidence of diapause in the European corn borer is shown in Figure 2. The effective range of photophases was between 8 and 16 hours, with diapause incidence exceeding 90% only under photophases of from 10 to 14 hours. The borer is a "long-day" insect, in the classical sense of the term (Dickson, 1949; Lees, 1952; Otuka and Santa, 1955). It should be noted, however, that the term "long-day insect," in the sense of an insect that develops without interruption under the influence of 24-hour photoperiods containing a rela- tively long photophase (> 15 hr. ), is a misnomer. Abnormally short photophases ( < 9 hr.) also result in uninterrupted development in this and some other species. The diapause incidences obtained (Fig. 2) were in agreement with those reported TABLK I Effect of photophase : scotophase ratios on incidence of diapause among European corn borer larvae Photophase : Scotophase =1:1 Photophase: Scotophase =2:1 Photophase : Scotophase =1:2 Photo, (hr.) Scoto. (hr.) Total (hr.) Diapause (%) Photo, (hr.) Scoto. (hr.) Total (hr.) Diapause (%) Photo, (hr.) Scoto. (hr.) Total (hr.) Diapause (%) 6 6 12 <5 10 5 15 <5 4 8 12 <5 8 8 16 <5 12 6 18 <5 5 10 15 32 9 9 18 78 14 7 21 <5 6 12 18 95 10 10 20 92 16 8 24 <5 7 14 21 65 12 12 24 100 20 10 30 31 8 16 24 <5 14 14 28 85 24 12 36 65 16 16 32 50 28 14 42 <5 18 18 36 19 20 20 40 <5 in earlier studies of diapause in the European corn borer (Mutchmor and Beckel, 1959; Beck and Hanec, 1960), and are of interest to the present study only in that they delimit the responses of the insect to naturally occurring photoperiods. Diapause incidence under 24-hour photoperiods approached a maximum when the photophase : scotophase ratio was approximately 1:1, and approached a mini- mum at ratios of either 1:2 or 2:1. Diapause incidence under photoperiods dis- playing different photophase : scotophase ratios is shown in Table I. One-to-one ratios were ineffective when the phase durations were either under 8 or over 16 hours. Diapause incidence was very high when both photophase and scotophase were from 9 to 14 hours, which is similar to the responses obtained in the series of 24-hour photoperiods (Fig. 2). When the duration of the photophase was twice that of the scotophase (2:1 series, Table I), diapause was induced only when the scotophase was from 10 to 12 hours. When the scotophase was twice as long as the photophase (1:2 series, Table I), diapause was evoked only by scotophases of from 10 to 14 hours. Total photoperiod duration did not appear to influence the incidence of diapause in these experiments ; nor did the phase ratios appear to be of any significance. Diapause induction was closely associated with scoto- PHOTOPERIOD AND DIAPAUSE phases of from 10 to 14 hours, with maximum effectiveness at 12 hours. These findings are in agreement with those of Danilyevsky and Glinyanaya (1950), who worked with Acronycta spp. Otuka and Santa (1955), experimenting with Barathra brassicac L., concluded that, although the absolute length of the phases was of importance, phase ratio also influenced diapause induction ; their data on this point are not all convincing, however, as they tested only 1 : 1 ratios. A series of experiments was run in which photophases of 10, 12, and 14 hours were combined with a wide range of scotophases (Fig. 3). At these photophases, the range of diapause-inducing scotophases was relatively narrow ; scotophases of from 10 to 14 hours were required to produce a diapause incidence of 90 or more 100 _ 90 so LJ (J LJ a. Z 70 a <60 50 40 10 SCOTOPHASE (hr) 15 FIGURE 3. The effect of scotophase duration on the incidence of diapause in European corn borer larvae reared under photophases of 10, 12, or 14 hours. per cent. The three response curves shown in Figure 3 did not prove to be different to a statistically significant degree. Maximum diapause incidence was induced by a 12-hour scotophase in each case. Diapause incidence among borers reared under scotophases of 10, 12, and 14 hours, combined with a wide range of photophases, was also measured (Fig. 4). A 12-hour scotophase was significantly more effective than 10- or 14-hour scoto- phases. The 12-hour scotophase induced a high incidence of diapause when it was combined with any photophase of from 5 to 18 hours duration. A photophase of more than 28 hours was required to reduce the diapause incidence to below 50%, when the experimental photoperiod included a scotophase of 12 hours. Phase durations required to induce a diapause incidence of 90 or more per cent are shown in Figure 5. The 90% response curve describes an ellipse, within which STANLEY D. BECK 100 90 Z 80 LJ (J 70 a W 60 D a I" 40 30 12 16 20 PHOTOPHASE (hr) 24 28 32 FIGURE 4. Effect of photophase duration on the incidence of diapause among European corn borer larvae reared under scotophases of 10, 12, or 14 hours. diapause incidence was greater than 9Q c /c. The long axis of the ellipse lies along the 12-hour scotophase coordinate, and the widest part of the response zone falls between the 11- to 13-hour photophase and 10- to 14-hour scotophase coordinates. These results, combined with those shown in Figures 3 and 4, suggest that diapause is most efficiently induced by a 24-hour photoperiod containing a 12-hour scotophase. The characteristics of photoperiods inducing a high incidence of diapause in the 20 90 PER CENT DIAPAUSE O \ O O \ O ~ is LJ a O (O J L 10 15 PHOTOPHASE (hr) 20 FIGURE 5. Photoperiodic requirements for the induction of a 90 or more per cent incidence of diapause among European corn borer larvae. (Plotted points represent photoperiods tested experimentally.) PHOTOPERIOD AND DIAPAUSE 7 European corn borer cover a much wider range of phases than has been found in some other lepidopterous species (Dickson, 1949; Bull and Aclkisson, 1960). The response range of the mite, Metatetranychus uhni, is fundamentally different from those of the lepidopterous species tested, in that long scotophases (12 to 24 hours) tended to promote diapause incidence, and long photophases tended to suppress diapause. A 12-hour scotophase was found to be as effective as a 24- hour scotophase, indicating a very broad range of maximum effectiveness (Lees, 1953. 1955). PHOTOPERIOD DIAPAUSE >90 r 1 1 r j r ir~ r 4 I < 5 < 5 36 R'J I 88 BM 1 28 < 5 :'-'-i-J E- I I I I I I I I 4 8 12 16 20 24 HOURS FIGURE 6. Effect of one-hour light interruptions of the scotophase on diapause induction in the European corn borer (24-hour photoperiod). Photoperiodic induction of diapause in the European corn borer is dependent upon the actual duration of the photoperiodic phases, with the duration of the scotophase being far more critical than that of the photophase. This situation appears to prevail in other photoperiodically responsive arthropods as well. Lees (1953) found that a scotophase of 12 hours would induce the production of dia- pause-eggs in Metatetranychus uhni even when accompanied by photophases up to 36 hours long. Tanaka (1950), working with the Chinese Tussar-silkworm Antheraea pernyi, reported that a 13-hour scotophase was effective in inducing diapause under photoperiods containing photophases as long as 59 hours. All of these results indicate that, although minor species differences exist, diapause in STANLEY D. BECK "long-day" insects is a response to scotophases of about 12 hours duration, within a surprisingly broad range of photophases. A series of experiments was run to test the effect of interrupted scotophases on the incidence of diapause (Fig. 6). In the first group of such experiments, an hour of light was inserted in the middle of the 13-hour scotophase of a 24-hour photoperiod. Diapause incidence was negligible under such a photoperiodic regime, whereas an uninterrupted scotophase of 12 or 13 hours induced a diapause incidence of well over 90 f /c. Interruption of the scotophase by as little as 0.5 hour of light greatly reduced diapause incidence. The sensitivity of the borer to an interruption of the scotophase is greater than that reported for a number of other insect species, in which interruptions of from two to three hours were required to reduce the diapause-inducing effect of a 12-hour scotophase (Dickson, 1949; Lees, 1953; Danilyevsky and Glinyanaya, 1950; Bunning, 1960). Scotophases of 17 hours in a 24-hour photoperiod were systematically inter- rupted by a one-hour light period (Fig. 6). Uninterrupted scotophases of 16 and 17 hours did not induce a significant level of diapause. When the one-hour light break came at the end of 10, 12, or 14 hours of dark, a significant incidence of 31 - TEMP. ( C ) 26^ R=l r ' Z^3 ~ A 21 15 Photo. Dark Dark 15 Photo. 15 Photo. PHOTOPERIOD (Hrs) 9 Scot o. => bcoto. 9 Scoto. DIAPAUSE (%) 64 16 16 96 15 FIGURE 7. Effect of combined photoperiod and thermoperiod on the incidence of diapause in the European corn borer. diapause was induced. Once again, the highest incidence was associated with 12 hours of uninterrupted dark. When the interrupting light occurred after either 6 or 8 hours of dark, no diapause was observed. After either two or four hours of dark, an hour of light followed by 14 or 12 hours of dark resulted in a diapause incidence of about 50 %. Apparently, a 12-hour dark period is less effective if preceded by a four-hour dark period than if followed by a four-hour dark period. Using the cabbage worm, Pier is brassicae, Bunning and Joerrens (1960) also determined the effects of interrupted scotophases on diapause incidence. Although the experimental insect was also a "long-day" form, systematic interruption of the scotophase with one-hour light periods produced a different pattern of response. They found that diapause was prevented by light interruptions occurring about 15 hours after the beginning of the photophase, regardless of the total length of the scotophase or the position of the interruption within the scotophase. The results with the European corn borer (Fig. 6) show no such relationship, and diapause incidence is obviously dependent upon an uninterrupted dark period of about 12 hours, with the response being only slightly influenced by the position of the 12- hour dark period within the total photoperiod. The previous finding that diapause incidence under a given photoperiod is temperature-sensitive (Beck and Hanec, I960), and the finding that diapause incidence is closely dependent upon the duration of the scotophase, led to the PHOTOPERIOD AND DIAPAUSE 9 hypothesis that the temperature sensitivity of the diapause reaction is associated with the temperature during the scotophase. This was tested by the experimental series shown in Figure 7. Low temperatures during the 9-hour scotophase caused a very high incidence of diapause, whereas low temperatures during the photophase produced no more diapause than did the control conditions of continuous dark. The results clearly imply that diapause induction in the European corn borer involves a scotophasic temperature sensitivity. Because of the reported effects of light and photoperiodicity on growth rates and a variety of morphogenic processes (Muller, 1957, 1958, 1960; Ball, 1958) growth records were maintained in all experiments conducted in this study. In an earlier paper (Beck and Hanec, 1960) it was reported that the series of 24- hour photoperiods tested had no measurable effects on the rate of larval growth of the European corn borer. In the present study, a much wider range of photo- periods was studied, but no growth rate effects were detected. At the rearing temperature employed (30 C.,) the larvae attained the fifth instar at an age of about 10 days, prepupal stage at 11 to 13 days; pupation occurred from the thirteenth day, and pupation of the non-diapause portion of the experimental populations was 50% completed by about the seventeenth or eighteenth day. This developmental schedule did not appear to be materially altered by any photoperiod tested, except, of course, that the larvae entering diapause did not pupate. DISCUSSION From the results of this study, some of the characteristics of the photoperiodic reactions of the European corn borer may be deduced. The induction of diapause is obviously associated with the periodic occurrence of 12-hour scotophases during the 11 -day developmental period of the larva (at 30 C.). Within rather broad limits, the actual duration of the scotophase w r as found to determine the incidence of diapause, regardless of the relative proportion of the photoperiod occupied by the scotophase. Following exposure to a 12-hour scotophase, the borer is apparently refractory to a second dark "stimulus" for a period of from 4 to 5 hours. This interpretation is based on the results of two types of experiments. First, the shortest diapause- inducing photophase was found to be between 4 and 5 hours in duration (Fig. 4). This might be interpreted as meaning that the "light reaction" requires from 4 to 5 hours for effective activation, were it not that the results of the interrupted scotophase experiments are inconsistent with such a conclusion (Fig. 6). These experiments showed that light durations of an hour or less are effective. They also indicate the existence of a dark-refractory period following a 12-hour scoto- phase; the photoperiodic regime of 12 hours dark, 1 hour light, 4 hours dark, and 7 hours light resulted in a diapause incidence of 88%. Statistical analyses showed that such a response was not significantly different from the response to a simple photoperiod composed of a 12-hour scotophase and 8-hour photophase. It would seem that the borers failed to respond to the added four hours of dark. The effect of a photophase of several hours is apparently no more than the effect of a photophase of but one hour. Diapause incidences of about 50% were obtained when a 17-hour scotophase was divided into a short period (two- or four- hour), an hour of light, and a long dark period (14- or 12-hour) (Fig. 6). Since 10 STANLEY D. BECK a significant incidence of diapause was observed, the short dark period was not followed by a dark-refractory period. The long dark periods had a diapause- inducing effect, as expected, but were followed by a second effective period of light. The over-all effect of such a photoperiodic regime was that there were two effective periods of light (1 and 7 hours) for every effective scotophase, thus reducing the diapause-inducing effect to the photoperiod. Even with 12-hour scotophases, diapause incidence declined slowly as the length of the total photoperiod exceeded 30 hours (Fig. 4). This effect is prob- ably determined by the number of photoperiods occurring within the 11-day larval developmental period. A 12-hour scotophase combined with a 5-hour photophase produces a 17-hour photoperiod, and 15.5 such photoperiods can occur in 11 days (264 hours). However, when a 12-hour scotophase is combined with a 32-hour photophase, the photoperiod is 44 hours, and only 6.0 such photoperiods can occur in 1 1 days. The gradual decline of diapause incidence observed with photoperiods of increasing length is interpreted as being explicable on the basis that the incidence of diapause tends to be proportional to the number of 12-hour scotophases experi- enced during larval development. Assuming that the induction of diapause is dependent upon the occurrence of a required number of effective scotophases and photophases, it should be possible to devise a system of predicting the incidence of diapause under any given photo- periodic regime. Such a predictive scheme would be based on the accumulation of diapause-effective hours of light and dark over the period of larval development, in a manner comparable to the temperature accumulations that have widespread use in phenological predictions. A number of systems of photoperiodic accumula- tion were devised. Highly significant coefficients of correlation between per cent diapause and photoperiodic accumulations were obtained by nearly all of the methods tested. However, the schemes were devised to fit the data at hand, and in spite of such bias, serious discrepancies were apparent between "predicted" and observed results in a few crucial experiments. For these reasons, the ac- cumulation methods tried were concluded to be without real meaning. Until more is known about the dynamics of both the "light" and the "dark" reactions, an em- pirical system of predicting diapause incidence is not likely to be of much funda- mental significance. The over-riding importance of a scotophase of about 12 hours and an effective light flash of only one hour in the induction of diapause may be interpreted to support the conclusion of Lees (1960) that the physiological mechanisms involved behave as an interval timer. The diapause-inducing photoperiods tested in this study were not confined to a periodicity of 24 hours, or any multiple thereof, or to any other specific cycle duration. These findings lend little support to the hypothesis that diapause is in response to the effects of photoperiods on an endogenous circadian rhythm. On the other hand, Figures 3, 4, and 5, above, may be interpreted as lending at least feeble support to the hypothesis that photoperiodic induction of diapause involves a circadian function, in that the most effective range of photoperiods centered around a 12-hour photophase as well as a 12-hour scoto- phase. The tendency for diapause induction to be associated with a 24-hour periodicity was much less well defined than was its dependence on a 10- to 14-hour scotophase. PHOTOPERIOD AND DIAPAUSE 1 1 Diapause, itself, cannot be a rhythmic function : it occurs but once per indi- vidual. Whether or not one or more of the contributing physiological events involves a circadian rhythm has not been demonstrated. Feeding behavior activity cycles have been implicated in the induction of diapause in the beetle, Leptinotarsa deccmlincata (DeWilde et al., 1959), but the photoperiodic effects could not be fully explained solely on the basis of the effects of photoperiod on feeding. Modifi- cation of photoperiodic effects by experimental alteration of feeding behavior was also reported by Muller (1957), who worked with a number of species of the homopterous genus, Euscclis. Circadian rhythmicity and an influence of photo- period on such activities as locomotion and eclosion have also been reported (Harker, 1960a; Pittendrigh and Bruce, 1959). A relationship between circadian cycles, neurosecretory cycles, and diapause induction has thus far eluded demon- stration, however. Whether diapause is the result of photoperiodic induction of elaboration of a "diapause hormone" (Hasegawa, 1957; Lees, 1959b; DeWilde and Boer, 1961), or a photoperiodically induced biochemical failure in the morpho- genic chain of events remains to be determined. SUMMARY 1. The European corn borer, Ostrinia nubilalis, is a so-called long-day insect, larval diapause being induced by naturally occurring photoperiods con- taining scotophases of from 10 to 14 hours. 2. Diapause induction was found to be dependent upon the actual number of hours of the photoperiodic phases. The duration of the scotophase was far more critical than that of the photophase. A 12-hour scotophase was of maximum effectiveness when combined with photophases of from 5 to 18 hours. Sig- nificant incidence of diapause occurred when a 12-hour scotophase was combined with photophases of from 4.5 to 32 hours. 3. Diapause induction is a temperature-sensitive phenomenon, with the inci- dence of diapause tending to be inversely proportional to the ambient tempera- tures occurring during the scotophase. 4. Interruption of the scotophase by a one-hour period of light modified the photoperiodic response, the effect depending on the position of the light inter- ruption within the scotophase. The effects were interpreted as a demonstration that the insect's photophasic requirement is satisfied by a one-hour light period, but that longer photophases are normally required because of a dark-refractory period following 12-hour scotophases. LITERATURE CITED BALL, H. J., 1958. The effect of visible spectrum irradiation on growth and development in several species of insects. /. Econ. Ent., 51 : 573-578. BECK, S. D., AND J. W. APPLE, 1961. Effects of temperature and photoperiod on voltinism of geographical populations of the European corn borer, Pyransta nubilalis (Hbn.). /. Econ. Ent., 54: 550-558. BECK, S. D., AND W. HANEC, 1960. Diapause in the European corn borer, Pyrausta tiubilalis (Hbn.). /. Insect Physiol., 4: 304-318. BECK, S. D., AND E. E. SMISSMAN, 1960. The European corn borer, Pyraitsta nubilalis (Hbn.), and its principal host plant. VIII. Laboratory evaluation of host resistance to larval growth and survival. Ann. Ent. Soc. Amcr., 53 : 755-762. BULL, D. L., AND P. I. ADKISSON, 1960. Certain factors influencing diapause in the pink bolhvorm, Pcctinophora gossypiclla. J. Econ. Ent., 53: 793-798. 12 STANLEY D. BECK BUNNIXG, E., 1960. Circadian rhythms and the time measurement in photoperiodism. Cold Spring Harbor Symp. Quant. Biol., 25: 249-256. BUXXING, E., AND G. JOERRENS, 1960. Tagespcriodische antagonistische Schwankungen der Blauviolett- und Gelbrot- Empfindlichkeit als Grundlage der photoperiodischen Diapause-Induktion bei Pieris brassicae. Zeitschr. Naturforschung, 15B : 205-213. DAXILYEVSKY, A. S., AXD Y. I. GLIXAXAYA, 1950. On the influence of the rhythm of illumi- nation and temperature on the origin of diapause in insects. C. R. Acad. Sci. U. R. S. S., 71 : 963-966. DICKSOX, R. C., 1949. Factors governing the induction of diapause in the oriental fruit moth. Ann. Ent. Soc. Amcr., 42: 511-537. HARKER, J. E., 1960a. The effect of perturbations in the environmental cycle of the diurnal rhythm of activity of Periplaneta americana L. /. E.rp. Biol., 37: 154-163. HARKER, J. E., 1960b. Endocrine and nervous factors in insect circadian rhythms. Cold Spring Harbor Symp. Quant. Biol., 25: 279-286. HARKER, J. E., 1961. Diurnal rhythms. Ann. Rer. Ent., 6: 131-146. HASEGAWA, K., 1957. The diapause hormone of the silkworm, Boinby.r niori. Nature, 179: 1300-1301. KOGURE, M., 1933. The influence of light and temperature on certain characters of the silk- worm, Bomby.r niori. J . Dcpt. of Agric., Kyushu Univ., 4 : 1-93. LEES, A. D., 1952. The physiology of diapause in the fruit tree and spider mite. Trans. IXth Int. Congr. Eng., 1: 351-354. LEES, A. D., 1953. The significance of the light and dark phases in the photoperiodic control of diapause in I\Ietatetranychns ulini. Ann. Appl. Biol., 40: 487-497. LEES, A. D., 1955. The Physiology of Diapause in Arthropods. Cambridge University Press, Cambridge. LEES, A. D., 1959a. Photoperiodism in insects and mites. A. A. A. S. Public., 55: 585-600. LEES, A. D., 1959b. The role of photoperiod and temperature in the determination of parthenogenetic and sexual forms in the aphid Megoura viciac Buckton. I. The influence of these factors on apterous virginoparae and their progeny. /. Insect Physiol, 3: 92-117. LEES, A. D., 1960a. The role of photoperiod and temperature in the determination of partheno- genetic and sexual forms in the aphid Mcgonra viviac Buckton. II. The operation of the 'interval timer' in young clones. /. Insect. Physiol., 4: 154-175. LEES, A. D., 1960b. Some aspects of animal photoperiodism. Cold Spring Harbor Symp. Quant. Biol., 25: 261-268. Mn.LER, H. J., 1957. Die Wirkung exogener Faktoren auf die zyklische Formenbildung der Insekten, insbesondere der Gattung Euscclis (Horn., Auchenorrhyncha). Zool. Jahrb.. 85: 317-430. MULLER, H. J., 1958. 1'ber den Einfluss der Photoperiode auf Diapause und Korpergrosse der Delphacide Stcnocranus ininittns Fabr. ( Homoptera Auchenorrhyncha). Zool. Anzeig., 160: 294-311. MULLER, H. J., 1960. Die Bedeutung der Photoperiode im Lebensablauf der Insekten. Zeitschr. ang. Ent., 47: 7-24. MUTCHMOR, J. A., AND W. E. BECKEL, 1959. Some factors affecting diapause in the European corn borer, Ostrinia nubilalis (Hbn.). Canad. J. Zool., 37: 161-168. OTUKA, M., AND H. SANTA, 1955. Studies on the diapause in the cabbage armyworm, Barathra brassicae L. III. The effect of the rhythm of light and darkness on the induction of diapause. Bull. Natl. Inst. Agric. Sci. (Japan) Ser. C, 1955: 49-56. PITTENDRIGH, C. S., AND V. G. BRUCE, 1959. Daily rhythms as coupled oscillator systems and their relation to thermoperiodism and photoperiodism. A. A. A. S. Public., 55: 475-505. TAXAKA, Y., 1950. Studies on hibernation with special reference to photoperiodicity and breeding of the Chinese Tussar-silkworm. /. Seric. Sci. Japan, 19 : 358. DEWILDE, J., AND J. A. DEBoER, 1961. Physiology of diapause in the adult Colorado beetle. II. Diapause as a case of pseudo-allatectomy. /. Insect Physiol., 6: 152-161. DE\VILDE, J., C. S. DUIXTJER AXD L. MOOK, 1959. Physiology of diapause in the adult Colorado beetle (Leptinotarsa deccmlineata Say). I. The photoperiod as a con- trolling factor. /. Insect Physiol., 3: 75-85. AGGREGATION TERRITORIES IN THE CELLULAR SLIME MOLDS 1 JOHN TYLER BONNER AND MARYA R. DODD Department of Biolncjy, Princeton University, Princeton, N. J. The size of the area or territory which encompasses the amoebae that enter into an aggregate of a cellular slime mold is related to the process of initiation of aggregation as well as the size of the sorocarp. Each territory represents one aggregation and therefore one initiation event, and the size of any one fruiting body, or sorocarp, is primarily determined by the number of amoebae that enter an aggregation center. It is true that some of the cells may divide after the beginning of aggregation (Wilson, 1952; Bonner, 1960) but since all intake of food ceases some time before aggregation, there is no increase in bulk during the morphogenetic phases of the life-cycle, that is, while the aggre- gated cell mass progressively becomes transformed into a fruiting body. It will be shown that regardless of the density of the amoebae, the territory size, under controlled environmental conditions, remains constant for any one species. This means that sorocarp size is largely controlled by cell density and that the problem of initiation is identical with the problem of the establishment of these rigid aggregation territories. MATERIALS AND METHODS A simple method was used for the control of the amoeba density within a culture by controlling, in turn, the amount of bacterial food supply. Two per cent Bacto agar containing 6.2 /Ag./ml. of dihydrostreptomycin sulphate (Lilly) was poured into plastic Petri dishes which were marked on the bottom surface with squares of .102 cm 2 . A few loopfuls of Escherichia coll taken from a stock culture (grown on \% dextrose, \% peptone, 2% Bacto agar) were placed in 2 to 7 ml. of sterile distilled water. The density of the suspension was then de- termined in a Bausch & Lomb "Spectrometer 20" at 550 m//,. Appropriate dilutions were made to achieve a particular density and .09 ml. of this final suspension was evenly spread (with the help of an electric turntable and a sterile, bent glass rod) on the surface of one of the streptomycin agar Petri plates. In this way it was possible to obtain a range in food supply on the plates from approximately 500,000 bacterial cells/mm. 2 (= an optical density of 6.0) to 5000 bacterial cells/mm. 2 (= an optical density of 0.1). The spores of the slime mold were inoculated in three marked spots on each plate with a very fine, glass needle with a tip rounded in a small glass bead about .5 mm. in diameter, and consequently the inoculation points were confined to small, limited areas. Except for special 1 This stud}' was supported in part by a grant from the National Science Foundation and in part by the funds of the Eugene Higgins Trust allocated to Princeton University. The authors are indebted to the following individuals who read the manuscript and provided helpful suggestions : M. Krichevsky, K. B. Raper, B. M. Shaffer, B. E. Wright. 14 JOHN TYLER BONNER AND MARYA R. DODD experiments involving controlled temperatures and light, the dishes were incubated on a table in the center of the laboratory (25 3 C.). This was done after determining that the results were identical with those done at constant 24 C. conditions in continuous light, and in 12 hours of light alternating with 12 hours of darkness. Also in order to check the possibility that the streptomycin might be affecting the results in some way, some controls were run without the streptomycin, and no difference could be observed in the morphology or the size of the aggregation territories. The counts of fruiting body density were made by inverting the Petri dish under a dissecting microscope and counting the number of fruiting bodies in 5 separate .102-cm. 2 areas and averaging the results. The mean aggregation territory was calculated directly from the sorocarp density. The radius of the aggregation ter- ritory was determined by considering each territory to be a circle. The size of fruiting bodies was determined two ways : in the case of very small ones the number of spore and stalk cells w^as counted directly, and in larger ones a camera lucida drawing was made of the stalk so that stalk lengths could be accurately determined. RESULTS Sorocarp size According to the original description of the various cellular slime molds, each has its Characteristic size. These sizes are usually indicated as stalk lengths, and under normal culture conditions this is an appropriate measure, although it is pos- sible, by increasing the humidity of the air and decreasing the solute concentration of the agar, to obtain stalks of great length (Bonner and Shaw, 1957). For instance, in Dictyostelium mucoroides the largest are described as having stalks of 10 to 50 mm. or more (Raper, 1951) while under the specially humid conditions we obtained stalks up to 220 mm. in length. But this involved no increase in cell number ; it is merely that the humid conditions favor prolonged migration and the transfer of the majority of amoebae into stalk cells, leaving very few spores. We are not concerned here with this aspect of size, but rather with size changes which reflect changes in cell number (or in dry weight). As Raper (1951) has stressed, D. mucoroides may best be described as a complex, rather than a species, for the variability among strains is great, size being one of the significant variables. At one extreme there are the large forms typified by D. giganteum of Singh (1947) and at the other there is D. minntum of Raper (1941). The range in stalk height (under conditions which reflect cell number) is from .5 to 50 mm. In the case of D. discoideum all the naturally occurring strains described thus far are all roughly the same size (a stalk height of 1 to 3 mm.) although Sussman (1955) has produced a mutant ("fruity") by U.V. radiation that is considerably smaller. The genus Polysphondylium contains two species which differ significantly in size. The larger P. violaceum has a stalk length of about 5 to 30 mm. (Olive, 1902) while P. pallidum is roughly half the size. D. lacteum, which has round rather than elliptical spores, is also a small species (.5 to 1.5 mm.). Acytostelium leptosoinnin (Raper and Ouinlan, 1958), another AGGREGATION IN SLIME MOLDS 15 round-spored form, is small (1 to 1.8 mm.) but presents rather a special case in that the stalk is non-cellular and consists of a simple delicate cylinder of cellulose. The ultimate in small-size in the cellular slime molds is Protostelium (Olive and Stoianovitch, 1960) which consists of one cell which builds its own stalk (.07 mm. in height). In all these cases the stalk length reflects size under optimum culture conditions. In the case of the larger forms this is a rich medium which produces a profusion of bacterial food supply. It is a well known and every-day observation that if a medium is used which supports poor bacterial growth, these large species will produce smaller sorocarps. Arnclt ( 1 937 ) . in fact, made a series of observations 10- 10- < CO O .01 I . * " P. pallidum D. purpureum 0. discoideum | I I I I I 1 I I | I | I I I I I I 1 1 | 1 I I I I 1 I I I 1 I I 1 1 I 1.0 20 I l.O 5.0 .1 .5 1.0 LOG MEAN STALK HEIGHT IN MM. FIGURE 1. A graph showing the density of the bacterial food supply plotted against the mean stalk height which is used as an index of sorocarp size. Each point is an average of 20 stalks. of cultures with different densities of bacterial food supply, and came to the conclusion that the size of the fruiting body was directly dependent upon amoeba density. As will be seen presently, this study is partly a quantitative exploitation of this commonplace observation. Another frequent observation, repeatedly emphasized by Raper (e.g. 1951). is that the smaller species will not grow in rich media, but require dilute media. This is true for all the small forms mentioned above (D. minutinn, P. pallidum, D. lacteum, A. leptosomum). In other words, for these species, the maximum size seems to be determined by the density of the bacterial food supply ; there is an upper limit in the amount of food, above which all development is inhibited, pos- sibly because of the production of inhibitory substances by the abundant bacteria. With this in mind, different species were grown on different known concentra- 16 IOHN TYLER BONNER AND MARYA R. DODD tions of E. coll on streptomycin agar. As can be seen from Figure 1, the size of the sorocarps increases with the increase in bacteria. The range indicated here only shows the lower limit of the species tested, but fails to show the upper limit. Note that for the three species shown in Figure 1, each has a different minimum threshold. Some studies were also made of D. ininutum which, as just mentioned, has a low maximum threshold (i.e., does not develop with a rich food supply, prob- ably because excessive bacterial growth produces inhibitory substances). It was a surprise to discover that it had a high minimum threshold necessary for aggrega- tion. In other words it is a form which can only develop in a very restricted range of food densities. Other forms, such as D. purpurcum, are much less particular and I0u D B FIGURE 2. Camera lucida drawings of small sorocarps. A. D. pnrpitrcniti with 7 stalk cells and 45 spores (not shown). Note the wisp at the end of the stalk. B. P. pallidum with 5 stalk cells and 6 spores. C. The smallest sorocarp obtained. It is P. pallidum with ^ stalk cells and 4 spores. D. D. lactcum showing a small stalk which midway becomes acellular. can adapt to a wide range of concentrations, and can produce, as a result, a wide range of size in their fruiting bodies from ones much smaller than D. mhnitum to ones very much larger. This leads to an important point as far as D. minutum is concerned : one of the reasons for its small size is the fact that it normally develops in a restricted range of food density and because of the low value of this particular range, the mean sorocarp size is correspondingly small. There are undoubtedly other factors which also ma}' limit size in D. ininutum, a matter which is under further investigation. In general, on any one culture dish, at any one bacterial density, the sorocarps were remarkably uniform in size. At the minimum thresholds for each species it was of interest to examine the morphology of the smallest sorocarps for any AGGREGATION IN SLIME MOLDS 17 possible effects caused by the size reduction. In some species the stalk cells ap- peared somewhat coarse and club-like (D. mucoroidcs, D. purpurewn, D. discoi- deum) while in others the cells were beautifully tapered (P. pallidwn, P. violaceum, D. lactewn} (Fig. 2). In the case of D. lactewn the smaller sorocarps had an acellular tip, giving them, at least at their anterior end, an appearance almost exactly resembling Acytostelium. This is not true of the larger sorocarps of D. 1000 100 o Q. o z 10 I I 10 too NO. OF STALK CELLS 1000 FIGURE 3. The log of the number of spores is plotted against the log of the number of stalk cells. These points are a composite of D. purpureum, D. mucoroides, and P. pallidwn, as they showed no discernible difference among them. lactewn. Occasionally in D. pur pur cum it was possible to observe sharp acellular wisps of stalk material at the anterior end of the stalk. The smallest fruiting body encountered in these studies of all the species examined was a seven-cell sorocarp of D. purpureum (three stalk cells, four spores). Since it was possible to count the cells in small fruiting bodies, a study was made of the proportions of stalk to spore cells, this method having many advantages over the less direct ones previously devised (Bonner and Slifkin, 1949; Bonner, 1957). A number of counts were made for D. mucoroides, D. purpureum and P. pallidum (these latter ones were unbranched because of their small size, but as they showed no significant difference among them, they have all been plotted 18 JOHN TYLER BONNER AND MARY A R. DODD together (Fig. 3). The character of the allometric relation is identical to the one previously described for larger pseudoplasmoclia (Bonner, 1957). Aggregation territories While size and food supply (which is a reflection of amoeba density) are directly proportionate, fruiting body density is independent of size. That is, regardless of the density of amoebae, the size of an aggregation territory remains constant (Fig. 4). It is of interest to note that Arndt (1937) came to a similar conclusion, al- though he made no quantitative determinations to support his observations. To emphasize the point, it is space, not the number of cells, that is important in de- termining the size of an aggregation territory. If there are 100 cells in a unit '^ *v 5 5.0 -\ ft a o O 1.0- .5 t * P pallidum 0, purpureum p discoidtfutr .8 1.3 .4 8 1.3 .4 MEAN RADIUS OF AGGREGATION TERRITORY IN MM. 1.2 FIGURE 4. Five graphs for different species, showing the radius of the aggregation territory under conditions of different bacterial food densities. The top points labeled cone. are on 2% agar without streptomycin and a heavy layer of bacterial paste smeared evenly over the surface. Each point is an average of 5 squares (.102 cm. 2 ) on one culture dish. space, they will produce one fruiting body of 100 cells ; if there are 1000 cells then they will produce one fruiting body of 1000 cells. Of course this statement is oversimplified as it ignores the possibility of cell division and the fact that not all the cells may enter the aggregate but some are left behind (and in fact their number can be accurately determined ) . Some experiments were also run at higher concentrations of bacteria to de- termine if the constancy of the size of the aggregation territory was affected by very high amoeba densities. This was first done on non-nutrient plates covered with a heavy layer of bacterial paste for all five species, and as can be seen in Figure 4 results compare with the more dilute plates. A further experiment was run on D. purpureum using full nutrient agar (10 gm. peptone, 10 gm. dextrose, .96 gm. Na 2 HPO 4 -12HoO, 1.45 gm. KH 2 PO 4 , 1000 ml. H 2 O, 20 gm. agar) and the same diluted in half and by one fourth (with the exception of the agar content AGGREGATION IN SLIME MOLDS 19 which remained at 2%). The results showed no obvious difference between the three concentrations of nutrients, and the range of twelve runs extended from a territory radius of .41 to .56 mm. which overlaps, but is significantly lower than the range using non-nutrient agar. This discrepancy is unexplained although there are a number of possibilities, such as the specific effects of the nutrients themselves on territory size, a possibility which will come up again in discussing the environ- mental factors which affect the aggregation territory. As a check to determine if the same result could be obtained by using radically different techniques,, two further experiments should be mentioned. In one the amoebae of D. discoideum were centrifuged free of bacteria and suspended in a salt solution (see Bonner, 1947) in van Tieghem cells with one end sealed with a coverslip. A microscope slide was then sealed to the other end of the cell, and the whole preparation inverted after the amoebae had settled on the coverslip. They were thus attached to glass and retained in small moist chambers which were incubated at 20 C. in the dark (except for hourly examinations under the micro- TABLE I An experiment on territory size using D. discoideum on coverslips in small moist chambers (van Tieghem cells) Amoeba density: amoebae/mm 2 . Mean sorocarp height in mm. Total number of sorocarps per coverslip Mean radius of the aggre- gation territory in mm. 246 .76 26 1.32 318 .77 23 1.41 618 1.08 22 1.44 1040 1.05 28 1.28 2570 .86 23 1.41 scope ) . At very high and very low amoeba concentrations the fruiting bodies were few and small, but in a large range of intermediate concentrations the territory size remained constant, while the mean stalk length tended to increase, as would be expected from previous results (Table I). In the other experiment the liquid culture technique of Gerisch (1960) was employed and after the growth phase the amoebae of D. purpureum were harvested, plated out in different concentrations on the streptomycin agar plates and incubated in room conditions. This method has the advantage of making it possible to de- termine amoeba densities directly. If Figure 5 and Figure 4 are examined, it is obvious that the ranges of the territory size using the two techniques are comparable. In order to examine the formation of aggregation territories in more detail, time lapse motion pictures were taken of the aggregation of D. purpureum at four different bacterial food densities on streptomycin agar. It was possible, in each of the cases, to determine the amoeba densities which were 54, 147, 174, and 356 amoebae/mm. 2 , respectively (included in Figure 5). The only significant difference that could be observed among the four cases was that the resulting fruiting bodies were progressively larger as the amoeba density increases. As far as the time of formation of the aggregates and the general blocking out of the territories, they were all similar. For instance in none of the cases was there any 20 JOHN TYLER BONNER AND MARYA R. DODD o OF TERRITORY IN MU. FIGURE 5. A graph showing the radius of the aggregation territory at different known amoeba densities for D. purpurcum. The solid dots involve the Gerisch (1960) liquid culture technique and the hollow circles were taken from a motion picture of cultures prepared by the streptomycin agar technique described in Materials and Methods. evidence of the formation of more numerous territories, some of which would then disappear. Each center, once formed, remained stable and there was no subsequent disintegration. If the rate of appearance of new centers is plotted against time for each of the amoeba densities, it is clear that they do not appear steadily, but with considerable irregularity. However, under these conditions of constant illumination there was no evidence of bursts of aggregation such as Shaffer (1958) obtained with alternate darkness and light. As is evident from Figure 4, each species has a characteristic territory size. They are listed in order of magnitude for the various species in Table II. A few determinations were also made of D. minutum and D. lactcum, which indicated that they are both on the lower end of the scale. This means that to a minor extent the small size of these two species and P. pallidum might be accounted for by their small territory size. The fact that environmental conditions affect the density and size of fruiting bodies has been appreciated for a long time. Potts (1902) and Harper (1932) showed that light produced more numerous and smaller fruiting bodies. This was confirmed by Raper (1940), who also showed that a slight decrease in humidity TABLK 1 1 The average radius of the aggregation territories of different species Species No. of cases Mean radius in mm. D. dixcoidriini 25 1.27 D. purpui'i'iun 36 .73 P. viol a ecu in 39 .63 D. mucoroidcs 24 .53 P. pallidum 31 .49 AGGREGATION IN SLIME MOLDS 21 would elicit the same effect. Bradley, Sussman and Ennis (1956) studied the influence of various chemical agents upon aggregation and they found that histidine in suitable concentrations produced more numerous smaller fruiting bodies, while adenine had the reverse effect. The effect of histidine has been confirmed by recent studies of Krichevsky and Wright (personal communication). Using the test system described in this paper and D. purpiireum as the test organism, Heller and Miles (1961 ) have shown that light is exceedingly effective and that as the intensity increases the aggregation territory becomes correspondingly reduced. They also showed that humidity, in the dark, had a very small effect as compared to the light effect, and that while the aggregation territory became reduced in size with decreasing humidity it soon reached a minimum (very roughly in the neighborhood of 98% relative humidity) and then increased rapidly as the relative humidity of the surrounding atmosphere approached 95%. In a parallel study Opderbeck (1961 ) examined the effect of histidine and also successfully repeated the results of the previous workers in the dark, but in the light he found that the histidine had the reverse effect ; that is, it produced, in all the concentrations tried (from 10' 1 to 10~ 3 M'), an increase in the size of the aggregation territories as compared to the controls. It is hoped that ultimately further study of these effects (which is in progress) may shed some light on the factors which are responsible for the delineation of the aggregation territory. DISCUSSION At the moment the mechanism of territory formation is unknown. It is, of course, possible to suggest many hypotheses, two of which will be mentioned here. ( 1 ) Any cell could become so altered in its state that it would produce an inhibiting substance which diffuses outward and prevents any other cell, within a given radius, from achieving the same state. Clearly the distance the substance diffuses would be independent of the number of cells within a territory. One must also presume that the cell producing the inhibitor lies at the center of the future aggregation pattern, similar to Shaffer's (1961 ) "founder cell" in P. violaceum. (2) The above hypothesis assumes two separate functions for the keystone cell: inhibition and the subsequent initiation of acrasin production. It might be possible to explain the whole phenomenon solely on the basis of acrasin diffusion. If we assume that the original puffs of acrasin are not carried outward from cell to cell by the Shaffer (1957) relay system, but diffuse from one cell or a small group of cells, and if we assume that a certain concentration of acrasin prevents other cells from becoming acrasin-emitting pace-makers, then this original diffusion gradient of acrasin can be unaffected by the number of cells in a territory and therefore delineate the aggregation territory. Unfortunately, these hypotheses and any other we might invent have far too many assumptions and are in urgent need of testing by experiment. In relating this work to those of others it first should be mentioned that although Sussman and Noel (1952) and Sussman and Sussman (1961) have made a study of the relation of amoeba density to the number of fruiting bodies, they did not measure either territory size (fruiting body density) or sorocarp size, and there- fore it is impossible to compare their work with the present study. 22 JOHN TYLER BONNER AND MARYA R. DODD On the other hand, the aggregation territories are obviously related to the problem of the initiation of aggregation and it is pertinent to examine the "initiator cell" hypothesis of Sussman and his group (e.g. Sussman and Ennis, 1959; Ennis and Sussman, 1958). This hypothesis assumes that special "initiator cells" are in a fixed proportion to the total cell number : in D. discoid cum one cell in every 2200 is presumed to be an "initiator cell," while in D. purpurcwn it is one cell in every 300. We have made some cell counts for these two species under threshold conditions where the amoeba density is just sufficient to produce aggregation and fruiting. If the total number of cells per territory is determined (i.e., the number of cells for the sorocarp as well as the number of cells that failed to enter the aggregate) they average 1032 for D. discoidcum and 150 for D. purpurcum. In other words there are roughly twice as many centers as there are "initiator cells." However, there are so many other ways of showing that aggregation occurs in small populations of cells below the number predicted from the "initiator cell" hy- pothesis that the hypothesis may no longer be considered tenable (Bonner, 1960; Konijn and Raper, 1961 ; Gerisch, 1961). But perhaps the far more important point is that contrary to the "initiator cell" hypothesis, aggregation is not determined by a cell which holds a strict propor- tion to the other cells in a population ; it is completely independent of the other cells (provided a sufficient cell density is maintained). The only factor which clearly and absolutely controls the initiation process is space : the aggregation territory is, for each species under given environmental conditions, a fixed entity. These conclusions are entirely in keeping with those of Shaffer (1961), who has shown that existing aggregations in P. riolaccum are capable of inhibiting the further production of founder cells, even in populations of cells that are not entering streams. According to Samuel (1961) the earliest manifestation of the establish- ment of the aggregation territory (i.e., initiation) is a regional depression in the rate of cell movement. This is followed, as Shaffer (1961 ) has shown, by the appearance of a cloud, an area of relatively dense amoebae. In D. mucoroides and D. purpurcum a true aggregation center, with the eventual appearance of incoming streams of amoebae, is only evident after these two initial stages. To completely understand the factors which control initiation and the distribution of the aggre- gation territory, it will be necessary to provide an explanation for all the events that lead up to the aggregation process itself. One final point that may have some bearing on future experiments : in previous studies (Bonner, 1960) it was shown that if a small group of cells is isolated by scraping away the cells all around, they often dispersed or disintegrated after aggregation, and in D. piirpurciim the aggregates showed a tendency to produce abnormal sorocarps. In the present study sorocarps of comparable size never disintegrated or showed signs of abnormality. The reasons for this difference are not known but they raise the interesting possibility that if an aggregation territory is isolated and not surrounded by other aggregation patterns, it lacks certain peripheral chemical influences, which results in disintegration or abnormality. The question really is whether or not the cells in neighboring territories remain in communication one with another during aggregation and the later states of morpho- genesis. AGGREGATION IN SLIME MOLDS SUMMARY AND CONCLUSIONS 1. The area of the aggregation territory in the cellular slime molds is constant at different cell densities and therefore the number of amoebae that aggregate in any one territory varies directly with the cell density. As a result sorocarp size in the cellular slime molds is a function of the density of the amoebae prior to aggregation. 2. The mechanism whereby the territory size is determined is not known, although clearly the problem of the initiation of aggregation is related to the estab- lishing of fixed territories. Since their establishment is independent of cell number we may propose the hypothesis that initiation is determined solely by space or distance. 3. There are a number of conditions which frame these general conclusions. The territory size is characteristic for each species and is constant only under a particular set of environmental conditions. Also the relation obviously only ap- plies when the amoeba density is sufficient for aggregation, and each species has a specific range of densities which permit aggregation and fruiting. LITERATURE CITED ARNDT, A., 1937. Untersuchungen iiber Dictyostelium mncoroidcs Brefeld. Arch. f. Entiv., 136: 681-747. BONNER, J. T., 1947. Evidence for the formation of cell aggregates by chemotaxis in the development of the slime mold Dictyostelium discoideum. J. Exp. Zool., 106: 1-26. BONNER, J. T., 1957. A theory of the control of differentiation in the cellular slime molds. Quart. Rev. Biol, 32: 232-246. BONNER, J. T., 1960. Development in the cellular slime molds : the role of cell division, cell size and cell number. 18th Growth Symposium. (Developing Cell Systems and Their Control, Ed. by D. Rudnick.) Ronald Press, N. Y., pp. 3-20. BONNER, J. T., AND M. J. SHAW, 1957. The role of humidity in the differentiation of the cellular slime molds. /. Cell. Comp. Physiol., 50: 145-154. BONNER, J. T., AND M. K. SLIFKIN, 1949. A study of the control of differentiation : the proportions of stalk and spore cells in the slime mold Dictyostelium discoideum. Amer. J. Bot., 36: 727-734. BRADLEY, S. G., M. SUSSMAN AND H. L. ENNIS, 1956. Environmental factors affecting the aggregation of the cellular slime mold, Dictyostelium discoideum. J. Protozool., 3 : 33-38. ENNIS, H. L., AND M. SUSSMAN, 1958. The initiator cell for slime mold aggregation. Proc. Nat. Acad. Sci., 44: 401-411. GERISCH, G., 1960. Zellfunktionen und Zellfunktionswechsel in der Entwicklung von Dicty- ostelium discoideum I. Arch. f. Entw., 152 : 632-654. GERISCH, G., 1961. II. Develop. Biol.. 3 : 685-724. HARPER, R. A., 1932. Organization and light relations in Polysphondylium. Bull. Torre\ Bot. Club, 59: 49-84. HELLER, S. A., AND M. C. MILES, 1961. The effect of humidity and light on sorocarp density in Dictyostelium purpureiim. Senior thesis, Princeton University. KONIJN, T. M., AND K. B. RAPER, 1961. Cell aggregation in Dictyostelium discoideum. Develop. Biol.. 3: 725-756. OLIVE, E. W., 1902. Monograph of the Acrasieae. Proc. Bost. Soc. Nat. Hist., 30: 451-510. OLIVE, L. S., AND C. STOIANOVITCH, 1960. Two new members of the Acrasiales. Bull. Torrcv Bot. Club, 87 : 1-20. OPDERBECK, C. T., 1961. The effect of histidinc on aggregation in Dictyostelium purpureiim. Senior thesis, Princeton University. POTTS, G., 1902. Zur Physiologic des Dictyostelium mucoroidcs. Flora, 91 : 281-347. 24 JOHN TYLER BONNER AND MARYA R. DODD RAPER, K. B., 1940. Pseudoplasmodium formation and organization in Dictyostelium discoidewn. J. Elisha Mitchell Sci. Soc., 56 : 241-282. RAPER, K. B., 1941. Dictyostelium minntum, a second new species of slime mold from decaying forest leaves. Mycologia, 33: 633-649. RAPER, K. B., 1951. Isolation, cultivation, and conservation of simple slime molds. Quart. Rev. Bio!,, 26: 169-190. RAPER, K. B., AND M. S. QUINLAN, 1958. Actyostelium leptosoimtm : A unique cellular slime mold with an acellular stalk. /. Gen. MicrobioL, 18: 16-32. SAMUEL, E. W., 1961. Orientation and rate of locomotion of individual amoebae in the life cycle of the cellular slime mold Dictyostelium mucoroides. Develop. Biol., 3: 317-335. SHAFFER, B. M., 1957. Aspects of aggregation in cellular slime molds. I. Orientation and chemotaxis. Amcr. Nat.. 91 : 19-35. SHAFFER, B. M., 1958. Integration in aggregating cellular slime moulds. Quart. J. Micr. Sci., 99: 103-121. SHAFFER, B. M., 1961. The Acrasina. Adv. in Morphogen., 2 (in press). SINGH, B. N., 1947. Studies on soil Acrasieae: I. Distribution of species of Dictyostelium in soils of Great Britain and the effect of bacteria on their development. /. Gen. Micro- bioL, 1: 11-21. SUSSMAN, M., 1955. "Fruity" and other mutants of the cellular slime mold, Dictyostelium discoidcuin: a study of developmental aberrations. /. Gen. MicrobioL, 13: 295-309. SUSSMAN, M., AND H. L. ENNIS, 1959. The role of the initiator cell in slime mold aggregation. Biol. Bull., 116: 304-317. SUSSMAN, M., AND E. NoiiL, 1952. An analysis of the aggregation stage in the development of the slime molds, Dictyosteliaceae. I. The population distribution of the capacity to initiate aggregation. Biol. Bull., 103: 259-268. SUSSMAN, M., AND R. R. SUSSMAN, 1961. Aggregative performance. E.vp. Cell Res., SuppL, 8: 91-106. WILSON, C. M., 1952. Sexuality in the Acrasiales. Proc. Nat. Acad. Sci., 38: 659-662. THE GENETICS OF ARTEMIA SALINA. I. THE REPRODUCTIVE CYCLE 1 SARANE THOMPSON BOWEN Department of Biology, San Francisco State College, San Francisco 27, California The brine shrimp Artemia salina is a branchiopod crustacean found in saline lakes and the evaporating ponds of commercial salt works. It is of interest to geneticists because amphigonic races have been reported to be diploid (2n = 42) or tetraploid. Parthenogenetic races have been found to be diploid, triploid, tetra- ploid, pentaploid, and octaploid. Barigozzi (1957) has reviewed the cytological studies of these races. No previous attempt has been made to analyze traits which are governed by a single locus. Artemia is easily cultured in the laboratory because it is resistant to environ- mental stresses. In a study of a California salt pond containing brine shrimp, Carpelan (1957) found diurnal changes of 12 C. in the water temperature in August. Provasoli and Shiraishi (1959) raised nauplii to adulthood in a sterile medium. Lochhead (1941) reported that females reproduce viviparously or ovi- parously. The thick-shelled egg contains a blastula and withstands desiccation for as long as 15 years. Therefore, mutant stocks might be conveniently stored in the form of thick-shelled eggs (cysts) without need of repeated subculture. Because the shrimp is transparent, the effect of genes upon cellular differentia- tion may be studied throughout the development of one individual. Weisz (1946) has pointed out the advantages of studying morphogenesis in this primitive crus- tacean which has nineteen body segments but few specialized structures to obscure the principles of development. He has stated (in 1947, p. 87) that the "... his- tological sequences are found to be governed by a continuous overall pattern of metameric development, precisely defined in relative time and in space. ..." In 1959, the author began a study of Artemia in the hope of developing a method for raising shrimp through many generations in pedigreed cultures. This paper describes a successful culture method and a series of experiments which test for sperm storage by the female and reproduction by parthenogenesis, paedogenesis, and pseudogamy. MATERIALS AND METHODS The cysts of the California race were collected from salt works on San Fran- cisco Bay ; those of the Utah race were from Great Salt Lake. The dried cysts are routinely stored in glass bottles and hatched in sea water. In 24 to 36 hours, the shells burst and each embryo emerges enclosed within a transparent membrane. In another eight hours, the embryo hatches out as a free-swimming nauplius. The 1 This research was supported by a grant from the National Science Foundation (NSF G-13219). 26 SARANE THOMPSON BOWEN nauplii are transferred directly to the culture medium made by adding 50 grams of Nad to one liter of filtered sea water. All cultures are maintained in 5 cc. of culture medium in shell vials 21 mm. in diameter and 70 mm. high. A standard yeast suspension (SYS) is made by mixing 1 cc. of dry Brewer's yeast with 9 cc. of medium. It is dispensed into the shell vials by means of a pipette once per week according to this standard feeding schedule: First day. Nauplii are separated from their parents (in the case of labora- tory stocks) or from the shells (if the cysts have been collected from salt ponds). One to three nauplii are put in each vial, 0.05 cc. of SYS is added, and the vial is tightly corked. Eighth day. Again, 0.05 cc. of SYS is added to each vial, irrespective of the number of surviving metanauplii. Fifteenth day. Five-one hundredths cc. of SYS is added for every shrimp present in the vial. Many will have reached sexual maturity. Males are easily identified by their larger antennae. Twenty-second day. Five-one hundredths cc. of SYS is added for every shrimp in the vial. Mated pairs may have produced a brood of nauplii. Twenty-ninth day, etc. The weekly feedings of 0.05 cc. of SYS per adult are continued. The stocks are maintained at room temperature (20-28 C.) in a room illuminated during the day by artificial light. No attempt is made to aerate the medium nor to remove waste materials. When raised by this method, more than half of the nauplii born to laboratory stocks reach sexual maturity and a few shrimp have reached the age of nine months. Mated females give birth to free- swimming nauplii ; virgin females release transparent thin-shelled eggs which sink to the bottom of the vial and do not hatch. Opaque thick-shelled cysts are not produced unless the culture method is modified. RESULTS AND DISCUSSION 1. Effect of food quantity upon "viability and fertility The standard feeding schedule described above was adopted because shrimp can reach maturity if the standard amount of food is either doubled or cut in half ; i.e., it allows margin for measurement errors. It also results in optimum viability, as the following experiment demonstrates. One first-instar California nauplius was placed in each of 276 vials containing 5 cc. of culture medium. All nauplii received the standard amount of SYS until the fifteenth day when the 198 survivors were divided into three equal groups of 66. Each group then was put on a dif- ferent regimen: 0.025 cc., 0.05 cc., or 0.10 cc. of SYS per shrimp each week. The data in Table I indicate that viability was highest in those shrimp receiving the standard quantity (0.05 cc.) of SYS. Females were examined by transmitted light under a binocular microscope to see if opaque yolk granules were in the eggs. Because vitellogenesis was first seen in females receiving 0.1 cc. of SYS per week (Table II), we may conclude that females fed twice the standard amount mature faster than those on the standard schedule. REPRODUCTION IN ARTEMIA 27 TABLE I Effect of quantity of food upon viability Amount of SYS each week Number of living shrimp Age (days) 15 22 29 36 57 .025 cc. 66 41 30 24 14 .05 cc. 66 44 38 34 28 .10 cc. 66 , m. 44 38 35 20 2. Relation of age and fertility When different males were mated successively to a fertile female, none was found to be sterile. But many females consistently produced either small broods of nauplii or broods of thin-shelled eggs which did not hatch. About one-half of both the wild and the mutant females had high fertility records like those shown in Table III. Note that the number of nauplii per brood is not correlated with the age of the female. The brood size may be influenced by uncontrolled factors such as the type of bacterial flora in the vial or the time of feeding in relation to the reproductive cycle. In another experiment, four pairs of shrimp, all of which were over five months of age, produced broods of normal size (39-76 nauplii) and the nauplii had normal viability. Fertility evidently does not decline throughout the first five months of life. 3. Tests for parthenogenesis and f>acdogcnesis To test for parthenogenesis, 20 immature females of the Utah race and 80 of the California race were isolated, one in each vial. They did not produce nauplii but laid broods of transparent eggs every four or five days. A control group was kept for the same period of one month in the presence of males. All of the 20 Utah controls and 64 of the 80 California controls gave birth to nauplii. More than 200 females of each race have been isolated and parthenogenesis has never been observed. Nauplii are born three to five days after the females have TABLE II Effect of quantity of food upon fertility Amount of SYS each week Females showing vitellogenesis /total females Age (days) 22 29 36 57 0.025 cc. 0/20 0/15 0/10 3/5 0.05 cc. 1/22 6/18 11/16 11/12 0.10 cc. 10/22 17/20 16/18 11/12 28 SARANE THOMPSON BOWEN mated and at no other time. All attempts to hatch the eggs of virgin shrimp have failed. These findings are in agreement with those of Lochhead (1941), who found that fertilization was essential for reproduction in the California race. However, both Jensen (1918) and Relyea (1937) reported that the Utah race could reproduce parthenogenetically. These two authors did not provide suffi- ciently detailed accounts of their experiments to permit an attempt to replicate them in this present study. In order to test for paedogenesis, California nauplii were allowed to grow to adulthood in the presence of their mother. In the fourteen cultures observed, the male nauplii were unable to fertilize their mother or their female sibs until they reached the tenth instar of Heath (1924) when their antennae took on the adult shape, enabling them to clasp the femal^l In 24 cultures, California fe- TABLK III Relation of age and fertility in three Utah females Number of nauplii in brood T- 1 Age 4-7 weeks Age 8-12 weeks A Q 19 16 31 70 38 .,0 B 17 45 15 56 20 19 36 41 70 C 69 3 42 58 71 92 2 males raised with their fathers also failed to reproduce until they reached adult- hood and vitellogenesis was present. Hundreds of nauplii from both races have been paired during routine maintenance of stock cultures, with no evidence of paedogenesis. 4. Description of the gene for red eye The first red-eyed shrimp was found by Miss Jean Hanson in the fall of 1960, in the progeny of a brother-sister mating in the Utah stock. The data in Table IV indicate that the gene for red eye, r, is a recessive and has complete penetrance in the homozygote. Because the reciprocal crosses of the type RR X rr yield no red progeny, we may conclude that the gene is not sex-linked. Only 138/682. or 20%, of the F 2 shrimp had red eyes. The deviation from the expected 25 % is highly significant (P < 0.01) and suggests that the red-eyed shrimp may have a lower viability than that of the black-eyed shrimp. The data from the ten backcross (Rr X rr) matings also suggest a lower viability of the red-eye pheno- type although the deviation from the expected 1 : 1 ratio is not significant (P = 0.50-0.30). Because the backcross data favor a single factor hypothesis, we may conclude that the red eye characteristic is governed by one locus. The three phenotypes may be described in this manner : RR (black). The'ocellus is pale red in the first instar of Heath (1924), black in subsequent instars. The two lateral eyes are black from the time they are first pigmented in the third instar and remain black throughout the life of all wild type shrimp. REPRODUCTION IN ARTEM1A Rr (black). The ocellus is pale red in the first instar, black in subsequent instars. The eyes are black throughout the life of most Rr shrimp. In rare instances, the eyes turn a deep ruby for a few days in the second week of life but revert to black. rr (red). The eyes and ocellus are pale red for the first ten days; the pigment is so sparsely distributed throughout the first five instars that the three pigmented areas cannot be seen when the metanauplius is examined under the binocular microscope (7x). At the end of the second week, the eyes and ocellus are bright red. The eyes darken progressively after the shrimp has reached sexual maturity ; by the twenty-second day, the eyes are dark ruby or black. The ocellus remains red for a longer period but may also turn ruby or btjfcpk. 5. Test for pseudogamy Pseudogamy is defined as the development of an egg parthenogenetically after the initial stimulus of penetration by a sperm. (The sperm nucleus then de- generates and has no effect on the genotype of the offspring. ) Because the 39 TABLE IV Segregation of the gene r in the Utah race Progeny Number of matings Mating Black Red Total 56 rr X rr 1793 1793 in Rr X rr 161 149 310 32 Rr X Rr 544 138 682 39 tfRRX 9rr 720 720 12 &rr X 1RR 236 236 matings of the type $RR X $rr (listed in Table IV) produced only black-eyed offspring, we may conclude that pseudogamy is not the normal form of reproduc- tion in Utah shrimp reared under standard laboratory conditions. 6. The sequence of events in the reproductive cycle The reproductive system of the female consists of two ovaries, two pouch-like oviducts, and a ventral median uterus. The following events normally take place in a 24- to 48-hour period. The female expels from the uterus the first egg gen- eration (brood A) as either virgin eggs or nauplii. The birth process takes from two to ten hours. She molts in a few seconds and then the next egg generation (brood B) passes from the ovaries into the oviducts in less than two hours. They remain there from one to 40 hours, whether copulation occurs or not. They then pass into the uterus, the process taking less than 30 minutes. The eggs remain in the uterus for three to five days, irrespective of whether or not they are fertilized. The cycle is normally completed in from four to six days. However, in three exceptional females, the eggs lodged in the oviducts for ten days and the cycle was prolonged. 30 SARANE THOMPSON BOWEN Lochhead (1941) correctly stated this sequence of events but did not publish the evidence for his conclusion that copulation occurred when the eggs were in the oviducts. Therefore, nine rr females were successively mated to males of rr, RR, and rr genotypes but the RR male was present only at the time when the eggs were seen to be in the oviducts. The RR males often failed to clasp during this short period and the females then laid eggs. Nauplii were obtained from three females ; observations on one of them are in Table V '. In all three cases, the RR male was present only during the time when the eggs were in the oviducts yet all the progeny were of the Rr genotype. Fautrez-Firlefyn (1957) and Goldschmidt (1952) have reported that eggs in the oviducts are in metaphase of the first meiotic division. TABLE Y Observations on the reproductive cycle of one rr female Duration of period Events 4 days The first male (rr) clasps the female. Egg generation A undergoes segmentation in the uterus. Egg generation B becomes visible in the ovaries due to accumulation of opaque yolk. 10 hours 70 nauplii (brood A) are expelled from the uterus. The rr male continues to clasp and attempts unsuccessfully to copulate. The female molts. The clasping pair is transferred to a slide and the male is pulled away. The female is returned to the vial. 55 minutes Egg generation B passes into the oviducts. The second male (RR) is added. He clasps and copulates. Afterward, one seminal vesicle is transparent ; the other is opaque due to the presence of sperm. The clasping pair is transferred to a slide and the male pulled away. The female is returned to the vial. Egg generation B passes into the uterus. 4 days The third male (rr) is added. He clasps the female within twenty minutes after the eggs have entered the uterus. Four days later, 115 nauplii (brood B) are born. 58 nauplii of brood A survive to an age when they can be classified. All have red eyes. The 98 survivors of brood B have black eves. 7. Studies of the female reproductive cycle with tests for sperm storage Observations on more than 200 females of each race indicated that if they mated once, they produced a single brood of nauplii and thereafter laid thin-shelled eggs which did not hatch. This suggests that Artemia females do not store sperm as do Drosophlla females. However, this evidence is not conclusive because the acts of clasping or copulation might in some way be essential for egg maturation. (For example, copulation might be the stimulus needed to bring about the reflex secretion by the oviduct of a substance which would cause the eggs to complete the first meiotic division.) This possibility is remote but, if true, it would in- validate the previously described tests for parthenogenesis as well as those for sperm storage. Therefore, the following experiments were designed to rule out this REPRODUCTION IN ARTEMIA 31 FIRST MALE (RR) SECONDMALE THIRD MALE (rr) (RR) Days 10 20 v^~ 3 Y v III 1 III II 1 1 1 1 1 till 1 1 1 i 1 i 1 1 i 1 1 1 1 t t t t \ t t t t Rr R r Rr Rr R r r r r r r r Rr (14) (37) (9) (49) (1 7) (17) (33) (33) (65) FIGURE 1. Genotypes of nine successive broods of nauplii produced by one rr Utah female mated to three different males. The numbers in parentheses indicate the number of progeny in each brood which survived to an age when their eye color could be classified. possibility and to test for parthenogenesis and for sperm storage in a female which was at all times in the presence of a male. RR males were hatched from cysts collected from Great Salt Lake; rr shrimp were selected from the Utah laboratory cultures. Twenty rr females were studied ; each was alternately mated to males of RR and rr genotypes. The record of one of these females is seen in Figure 1. On the twenty-fifth day, a brood of Rr nauplii was produced as the result of a mating with an RR male. On the twenty- sixth day, the first male was removed, a second male with rr genotype was added, the eggs moved into the oviducts, and copulation occurred. These observations on the transparent female are confirmed by the fact that the nauplii born on the thirty-first day had the rr genotype. On the thirty-sixth day a brood of rr nauplii was born, the next generation of eggs passed into the oviducts, and the female again mated with the second male. On the thirty-seventh day, a third male was added, he attempted to copulate, but was unable to affect the genotype of the next brood because the eggs were now in the uterus. In Figure 1 , two changes in brood paternity may be seen : one between the twenty-fifth and thirty-first days and another between the forty-second and forty- sixth days. Another six changes in brood paternity are summarized in Table VI ; in each case the two broods (A and B) are less than six days apart. Note that in the second (B) brood, all the nauplii have the same genotype because sperm are not stored by the female from one cycle to the next. Corroborative data were obtained from the other females, but in each case a brood of virgin eggs TABLE VI Comparison of pairs of broods of different paternity born to six rr females Brood Number of nauplii classified and their genotype Female A B 73 Rr 56 rr 30 rr 46 Rr 31 Rr 41 rr 58 Rr 33 rr 17 rr 26 Rr 22 rr 12 Rr 32 SARANE THOMPSON BOWEN separated the two broods of different paternity because the second male failed to clasp in the short period when the eggs were in the oviducts. The author wishes to express her gratitude to three students who subcultured the stocks used in this study : Carol Cleminshaw, Jean Hanson, and John Parker. Thanks are due to Dr. John S. Hensill, who suggested the use of Artemia in genetic experiments, and made many valuable suggestions during this investigation. SUMMARY 1. This paper reports the first analysis of an inherited trait governed by one locus in the brine shrimp, Artemia salina. The autosomal gene, r, for red eyes arose spontaneously in a Utah race. It is recessive to the wild type allele, R, for black eyes. It has complete penetrance in rr shrimp. 2. The standard culture method outlined here has successfully carried the mutant stock through ten generations in a one-year period. 3. Reproduction was studied in two races from California and Utah. Neither paedogenesis nor parthenogenesis was observed in these shrimp which were raised by the standard culture method. This observation conflicts with the reports of Jensen and of Relyea that the Utah race could reproduce parthenogenetically. 4. Matings of RR males to rr females produce only black-eyed progeny. This indicates that when raised by the standard method the Utah shrimp do not normally reproduce by pseudogamy. 5. Studies of the sequence of steps in the female reproductive cycle confirm the observations of Lochhead. Genetic experiments have demonstrated that al- though the adults may clasp continuously throughout the cycle, copulation is ef- fective only when the eggs are in the oviducts. 6. Females do not store sperm from one reproductive cycle to the next. If an rr female is alternately mated in different cycles to males of RR and rr genotype, all the nauplii in one brood have the same genotype. LITERATURE CITED BARIGOZZI, C, 1957, Differenciation des genotypes et distribution geographique d' Artemia salina Leach: donnees et problemes. Ann. Biol., 33: 241-250. CARPELAN. L. H., 1957. Hydrobiology of the Alviso salt ponds. Ecology, 38: 375-390. FAUTREZ-FIRLEFYN, N., 1957. Protcines lipides et glucides dans 1'oeuf d'Artemia salina. Arch. Biol., 68: 249-296. GOLDSCHMIDT, E., 1952. Fluctuation in chromosome number in Artemia salina. J. Morph., 91: 111-133. HEATH, H., 1924. The external development of certain phyllopods. /. Morph., 38: 453-483. JENSEN, A. C., 1918. Some observations on Artemia gracilis, the brine shrimp of Great Salt Lake. Biol. Bull, 34: 18-32. LOCHHEAD, J. H., 1941. Artemia, the "brine shrimp." Turtox News, 19: 41-45. LOCHHEAD, J. H., 1950. Artemia. In: Selected Invertebrate Types (pp. 394-399). Edited by F. A. Brown, Jr. John Wiley and Sons, Inc., New York. PROVASOLI, L., AND K. SHIRAISHI, 1959. Axenic cultivation of the brine shrimp Artemia salina. Biol. Bull., 117: 347-355. RELYEA, G. M., 1937. The brine shrimp of Great Salt Lake. Amer. Nat., 71 : 612-616. WEISZ, P. B., 1946. The space-time pattern of segment formation in Artemia salina. Biol. Bull., 91: 119-140. WEISZ, P. B., 1947. The histological pattern of metameric development in Artemia salina. J. Morph., 81 : 45-95. SURVIVAL AND GROWTH OF LARVAE OF THE EUROPEAN OYSTER, O. EDULIS, AT LOWERED SALINITIES HARRY C. DAVIS AND ALAN D. ANSELL ' U. S. Intrcau nf Commercial Fisheries, Biological Laboratory, Milford. Connecticut, and University of Southampton, England The European oyster. Ostrca edulis, in its native habitat, is found primarily in oceanic or nearly oceanic salinities. Its native range extends from the southern coast of England and the Scandinavian countries to the Mediterranean, but re- cently it has been successfully introduced into New England waters (Loosanoff, 1951, 1952, 1955; Welch, in press). Initially, Loosanoff (1955) proposed this oyster for introduction into those areas where the summer water temperatures rarely, if ever, are high enough for American oysters to reproduce. More recently, because of heavy mortalities of American oysters in several states, he has suggested that O. edulis might be introduced as a second commercial oyster into these and other oyster-producing areas of the United States. Since European and American oysters belong to different genera, they will not interbreed and, what could be extremely important, European oysters might not be susceptible to some of the diseases now affecting American oysters. Moreover, since European oysters are larviparous and their larvae are usually 175 /u, to 185 /JL at the time of release, the food requirements of the larvae are not as re- stricted as those of the young larvae of American oysters. Therefore, good sets of European oysters might frequently be obtained in seasons when setting of American oysters fails. Korringa (1941) reviewed the literature on the effects of salinity on eggs and larvae of several species of oysters, including 0. edulis. He found no correla- tion between rate of growth of larvae or intensity of their setting and differences in average salinities in the Oosterschelde, Holland, ranging from 25 to 35 ppt. He quotes the assumption of Gaarder (1932, 1933) and Gaarder and Bjerkan (1934) that 24 ppt is the lowest salinity for satisfactory growth of larvae of 0. edit/is. Korringa also states that changes of salinity, within the range found in the Oosterschelde, cannot be held responsible for the success or failure of spatfall of O. edulis in this area, and points out that experiments in vitro had not yet been carried out to determine the effect of salinity upon setting. Walne (1956) reported experiments on rearing 0. edit! is larvae in salinities initially adjusted to 21.1, 25.9, 27.9 and 31.3 ppt. The water in the larval cultures was not changed during the course of his experiments and, apparently due to evaporation and the addition of algal food suspension, the salinity in all cultures increased. Thus, in cultures initially at 21.1 ppt the salinity increased to 25.9 and 26.2 ppt. Under these conditions he found that the larvae survived and grew 1 Present address : University of Southampton, Plankton Laboratory, c/o Poole Generating Station, Rigler Road, Poole, Dorset, England. 33 34 HARRY C. DAVIS AND ALAN D. ANSELL throughout the salinity range tested, but he did not obtain any spatfall in the cultures started at 21.1 ppt. The present study was undertaken to obtain more precise data on the effect of low salinities on egg production and incubation in O. edulis, upon survival and growth of larvae, and upon setting of mature larvae. Since several of the oyster- producing areas in the United States, where we might wish to introduce 0. edulis, are characterized by relatively low salinities, such information is necessary. METHODS The methods employed were essentially the same as those used by Davis (1958) in a similar study of the salinity tolerance of eggs and larvae of the American oyster. However, since O. edulis is larviparous, to obtain larvae from these oysters it was necessary to hold adults in aquaria during the period of gonad development, spawning and incubation. Approximately ten adult oysters were kept in each aquarium in 60 liters of water. One group of oysters was kept at our normal salinity (27 ppt) ; another group was kept at a salinity of 20 ppt; and a third group was kept at 17.5 ppt. We used sea water, filtered through an Orion filter designed to remove particles above 15 /A in diameter, and lowered the salinity to the desired level with demineralized tap water. Water in these aquaria was changed daily and three to four liters of algae were added during each change to supplement the food present in the water. The adult oysters kept at normal salinity provided the larvae used in these experiments. As soon as larvae were noticed after release, they were collected by draining the water from the aquarium through a 250-mesh screen. By using care to drain without disturbing the sediment on the bottom of the aquarium, healthy larvae were collected relatively free of debris. These larvae were then resuspended in a three-liter Pyrex jar, the number of larvae per ml. determined, an appropriate volume pipetted into each of a series of one-liter polyethylene beakers, and the volume made up to one liter of the desired salinity. Two beakers of larvae were set up at each of the following salinities: 27 ppt (control), 25 ppt, 22.5 ppt, 20 ppt, 17.5 ppt, 15 ppt, 12.5 ppt and 10 ppt. In the first experiment we used 6400 and in the second, approximately 13,000 larvae per beaker. Following the procedure of Davis (1958), to hold salinities constant, the one-liter cultures received food only every second day when the water was changed, instead of daily as is the usual practice. All the cultures were covered to prevent excessive evaporation and kept in a constant temperature bath at 23 C. 1 C. In the two experiments to determine the effect of low salinities on setting of mature larvae, we reared the larvae in 15-liter culture vessels at 27 ppt until their average size was 250 p. and many were already in the 275 to 300 /A range. Ap- proximately 9000 of these larvae were pipetted into each of a series of one-liter beakers filled with water adjusted to the desired salinities. In the first experiment, because of the limited number of mature larvae available, we used a single culture at each salinity, but in the second, mature larvae were abundant and duplicate cultures were used at each salinity. A single oyster shell in each beaker was used as cultch. These shells were replaced every second day, as the water was changed, and the shells removed OYSTER LARVAE AT LOW SALINITIES 35 were examined under a dissecting microscope to determine the number of spat caught. Only those setting on the smooth, white, inner surface of the shell were counted. Since many oysters also set on the rough, dark, outer surface of the shell, on the walls of the container, and on small shell fragments or other debris where it is impossible to count, the number of spat recorded is only a rough index of the total number setting. RESULTS Effects of lowered salinities on growth oj larvae The results of the first experiment showed that growth of larvae was virtually normal in salinities as low as 22.5 ppt (Fig. 1). At 20 ppt, although larvae grew 1ST EXPERIMENT GROWTH-RELEASE TO 10 TH DAY 270 PPT (CONTROL) 250 PPT 22 5 PPT V///////////////^^^^ 20.0 PPT 12.5 PPT | 90% MORTALITY BY 10 TH DAY 10.0 PPT 100 V. MORTALITY BY 4 TH DAY 2ND EXPERIMENT GROWTH - RELEASE TOI4TH DAY 27.0 PPT (CONTROL) /////m^^ 50 '/. MORTALITY BY IOTH DAY 15.0 PPT ffl 90 '/. MORTALITY BY IOTH DAY 12.5 PPT 100 V. MORTALITY BY IOTH DAY 10.0 PPT 100 V. MORTALITY BY 4 TH DAY 170 180 190 200 210 220 230 240 250 260 MEAN LENGTH IN MICRONS FIGURE 1. Growth of O. cditlis larvae at different salinities. Mean lengths are based on measurements of 100 larvae from each of duplicate cultures at each salinity. 36 HARRY C. DAVIS AND ALAN D. ANSELL at a considerably slower rate, some were reared to setting stage and metamorphosed. At 17.5 ppt growth was extremely slow ; no larvae reached the setting stage and eventually all died. At 15 and 12.5 ppt. growth of larvae was even slower, but some lived at each salinity for ten days or more. Larvae kept at 10 ppt all died within four days. The larvae used in the second experiment were about 11 //, smaller at the time of release and grew at a considerably slower rate than those used in the first ex- periment (Fig. 1). Since larvae from the brood used in this second experiment grew normally in other cultures set up at the same time, we believe the slower growth was the result of the higher concentration of larvae (13,000 per liter as opposed to 6000 per liter in the first experiment). This would indicate that to assure good growth these larvae must be kept in much lower concentrations than larvae of either Venus inercenaria or Crassostrea virginica, both of which grow quite well at 13.000 individuals per liter. TABLE I Effect of reduced salinities on setting of larvae reared almost to setting stage at 26-27 f>f>l Average number of spat per culture Salinity in parts per thousand 1st experiment 2nd experiment 26-27 (Control) 147 134 25.0 504 145 22.5 120 171 20.0 114 -S8 17.5 24 5 15.0 4 0.5 12.5 10.0 7.5 The pattern of growth was generally the same in both experiments, even though growth of larvae was much slower in the second experiment. Growth of larvae was not greatly different from that in control cultures at salinities of 20 ppt and higher. As in the first experiment, the rate of growth dropped off sharply between 20 ppt and 17.5 ppt. Mortality was high and growth negligible at all lower salinities. Because of the very slow growth of larvae, the second experiment was discontinued before larvae in any of the cultures reached setting stage. Effects of lou'cred salinities on selling Two experiments were run to determine the minimum salinity at which larvae of 0. ednlls could set. In both, larvae that had been reared almost to metamorpho- sis at our normal salinity of 26-27 ppt were transferred directly to lowered salinities and their setting recorded (Table I). Although a salinity of 20 ppt was the lowest at which larvae could be reared from release through metamorphosis, some mature larvae transferred to salinities of 17.5 and 15 ppt did set (Table I). OYSTER LARVAE AT LOW SALINITIES 37 It is significant that all of the set obtained at 17.5 and 15 ppt occurred within four days after the larvae were transferred to these salinities, while setting con- tinued for as long as 14 days in higher salinities. This indicates that only those larvae that were almost ready to set at the time of transfer were able to complete growth and metamorphose at these lower salinities, while all less developed larvae died. However, even those larvae that were ready to set at the time of transfer were unable to complete the process of metamorphosis at salinities of 12.5 ppt or lower. A similar record of the number of set obtained from larvae reared at lowered salinities showed that while there was no significant mortality within ten days in a salinity of 15 ppt nor in an}- of the higher salinities, none of the larvae grown at 15 and 17.5 ppt survived to set. Even in cultures kept at 20 and 22.5 ppt, sig- nificantly fewer larvae succeeded in completing metamorphosis than in cultures reared at 25 ppt or in control cultures. Effects o\ lower salinities on gonad development, spatvning and incubation Because it was found (Davis, 1958) that eggs from C. I'irginica that had de- veloped gonads at salinities of about 8.75 would develop into straight hinge larvae at considerably lower salinities than would eggs from parents that had developed gonads at 26-27 ppt, we thought it worth while to attempt to induce gonad de- velopment, spawning and incubation of O. edit I is at lowered salinities. We hoped to use larvae from oysters kept at salinities of 20 ppt and 17.5 ppt in salinity experiments parallel with those on larvae from oysters kept at our normal salinity, to determine whether the salinity tolerances of the larvae were altered by the salinity in which the parent oysters had developed gonads and spawned. The adult 0. ednlis that were kept in aquaria at 20 ppt and t 17.5 ppt failed to release any normal living larvae. In other respects, however, transfer to these lowered salinities appeared to affect them for only a short time. Oysters trans- ferred to 17.5 ppt, for example, failed to feed normally for only a few days, but thereafter cleared the water of the algae added as food, as did those kept at higher salinities. These oysters were kept in these salinities for 60 days and appeared normal in every respect, except that they produced no normal larvae. Evidence of spawning of at least one female, in the group kept at 20 ppt, was observed, but no living larvae were recovered. A few empty larval shells were found on several occasions, but there were not enough of these to account for a normal brood from a single female. Although spawning was not observed in the group kept at 17.5 ppt. on several occasions highly abnormal living larvae were recovered. These larvae either had no shells at all or had small, highly abnormal ones. Nevertheless, many of them were capable of taking food and lived for several days after transfer to normal sea water. Hecause there were only a few of them, their further development could not be followed. DISCUSSION The results of our experiments indicate that culturing of 0. edit! is, in areas where the salinity is 20 ppt or lower, cannot be successful because these salinities are too low for reproduction of this species. It is possible, nevertheless, that in 38 HARRY C. DAVIS AND ALAN D. ANSELL situations where there is a gradation of salinities, mature larvae, developing and growing to setting size at salinities of about 22.5 ppt or higher, could be carried by currents to salinities as low as 15 ppt and set and grow there. Since some of the larvae reared at 20 ppt did metamorphose, it is possible that culture of O. edulis at this salinity might be successful, but growth of larvae would be slow and the intensity of setting reduced. Moreover, adult oysters kept at this salinity failed to give any living larvae. While it is possible that we might have had different results if we had used more oysters, the number of oysters (only 10) kept at this salinity was the same as the number kept at our normal salinity that released the several million larvae used in these experiments. The release of abnormal veligers by oysters kept at 17.5 ppt is, we believe, positive evidence that this salinity is too low for normal larval development. It is possible that oysters acclimated to this salinity over a longer period of time might have given viable larvae. However, the results of Davis (1958) showed that keeping adult American oysters, which had developed gonads at 8.75 ppt. for only a few days at salinities of 7.5, 10 and 15 ppt altered the salinity tolerance of their eggs. Previous investigators using, for the most part, field data have shown that 0. edulls larvae can grow and set at salinities as high as 34 to 39.5 ppt (Mazzarelli, 1924). Korringa (1941) also states that, "It is my opinion that variations in salinity between 25 ppt and 35 ppt probably have little or no influence on larval growth and development in Ostrea c dulls" (pp. 133-134). Our results cannot be compared directly with those of Walne (1956) because he did not change the water and his salinities were not held constant. Using our methods, however, larvae were reared to metamorphosis and some set was ob- tained at salinities of 20 and 22.5 ppt, both lower than the lowest initial salinity (25.9) from which Walne obtained spatfall. Our results confirm Korringa's opinion that a salinity of 25 ppt has no appreci- able adverse effect on growth of larvae. They further show that the lower limit for good growth and setting is about 22.5 ppt, although larvae can grow to metamorphosis at 20 ppt and mature larvae can set at even lower salinities. The authors wish to express their appreciation to Dr. V. L. Loosanoff, Director of Milford Laboratory, for suggesting the problem and editing the manuscript, to Mr. Herbert Hidu for assistance in caring for the cultures and making many of the measurements, to Mr. Manton Botsford for preparing the figure, and to Miss Rita Riccio and her staff for their care in editing and preparing the manuscript. Thanks are also given to the John Murray Committee of the Royal Society for the travelling grant to one of us (A. D. A.) which made this cooperation possible. SUMMARY 1. At salinities of 25 and 22.5 ppt growth of larvae of O. edulis and intensity of setting was not significantly different from that in control cultures at our normal salinity of 26-27 ppt. 2. At 20 ppt growth of larvae was appreciably slower than at higher salinities and the intensity of setting was reduced. OYSTER LARVAE AT LOW SALINITIES 39 3. At 17.5 and 15 ppt larvae lived for some time and showed appreciable growth, but they all died prior to metamorphosis. 4. At 12.5 ppt larvae showed no growth and by ten days after swarming they had suffered 90% or higher mortality. 5. At 10 ppt all larvae died in less than four days. 6. Some larvae that had been reared to setting size at a salinity of 26-27 ppt were capable of setting in salinities as low as 15 ppt. 7. No normal larvae were obtained from adult oysters kept at salinities of 20 or 17.5 ppt. LITERATURE CITED DAVIS, H. C., 1958. Survival and growth of clam and oyster larvae at different salinities. Biol. Bull, 114: 296-307. * GAARDER, T., 1932. Untersuchungen iiber Produktions und Lebensbedingungen in Nor- wegischen Austernpollen. Bergcns Mus. Arbok 1932 Naturv. Rckke No. 3. * GAARDER, T., 1933. Austernzucht in Norwegen. Chemisch biologische Untersuchungen in Norwegischen Austernpollen. Intern. Rev. d. gesammt. Hydrobiol. u. Hydrogr. Bd. XXVIII, Heft 3/4, 250-261. * GAARDER, T., AND P. BJERKAN, 1934. 0sters og 0sterskultur i Norge. Bergen 1934. A. S. John Grieg. KORRINGA, P., 1941. Experiments and observations on swarming, pelagic life and setting of the European flat oyster, Ostrca cdulis L. Arch. Neerl. Zool., 5 : 1-249. LOOSANOFF, V. L., 1951. European oyster, O. cdulis, in the waters of the United States. Anat. Rec., Ill: 126. LOOSANOFF, V. L., 1952. Imported European oysters thriving. Atlantic Fisherman, December, p. 45. LOOSANOFF, V. L., 1955. The European oyster in American waters. Science, 121 : 119-121. *MAZZARELLI, G., 1924. Note sulla biologia dell' ostrica (Ostrea edulis L.). IV. Durata e andamento del periodo reproduttiva delle ostriche del Lago Fusaro. Boll. Soc. Nat. NapoU, 36: 158-178. WALNE, P. R., 1956. Experimental rearing of the larvae of Ostrca cdulis L. in the laboratory. Min. Agric., Fish. & Food, Fish, hives. Series II, 20: 1-23. WELCH, WALTER R., in press. The European oyster, Ostrea edulis, in Maine. Conv. Address, Nat. Shellfish. Assoc. * Reviewed by Korringa, 1941, cited above. THE BIOLOGY OF ASCIDIA NIGRA (SAVIGNY). I. SURVIVAL AND MORTALITY IN AN ADULT POPULATION IVAN GOODBODY Department of Zoology, University College of tlie ll'est Indies, Jamaica Ascidia nigra is a solitary ascidian with a jet black test and measuring 10 to 12 cm. in length when fully grown. The colour of the test is due to pigment granules which migrate out through the test and form a thin layer at the surface where the pigment continuously sloughs off, together with some of the test substance. This process, which was first described by Hecht (1918) and has been confirmed by the writer, makes it nearly impossible for other sedentary organisms to settle and grow on the test. Unlike many other ascidians, therefore, A. nigra has a clean test throughout its life and this, combined with its conspicuous colouration, makes it an ideal animal for ecological study ; against its background it stands out con- spicuously and the history of individuals may be recorded easily. The species is widely distributed throughout the warm waters of the western Atlantic ranging from Florida (25 N.) to Sao Sebastian, Brazil (24 S.). It occurs in Bermuda (32 N.) (Van Name, 1945) and is recorded from the Red Sea, Gulf of Aden and Gulf of Guinea (Millar, 1958). Throughout the West Indies it is one of the commonest inshore shallow-water ascidians, usually confined to the sheltered waters of harbours and mangrove lagoons ; it is only rarely found on coral reefs. In the harbours and lagoons it is found attached to mangrove roots, piers, pilings, buoys and ship bottoms from surface level to about 25 feet ; it also occurs sometimes on the sea floor where a suitable hard substratum, such as a stone, enables it to settle. In general it is a primary coloniser, developing large popula- tions on new or cleaned surfaces but ultimately being replaced by other organisms. The primary sessile community in Kingston Harbour, Jamaica, is dominated by algae (EnteromorpJia) , cirripedes (Balamts amphitrite Darwin), hydroids, ser- pulids and the ascidians 1 A. nigra, Didcinnniii candidum, Diplosonia macdonaldi. Symplcgma viridc and Polyclinum constcllatniii. This gives way by various stages to a climax community dominated by sponges (principally Haliclona sp., Tcdania ignis (Duchassaing and Michellott) and Mycalc cccilia deLaubenfels ) , anemones (Aiptasia tagetes (Duchassaing and Michellott)), ophiurans (Ophioihrix angiilata Say) and the ascidians Microcosmus e.vasperatns, Hcrdinania inoinus and Pyura vittata, but in which Ascidia nigra is relatively rare. A detailed account of suc- cession in this community will be published in a latter paper. The present paper describes the history of a population of Ascidia nigra which settled and developed naturally on panels suspended in the sea at Port Royal in Kingston Harbour, Jamaica (Station C in Goodbody, 1961a). During the period of observation the population suffered a density-independent mortality following 1 Nomenclature of ascidinns is taken from Van Name (1945). 40 BIOLOGY OF ASCID1A NIGRA 41 an influx of fresh water to the harbour ; the details of this are discussed below and in the paper referred to above. METHODS A raft (or pontoon) four feet square was moored 20 feet away from the sea wall at Morgan's Harbour, Port Royal, where the water is 23 feet deep. On each of the four sides of the raft two panels were suspended one below the other on galvanised chains so that the upper panel had its upper edge 4 feet below the water surface and the lower one 7.5 feet below the surface. The panels hung so that the settling surfaces were vertical in the water. In this way 8 panels were suspended from the raft, 4 at each vertical level. Each panel was made of a rectangular sheet of Vs-inch-thick "Tufnol" - 18 X 15 inches, bound along each side of each long edge by a thin strip of steel 1.5 inches wide. This reduced the area of "Tufnol" available for settlement to 18 X 12 inches. Both sides of each panel were available for settlement and were desig- nated Front (F) and Rear (R) according to whether it faced out from the raft (F) or in towards the mooring chain (R). The raft was moored by a single central chain and swivel so that it could turn horizontally about a central axis. Though undesirable from some points of view, this and the small size of the raft were essential parts of an attempt to minimise the risk of destruction by hurricanes. The 8 panels were placed in position on August 1, 1957, and were not finally removed until May 6, 1960. At intervals the panels were removed from the raft and taken to a tank in the marine laboratory. The maximum time any panel was out of the water at any one time was one minute. In the laboratory the position of each specimen of A. nigra was plotted on a map of the panel and a colour photograph of each side of the panel was taken. From the photographs permanent maps were made and each animal designated by a serial number. In this way a permanent record was obtained of the history of each animal whose position was known on the map from its first appearance to its ultimate death or disappearance. In the course of lifting and examining panels a total of five animals were accidentally killed. These have been excluded from the analysis of data except in Table II which is concerned solelv with new settlements. j COLONISATION AND POPULATION GROWTH Table I shows a summary of the history of the population, and in columns 4 and 5 and in Figures 1 and 2 are shown the number of new animals appearing on the panels and the total population present. It will be seen that the majority of animals colonised before September 26, 1957 (57 days after the start of the experiment) and very few animals colonised after October 25, 1957. Newly settled A. nigra do not begin to develop the black pigment until about 20 days after metamorphosis and hence are only just visible to the naked eye as small black "Tufnol" is the trade name for a synthetic resin bonded fabric sheet conforming to British Standard No. 972: 1941. It is ideal material for experimental panels as it is the colour of dark wood and is completely unaffected by sea water or by boring animals as Teredo or Linnioria. 42 IVAN GOODBODY objects when 4 weeks old (unpublished observations). It is probable, therefore, that the majority of animals settled on the panels in the first 4 weeks but were not all visible when they were first examined on August 28, 1957. It is clear from these figures that once the primary colonisation is complete few new animals settle in the community. Elsewhere (Goodbody, 1961c) I have shown that Ascidia nigra breeds throughout the year in Jamaica, so that the fall-off in numbers 2OO- O 2' 3 'A ' 5 6 7 8 9 IO II 12 13 14 15 Ib 17 PERIOD FIGURE 1. The number of new Ascidia nigra appearing on panels (settlers) at eacli inspection throughout the life of the population. The length of each period is not the same ( see Table I). of new animals appearing on the panels cannot be due to a seasonal drop in the number of larvae available. The fall-off in numbers must therefore be due to competition, either inter- or intraspecific, most probably the former. If intraspecific competition was important in determining settlement and growth of new individuals it would be expected that late settlers would tend to appear on panels with few other A. nigra. BIOLOGY OF ASCID1A NIGRA 43 Table II examines this possibility and analyses data for 27 new settlements between November 1, 1957, and October 3, 1958. It is clear from this table that there was no tendency for new settlements to occur on panels with few, as opposed to many, other A. nigra. It is more probable that the suppression of later colonisa- tions is an effect of the whole community, in fact a combination of inter- and intra- specific competition. Most of the organisms on the panel are leptopel feeders competing for similar food to that of A. nigra, and any newcomer must have difficulty in obtaining sufficient food for growth. Furthermore, sponges may actually inhibit the development of other sessile organisms (Goodbody, 1961b). 35O- g250- o Q_ I5O- O 5O- \ \ O 100 2OO 30O 4OO DAYS 5OO 60 O 7OO FIGURE 2. The total population of Ascidia nigra present on panels throughout the life of the population. The sharp drop after day 435 is due to fresh-water floods (see text). Figure 2, which illustrates the growth of the total population, also illustrates the rapid rate of colonisation of the new panel. Subsequent to the initial rapid rise in population size, the population becomes stable for a long period, and since few new colonisations are taking place this indicates also a low mortality rate which is discussed in the next section. The data outlined above, depicting A. nigra as a primary coloniser in sessile communities, are substantiated by numerous field observations of natural situations. In recent years several mass mortalities have occurred among sessile communities in Kingston Harbour (Goodbody, 1961a) and subsequently large areas of cleaned 44 IVAN GOODBODY TABLE I Total settlement, mortality and mortality index. Mortality index is found bv 100 Diiiltiblying percentage mortality by - ' no. days Period Date Interval in days Total population No. new animals settling No. animals dying % Mortality Mortality index 1. 28. 8. 57 28 117 117 2. 26. 9. 57 29 327 210 3. 25. 10. 57 29 372 50 5 1.5 5 4. 27. 11. 57 33 379 11 4 1.1 3 5. 8. 1. 58 42 381 4 2 0.5 1 6. 6. 2. 58 29 373 3 11 2.9 10 7. 12. 3. 58 34 366 1 8 2.1 6 8. 9. 4. 58 28 363 1 4 1.1 4 9. 6. 5. 58 27 358 5 1.4 5 10. 25. 6. 58 50 350 3 11 3.1 6 11. 30. 7. 58 35 337 13 3.7 11 12. 3. 10. 58 65 320 3 20 5.9 9 13. 27. 10. 58 24 120 2 202 63.1 263 14. 8. 12. 58 42 110 2 12 10.0 24 15. 13. 1. 59 36 90 1 21 19.1 53 16. 26. 2. 59 44 60 30 33.3 76 17. 26. 3. 59 28 49 11 18.3 65 18. 30. 4. 59 35 19 30 61.2 175 19. 27. 5. 59 27 4 15 78.9 292 20. 10. 6. 59 14 2 2 50.0 357 21. 7. 7. 59 27 2 100 370 surfaces became available for larval settlement on mangrove roots, piers, pilings, etc. A. nigra has always been one of the first colonisers of these surfaces, some- times developing as dense clusters of several hundred individuals. Similarly, whenever new piles are driven this species develops in dense clusters in the early stages. However, as the sessile community develops on these surfaces, A. nigra is slowly replaced by dominants such as M. e.vasperatiis, P. I'ittata and H. nwtnus, together with sponges, anemones and lamellibranchs. Why A. nigra should succeed only as a primary coloniser is not clear at present, but it is pertinent to point out that at all stages of its growth the siphons project TABLE l( Settlement of new animals, in relation to density of A. nigra already on panel, between 25th October 1957 and 3rd October 1958 Panel density No. of panels No. of settlements No. of settlements per panel 11-15 4 6 1.5 16-20 2 2 1.0 21-25 4 6 1.5 26-30 3 7 2.3 31-35 2 5 2.5 36-40 1 1 1 BIOLOGY OF ASCIDIA NIGRA 45 out into the water beyond the level of the remainder of the community, including other ascidians. This suggests that competition for food may be of paramount importance and that by keeping the siphons projecting it can tap the food supply before it reaches the other members of the community. 300- a: O IOO- o IOO ZOO 3OO 4OO DAYS 500 6OO 7OO FIGURE 3. Mortality rate expressed as mortality index (see text). The dotted line connects the two periods before and after the fresh-water floods. MORTALITY AND SURVIVAL The intervals between successive inspections of the panels were not constant and so it is not possible to construct the usual form of life table. The succeeding analysis is therefore confined to two functions of the life table, mortality rates and survival. In place of the usual "mortality rate" (1000 q.r) I have substituted a "Mortality Index" which may be expressed as where / is the interval in days, D is the number of deaths during the interval and P the total population at the beginning of the interval. The Mortality Index provides a means of comparing the mortality rate over the whole life span of the population. The mortality index for the whole population is shown in Table I and Figure 3. In calculating this function no allowance has been made for the small increment in 46 IVAN GOODBODY the population subsequent to the establishment of the main population. It is clearly demonstrated here that from the time animals first appeared as small black ascidians, the mortality rate was low until the population was about 15 months old, after which it climbed rapidly until the last animal died when the population was 23 months old. The greatest life span attained by any one animal was an IOOO-I 50O- \ \ 25O- 00 C O a; Z) oo 3- \\ ' * \ \ \ \ -^v_ i l i J ..... i ii i O i i i i i f I T" ISO 3OO 45O 6OC ) DAYS FIGURE 4. Corrected survival curves for three groups of ascidians which first appeared on the panels in August, September and October, 1957 (see Table III). : Group A, appearing in August. : Group B, appearing in September. : Group C, appearing in October. For the sake of clarity curves B and C have not been carried back to the origin. individual which appeared at the first inspection, hence settled in the first week of August, 1957, and died between May 27 and June 3, 1959. It must therefore have been between 94 and 96 weeks old (22 months) at the time of death. The curves shown in Figures 2 and 3 are complicated by a density-independent mortality due to fresh- water floods occurring between October 3 and 27, 1958. Following heavy rain a layer of low-salinity water penetrated the harbour and caused extensive mortality among sedentary organisms (Goodbody, 1961a). On panels at the 4-foot level 86% of the A. nigra were killed and at the 7-foot level 33% were killed. Subsequent to this mass mortality the mortality curve climbs BIOLOGY OF ASCIDIA NIGRA 47 steadily upward, but there is no reason to believe that this is not normal or that it is in any way connected with the floods. Data from 18 animals in a pilot experiment run in 1956-58 show that 50% had died in 12 months and all were dead after 85 weeks (19-J months). TABLE III Survival rale (Ix) in three separate groups of A. nigra showing the observed rate and the re- calculated rate for a population free from density independent mortality (see text). Group A settled in Aiigust, 1957, B in September and C in October. Figures in parentheses show actual number of animals in each group at outset. For dates corresponding to period numbers see Table I. Periods added to this table are 18a: 6. 5. 59, 18b: 13. 5. 59, 18c: 20. 5. 59, 19a: 3.6.59 Period No. Observed Survival (Ix) Recalculated Survival (Ix) Group A (117) Group B (210) Group C (50) Group A (117) Group B (210) Group C (50) 1. 1000 . 1000 2. 1000 1000 - 1000 1000 3. 1000 976 1000 1000 976 1000 4. 991 967 980 991 967 980 5. 991 957 980 991 957 980 6. 974 938 920 974 938 920 7. 974 919 920 974 919 920 8. 957 914 920 957 914 920 9. 940 900 920 940 900 920 10. 915 876 860 915 876 860 11. 897 848 840 897 848 840 12. 838 810 780 838 810 780 13. 350 262 300 811 769 755 14. 325 224 280 753 658 704 15. 265 176 240 614 517 603 16. 179 124 160 500 364 402 17. 145 90 160 405 264 402 18. 77 24 40 215 70 100 18a. 51 5 20 142 15 50 18b. 43 20 120 50 18c. 9 20 25 50 19. 9 20 25 50 19a. 20 50 20. Since settlement of young animals goes on over a period of three months with small increments in later intervals, the population being dealt with cannot be con- sidered as homogeneous in age structure. The information in Table I is therefore of limited interest only as a history of the entire population. It is more informative to consider separately the life table data for each of the three groups of animals settling in the first three months and to see if any differences exist between them. This is accomplished by a comparison of survivorship (Ix) data and is shown in Table III and Figure 4 for the groups of animals which first appeared on panels in August, September and October, 1957. 48 IVAN GOODBODY Table III has been constructed so as to show the actual survival, and calculated survival data in which the effect of the density-independent mortality has been eliminated. To calculate the latter it has been assumed that the mortality index between October 3 and 27, 1958, would normally have been the mean of that pre- vailing in the periods immediately preceding and succeeding it; from this a value for the percentage mortality in period 13 has been calculated for each of the three groups. By substituting this value in place of the observed mortality in period 13 the Ix values for periods 13-21 can then be re-calculated to give a picture of the survival curve for a population free from density-independent mortality. Figure 4 shows corrected survival curves for the three populations. In con- structing this figure it has been assumed that population A settled on August 1, B on August 28, and C on September 26, and the ordinate plots the number of days since settlement. In this way the three survival curves can be directly com- pared. This curve can be only an approximation to the true survival curve, but it is probably reasonably close. The obvious sources of error are in the calculation of the per cent mortality for period 13 and in assuming that all the animals in a group settled at the chosen dates. Furthermore, this does not include mortality in the first 28 days and is only a survival curve for animals of 28 days and over. Two features of interest are apparent in Figure 4. First the survival curve in each case is of the "negative skew" rectangular type (Deevey, 1947) in which few deaths occur until very near the maximum length of life and most of the animals die in a short space of time near the end. This suggests senescence and physio- logical longevity and is discussed further below. Secondly, although the form of the curve is similar in all three groups the first settlers appear to have lived for longer than the other two groups. The 50% level was reached after 550 days in Group A, 480 in Group B and 465 in Group C. There are three possible explanations for this difference. (1) The population might really be homogeneous, all having settled in the first week of August, and the early development of some of them may have been greatly retarded by, for example, food shortage, so that they appeared very much later than the others. This possibility, though real, can probably be ruled out. The maximum error that could have arisen this way is 58 days, while the difference in longevity is of the order of 85 days. (2) The first settlers may be genuinely more successful than later settlers and consequently live longer. (3) The sharp drop in the curve may be the result of a change in the environment and be ecological mortality and not senescence. It is not possible at present to decide definitely between these two latter possibilities but in assessing them the following points may be considered. Life spans of 18-20 months have been calculated for several temperate water species of ascidians by Millar (1952, 1954) and Sabbadin (1957, 1958), and animals in a pilot experiment with A. nigra in 1956-58 had a total life span of 19-20 months. This supports the contention that the curve may be natural and senescent. On the other hand I have shown elsewhere (Goodbody, 1961c) that the climax sponge/ anemone/ophiuran community inhibits the development of the early stages of the primary colonising community, possibly as a result of toxins produced by sponges or by competition for food. This "influence" might extend to the adult population of A. nigra and be responsible for its death. In support of this we may note BIOLOGY OF ASCIDIA NIGRA 49 that the final death of all three populations occurred at approximately the same time (Table III), and that this coincided with the time when the climax com- munity was reaching full development. This problem can be resolved only by further experiment. While there is a high rate of survival in populations settling in the first few months, the same does not appear to be true for animals which settle at later times. In Table IV the survival of animals to October 3, 1958 (i.e., immediately before the floods) is shown for animals settling at different periods. There is a considerable and significant difference between those settling in August, Sep- tember and October and those settling from November to June. Here is further evidence that A. nigra can survive only as a primary coloniser in the community, and that new individuals do not appear in the community because they cannot compete in it. TABLE IV Showing different rates of survival in groups of Ascidia nigra settling at successively later stages. Survival has been taken to 3. 10. 58 (approximately one year) as density-independent mortality occurred on 6. 10. 58 Settlement date Number settling Number surviving 3. 10. 58 % Surviving Aug., 1957 117 98 84 Sept., 1957 210 170 81 Oct., 1957 50 39 78 Nov., 1957 11 5 45.5 Dec., 1957 to June, 1958 12 5 42 I know of no active predator on the adult ascidian and it appears that the test substance is distasteful. Fishes in an aquarium will eat the body tissue of an Ascidia nigra but examine and reject the test substance if that is fed to them. A small amphipod, Erichthonius brasiliensis (Dana) is common on the test and in and around the siphonal margin, but this is probably a true commensal arrange- ment. Several copepod commensals occur in the branchial sac and also a large pea-crab, Pinnotheres moseri Rathbun. The incidence of this crab in animals is very variable but there seems to be never more than one in any ascidian. There is no evidence that it harms the ascidian and all infected animals have appeared normal and healthy. Externally a galatheid crab and the ophiuran, Ophiothrix angulata, are com- monly found wandering on the test but have probably no effect other than to help keep the test clean. A large xanthid crab (Menippc nodijrons Stimpson) occa- sionally makes tunnels around the base of sessile organisms in this community and in a later experiment was suspected of actually dislodging some ascidians from a panel, but this is rare. DISCUSSION Previous studies of the life cycle of solitary ascidians have been published by Millar and Sabbadin, both of them concerned with temperate-water species. Mil- 50 IVAN GOODBODY lar (1952) studied Ascidiella aspcrsa and Ciona intestinalis in the Clyde, Scot- land, and found the pattern of the life cycle of both species to be fairly similar. There is a single breeding season in summer, involving several generations of larvae; and the young animals grow until late autumn when growth ceases. In the following spring growth re-commences and these animals form the breeding stock for the subsequent summer spawnings. This adult population died off in the following winter, thus having a total life span of 12-18 months. In a later paper Millar (1954) studied populations of Dendrodoa grossularia from the Clyde and from Essex, England. This species has a similar form of life cycle, living for from 18 to 24 months, but in Essex has two distinct breeding seasons in each summer. Sabbadin (1957, 1958) studied Ciona intestinalis, Molgula manhattensis and Styela plicata in Venice, Italy. Breeding is continuous from early spring to late autumn in all three species. Animals metamorphosing in early spring may breed during the ensuing summer and have completed their growth within the year. Later settlers cease growing during winter, and do not complete their growth and breed until the following spring. The total life span is reported to be "about one year" (Sabbadin, loc. cit.}. Except for the similarity in total life span, the pattern of annual cycle in these temperate-water species differs from that of Ascidia nigra. Whereas the former have annual breeding seasons, A. nigra breeds throughout the year so that the population is composed of animals of all ages. Present information suggests that A. nigra commences breeding when it is about 85 days old and thereafter may spawn at intervals of about 60 days. Growth in the temperate-water species is arrested during the winter months and according to Sabbadin (1957) is definitive. However, this author studied ascidian populations by means of monthly samples and his observations on growth are based entirely on these samples. From such data it is not possible to say with certainty that growth is not continuous throughout life. Unpublished data on A. nigra suggest that in Jamaica this species may continue to grow throughout life though at a progressively retarded rate. This point is important in relation to senescence and the problem of why ascidians die. Life tables for other animals have been extensively reviewed by Deevey (1947) and by Alice et al. (1949). The point of interest here is in the type of survival that is illustrated. The curve shown in Figure 4 is typical of a population which exhibits senescence (Comfort, 1956) and very little ecological mortality. This raises again the question of why ascidians die. It has been mentioned earlier that the decline in the ascidian population is coincident with the rise of the climax sponge community. Experiments are now in progress to try and determine whether one is dependent on the other or whether the ascidians die for physiological rather than ecological reasons. In the studies by Millar and Sabbadin (loc. cit. ) populations died over the winter months when water temperatures were cool. In these areas declining temperature may be the cause of, or hasten, the onset of death and thus obscure the incidence of senescence. In Jamaica, water temperatures vary about 6 C. throughout the year (Goodbody, 1961c) and are unlikely to be concerned in any way with death. Finally it should be emphasized again that these data deal only with ascidians from the twenty-eighth day of life onwards and more drastic mortality is to be * * v BIOLOGY OF ASCID1A XKiKA 51 expected in the first four weeks. Data on this period are at present accumulating and will form the subject of a later paper in the series. This work was supported by a grant from the Xuffiekl Foundation to whom acknowledgment is made. I am also grateful to Mr. G. Hechtel for the identifica- tion of sponges, and to Dr. T. E. Bowman for identifying the amphipod. SUMMARY 1. A "marked" population of Ascidia iiigra has been followed throughout life from their first appearance to the time of death. 2. Mortality in the first four weeks after metamorphosis has not been studied. Thereafter few animals die until they are 18 months old and all are dead at 22 months old. 3. A mass mortality due to fresh- water floods occurred in the period of ob- servation. A corrected survival curve has been calculated for a situation in which this did not occur. This suggests that senescence may occur. 4. A. nigra is a primary coloniser. When colonising new surfaces, the earliest colonisers survive longer than the later ones. Very few animals can colonise after the primary sessile community is three months old. LITERATURE CITED ALLEE, W. C, O. PARK, A. EMERSON AND K. P. SCHMIDT, 1949. Principles of Animal Ecology. Philadelphia : Saunders. COMFORT, A., 1956. The Biology of Senescence. London : Routledge and Kegan Paul. DEEVEY, E. S., 1947. Life tables for natural populations of animals. Quart. Rev. Blol., 22 : 283-314. GOODBODY, I., 1961a. Mass mortality of a marine fauna. Ecology, 42: 150-155. GOODBODY, I., 1961b. Continuous breeding in three species of tropical ascidians. Proc. Zool. Soc. London, 136: 403-409. GOODBODY, I., 1961c. Inhibition of development of a marine sessile community. Nature, 190: 282-283. HECIIT, S., 1918. The physiology of Ascidia aim Lesueur. I. General physiology. /. Exp. Zool., 15: 229-259. MILLAR, R. H., 1952. The annual growth and reproductive cycle in four ascidians. /. Mar. Biol. Assoc., 31 : 41-61. MILLAR, R. H., 1954. The annual growth and reproductive cycle of the ascidian Dendrodoa grossnlaria. J. Mar. Biol. Assoc., 33: 33-48. MILLAR, R. H., 1958. Some ascidians from Brazil. Annals Mag. Nat. Hist., Scr. 13, 1 : 97-514. SABBADIN, A., 1957. II ciclo biologico di Ciona intcstinalis (L), Molyula manhattensis (De Kay) e Styela plicata (Lesueur) nella Laguna Veneta. Archiv. Ocean. Liinnol. Venc- sia, XI: 1-28. SABBADIN, A., 1958. Sur les caracteristiques du cycle biologique de quelques ascidies dans la lagune de Venise, en rapport avec le regime thermique. Commission Internat. E.rplor. Scicntifiquc dc la mcr Mcditerranee, Rapports, XIV: 577-581. \ AN NAME, W. G., 1945. The north and south American ascidians. Bull. Aincr. Mns. Xat. Hist., 84 : 1-476. FREEZING RESISTANCE IN SOME NORTHERN FISHES 1 MALCOLM S. GORDON, BEN H. AMDUR 2 and P. F. SCHOLANDER 3 Department of Zoology, University of California at Los Angeles, Department of Chemistry, Harvard University, and Scripps Institution of Oceanography, University of California, San Diego Scholander et al. (1957) described two groups of marine fishes from Hebron Fjord, Labrador, which spend long periods at subfreezing temperatures. One group (deep-water fishes) appears to survive by remaining perpetually in the supercooled state. The second group (shallow- water fishes) is exposed to sub- freezing temperatures for about two-thirds of each year and survives these periods by combining supercooling with some increase in the osmotic concentration of their body fluids. The former group lives in water depths such that ice is absent even during the coldest winters. The latter group, however, lives in shallower areas and is therefore likely to encounter ice. Natural selection has presumably operated on this second group so that those fishes have survived best which have added some antifreeze to their blood, thereby reducing the possibility of their fatally freezing as a result of an accidental collision with some ice. Several species of Norwegian boreal and arctic fishes apparently also respond to subfreezing tem- peratures in much the same way as the shallow water fishes from Labrador (Elias- sen ct al., 1960). The conditions of life faced by the shallow-water fishes in Labrador in winter give rise to several questions, such as : how can these fishes tolerate any degree of supercooling at all, living as they do in close proximity to ice (supercooled non- arctic fishes freeze rapidly when touched by a piece of ice (Scholander ct al., 1957) ) ; and what is the nature of the antifreeze substance? It is not NaCl. The present paper describes new observations bearing on these subjects. MATERIALS AND METHODS In March, 1959, the present authors made a return visit to the Hebron Fjord. A small prefabricated laboratory hut was set up on the ice near Hebron settlement, 1 These studies have been supported by contracts between the Office of Naval Research, Department of the Navy, and the Arctic Institute of North America ; also by a research grant from the National Science Foundation (G8802). Reproduction of this paper in whole or in part is permitted for any purpose of the United States Government. -Present address: Forsyth Dental Infirmary, 140 Fenway, Boston 15, Massachusetts. 3 The authors' thanks for aid and cooperation are due the following people : In Labrador, Rev. and Mrs. Hettasch and Mr. and Mrs. T. Baird. At the Biological Station, Fisheries Research Board of Canada, St. Andrews, New Brunswick, Dr. J. L. Hart, Mr. L. R. Day, the crew of the station trawler "Mallotus," and the station staff. At UCLA, Dr. J. Mead, Department of Nuclear Medicine and Radiation Biology, and Dr. K. Allen, Department of Zoology. Arctic cold weather equipment and sleeping bags for the Labrador expedition were kindly furnished by the U. S. Army Quartermaster Research and Development Command, Natick, Massachusetts. Valuable technical assistance was provided by David Hall, Anita Battiste, Cynthia Roscnblum and Gordon Engel. 52 FREEZING RESISTANCE IN NORTHERN FISHES 53 and large numbers of fishes were caught by jigging with hand lines through holes in the ice over 2-10 meters of water. Water temperatures were between - - 1.68 and -1.81 C. Two species of fishes were taken: the short-horned sculpin (Myoxocephalus scorpiits) and the Fjord cod (Gadiis ogac). The sculpins weighed 200-800 gm. ; the cod weighed 400-1000 gm. Observations made on numbers of intact, living sculpins and cod were : body temperature, measured by thermometer inserted several centimeters into the cloaca ; and survival time following contact with ice, either on the outer body surface, on the gills, or in the main body muscle mass by insertion of crystals under the skin. Other fishes were immediately taken into the heated laboratory and blood samples taken by heart puncture. Heparin was added to a few samples, but most were simply allowed to clot in plastic centrifuge tubes, then stored on ice (for a maximum of three hours ) until they could be centrifuged on an electric centrifuge. Following removal of aliquots for the analytical procedures carried out in the field, the plasma or serum samples were sealed into Pyrex glass ampules, frozen and returned in the frozen state to Gordon's laboratory. Frozen serum samples were obtained from 32 sculpins and 6 fjord cod. Analyses made at Hebron were : freezing point depression, using the method described in Scholander ct al. (1957) ; electrical conductivity, measured on 1:10 dilutions of serum with an Industrial Instruments, Inc., Model RC-16B conduc- tivity bridge calibrated against known NaCl solutions ; urea plus ammonia nitrogen, measured on two sculpin samples by the Conway microdiffusion technique ; glucose, measured on two sculpin samples by the glucose oxidase method ; glycerol, spot- tested on several sculpin samples using the acrolein reaction. In April, 1960, Gordon visited the Biological Station of the Fisheries Research Board of Canada. St. Andrews, New Brunswick. Sea water salinities at St. Andrews were similar to those in the study area in Labrador. Water temperatures were 3-4 C. Winter water temperatures around St. Andrews rarely are as low as C. Eighteen short-horned sculpins were used to determine the tolerance of a non-arctic population of this species to water refrigerated to - - 1.5 C. Four re- maining sculpins were kept at 4 C. and serum samples were carried frozen to California. Forty-two tomcod (Microgadits tomcod a close relative of Gadns ogac ) were used in several groups for low temperature tolerance experiments involving cooling from 4 to - 1.5 over periods of 12-48 hours. Blood samples were obtained from five tomcod which survived at - 1.5 for 91 days. Sixteen other tomcod maintained at 4 C. were also used for blood samples. The serum obtained from all of these fish was also carried frozen to California. The serum and plasma samples were used in attempts to identify the antifreeze substance. Samples were stored frozen at about - 20 C. and gradually used for chemical analyses over a two-year period. A series of general analyses (e.g., freezing-point depression, Na, K, Cl. non-protein nitrogen) were carried out on groups of samples from each species ; then, as what appeared to be promising leads developed, more specific techniques were used. These specific techniques 54 MALCOLM S. GORDON, BEN H. AMDUR AND P. F. SCHOLANDER TABLE I Analytical procedures used on fish plasma and scrum samples Analysis Procedure I. General analyses done on whole serum or plasma: Freezing-point depression a ' b ' c Chloride 8 ' 6 ' Sodium a ' b ' c Potassium a ' b ' c Total phosphorus a ' b Urea a - b Cryoscopy (Ramsay and Brown, 1955) Voihard AgNO, - SCN titration Flame photometer on diluted samples Flame photometer on diluted samples Colorimetry (Feigl, 1947, p. 317; Chen el al, 1956) Conway microdiffusion technique (Natelson, 1957, p. 387 II. Analyses done on samples deproteinized with trichloracetic acid or (usually) ethanol: diethyl ether (3:1, V/V), some samples also desalted, either by electrodialysis or with pyridine: Xon-protein nitrogen a - b - c Amino nitrogen a ' b Glycerol a Ascorbic and dehydroascorbic acids" Reducing sugars" Xon-reducing sugars" Aldehydes and ketones* -diketones a Carboxylic acids* Amino acids a - b - c Aromatic compounds* Primary alkyl amines b Secondary amines b Tertiary amines b Purines and pyrimidines a Nesslerization (Natelson, p. 272) Colorimetry (Natelson, p. 93) Resorcinol test (Jones, 1947); paper chromatography (Block et a!., 1958, pp. 178, 182) Paper chromatography as for glycerol ; colorimetry (Schaf- fert and Kingsley, 1955) Paper chromatography as for glycerol ; Folin-Wu method (Natelson, p. 205) Paper chromatography (Block et al., p. 185) Paper chromatography (Block et al., p. 340) o-dinitrobenzene test (Feigl, 1956, p. 24) Paper chromatography (Block et al., pp. 216, 231) Paper chromatography (Scherbaum et al., 1959) A1CU test (Shriner and Fuson, 1948, p. 89) Rimini test (Cheronis and Entrikin, 1957, p. 260); 2, 4-dinitrochlorobenzene test (Smith and Jones, 1948, p. 1 10) Nickel-dithiocarbamate test (Duke, 1945) N-bromosuccinimide test (Cheronis and Entrikin, p. 258) Paper chromatography (Block et a/., p. 285); ultraviolet spectroscopy (see III below) III. Analyses done on sculpin samples deproteinized as in II, vacuum-distilled to dryness at 30 5 C., then redissolved in water or various organic solvents in some cases two or three solvent extraction stages, with vacuum distillation to dryness between stages. pH a Ultraviolet absorption spectra a ' b - c Infrared absorption spectra a ' b - c Simpler carboxylic acids and esters, including lipids a Amino acids a pH meter on diluted samples UV spectrometer on pH 2, 7 and 11 water extracts dried and re-dissolved in ethanol :diethyl ether (3:1, v/v) IR spectrometer, on extracts dissolved in absolute ethanol, CHC1 3 , CC1 4 , CS 2 and diethyl ether in various sequences Gas chromatography of methyl esters in ethanol, diethyl ether, CHC1 3 and CC1 4 extracts (courtesy J. Mead) As under III above, also by column chromatographic fractionation with identification on paper (courtesy K. Allen) ,b,c. Species on which analyses done. a M. scor pins', b G. otjor; c .!/. tomrod FREEZING RESISTANCE IN NORTHERN FISHES 55 TABLE 1 1 Plasma concentrations in the short-horned sculpin (Myoxocephalus scorpiusi Concentration [x S. E. (N)] Labrador summer Labrador spring \Vw Brunswick spring 4-7 C. -1.7 C. +4" C.f A (mOsm/1.) 430 10 (6)* 672 15 (17) 450 5 (4) 775 40 (6) Cl (meq./l.) approx. 200* 234 3 (6) 184 2 (4) Na (meq./l.) 216 4 (6) 276 2 (4) K (meq./l.) 4.3 1.7 (6) 6.4 0.3 (4) Total P (gm./l.) 0.55 0.05 (5) NPN (gm./l.) 1.3 0.2 (5) 1.7 0.2 (4) 0.9 (1)* Urea-N (gm./l.) 0.4 db 0.1 (3) Amino-N (gm./l.) 0.15 0.03 (4) * Data from Scholander ct al. (1957). t Single pooled sample. are summarized in Table I. This chemical identification effort was terminated with the exhaustion of the supply of samples. RESULTS Resistance to freezing in Labrador fishes Data on osmotic concentration of the blood of fishes captured in Labrador in 1959 are included in Figure 1 and Tables II and III. These spring fishes were significantly more concentrated than the summer fishes studied by Scholander ct al. TABLE III Plasma concentrations in codfish Concentrations [x S. E. (N) ] Substance Gadus ogac (Labrador) Mtcrogadus tomcod'f (New Brunswick) Summer, 4-7 C. Spring. -1.7 C. Spring. 4 C. Spring, 1.5 C. A (mOsm./l.) 430 10 (6)* 505 db 10 (5) 440 10 (14) 525 5 (5) 790 10 (8)* Cl (meq./l.) approx. 200* 243 19 (3) 142 2 (14) 166 3 (5) Na (meq./l.) 216 4 (3) 231 3 (14) 246 1 ; K (meq./l.) - 5.5 1.4 (3) 5.1 1.4 (14) 8.3 0.3 ( 5 ) Total P (gm./l.) 0.72 0.03 (2) NPN (gm./l.) 4.0 0.2 (3) 1.0 0.2 (ID 1.3 0.2 (5) Urea-N (gm./l.) 0.7 0.1 (1) Amino-N (gm./l.) 1 0.25 0.02 (3) ' 1 * Data from Scholander et al. (1957). t Analyses on three pooled samples for 4 C. fish, one pooled sample for 1.5 fish. 56 MALCOLM S. GORDON, BEN H. AMDUR AND P. F. SCHOLANDER The temperature of the water from which these fishes were taken varied from 1.68 to 1.81 C. The body temperatures of fresh-caught fishes (measured within 30 seconds of their removal from the water) were uniformly - 1.50 C. for each of three sculpins, -- 1.50 to -- 1.75 C. for five cod. 1.50 - 1.00- u o 0.50- 0.00 WATER 7 22, ^ s v M. scorpius G. ogac M. tomcod 1 1 L. VV H \ t~ 1 ~~\ 1 :| | 'I i ex; 1 3d SUMMER SPRING SPRING SUMMER SPRING SPRING SPRING 4-7C. -I.7C. 4 C 4-7C -I.7C. 4 C. -I.5C. *~ ~" NEW "- ~" "~ ~* LABRADOR BRUNSWICK LABRADOR NEW BRUNSWICK FIGURE 1. Blood serum concentrations of major constituents in groups of variously thermally acclimatized fishes of three northern species : Myoxocephalus scorpius. Gadits ogac and Microgadus tomcod. Measured concentrations converted to equivalent freezing point depression using 1 M=l.S6 C. Total phosphorus (total P) and non-protein nitrogen (NPN) freezing point depressions calculated assuming one P or one N atom per molecule, respectively. Analyses for each component carried out on samples from 1-17 fishes in each group (cf. Tables II and III). Horizontal arrows alongside two M. scorpius bars and one G. ogac bar indicate meas- ured freezing point depressions from 1959 (Labrador) and 1960 (New Brunswick) expedi- tions. Area above arrow outlined by dashes is, for spring Labrador M. scorpius, average measured freezing point depression from 1957 (Labrador) expedition. For spring New Brunswick M. scorpius and spring Labrador G. ogac similar areas above arrows indicate freezing point depression in excess of measured values calculated from chemical data. FREEZING RESISTANCE IN NORTHERN FISHES 57 It is evident that all of these fishes were supercooled by small but significant amounts an average of 0.25 C. for the sculpins, about 0.75 C. for the cod. We therefore carried out experiments on the ease with which freezing could be induced in these fishes by seeding them with ice. Three sculpins, damaged only very slightly by being jigged, were kept in the sea water in the fishing holes cut in the ice and observed for signs of freezing due to contact with the walls of the ice hole and with the ice particles carried over their gills by their respiratory movements. One sculpin died, showing ice crystals in one eye, within 30 minutes. The second sculpin died similarly after about one hour. The third sculpin showed ice crystals in one eye transitorily between 30 and 50 minutes after capture, then showed no further visible signs of freezing until its death after four hours. Death was unaccompanied by convulsions in these fishes. Three cod similarly kept in an ice hole survived with no apparent freezing until observations ceased after 30 minutes. Five other cod handled similarly survived over an observation period of two hours. A sixth fish, seeded with snow and ice rubbed into a small incision in the skin of the back, died after about an hour with no visible signs of freezing. No convulsive movements were noted in any of these fishes. Note should be made of the differences in mean osmotic concentration of the blood of the sculpins and cod taken in 1959 and those studied earlier by Scholander ct al. We are sure that these differences are real, as the same procedures and some of the same people were involved in making both sets of analyses. The 1959 fishes were apparently significantly lower in internal osmotic concentration than were the fishes of the same species living in the same area two years earlier. Thus, while all shallow-water fishes seem to add some antifreeze to their blood during the winter, the amount added appears variable. Resistance to freezing in Nezv Brunswick fishes Data on osmotic concentration of the blood of New Brunswick short-horned sculpins and tomcod are presented in Figure 1 and Tables I and II. The internal osmotic concentrations of these fishes were almost the same as the osmotic con- centrations of Labrador sculpins and cod in summer, even though the New Brunswick fishes were sampled at the same time of year as were the springtime Labrador fishes. The addition of antifreeze to the blood of the sculpin, also prob- ably the fjord cod, therefore appears to be a true response to low temperature and not simply a seasonal phenomenon. It is possible, however, that the New Bruns- wick fishes are physiologically different from the Labrador fishes. New Brunswick fishes survived with difficulty when subjected to water tem- peratures as low as the environmental temperatures easily endured by Labrador fishes at the same -time of year. At least the sculpins, however, probably can greatly improve their tolerance through gradual acclimatization. Eighteen sculpins were transferred from sea water at 4 to sea water at - - 1.5, either directly or with some acclimation over periods longer than 12 hours. Seven- teen of these froze upon coining in contact with bits of ice, whether this contact occurred immediately or after periods of up to 150 hours at - 1.50. Only one fish, after 100 hours at - 1.5, showed no signs of distress and did not freeze MALCOLM S. GORDON, BEN H. AMDUR AND P. F. SCHOLANDER when vigorously rubbed with bits of ice at intervals over a period of an hour. This fish froze immediately, however, when cooled to 1.7 in a bucket of sea water. It therefore seems that only a small fraction of the New Brunswick sculpin population (one of 18 fishes tested ) is able to develop resistance to freezing similar to that possessed by the entire Labrador fish population. More gradual acclima- tion might have increased this proportion somewhat, but it seems probable that New Brunswick sculpins are physiologically different from their arctic relatives on a population basis. Somewhat similar observations by Eliassen ct al. (I960) on supercooled Norwegian Coitus scorpius (probably the same species as M. scorpius) indicate the existence of the same situation in non-arctic eastern Atlantic fishes as well. Though closely related, the tomcod seems to be much less resistant to low temperatures and freezing than G adits ogac. Only six fishes, of 42 cooled from 4 to - 1.5 over periods of 12-48 hours, survived unfrozen at temperatures lower than - 1.2. One of these six survivors froze and died after 7 days at - 1.5. The other five survived until the experiment was terminated with removal of blood samples after 9^ days. It seems probable, from the freezing point of the blood of these surviving fishes (Figure 1 and Table III), that the reason for their survival was chance avoidance of any contact with the few small bits of ice which formed in their tank. The surviving tomcod at - - 1.5 showed three differences from controls main- tained at 4. These were: (a) volume of blood obtainable by both cardiac and caudal artery puncture about four times larger in the warm- as opposed to the cold-acclimatized fish; (b) the blood of the cold-acclimatized fish appeared to have a higher hematocrit and clotted much more rapidly at room temperature than the blood of the warm-acclimatized fish; (c) the stomachs and intestines of only the five cold-acclimatized fish were swollen with what appeared to be sea water which the fish had drunk. These differences between the high- and low-temperature groups of tomcod, combined with the higher blood concentrations of the low temperature fishes (Table III), make it reasonable to infer that the temperature of 1.5 produced an osmoregulatory disturbance in the tomcod. This disturbance may have been due to an increase over normal levels of the permeability to water of the gill mem- branes or integument, or may have been a result of a slowing of rates of solute excretion. Chemical nature of the antifreeze substances in Labrador fishes The general comment summarizing the results of the chemical analytical work carried out is that we have not been able to specifically identify the antifreeze sub- stance in either the sculpin or the cod. We have, along with Eliassen et al. (1960), verified the fact that the winter (low temperature) increase in blood concentration in both forms is not due to increased concentrations of NaCl. We have also eliminated from further consideration many classes of possible compounds and have some indications as to the directions in which work should go when addi- tional material becomes available. The sum of blood Na, Cl, and K concentrations in New Brunswick sculpins at 4 was almost the same as the sum of these same FREEZING RESISTANCE IN NORTHERN FISHES 59 concentrations in the blood of Labrador sculpins at - 1.7 (466 and 454 meq./L, respectively, Fig. 1 and Table II J. The osmotic concentrations of these two groups of blood samples differed, however, by over 200 mOsm./l. The picture in the fjord cod is not so clear, due to the lack of complete data from cod at higher temperatures. However, plasma Cl concentrations in winter cod were no more than 40 meq./l. higher than in summer fish. Osmotic concen- trations differed by 75-360 mOsm./l. An important biochemical difference between the short-horned sculpin and the fjord cod, a difference probably indicative of the nature of the antifreeze sub- stance in the cod, is that shown by the non-protein nitrogen values for each form. Both species seem to have a great deal more NPN in their blood than do most other teleost fishes (NPN concentrations in the blood of many species of marine teleosts are in the range 0.04-0.73 gm./l. (Cordier and Chanel, 1958; Denis, 1922, Drilhon, 1952; Jonas and MacLeod, 1960), with only the Japanese eel in summer reaching a level as high as 1.25 gm./l. (Kawamoto, 1929)). In addition, however, the fjord cod seems to have about three times as much NPN as does the sculpin. Assuming all NPN substances in both species are in the form of molecules con- taining only one N atom, this fraction could supply more than enough solute to account for the wintertime increase in osmotic concentration in the fjord cod. It thus seems possible that the antifreeze of the fjord cod is a part of the NPN frac- tion. This is. however, not true for the sculpin, and NPN levels do not seem to vary significantly with temperature in this latter species. Osmotically significant amounts of the compounds and groups of compounds listed in Table I were absent from the fjord cod blood samples tested. All amine tests were negative, even when samples had been treated with powdered zinc in order to reduce any oxides which may have been present (e.g., trimethylamine oxide). A point of interest concerning the fjord cod samples is the identity of the commonest free amino acids which were present. These were : aspartic and glutamic acids, threonine and monoiodotyrosine. Small quantities of samples were used in all of these analyses, so it is probable that acids present in very low con- centrations were missed. The identity of the sculpin antifreeze is presently completely obscure. There is a distinct possibility that no one compound is the antifreeze. If it is a single substance, it is apparently not a part of the NPN fraction and is probably not a phosphorus-containing compound. Other compounds and groups of compounds tested for but not detected in osmotically significant amounts are listed in Table I. Analysis of the amino acids present in sculpin blood showed alanine, methionine and taurine to be present in largest quantities. Chemical responses to loiv temperatures in Ne-w Brunswick tomcod It is interesting to note that plasma NPN concentrations in tomcod maintained at - 1.5 apparently increased by a proportionately larger amount than did the freezing point depression of the blood or the concentrations of the commonest inorganic ions. Plasma NPN levels rose about 30% while osmotic concentration increased only about 20%, chloride increased about 15% and sodium increased about 6% (Fig. 1 and Table III). The compounds involved in this increase in 60 MALCOLM S. GORDON, BEN H. AMDUR AND P. F. SCHOLANDER concentration are unidentified. There is a striking similarity between the torn- cod's response to low temperature and that shown by the fjord cod. DISCUSSION The degree of supercooling occurring in all arctic fishes studied to date is quite small. However, even this slight degree of supercooling carries with it significant danger of fatal freezing if seeding with ice should occur. In order for the body fluids of the Labrador sculpins to come to thermodynamic equilibrium with the fishes' own body temperatures, approximately 20% of their free body water would have to freeze. For the fjord cod the equivalent figure would be about 40-50% of free body water. We have no way of estimating how much, if any, of the body water actually did freeze in the Labrador sculpins and fjord cod tested for survival in the presence of ice. All we know is that these fishes did survive for periods of hours in contact with ice. There were none of the usual visible signs of freezing which always occur almost immediately in non-arctic fishes subjected to similar treat- ment (cf. Scholander ct al., 1957). It therefore seems probable that if there was any freezing of body fluids, it occurred slowly and probably did not spread very far through the tissues of the fishes' bodies. A similar situation apparently existed in some of the fishes studied by Eliassen ct al. (1960). Several theoretical alternatives seem reasonable as possible explanations for these observations. First, one might postulate that the skin, gills, etc., of the Labrador sculpins and cod are less open to penetration by ice crystals than the integuments of non-arctic fishes were shown to be by Scholander ct al. (1957). Second, it is possible that the skins of arctic fishes are no less penetrable by ice than are those of non-arctic forms, but that their body fluids possess special proper- ties. One such property might be a significant slowing of ice crystal growth. The antifreeze substances themselves might confer this property, as do glycerol and other alcohols and various sugars, also some proteins in pure solutions (Lusena, 1955). Another special property might be the occurrence of larger amounts of bound water around tissue proteins, etc., than are present in non-arctic forms. A third possibility is that neither of the above suggestions is correct, but that instead the spread of ice crystals through the bodies of our fishes was inhibited by the cell membranes of the fishes' tissues, and no damaging intracellular freezing occurred. The cell contents in this circumstance would have to tolerate some dehydration. Possible support for this idea comes from the observations of Cham- bers and Hale (1932) on the efficiency of frog sarcolemmae and amoeba cell membranes as barriers to propagation of ice crystals. In both of the last two situations it would seem possible for fishes which had been seeded by a chance encounter with ice, as perhaps in an effort to escape a pursuing predator, to rid themselves of such ice as may form internally by becoming physically active enough temporarily to raise their body temperatures by a few tenths of a degree. The body temperatures we measured on the Labrador sculpins indicate that they may generally be a few tenths of a degree warmer than their environment. Observations by Britton (1924) indicate that similar small differ- ences between body and ambient temperatures may be a year-round feature of Al. scorpms even in non-arctic areas. FREEZING RESISTANCE IN NORTHERN FISHES 61 SUMMARY 1. The occurrence of small degrees of supercooling and the presence in the blood of organic antifreeze compounds are confirmed in arctic populations of short-horned sculpins (Myo.roccphalus sc or phis) and fjord cod (Gadiis ogac) captured at Hebron Fjord, Labrador, in early spring. The quantity of antifreeze added seems variable, however. 2. Although significantly supercooled, these same fishes were found to be very resistant to freezing even though seeded with ice. Possible explanations for this resistance are discussed. 3. Non-arctic populations of the sculpin and of the tomcod (Microgadus toni- cod, a close relative of the fjord cod) also studied in early spring were found to lack both the resistance to freezing when supercooled and also the antifreeze sub- stances found in the arctic fishes. Very few of the non-arctic sculpins were able to develop any resistance to freezing even after several days' exposure to arctic water temperatures. 4. The antifreeze substance added to the blood of the fjord cod is indicated to be a member of the non-protein nitrogen fraction. There is no evidence that it is an amino acid, an amine or an amine oxide. 5. The antifreeze substance added to the blood of the short-horned sculpin is also unidentified. It apparently contains neither nitrogen nor phosphorus, is not glycerol or a related alcohol and probably is not a reducing or a non-reducing sugar, an aldehyde or ketone, a carboxylic acid (lipid or other), an ester or an aromatic compound. 6. Tomcod exposed to arctic water temperatures show increases in plasma non-protein nitrogen levels similar to those which occur in the fjord cod in winter. LITERATURE CITED BLOCK, R. J., E. L. DURRUM AND G. ZWEIG, 1958. A Manual of Paper Chromatography and Paper Electrophoresis. Academic Press, N. Y., 2nd ed. BRITTON, S. W., 1924. The body temperature of fishes. Contr. Canad. Biol. N. S., 1 : 413-418. CHAMBERS, R., AND H. P. HALE, 1932. Formation of ice in protoplasm. Proc. Rov. Soc. London, HOB: 336-352. CHEN, P. S., JR., T. Y. TORIBARA AND H. WARNER, 1956. Microdetermination of phosphorus. Anal. Chcm., 28: 1756-1757. CHERONIS, N. D., AND J. B. ENTRIKIN, 1957. Semimicro Qualitative Organic Analysis. Inter- science, N. Y. CORDIER, D., AND J. CHANEL, 1958. Variations du taux de 1'azote non proteique total, de 1'azote non proteique non polypeptidique et de 1'azote polypeptidique dans le serum de la Rascasse (Scorpaena porcus L.) sous 1'influence de 1'asphyxie. C. R. Soc. Biol., 152: 787-790. DENIS, W., 1922. The non-protein organic constituents in the blood of marine fish. /. Biol. Chcm., 54: 693-700. DRILHON, A., 1952. Acides amines libres des sangs des poissons : etude chromatographique. C. R. Acad. Sci., 234: 466-469. DUKE, F. R., 1945. Metallo-organic complexes in organic analysis qualitative tests for amines. Ind. Eng. Chcm., Anal. Ed.. 17: 146-149. ELIASSEN, E., H. LEIVESTAD AND D. MOLLER, 1960. The effect of low temperatures on the freezing point of plasma and on the potassium/sodium ratio in the muscles of some boreal and subarctic fishes. Arbok Univ. Bergen, Mat-Natm~i>. Scrie, No. 14: 1-24. FEIGL, F., 1947. Qualitative Analysis by Spot Tests. Elsevier, N. Y., 3rd cd. 62 MALCOLM S. GORDON, BEN H. AMDUR AND P. F. SCHOLANDER FEIGL, F., 1956. Spot Tests in Organic Analysis. Elsevier, N. Y., 5th ed. JONAS, R. E. E., AND R. A. MACLEOD, 1960. Biochemical studies on sockeye salmon during spawning migration. X. Glucose, total protein, non-protein nitrogen and amino acid nitrogen in plasma. /. Fish. Res. Bd., Canada, 17: 125-126. JONES, C. N., 1947. Report on the detection of preservatives and bacteriostatic agents in ampul solutions. /. Assoc. Offic. Agr. Chemists, 30: 486-488. KAWAMOTO, N., 1929. Physiological studies on the eel. 1. The seasonal variation of the blood constituents. Sci. Kept. Tohoku Imp. Unh>., Scr. 4, 4: 635-642. LUSENA, C. V., 1955. Ice propagation in systems of biological interest. III. Effects of solutes on nucleation and growth of ice crystals. Arch. Biochem. Biophvs., 57: 277-284. NATELSON, S., 1957. Microtechniques of Clinical Chemistry for the Routine Laboratory. C. C. Thomas, Springfield. RAMSAY, J. A., AND R. H. J. BROWN, 1955. Simplified apparatus and procedure for freezing- point determinations upon small volumes of fluid. /. Sci. Instr., 32 : 372-375. SCHAFFERT, R. R., AND S. R. KiNGSLEY, 1955. Determination of reduced dehydro- and total ascorbic acid in biological material. /. Biol. Chem., 212: 59-68. SCHERBAUM, O. H., T. W. JAMES AND T. L. JAHN, 1959. The amino acid composition in relation to cell growth and cell division in synchronized cultures of Tetrahymena pyrijormis. J. Cell. Comp. Physiol, 53: 119-138. SCHOLANDER, P. F., L. VAN DAM, J. KANWISHER, T. HAMMEL AND M. S. GORDON, 1957. Supercooling and osmoregulation in arctic fish. /. Cell. Comp. Physiol., 49 : 5-24. SHRINER, R. L., AND R. C. FUSON, 1948. The Systematic Identification of Organic Com- pounds. John Wiley, N. Y., 3rd ed. SMITH, F. J., AND E. JONES, 1948. A Scheme of Qualitative Organic Analysis. Blackie, London. A HISTOCHEMICAL STUDY OF DIGESTION AND DIGESTIVE ENZYMES IN THE RHYNCHOCOELAX LINEUS RUBER (O. F. MULLER) J. B. JENNINGS Department of Zoology, The University of Leeds, England It has been shown in a previous account (Jennings, 1960) that digestion in the rhynchocoelan Linens ruber is the result of both extracellular and intracellular processes. The food, which consists of animals such as small annelids and crus- taceans captured by means of the eversible proboscis, is swallowed whole and, after being killed by acid secretions poured on to it during its passage through the foregut, is broken down into a semi-fluid mass within minutes of arrival in the intestine. The enzyme bringing about this initial, extracellular, breakdown is produced by gland cells scattered throughout the intestinal gastrodermis and operates at a pH of 5.0-5.5. The fragmenting food is phagocytosed by other gas- trodermal cells and digestion completed intracellularly. In the present work the course of digestion has been examined in greater detail and an attempt made to identify some of the enzymes concerned in both the breakdown of the food and the general metabolic activity of the gastrodermis. MATERIALS AND METHODS Individual Linens ruber were isolated and starved for seven days to clear the gut of all traces of previous meals. Individual isolation was necessary since after five or six days without food cannibalism often occurs. To study the course of digestion, and to locate and identify the enzymes con- cerned, starved Linens were fed upon inert test foods such as clotted frog blood, either alone or mixed with cooked beef fat or starch paste. These foods were used in preference to the natural living food to eliminate any possibility of enzymes contained in the latter being mistaken for those produced by the Linens gut. Series of Linens were killed for examination after seven days' starvation and at progressive intervals up to 48 hours after an observed meal on one or other of the test foods. In earlier experiments the Lineus were killed by freezing in isopentane cooled by liquid nitrogen to - 160 C. and subsequently dehydrated for 48 hours at - 40 C. under a vacuum of 10~ 3 mm. mercury. Such specimens were embedded in paraffin wax (melting point 42 C.), sectioned at 8 p. and examined by the histochemical techniques listed below. During the course of the work, however, it was found that specimens fixed for 12 hours at 4 C. in 10% formalin in sea water, buffered to pH 7.0, and then rapidly dehydrated in absolute acetone at the same temperature, cleared in xylol at room temperature and embedded in 42 C. wax showed no significant decrease in enzyme activity when compared with the freeze dried specimens. The duration of dehydration, clearing and 63 64 J. B. JENNINGS embedding was kept to the absolute minimum consistent with the size of the specimen. This method of preparing sections for histochemical examination was adopted for the bulk of the work as it enabled many more specimens to be dealt with in a given time and W 7 as far less troublesome in operation. The sections were mounted with albumen and after drying at 20 C. for six hours were dewaxed in xylol and passed through two changes of absolute acetone before being transferred directly into the various reagents for visualising enzyme activity. Carbonic anhydrase activity, known to be associated with production of hydro- chloric acid in the mammal stomach, was visualised by the cobalt sulphate-bi- carbonate method given by Hausler (1958). Sections were incubated for three hours at 20 C. and it was found that for optimum results the layer of substrate solution upon the sections must not exceed 1 mm. in depth. The medium was buffered to pH 8.0 and control sections incubated in the presence of 4 X 10~ 3 M Diamox sodium (2 acetylamino-l,3,4-thiadiazole-5-sulfonamide sodium), a specific inhibitor for carbonic anhydrase. As a further control, and to check the reliability of the method, sections of mouse stomach were similarly treated. Proteolytic enzymes were investigated by the methods of Hess and Pearse (1958) for cathepsin C type enzymes and Burstone and Folk (1956) for leucine aminopeptidase. To detect cathepsin C type activity, sections were transferred from acetone into a 10~ 5 M solution of E-600 (diethyl-p-nitrophenyl phosphate) to inactivate esterases whose presence would otherwise give false positive reactions. The sections were then incubated for three to four hours at 20 C. in a standard indoxyl acetate medium buffered at pH 5.0. Control sections were immersed for one hour, before incubation, in 1 X 10~ 3 M cysteine solution which activates cathep- sin C so that sections treated in this way show a more intense enzymatic action than do non-treated ones. Further controls were performed by immersing sections in 1 X 10~ 3 M lead nitrate solution for one hour, or in water at 90 C. for three minutes, before incubation. For leucine aminopeptidase activity sections w r ere in- cubated for six hours at 20 C. in a medium containing L-leucyl-/3-naphthylamide as substrate and Garnet G.B.C. as a simultaneous coupler. The medium was buffered to pH 7.2 and heat-inactivated sections used as controls. Lipolytic activity was investigated by the method of Gomori (1952) using Tween 80 as the substrate in a medium buffered to pH 7.2. Sections of Linens fed on blood mixed with beef fat were incubated for twelve hours at 20 C. and again heat-inactivated sections were used as controls. Attempts were made to detect carbohydrase activity by using the ferric-8- hydroxyquinoline method for /3-glucuronidase as modified by Billett and McGee- Russell (1955). This method failed to give satisfactory results and observations on carbohydrate digestion were limited to tracing the fate of starch meals by the Lugol's iodine technique. Alkaline phosphatase activity was demonstrated by the calcium phosphate method (Gomori, 1952). Sections were incubated for two hours at 20 C. in a medium containing sodium /?-glycerophosphate buffered to pH 8.0 and control sections inactivated by heat prior to incubation. Sections were also examined for acid phosphatase activity (Gomori. 1952), again using sodium /3-glycerophos- phate as substrate. DIGESTION IN LINEUS RUBER 65 OBSERVATIONS Histology of the gut The histological structure of the gut in Linens has been described in detail elsewhere (Jennings, I960). Briefly, the gut consists of three regions, the mouth and buccal cavity, the foregut, and the intestine. It is ciliated throughout its length and lacks both multicellular glands and musculature. The buccal cavity is lined by ciliated cuboidal cells backed by masses of acidophil and basophil gland cells, the majority of which are Alcian blue- and periodic acicl-Schiff-positive and produce mucus to facilitate ingestion. The foregut is lined by similar ciliated cells, interspersed with acidophil gland cells, lying upon tissue with indistinct cell walls containing numerous acidophil and basophil glands, many of which are mucus- producing. The intestine forms the major portion of the gut and bears paired serially repeated lateral pouches throughout its length. The intestinal wall, or gastrodermis, is made up of two types of cells which stand in a single layer upon a thin basement membrane. The larger and more numerous cells are columnar and the cytoplasm usually contains phagocytosed food particles in various stages of digestion. The second type of cell is glandular and contains up to 30 acidophil proteinaceous spheres which are discharged into the gut lumen when food enters. Enzymes produced in the gut Carbonic anliydrasc Carbonic anhydrase activity was found in some 10 I5 c /r of the acidophil gland cells of the buccal cavity and foregut (Fig. 1 ). In the latter, gland cells in both the lining epithelium and the backing syncytial tissue gave positive reactions. The number of reactive cells, and the degree of activity, could not be related to the time of feeding and it would appear that the enzyme is always present in an active condition. Control sections incubated in the presence of the specific inhibitor, Diamox sodium, gave no reaction. Sections of mouse stomach incubated at 37 C. showed intense activity in the acidophil oxyntic cells of the fundic glands, which are known to be the source of the gastric hydrochloric acid. This activity, as in Linens, \vas inhibited by Diamox sodium. Cathepsin C During starvation the gland cells of the gastrodermis give an intense positive reaction to the Hess and Pearse method for cathepsin C (Figs. 2 and 3). The individual spheres within the gland cells stained so intensely that often other details of the cell, such as the nucleus, were obscured. Sections immersed in 10~ 3 M cysteine solution, a known activator of cathepsin C type enzymes, before in- cubation, reached the maximum density of staining in approximately half the incubation time needed by non-activated sections. Sections inactivated by heat or by immersion in 10 '' M lead nitrate solution gave no reaction whatsoever. Sections prepared within 30 minutes of feeding showed that the majority of the gland cells had discharged their spheres and such cells failed to give any reaction. This condition ( Fig. 4 ) persisted for two to three hours and then the 66 J. B. JENNINGS FIGS. 1-7. DIGESTION IN LIXKUS RUBER 67 gland cells gradually filled up again with active spheres so that about six hours after feeding they were back in their normal condition. After a meal of frog blood haemolysis occurred as the food entered the intestine, and within 30 minutes the digesting mass in the lumen gave a fairly strong reaction for cathepsin C. About this time the columnar cells of the gastrodermis commenced phagocytosis of the food and the phagocytosed material continued to give a cathepsin reaction within the cells (Fig. 4). This, however, decreased rapidly with time and it was clear that the reaction was the result of enzyme being phagocytosed with the food and remaining active for a short time within the cells. There was no evidence of intracellular production of cathepsin C. One and a half hours after feeding, the gastrodermis was loaded with phagocytosed material and the apparently intra- cellular cathepsin activity had disappeared. The optimum pH for visualisation of the enzyme was 5.0 and this agrees with previous in 1'k'o observations using indicator-stained food, which showed that the early stages of digestion went on at pH 5.0-5.5 (Jennings, 1960). Lend nc aminopeptidase Sections of starved Linens, and of specimens fixed within one and a half hours of feeding, gave no reaction for leucine aminopeptidase. About two hours after feeding, however, phagocytosed material started to show a faint positive reaction. The intensity of the reaction increased rapidly with time, and six hours after feeding all the material in the columnar cells of the gastrodermis gave an extremely strong and vivid reaction (Fig. 5). Heat-inactivated control sections showed no such activity. At no time was any leucine aminopeptidase activity found in either FIGURE 1. Longitudinal section of the foregut in Linens, showing carbonic anhydrase activity in the gland cells of the ciliated epithelium. Hausler cobalt sulphate-bicarbonate method. Scale : 1 cm. = 50 /u. FIGURE 2. Transverse section of an intestinal pouch in a starved Linens, showing the positive cathepsin C type reaction given by the gland cells of the gastrodermis. The gland cells are seen as dark streaks. Hess and Pearse E600-indoxyl acetate method. Scale : 1 cm. = 50 M. FIGURE 3. Transverse section of a portion of the gastrodermis of a starved Linens, show- ing two gland cells packed with enzymatic spheres giving an intense cathepsin C type reaction. Hess and Pearse method. Scale : 1 cm. = 25 M- FIGURE 4. Longitudinal section of the gastrodermis in Linens 30 minutes after feeding. Hess and Pearse method. The gland cells have discharged their spheres and are no longer apparent by this technique. The columnar cells have commenced phagocytosis, and the newly engulfed food, seen in the distal regions of the cells, shows cathepsin C type activity retained from the initial extracellular phase of digestion. Scale : 1 cm. = 50 M- FIGURE 5. Transverse section of two intestinal pouches in Linens six hours after feeding. The gastrodermis is swollen and loaded with food vacuoles all exhibiting intense leucine aminopeptidase activity. Burstone and Folk L-leucyl-/3-naphthylamide G.B.C. method. Scale : 1 cm. = 50 p. FIGURE 6. Transverse section of an intestinal pouch in Linens four hours after a meal of frog blood and beef fat. The gastrodermis is loaded with food vacuoles, many of which show lipolytic activity (seen as black spheres). Gomori Tween 80-lead sulphide method. Scale : 1 cm. = 50 /*. FIGURE 7. Transverse section of a portion of the Linens gastrodermis within 5 minutes of feeding, showing intense alkaline phosphatase activity around the gland cells. The gut lumen (top right) contains haemolysing frog erythrocytes. Gomori calcium phosphate method. Scale : 1 cm. = 25 n. 68 J. !'>. JENNINGS the gland cells of the gastrodermis or the gut lumen, and this enzyme would appear to be entirely intracellular and to be concerned only in the later stages of digestion. The optimum pH for the visualisation was 7.2 and this difference in pH optima between the lumen-acting cathepsin C and the intracellular amino- peptidase is probably the reason why aminopeptidase activity does not commence immediately phagocytosed material enters the cells. It has been seen that newly phagocytosed food continues to show cathepsin C activity for over one hour after feeding and so is presumably still at, or near, the acidic pH value necessary for this. The decrease of cathepsin C activity and its gradual replacement by amino- peptidase will be accompanied by an increase in pH value up to the optimum 7.2 and this apparently takes a little time to achieve. Leucine aminopeptidase activity continues whilst food remains in the gastro- dermis and finally disappears 9 to 12 hours after feeding, depending upon the amount of food taken. It would appear, therefore, that this enzyme is normally present in an inactive form, unlike the cathepsin C of the gland cells, and only becomes active (i.e., as shown by the present method) when it is secreted from the cytoplasm into a food vacuole. Li pas e Small amounts of lipolytic activity were found in the gastrodermis and paren- chyma during all stages of starvation. These were probably the result of the starved Linens utilising its fat reserves which are laid down in these sites (Jen- nings, 1960). No significant increase in activity occurred until two to three hours after a meal of blood and beef fat. About this time a few food vacuoles resulting from phagocytosis of the meal showed a positive reaction. The number of such vacuoles then increased rapidly and within a further hour they could be found throughout the gastrodermis ( Fig. 6 ) . As intracellular digestion progressed the number of reactive vacuoles diminished but some could be found up to 48 hours after feeding. This was a consequence, no doubt, of the unusually large proportion of fat in the meal. The optimum pH for the reaction was 7.2-7.4 and again, as in the case of aminopeptidase, this is probably the reason for the time lag between the onset of phagocytosis and the appearance of lipolytic activity. Control sections inactivated by heat showed no lipolysis. Carbohydrate actirit\ Attempts to localise /8-glucuronidase activity by the Killett and McGee-Russell method during starvation and after meals of blood and starch paste were unsuc- cessful. The end product of the histochemical reaction, ferric-8-hydroxyquinoline, was precipitated indiscriminately over the entire sections, both experimental and control. Sections stained with Lugol's iodine, however, showed progressive diges- tion and disappearance of phagocytosed starch parallel in time with the amino- peptidase and lipase activity, and it was concluded that the unknown carbohydrases act at a similar slightly alkaline pH. There was no extracellular carbohydrate digestion. DIGESTION IN LINEUS RUBER Alkaline plwspliatase The gland cells of the gastrodermis normally show no alkaline phosphatase activity but immediately after the Linens has fed, when they are discharging their enzymatic spheres into the gut lumen, intense alkaline phosphatase activity appears around the cell walls (Fig. 7). This localised activity disappears during recon- stitution of the gland cells and so appears to he associated with their secretory rather than recovery phase. The columnar cells of the gastrodermis show at all times intense alkaline phos- phatase activity along their free ciliated borders (Fig. 8). The zone of activity varies in depth from a narrow band 3-5 /JL deep immediately below the cilia, to a broad belt which mav extend down into the cells for as much as one-third of 8 FIGURE 8. Longitudinal section of a portion of the foregut (left) and intestine (right) of a starved Linens. Gomori method for alkaline phosphatase. The cells of the foregut show no activity, in marked contrast to those of the intestinal gastrodermis which show intense activity distally. Scale : 1 cm = 50 //. FIGURE 9. Transverse section of an intestinal pouch of Linens four hours after feeding. All the food vacuoles and some of the cytoplasm surrounding them show intense alkaline phosphatase activity. Gomori method. Scale : 1 cm. = 50 fj.. their depth. This activity is obviously concerned with some aspect of the digestive function of the gastrodermal cells, for it is completely absent from the cells of the foregut wall (Fig. 8). The distal band of alkaline phosphatase activity persists during extracellular digestion but as phagocytosis and intracellular digestion advance, it spreads down- wards until the entire cytoplasm of the columnar cells gives a positive reaction. The activity is more concentrated around and within the food vacuoles ( Fig. 9 ) and reaches its peak at about the same time as aminopeptidase and lipase activity. Thus it would appear to be connected with the secretion of these enzymes and the absorption of the products of intracellular digestion. This condition persists until digestion is completed and then the activity decreases in amount until only the normal distal zone remains. . Icid phosphatase No trace of acid phosphatase activity was found in any region of the gut 70 J. B. JENNINGS DISCUSSION Whilst it is clear that many more enzymes than the ones located in this work will be concerned in digestion in Linens, sufficient information has been obtained for the sequence of digestive processes and the part played by the different types of enzymes to be understood. The presence of carl tonic anhydrase in acidophil gland cells of the buccal cavity and foregut would suggest that these cells are the source of the acid secre- tions poured on to the food during ingestion to assist in killing it and to provide a medium of suitably low pH value for the initial stages of digestion, taking into account the known association of this enzyme with hydrochloric acid secretion in the oxyntic cells of the mammal. The inhibition of the enzyme by the specific inhibitor for carbonic anhydrase and the similar results obtained with mouse oxyntic cells leave little doubt as to its identity. It was suggested previously that certain basophils in the foregut wall produced the acid secretions (Jennings. 1960) but since these fail to show carbonic anhydrase activity it may be that they have some other, unknown, function. The early, extracellular, stages of digestion are effected by the proteolytic en- zyme discharged from the gland cells of the gastrodermis. This has been identified as cathepsin C, or, at least, a cathepsin C t\[>c enzyme, and the identification con- firmed by use of specific activators and inhibitors. Cathepsin C is an endopeptidase and attacks inner portions of protein chains to break down the molecule into simpler polypeptides and peptides. Thus its function in Linens is to initiate proteolysis and break down the food to a condition suitable for entry into the gut cells where digestion is completed. In this respect the enzyme has an adaptive significance comparable to that of the elaboration of the feeding mechanism in the triclad Turbellaria, where purely mechanical means are used to make the food available for phagocytosis, and extracellular digestion does not occur (Jennings, 1957). In Linens, and presumably most rhynchocoelans. a simpler type of feeding mecha- nism means that the food is swallowed whole and consequently there must be some other provision for its breakdown before intracellular digestion can begin. The intracellular proteolysis appears to be effected by aminopeptidases. of which one example, leucine aminopeptidase, has been identified. These enzymes are exopeptidases and remove terminal amino acids from polypeptides resulting from endopeptidase activity and so complete digestion of the protein content of the food. They function at a slightly alkaline pH, in contrast to the extracellular!) - acting endopeptidase. cathepsin C, which requires a fairly strongly acidic medium. Thus proteolysis in Linens resembles that in most other animals in that it occurs in two distinct phases, the first acidic, the second alkaline. The extracellular and intracellular proteolysis makes the fat and carbohydrate content of the food available for digestion by breaking down tissue and cell mem- branes, and the digestion of these food elements is entirely intracellular. Lipolytic activity appears at about the same time as aminopeptidase and the enzyme re- sponsible operates at a similarly slightly alkaline pH. It is probable that the en- zyme visualized here is the "true lipase" of Gomori (1952), homologous with mammalian pancreatic lipase, since an unsaturated substrate ( Tween 80) was used and this, according to Gomori, is attacked only by pancreatic type lipase and not by other esterases. DIGESTION IN LINEUS RUBER 71 It was not possible to identify any carbohydrases but the intracellular digestion of starch showed that these are present and working at a pH similar to that needed for aminopeptidase and lipase activity. Rosenbaum and Rolon (I960) located /S-glucuronidase activity in the phagocytic gut cells o f the not too distantly related triclad flatworm and it may be that this enzyme is also present in Linens but failed to survive fixation. The intense alkaline phosphatase activity observed in the distal region of the columnar gastrodermal cells appears to be related to the part played by the cell wall and its cilia in the phagocytic uptake of food. The cilia coalesce into pseudo- podia-like processes which engulf the fragmenting food (Jennings, 1960) and the alkaline phosphatase no doubt plays some part in this modification and later re- covery of the cilia. This interpretation is supported by the absence of phosphatase activity from the foregut cells, which are not concerned in uptake of food and whose cilia show no such modification. The alkaline phosphatase activity de- veloping deeper within the gastrodermal cells during intracellular digestion would seem to lie related to secretion from the cytoplasm into the food vacuoles of amino- peptidases. lipase and carbohydrases since these appear simultaneously with the phosphatase. SUMMARY 1. Digestion in the rhynchocoelan Linens rnher is both extracellular and intra- cellular. The extracellular phase is entirely proteolytic and is brought about by an endopeptidase acting in an acid medium. The semi-digested food is then phagocytosed and digestion is completed in the second, intracellular phase by exo- peptidases, lipases and presumed carbohydrases, all operating in an alkaline medium. 2. The following enzymes have been located and identified by histochemical methods: carbonic anhydrase, a cathepsin C type protease (endopeptidase), leucine aminopeptidase ( exopeptidase ) , lipase and alkaline phosphatase. 3. Carbonic anhydrase occurs in acidophil gland cells in the buccal cavity and foregut. It is believed to be associated with production of acid used to kill the food and provide a suitable medium for the extracellular phase of digestion. 4. The cathepsin C type protease is produced by gland cells in the gastrodermis and is discharged into the gut lumen to bring about the initial extracellular proteolysis. 5. Leucine aminopeptidase is produced within the phagocytic cells of the gastro- dermis when food vacuoles are present and is concerned in completion of protein digestion. 6. Lipase, identified as "true lipase" homologous with mammalian pancreatic lipase, is formed within the phagocytic cells at the same time as leucine amino- peptidase and attacks the fat content of the food. 7 '. Alkaline phosphatase is present in an active form in the gastrodermis and appears to be concerned with phagocytosis. The amount present increases when intracellular digestion occurs and this increase is apparently associated with se- cretion of aminopeptidase and lipase into the food vacuoles. LITERATURE CITED BILI.KTT, F., AND S. M. McGEE-RussELt, 1955. The histochemical localisation of /8-glucuronidase in the digestive gland of the Roman snail (Hcli.v pomatiu). Quart. J. Micr. Sci., 96: 35-48. 72 J. B. JENNINGS BURSTONE, M. S., AXD J. E. FOLK, 1956. Histochemical demonstration of aminopeptidase. /. Ilistochcin. Cytochcin., 4: 217-226. GOMORI, G., 1952. Microscopic Histochemistry. Univ. of Chicago Press, Chicago. HAUSLER, G., 1958. Zur Technik und Spezifitat des histochemisehen Carboanhydrasenachweises im Modellversuch und in Gewebsschriitten von Rattennieren. HistocJicmic, 1 : 29-47. HESS, R., AXD A. G. E. PEARSE, 1958. The histochemistry of indoxylesterase of rat kidney with special reference to its cathepsin-like activity. Brit. J. E.\-p. Path., 39: 292-299. JENNINGS, J. B., 1957. Studies on feeding, digestion and food storage in free-living flatworms < Platyhelminthes : Turbellaria ). Biol. Bull.. 112: 63-80. THXNIXGS. J. B., 1960. Observations on the nutrition of the rynchocoelan Linens rubcr ( O. F. Miiller). Biol. Bull, 119: 189-196. ROSEXBAUM, R. M., AND CARMEN I. RoLON, 1960. Intracellular digestion and hydrolytic en- zymes in the phagocytes of planarians. Biol. Bull.. 118: 315-323. NEUROSECRETION AND CRUSTACEAN RETINAL PIGMEXT HORMONE: DISTRIBUTION OF THE LIGHT-ADAPTING HORMONE ' L. H. KLEINHOLZ, P. R. BURGESS, D. B. CARLISLE AND O. PFLUEGER '/'//( Biological Laboratories, Reed College, Portland 2, Oregon; the Marine Biological Labora- tory, ll'oods Hole, Mass.; the Kristineberg Zoological Station, Fiskchiickskil, Sweden Pigmentary effectors of crustaceans, the chromatophores and retinal pigments, have long been known to be regulated by hormonal substances originating in the eyestalks (Perkins, 1928; Kleinholz, 1936; Fingerman ct /., 1959). Localization of the origin of these substances within the eyestalks is less well known. Hanstrom (1937) put these physiological observations on a firmer morphological basis by describing two structures in the crustacean eyestalk that might be in- volved in regulating the integumentary chromatophores : ( 1 ) the X -organ, com- posed of modified sensory cells of a rudimentary eye papilla, and ( 2 ) the sinus gland, usually located dorsally between the two middle optic ganglia. After test- ing extracts prepared from portions of eyestalks containing one or the other of these two glandlike structures, Hanstrom (1937) and Carlson (1936) proposed the sinus gland rather than the X-organ as the source of chromatophorotropic ac- tivity. Initial observations by Brown (1940) comparing activities of sinus gland extracts, of extracts prepared from entire eyestalks, and of extracts prepared from eyestalks from which sinus glands had been removed, indicated that nearly all the chromatophorotropic activity could be accounted for by the sinus gland. Later. however, more quantitative studies showed that extracts prepared from sinus- glandless eyestalks or from the four optic ganglia were as active on crustacean chromatophores as those prepared from sinus glands alone (Brown, 1950; Sandeen, 1950). Little is known about the localization of the light-adapting retinal pigment hormone. Welsh (1941 ) found extracts of the sinus gland and of the medulla tenninalis to be active, while those of supraesophageal ganglia were not ; the ac- tivity of medulla terminalis extracts was explained as possibly due to residual tissue from the sinus gland, or to material that had escaped from the sinus gland during its removal from the eyestalk tissue. Kleinholz (1958), however, found little or no retinal pigment activity in sinus gland extracts in some species, while substantial amounts of this hormone were found in components of the eyestalk other than the sinus gland, and in other ganglia of the central nervous system. Cytological examination of the sinus gland had disclosed a surprising lack of cellular organization in what was supposedly an active endocrine gland. Largely as a result of the studies of Passano (1953) and of Bliss and Welsh (1952) a revised explanation of the morphological relationships of the sinus gland was pro- 1 Aided by grants to L. H. K. from the National Science Foundation (G-1395, G-3986) and from the National Institutes of Health (B-2606). 73 74 KLEINHOLZ, BURGESS, CARLISLE AND PFLUEGER posed, based on the description of a new X-organ (Carlisle and Passano, 1953). differing in its histology and in its location in the eyestalk from the one originally described by Hanstrom. This new X-organ, located in the fourth optic ganglion, is composed of modified neurons, the axons of which constitute the so-called sinus gland nerve (Bliss and Welsh, 1952; Passano, 1953). Passano has interpreted the sinus gland as consisting of an aggregation of these axonal terminals, swollen by the accumulation of neurosecretory materials originally elaborated by such modified neurons and transported to their final site by axoplasmic flow. Bliss, Durand and Welsh (1954) have reported that axons from neurosecretory cells at other locations in the central nervous system also contribute to the formation of the sinus gland. Tn view of the variety of hormonal functions now attributed to this sinus gland- X-organ complex, and the inadequacy of information on the origin of such hormones, it was believed desirable to undertake such localization studies with the hope that correlation of particular hormones with specific cell types in this anatomical com- plex might subsequently be possible. MATERIALS AND METHODS A variety of decapod crustaceans was used in this study. Donor animals whose eyestalks and eyestalk components were tested for distal retinal pigment hormone included the following: Libinia ananjinata Leach. Callincctcs sapidus Rathbun. Carchuis inacnas Linnaeus, Pandalus borcalis KroVer, A T cf>hrof>s norrec/icits Linnaeus, Hoinanis ainericaiuts Milne Edwards, Orconcctcs t'irilis Hagen, and Calocans macandrcac Bell. The two species on which the various extracts were tested by injection were Palacmonctcs vitlyaris and Palacnwn adspersus.' 2 Particular glandular or ganglionic portions of eyestalks, ablated from light- adapted donors, were separated with the aid of a binocular dissecting microscope. Extracts of such tissues (whole eyestalks, sinus glands, sinusglandless eyestalks. specific optic ganglia, etc. ) were prepared by triturating the pooled components with sand and sufficient solvent (sea water or distilled water) to give a final con- centration of 10 units per 1.0 nil. for most of the tissues. The resulting tissue brei was heated in a boiling water bath for 2-3 minutes and centrifuged to remove extraneous material and coagulated proteins ; in some cases the mixture was centrifuged without such heating and in still others comparison of activity was - The vicissitudes of systematic nomenclature are frequently a source of confusion to ex- perimental biologists. Several colleagues have called our attention to the revised synonomy of the genera with which we have worked. Dr. Fenner A. Chace, Jr. of the U. S. National Museum, Smithsonian Institution, lias advised us that most of the European members of the genus Lcatidcr have been returned to the genus Palaciinni ; L. udspcrsiis is therefore now re- ierred to as Palaemon adspcrsits. Holthius ( 1952) has recently separated from Palacmonctcs vulgaris of the Woods Hole region two additional species, P. pnc/io and P. intcriiicdins. Examination of 100 Palaemonctes, randomly selected from animals furnished by the collectors, tor differences in their rostral teeth, the character on which distinction between the three species can be most readily made, showed two P. f>ii(/io and no P. intcrmcdius. The Palaemonetes in tin's study were therefore used without further regard to species distinction. Fingerman and Mobberly (1960) reported 8% P. />> and 92% P. ritli/uris in a similar examination at Woods Hole. They observed, in addition, no physiological distinction in retinal pigment responses between these two species. CRUSTACEAN RETINAL PIGMENT HORMONE 75 made between heated and unheated extracts of the same tissue. The particulars of such preparations are described for each donor species. The test and control animals, isolated in individual containers, were dark- adapted for 3 to 10 hours before injection. Each test animal was injected, at measured intervals, with 0.05 ml. of the prepared extract by the dim light from a red lamp. Uninjected control animals were exposed to the same light for com- parable periods. Since it has been shown (Kleinholz, 1936, 1938) that maximum response of dark-adapted distal retinal pigment cells both in Palaciiioiietcs I'ulgaris and in Palaciiion adsf>ersus occurs between 30 and 45 minutes after injection of eyestalk extract, measurements in the tests with Pandalus and with Ncphrops tissues were made in this interval ; in all others the position of the distal retinal pigment of each FIGURE 1. An eyestalk of Palaemonetes rult/uris. from the dorsal surface, showing the dimensions of the retina measured for calculation of the distal retinal pigment index, //>. test animal was measured 45 minutes after injection. This was done with a com- pound microscope furnished with an ocular micrometer. The two eyestalks of the prawn were ablated and oriented with the dorsal pigment spot uppermost in a small cell containing water. Two readings from the margin of the cornea were taken of each retina : one to the margin of the distal retinal pigment and the second to the level of the dorsal pigment spot ( Fig. 1 ) . From these readings the distal retinal pigment index was calculated (Sandeen and Brown, 1952). Such measurements were made within two minutes after removal of the eyestalks ; this period of ex- posure to the light from the microscope lamp had no perceptible effect on the position of the distal retinal pigment. Retinal pigment indices for the series of prawns injected with similar extracts were averaged and the standard deviations calculated for each series to show the variance among the results. Student's / distribution was used in finding the 76 KLEINHOLZ, BURGESS, CARLISLE AND PFLUEGER .30 .ou E.S- h (E.S.-S.G) h E.S. E.S.-SG. 1X33 It .03 1.02 A .03 .20 _ S.G. h .04 36 1.04 X UJ 17 20 24 TG. Q 1.03 .10 - 31 37 17 Con. h- 02 Ul !S> 26 51 nn .20- o- LIBINIA .00 ES .03 E.S- SG. - .06 34 36 - SG- ^.03 Con -02 38 33 CALLINECTES .3 Or h- UJ cc ES-MT E.S-R. E.S. E S-SG 1.01 ">* E S- O.G t03t i.04 i03 .20 - H Q 50 34 16 20 .1 SG MT. QG. 18 1.06 .02 38 .02 17 R 20 Con. 20 oz /"\ /^ 70 CARCINUS FIGUKE 2. Distal retinal pigment light-adapting hormone in brachyuran nervous tissue, ^s shown by the effects of injected extracts on the degree of distal retinal pigment light-adapta- lion in test Palaemonetes. ES, extract of whole eyestalk ; ES h , extract of whole eyestalks, heated briefly in a water-bath ; SG, sinus glands ; SG h , heated extract of sinus glands ; ES-SG, CRUSTACEAN RETINAL PIGMENT HORMONE 77 statistical significance between average distal retinal pigment indices of experi- mental series. OBSERVATIONS A. Brachyura 1 . Carcinus niacnas This study was begun with extracts prepared from eyestalks and eyestalk components of Carcinus niacnas and tested at Woods Hole on the dark-adapted retina of Palaenwnctcs. The donor Carcinus were light-adapted males and aver- aged 5.4 cm. in maximum carapace width. Tissues, in concentrations of 10 units per 1.0 ml. of sea water, were triturated, as described above; the supernatant was injected in 0.04-ml. doses into the abdominal musculature of each dark-adapted test animal. The distal retinal pigment indices of test animals injected with these extracts are summarized in Figure 2. Extracts of sinus gland produced retinal pigment indices higher than those for the uninjected control Palaenwnctcs (0.11 as com- pared with 0.05) but much lower than those of test animals injected with extracts of entire eyestalks (0.24). Extract prepared from sinusglandless eyestalks, how- ever, resulted in responses as marked as those given by extracts prepared from entire eyestalks. In similar fashion, extracts of medulla terminalis, of retina, and of the second and third optic ganglia, each produced some degree of light- adaptation of the distal retinal pigment as evidence for the presence of retinal pigment hormone in the tested extract. Extracts of the remainder of the eyestalk from which these particular components had been removed evoked responses similar to that produced by the entire eyestalk. 2. Callincctcs sapidits In an earlier account (Kleinholz, 1936) the slight efifect of Callincctes eyestalk extract on the distal retinal pigment of dark-adapted Palaenwnctcs was complicated by high mortalities among the test animals. In the present study, eyestalks were removed from light-adapted male crabs, averaging 12.4 cm. in carapace width, measured from tip to tip of the lateral spines. Extracts in concentrations of 6 units (i.e. whole eyestalks, sinus glands, etc.) per 1.0 ml. of sea water were used without untoward effects on the survival of the injected test animals if the crude extract was heated for two minutes in a boiling water bath, and the coagulated proteins removed by centrifugation, It is apparent from Figure 2 that injected sinus gland extracts cause only slight light-adaptation ( the retinal pigment index is 0.083 compared with the eyestalks from which the sinus glands had been removed; (ES-SG)*, heated extract of sinusglandless eyestalks ; MT, medulla terminalis or fourth optic ganglion ; ES-MT, eyestalks from which medulla terminalis had been removed; R, retina, including portions of the lamina ganglionaris or first optic ganglion ; ES-R, eyestalks minus the retinal region ; OG, the second and third optic ganglia, not including the sinus gland ; ES-OG, eyestalks, including the sinus glands, but minus the second and third ganglia; IG, thoracic ganglia; CON. control, uninjected dark-adapted test animals, exposed briefly to the light from a red photographic lamp. The numbers near the top of each bar are the standard deviations (rounded to the nearest hundredth) from the average distal retinal pigment index. The numbers near the middle of each bar indicate the number of retinas measured in calculating the average index. KLEINHOLZ, BURGESS, CARLISLE AND PFLUEGER 0.057 of the uninjected controls) whereas extracts of eyestalks without the sinus glands produce a response almost that obtained with extracts of whole eyestalks (indices of 0.219 and 0.235, respectively). 3. Libinia emarginata Nervous tissues, taken from light-adapted males weighing about 500 grams, were ground and extracted with sea water to a final concentration of 10 units per 1.0 ml. One set of extracts of whole eyestalks. of sinus glands, and of eyestalks without sinus glands was heated for two minutes at 100 C., and the activity in distal retinal pigment hormone compared with that of the like series of unheated extracts. The fused mass constituting the thoracic ganglion also \vas extracted for injection. Since the wet weight of such a ganglionic mass is approximately equal to that of the soft tissue of an eyestalk from the same animal, extracts of thoracic ganglia were prepared in concentrations of 10 per 1.0 ml. The results of tests with such tissue extracts from Libinia are shown in Figure 2. Sinus gland extracts, although producing substantial migration of the distal retinal pigment in the injected test animals (Retinal Pigment Index = 0.166), are appreciably less active than extracts of entire eyestalks (RPI =0.261) or extracts of sinusglandless eyestalks (RPI : =0.255). Thoracic ganglia extracts produce a lesser degree of light-adaptation of the distal retinal pigment cells (RPI = 0.138) than do the other tissue extracts. The retinal pigment indices effected by injec- tion of heat-treated extracts (RPI for whole ES = 0.268 ; RPI for SG = 0.178; RPI for ES-SG = 0.265 ) are not significantly different from those produced by the unheated extracts. In the three species of crabs examined, the activity of retinal pigment hormone in sinus gland extracts is consistently less than in extracts of sinusglandless eye- stalks. The differences in degree of light-adaptation of the distal pigment cells produced by sinus gland extracts from these brachyurans have not been examined more closely; they may be due to size differences between the donor species, or to the number of sinus glands used in preparing the extracts. P.. Caridea ' 1 . Pandahis borcalis Similar studies with comparable extracts prepared from the eyestalks of de- capod crustaceans other than brachyurans seemed desirable, and advantage was taken of an opportunity to make such tests with the relatively large eyestalks of Paudalns at the Kristineberg Zoological Station. P. borcalis is a protandrous hermaphrodite whose biology has been described by Rasmussen ( 1953 ) and by Horsted and Smidt (1956). Carlisle ( 1959) has reported on the histology of the neurosecretory elements in the eyestalk of Pandalus. The components of the eye- stalk tested for retinal pigment hormones are shown in Figure 3. 2 These studies with tissue extracts of Ptindalus borcalis were made in 1957 in collaboration with Dr. David B. Carlisle at the Kristineberg Zoological Station in Sweden. D. B. C. dissected and separated the various eyestalk components of Pandalits; activity of the extracts was tested by L. H. K. and D. B. C. We are indebted to Dr. G. Gustafson, then director of the Station, for making laboratory accommodations available to us for this study. CRUSTACEAN RETINAL PIGMENT HORMONE 79 Extracts of such components, either singly or in various combinations, were prepared in concentrations of 10 units per 1.0 nil. of sea water and heated for two minutes at 85 C. Similar preparations of sinus glands and of eyestalks minus sinus glands, without such heating, were tested in comparison with the heat-treated extracts. In addition, extracts of 10 ventral nerve cords (from telson to circum- esophageal connectives) per 1.0 ml. of walking legs and svvimmerets for control tests, and of eyestalk preparation of erythrophore-concentrating hormone, gener- ously given us by Drs. Ostlund and Fange. were tested for retinal pigment activity. X.O.C. MI. Fici'Kii ,1. Dissection of a left eyestalk of Pandalits borculis, approaching somewhat obliquely from the dorso-abaxial aspect, so that the mid-dorsal line is towards the right-hand side of the drawing. Only the nervous tissues are shown. Ganglionic X-organs show in their natural appearance as whitish patches against their corresponding medullae ; from each of them a neurosecretory tract runs to the sinus gland. A tract also runs from the brain to the sinus gland. In life these tracts also appear as white lines against the nervous tissue. Dissection by D. B. Carlisle. LG, lamina ganglionaris ; MF.. medulla externa, or second optic ganglion; MI, medulla interna ; MT. medulla terminalis; SG, sinus gland; SPX, sensory pore X-organ (Hanstrom's X-organ) ; XOC, X-organ connective. The unavailability of Palacnionctcs i'iil(/aris at Kristineberg necessitated the use of Palacinon adspcrsns as the test animal. The responses of the distal retinal pigment of dark-adapted Palacinon adspcrsns, measured between 30-45 minutes after injection of various concentrations of eyestalk extracts (Kleinholz, 1938). are similar to those shown by Palacnionctcs, in that the responses are graded ac- cording to concentration. The distal retinal pigment indices, however, are slightly different from those in Palacnionctcs. In dark-adapted, uninjected Palacinon the distal retinal pigment index averaged 0.029, while that for animals light-adapted for three hours in direct sunlight averaged 0.257. so KLEINHOLZ, BURGESS, CARLISLE AND PFLUEGER .20 SG. E.S. 04 N.C. 03 ES.-S.G. aG.+XO (SG.+XO) E.S.-X.O. 03 S.G. U 04 O4 04 O5 E.S- 03 (E.S-SGJ U RT (SG.X.Q*M.T.) If\ xo. O4 03 03 UJ Q 28 20 16 02 12 20 20 10 37 20 20 20 20 QK 1- WL. Con OI UJ 01 01 cr> on 12 18 42 PANDALUS _i .30 UJ o: ES ES b 02 oz a o vj. .03 .20 h- SG. Q E S E s-SG ES-sa 03 10 14 16 01 O4 02 .1 - - - SG. 36 SG Con. 10 O2 16 Con. 22 O3 101 18 OI Con r\ r\ 16 12 12 01 HOMARUS ORCONECTES NEPHROPS FIGURE 4. Distal retinal pigment light-adapting hormone in caridean ( Paiidalns] and macruran nervous tissues. Extracts of Pandalus and of Ncphrops tissues were tested on Palactnou adspcrsits; tissues from Homanis and from Orcoiicctcs were tested on Palacmonctcs. ES, extract of whole eyestalks ; ES a , extracts of whole eyestalks of Homanis, in a concentration of 6 ES per 1.0 ml. ; RSn. extracts of whole eyestalks of Howarus. in concentration of 3 ES per 1.0 ml.; SG, sinus gland; SG,,. unlu-ated sinus gland extract from Pandalus: ES-SG, sinus- CRUSTACEAN RETINAL PIGMENT HORMONE SI The results of these tests with Pandalits tissues are summarized in Figure 4. Several interesting features are apparent in this summary. First, extracts of Pandahis sinus gland, in contrast with brachyuran sinus gland extracts, effect as much light-adaptation in the test animals as do extracts of whole eyestalks (distal retinal pigment indices of 0.191 and 0.175, respectively). Second, extract of eye- stalks from which sinus glands have been removed, as with the brachyurans. also produces high distal retinal pigment indices, the average being 0.159. Third, extracts of ventral nerve cord contain light-adapting hormone, the average distal retinal pigment index being 0.172. Fourth, as occurred also in tests with Carcinus tissue, extracts prepared from components of Pandahis eyestalk are effective in bringing about various degrees of light-adaptation of the distal retinal pigment. Fifth, the erythrophore-concentrating hormone preparation of Eclman, Fange and Ostlund (1958), which we tested and found active on the small erythrophores of destalked Palacnion adspcrsus in a concentration of 1 : 100,000, was only slightly effective in causing light-adaptation in a concentration of 1 : 10,000, the resulting distal retinal pigment index being 0.040. Sixth, the differences in distal retinal pigment index between heated extracts (of sinus gland, 0.191 ; of eyestalk minus sinus gland, 0.159) and similar unheated extracts (of sinus gland, 0.148; of eyestalk minus sinus gland, 0.127) are significant. The difference for the sinus gland ex- tracts is significant at the 0.001 level, that for the sinusglandless eyestalk extracts is significant at the 0.05 level but not at the 0.01 level. C. Macrura 1. NepJirops nonrr/iciis Eyestalk tissue of Nephrops nori'c. Zool., 50: 71-105. RASMUSSEX, B., 1953. On the geographical variation in growth and sexual development of the Deep Sea Prawn. Rep. Norwey. Fish. Invest., vol. 10, No. 3, 160 pp. SANDEEX, M. L, 1950. Chromatophorotropins in the central nervous system of Uca ptii/ilator, with special reference to their origins and actions. Pliysiol. Zool., 23 : 337-352. SANDEEN. M. L, AND F. A. BROWN, JR., 1952. Responses of the distal retinal pigment of Palaemotictes to illumination. Pliysiol. Zool.. 25 : 222-230. \\~RLSH, J. H., 1941. The sinus glands and 24-hour cycles of retinal pigment migration in the crayfish. /. E.vp. Zool.. 86: 35-49. GAMETOGENESIS AND SPAWNING OF THE EUROPEAN OYSTER. O. EDULIS, IX WATERS OF MAINE V. L. LOOSANOFF L'. S'. Bureau <>f Commercial Fisheries. Jiioloyical Laboratory, MilforJ. Connecticut In the late '40's a heavy mortality of an epizootic nature occurred among the populations of the soft clam, Mya arcnaria. along the New England coast. The mortality was especially serious in the waters of Maine where, in many areas, these clams almost completely disappeared. Since .17. arcnaria was virtually the only commercial mollusk of that state, many shore communities were seriously affected economically- Realizing that economy established on a single shellfishery is insecure, we suggested that the local shellfisheries resources be supplemented by the introduction of another mollusk of commercial promise. We had in mind the European oyster, Ostrca edit! is. which propagates at a somewhat lower temperature than our native oyster, Crassostrea virgin ica. The European oysters were shipped to Milford from the Oosterschelde, Hol- land, in the fall of 1949. Large oysters arrived in good condition and, after resting in Milford Harbor, were planted in Boothbay Harbor and three other locations in Maine ( Loosanoff , 1951, 1955). In Boothbay Harbor the oysters grew well and showed comparatively low mortality. Gonadal samples, taken from this group at regular intervals, provided the material for the histocytological studies on which the present paper is based. The tissue was preserved in Bouin's fluid, sectioned at 5 p.. and stained with Heidenhain's iron hematoxylin and eosin. A description of the sexual phases of the European oyster is not the purpose of this article. This matter has been considered in a number of extremely in- formative publications, starting with Hoek ( 1883 ) ; Sparck ( 1925 ) ; and, especially, Orton (1927, 1933). Later it was reviewed by Coe (1936). Obviously, it is not necessary to go into the details of this complex problem again. Therefore, this article describes only two important aspects of the behavior of European oysters gametogenesis and spawning under the new set of ecological conditions encountered in Boothbay Harbor. It may be appropriate to mention here, chiefly as a reminder. Orton's discovery that the European oyster is protandric. with the sexual phases alternating regularly, in most individuals, after the initial male phase. Under favorable conditions each adult oyster completes one male and one female phase each year. Some oysters function as males early in the spawning season and later change to the female state. Others have the opposite sequence of phases. Because of this situation, individuals of both sexual types are found in the population during the entire spawning season (Cole, 1941). Orton (1927) also suggested dividing the oysters into approximately eight arbitrary categories based upon the relative numbers of cells of different sexes in their gonads. These categories, which were established principally on studies 86 GONADS OF O. EDULIS IN MAINE 87 of gonadal smears, do not, however, appear to be precise enough if compared to results of similar studies in which regular histological preparations are used. The latter, based on a large number of sections, would show that different portions of a gonad of the same individual may often give an entirely different characteriza- tion of the sexual conditions of that individual. In other words, while some portions of the gonadal tissue of an individual may characterize it, according to Orton's standards, as a pure male, another and often closely adjacent portion of the gonad may display purely female conditions. Sometimes, in the same individual the adjacent follicles might contain cells of opposite sexes although the individual, as a whole, is, obviously, ambisexual. Since virtually all of the Boothbay Harbor oysters studied were ambisexual to some extent, we hesitate to accept the classification of Orton, who assumed that there are pure male and pure female individuals which, presumably, contain ex- clusively male or female cells, respectively. Accordingly, in this study we shall designate all individuals as representing three chief categories, namely, ( a ) strongly ambisexual (Fig. 1); (b) predominantly male (Fig. 2); and (c) predominantly female ( Fig. 3 ) . It is convenient for several reasons to begin the discussion of the seasonal sexual changes of the European oysters kept in Boothbay Harbor with the spring gametogenesis. Although in some instances it may begin in April, the population in general does not display early stages of gametogenesis until the latter part of May. Even at that time, approximately 10% of the oysters still possess typical winter follicles, which are small, contracted and surrounded by large masses of connective tissue (Fig. 4). At this time they contain only indifferent cells or gonia, although phagocytic cells are often present in the lumina. In oysters already undergoing spring gametogenesis, follicles in the same individuals are often in widely different stages of development. Although, in general, they are still small and surrounded by connective tissue, which occupies most of the space between digestive gland and outside membrane of the oyster body, a few sperm-balls are already present in males. In the majority of cases, however, only spermatids or spermatocytes have been formed so far. At this period, developing eggs are usually small, measuring under 50 p. (Fig. 5). Two basic differences can be noticed between the general structure of the gonads of American and European oysters at this time and, also, later in the season. One is that the tissue layer for potential development of gonad material is significantly thicker in American than in European oysters of the same size. In the latter, the layer of gonadal tissue, i.e., the area demarcated from the outside by the body wall and from the inside by the digestive gland, may very often con- sist of a single layer of follicles. Actual measurements show that while in Ameri- can oysters 3" to 4" in size the thickness of gonadal layer may often be 3.5 to 4 mm., in European oysters of the same size it rarely exceeds 1.25 mm. Another basic difference is that while in American oysters the follicles are surrounded by a large number of leucocytes (lymphocytes) which probably act as carriers of nutritive material to the developing follicles ( Loosanoff , 1942), cells of this type are entirely absent in O. cdiilis (Fig. 6). By the middle of June the gonadal layer is still comparatively thin and only in exceptional cases do proliferating follicles extend all the way to the digestive gland. 88 V. L. LOOSAXOFF FIGURE 1. Ripe ambisexual gonad of C. cdnlis. Sperm-balls and groups of cells in earlier of spermatogenesis are in tbe center of the follicles. Note connections between blood vessels and apices of the follicles. X 80. I-K.URE 2. Ripe predominantly male gonad of O. cdnlis. Many sperm-balls can be seen jn genital canal. ( 80. FIGURE 3. Ripe predominantly female gonad of O. cdnlis. X 80. FIGURE 4. Winter follicles of O. cdnlis confined between body wall and digestive gland GONADS OF O. EDULIS IN MAINE 89 Nevertheless, more than half of the follicles with male cells contain what appear to he fully developed sperm-halls (Fig. 7). Spermatids become more numerous while primary and secondary spermatocytes begin to decrease in numbers. Eggs measuring between 75 and 100 p. are already found in some individuals, but there is no evidence of spawning. At the end of June and during the first week of July sperm-balls can be found in the genital canals of some oysters, yet no mass spawning has occurred. Spawning begins during the second or third week of July and continues until about the end of August ( Fig. 8 ) . However, during the first part of this period manv oysters are still unripe and gonadal follicles occupy onlv about half the available space. In many males, even those which have discharged some sperm- balls, large groups of spermatids, and even spermatocytes, are still present. Toward the end of July unripe oysters are still common but in the group representing only 5 to I0 c /c of the population, spawning is already completed and resorption of gonads is beginning. In the latter individuals shrinking of follicles is in progress and invasion of connective tissue is rapidly taking place. Most of the individuals that have completed spawning are males but, strangely enough, re- gardless of early completion of the male phase, there is no cytological indication that a female phase will follow. It is probable, therefore, that in many instances, because of our short, cold summers these oysters can complete only one reproductive phase each season. By the middle of August approximately 75 c /f of the oysters are either partially ( Fig. 9 ) or completely spawned. The majority of the completely spawned in- dividuals, however, are predominantly males. As observed earlier in the summer, it is apparent that the male phase would not be followed by the completed female phase. Only in exceptional cases are small ovogonia or ovocytes seen developing along the walls of the follicles but. due to the lateness of the season, their develop- ment cannot progress very far before the onset of winter and relative inactivity. After completion of the major spawning activities, the rest of the period of comparatively high temperature, confined chiefly to September (Fig. 15 ), is charac- terized by resorption of gonads and accumulation of glycogen. Resorption is brought about principally In phago-leucocytic cells which enter the follicles not through the follicular walls, as in American oysters (LoosanorY, 1942), but directly through the blood vessels (Figs. 10 and 11 ). Accumulation of glycogen is demon- strated by development of masses of connective tissue containing it, which rapidly fills almost the entire area between the digestive gland and the body wall of the oyster. The follicles, situated as islands within this tissue, are small, contracted and often filled with phagocytic cells (Fig. 12 ). Strangely enough, simultaneously with completely spawned individuals there are still apparently normal oysters that have undischarged sex cells. These vary from those that retain only few cells in the follicles, or at least in genital canals, to those that carry virtually an undischarged supply of gametes. These individuals and surrounded by large areas of connective tissue. >< 80. FIGURE 5. Ambisexual gonad in early stages of spring gametogenesis. K 80. FIGURE 6. Predominantly female gonad during spring gametogenesis. Note absence of leucocytes (lymphocytes) outside of follicular wall, and connections between blood vessels and follicles. X 80. 90 V. L. LOOSANOFF 12 FIGURE 7. Portion of male follicles showing groups of cells in different stages of de- velopment. Sperm-balls, characterized by presence of tails of spermatozoa, are seen in the center. X 345. FIGURE 8. Ambisexual gonad during last stages of spawning. Remnants of male cells and some undischarged eggs can be seen in shrunkcMi follicles, while a number of ripe eggs in the process of discharge are concentrated in the genital canal. X 80. GONADS OF O. EDULIS IN MAINE ( H may belong to all categories, including both strongly defined groups, i.e., those predominantly males and those that are almost pure females. A still more confusing aspect of the situation is that in some individuals with definitely male characteristics, spermatogenesis appears to be still in progress. Furthermore, in the same individuals, and sometimes even within the same micro- scopic field, follicles may be located side by side, some of which are completely discharged and already resorbed, others, only partly discharged, and the third group may contain male or female cells that seem to be undergoing healthy gameto- genesis. Such situations can be encountered in samples collected by the end of October and, in a few instances, even later in the year. According to Orton (1927) the proportion of females remaining unspent at the end of the summer in English waters varies from to 5%. In our case, however, the proportion was significantly higher, in some samples exceeding 25 %. In discussing the fate of unspawned eggs, Orton (1927, p. 974) suggested that the oysters eventually dispose of these cells in two ways : some ova may be either retained in the gonad and become resorbed or "they may be included in egg-cysts and extruded in masses and excreted en bloc on to the internal face of the shell and covered over with nacreous or horny matter in the form of an excretion blister." While we agree with the first method suggested by Orton, we cannot accept his second suggestion which has not been supported by any reliable observations. Striking differences in the condition of the gonad of the European oyster at the end of the spawning period in Boothbay Harbor probably suggest that we did not deal with a homogeneous population, but a group composed of representa- tives of physiologically-different races. Loosanoff and Engle ( 1942 ) and Loosan- off and Nomejko (1951 ) have demonstrated the existence of such physiologically- different races of the American oyster, C. virginica, along our Atlantic coast. These races require different minimum temperatures for development of gonads and inducement of spawning. In general, the breeding temperature requirements of the northern oyster are lower than those of the southern group. Korringa (1957) came to a similar conclusion regarding European oysters. He stated that the general population of that species is composed of several distinctly different physio- logical races which require different temperatures to carry on normal propagation activities. Since French seed oysters are sometimes imported for cultivation in Holland, it is quite probable that within the sample sent to us there were individuals the genetic complex of which was such that their temperature requirements for repro- duction were higher than those of the true northern oysters of Holland. As a result. because of the short, cold summers of Boothbay Harbor, these oysters were unable FIGURE 9. Partly discharged predominantly male gonad. Groups of sperm-balls are seen in the genital canal, while many follicles are contracting and are invaded by phago-leucocytic cells. X 80. FIGURE 10. Resorption of gonad of O. cdtilis after spawning. Note shrinking follicles and connections between follicles and blood vessels. >< 80. FIGURE 11. Migration of phago-leucocytic cells from blood vessels into apices of follicles. X775. FIGURE 12. Follicles in last stages of resorption. Large group of phago-leucocytic cells occupying lumen. Young sex cells along follicular walls can be seen. 345. 92 V. L. LOOSANOI-F \ . 14 FIGURE 13. Cytolysis of predominantly female gonad that failed to ripen and be dis- charged. Note differences of degree of cytolysis of ovocytes in different follicles and also absence of phago-leucocytic cells outside follicular \valls. X 80. FIGURE 14. Advanced stages of resorption of predominantly female gonad. Follicles are filled with large numbers of phago-leucocytic cells. X 80. 60 55 50 UJ K 5 45 K UJ a. Z u I- 40 35 30 25 MAXIMUM MEAN MINIMUM JAN FEB MAR APR MAY JUNE JULY MONTH AUG SEPT OCT NOV DEC FIGURE 15. Mean, maximum and minimum monthly water temperatures at Boothbay Harbor, Maine during the period from 1906 to 1948. GONADS OF O. EDULIS IN MAINE ( >3 to discharge their spawn. If the original shipment of European oysters, from which our samples for histological studies were collected, did include such races, it is quite probable that only the individuals carrying certain combinations of genes were able to propagate in Boothbay Harbor and are, therefore, responsible for the new population, which is now found in inshore areas of Maine (Welch, in press). Studies of gonads collected late in November and December indicate that oysters can roughly be classified into two groups. In the first, resorption is almost or already completed. The follicles may be still full of phagocytes but, neverthe- less, some young cells may be found along their walls. However, these cells are not differentiated. The second group consists of individuals in which resorption of unspent sex cells is far from complete. Superficial examination may suggest that gonads are in the middle or at the beginning of the spawning period, but more detailed studies will show that cytolysis of undischarged cells, especially eggs, is in progress (Fig. 13). Thus, at this time of the year, as during other periods, there are pronounced individual differences between the condition of gonads of different oysters even if they belong, in general, to the same sex group. Resorption of unspent cells continues through the winter. In some cases, it extends into April and perhaps even the first week of May, apparently depending upon individual conditions of the animals at the end of the preceding spawning season. However, at the same time and in the same oysters, in some of the follicles where resorption has already been completed, spring gametogenesis may be in progress. Consequently, two processes are going on simultaneously in the same individual constructive gametogenesis and cytolysis. The latter is most active in female follicles where large groups of phagocytic cells attack remnants of the eggs (Fig. 14) which, in their appearance, greatly differ from each other, depending upon the stage of cytolysis. As a rule, during early May no fully formed sperm-balls are found in the male follicles, but spermatids are occasionally seen. In female follicles normally devel- oped ovocytes are clearly distinguishable, and in well advanced females, follicles may be seen proliferating almost to the digestive gland. In this way the annual sexual cycle of O. cdulis is completed and begins anew. These studies were made possible through the cooperation of Messrs. John B. Glude and Walter R. Welch, of the U. S. Bureau of Commercial Fisheries, who were kind enough to take care of the European oysters kept in Boothbay Harbor and send me preserved samples of gonads. Special thanks are due to Messrs. Richard E. Reed, former Commissioner, Sea & Shore Fisheries, Maine, and Dana E. Wallace, who were extremely helpful in overcoming a number of difficulties in connection with the introduction of 0. cdnlis into this country. SUMMARY 1. Active spring gametogenesis begins in May. 2. General spawning begins approximately during the second or third week of July and continues until about the end of August. 3. Because of short, cold summers, usually only one sex phase is completed bv an individual ovster. 94 V. L. LOOSANOFF 4. Resorption of gonads is carried principally by pbago-leucocytes which enter the follicles not through the follicular walls, as in American oysters, but directly through the blood vessels. 5. Resorption of gonads may continue throughout the winter and early spring. 6. Differences in conditions of gonads of 0. cditlis planted in Boothbay Harbor suggest that the original group was not composed of a homogeneous population, but consisted of different physiological races. LITERATURE CITED CUE, W. R., 1936. Sex ratios and sex changes in mollusks. Mcinoircs an Mnscc Royal d'Histoirc Naiurcllc dc Bclgiquc. 12 : 69-76. COLE, H. A., 1941. The fecundity of Ostrca cdnlis. J. Mar. Biol. Assoc.. 25: 243-260. HOEK, P. P. C., 1883. De Voortplantingsorganen van de Oester : Les organes de la generation de 1'huitre. Tijdschr. Nicd. Dicrknnd. I'crccniging, 1: 113-253. KORRINGA, P., 1957. Water temperature and breeding throughout the geographical range of Ostrca cdnlis. Ann. Biol., 33: 109-116. LOOSANOFF, V. L., 1942. Seasonal gonadal changes in the adult oysters, Ostrca virginica, of Long Island Sound. Biol. Bull.. 82: 195-206. LOOSAXOFF, V. L., 1951. European oyster, O. cdnlis, in the waters of the United States. Anat. Rec.. Ill: 126. LOOSAXOFF, Y. L., 1955. The European oyster in American waters. Science, 121: 119-121. LOOSAXOFF, V. L., AND J. B. ENGLE, 1942. Accumulation and discharge of spawn by oysters living at different depths. Biol. Bull.. 82: 413-422. LOOSANOFF, V. L., AND C. A. NOMEJKO, 1951. Existence of physiologically-different races of oysters, Crassostrca viryinica. Biol. Bull., 101 : 151-156. OKTON, J. H., 1927. Observations and experiments on sex-change in the European oyster (O. cdnlis). J. Mar. Biol. Assoc.. 14: 967-1045. ORTON, J. H., 1933. Observations and experiments on sex-change in the European oyster (O. cdulis) . Part III. On the fate of unspawned ova. Part IV. On the change from male to female. /. Mar. Biol. Assoc., 19 : 1-53. SPARCK, R., 1925. Studies on the biology of the oyster (Ostrca cdnlis) in the Limfjord, with special reference to the influence of temperature on the sex change. Rcpt. Dan. Biol. Sta., 30: 1-84. WELCH, WALTER R., in press. The European oyster, Ostrca cdnlis. in Maine. Conr. Address. Nat. Shellfish. Assoc. SUBLITTORAL ECOLOGY OF KELP BEDS OF THE OPEN COAST AREA NEAR CARMEL, CALIFORNIA JAMES H. McLEAN 1 Hopkins Marine Station, Pacific Grorc, California The intertidal fauna and flora of the Monterey Bay area in central California are well-known, but the immediate subtidal associations, especially on the open coast where Nereocystis kelp predominates, are largely unexplored in places exposed to surf action. Free diving with the aqualung now makes possible direct observation of sublittoral zones (Drach, 1958). METHODS Diving along the open coast was undertaken in selected locations approximately nine miles south of Carmel. near Granite Creek (36 26.5' N. Lat., 121 55.5' \Y. Long.), between August, 1959, and September, 1961. This report is based on observations and collections made on 30 dives, amounting to approximately 20 hours underwater. Eleven dives were made during the summer of 1959, nine were made during the summer of 1960, and ten dives were made in alternate months during the fall, winter and spring of the two-year period of observation. Diving from protected shore areas is possible in the summer during the calm periods associated with most low tides, but during the winter months continuous heavy swells make diving possible only on exceptionally calm days. After each dive, the collected material was identified and all observations recorded. During calm weather visibility may exceed 40 feet, but following rough weather or during phytoplankton blooms, it may be restricted to less than 10 feet. The profusion of forms found in sublittoral zones requires the diver to make man}- successive dives; only when most of the fauna is recognizable is it possible to note different patterns of distribution and to measure or estimate abundance. OPEN COASTAL CONDITIONS The coastal area from the Monterey Peninsula south to Point Sur and further south to San Simeon is characterized by steep granite cliffs (Fig. 1). Relatively deep water is found inshore ; beaches are few and are usually limited to small coves. The shoreline is highly irregular. Uneven weathering of the granite has left pinnacles, some of which are submerged or constitute small islands. These small islands and projecting reefs afford protected areas in which the water between the open ocean and the shore may be 30 feet deep. Upwelling is characteristic of this coastal area, producing cool temperatures, especially in the spring and summer. Monthly temperature readings have been made at Soberanes Point, one mile to the north of Granite Creek, by personnel of 1 Present mailing address : Department of Biological Sciences, Stanford, California. 95 96 JAMES H. MrLEAN Hopkins Marine Station. Averages of monthly temperatures taken from 1956 through 1960 show a minimum of 10.8 C. in April and a maximum of 13.4 C. in September. The extremes over this five year period were 8.8 C. in June of 1959 and 14.6 C. in September. 1957. At Mussel Point in Monterey Bay, averages of daily readings for the same period reached a minimum of 12.5 C. in February and a maximum of 15.1 C. in September. These readings are approximately two de- FIGURE 1. The diving area at Granite Creek, looking north. Inshore walls are steep and the small islands afford protected diving areas. Ncrcocystis kelp, shown at the surface, is anchored in depths of 40 feet. grees higher than on the outer coast to the south where upwelling is prevalent. Salinities in the Monterey Bay area usually vary between 33.5'ir and 34.0/9 60 UJ o 40 UJ o 20 10 HEREDITY VARIANCE DOSE VARIANCE HEREDITY X DOSE VARIANCE SEX VARIANCE ENVIRONMENTAL VARIANCE VARIANCE OF SEX INTERACTIONS 12 26 40 60 DAYS POST- PARTURITION 75 FIGURE 4. Irradiation at 101 days. Breakdown of variation in body weights. Components expressed as a percentage of total variation. as arising from the differential responses of the genotypes or sexes from one level of irradiation to the next. The term, E, is considered due to uncontrollable varia- tion, and represents random variation of individual differences of mice of the same An analysis of variance was obtained sex within a litter given the same treatment. 126 DONALD J. NASH AND JOHN W. GO WEN for each of the embryological ages separately. The results of the component analy- sis are shown in Figures 3-7. Although there is considerable variation among the different embryological ages in the percentages of total variation that are attributable to the various factors 100 90 80 60 50 UJ o I UJ o a: UJ a. 40 30 20 10 HEREDITY VARIANCE DOSE VARIANCE HEREDITY X DOSE VARIANCE SEX VARIANCE ENVIRONMENTAL VARIANCE VARIANCE OF SEX INTERACTIONS 12 26 40 60 DAYS POST- PARTURITION 75 FIGURE 5. Irradiation at 14* days. Breakdown of variation in body weights. Components expressed as a percentage of total variation. operative in this experiment, there are some general patterns that hold true for all ages. Within those embryological ages where progeny show the greatest body weight response, not including birth weight response, to irradiation the effects of irradiation do not reach their maximum contribution to total variation until usually EFFECTS OF X-IRRADIATION UPON GROWTH 127 40 days or more after birth. Furthermore, this effect declines little, if at all, by 75 days. The heredity influences on variance (Table IV) contributed about 20% of the variances for the different growth stages as separated by the x-ray treatments at 100 90 80 270 60 50 P 40 2 UJ O cc LU ,_ Q. 30 20 10 HEREDITY VARIANCE DOSE VARIANCE HEREDITY X DOSE VARIANCE SEX VARIANCE ENVIRONMENTAL VARIANCE - VARIANCE OF SEX INTERACTIONS \ 12 26 40 60 DAYS POST- PARTURITION 75 FIGURE 6. Irradiation at 17* days. Breakdown of variation in body weight. Components expressed as a percentage of total variation. the specified embryonic stages of development. This average hereditary effect was almost identical with that observed for the mice treated directly after birth, 19.9%. The highest genotypic effects were observed for the 10-|- and 17^-embryonic day treatments. Lower values were observed for the embryos exposed at 144 and 6| 128 DONALD J. NASH AND JOHN W. GOWEN days after fertilization. The hereditary effects were strongest at the 12-, 26-, and 40-day growth periods. The decrease from 40 to 60 and 60 to 75 days post partum corresponded to the point in development where the infant mice were changing from their dependence on hoth their own and their mother's inheritances (in terms of lOOr 90 80 I 70