THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N, CALKINS, Columbia University E. E. JUST, Howard University E. G. CONKLIN, Princeton University FRANK R. LlLLIE, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University EDMUND B. WILSON, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor VOLUME LXXII FEBRUARY TO JUNE, 1937 Printed and Issued by LANCASTER PRESS, Inc. PRINCE 8t LEMON STS. LANCASTER, PA. 11 THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: \Yheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between June 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the re- mainder of the year. Entered October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. LANCASTKR PKKSS, INC., LANCASTKR. PA. CONTENTS No. 1. FEBRUARY, 1937 PAGE VON BRAND, THEODOR Observations upon the Nitrogen of the Particulate Matter in the Sea 1 DETHIER, VINCENT G. Gustation and Olfaction in Lepidopterous Larvae 7 KLEINHOLZ, L. H. Studies in the Pigmentary System of Crustacea. I. Color changes and diurnal rhythm in Ligia baudiniana 24 CAMERON, JOHN ANDREW The Mitotic Rate in Tadpole Skin after Repeated Injury. ... 37 AREY, LESLIE B. Observations on Two Types of Respiration in Onchidium. ... 41 CLANCY, C. W., AND G. W. BEADLE Ovary Transplants in Drosophila melanogaster : Studies of the Characters Singed, Fused, and Female-Sterile 47 WELSH, J. H., AND F. A. CHACE, JR. Eyes of Deep Sea Crustaceans. I. Acanthephyridae 57 GUTTMAN, S. A. Evidence for the Production of Accelerator and Depressor Substances by Ultraviolet Radiation of Limulus Muscle 75 ODLAUG, THERON O. Notes on the Development of Gorgodera amplicava in the Final Host 80 MORGAN, T. H. The Behavior of the Maturation Spindles in Polar Fragments of Eggs of Ilyanassa Obtained by Centrifuging GILCHRIST, FRANCIS G. Budding and Locomotion in the Scyphistomas of Aurelia. ... 99 DAWSON, J. A., WALTER R. KESSLER AND JOSEPH K. SILBERSTEIN Mitosis in Amoeba proteus 125 No. 2. APRIL, 1937 TURNER, C. L. Reproductive Cycles and Superfetation in Poeciliid Fishes. . 145 iii iv CONTENTS PAGE VON BRAND, THEODOR, NORRIS \V. RAKESTRAW AND CHARLES E. RENN The Experimental Decomposition and Regeneration of Nitro- genous Organic Matter in Sea Water 165 KLEINHOLZ, L. H. Studies in the Pigmentary System of Crustacea. II. Diurnal movements of the retinal pigments of Bermudan decapods. . . 176 RENN, CHARLES E. Bacteria and the Phosphorus Cycle in the Sea 190 SONNEBORN, T. M. Induction of Endomixis in Paramecium aurelia 196 HODGE, CHARLES, 4m Some Effects of Diet on the Gastric Epithelial Cells of the Grasshopper, Melanoplus differentialis Thomas 203 FAURE-FREMIET, E. Licnophora lyngbycola, a New Species of Infusorian from Woods Hole 212 ALLEE, W. C., AND GERTRUDE EVANS Some Effects of Numbers Present on the Rate of Cleavage and Early Development in Arbacia . . 217 FULLER, JOHN L. Feeding Rate of Calanus finmarchicus in relation to Environ- mental Conditions 233 GRAFFLIN, ALLAN L. Cyst Formation in the Glomerulnr Tufts of Certain Fish Kidneys 247 No. 3. JUNE, 1937 SHULL, A. FRANKLIN The Production of Intermediate-winged Aphids with Special Reference to the Problem of Embryonic Determination . . . 259 PARKER, (',. II. The Lopin^ of Land-Snails 287 1 IAM.I.K, Akiiifk I). The Physiology of Digestion in Plankton Crustacea. II. Further -indies on the digestive enzymes of (A) Daphnia and Polyphemn-. ( ' M) Diaptomus and Calanus 290 COONFIELD, H. l\. Symmetry and Regulation in Mnemiopsis leidyi, Agassiz . . . . 299 DALCO, A., AND G. YANDEBROECK On the Significance <>f the Polar Spot in Ripe Unfertilized and in Fertilized Ascidian Eggs 311, CONTENTS v PAGE HYMAN, LIBBIE H. Reproductive System and Copulation in Amphiscolops langer- hansi (Turbellaria Acoela) ...319 MILES, SAMUEL STOCKTON A New Genus of Hydroid and its Method of Asexual Repro- duction 327 HOPKINS, DWIGHT LUCIAN The Relation between Food, the Rate of Locomotion and Re- production in the Marine Amoeba, Flabellula mira 334 ABRAMOWITZ, ALEXANDER A. The Chromatophorotropic Hormone of the Crustacea 344 SCHECHTER, VICTOR Calcium Reduction and the Prolongation of Life in the Egg Cells of Arbacia punctulata 366- CHASE, HYMAN Y. The Effect of Ultra-violet Light upon Early Development in Eggs of Urechis caupo 377- CHURNEY, LEON AND HERBERT M. KLEIN The Electrical Charge on Nuclear Constituents (Salivary Gland Cells of Sciara coprophila) 384 LOOSANOFF, VICTOR L. Development of the Primary Gonad and Sexual Phases in Venus mercenaria Linnaeus 389 LOOSANOFF, VICTOR. L. Seasonal Gonadal Changes of Adult Clams, Venus mercenaria (L.) 406 ,Fox, DENIS L., H. U. SVERDRUP AND JOHN P. CUNNINGHAM The Rate of Water Propulsion by the California Mussel 417 INDEX . 439 Vol. LXXII, No. 1 February, 1937 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY OBSERVATIONS UPON THE NITROGEN OF THE PARTICULATE MATTER IN THE SEA ' THEODOR VON BRAND (From the Woods Hole Oceanographic Institution) The present investigation has been undertaken in order to secure data on the amount of organic matter present in participate form in dif- ferent regions of the sea. Such information is of importance because in the cycle of life in the sea a very large part of the living matter must occur as nanoplankton. These organisms, together with the more re- sistant portions of decomposing organisms, detritus, make up the par- ticulate matter which has been studied. Data have accumulated rapidly concerning the dissolved substances such as nitrate, phosphate, etc. which appear in the later stages of the cycle of decomposition but at present almost nothing is known of the quantitative occurrence of these materials while they are bound up in organized matter. For practical analytical reasons, as well as because of the supreme importance of nitrogenous compounds in the biological cycle, the dis- tribution of particulate matter has been measured in terms of the nitro- gen content. Water samples of one to four liters were collected with Nansen bottles and preserved by the addition of 20 cc. of formalin per liter. The particulate matter was concentrated according to the cen- trifugation method of Steemann-Nielsen and von Brand (1934). The nitrogen content was determined according to von Brand's (1935) modification of Krogh and Keys' (1934) method. In every case dupli- cate analyses were performed. With a few exceptions, in the case of very low values, they agreed in the range of 10 to 15 per cent. All organisms of about the size of copepods or above were removed. The values obtained thus represent the nitrogen content of the nanoplankton and detritus. Throughout the whole procedure every care was taken in order to avoid contamination with nitrogen-containing substances. There were 1 Contribution No. 123 of the Woods Hole Oceanographic Institution. 1 THEODOR VON BRAXD no indications of any great amount of contamination, but the possibility of a certain degree of contamination cannot be excluded with certainty (for example, from the wire or the closing mechanism of the Nansen bottles). It is not impossible, on the other hand, that during the con- centration a certain amount of the ^articulate matter was lost. The method of concentration gave very satisfactory results with nordic plankton (Steemann-Nielsen and von Brand, 1934) ; but with Mediter- ranean plankton considerable losses were reported (Bernard and Page, 1936). This may, however, be due to slight modifications of technique. Dr. John Fuller (personal communication) tested the method with a diluted culture of Nitschia closteriuin. He compared the values ob- tained after centrifugation with those calculated from previous direct counts on the undiluted culture. He recovered 94 per cent, whereas Bernard and Fage lost 50 per cent of Nitschia longissiina. In Table I are .summarized the analytical data from which the re- sults given in Table IV are derived. They demonstrate the order of TABLE I Accuracy of the nitrogen method X.iphthyl red in milliliters •yN in particulate matter Water Naphthyl centrifuged for the single I. Analysis II. Analysis red in milliliters analysis in corresp. to I. ii. milliliters l-y N Analysis Analysis Control Organ. Diff. Control Organ. Diff. precip. ip. precip. precip. 200 2.88 1.30 1.58 2.70 1.00 1.40 0.45 3.5 3.1 100 3.40 2.45 0.95 3.39 2.19 1.20 0.44 2.2 2.7 200 3.27 1.83 1.44 3.31 1.63 1.68 0.44 3.3 3.7 100 2.75 1.34 1.41 2.75 1.46 1.29 0.47 3.0 2.7 100 3.24 1.08 2.16 3.00 0.50 2.50 0.50 4.3 5.0 100 3.18 1.59 1.59 2.96 1.46 1.50 0.45 3.5 3.2 200 3.23 0.65 2.58 3.17 0.32 2.85 0.45 5.7 6.3 200 3.03 0.34 2.69 2.73 0.30 2.43 0.44 6.1 5.5 accuracy which ran be expected from the nitrogen method. The per- centage error is about the same in most of the data reported in the other tables. The N values actually 28). but none of these can be used with caterpillars. Some are designed for the use of different kinds of odors rather than different concentrations of the same odor; others, for winged insects only; and all, for adult insects only. After working with a great many caterpillars I have come to the conclusion that they possess a short-range olfactory sense. On account of this all olfactometers so far designed are inadequate since the source of the odor is distant from the chamber containing the animals, and by the time the odor reaches the larva- it is too dilute to be detected. TASTE AND SMELL IN LEPIDOPTEROUS LARVAE 9 The most successful method was adapted from that used by Parker and Stabler (1913) for human beings. Tt consisted, in short, of evapo- rating a known weight of ethyl alcohol in a known quantity of air. Before the experiment the container was cleaned thoroughly and rid of all odor (as far as the experimenter could determine). Four- hundredths of a cc. of 10M alcohol (460.0 gm. of alcohol to 1000.0 cc. of water) was placed in the glass jar which was then sealed, inverted, and placed in the sunlight. The jar was shaken frequently. Twelve hours were judged adequate time for complete evaporation and nearly complete diffusion. At this time one Euchcetias egle larva reposing on an odorless strip of paper was quickly thrust into the jar. Every at- tempt was made to prevent the air mixture from becoming too greatly disturbed. The time necessary for a response was usually three to four minutes. The best responses occurred when the larvae were more or less motionless. The typical response to an odor, particularly to a disagreeable one, consists of characteristic movements of the mouthparts. The labrum is retracted ; the labium, extended forward and downward. The maxil- lae, since they are more or less fused to the lower lip, of necessity follow a like course. The mandibles open, and then the whole process is reversed, beginning with the closing of the mandibles. A rapid alter- nation of this " spitting " motion often accompanied by a violent trembling of the antennae and maxillae constitutes a perfect response to an odor. The threshold was taken as the lowest concentration which would cause a response. Control experiments were run using one drop of water under similar conditions. The threshold of taste is so very high that all odors used in this experiment were below the threshold of taste; therefore, responses were truly due to olfaction. Parker and Stabler computed the threshold concentration of ethyl alcohol for man as 5.750 mg. ethyl alcohol/liter of air. Wirth arrives at an ethyl alcohol threshold concentration of 5-20 mg. /liter for an ichneumonid. On the basis of similar calculations I find the threshold of E. egle to be 55 mg./liter. This is roughly four times more than that of the ichneumonid and 9.5 times more than that of man, while that for the ichneumonid is 1.5 times more than that of man. It seemed obvious from the beginning that the caterpillar's organs of taste were located in the head and a series of experiments was begun in an attempt to locate these organs. Two procedures were employed : either to stimulate areas locally and so determine the seat of taste re- ception, or to remove appendages and areas of the head capsule one 10 VINCENT G. DKTH1ER by one or in various combinations until a point was reached where the larva no longer responded to taste. A sharp lookout was kept for any rise in the threshold as various areas were removed. Larvae of /. isa- bclla and E. cgle were used (cf. Fig. 1). FIG. 1. Ventral view of the mouthparts of the larva of Isia isabrlla. Lr, labrum ; Ep, epipharynx ; Hb, hypopharyngeal lobe: M, mandible; Al, A2, A3, first, second and third segments of the antenna; E, eye; H, hypopharynx ; Lm, labium ; MX, maxilla ; M p, maxillary palpus ; Sp, spinneret. When a larva was to be stimulated locallv, it was tied down on its back to a piece of cardboard by means of two loops of string, one around the neck, the other around tin- sixth or seventh abdominal seg- ment. Each loop passed over the caterpillar and then through two holes TASTE AND SMELL IN LEPIDOPTEROUS LARVAE 11 in the cardboard to the bottom side where the two ends were tightly tied. This " operation board " was then strapped to the stage of a binocular dissecting microscope by means of two rubber bands. A fine needle was used to bring a drop of solution to the mouthparts. After a study of the types of responses due to a dry needle or to plain water had been made, it was possible to distinguish no response from a positive response (to a test solution) on the basis of the follow- ing criteria : 1. Response to a mechanical stimulus may or may not occur as a single motion of the organ away from the source of stimulus. 2. Response to water is either neutral or passive, i.e., no response or a feeble imbibing. If the larva has previously had its fill of water, the latter does not occur. 3. Response to sweet solutions consists of a greedy drinking. 4. Response to a salt, sour, or bitter solution consists of a decisive attempt to turn away and of a vigorous spitting motion of the mouth- parts. TABLE I Local stimulation of head appendages. Minus sign signifies a negative response and plus sign signifies a positive response. Solution Maxilla Epipharynx Hypopharynx Antenna HCl (.05M) + + _ HoO Sucrose (1M) + + _ After treatment with a solution the mouthparts were rinsed with water and dried with a bit of paper towel, and the test was not re- peated until the larva had quieted down and become motionless. An examination of Table I will show the areas stimulated, the solutions employed, and the results observed. The next procedure consisted in removing various areas. This was accomplished by means of an electric cautery needle in some cases and by amputation with forceps in others. In either case the wounds closed satisfactorily, and the shock of the operation passed in from fifteen to thirty minutes. As an extra precaution, however, no test was made until a recovery period of twenty-four hours had elapsed. Fol- lowing this interval tests were made in the usual manner. It was found that taste could be abolished entirely only by removing 12 VINCENT G. DETHIER both the epipharynx and hypopharynx. The removal of either one caused the responses to be weaker but did not raise the threshold; removal of any of the other appendages (antennae included) either singly or in combination produced no effect whatsoever. These ex- periments indicated that the organs of taste are located in the epi- pharynx and hypopharynx. In order to check the conclusions deduced from the above experi- ments, namely, that organs of taste are located primarily, if not entirely, in the epipharynx and hypopharynx, a histological examination was made of the labrum and labium of first instar tent caterpillars (Mala- cosoma amcricana} and of last instar flour moth larvae (Ephcstia kuchniclla) . The various types of innervated organs found on these two ap- pendages are listed in Table II. TABLE II Sense organs of the epipharynx and hypopharynx. Epipharynx Hypopharynx Sensilla placodea Large spines Sensilla trichodea Epipharyngeal setae Sinus organs Large spines Small spines Small spines I have not been able to find any account of the organs present on the hypopharynx of caterpillars. The organs occurring ventrally on the epipharynx have been studied externally in Bombyx mori (Grandi, 1923) and externally and histologically in Orthosia lota (Henig, 1931). I have studied the external anatomy and histology of the same organs in Malacosoma americana, Ephestia kuchniella, Isia Isabella, EncJia-tias ecjle, and Protoparcc quinqiiemaculata. The number and location of these organs is nearly identical in all of these species. Sensilla placodea. — These are thin flat circular plates in the integu- ment. There is no aperture present. Each is innervated by a single bipolar sensory cell. These "sense domes" (Henig) or "pores" (Mclndoo) occur in two groups of three on either side of the median line. Sensilla trichodea. — Typical hairs are found universally over the bodies of caterpillars. Usually from ten to fourteen are distributed over the ventral surface of the labrum. TASTE AND SMELL IN LEPIDOPTEROUS LARVAE 13 Epipharyngcal Setce. — These occur in two groups of three on the anterior edge of the labrum to either side of the median line. They are wide, flat, hollow organs which do not usually articulate in a socket. Sinus Organs. — Henig (1931) seems to have been one of the first to mention these two strange organs occurring close to the ventral sur- face of the posterior part of the labrum. The sinus organs (cf. Fig. 2) FIG. 2. 1, Terminal end of the third segment of the antenna of an A. plexip- pus larva. 2, End of the third segment of the antenna of a S. cecropia larva. 3, Diagram of a longitudinal section through a larval head showing where the part of the head anterior to the line A-B was removed. 4, Surface and profile views of a typical pore. 5, Section through the head of M. aincricana parallel to the face showing the sinus organs. 6, End of the third segment of the antenna of a P. quinquemaculata larva, h, hair ; sb, sensillum basiconicum ; st, sensillum styloconicum ; hp, headpiece ; C, cuticula ; B, brain ; E, eye ; Ad, adductor muscles ; P, pharyngeal muscles ; S, sinus organ ; Hd, hypodermis ; Lr, labrum. are marked by no external differentiation of the chitin. In the species which I examined these organs occurred as round diffuse bodies at the end of a rather prominent nerve trunk. They are approximately twenty micra thick, staining dark blue with Mallory's triple connective tissue stain and dark red with Delafield's hematoxylin and eosin. 14 VINCENT G. DETHIER Spines. — Both large and small spines occur rather generally over the surfaces of the epipharynx and hypopharynx. The two types differ in size and thickness of wall only, but both are so small that it is diffi- cult to determine much concerning their structure. They appear to be hollow and project up from the even surface of the surrounding chitin, i.e., they are neither raised on bases nor set in sockets. All of them appear to be innervated, but Henig (1931) did not describe them. We see from this list that there is an abundance of sensilla in which the sense of taste might be located. One fact, however, throws out a huge number of possible organs, namely, that caterpillars with the labrum removed can still taste. Obviously there must be organs of taste in the hypopharynx and only large and small spines occur here. Of all the organs in this list the hairs are decidedly tactile; the epi- pharyngeal setae, rather too thick-walled to be gustatory in function ; the " pores," too widely distributed to be gustatory (they also occur uni- versally over the body and its appendages). The sinus organs offer a tempting possibility. Since only spines occur on the hypopharynx, it seems as though we should attribute the perception of gustatory stimuli to them. I do this with reservations because I rather favor the hy- pothesis that the function of these spines is to prevent food from slip- ping within the mouth. Future study of the structure of the sinus organs will reveal whether or not they are adapted for receiving gustatory stimuli. It was observed in repeated experiments that larva? responded to odors only when the source of the odor was near the head of the animal. The close proximity of odors to any portion of the body except the head brought no response. When a larva was cut in half, only the anterior portion responded to the odor. If the animal was severed be- tween the thorax and the abdomen, the same was true. The isolated head responded vigorously to odors; the remainder of the body did not. If merely the face of the larva was removed along the line A-B (cf. Fig. 2), there was still no response to odors. (In this case since the mouthparts were removed, a typical response would have been simply a turning away from the source of the odor.) This would seem to indi- cate that the organs of olfaction were localized in that portion of the head anterior to the line A-B. As had been done in the case of taste, the olfactory responses were determined after the removal »>f various appendages by the electric cautery or by fine forceps. A twenty-four-hour period elapsed between the time of operation and the time of the test. In separate experiments the labrum, hypo- pharynx, these two together; the antenna-, maxilla1 and these two to- TASTE AND SMELL IN LEPIDOPTEROUS LARVAE 15 gether, were removed. Only when both the antennae and maxilla; were removed was the response absent; in all other cases ol faction was not at all impaired. The following species were tested : Porthetria dispar, Isia Isabella, Anosla plexippus, Enchcctias egle, Samia cecropia, Pieris rape?, Papilio poly.vcncs, Papilio tiinnts, liphcslia kuchniella, and two noctuids. The location of the seat of smell was further determined in greater detail by removing the parts as shown in Table III. These experiments indicate that the seat of olfaction lies in the terminal segments of the antennae and maxillae. It now remains to be seen whether these segments bear sensilla that could possibly be olfac- tory in function. TABLE III Olfactory responses after removal of various appendages. Parts removed Responses observed a. Segments I, II, + III of the maxillae with Segment III of the antennae None b. Segments I, II, + III of the antennae with Segment III of the maxillae None c. Segment III of both antennae and maxillae None The antennae of caterpillars are located laterad of the bases of the mandibles, usually nearer their anterior articulations. It is generally agreed that the antenna has three segments. Segment I (basal), large and membraneous, contains no sense organs. Segment II is reduced to a mere ring distinguished solely by four sensilla placodea. Segment III (cf. Fig. 2) is characterized by a galaxy of sensilla differing in structure, number and distribution with each species. I have examined a wide variety of species including representatives of Noctuidae, Sphingidac, Arcti'idse, Lasiocampidse, Saturni'idae, Hesperi- idae, and Danaidae. A generalized outline of sensilla occurring on the terminal segment follows : 1. Two large hairs (sensilla trichodea), one of which is approxi- mately four times longer than the other. 2. A palpus-like headpiece. 3. Four sensilla located on the headpiece, two of which are blunt or acute cones (sensilla basiconica) either appreciable in size or barely discernible projections of the chitin, one an acute peg mounted on a conical base (sensillum styloconicum), and finally a rather large cone (sensillum basiconicum). 16 VINCENT G. DETHIER 4. Two large cones (sensilla basiconica) located to either side of the headpiece. 5. Two small rather inconspicuous cones (sensilla basiconica) lo- cated near the base of the headpiece. 6. A pore plate (sensillum placodeum) situated almost at the basal extremity of the terminal segment. I n some species examined the three large sensilla basiconica were FIG. 3. Maxilla of a P. quinqiicmaculata larva. h, hair; st, sensillum styloconicum ; hp, headpiece; p, pore (sensillum plat-odeum); sh, sensilla basi- conica; Mp, maxillary palpus; St, stipes; I, II, III, segments one, two, and three of the maxillary palpus. found to be characteristically sculptured (cf. Fig. 2). One type ap- pears as spiral ridges on the outside of the cone while the other type exists as internal ridges running from top to bottom and connected by fine cross striations. The purpose of these ridges and striations is not known, nor have they ever been reported in the literature. Character- istic of all these organs (hairs excepted) is the fact that they arc hollow and their walls except at the bases are less than one micron in thickness. TASTE AND SMELL IN LEPIDOPTEROUS LARVAE 17 Henig (1931) has mapped the nervous system in the antennae of O. lota. This can he taken as representative of a typical nerve map since it coincides with the nervous arrangement in the species which I have studied. The maxillae are located near the sides of the labium and fused to it at their bases. I follow here the nomenclature of parts employed by Henig (1931). Each maxilla has a large and fleshy base, the stipes (cf. Fig. 3). Distal to this is the palpiger. The remaining part of the appendage is the maxillary palpus. The first segment supports a conspicuous protuberance, the maxillary headpiece. Segment II is the same size as the headpiece, and Segment III is a minute barrel-shaped structure. Only the headpiece and the three segments contain any sense TABLE IV Types of sensilla occurring on the antennae and maxilla. Antennae Maxilla Sensilla basiconica Sensilla basiconica Large Small Small Sensillum styloconicum Sensillum styloconicum Small Large Sensilla placodea Sensilla placodea organs other than the usual types of tactile hairs. Following is a gen- eralized outline of the sense organs occurring on these locations : 1. Headpiece — three true hairs (sensilla trichodea), one to three very small cones (sensilla basiconica), two large sensilla styloconica. 2. Segment III — one true hair (sensillum trichodeum), one sensil- lum placodeum on the side opposite the location of the headpiece. 3. Segment II — one sensillum placodeum. 4. Segment I — one sensillum placodeum, five to seven minute cones (sensilla basiconica) on the extreme end of the segment. This outline may serve as a description of any one of the species ex- amined since the arrangement and structure of the sense organs is very similar. The descriptions of Henig for O. lota and of Grandi for B. mori differ but slightly. The type of innervation of the maxillae is similar to that of the antennas and need not be described here. From this survey we see that there are certainly many sense organs to which an olfactory function might be attributed. Table IV sum- marizes the types found. It seems improbable that sense pores (sensilla placodea) are olfac- tory in function, although thus far only one direct experiment has been 18 VINCENT G. DETHIER made to test this point. The maxillae and the terminal segment of the antennas of fifteen E. kuchniclla larvae were carefully removed by cutting. This operation left the larvee with sense pores on the stump of the antennae in addition to many located on the thoracic legs. The larvae completed a normal life cycle and not once during this time did they respond to an odor. Brues (1920) suggested that the odor of a plant may he one factor influencing food plant choice by lepidopterous larva?. It is now pos- sible to test this suggestion by direct experiment and, moreover, to check the results by ascertaining whether they are consistent with the various characteristics of the chemoreceptors of caterpillars as they have been described in the foregoing sections. I chose to work, for the most part, with the larvae of Anosia plexip- pus. The first series of experiments deal with the role of olfaction on plant choice. Both sides of the leaves of a clean vigorous milkweed plant (Asclepias syriaca L.) were liberally painted with a cheap perfume having a characteristically penetrating odor. As soon as the perfume on both surfaces of the leaves had become dry the leaves were stripped from the plant and placed on the floor of a clean, odorless, wooden breeding box. The leaves still retained the pungent odor of the per- fume. A larva of Anosia ple.vif~pns was then introduced into the cage. Under no circumstances would it eat the leaf. It crawled about and acted toward the leaf as it would towards any foreign leaf. The same experiment was repeated using methyl alcohol. At first the larva did not approach the leaf, but after a short time approached and began to nibble the leaf and eventually ate it. The leaf at first gave off the odor characteristic of methyl alcohol. This odor is very pronounced to our own senses ; therefore, on the basis of threshold experiments for olfac- tion it must have been perceptible to the larva.1 When the larva finally nibbled the leaf, I found that no odor remained and that there was no detectable alcohol taste on the leaf. Since man's threshold is lower than that of the larva, it would seem that the larva likewise was not able to detect any alcohol taste, and, therefore, ate the leaf. It is common experience that when larvae are placed in a breeding cage, they come to rest on the leaves of their food plant. The follow- ing experiments are based upon this observation. An assortment of various leaves was cut into uniform squares and placed in a breeding cage. Eighty-five per cent of the total number of 1 Experiments indicate that the relationship between the olfactory thresholds of man and caterpillars remains the same for practically all odorous substances. Dimethyl phthalate, which is odorless to man, appears to be odorless to cater- pillars. Larvae show absolutely no sign of being able to perceive this substance. TASTE AND SMELL IN LEPIDOPTEROUS LARVAE 19 Monarch larvae came to rest on the squares of milkweed only. If, how- ever, the squares of milkweed were treated with perfume or turpentine, the larvae were found resting indiscriminately on any square. If every square was treated with either of the two substances mentioned, the larvae again were found resting on any. square, without so much as nibbling a piece of the leaf. Here it appears that the perfume and turpentine obscured that property of the milkweed by which the larvae recognized it. The following experiments designated as " screen tests " prove definitely that this property is odor. Whatever leaves were used in the tests were laid on the floor of the breeding cage. A square of clean wire screen whose area coincided i FIG. 4. 1 and 2. Diagrams of typical paths of A. plexippus larvae over their food plant (milkweed) during a " screen test." 3 and 4. Diagrams of typical paths of A. ple.rippits larvae over a foreign leaf (oak) during a "screen test." The grouped arrows indicate locations where the larvae felt around in all directions. with that of the cage floor was pressed over the leaves and fastened tightly into position. Care was taken that no portion of the leaves projected through the screen. A larva was then placed in the cage. As it crawled about its path was plotted and various movements of its head noted. The diagrams represent typical paths (cf. Fig. 4). Monarch larvae passing over '' foreign " leaves maintained a straight path. When they passed over a milkweed leaf, they described a zigzag path, a sort of " feeling about." They never, with one exception, tried to bite through the screen or to get at the leaves in any other manner. When they tended to pass off the leaf, however, they raised the anterior portions of their bodies and felt about. In 50 per cent of the cases the 20 VINCENT G. DETHIER larvae turned back over the leaf. The paths for various species varied only slightly. The most significant experiment in this series consisted in placing the following leaves under the screen: mullein (Vcrbascum Thapsns L.), oak (Qucrcns ilicifolia). plantain (Plantago major L.), mullein coated with milkweed latex, oak coated with milkweed latex, and plantain coated with milkweed latex. The paths of the larva; showed definitely that the caterpillars recognized the latter three leaves im- mediately. If, as these experiments seem to indicate, larvae recognize the milk- weed by some odor that the leaves emit, leaves coated with some odor- less but disagreeable (to the larva;) tasting solution should, nevertheless, attract the larva;. All of the foregoing experiments were repeated. Leaves were covered with either sucrose solutions, sodium chloride solu- tions, or hydrochloric acid solutions. Although larva; immediately recognized milkweed leaves coated with a sodium chloride solution, they would begin to eat such leaves but never continue eating them. The leaf had the correct odor but the wrong taste. It is significant to note that although larvae would not eat salt- covered leaves, they, nevertheless, came to rest on them in preference to foreign leaves. In conclusion, two more types of tests were tried. The first con- sisted of painting various substances such as oak leaves, cherry leaves, lettuce leaves, and filter paper with milkweed latex. The second con- sisted of stripping of both surfaces of a milkweed leaf and gluing them on either side of some other kind of leaf. The sticky milkweed latex was used as a glue. In the first case the larva; would attack the treated substance and reject it after a few mouth fuls. At intervals they would attack it afresh but would starve rather than eat it. In the second case the same was true. The normal behavior of the larvae was to attack the edge of the leaf. When they did this, however, they soon tasted the foreign leaf sandwiched in between the milkweed. A few larvae managed to survive by eating both surfaces and leaving the middle leaf untouched. We may conclude that the odor of the leaf and its taste are primary factors influencing food plant choice in lepidopterous larvae. DISCUSSION Experiments have been presented which show definitely that ol fac- tion and gustation are the primary factors influencing food plant choice by lepidopterous larva;. The threshold, location, and probable end organs of each of these two senses have been determined. TASTE AND SMELL IN LEPIDOPTEROUS LARVAE 21 Smell in caterpillars is a sense of exceedingly " short range " since, in general, the threshold for food plant odors is very high. A larva, therefore, cannot scent its food plant from afar. The caterpillar, if it was not hatched on the food plant, must find its food entirely by chance. It is a common occurrence in nature for larvae to be knocked or blown from the plant upon which they were feeding. Gravity and light direct them to where their food plant would normally be. For example, larva? which feed on a ground plant such as plaintain or clover are positively geotropic and negatively phototropic, while larvae which feed on trees, shrubs, or herbaceous plants which grow upright are negatively geotropic and positively phototropic. Excellent proof that the first move in the direction of the food plant is a chance movement can be found in the case of the larvae of Echcetias egle which feed on milkweed. After vigorous wind storms the brightly colored larvae are found on practically every type of upright vegetation in the vicinity of a milkweed patch. Those which are not lucky enough to chance upon a milkweed plant most certainly die unless after another tumble to earth they ascend the proper plant. Similar conditions throughout nature must account for a large part of the mortality of caterpillars. Having found a plant, the caterpillar crawls over the leaf, waving its head from side to side. The larva's " short range " sense of smell is particularly fitted to this environment since the larva moves about with its olfactory receptors, the antenna; and maxillse, exceedingly close to the source of stimulus, the leaf. We recall that the olfactory threshold of caterpillars appears to be no better than our own. We also find by pressing various leaves to our noses that many of these leaves which we previously thought of as possessing no odor, do emit a rather characteristic one. The caterpillar, then, with its poor thresh- old perceives many plant odors because their sources are very close at hand. Plants which appear odorless to us are probably odorless to caterpillars. The .world of the Monarch larva is probably composed solely of " milkweed odor "' and " not milkweed odor." It is pos- sible that the odor of milkweed is the stimulus which sets off a reflex action, namely a biting reflex. At this point the larva " tastes " the leaf. Probably the taste substance in the milkweed is a stimulus which causes additional reflexes culminating in feeding. When we attempt to explain the feeding habits of polyphagous larvae, we encounter various difficulties. The experiments presented proved that the senses of gustation and olfaction are just as acute in this type of caterpillars as in monophagous and oligophagous species. There are, however, many plants whose odors are undoubtedly below the thresholds of the larvae. Then again in these larvae a specific odor VINCENT G. DETHIER may not set off a biting reflex as postulated for monophagous species. Any plant, therefore, which does not have repelling odor or whose odor is below threshold is attacked. Likewise a plant which does not have a repelling taste substance or whose taste substance is below threshold is attacked. If there is a similarity of odor and taste in several plants, as for example, cabbage and radish, it is conceivable that a fundamentally monophagous caterpillar is deceived into feeding on both species. I am inclined to classify as oligophagous only those insects which feed on plants with similar odors and tastes. An example would be Picris rapcc whose larvae feed on cabbage and radish, both of which taste and smell alike to humans. All other species are polyphagous. They eat any plant which does not contain a repellent. A polyphagous larva is not one which recognizes many odors, but rather a larva in which a par- ticular taste and odor is not required to start the chain of feeding reflexes. When we attempt to find what constituent of the plant leaf attracts the larva, we must remember certain basic facts : The constituent must most probably be one which will stimulate the human senses of taste and smell because experiments seem to indicate that the thresholds of caterpillars are higher than those of humans ; many of the constituents of a leaf vary with the season, the age of the leaf, and the time of the day and the attractant must be a constituent of the leaf which does not vary under the above conditions. The latter statement needs explanation. To human beings cabbage always tastes and smells like cabbage, lettuce like lettuce and so forth, regardless of the time of day or season the vegetable is grown or eaten. Obviously the substance in such a vegetable which gives it its character- istic taste and odor must remain relatively constant. It may vary slightly in concentration but must never fall below threshold concentra- tion. Since it appears that caterpillars can neither taste nor smell better than we can, the substance in the food plant by which the larva recognizes it undoubtedly remains quite constant in concentration. Thus in examining a plant for this unknown, we may reject all sub- stances which are peculiar to particular times and conditions. Using this method we may hope to determine what substances at- tract different larvae to different plants. In the case of the milkweed butterfly experiments seem to indicate that the substance determining the choice of this species of larva is the milkweed latex or some sub- stance contained therein. The compound or compounds in the latex which give it its characteristic taste and odor arc the attractants which we seek. TASTE AND SMELL IN LEPI DOPTEROUS LARVAE LITERATURE CITED BRUES, C. T., 1920. The selection of food-plants by insects, with special refer- ence to lepidopterous larvae. Am. Nat., 54: 313. GRANDI, G., 1922-23. Studi sullo svilluppo postembrionale delle varic razze del Bombyx mori L. Bollettino del Laboratorio di Zoologia General c Agraria, 16: 137; 17: 1. Portici. HENIG, B., 1931. Ueber die Innervierung der niederen Sinnesorgane der Schmet- terlingsraupen. Trav. Soc. Sci. Wilno, 6: 41. MclNDOO, N. E., 1919. The olfactory sense of lepidopterous larvae. Ann. Ent. Soc. Amcr., 12 (2) : 65. MclNooo, N. E., 1927. Smell and taste and their applications. Scientific Monthly, December, p. 481. PARKER, G. H., AND E. M. STABLER, 1913. On certain distinctions between taste and smell. Am. Jour. Physiol., 32: 230. RIPLEY, L. B., AND G. A. HEPBURN, 1929. Studies on reactions of the Natal fruit-fly to fermenting baits and a new olfactometer successfully used with fruit flies. Union S. Africa Dcpt. Agric. Ent. Mem., 6: 19, 55. WIRTH, W., 1928. Untersuchungen iiber Reizschwellenwerte von Geruchsstoffen bei Insekten. Biol. Zcntralblatt, 48: 567. STUDIES IN THE PIGMENTARY SYSTEM OF CRUSTACEA I. COLOR CHANGES AND DIURNAL RHYTHM IN LIGIA BAUDINIANA L. H. KLEINHOLZ (Prom the Bermuda Biological Station for Research, Inc., and the Biological Laboratories, Harvard University)1 In the summer of 1936, while engaged in collecting near the bio- logical station at St. George's West, I noticed distinct color differences in specimens of an isopod that was very common on the rocky ledges along the shore. The animals on the black porous rock above the high- tide mark were dark-grey or black, while those feeding on the algae and other plant material covering the limestone of the intertidal zone were always yellowish-white in color. Closer observation led to the belief that this difference in coloration was due, not so much to morpho- logical variations in pigment distribution, as to an active physiological concentration and dispersion of pigment within cells. It was therefore decided to investigate more fully the chromatophoral behavior in these isopods which were identified as Ligia baudiniana. The mechanism of chromatic change in the isopod crustaceans has not been definitely established. Following the early studies of Pouchet (1876) on the color changes of decapod crustaceans, Matzdorff (1883) published a report on the coloration of Idotca. He was of the opinion that chromatophoral activity was under the control of the nervous system, but decided that the inconclusive results obtained by severing the ventral nerve cord were to be attributed to injury brought about by the operative manipulation. Menke (1911), in a study of the rhythmic activity of color changes in isopods, believed that the mecha- nism for such responses was based upon physiologically innervated melanophores, referring in support of this to a figure of an innervated chromatophore from the integument of a young Philoscia in Weber's (1881) paper. It appears from a study of the figure that the nerve fiber supplying the chromatophore is a process of a peripheral nerve cell, the other processes of which supply the sensory (?) hairs on the body. The relationship, however, is not very evident and no further anatomical detail is given. The implication drawn from the illustration 1 This study was made possible by a prant from the Tames F. Porter Fund of Harvard University. I wish to thank Dr. J. F. G. Wheeler, Director, for the many kindnesses shown me during my stay at the Bermuda Station. 24 COLOR CHANGES AND DIURNAL RHYTHM IN LIGIA 25 was that chromatophoral activity depended upon a local receptor-ef- fector mechanism, a condition which subsequent study of blinded animals failed to confirm (Tait, 1910). Alenke reported that section of the ventral nerve cord had no effect on metachrosis, but that the chromatophores became dispersed if the dorsal side of a segment were cut without damaging the dorsal vessel. Several years later, Perkins (1928) and Roller (1928) showed that pigmentary changes in the decapod crustaceans were under the control of an endocrine substance that had its origin in the eye-stalks. When appropriate extracts of the eye-stalks of Pal&monetes were injected into shrimps which had become dark as a consequence of their having been kept upon a black background, the animals soon became light due to a concentration of pigment within the integumentary chromatophores. Perkins' studies on Palamonetes were confirmed and extended by Brown (1935), who believes that the chromatic components of the pigmentary system are controlled by separate hormones. Extensive experiments by Kropp and Perkins (1933) and by Han- strom (1935) established the presence of a humoral chromatophore activator in the eye-stalks of a wide variety of crustaceans. Hanstrom, in addition, has located two organs suspected of glandular function in the eye-stalks of many decapods, and one of these, the blood gland (Bliitdruse), he believes chiefly responsible for the control of pigmen- tary activity. In some crustaceans, Gebia affinis and Hippa talpoida, these glands are present, not within the eye-stalk, but in the head on the surface of the brain. This was confirmed by experiments which showed that extracts of the eye-stalks of these two crustaceans when injected into blinded Paltriuoiictcs, had no effect upon the dispersed chromatophores, while extracts of the heads were as active in effecting concentration of the pigment cells as were preparations from the eye- stalks of Pal&inonetes. In view of these advances in the study of color changes of decapod crustaceans, it was thought advisable to study melanophore activity in the isopods with regard to endocrine control. MATERIALS AND METHODS Specimens were collected along the shore where the isopods are found in large numbers, feeding upon the plant material that is un- covered by the ebbing tide. The animals were obtained by lifting rocks and dropping them into a wooden bucket, the impact being sufficient to jar the individuals loose from the under sides of the stones. Col- lections were made daily to insure a supply of normal, active individuals. 26 L. H. KLEINHOLZ In following the responses to changes in color of background, a large white porcelain dish, the bottom of which was covered with moist white sand, served as a white background. Fur adaptation to black backgrounds a glass bowl, the outside of which had been covered with black paint, was used, the sand for the bottom of this vessel being mixed with an equal amount of pulverized coal. The containers were illumi- nated by light from a 60-watt lamp at a distance of 18 inches. It became necessary during the course of these experiments to observe the reactions of blinded Liyia. Blinding by extirpation of the sessile eyes was unsatisfactory because the ensuing hemorrhage invari- ably caused the death of operated individuals. Such operations were eventually abandoned and blinding was accomplished by covering the eyes with an opaque mass, obtained by mixing plaster of Paris and lampblack with a little water. This mixture was applied over the head so that the eye- were completely covered, and, after being allowed to dry, was coated with a thin layer of waterproof paste to prevent moistening and crumbling. Extracts of the heads of Lif/ia were prepared in various concentra- tions to determine the possibility of an endocrine factor in pigmentary changes. These were prepared in two ways : in one, the heads were crushed and ground with 1.0 cc. of sea-water in a mortar, most of the coarse detritus \vas separated off, and the remaining fluid drawn directly into a hypodermic syringe for injection; the second method was essentially the same, except that the triturated heads were trans- ferred to a test tube and brought to a boil. The heat was sufficient to clump most of the solid material so that the supernatant fluid \vas almost water-clear. The solution was allowed to cool and was then drawn into the syringe. In the experiments where such extracts were injected, the pigmen- tary condition of the specimen was first examined by means of a dis- secting microscope, and observations were again made within 10 minutes after treatment. Care was taken, when injecting, to insert the needle dorsally into the body spaces, well anterior to the heart (usually between the fifth and sixth thoracic segments), to avoid loss of body fluid. Control injections consisted of both boiled and ordinary sea water. COLOR PATTERN The dominant and most obvious component of the chromotophore system in Liyia consists of cells containing a black pigment, possibly a melanin. These pigment cells are distributed over the entire surface of the animal, being more numerous and apparently smaller in size COLOR CHANGES AND DIURNAL RHYTHM IN LKJIA 27 on the dorsal side, especially in the region of the mid-line; they are less densely aggregated near the lateral margins of the tergites (Fig. 2). A second component of this system consists of white pigment. This occurs in many individuals as rather large clusters on the posterior, dorsal surface, and, by examination with the low powers of the dis- secting microscope, does not appear cellular. In addition to massed TABLE I Responses of the melanophores of Ligia baudiniana to changes in background. Series I Condition of melanophores Time after transfer of 7 white background 0 A B c 5 minutes 7 10 minutes . . . 1 6 20 minutes . . . 2 4 35 minutes 1 5 Series II Time after transfer of above white specimens to a black background 0 A B c 17 minutes 1 5 95 minutes 2 4 240 minutes 4 1 1 300 minutes 6 Series III Time after transfer of 7 dark specimens to a 0 A B C white background 5 minutes 7 30 minutes 4 3 pigment there are, however, definite cells containing this white sub- stance. Such " guanophores " appear to show a limited activity in the concentration and dispersion of their pigment, but close observation of their behavior was not undertaken in this study. A yellow pigment of some sort is also present and is most noticeable in preserved speci- mens. This combination of body colors is very effective in maintaining a concealing coloration of the animals in their native habitat. 28 L. H. KLEIXHOLZ TABLE II Responses of Ligia to injection of extracts. The concentration per cc. designates the number of heads triturated in 1.0 cc. of sea-water. £«., extract prepared from white-adapted specimens. Eb, extract prepared from black-adapted Ligia; Ab, black-adapted isopods were injected; Av, white-adapted animals were injected. In Series I, the extracts were prepared from background-adapted animals, in Series II extracts were similarly prepared and boiled. Extracts for Series III were pre- pared unboiled, from specimens during the two conditions of diurnal activity, while in Series IV similar extracts were boiled previous to injection. Series I Concentra- tion per cc. EU- into .-U Eb into Ah Eh into Aw No. of Lifiia inject. n A B C No, of Ligia inject. 0 A B c No. of Ligia inject. 0 A B C 4. . 11 10 9 10 10 2 2* 1 1 3 3 4 1 4 6 5 5 8 5 14 1 13 3 3 4 4. . 4 8 4 Total 50 4 15 29 14 1 13 3 3 Series II 25. . .. 3 3 8 5 1 3 1 4. 10 7 8 10 4 4 4 4 Total . 18 1 8 9 4 4 4 4 Series 1 1 1 4 1 1 1 1 4. 10 10 4. 8 7 1 4. . 8 3 5 1 14 1 5 8 4. 5 1 4 Total . 2] 71 35 12 1 1 1? COLOR CHANGES AND DIURNAL RHYTHM IN LIGIA 29 TABLK II (cont.) Series IV Concentra- tion per cc. EV into .4;, EI, into Ab Eb into j4u, No. of Ligia inject. 0 A B c No. of Ligia inject. 0 A B C 3 3 4 2 1 2 4 No. of Ligia iniivt . 4 0 A B C 4 10 4 4 4 4 4 1 1 4 3 3 4 4 6 4 4 4 4 4 4 4 4 4 4 5 8 2 1* 1* 1 2* 3 1 2 1* 1 2 4 3 3 1 1 1* 3 2 4 2 1 2 1 1 2 5 2.5 1.25. . 10. . 5 2.5' 10 5 2.5 1.25 . . 4. 2 1 0.5 0.25. . 0.12 0.06 4 Total . . . 16 6 10 67 3 19 20 19 4 4 * Specimen died. COLOR ADAPTATIONS TO BACKGROUNDS AND EFFECTS OF BLINDING The surmise that there was a physiological color change in adapta- tion to the color of the background was confirmed by testing the re- sponses of specimens in the laboratory on black and on white back- grounds (Table I, and Figs. 2 and 3). The melanophore changes of the black-adapted Ligia of Series I and Series III were recorded after the animals were transferred to a white background. The specimens of Series II were those of Series I which had become adapted to the white background and were then transferred to the black vessel as a converse experiment. The conditions of the melanophores recorded in Tables 1 and II are designated by the symbols used in Fig. 1. The color changes 30 L. H. KLKIXHoLZ FIG. 1. Outline drawings of melanophores, showing the conditions of the cells in four stages from dispersion to concentration of pigment. O. the maximally dispersed state, with many delicate processes visible; ./, beginning of concentra- tion, the distal processes of the cells losing their delicate tracery, and ending in blunt rounded knobs ; B, the stellate condition ; C ', the punctate state, with the nucleus obscured by the pigment granules. in response to background are seen (Table I) to be more rapid during concentration than during dispersion, being in this respect similar to the behavior of the chromatophores in I'ahciuonctcs. It became desirable after these preliminary experiments to observe the color reactions of blinded Ligia. For the isopod crustaceans, Tait (1910) reported darkening of Ligia occanica when the eyes were covered with an opaque mass, and Pieron (1914) obtained similar re- sults with blinded Idotca. In the decapods, however, differing types of responses have been found, depending upon the method of blinding em- COLOR CHANGES AND DIURNAL RHYTHM IX I.K.I \ 31 ployed by the investigator, and upon the crustacean studied. In operat- ing upon the pedunculate eyes of the decapods, two methods of blinding are possible. One procedure is to remove or destroy the retina only; the second method is to excise the entire eye-stalk, therein- removing Hanstrom's blood-gland in addition to the retina. The pigmentary changes in the animal resulting from the second type of operation are due more to deficiency of the chromatophorotropic hormones from the blood stream than to destruction of the retina. With Pahcinouctes, which does not possess a melanophore system in the body pigment, Brown found that cauterization of the retina re- sulted in a loose concentration of the red pigment and a half-way dis- persed condition of the yellow pigment cells. Ablation of both eye- stalks, however, effected a full dispersion of the red and the yellow chromatophores and random variations in the white pigment cells. Carl- son (1935, 1936) in studying Uca pugilator, which possesses melano- phores in the pigmentary system of its integument, found that removal of the distal thirds of both eye-stalks had no effect on the pigment cells, whereas total removal of both stalks (the blood-gland is located in the middle third of the eye-stalk) caused a permanent concentration of the melanophores and the erythrophores, while " the yellow chromatophores were slightly more contracted than the stellate state, and the white ones a little more expanded than that state." Abramowitz (1935) also re- ported concentration of the black pigment cells and dispersion of the guanophores of Portunus anccps, following total removal of the eye- stalks. The pigmentary reactions of the two brachyurans to these op- erations parallel the color changes of many of the lower vertebrates after hypophysectomy. There is a striking similarity in physiological effect between crustacean eye-stalk extract and the melanophore-dispers- ing principle of the vertebrate hypophysis, as reported in recent studies by Abramowitz (1936a, 19366). In preliminary tests a number of Ligia were blinded by removal of the eyes with a spear-point needle. Such specimens, with their melano- phores initially punctate, became darker after the operation, but it was thought advisable to repeat the experiment, using a technique that in- volved less injury to the animal. Similar results were obtained when 10 white-adapted Ligia were blinded by covering the eyes with an opaque mass. The melanophores of all 10 isopods became maximally dispersed within an hour; five minutes after this last observation (3:45 P.M.) the specimens were placed in the dark-room, and when examined later in the evening (5:30 P.M.) 4 of them were light, 3 were inter- mediate, and 3 were still dark. The next morning (at 8:00 A.M.) of 5 surviving animals, 4 were dark and 1 was intermediate in color. 32 L. H. KLEIXHOLZ These confusing results became more intelligible when further ob- servations showed that the isopods underwent a diurnal rhythm in melanophore activity in constant darkness. Under such conditions the pigment in the black cells was dispersed during the day and conccn- trnk-d at night. Upon an illuminated black background, however, the rhythm did not appear at night ; the isopods remained dark. The fol- lowing notes on a series of animals in the- dark-room indicate the pig- mentar condition of the June 23 10:30 I'.M. Six white /,////« with melanophores punctate wc-re placed in the dark-room. June 24 9:00 A.M. Two specimens are still light; 4 are dark with the melanophores dispersed. June 24 6:30 P.M. Same as at 9:00 A.M. June 24 10:45 P.M. Six specimens are light: in 3 the melanophores arc- punctate, while in the remaining 3 they are punctate and stellate. June 25 10:15 A.M. All specimens are dark: 4 with melanophores maximally dispersed ; 2 show them stellate and slightly more dispersed. June 26 12:30 A.M. All 6 isopods are light with the melanophores punctate. June 26 10:30 A.M. Six specimens are dark. June 26 4:30 P.M. Same as at 10:30 A.M. June 26 10:15 P.M. All the animals are light. L'n fortunately the critical times during the day when these rhythmic changes were initiated could not be determined. It became evident that considerable variation existed in the onset of the changes, and that such variation might be due to the effects of captivity. There could, how- ever, be no doubt of the existence of a pigmentary rhythm, since it was repeatedly both under laboratory conditions and at night in natural habitat of the isopods. EFFE< is OF I NJECI ING KM KACTS llanMn'nn (1*>35) showed that the activity of crustacean eye-stalk extract^ in concentrating the dispersed chroma! (Chores of blinded Palamonetes was correlated with the presence of the blood-gland in the EXPLANATION OF PLATE I l-ii,. 1. On tin left i a i"rimen that has been darkened by exposure to a black background; on the ri.nht an isnpocl adapted to a white background. The photograph was taken from /./•// presumed to be of endocrine function are located on the surface of the brain. Since the eyes of Ligia arc sessile, entire heads of specimens were used in preparing extracts. The bodily changes in color following the injection of extracts prepared from speci- mens in different pigmentary conditions are indicated in Table II. It is evident that black-adapted specimens responded to such treat- ment by a concentration of their melanophores, while white-adapted isopods showed no change. Control injections of sea-water into 42 black-adapted Ligia were without effect on the dispersed melanophores of 37 individuals, while 5 animals became perceptibly lighter in color. These 5 specimens were part of a group of 14 isopods that were in- jected at night (11:15-11:45 P.M.), a time when specimens in the dark-room are light because of the diurnal periodicity. The behavior of the black pigment cells, following injection of the extract into Ligia. is in striking contrast to that of the melanophores of I'ca. Carlson (1936) and Abramowitz (1936&) have shown that the dark coloration may be restored to blinded Uca by the injection of eye- stalk extracts. The responses of the melanophores of the two brachy- urans when compared with the diametrically opposite behavior of those in Ligia may be due to some fundamental difference in the nature of the pigment cells (vide, Kigney, 1919, on the responses of retinal and body pigments of frogs to adrenalin), or it may be due to the existence of two different hormones, one causing concentration and the other effecting dispersion of the black pigment. Unfortunately, critical ex- ]>36), Pieron's explanation should be revised to allow for the humoral factor. Several interesting speculations as to the basis for this diurnal ac- tivity have been put forward by Welsh (1936) and by myself. Welsh COLOR CHANGES AND DIURNAL RHYTHM IN LIGIA 35 suggests, " There may be a rhythmic secretory cycle in the gland which continues under constant conditions or the situation may he much more complex and the rhythm in the eye may only accompany a general rhythmic activity which results from a series of changes involving the nervous-endocrine systems." A third possibility, supplementary to the first suggestion, is that the rhythm may be due to a diurnal cycle of ex- haustion and elaboration of the secretory material when the animal is maintained under constant conditions. Physiological tests fail to substantiate this last possibility. Ex- amination of the data in Table II shows that extracts prepared from specimens of Ligia in the two diurnal pigmentary conditions are practically equally effective in causing concentration of the dispersed pigment in the melanophores. The greater activity of boiled extracts has been reported in similar observations by Perkins and Snook (1931) and by Hanstrom (1935). There is as yet no direct evidence favoring either of the two remain- ing possibilities. More complicated reactions than are at present indi- cated may be involved in such periodic pigmentary changes. While neither of these hypotheses really clarifies the means by which the rhythmic activity originates, such assumptions are of assistance in nar- rowing down the number of systems to be studied in the hope that eventually more light may be thrown upon the nature of this phe- nomenon. SUMMARY 1. The bodily changes in color of Ligia baudiniana upon black and upon white backgrounds are due chiefly to a dispersion and concentra- tion of pigment granules within melanophores. 2. When the animals are kept in constant darkness, there is a diurnal rhythm in pigmentary activity, the isopods being dark during the day, and light at night. 3. Injection of aqueous extracts of heads into the body spaces of dark Ligia brings about lightening in color by a concentration of the melanophores. 4. Extracts from the heads of dark and of light specimens in the two conditions of diurnal rhythm are practically equally effective in concentrating the melanophores of dark isopods. It may be concluded from this that the diurnal pigmentary activity is not due to a cycle of exhaustion and elaboration of secretory material in the endocrine gland controlling the color changes. 36 L. H. KLEIXHOLZ LITERATURE CITED ABRAMOWITZ, A. A., 1935. Color changes in cancroid crabs of Bermuda. Proc. Nat. Acad. Sci., 21: o77. ABRAMOWITZ, A. A., 1936a. Action of crustacean eye-stalk extract on melano- phores of hypophysectomized fishes, amphibians, and reptiles. Proc. Soc. Exp. Biol. Mcd.. 34: 714. ABRAMOWITZ, A. A., 193o/>. The action of intermedin on crustacean melanophores and of the crustacean hormone on elasmobranch melanophores. Proc. Nat. Acad. Sci.. 22: 521. BENNITT, R., 1932. Diurnal rhythm in the proximal pigment cells of the crayfish retina. Physiol. Zool., 5: 65. BIGNEY, A. J., 1919. The effect of adrenin on the pigment migration in the melanophores of the skin and in the pigment cells of the retina of the frog. Jour. Expcr. Zool., 27: 391. BROWN, F. A., JR., 1935. Control of pigment migration within the chromatophores of Palaemonetes vulgaris. Jour. E.i-pcr. Zool.. 71: 1. CARLSON, S. PH., 1935. The color changes in Uca pugilator. Proc. Nat. Acad. Sci., 21: 549. CARLSON, S. PH., 1936. Color changes in brachyura crustaceans, especially in Uca pugilator. Kitni/l. Fysiograf. Sdllskap. Lund Fb'rhand.. 6: 1. 1 1. \NSTROM, B., 1935. Preliminary report on the probable connection between the blood gland and the chromatophore activator in decapod crus- taceans. Proc. Nat. Acad. Sci., 21: 584. KKKBLK, F. W., AND F. W. GAMHI.K. 1X99. The colour-physiology of Hippolyte varians. Proc. A'<»y. Soc. Land.. 65: 461. KLEINHOLZ, L. H.. 1936. Crustacean eye-stalk hormone and retinal pigment mi- gration, liiol. null., 70: 159. KOLLER, G., 1928. Versuche iiber die inkretorischen Vorgiinge beim Garneelen- farbwcelisel. Zeitschr. Tcrgl. Physiol.. 8: 001. KROPP, B., AND E. B. PERKINS, 1933. The occurrence of the humoral chromato- phore activator among marine crustaceans. /.'/»//. sci. dc la l:rancc ct Bclg., 48: 30. POUCHET, G., 1X76. Des changements de coloration sous 1'influence des nerfs. Jour, dc I'Anat. ct Physiol. , 12: 1, 113. TAIT, J., 1910. Colour changes in the isopod, Ligia oceanica. Jour. Ph\siol., 40: xl. \VI.IIF.K, M., 1X81. Anatomisches iiber Trichonisciden. Arch, mikros. Anat., 19: 579. WKI..-II, J. H., 1930. Diurnal rhythm of the distal pigment cells in the eyes of certain crustaceans. Proc. Nat. Acad. Sci., 16: 3X6. WELSH, J. H., 1935. Further evidence of a diurnal rhythm in the movement of pigment cells in eyes of crustaceans. liiol. null.. 68: 247. WELSH, I. H., 1936. Diurnal movements of the eye pigments of Anchistioides. B'wl. Jiull., 70: 217. THE MITOTIC RATE IN TADPOLE SKIN AFTER REPEATED INJURY1 JOHN ANDREW CAMERON2 (From flic Department of Zoology, University of Missouri, and from the Marine Biological Laboratory, Woods Hole. Mass.) The epidermis of frog tadpoles usually shows a very low rate of mitotic division (Cameron, 1936a, 1936ft)- The present report is based on an attempt to set up a situation highly favorable to active mitosis. It was thought important to ascertain whether the consistently low rates obtained had been determined by specific inhibiting factors related to the age and development of the animal, the time of year, the existence of mitotic rhythms, or the manner of feeding and maintaining the tad- poles used in previous work. MATERIAL AND METHODS Bullfrog tadpoles about six centimeters long were used. Care was taken to reject those showing signs of metamorphosis. Each was kept in a one-liter beaker half full of water, the beakers being surrounded by a water bath at 25° C. Eight 150-watt Mazda lamps, inside frosted. with reflectors, were grouped 18 inches above the water level in the beakers. Constant illumination was provided from 6:00 A.M. to 9:00 P.M. daily. Black curtains excluded stray light during the remaining hours. The bath and lights, with their electrical control system, were kindly supplied by Professor Albert Saeger. After the tadpoles had been 24 hours in the bath, the posterior half- centimeter of the tail of each was cut off and fixed in Bouins fluid. Successive half -centimeter pieces were taken at 24-hour intervals. Each cut was approximately at right angles to the tail axis. Beginning with the second amputation each piece of tissue had for its anterior face a freshly cut surface and for its posterior face a surface which had been cut 24 hours earlier. The four lateral faces were covered with original epidermis which had contributed cells for the covering, by migration, of one or more " posterior face " surfaces. 1 This study was aided by a grant from the Committee on Radiation of the National Research Council to W. C. Curtis. - Research Fellow in Biology, Harvard University. Fellow of the General Education Board. 37 JOHN A. CAMERON Each block of tissue was cut into vertical sagittal sections 10/i thick and stained with Mayer's luemalum and orange " G." Thus each microscopic section of any piece, after the first or tail-tip piece, had along its posterior border a surface over which cells had migrated dur- ing the previous 24 hours, unless the. section was a lateral surface section. TABLE I Data from a tadpole in which each injury was covered by migration within twenty-four hours. Days Previoii- injin Average cells per section Sei i ion- counted Mitoses counted Mitoses per 10,000 cells 1 0 12,800 35 51 1.1 2 1 2,100 26 6 1.1 3 2 2,340 30 9 1.3 4 3 4,000 36 18 1.3 5 4 1,800 49 8 0.9 6 5 2,400 45 11 1.0 7 6 950 60 5 0.8 8 7 2,300 64 12 0.8 TAHI.K 1 1 Data from a tadpole in which all but the last two injuries were covered by migration within twenty-four hours. Days Previous injuries Avrrau" rrlls per section Sections counted M it oses counted Mitoses per 10,000 cells 1 0 10,800 14 56 3.7 2 1 10,040 11 35 3.2 3 2 4,800 12 18 3.1 4 3 2,150 13 5 1.8 5 6 4 5 3,800 5,100 21 22 15 22 1.9 2.0 7 6 3,800 31 151 13.2 8 7 3,50(1 42 732 50.0 The mitotic figures in all the epidermis of every fifth section of each piece were counted at a magnification of 660 X- Mitotic figures were fiiund in the original epidermis and very rarely in "new'' epidermis formed by migration within 24 hours after injury. Kates per 10,000 cells have therefore- been taken from the ratios of total mitoses to nuin- b'-r of Dun migrated epidermal cells. Average numbers of cells per lion were obtained by counting selected sections from the region midway between the first lateral and the center sections of each piece of tissue. MITOTIC RATE AFTER REPEATED INJURY 39 Findings Tables I and IT are samples of the results obtained, one for each of the two classes into which the tadpoles studied can be logically separated. DISCUSSION The experimental animals, living under conditions favoring a high rate of general metabolism, were subjected to successive injuries each requiring a greater number of cells to cover it on account of the in- creasing diameter of the tail cephalad from the tip. The epidermal cells anterior to any injury were subjected to some degree of stimula- tion to migrate over the cut surface and to divide, assuming that in- jury is the source of a stimulus to mitosis. The increase, if any, in relative frequency of mitosis in regions anterior to and near the suc- cessive wounds might be taken as an index of the degree to which mitosis supplied the new cells required. In the first animals studied no increase in the mitotic rate was found. In cases where seven injuries were followed by seven complete coverages (see Table I), the rate remained around 1 in 10,000. Table I is a sample of counts made by K. O. Mills. The report (Cameron and Mills, 1936), made at the General Meeting of the Marine Biological Laboratory for 1936, was based on these data. The rate here is definitely lower than the rate found in adult frog skin adjacent to areas injured by X-rays (Cameron, 1936fr, Table I 3). There were some individuals in which the latest and most extensive injuries were not completely covered by " new ': epidermis after 24 hours. Counts of these specimens show that the mitotic rate rises sharply in the neighborhood of areas not covered in the usual manner. Table II is an example of records of this class. The injured areas of the second through the sixth pieces were covered, the seventh had an uncovered area about one-third the diameter of the notochord, and the eighth had an uncovered area about the diameter of the notochord. The mitotic rate in the seventh is about five times the average rate of the previous six, and the rate of the eighth about twenty times the same average. The conditions associated with incomplete coverage within 24 hours may then be credited with a twenty- fold increase in the mitotic rate, and it is inferred that the low rates in the other cases indicate that cells were being supplied through migration. The same general picture is found in other specimens. 3 This table is for 10,000 cells, not for 1,000. The figure 1,000 in the title is~a misprint. 40 JOHN A. CAMERON It is also possible that there is simply a cumulative or additive effect of all the injuries to a given tadpole which sets off a period of division as soon as the required threshold is attained. Certainly the conditions leading to failure of rapid coverage of the injured surface are closely related to those producing an increased mitotic rate. It also seems clear that even skin which has maintained a very low rate for a long period can be stimulated to active proliferation. SUMMARY Successive half-centimeter pieces were cut each day for eight days from the tails of bullfrog tadpoles. The epidermis maintained a very low rate of epidermal mitosis despite the great loss of cells by migration over the injured surfaces. Conspicuous exceptions with relatively high rates were found in cases where the epidermis failed to cover the seventh or eighth wound within twenty-four hours after injury. Here the mitotic rate reached a value twenty times the previous average rate. LITERATURE CITED CAMERON, J. A., 1936a. The origin of new epidermal cells in the skin of normal and X-rayed frogs. Jour. Morph., 59: 327. CAMERON, J. A., 19366. Mitosis during the healing of X-ray burns. Radiology, 27 : 230. CAMERON, J. A., AND K. O. MILLS, 1936. Behavior of frog tadpole epidermal cells duriiiii seven successive regeneration periods. (Abst.) Biol. Bull., 71: 405. OBSERVATIONS ON TWO TYPES OF RESPIRATION IN ONCHIDIUM * LESLIE B. AREY DEPARTMENT OF ANATOMY, NORTHWESTERN UNIVERSITY MEDICAL SCHOOL Onchidiuni is a small, naked, pulmonate mollusk which is remarkable because of several conspicuous peculiarities. The habitat of Onchidiuni is littoral-marine and the animal divides its existence between life in the water and in the air. A daily sally from the cavernous nests in the eroded rock in which these animals live is followed by a return governed by precise homing behavior (Arey and Crozier, 1921). The mantle of some species bears eyes with retinulae of the inverted type — a condition unique among gastropods (Semper, 1877; Stantschinsky, 1908; Hir- asaka, 1912). The presence of lungs, except as analogues adapted from a portion of the kidney, was long denied, but this error has since been corrected. Cuvier (1805) was the first to describe lung cavities in Onchidiuni, but this observation gave way later to other interpretations. Milne- Edwards (1857) considered the pouch hitherto described as lung to be kidney, like that of other gastropods. This view was revived by v. Ihering (1877), who interpreted the so-called lung of Onchidiuni as being comparable to the broadened cloacal portion of the kidney of some other marine forms. This he thought had undergone a partial func- tional change from a primarily secreting organ to a respiratory sac. The conclusions of v. Ihering were enthusiastically supported by Joyeux-Laffuie (1882). On the other hand, Semper (1876) argued for the existence of two separate though juxtaposed organs and in this opinion he was joined by Bergh (1895) and especially by v. Wissel (1898). As the results of the investigations of these latter workers, the Onchidiuni family became removed from the nudibranchs and now rests securely once more among the pulmonates. As to the mode of respiration employed, Cuvier considered it to be solely pulmonary. In opposition stand Ehrenberg (1831), Milne-Ed- wards (1857), v. Ihering (1877), Vaillant (1871), Joyeux-Laffuie (1882) and others who have argued for two types of respiration, cu- taneous and pulmonary, but with the latter playing a subordinate role. 1 Contribution No. 241. The observations were made at the Bermuda Biologi- cal Station for Research, Inc. 41 42 LESLIE B. AREY Only v. Wissel has considered the lungs to be the major organs of respiration when the animal is out of water. Recently an opportunity has come to supplement earlier studies (1919o ; 1919/> ; 1921 ) on the natural history and behavior of Onchidimn floridanum Dall, an Antillean species common at Bermuda. Among the new observations are some dealing with respiration. Since the opinions expressed hitherto have not always been drawn from the living animal, the recording of certain direct observations should be of value and in- terest. Onchidimn floridanum spends the major part of its existence in eroded cavities in the soft Eolian limestone which constitutes the rocky shores of Bermuda. These cavernous nests, often the size of one's fist or larger, are intertidal in position and communicate to the outside by small, cleft-like apertures so overgrown with Mod Joins as to be wholly inconspicuous. It is remarkable to observe how Onchidia, some 2 cm. in length, can insinuate themselves through such narrow openings. Once a day, but only during daylight hours, the animals of a colony leave their iK'st to wander about on the neighboring intertidal rocks, feeding on the felt-work of alga which carpets them. Emergence is at a fixed time when the receding tide has dropped a definite distance below the nest. The feeding period lasts for an hour or so, after which a return is made to the appropriate home nest through the operation of precise homing activities, as has been related elsewhere (Arey and Crozier, 1921). The emergence and return of a colony involves a roughly simultaneous group activity, although it seems clear that not all members of a communal group come out at every feeding period. Cer- tain it is that on some occasions relatively few individuals are to be seen in the open, while during stormy weather or when the proper tidal period comes too close to darkness all animals may stay within their places of refuge. During the periods of several hours, twice a day, when the nests are left above tidal level, the Onchidia are for the most part ex- posed to the air, although retained water and water seepage from the porous rock keep the nest very wet. By contrast, during the feeding period the animals are commonly exposed to the direct rays and baking heat of the semitropical sun. The mantle of Onchidimn flondanuni, like most of its allied species, U thickly studded with large numbers of tiny, rounded elevations which ran properly be designated as short papilke. These are far smaller than the long, branching ones figured by Plate (1894) for 0. sai'i of the constitution sn/-\- give rise to reduced eggs, sn in constitution, which develop normally after fertilization with either sn+ or sn sperm. It is clear, then, that the two types of eggs, sn produced by a sn female and sn produced by a sn/-{- female, are differentiated by the constitution of the mother. This differentiation might be brought about directly in the developing oocyte or indirectly by means of an influence of other maternal tissues. The experiments reported below were carried out in an attempt to differentiate between these two possibilities. Ovaries from wcsn larvae were transplanted to Ft wild type larvae derived from the cross of the two inbred wild stocks. Florida and Swedish-c. From 37 larvae into which had been injected wfsn ovaries, 27 gave rise to adult females. These were mated individually to w''sn male-. Two females were lost before any record of the eggs was made, and ( .IK before dissection. Of the 24 females on which complete rec- ords were secured, it was found that 13 laid two types of eggs, one type judged by appearance to be sn, the other wild type. The remain- ing 1 1 females laid only wild type eggs. When these females were OVARY TRANSPLANTS IN DROSOPHILA 49 dissected, it was found that the attachment of the ovaries to the ducts agreed in every case with the egg-hying record (Tank- I). Of the 11 females which laid only wild type- eggs, 5 had no developed implanted ovary. Only eggs of normal wild type appearance hatched and all hatched eggs that were tested by growing the larvae to maturity gave rise to wild type flies. It is clear that none of these normal eggs orig- inated in the implanted sn ovary. TABLE I Summary of the results of transplanting singed ovaries to wild type female hosts and the reciprocal. In this and following table wild type is designated by +. Number Total progeny Result of dissection, implant Implant Host of mature Eggs laid females Phenotype Number Attached Free Not de- termined wesn + 13 + and sn + 746 13* 0 wesn + 6 + + 624 It 5 + wesn 6 + and sn + 321 4 0 2 + Wsn 4 + + 26 2 0 2 + Wsn 3 sn — 0 0 2 1 * In one female three ovaries were attached to ducts, in a second, only one wild type ovary present; see footnote to Table II. t Three ovaries attached to ducts; see footnote to Table II. In making the reciprocal transplants, wild type ovaries transplanted to w'-'sn larvae, 34 operations were made. Seventeen adult females were obtained which were mated to ivcsn males. A record of the eggs laid was secured on 13, and the attachment of the ovaries on 8. The wcsn females were less vigorous than wild type females and were inclined to stick to the food mass or on the moist walls of the vials. As indi- cated in Table I, the egg-laying record and the dissection records agree in all cases. Here again only the eggs which were judged to be wild type gave rise to offspring and these were all wild type ; they must have originated in wild type implanted ovaries. This result shows that, aside from the production of abnormal eggs, the reproductive apparatus of a singed female is able to function in an essentially normal way, i.e., the genital ducts, accessory glands, and external genitalia of such a female are functionally normal. It is evident from the results just described that there was no de- tected influence of host on implant or of implant on host, i.e., the im- plants developed according to their own genetic constitutions with no apparent influence of the genetic constitution of the host tissues. So far as the experiments go, the differentiation is intragonadal in nature, 50 C. W. CLANCY AND G. W. BEADLE but the possibility of extra-gonadal influences, acting before the time at which the transplants were made, is not excluded. TRANSPLANTATION OF FUSED OVARIES Fused females lay eggs of normal appearance but which fail to hatch when fertilized by sperm from fused males. According to Lynch (1919) such eggs show signs of development and in very exceptional cases (2 in several hundred) may give rise to larva1 which die at an early stage of development. Lynch (1. c.) observed, however, that if such eggs are fertilized by X sperm from a wild type male (or other not-fused males) normal development occurs. Absence of males from such matings indicates that fertilization by a Y sperm does not bring about normal development. Eggs from fu/-\- females, even though they be fit in constitution after reduction, show normal development when fertilized with normal X, fused X. or Y sperm. Fused males are apparently normal in fertility. As pointed out by Lynch, a fu egg can be made good by something which happens to it before fertilization (development from a -\-/fu oocyte) or, if it develops from a fu/fu oocyte, by being fertilized by a not-fused X sperm. Assuming that the effect of the fu* gene is a positive one, we might express this in terms of gene activity in the following way. A fu egg arising from a fu/fu oocyte lacks something essential for its normal development. If a fit egg arises from a fu/-\- oocyte, this deficiency has been made good before fertilization by the activity of the fii+ gene directly in the oocyte (possibly after fertilization by the activity of the fu* gene in the polar body nuclei), or indirectly by the activity of the fu* gene in cells other than the oocyte. Whatever is deficient in a fu egg arising from a fu fu oocyte can evidently be compensated for by the activity of the fu* gene brought in by an X sperm. The results of transplanting a fu ovary to a not-fused female .should answer the question of whether the fu* gene can influence the oocytes indirectly through tissues outside the ovary. White fused (w fu} ovaries were implanted in W'sn hosts. Singed females were used as hosts so as to be able to distinguish the eggs of an implanted ovary from those of the ovaries of the host. Of 37 w'sn females in which fused ovaries had been implanted, 18 laid both singed and fused eggs, 3 fused eggs only. 14 singed eggs only, and in 5 cases no eggs at all were laid. Females laying fused eggs and mated to fused males gave no offspring. After three days, 6 females that had been laying fused eggs were isolated and remated to wild type males. From the eggs laid during the next three days some larva* hatched. OVARY TRANSPLANTS IX DROSOPHILA 51 These were collected and placed in a culture bottle en masse. In all, 28 flies emerged from larvae collected in this manner ; all were wild type females. In this experiment, as in the case of the analogous experiment with singed ovaries, the implanted fused ovaries behaved in all respects in the same way as would have been expected had they completed develop- ment in their original environment. Conclusions similar to those ar- rived at in the experiments with singed are indicated. EXPERIMENTS WITH FEMALE-STERILE Females homozygous for the gene fcs have rudimentary ovaries ; they can be distinguished from normal females by dissection one day or more after eclosion. Apparently the oocytes fail to grow normally. Homozygous fcs males are fertile. Transplantation of wild type ovaries to fcs females show that such an ovary can become attached to an oviduct of the host and function normally. Of 7 such females in which the implanted ovary had de- veloped, 2 produced wild type eggs which, when fertilized with sperm from fcs males gave rise to wrild type adult flies. Five females showed the implanted ovary normally developed but unattached. Of 4 wild type females in which implanted female-sterile ovaries had been implanted, 2 had the implant attached and in 2 the implant was unattached. In both cases, after aging of the females, the im- planted ovaries showed no more development than is characteristic for the ovaries of normal fes females. It is clear from the two instances in which an implanted fcs ovary replaced a normal wild type ovary of the host that fcs ovaries are capable of competing successfully in attachment with normal ovaries. The experiments with female sterile ovaries show that, under the conditions of the experiments and with respect to the characters under consideration, the development of ovary implants is autonomous or independent of the genetic constitution of the host. In experiments in which it is desired to recover eggs from ovaries grown in hosts of a different genetic constitution, the character female- sterile promises to be of considerable value. When one uses fcs fe- males as hosts, no eggs develop in the ovaries of the host and all recov- ered eggs therefore originate in the implanted ovary. Furthermore, the limited development of fcs ovaries minimizes the unfavorable effects of mechanical crowding often apparent in females with three normally developed ovaries. Females homozygous for the fes gene with an im- planted normal ovary have been observed to lay an average of more 52 C. W. CLANCY AND G. \V. BEADLE than 25 eggs per day for an interval of 10 days with no evident signs of decreased production at that time. FREQUENCY OF ATTACHMENT OF IMPLANTED OVARIES As pointed out ahove it is possible, by using donors and recipients which carry a gene difference at the singed locus, to determine by dissection of aged females whether or not an implanted ovary has established connection with an oviduct of the host. Experiments were carried out which have a bearing on three ques- tions: (1) Is a singed ovary on a par with a wild type ovary in estab- lishing a connection with an oviduct? (2) Do variations in the differ- ence in genetic constitution between ovaries competing for attachment influence the result? (3) What is the effect, on competition for at- tachment, of a difference in development of competing ovaries (age difference between donor and recipient)? Three stocks were used: 1. F, K'''SH females from the cross of two distantly related wrsn stocks, ic'sn/ClB females mated to w''sn males from the stock yy X W'SH. 2. F, wild type females from the cross of Florida and Swedish-c inbred wild type stocks. 3. Inbred wild type stock Oregon-R-c, made homozygous an unknown number of generations previously by the standard inversion technique. The results of two sets of reciprocal transplants involving these stocks are summari/ed in Table II. In the first pair of reciprocal trans- plants, involving the two outcrossed stocks, the percentages of attach- ment are 75.0 and 65.3. Neither of these is significantly different sta- tistically from the 66.7 per cent expected on the basis of random at- tachment, and both are higher than the percentage (45.4) found by Kphrussi and Beadle (1935) in a series of miscellaneous experiments. Furthermore, the difference between the two values is not statistically significant. The second pair of reciprocal transplants, involving the out- crossed U'rsn stock and the inbred ( >n iMni-R-c wild type stock give per- centages of attachment of implanted ovaries of 44.7 and 41.7. These two values, again approximately equal, are significantly lower than the ci.rn^ponding values obtained in the first pair of reciprocals, and are significantly lower than the 66.7 per cent expected on the assumption that attachment is random. It can be seen that in each pair of reciprocal transplants the fre- quency of attachment of wild type implants is higher than that of singed implants. However, in each instance the difference is of doubtful sta- tistical >ignificance. Combining the two series so as to compare wild OVARY TRANSPLANTS IN DKOSOPHILA 53 type in W'sn with Wsn in wild type, values of 65.3 and 53.6 per cent are obtained, a difference aproximately 2.5 times its probable error. \Yhile it cannot be concluded that there is no inherent difference between singed and wild type ovaries with respect to the chance of their becom- ing attached to oviducts of the host, the experiments fail to demonstrate such a difference and show that, if it does exist, it must be of relatively little importance as compared with other differences. From the data tabulated in Table II it is evident that the results are different with the outcrossed and inbred wild type stocks. Further- more, reciprocal transplants give approximately equal frequencies of TABLE II Summary of data from experiments on frequency of attachment of implanted ovaries. In all instances listed at least two ovaries were attached to the two lateral oviducts of the host. In the calculated frequencies of attachment only the two classes with three ovaries developed, two attached, and one free, were taken into account. Probable errors are given with the calculated percentages. Three ovaries Two ovaries * Percentage Implant Host attachment Implant Implant All Implant No of implant free attached attached * attached implant +Fla/+S-c.... wcsn 15 45 5 2 12 75.0±3.8 wcsn +Fla/+S-c 17 32 6 1 13 65.3±4.6 +Ore-R-c wesn 26 21 4 2 2 44.7±4.9 wesn . . +Ore-R-c 28 20 1 0 7 41.7±4.8 * In this and Table III, 26 instances are recorded of attachment of three ovaries. Two additional cases of such attachment are recorded in the footnotes to Table I. In many of the attachments of this nature one ovary appeared to be imperfectly attached. In no instance was a clear bifurcation of a lateral oviduct observed; often two ovaries appeared to have a common attachment to a single lateral oviduct. In several recorded instances only two ovaries were developed, sometimes both from the host (failure of the transplantation operation or possibly failure of the implanted ovary to develop), sometimes only one from the host (presumably injury to or destruction of a normal ovary during operation). attachment of the implant, high in one case and low in the other. It is pointed out again that in these experiments the donors and recipients were approximately equal in absolute age (ages controlled to within a period of 2 hours or less). It seems reasonably safe to assume that the differences in frequency of implant-attachment observed between the two wild type stocks is to be attributed to differences in genetic constitution. The difference be- tween the two series and the approximate equality in attachment fre- quency in reciprocal transplants suggest, and are consistent with, the1 following assumptions : 54 C. W. CLANCY AND G. \V. BEADLE 1. Which two of three ovaries will become attached is a matter of chance if the ovaries of the host and the implanted ovary are at the same developmental stage at the time of attachment. 2. If the implanted ovary is at a different developmental stage from the ovaries of the host (either more or less advanced) at the time of at- tachment, the implant will be at a disadvantage in competition with the ovaries of the host. 3. Development of ovaries takes place at different rates relative to absolute age in females of different genetic constitutions. It should be possible to test directly the first two of these assump- tions by varying the relative ages of donors and recipients. There is no way of determining except by trial whether the low value obtained in the series involving the Oregon-R-c wild type stock is the result of TABLE III Results of transplants involving wfsn and Oregon-R-c (Ore-R-c) with an age difference between donor and recipient. Three ovaries Two ovaries Implant Host Implant ymmKer or older Age diffei - ence Im- plant Im- plant All at- Im- plant No im- Percentaue attachment of implant free at- i.n 'ii-d tached at- tached plant hours Ore-R-c. . . wesn Younger 16-23 6 27 2 2 11 81.8±4.5 Ore-R-r Older 16 24 o 1 1 0 (} * wesn Ore-R-c Younger 21-26 53 53 7 0 16 25.4 ±3. 5 * Hosts died after pupation, see text. slower or of faster development of the ovaries of Oregon-R-c relative to those of the wcsn stock. If the former is the explanation, then by implanting ovaries from Oregon-R-c females into older wrsn females, the frequency should be decreased. I f the latter is the case, then ovaries from Oregon-R-c females implanted into older i^'sn females should >how an increased frequency of attachment. Four combinations with a given age-difference are obviously possible with two given stocks. Table III sumniari/.e* the attachment frequencies ot implants in ex- periments made with the Oregon-R-c and w''sn stocks where the aver- age age difference between donor and recipient was approximately 20 hour-, h is seen that Oregon- l\-c ovaries transplanted to older i^'sn females njve a significantly higher frequency of attachment than was obtained in the comparable experiment without an age difference (81.8 per cent a> compared with 44.7 per cent ). The reciprocal of this trans- OVARY TRANSPLANTS IN DROSOPHILA 55 plant, wcsn ovaries in younger Oregon-R-c females, should, following the assumptions made above, likewise give a high attachment frequency. However, when the experiment was made it was found that the hosts lived for more than 24 hours, appeared to pupate normally, but for the most part died before maturity. Four sets of transplants were made at separate times, and in a total of 149 operations, only two females reached maturity. As compared with this high mortality, the average mortality for all other experiments involving an age-dim-mm- was 32.2 per cent; the mortality for the transplantation operations involving no age-difference was 45.5 per cent. Apparently, for some reason, the im- plant was lethal to the host in this particular case. In the only other combination attempted, wesn ovaries implanted in older Oregon-R-c hosts, the frequency of attachment was lower than in the same combination without an age difference (25.4 as compared with 41.7 per cent, a difference of approximately 2.9 times its probable error). The two successful combinations with an age difference are con- sistent with the assumptions listed above and indicate that, if these as- sumptions are correct, the ovaries of Oregon-R-c females are develop- mentally further advanced than are those of u>csn females of a corre- sponding absolute age. However, in the absence of the reciprocals of these age-difference experiments, the assumptions are by no means proved to be correct. The lethal result in the one combination is quite unexplained and suggests either that the assumptions are incorrect or that they fail to take into account all of the factors concerned. SUMMARY The development of singed ovaries transplanted to wild type fe- males in the late larval stage shows autonomous development. Eggs recovered from such females have the characteristic shape of eggs from singed females and they fail to give rise to larvae. Wild type ovaries grown in singed hosts likewise show autonomous development. Viable eggs can be recovered from such ovaries ; they give rise to wild type offspring (females heterozygous for zuesn) when fertilized by wesn sperm. Fused ovaries grown in singed hosts have characteristics not de- tectably different from such ovaries grown in their normal position. Recovered eggs fertilized by fu or by Y sperm fail to hatch, but those fertilized by not-fu X-carrying sperm give rise to normal females hetero- zygous for fu. Ovaries from female-sterile females grown in wild type hosts may become attached to the oviducts of the hosts, competing successfully 56 C. W. CLANCY AND G. \V. BEADLE with normal ovaries, but they remain rudimentary as they do in their normal genetic surrounding. Wild type ovaries grown in female-sterile females show autonomous development. Viable eggs giving normal development are recovered following attachment of the implant to the oviduct of the host. Using a single outcrossed stock of w''sn and two wild type stocks, one outcrossed and one inbred, reciprocal ovary transplants show : ( 1 ) that the frequency of attachment of tin- implant varies with different genetic stocks, and (2) that, under the conditions of the experiments and with the numbers involved, there is no statistically significant dif- ference in the frequency of implant-attachment in reciprocal transplants. Transplants in which donors and recipients were different in abso- lute age show that the frequency of attachment of the implant can be varied, either increased (in certain combinations) or lowered, by vary- ing the relative age> of donors and recipients. In one combination in which ovaries were implanted to hosts younger than the donors, there wa> apparently a lethal interaction such that most of the hosts died after pupation. The bearing of the age-difference experiments on the differ- ences in frequency of implant-attachment observed with different stocks is considered. The application of certain of the results summarized above to the n>e of gonad transplants in ftrosopliila as an experimental tool an- pointed out. LITERATURE CITKD Epnurssi. I!., AND G. W. BKADI.K, 1935. La transplantation dcs ovaires chez la Drosophile. Hull. Hiol. Fr. /*,-/./.. 69: 492. EPHRUSSI, I'.., AND (i. \V. I'.i \m i . l''.if>. A technique of transplantation l<>r Drosophila. Am. Nat., 70: 21S. LYNCH, C. .!., 1919. An analysis of certain cases of intra-specific sterility. ( it-net ics. 4: 501. MOHR, O. L., 1922. Cases of mimic mutations and secondary mutations in the X-chroiiiiiM>mc of I )n>sopliila inelaim^aster. Zci/sclir. inil. slhst. 11. J'crerb., 28: 1. EYES OF DEEP SEA CRUSTACEANS I. ACANTIIEPHYRID^E J. H. WELSH AND F. A. CHACE, JR. (From flic U'oods Hole Oceanographic Institution 1 and the Biological Laboratories, Harvard University) The depths of the ocean impose upon the animals which live there a very different set of environmental conditions from those found at the surface. Absence of sunlight, low temperature, high pressure and vis- cosity, and a somewhat different chemical composition of the sea water necessitate morphological and physiological modification in the organ- isms which inhabit deep water. Some of these modifications are well known. For example, if we select the more obvious features which are associated with low light intensity, or even complete absence of sunlight, we must include the presence of photophores, peculiar pigmentation of the body, and adaptation of the eyes for vision at a very low intensity of illumination. The photophores and eyes of certain groups of deep- sea crustaceans and fishes have been studied in some detail in the past and are reasonably familiar to biologists. One of the first and best known studies in which the structure of the eye was related to depth was that of Beddard (1884) on the isopod genus Serolis. A similar study was made later (Beddard, 1890) on the genus Arcturus. Smith (1886) and Henderson (1888) had already noted the tendency toward degeneration in the eyes of deep-water crustaceans, but neither had made a detailed study of these eyes. Chun's (1896) study of the very unusual eyes of deep-water euphausiids was made from collections of the " Challenger " as were Beddard's. From the collections of the " Valdivia," Doflein (1904) obtained material for a valuable study of the eyes of deep-sea crabs. Dohrn (1908) described the eyes of a few other crustaceans taken by the 'Valdivia." More recently Hanstrom (1932-33) has had an oppor- tunity to study the eyes of deep-water crustaceans. The problem which confronted these workers, and many others who have made more casual observations, was to explain why some forms found in deep water have well-developed and apparently functional eyes while others are completely blind or have very degenerate eyes. This still remains the most interesting problem in connection with a con- 1 Contribution No. 126. 57 58 J. H. WELSH AND F. A. CHACE, JR. sideration of the effect of depth and hence of diminishing light on the eye. There are many more problems which the older investigators would have doubtless solved if they had had more material from known depths, living material, and a more constant supply. The regular operation of the research vessel " Atlantis " of the Woods Hole Oceanographic Institution has made- it possible to study further the animals of the deep sea and has led to the planning of a program in which the first step has been to investigate more fully the modifications with depth in the eyes of certain of the crustaceans. What happens to the size of tin eye. the number of ommatidia, the amounts of reflecting and screening pigments, pigment migration, and the structure of the rhabdomes with increasing depth? What is tin- relation between the development of the eye and the presence or absence of photophores? Such questions, dealt with in an introductory manner, using for material certain representative deep-sea acanthcphyrids or prawns, are those we wish to discuss in the present paper. Final con- clusions must await the study of many more types of crustaceans from known depths, but certain tendencies are noticeable in the work thus fat- carried out. The authors are indebted to several members of the Woods Hole Oceanographic Institution, particularly to Dr. 11. li. Bigelow, and to Mr. 1>. B. Leavitt from whose collections, made in a study of the verti- cal distribution of deep-water plankton, some of the material for the present study was obtained. The expense of preparing the eyes histo- logically has been cared for in part by a grant from the Milton Fund of Harvard University. METHODS The first problem in understanding the effect of the physical en- vironment on deep-sea animal is to obtain a series of a given species from known levels, sufficiently large so that it is possible to determine the average level at which such a species lives. The use of open nets is unsatisfactory as animals are caught at all levels, and only after a large number of hauls at a given station can an idea of tin- vertical distribution of a species be determined. Xets which may be sent down to the de- sired level closed, then opened and towed at that level, and finally closed before being brought to the surface are almost a necessity. To operate two to live such nets, particularly when towing at depths of one to two miles, presents mechanical problems which have been quite adequately solved. The closing net used in deep-water tows t mm Atlantis previous to l''.V, is described in a paper by Leavitt (1935). The type EYES OF DEEP SEA CRUSTACEANS 59 of net used during- the 1936 cruise was modified somewhat and was found to he well adapted for collecting material alive or in good con- dition. It will he descrihed in a later paper. The locations of stations from which material was obtained in 1933 and 1934 are given by Leavitt (1935). During July, 1935, two stations were made in the Sargasso Sea: Station No. 2462 (42° 29' X. and 70° 21' W.) and No. 2463 (42° 27' N. and 70° 14' W.). During late Au- gust, 1935, one station was made on the inshore side of the Gulf Stream: No. 2475 (38° 25' N. and 71° 04' W.). In September, 1936, collections were made at Station 2666 (39° N. and 70° W.) and at Sta- tion 2667 (35° 40' N. and 69° 36' W.). Obtaining deep-water animals living and in good condition is a second difficulty. Where there may be a warm surface layer, as on the slope side of the Gulf Stream, the animals coming from deeper layers are subjected to a considerable increase in temperature which, it is be- lieved, is much more damaging than the reduction in pressure. The following temperatures were found in early September, 1936. Station 2666 Station 2667 Slope side of Gulf Stream Sargasso Sea Depth Temp. °C. Temp. °C. Surface 21.0° 26.2° 200m 9.7° 18.3° 400m 5.4° 17.8° 800m 4.5° 14.9° 1000m 4.2° 11.7° 2000m 3.6° 4.3° It may be seen that while there was a difference of 5.2° C. at the sur- face, at 400 m. there was a difference of 12.4° C. Many more of the larger crustaceans were living when they arrived at the surface at Sta- tion 2667 than at Station 2666. During the two trips in 1935 an opportunity was had for the first time to study living forms. Among the acanthephyrids only one spe- cies, Systellaspis dcbilis, was obtained alive in any considerable numbers. The only requirement for keeping this form living was found to be a low temperature. At 10° C. they survived for three days, at the end of which time they were killed ; therefore, at that time it was not known how much longer they would have lived. During the 1936 cruise a cooling system was on board which made it possible to maintain tanks of sea water at 5° C. Many more forms were taken alive and several species of crustaceans were kept living for the duration of the cruise. Apparently it is possible to maintain many deep-water forms alive if 60 j. H. WELSH AND 1-. A. CHACE. JR. they arc kept at a temperature relatively the same as that where they normally live. The material used for histological study was fixed in Benin's im- mediately on removal from the nets or, in cases where it was desired to observe pigment migration, was kept illuminated for a period of time before fixing. It was found quite necessary to cut through the carapace in order to obtain sufficiently rapid penetration of the fixative. The paraffin method was used in sectioning the eyes. In estimating relative amounts of pigment it was found desirable to leave some sec- tions of each eye unstained. Ehrlich's luematoxylin and a counter of eosin were employed for staining those sections which were used in studying the general structure. MATERIAL Although certain species of the Acanthephyridae are so numerous that the possibilities of fishing for these prawns on a commercial scale have been suggested, it was as late as 1S81 before more than two species of the family were known. Today, due to the many deep-sea expedi- tions in the past fifty years, the family is represented by six genera and forty-five species. The acanthephyrids normally inhabit the deeper parts of the sea where the penetration of sunlight is practically im- measurable— the so-called " red prawn-black fish " region. Like a number of other deep-sea animals which have been known to M-icnce for only a short time, the- acanthephyrids are a comparatively primitive group. The biramous form of the legs, one of the most dis- tinctive characters of the family, ranks them among the most primitive members of the decapod Crustacea. Their color, with but a few ex- ceptions, is a uniform dee]) crimson red. In other respects the species may differ so noticeably from each other that a cursory examination would scarcely lead to their inclusion in a single family. The integu- ment varies from a hard, polished, armor-like shell to a membranous skin which is displaced or torn from the slightest handling; the rostrum or "head-spine" is typically long and slender, but in many species it is almost entirely lacking; the legs mav be short and comparatively stout. or long, slender and fragile, or they may be modified, as in Ephyrina, into broad, lamellate appendages; and the eggs are cither so small that mam hundreds may be attached to a single teniale, or so large that twenty-live would be a burden to even the larger species. It will soon be seen that this diversity of form is likewise illustrated by the eyes of these' prauiis. Since most of the species are strictly bathypelagic or- •EYES OF DEEP SEA CRUSTACEANS 61 ganisms, they probably obtain tbeir food from the detritus that is con- stantly raining down from the swarms of minute plants and animals at the surface of the sea. They arc peculiarly adapted to strain this food from their surroundings; the thoracic legs are all provided with numerous long hairs and spines which apparently form a very efficient, sieve-like basket when the legs are held curved beneath the body. In the genus Ephyrina the legs are strikingly wide and flat so that, when held in position, they perform their function in much the same wax- as the baleen of the whale-bone whales. This modification of the legs for food-gathering is of particular interest when one realizes that the thoracic limbs of most shrimp-like Crustacea are used primarily for walking on the sea-bottom and play no direct part in the swimming movements of the animal. Since many of the acanthephyrids probably spend their entire lives far above the sea-floor, the thoracic legs would be only a hindrance to the progress of the prawn through the water if they were not modified to perform a function entirely apart from that for which they were originally designed. There are, of course, few barriers to the dispersion of bathypelagic organisms which inhabit a world-wide zone of comparative uniformity of temperature and salinity. Many of the Acanthephyridse, as in most families containing a like number of forms, are known from a very few specimens and little can be concluded about the distribution of those forms at present ; but of the commoner ones, some are practically cos- mopolitan, some seem to have a discontinuous range, while still others are confined to a reasonably small area. The Indo-Malayan region has the largest representation of these species, if our present records are reliable, with the North Atlantic second in importance. It is a curious fact that there are at least three reasonably common species in the Indo- Malayan region which so far have not been encountered elsewhere, but all three of these species show a very close relationship to three other forms which have been found almost everywhere except in the Indian Ocean. Much more data must be accumulated before the factors re- sponsible for the specific isolation of bathypelagic organisms are known. As regards the vertical distribution of the group the available data are likewise incomplete. However, of the forty-five species, thirty have been taken with mid-water nets and must be bathypelagic for at least part of the time. Further investigation will undoubtedly reveal that at least some of the remaining fifteen species seldom go down to the sea-floor. On the other hand, three stout-legged, heavy-bodied forms which have always been taken with the dredge or trawl can safely be termed benthonic animals. 62 J. H. WELSH AND F. A. CHACE, JR. Investigations undertaken in the past few years by Mr. Leavitt on ' Atlantis " have yielded invaluable information regarding the actual depths at which we may expect to find acanthephyrids in the North At- lantic. The accompanying tables (Tables 1. II. and III) give the depths at which three of the most common Atlantic species have been taken with closing nets. Incomplete as these records are, it is quite evident that Hymenodora glacialis normally frequents a deeper zone than either of the other species, a fact which agrees with earlier data. Acanthc- TABLE I Depths at which specimens of Acanthephyra purpurca have been taken with closing nets by the "Atlantis." Deptli in meters "Atlantis" Station Number of specimens Total specimens 300 1737 1 1 400 2263 10 10 800 2263 2462 2463 65 7 2 74 920 2216 10 10 1000 2260 2263 1 2 3 1400 2263 1 1 1600 2263 1 1 1800 2216 2 2 2200 2263 1 1 2600 2260 2 2 phyra purpurca has been taken in open nets from the surface, where one -I'ccimen was collected with a dip-net down to 2800 meters, but tin- majority have been found between 400 and 1000 meters. //. (jlaclaUs has been found on two occasions at the surface in the Arctic, and some numbers ot that species have been found in the stomachs of arctic sea- birds. The shallowest trustworthy record for the species taken with midwater nets, however, is one- in which ten specimens were taken in 750 met ••]-,, and by far the largest number of specimens have come from EYES OF DEEP SEA CRUSTACEANS 63 TABLE II Depths at which specimens of Systellaspis debilis have been taken with closing nets by the "Atlantis." Depth in meters "Atlantis" Station Number of specimens Total specimens 400 2263 15 15 600 2260 1 1 800 2462 9 9 1800 2263 1 1 depths greater than 1000 meters. This species has also been taken with a closing net by the " Valdivia " Expedition in over 4200 meters in the South Pacific. Systellaspis debilis apparently frequents slightly shal- lower depths than the other two species, regardless of the fact that it is the only one of the three which has not been found at the surface. It has been collected with open nets between 32 and 2000 meters and ap- pears to be most abundant between 150 and 500 meters. At present little is known of the daily vertical migration of these TABLE III Depths at which specimens of Hymenodora glacialis have been taken with closing nets by the "Atlantis." Depth in meters "Atlantis" Station Number of specimens Total specimens 1000 2463 2 2 1200 2462 3 3 1400 2263 4 4 1600 2263 2475 2 3 5 2000 1737 3 3 2200 2263 1 1 2600 2260 2 2 2800 to 3200 1739 1 1 64 J. H. WELSH AND F. A. CHACE, JR. species. The results of past expeditions seem to indicate a very slight movement toward the surface at night in ./. pnrpurca and A", debilis but absolutely none in //. glacialis (Murray and Hjort). An examination of the external features of the eyes of these and other acanthephyrids offers a partial explanation of the above findings. The form of the eyes, although extremely diversified in the group as (Figures l-o arc dorsal vieus of the right eye and eyestalk) IMC. 1. Acanthephyra pnrpurca. X 5. FIG. 2. Ephyrina bcnedicti, male, X 4. FIG. 3. Notiistantns Inngiroslris. small female, X 5. FIG. 4. Systellaspis debilis, X 7.5. FIG. 5. Oplophorus t/rinuildii. male, X 7.5. FIG. ft. JJyinciiodom ulacialis. male, X 7.5. a whole, remains fairly constant within the separate genera. The eyes of the species of Acanthephyra ( I;ig. 1 ) are of normal si/.e, generally well pigmented with a dark brown pigment and, as in the other genera, tin- dorsal surface of the eyestalk is provided with an incomplete, deeply pJLMiiented ocellus or accessory cornea, whose function is unknown. In some speeies the ocellus is complete and entirely distinct from the EYES OF DEEP SEA CRUSTACEANS 65 true cornea, while in others it may 1>e partially fused with the cornea or even entirely absent. The eyes of llphyr'uui (Fig. 2) and Notosto- nins (Fig-. 3) while similar in size and shape- to those of Acanthcphym have a jet-black pigment which remains for many vears even in alcohol. Tn Systcllaspis (Fig. 4) the cornea is considerably larger in proportion to the stalk than in the foregoing genera. The extreme size is attained in the species of Oplophorus (Fig. 5) in which the eye may be actually broader than long, and the cornea is generary set diagonally on the end of the stalk. It is worthy of note that these large-eyed species of Systcllaspis and Oplophorus possibly all possess photophores. The other extreme is found in the eyes of the two species of Hymenodora (Fig. 6). In these forms the corneal portion is reduced to a size which is considerably smaller in diameter than the eyestalk. On recalling the records concerning the vertical distribution, it will be seen that those species which have the largest eyes apparently frequent the shallower layers of water, while those with vestigial eyes, like Hymenodora, may be termed truly abyssal forms. The accompanying graph (Fig. 7), despite inaccuracies that can scarcely be eliminated in measuring tissues so subject to alteration after preservation, clearly indicates that the eyes increase in diameter in proportion to the length of the carapace. It is also quite obvious that the degenerate eyes of Hymenodora increase in size very little with the growth of the animal. Perhaps the most striking fact brought out by these measurements, however, is that the eyes of Oplophorus gri- inaldi'i and Systcllaspis del/His, two forms bearing photophores, are larger in relation to body size than are the eyes of those forms lacking luminescent organs. COMPARATIVE STRUCTURE OF THE EYES Fortunately the three most abundant species of acanthephyrids in the North Atlantic and therefore those whose vertical distribution is best known, represent three possible trends in the development of the eye. Acanthephyra purpurca is found in greatest numbers within the photic zone, and as it shows a diurnal migration this species must be influenced by the penetrating daylight, although of extremely low intensity. Even in the Sargasso Sea, where the water is very clear, the intensity of blue light at noonday may be reduced to 0.5 per cent of the light at the surface, at a depth of 180 meters (Clarke, 1933). The penetration of other components of daylight into Atlantic waters may be found in a paper by Oster and Clarke (1935). A. purpurca is 66 J. H. WELSH AND F. A. CHACE, JR. found far below 180 meters during the day ; hence it must be subjected to a very low intensity of illumination. Systellaspis d chilis has much the same vertical distribution as A. pnrpurca, but this form possesses numerous photophores, and, as has already been pointed out, there is an apparent correlation between the size of the eye and the presence or absence of photophores, for species 2 * li * I X OB/oaAori/jar/ma/a/t _7 r • Acan f/> e/ody 'ra ourourto + riumertff aera q/octo/rj JO IS of Loroaacf in 20 Fir,. 7. The diameters of the eyes of four series of acanthcphyrids are here shown plotted against carapace length. Oplophonis and Systellaspis have photo- phores. of Systellaspis and Oplophorus have larger eyes in proportion to body size than the other genera of acanthephyrids which lack photophores. The ability to emit clouds of luminous material is possessed by certain of the deep-sea prawns, as has been mentioned by Alcock (1902) and by Beebe, who observed this phenomenon frequently in his bathy- sphere descents. So little is known of this interesting method of il- luminating the surroundings that it is not possible to hazard a guess as to its effect on the development of the eye. EYES OF DEEP SEA CRUSTACEANS 67 The third species, II \incnodora i/lucia-Iis, inhabits a region well below the photic zone and does not possess photophores and it will be shown that the eyes of this species are very degenerate structurally. It will not be necessary to describe the general plan of the eye of acanthephvrids for it is basically like that of other decapod crustaceans, and such eyes have been described by Parker (1891), Patten (1887), Trojan (1913), to mention only a few of the earlier investigations of the histology of decapod eyes. From the standpoint of vision and visual acuity the most important features to be considered in any com- pound eye are the following: (a) number of ommatidia, (&) develop- ment of the rhabdomes (the receptor units), (r) amounts, distribution, and movements of screening and reflecting pigments. To these should be added the photosensitive material contained in the rhabdomes. So little is known of the nature of this material that it cannot be discussed at the present time. The average number of ommatidia in a longi- tudinal section of an eye may be taken as a measure of the total number of ommatidia and thereby the task of counting all the elements is avoided. In all counts the eyes of large, mature individuals were se- lected, although not the largest obtainable. In eyes of mature A. pur- piirca (Fig. 10) averaging 2.5 mm. in diameter the average number of ommatidia in a section was found to be 145. Eyes of mature S. dcbilis averaging 1.9 mm. in diameter had an average of 81 ommatidia. Eyes of H. glacialis averaging 0.6 mm. in diameter had an average of 22 om- matidia. The number of ommatidia, hence the number of rhabdomes, determines in part the visual acuity of an arthropod ; therefore the variation in number in these three forms must be significant. The rhabdomes, whence the fibers of the optic nerve arise, are in a sense the most important structures found in the compound eye. A single rhabdome is formed of parts of the seven functional retinular cells found in each ommatidium, and a longitudinal section presents a peculiar striated appearance. In none of the acanthephyrids which have been studied do the rhabdomes have the well-defined outlines seen in the majority of decapods, including certain deep-water forms. Of the three species being especially considered they are perhaps most definite in S. dcbilis, and may be seen in sections of the eyes of light- adapted specimens (Fig. 15) where they are outlined by the proximal pigment. In A. pur pur ca they are difficult to distinguish (Fig. 9). In all forms taken from the nets during either day or night they were never surrounded by pigment. This means that light entering a given ommatidium could reach the rhabdomes of neighboring ommatidia (Exner, 1891). 68 J. H. WELSH AND F. A. CHACE, JR. > •&'•>• -.¥?' ,^ ' '" ^ 10 I;IG. 8. Dorsal view of the head of A. pnrpurca, X 2. I-V,. 9. Photomicrograph of a section of the eye of a younii . /. purpurca, X 40. IMC. 10. Photomicrograph of an unstained section of the eye of a mature A. pia-pursa, X 40. Fi<;. 11. Reflecting pigment of the eye of a young A. pnrpurca as seen by reflected light, X 35. FIG. !_'. Ki-llrrtiiifj pigment of a mature A. pnrpurca, X 35. EYES OF DEEP SEA CRUSTACEANS 69 In //. glacialis the rhabclomes arc entirely lacking. The nuclei of the retinular cells remain, and a few nerve fibers leave their vicinity (Fig. 13). At best the eye of this form can be used only as a light receptor. While not as degenerate as the eyes of certain blind crayfish (Parker, 1890) there is certainly not the necessary visual apparatus for image reception. EYE PIGMENTS Unpublished results obtained by one of us from keeping certain decapods in constant darkness for a period of months indicate that there may be a gradual reduction in the amount of screening pigment in the eye. This is perhaps to be expected from the work of Odiorne (1933) on the effects of black and white backgrounds on the body pigments of certain fishes. Is there any correlation in the acanthe- phyrids under consideration between depth and the amounts of screen- ing and reflecting pigment ? In practically all decapod crustaceans from shallow water there are two sets of screening pigments. The so-called distal or irispigment is found around the cones, and in many forms such as Palcumonetes (Welsh, 1930) it migrates in and out under the influence of hormonal action (Kleinholz, 1936), forming the pseudo- pupil which helps to regulate the amount of light reaching the rhab- domes. In certain shallow-water prawns which are distinctly nocturnal in their habits such as Penceopsis goodei (Welsh, 1935) and species of Brachy car pits and Rlivnchocinetes, the eyes of which have not been described, the distal pigment forms a collar around the outer ends of the cones and remains in that position in light and dark-adapted eyes. In 5". dcbilis the distal pigment is present in an amount comparable to that found in nocturnal surface forms while it is slightly less dense in A. purpurea and completely lacking in H. glacialis. In specimens of S. dcbilis, light-adapted for three hours, there was no measurable mi- gration of the distal pigment. In specimens of both A. purpurea and S. dcbilis, fixed directly from the nets either during the day or night, the distal pigment was always in the extreme outer position. The second set of screening pigment, the proximal pigment, is found in the retinular cells. In dark-adapted eyes of decapods in general, the main mass is found below the basement membrane, while in the light it moves peripherally to surround the rhabdomes and migrates between the plates. In this position it absorbs a large part of the light which reaches the rhabdomes before it has penetrated to any considerable depth. In all specimens of A. purpurea and S. debilis taken from the nets during the day or night the proximal pigment never surrounded the rhabdomes and the major portion was in the position characteristic 70 J. H. WELSH AND F. A. CHACE, JR. of extreme dark adaptation. Tn adults of these two species this pig- ment is present in about equal amounts, but is much less dense than in nocturnal forms found in surface waters. In young of A. purpurca, however, the proximal pigment is very heavy in the lateral portion of the eye (Fig. 9). In //. glacialis the proximal pigment is entirely lack- ing as is the distal pigment. If we assume that the diurnal vertical migrations of A. purpnrca and S. dcbilis keep these forms in a region of very low but constant light intensity, it would be interesting to know whether the proximal pigment still retains the ability to migrate around the rhabdomes when these forms are subjected to illuminations of high intensity. It was possible to test this point with S. dcbilis and A. purpurca. Figure 16 is a photomicrograph of a section of the retina of a specimen of .S. debills which had been kept in darkness for three days. Figure 15 is of a similar region of the retina of a specimen kept in diffuse daylight for one hour. During this time the pigment had migrated for some dis- tance peripherally and had surrounded the proximal two-thirds of the rhabdomes. This indicates that the ability of the proximal pigment to migrate still persists even though it may never do so normally in the lifetime of the organism. The proximal pigment of A. purpurca is also capable of movement although after prolonged light-adaptation it seldom migrates far into the retinular cells. The reflecting pigment of decapod eyes is a layer of amorphous guanin which is concentrated around the bases of the retinular cells and acts as a mirror to reflect light back into the rhabdomes (Welsh, 1932). A reflecting or tapetal layer is found in most animals which are active during the night or which live in a region of low light in- tensity ; therefore it is not surprising that this pigment layer is well- developed in the eyes of deep-water decapods. EXPLANATION OF FIGS. 13-18 FIG. 13. Section of the eye of H. glacialis photographed with transmitted light, X 60. FIG. 14. Section of the eye of H. glacialis photographed by means of re- flected light, X 60. FIG. 15. Region of the retinular cells and rhabdomes of a light-adapted eye of S. dcbilis, X 175. FIG. 16. Region of the retinular cells and rhabdomes of a dark-adapted eye of 5". dcbHis, X 175. FIG. 17. Section of a portion of the eye and the papilla of O. griinaldii photo- graphed with transmitted light, X 50. FIG. 18. The same photographed with reflected light, X 50. EYES OE DEEP SEA CRUSTACEANS 71 13 14 • ^"-r J : \' ^' S •*--n.'-*^ ^15 t- It r 16 FIGURES 13-18 72 J. H. WELSH AND F. A. CHACE. JR. The eyes of .-1. pnrpurca and .V. dcbilis have this reflecting layer in addition to the usual screening pigments (Figs. 11 and 12) while II. glacial is has only reflecting pigment (Fig. 14). The quantity of re- flecting pigment in //. ghcialis is greater in proportion to the size of the eye than in the other t\vo forms under consideration. It is worth noting that in a series of five species of Sergestcs which have been studied in a preliminary way there is a rather striking correlation be- tween the amount of reflecting pigment and the depth at which the sev- eral forms were taken. DISCUSSION The gradual modification of an organ such as the eye as a result of environmental changes, or changes in habit, is one of the fascinating aspects of evolution. A comparison of the eyes of forms which are active at night with those which are active during the day reveals struc- tural and functional adaptations which are among the best examples of the manner in which living material may be modified by external con- ditions. The degeneration of the eye of cave crayfishes (Packard, 1888; Parker, 1890) is a striking illustration of the disappearance of a useless organ. In the sea we have a gradual reduction in the amount of sunlight which penetrates into the water until a depth is readied, which varies with the locality, at which there is a complete absence of sunlight. The animals of the deep sea have without much question evolved from shallow water forms. Their eyes have become modified depending on the depth to which they ha\e migrated. In regions of low light intensity they are, in general, so changed by an increase in size, loss of screening pigment, and in other ways that they are doubtless quite effective organs of sight. Below the level to which light penetrates some are degenerate and some are completely lacking. On the other hand, some are large and, structurally at least, well adapted for vision or for light recep- tion. This fact still remains as one of the most baffling problems asso- ciated with the biology of deep-sea animals. Can there be enough light produced by luminescence to account for the well-developed eyes of some abyssal forms, particularly those living on bottom? This is the question with which most discussions of the eyes of deep-sea animals have ended. It is our hope that if the opportunity remains to continue these studies this question may be satisfactorily answered. SUMMARY 1. Three species of acanthcphyrids have been taken in closing nets in the region of the Sargasso Sea and in slope water near the Gulf EYES OF DEEP SEA CRUSTACEANS 73 Stream in numbers sufficient so that their vertical distribution is quite accurately known. Acanthephyra pur pure a and Syttcllaspis dcbilis are found mostly within the photic zone. Hymenodora glacialis inhabits a region below that to which sunlight penetrates. 2. The eyes of A. pur pur ea and 5". debilis are quite similar structur- ally to the eyes of shallow-water prawns, except that there is less screen- ing pigment. 3. Species of Systellaspis and Oplophorus possess photophores and the eyes of these forms are larger in proportion to body size than the eyes of those acanthephyrids which lack photophores. 4. The eyes of H. glacialis are quite degenerate. The rhabdomes, and both distal and proximal pigments are lacking. The reflecting pig- ment layer is well developed. 5. Characteristic movements of the proximal pigment of S. debilis and A. purpurea occur as the result of light adaptation. LITERATURE CITED ALCOCK, A., 1902. A Naturalist in Indian Seas. John Murray, London. BEDDARD, F. E., 1884. Report on the Isopoda Collected by H. M. S. Challenger. Part I. The Genus Serolis. Zool. Chall. Exp., 11: Part 33. BEDDARD, F. E., 1890. On the minute structure of the eye in some shallow-water and deep-sea species of the isopod genus Arcturus. Proc. Zool. Soc. London, 26: 365. CHACE, F. A., JR., 1936. Revision of the bathypelagic prawns of the family Acanthephyridse, with notes on a new family, Gomphonotidse. Jour. Wash. Acad. Sci., 26: 24. CHUN, C., 1896. Biologische Studien iiber pelagische Organismen. VI. Leucht- organe und Facettenaugen. Bibliotheca Zoologica, 7: 191. CLARKE, G. L., 1933. Observations on the penetration of daylight into mid- Atlantic and coastal waters. Biol. Bull., 65: 317. DOFLEIN, F., 1904. Brachyura. Wissenschaftliche ergcbnissc der Deutschcn Tiejsce-Expedition., 6: 1. DOHRN, R., 1908. Ueber die Augen einiger Tiefseemacruren. Inaugural Diss., Marburg. EXNER, S., 1891. Die Physiologic der facettirten Augen von Krebsen und Insecten. Leipzig und Wien. HANSTROM, B., 1932. Neue Unterstichungen iiber Sinnesorgane und Nerven- system der Crustaceen. II. Zool. Jahrb., Anat., 56: 387. HENDERSON, J. R., 1888. Report on the Anomura Collected by H. M. S. Chal- lenger. Zool. Chall. Rep., 27: Part 49. KLEINHOLZ, L. H., 1936. Crustacean eye-stalk hormone and retinal pigment mi- gration. Biol. Bull., 70: 159. LEAVITT, B. B., 1935. A quantitative study of the vertical distribution of the larger zooplankton in deep water. Biol. Bull., 68: 115. MURRAY, J., AND J. HJORT, 1912. The Depths of the Ocean. Macmillan and Co., Ltd., London. ODIORNE, J. M., 1933. Degeneration of melanophores in Fundulus. Proc. Nat. Acad. Sci., 19: 329. 74 J. H. WELSH AND F. A. CHACE, JR. OSTER, R. H., AND G. L. CLARKE, 1935. The penetration of red, green and violet components of daylight into Atlantic waters. Jour. Opt. Soc. Am., 25: 84. PACKARD, A. S., 1888. The cave fauna of North America, with remarks on the anatomy of the brain and origin of the blind species. Mem. Aro/. Acad. Set., 4:' 3. PARKER, G. H., 1890. The eyes in blind crayfishes. Bull. Mas. Comp. Zool., 20: 153. PARKER, G. H., 1891. The compound eyes in crustaceans. Bull. Mus. Comp. Zool., 21: 45. PATTEN, W., 1887. The eyes of molluscs and arthropods. Jour. Morph., 1: 67. SMITH, S. I., 1886. The abyssal decapod Crustacea of the "Albatross" dredgings in the North Atlantic, 17: 187. TROJAN, E., 1913. Das Auge von Palaemon squilla. Dcnkschr. kais. Akad. Wiss., Wien, 88: 291. WELSH, J. H., 1930. The mechanics of migration of the distal pigment cells in the eyes of Palaemonetes. Jour. E.vpcr. Zool., 56: 459. WELSH, J. H., 1932. The nature and movement of the reflecting pigment in the eyes of crustaceans. Jour. E.rpcr. Zool., 62: 173. WELSH, J. H., 1935. Further evidence of a diurnal rhythm in the movement of the pigment cells in the eyes of crustaceans. Biol. Bull., 68: 247. EVIDENCE FOR THE PRODUCTION OF ACCELERATOR AND DEPRESSOR SUBSTANCES BY ULTRAVIOLET RADIATION OF LIMULUS MUSCLE S. A. GUTTMAN (From the Department of Physiology and Biochemistry, Cornell University Medical College, and the Marine Biological Laboratory, Woods Hole, Massachusetts) The object of this investigation was to determine the causes of the frequency changes after ultraviolet radiation of Liimdus muscle. Hin- richs and Genther (1930) reported that as a result of exposure to ultraviolet point radiation " in general, an increase in rate of beat was obtained by short exposures of a given region [heart of Limidus poly- phenms], and a decrease in rate following longer exposures." Guttman (1936a) reported that short periods of radiation (3 minutes) from a Cooper-Hewitt Uviarc caused a marked primary increase in frequency and amplitude of the Limidus heart. The heart of Liinuhts polyphcmus was excised along with the dorsal ganglion. The rhythmically beating heart was then placed in a finger bowl containing 200 cc. of sea water. Temperature changes were ob- served by means of a thermo junction. One blackened junction was placed on the ganglion at the level of the sixth segment. As the heart- beat of the Limulus is neurogenic (Carlson, 1904 and 1905 ; and Garrey, 1932), its frequency and amplitude depend on the action of the ganglion. Thus, if the temperature of the ganglion is recorded, it is easy to determine whether frequency changes are attributable to temperature variations of the pace-maker. During the course of this investigation temperature increases caused by the heat of the mercury arc were not responsible for the observed changes. The increases in temperature never exceeded 0.3° C. for direct radiation of the heart and markedly less for the shielded preparations as a result of a 3-minute period of radiation. This temperature change wras found to have a negligible effect. A series of 13 experiments were performed on hearts which were partially shielded. In this series of experiments the excised heart, as stated above, was placed in a finger bowl and a shield (layers of white paper, tin foil, and black paper — the white paper outermost) was placed over segments 3-9; i.e., only segments 1-2 received the radia- 75 76 S. A. GUTTMAN tion. A frequency increase was noted. After a length of time the entire heart was radiated, and finally segments 3-9 were radiated while 1-2 were shielded — see Fig. 1. Frequency increases were always great- est for total radiation, next for segments 3-9, and smallest for seg- ments 1-2. Another series consisted of 5 experiments, in each of which 2 excised hearts were placed in a linger howl with the usual 200 cc. of sea water. One of the hearts was shielded and 2-3-minute periods 60 r- 0 10 20 30 40 50 60 70 80 90 100110120130140150160170 TIME IN MINUTES Fir;. 1. Curve showing frequency changes of Linuiliis heart after radiation of various regions. • frequency ; • period of radiation; 1-2 radiation of seg- ments 1-2; 3-9 radiation of segments 3-9; 1-9 radiation of segments 1-9 (entire heart). • if radiation were givrn In the unshielded heart. The usual frequency changes (primary acceleration) were observed for the irradiated heart. The shielded heart also markedly increased its frequency. The in- crease in frequency (expressed in per cent) was usually greater in the unirradiated heart than in tin- radiated preparation — see Fig. 2. An explanation is suggested below. XYhen hearts which did not heat because of injury to the ganglion during dissection were irradiated for 2-3 minutes, the rhythmically heating hearts (shielded) always exhibited a marked frequency increase. ULTRAVIOLET RADIATION OF LIMULUS MUSCLE 77 In several series, following the same pmordure as above, masses of Lhmtlus skeletal muscle (7 experiments) and smooth muscle (intestine) (6 experiments) were irradiated and the heart was shielded. Here, too, the heart frequency increased after short periods of radiation of the various types of muscle — see Fig. 3. 45 40 35 H 30 W 0 25 u a, 55 H-H 20 15 u 10 u 5 a u 0 UH — 5 — 10 0 5 10 15 20 25 30 35 40 45 TIME IN MINUTES FIG. 2. Curves showing frequency of Limulus hearts (1) directly radiated, (2) bathed by sea water which surrounded the radiated heart. O frequency of heart in solution which bathed a radiated heart ; • frequency of radiated heart ; • period of radiation. Several other series of experiments were performed. It was found that placing a beating heart in irradiated sea water did not induce any frequency change. Radiation of sea water which previously contained unirradiated cardiac, smooth, and striped muscle had no effect on the frequency of Limulus hearts. Thus it is concluded that the radiation must be directly on a tissue in order to produce the primary acceleration and that irradiation of sea water is ineffective. 78 S. A. GUTTMAN E- 2 u u CL, fc c U D 160 140 120 100 80 60 40 20 0 U g-20 —40 I I I 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 TIME IN MINUTES FIG. 3. Curve showing frequency and I'^iiM has presente-d e-vide-nce- which indicates the probability of a potassium-calcium shift induced by ultra- violet radiation. This shift, however, probably is not responsible for the frequency changes observed. Guttman (1936fi) state-s, " It may be- suppoM-d. in the case of the Limulus heart, that the- primary action of the radiation | ultraviolet | is to increase the frequency of the- nerve cell discharge--, and also to increase the number of active- nerve cells; hence the number of nerve libers transmitting impulses to the myocardium is increased. The former supposition would account for the increased frequency, the latter for the increased amplitude." ULTRAVIOLET RADIATION OF LIMULUS MUSCLE 79 It appears that an increase in the frequency and number of nerve cell discharges may be caused by the liberation of some accelerator substance from Liinitlus muscle upon irradiation. It is also well known that nerve is very resistant to radiation and the radiant energy em- ployed could not possibly affect nerve directly. The effect of the substance usually disappears within 20 minutes after the cessation of radiation and then the secondary effect appears (the suppressed fre- quency and amplitude). It may be that the secondary effect is caused by a second substance produced upon irradiation but the evidence for this is not clear cut and further investigation is necessary. However, this secondary effect does not appear in some hearts which have been irradiated for 3 minutes and an even shorter time of irradiation seems to minimize the secondary effect. In the shielded hearts, stimulated indirectly, the secondary effect appears, apparently due to the diffusi- bility of the supposed secondary substance. Thus there is the possi- bility of the production of two chemical substances which are antag- onistic in action. It is a pleasure to acknowledge my indebtedness to Professor W. E. Garrey of Vanderbilt University Medical College, Nashville, Tennessee, for his interest and constructive criticism. LITERATURE CITED CARLSON, A. J., 1904. The nervous origin of the heart-beat in Limulus, and the nervous nature of co-ordination or conduction in the heart. Am. Jour. Physiol, 12: 67. CARLSON, A. J., 1905. Further evidence of the nervous origin of the heart-beat in Limulus. Am. Jour. Physiol., 12: 471. GARREY, W. E., 1932. Some aspects of the physiology of the heart of Limulus polyphemus. Collecting Net, 7: 137. GUTTMAN, S. A., 1935a. Influence of ultraviolet irradiation on clam heart sub- jected to potassium excess. Proc. Soc. Exper. Biol. and Med., 33: 363. GUTTMAN, S. A., 193S&. Effect of ultraviolet on heart of Rana pipiens and Alli- gator mississippiensis. Proc. Soc. Exper. Biol. and Mod., 33: 408. GUTTMAN, S. A., 1936Y/. Effect of ultraviolet radiation on the heart of Limulus polyphemus. Biol. Bull., 70: 279. GUTTMAN, S. A., 1936&. The influence of ultraviolet irradiation on frog and Limulus hearts subjected to potassium excess. Jour. Cell, and Com p. Physiol., 8: 37. HINRICHS, M. A., AND I. T. GENTHER, 1930. Rate of heart-beat in Limulus as affected by exposure to ultraviolet point radiation. Proc. Soc. Exper. Biol. and Med., 28: 121. NOTES ON THE DEVELOPMENT OF GORGODERA AMPLICAVA IN THE FINAL HOST THERON O. ODLAUG (From the Department «/ ttiology. University College, Nczv York University and the Marine Biological Laboratory, Woods Hole, Mass. ) INTRODUCTION Loschge (1785) first reported flukes in the bladder of Rana cscit- Icnta and Zeder (1800) named them Distomum cygnoidcs. On the basis of the number of testes, Looss (1902) separated the bladder rlukes of frogs into two genera. Cor god era and Gorgodcrina, the former genus possessing nine testes and the latter, t\vo. Bladder rlukes of frogs have been reported in Xorth America by Leidy (1851), Bensley (1897), Stafford (1902). fort (1912). Ingles and Langston (1933), and Ingles (1936). Seven species of gorgoderid \vorms have been described in North America: Gorgodera amplicava from Rana clainitans, R. cates- beiana, and A*. f>ipicns; Gorgodera minima from R. catesbeiana and R. pipicns ; Gorgodcrina simplc.r from A', catesb^ iana and Bufo lentigino- sus; Gorgodcrina translucida from Hufo lentiginosus and R. virescens; Gorgodcrina attcnuala from A', catesbeiana and R. virescens; Gorgo- derina inultilobata from A', hoyli and A', aurora; Gorgodcrina aurora from R. aurora. Krull (1934). \vho reported the principal stages in the life history of Gorgodera ainplicara. found that the clam, Musculium partumcium served as the1 first intermediate host, the snail Hclisoma antrosa as the second intermediate host, and the frogs R. clainitans and R. catesbeiana as the final hosts. The first life history studies on bladder tlukes were done by Ssinitzin (1905) who traced the development of Gorgodera cygnoidcs, Gorgodera pagenstechcri, and Gorgodera rarsoi'iensis. lie found that the cercarise were of the cystocercous type, that they were produced in bivalve mollusks, and that the metacercaria,- developed in the aquatic larvae of insects. Lutz (1926) indicated the intermediate hosts through which species of Gorgodcrina might pass in order to complete the life cycle. ll<- toiind the cercaria- in two small bivalves, Cyclas and Pisidimn or Sphccrium, and the metacercariae encysted in the esophagus of odonatan larvae. The .same author also stated that the parasites in the final host, which is some anuran, were often found in the ureters. He observed 80 DEVELOPMENT OF GORGODERA AMPLICAVA IN FINAL HOST 81 that "En effet, les conduits off emits minutes. The tops were left in the tubes 5 minutes longer and then transferred to sea water. At this time about half of the tops were about twice as long as wide; a few others were longer with a constriction near the middle, and the rest were oval or pear-shaped. Two of the living tops are drawn in Figs. 9, 10. The larger tops were dividing 13 minutes later (Figs. 11, 12, 13). At this time the second lobe had gone back in the control. The controls did not divide until 32 minutes later. Thus the division of the tops by the second spindle was a little delayed compared with the time of extrusion of the second polar body of the normal control eggs, but took place half an hour before cleavage of the latter. Sixteen minutes later all of the 56 tops had divided (Figs. 14, 15). An hour and a half later one of the two halves was dividing or divided (Fig. 16, 17, 18). The other half was undivided. Some of the eggs had been preserved soon after removal from the machine. A plate of chromosomes (the mitotic figure for the second maturation division) lay in the middle of the stained zone (Fig. 51). Other tops were preserved after the maturation division (Fig. 52), and others after one of the two halves had divided (first cleavage, Fig. 53). The other half contained a nucleus, but was undivided. The former half presumably contained the sperm nucleus. (6) This set of eggs was centrifuged at the time when the second lobe was about one-third out at 1850 r.p.m. for 5 minutes, then at 2970 r.p.m. for 12 minutes. The tops were off, and were left in the tube for an hour. \Yhen removed most of them were elongated and many constricted (Figs. 19, 20, 21, 22). Halt" an hour later both halves were more spherical (Fig. 23) and remained in this condition for 50 minutes (Figs. 24, 25, 26), when one of the halves began to divide, almost always the polar half (Figs. 27, 28, 29). Eggs killed 6 minutes later showed a variety of conditions; .some were still divided into two equal parts (Figs. 54, 55, 56, 57) ; two of these had a nucleus in each half; one had a division figure (anaphase stage, l-'ig. 5'») in one half and the chromosomes irregularly dispersed at the division plane; one (Fig. 57) had a bridge of chromosomes across the division plane. Two tops are drawn in Figs. 58, 59, in which one half has cleaved into two n 11s, the other half has a small group of deeply stained chromatin. Two others (Fig. 60) have a line of chromosomes extending across the division plane. One, Fig. 61, has a small cell with a large nucleus in the bridge between the larger halves, and a dark mass of chromatin in one half which may be either the sperm nucleus or a part of the mitotic figure. MATURATION SPINDLES CENTRIFUGED EGGS 93 The outcome of these experiments shows clearly that the mitotic figure of the second polar spindle may divide a fragment (itself about half the size of the egg or a little larger) into equal or nearly equal halves, provided the top is elongated at the time of division. This means probably that the normal division, at the second maturation, into a minute polar body and an egg is not due to any pecularity of polarity of the mitotic figure, but to its location on the surface of the egg. On the other hand, if the second spindle is driven into the whole egg, and the egg becomes spherical, it fails to divide the egg, unless, as previous work has shown (Clement, Morgan), the egg is elongated at this time. THE FIRST POLAR SPINDLE IN " TOPS " OBTAINED BY CENTRIFUGING In order to find out whether the first polar spindle could also divide the tops into equal parts, a number of sets of eggs were centrifuged at different stages in the development of the spindle. The results were negative. It was found that the spindle remained at the surface of the egg, or else moved there, and gave off a polar body. (7) A capsule, just laid, was opened, and the eggs centrifuged at once. A few eggs were preserved before centrifuging. These showed the chromosomes in metaphase in the polar hemisphere. After 24 minutes on the machine tops separated from bottoms. They were much elongated at first. Six minutes later they were still elongated (Fig. 30), but not so much so as at first. One of the bottoms is drawn in Fig. 31. The tops became spherical and polar bodies were given off in or near the oil caps. None divided into equal parts at the time of formation of the first or of the second polar body. Later they cleaved into 2 cells when the first (delayed) cleavage was due. (8) Another set of eggs from a capsule just laid was centrifuged, in all 47 minutes ; 5 minutes at 1850 r.p.m., 10 minutes at 2970 r.p.m., 6 minutes at 3450 r.p.m., and 26 minutes at 2970 r.p.m. When re- moved the tops were greatly elongated (Fig. 32). The first polar body had not been given off. Twenty-four minutes later a polar body appeared in the oil in some tops, and during the next 20 minutes appeared in more tops. The first lobe of the control appeared 16 minutes after beginning centrifuging; and the second when the tops were still on the machine. Sixteen minutes later, when the tops were more contracted, none had divided and several at least had two polar bodies. Most tops were constricted at this time, but no division oc- curred. There were, however, several fragments (in which no polar body was seen), which were deeply constricted, but as these did not contain the oil field they might be called middles. Later, after 26 min- J PLATE I FIGS. 1-33. Tups of CPUS (except Figs. 2 and 31) separated over pum arabic. These arc all drawn to the same scale from living material. For details see text. MATURATION SPINDLES CENTRIFUGED EGGS 95 utes, one third of these divided into e<|ual parts. The control did not divide at this time so that these may represent either the first or more probably the second polar mitosis. In two of these only one of the equal cells divided again, as is the case of tops divided by the second polar mitosis. Some of the tops of this set, killed at the time of removal from the machine (Figs. 63, 64), showed the polar chromo- somes in the oil cap, possibly in anaphase. The sperm nucleus was deeper in the protoplasm. If, in some of these tops, the oil with the polar chromatin had been driven off, the middles with the sperm nucleus might be expected to divide at the cleavage stage, but I have no record of such cases. (9) Another set of eggs, centrifuged 42 minutes before the first polar body was due, was not divided by the polar spindles. (10) Another set was centrifuged one hour and 15 minutes before the first lobes appeared. Three tops divided into two parts at the time of the second spindle, but 86 did not. (11) Another set centrifuged for 19 minutes, one hour before the first lobe, at 1850 r.p.m. for 5 minutes, then at 2970 r.p.m. for 12 minutes, and at 2970 r.p.m. for 2 minutes. When removed the tops were much elongated (Fig. 33), and no polar bodies had been given off. Seventeen minutes later polar bodies appeared in the oil, or at the side. Eighteen minutes later most of the tops were spherical ; some were still elongated. None had divided 35 minutes later. Tops killed when removed from the machine showed a metaphase plate in the polar part of the stained area (Fig. 62). Fifty minutes later two polar bodies were at the side of some tops, but none showed chromatin in the middle of the stained zone, i.e., the second spindle had come to the surface. The failure of the first spindle to divide the tops, as does the second spindle when driven from the surface, may be due to failure to have centrifuged at the right stage, but this seems improbable since a fairly wide range of times was tried. The failure may be due in part to the first spindle being driven into the oil region or near the surface, rather than into the protoplasmic zone. In other cases when the spindle was present in the top it moves to the surface of the oil or to the side of the top, and gives off one or two polar bodies there. It seems then that, compared with the behavior of the second spindle, the condition of the egg is such that the first spindle comes to the surface again while the second does not always do so. In what respect the conditions are different is impossible to state. The difference has nothing to do with the attachment of one pole to the surface in one case and not in the other, since in Ilyanassa it is the whole polar spindle that is carried 06 T. H. MORGAN PLATE II FIGS. 34-64. Tops of eggs drawn from preserved material. These are not on the same scale as Figs. 1-33. They were smaller, due to shrinkage in preserv- ing and mounting in balsam, and were enlarged and reduced on a different scale from those in Plate I. MATURATION SPINDLES CENTRIFUGED EGGS 97 into the interior, or else it is not moved at all. Neither a short, nor a long time in the centrifuge seemed to affect the results. It seems im- probable that the two spindles are in'themselves different with respect to their power of dividing a fragment. The failure of the first spindle to divide the top in these experiments seems rather to be due to a failure to drive it or hold it in the middle of the fragment. Under other conditions it may be found possible to do this. CONCLUSIONS When these results are compared with those obtained by Morgan (1935) and Clement (1935) where the eggs of Ilyanassa were elongated in the centrifuge but the tops have not pinched off, it is evident that in both it is the second spindle that caused the division of the egg or of the top, if at the time the whole egg or the top has retained to some extent its elongated shape. If the top contracts into the egg, the egg is not divided by the polar spindle, even although the second polar body is not extruded. It seems that the stretching of the egg is correlated with the division and is a contributing factor. I have discussed else- where whether there is any advantage in speaking of one of the parts after the division of the second spindle as a giant polar body and the other as the egg. It is even less apparent in the case of these tops that anything is to be gained by such a comparison. The main point is, I think, that the results show that the second spindle, removed from the surface by centrifuging, may divide the egg or fragment of the egg into two equal parts, provided the egg or the top is still elongated at the time of division. In the case of Crcpidula, Conklin ascribed the formation of the giant polar bodies to the forcible removal of one daughter group of second polar body chromosomes from the other group. In Ilyanassa, on the contrary, the division is brought about by the removal of the whole mitotic figure from the surface, and in this re- spect it may appear that the division that follows may be more nearly compared with the maturation division that gives rise to a polar body. It would seem that the eggs of Crepidula and of Ilyanassa behave dif- ferently on the centrifuge and consequently their subsequent behavior is different. BIBLIOGRAPHY CLEMENT, ANTHONY C, 1935. The formation of giant polar bodies in centrifugcd eggs of Ilyanassa. Biol. Bull, 69: 403. CONKLIN, E. G., 1917. Effects of centrifugal force on the structure and develop- ment of the eggs of Crepidula. Jour. E.vpcr. Zool., 22: 311. MORGAN, T. H., 1933. The formation of the antipolar lobe in Ilyanassa. Jour. Ex per. Zool., 64: 433. 98 T. H. MORGAN MORGAN, T. H., 1935a. Centrifuging tlie eggs of Ilyanassa in reverse. Biol. Bull., 68: 268. MORGAN, T. H., 1935b. The separation of the egg of Ilyanassa into two parts by centrifuging. Biol. Bull., 68: 280. MORGAN, T. H., 1935r. The rhythmic changes in form of the isolated antipolar lobe of Ilyanassa. Biol. Bull., 68: 296. MORGAN, T. H., AND ALBERT TYLER, 1930. The point of entrance of the spermato- zoon in relation to the orientation of the embryo in eggs with spiral cleav- age. Biol. Bull., 58: 59. MORGAN, T. H., AND ALBERT TYLER, 1935. Effects of centrifuging eggs of Urechis before and after fertilization. Jour. Exper. Zool., 70: 301. WILSON, EDMUND B., 1929. The development of egg-fragments in annelids. Arch, f. Entiv.-mccli., 117: 179. WILSON, EDMUND B., 1930. Notes on the development of fragments of the ferti- lized Chaetopterus egg. Biol. Bull., 59: 71. BUDDING AND LOCOMC TION IN THE SCYPHISTOMAS OF URELIA FRANCIS G. GILCHRIST (From the Hopkins Marine Station of Stanford University and the Scripps Institution of Oceanography) The Scyphozoa have three principal methods of reproduction : ( 1 ) sexual reproduction by means of gametes formed by the free-swimming medusae ; (2) lateral budding by the scyphistomas or polyps which de- velop from fertilized eggs; and (3) stabilization — a sort of transverse fission by which a scyphistoma gives rise to a series of young medusae. The studies here reported concern the second of these methods of re- production ; namely, lateral budding. They concern also the closely related phenomenon of locomotion by means of " sterile buds " or pedal stolons. Lateral budding in a scyphistoma was described one hundred years ago by Dalyell (1834), who also saw strobilization (1836) and noted that after giving off the young medusae (ephyrse), the basal portion of the strobilizing polyp (strobila) returns to the vegetative or scyphistoma stage. Thus, according to Dalyell, the scyphistoma or "' Hydra tuba," as he termed it, is a perpetual condition, able to maintain itself indefi- nitely by budding. Herouard (1908) has described budding and loco- motion (" metrotropism ") ; while a fuller account is given by Perez (1922) ; and more recently by Halisch (1933). The present account covers somewhat the same ground, but undertakes by observation and experiment, to discover the causal relations involved. In short, we shall attempt to be explanatory. MATERIAL The polyps used were presumably those of Aurclia, and were ob- tained in great abundance from the underside of an old float in a slough not far from Pacific Grove, California. They were obtained also from the hulls of destroyers which had lain at anchor for four years in San Diego Bay. I wish to acknowledge my indebtedness to the two Cali- fornia institutions whose guest I was while making these studies: the Hopkins Marine Station of Stanford University, located at Pacific Grove, and the Scripps Institution of Oceanography of the University of California, located at La Jolla. 99 100 FRANCIS GILCHRIST A scyphozoan polyp (or scyphistoma) consists of two germ layers, ectoderm and entoderm, separated from one another by a thick layer of gelatinous mesoglcea. Its body may be divided for purposes of de- scription into two parts: (a) a body portion proper, terminating distally in an oral disc (peristome) bordered by a circle of tentacles and sur- rounding the mouth. \Ye may call this terminal portion of the body, for want of a better name, the " hydranth." The lips of the mouth project somewhat, as a proboscis. The body is partially constricted in- ternally by four longitudinal folds of entoderm, the gastric ridges (taenioles). (b) Below the body is the stalk, a simple cylinder which ends in a base of attachment. Commonly there is a small amount of 3 mm. FIG. 1. Scyphozoan polyps (Aurclia] slum ing buds and pedal stolons, d and / arc typical " fig-type " buds, c and uds and stolons develop from the wall of the lower body and stalk. They may be classified as of four principal types; although variation within limits, rather than conformity to rule-, is the rule in scyphistomas. (1 ) The fig-lypc bud (Fig. 1, / and d) is an outgrowth of usually the lower body or upper stalk. Typically it is compressed from side to side, and it soon begins to constrict away from the side of the parent. BUDDING AND LOCOMOTION IN SCYPHISTOMAS 101 A mouth and circle of tentacles develop on the side of the bud which is towards the parent's tentacles, while a tendril-like stolon grows out on the side of the bud away from its attachment to the parent. Sometimes the hydranth forms before the stolon appears ; sometimes the reverse is the case ; but most frequently the tentacles and stolon grow out at about the same time. In the course of a few days the stolon elongates, at- taches by its tip, and contracts ; thus it draws the bud away from the parent (Fig. 2). A narrow strand of tissue composed of ectoderm only, remains for a time to connect the bud to its place of origin on the side of the parent. Budding of this sort has been described by Perez and Halisch. (2) A second type of outgrowth is the pedal stolon (Fig. 1, c and g). This is at first a blunt cone, which becomes more and more acute A C FIG. 2. Locomotion of buds. A is an adult polyp with older and younger stolons and from which a bud, B, has recently migrated. C is another bud mi- grating. until it is an elongated tendril of nearly uniform cross-section. Pedal stolons grow almost entirely from the upper stalk portion of the polyp. As a rule they are at first directed outward and upward, that is, away from the polyp's base. Shorter stolons which attach rather promptly may grow downward from the lower part of the stalk, especially in polyps which have been torn loose from their place of attachment. (3) An intermediate type of outgrowth is the stolonic-bud (not figured). This is similar to the pedal stolon in its manner of origin, and usually forms in about the same position. Its distal portion is in fact a pedal stolon ; but as it grows outward, its proximal portion draws to itself a more than usual amount of the body wall of the parent polyp. This gradually constricts away from the parent, and as it does so, it forms a mouth and tentacles on the side of the bud near its proximal 102 FRANCIS GILCHRIST end. Sometimes two hydranths are formed side by side, instead of one. The stolonic bud differs from the fig-type bud in that the development of the hydranth and the process of constriction away from the parent do not occur until after the stolon has grown out, and even after it has attached. (4) The fourth sort of outgrowth is the hydra-type bud (Fig. 1, i. but especially r). This is a cylindrical outgrowth, usually from tin- lower stalk. Its distal end forms a mouth and circle of tentacles. It does not soon form a pedal stolon, but remains attached to the base of the parent for a long time. Indeed, it may grow to be the size of its parent, with the result that small colonies of several individuals are sometimes formed. Ultimately, however, the buds separate by the formation of pedal stolons on the parent or on the bud, above the place of union. The stolon, after attaching and contracting draws the bud and parent apart. When the stolon is on the parent, the body of the parent may be drawn away, leaving the bud attached to the substrate by the old base of the parent. Such a method of separation is described as of regular occurrence in a scyphistoma of unknown genus (Renton, 1930). The day-by-day history of the fig-type buds, the sequence of tentacle formation, and the descriptive aspects of locomotion by means of pedal stolons have been given by Perez (1922) and independently by Ilalisch (1933). LOCOMOTION The Method of Locomotion Pedal stolons serve two purposes: for locomotion and to provide new bases of attachment. The first process is seen to best advantage in the migration of buds away from the parent (Fig. 2). The stolon grows out on the side of the bud away from the parent's body. It then attaches to the substrate, contracts, and so draws the bud after it. Then another stolon grows out on the side of the bud away from the parent ; and this in turn lengthens, attaches, and contracts. Thus step by step the bud moves from its place of origin. After perhaps six steps, the stolons appear less frequently and in less organized fashion; and al- though change of location still takes place, forward locomotion in one direction is not so consistently observed. I -".veil in adult polyps, however, forward locomotion in one direction is occasionally seen (Fig. 3). A number of polyps were removed from the oyster .shells to which they were attached, into a dish of sea water. Individuals were then chosen which showed neither buds nor stolons, BUDDING AND LOCOMOTION IN SCYPHISTOMAS and these were placed in individual dishes. Twenty-four hours later most of them had developed stolons, and again twenty-four hours later most of them had attached. (Possibly isolation favored the formation of stolons, although this is not proven.) Careful drawings were made of certain individuals from day to day; and it was found that some sent out two or more stolons, either simultaneously or in close succession, which thus anchored the polyps to one spot; while the majority formed stolons one by one, roughly at two-day intervals. There was in some instances a tendency for stolons to form successively on the same side, so that a more or less directed locomotion took place. A Sept > FIG. 3. Locomotion of an adult polyp. The same polyp was drawn on four successive days. The pedal stolons are numbered from 1 to 5 in the sequence of their formation. " Life History " of a Pedal Stolon Locomotion by stolons will become clearer if we consider the " life history " of a single stolon (Figs. 3 and 4). (a) The stolon begins as a blunt, conical projection from the upper stalk portion of a polyp (Fig. 3 A, stolon 4). (6) It then elongates and pushes outward and some- what upward away from the base of the polyp (Fig. 3B, stolon 4). It continues to grow in length until it has become a tendril-like process composed of an ectodermal sheath with a solid entodermal core, free to wave to and fro in the moving water. Indeed, it apparently has some movement of its own. (c) When fully elongated the tip of the stolon (ectoderm) develops glandular cells and becomes adhesive. Coming into contact with some surface it attaches (Fig. 3C, stolon 4). (d) 104 FRANCIS GILCHRIST Contraction follows immediately, and with so much power that the en- tire polyp is pulled toward the new point of attachment (Fig. 3D, stolon 4). (e) The stolon now becomes for a time molded over into the base of the polyp; its attached tip becomes the foot (Fig. 3/J, stolons 4 and 3). (/) In due time another stolon forms, elongates, attaches, and contracts; thus drawing the polyp onward. The old stolon (the base) C ,•- IMG. 4. Living pedal stolons as scon with the high power of the microscope (optical sections). A, a young stolon with specialized ectodermal tip and core of entoderm. 13, the same contracted as a result of mechanical stimulation. C, en- larged detail showing margin of a tip undergoing attachment to the substrate. /). a stolon under tension following attachment. The core of entoderm is migrating back into the stalk of the polyp in the direction of the arrow. E, attached stolon under tension. The adherent tip is anchored to the bottom of the dish by means of fibrils. The core of entoderm is partly withdrawn. /• , an old stretched stolon. The entoderm has become reduced to a thread with a few cells scattered along it. Cnidoblasts are shown in this and other figures. is thus stretched out into a thread composed almost solely of ectoderm, which for a time retains its attachment to the substrate (Fig. 3/1 to />. stolon 2). ((/) Soon, however, the old stolon breaks from its attach- ment, and is withdrawn into the side of the stalk of the- polyp, usually with sonic debris adhering (Fig. 3.-/ and H. stolon 1). The entire his- tory of a pedal stolon thus suggests strongly the behavior of a pseudopod r.UDDING AND LOCOMOTION IN SCYPHISTOMAS 105 of a rhizopocl protozoan; except, of course, that the stolon is composed of many cells. The above description applies \\lim one Molun is formed at a time. When several are formed, each one cannot become a base (Fig. 2A). Some develop tension without being able to shorten. After a day or so in this case also the entodermal core is withdrawn ; and still later the ectoderm pulls away from its attachment and is withdrawn. Role of the Germ Layers in Locomotion Observations under the high power of the microscope make it clear that the two germ layers of the polyp play diverse roles in the processes of budding and locomotion. We shall not discuss their parts in the development of a hydranth at this time, except to note that the entoderm comes very close to the ectoderm in the region where a mouth is form- ing, and in the regions where tentacles are about to grow out. (a) The first visible evidence of a bud or stolon, similarly, is a conical projection of entoderm as seen through the overlying transparent ectoderm. The tip of the entoderm seems to push away the intervening mesogloea and to come into actual contact with the ectoderm. (7;) As the stolon grows outward a rearrangement of the entoderm takes place (Fig. 4A). The entodermal cells of the wall of the stalk migrate actively into the form- ing process and rearrange themselves into a solid core. The ectoderm, meanwhile, undergoes little change in its appearance and in the arrange- ment of its cells, except that it grows thinner, (c) An exception to the above statement is the ectoderm of the tip of the stolon, which thickens and becomes glandular. It is worth noting again that it is in this re- gion that the ectoderm is in most close contact with the underlying entoderm. If an elongated stolon be mechanically stimulated, a contraction of the ectodermal sheath takes place (Fig. 4B). The tip does not con- tract, and the entodermal core appears to be only passively compressed. Its cells become disc-shaped, and give the appearance in lateral view of a pile of coins. Contraction of this sort, however, is only temporary ; and in a short while (an hour or two) the stolon is again fully elongated. (rf) Contact of the tip of the stolon with the substrate stimulates the glandular cells to discharge a cementing substance in the form of fibrils (Fig. 4C). On coming into contact with the solid surface the ends of these fibrils adhere firmly and so serve to anchor the tip of the stolon. Herouard (1911) has described and figured a process of the formation of " tonofibrilles " which differs in some details from the bove statement. As this process is taking place the tip of the stolon 106 FRANCIS GILCHRIST spreads out and flattens down in close apposition to the surface. A thin perisarc, a sort of mucus, is secreted around this area of attach- ment (Fig. 4£). (c) Contraction of the stolon follows. This begins with a flow of the core of entoderm toward the body of the polyp (Fig. 4D). The flow is most rapid in the middle, with the result that the entodermal cells become convex in the direction of the polyp's stalk. As the entoderm is withdrawn, the sprier between the ectodermal sheath and the entodermal core becomes wider and wider (Fig. 4£). If a stolon should be torn from its nn Hiring at this stage of contraction, a contraction of the ectoderm takes place, and the narrow entodermal core is thrown into a spiral coil within the ectoderm. (/) On reach- ing the stalk the entodermal cells rearrange themselves again into the entodermal layer of the body wall. This involves, of course, the de- velopment of a cavity between them. (g) As the entodermal core is withdrawn the ectodermal sheath develops tension. If the polyp is free to move, it is pulled toward the new point of attachment, which thus becomes the new base. If, how- ever, the polyp is not free to move by reason of other attachments, the entoderm becomes drawn out into a thin axial thread with a few ento- dermal cells scattered along it (Fig. 41'). After thus serving for a time as an anchoring line, the ectoderm finally pulls away from its attachment and is withdrawn into the >ide of the stalk. Morplnilla.vis of the Stalk From the descriptions just given, it will be seen that the stalk por- tion of a polyp is of remarkable plasticity. It continually changes its form and structure as one stolon after another is formed and resorbed. The process is not to be classed as growth in the usual sense, since it does not invoke an increase in the si/.e and number of cells. Rather, the cells change their shape and to some extent their specifications. First, for example, the upper stalk cells become stolon cells with the tendency to form a solid process. Then, after the stolon has attached, they become progressively transformed into stalk cells again. IMastic molding of this sort was termed by Morgan " morphallaxis." t'sually when a stolon attaches and contracts, the stalk becomes bent downward toward the new attachment (Fig. 3H). In recovering its upright position, material is withdrawn (especially entoderm) from the old base and stolon ( Fi-. .}(' and /)). Looked at grossly there is thus a flow ot materials of the stalk from the .side on which old stolons are being resnrbed to the side on which new stolons are forming. This movement was demonstrated in an experiment in which spots of vital (Jve BUDDING AND LOCOMOTION IN SCYPHISTOMAS 107 (Nile blue sulphate) were placed upon the side of the upper stalk. (The polyp was laid on a piece of paper and a crystal of the dye was touched to the surface for a few seconds. The- polvp was then returned to sea water and any adhering dye was washed off.) In some instances stolons grew out from the region of the spot, or from just above the spot. It was then observed that the spots elongated into the stolons. Later, A I mm. FIG. 5. Reconstitution from small pieces. A is a small piece taken from the oral disc (peristome). B, C, and D are tentacles with a small amount of adjacent material from around their base. £ is a piece of entoderm which has rounded but not undergone regeneration. F is a comparable piece of ectoderm. It has formed a complete polyp. G and H have regenerated from very small fragments of ectoderm. when the stolons were resorbed, the marked areas again became com- pact spots, but upon the side of the stalk opposite to that upon which they were originally placed. EXPERIMENTS Recoiistitntion Regeneration of Small Fragments. — Animals which possess the capacity for vegetative reproduction are as a rule also able to regen- 108 FRANCIS GILCHRIST erate themselves from pieces. This is true of scyphozoan polyps. In- deed, no part of this animal is without some power of restitution, al- though the power differs from region to region hoth in degree and in kind. Small pieces roughly one millimeter square isolated from different locations behave differently, (o) A piece cut from the oral disc (peristome), and including a part of the oral lip but no tentacles, will roll itself together and will regenerate one mouth, sometimes two mouths ; but usually no tentacles nor any part of the body below the tentacles are formed (Fig. 5A). (&) A similar piece taken from the region of tentacles will close and regenerate lips and oral disc, sometimes addi- tional tentacles, but usually no body (Fig. 5, B and C). If a tentacle has been cut off it will grow out again. Sometimes the stump of a tentacle will regenerate a bifurcated tentacle, (r) An isolated ten- tacle heals its cut end but otherwise undergoes little change. Tt may live for days, .swimming about as though it were a ciliated worm. Oc- casionally a tentacle cut close to its base regenerates a small mouth (Fig. 5D). (d) A piece of the upper body removed from just below the circle of tentacles promptly regenerates an oral disc and tentacles. (e) From the upper body region downward to the attached base there is a gradual decrease in hydranth- forming tendency. The hydranths which are formed at lower levels are smaller, of fewer tentacles, and require a longer time for their regeneration. Considerable variability in hydranth-forming power exists, however, especially at the lower body and upper stalk levels. Some pieces form whole, well-proportioned polyps, and do so quite promptly. Other pieces from the same level form single large stolons, or even two stolons. Xo doubt this variability reflects the bud-forming or stolon-forming tendencies which were present in this region of the polyp at the time the pieces were isolated. (/) Pieces taken from the basal end of the polyp or near it usually round up into a ball and secrete a loose perisarc about themselves, (g) Iso- lated stolons commonly attach by their tips and then, after contracting, regenerate entire small polyps (Fig. 9A-D). The gradient of hy- dranth-forming tendency, which has just been described, is illustrated in the behavior of transverse segments as shown in Fig. 6. Segments /' and c, which are from the upper body, have their hydranths fairly well reformed on the third day. Segments d, c, and /, which are from the lower body and upper stalk have small hydranths on the seventh day. Stdons are most advanced in segments c, d, and c. Segment y has rounded up and secreted perisarc. Later it may break from its and regenerate a very small polyp, observations show that invisible differentiation (segregation, BUDDING AND LOCOMOTION IN SCYPH1STOMAS 109 chemo-differentiation, determination) is present in the polyp. No two parts behave in the same fashion when tlirv are isolated. The invisible differentiation is not complete (stable or irreversible determination) except apparently in the oral lips and tentacles. Elsewhere large powers of regulation remain; or in other words, invisible differentiation is in- complete (labile determination), and ^differentiation is possible. In- visible differentiation reveals itself principally in formative tendencies; that is, in hydranth- forming tendencies which are strongest at the upper FIG. 6. Reconstitution from transverse segments. Note the decreasing hydranth-forming tendency from a to g; the pronounced stolon-forming tendency in c to e ; the base-forming behavior in g. a, b, and c were drawn on the third day ; d, on the fourth day ; c, f, and g on the seventh day. end of the polyp and decrease toward the lower end, and in base-forming tendencies which are strongest at the lower end. The power to form typical pedal stolons is greatest in pieces from the upper stalk area. Polarity in Regeneration Polarity in the scyphistomas is similar to the same phenomenon in Hydra and the hydroids. In general, transverse sections regenerate apico-basally ; that is. with the hydranth formed from the upper cut - Mi-face (Fig. f>, h to f). Sometimes, however, sections cut from just 110 FRANCIS GILCHRIST below the circle of tentacles regenerate biapically ; that is, with a hy- dranth from both the upper and the lower cut surfaces (Fig. 7 A, B, and C). This is most likely to occur if the section be thin. Longitudinal halves and quarters of polyps from which the hydranth has been removed regenerate asymmetric hydranths, which reveal both the inherent polarity of the piece and the orienting effect of wounding. If the piece be long the new hydranth is formed from the upper cut FIG. 7. Polarity in regeneration. A, B, and C. Thin transverse sections from the upper body region frequently regenerate biapically. D and E. A longi- tudinal quarter of a polyp (hydranth removed) regenerates obliquely, that is, with the hydranth facing upward and also inwards. /• and G. A quarter of a body segment regenerates with the hydranth facing inwards, at right angles to the original polar axis. //. Similar to G except that the large entoderinal gastric ridge prevented closure of the wound. Oral lips and tentacles have formed from the margin of the wound. end; but the hydranth is oblique, facing inward as well as upward (Fig. 7, D and £). If, however, the piece be short, the new hydranth may be entirely symmetric and face directly inward (Fig. 7, F and G.) The oral lips and tentacles are in this case formed from the cut margins of the piece. Indeed, the wound may never close, as in cases in which the entodcrm of the gastric ridge is disproportionately large (Fig. 7, //.) These facts suggest that polarity in scyphistomas is less a matter of BUDDING AND LOCOMOTION IN SCYPH'ISTOMAS HI orientation (for example, of molecules or of "'intimate structure") than it is of gradation. The upper body portion is strongly hydranth- forming. Moreover, the hydranth-forming potential is not greatly dif- ferent between the opposite surfaces of thin pieces taken from just below the tentacles. Were polarity primarily orientation, the propor- tion of biapical regenerates would not increase as sections are cut thinner. If the basis of polarity is orientation, it must be of a very labile nature, to account, for example, for the change in the direction of polarity which results from wounding. Unlike the hydranths, the pedal stolons do not form at cut surfaces. In the case of biapical regenerates they form from the side of the piece. In monapical regenerates they form from uninjured body wall just to one side of the healed lower cut surface. In longitudinal pieces the stolons form from the non-wounded surface (see later). It is obvious that stolon formation obeys different laws with respect to polarity and with respect to the effects of wounding, from the laws of hydranth formation. Regeneration from Ectoderm and Entoderm It is fortunately possible to separate pieces of ectoderm and entoderm and to observe the behavior of these two germ layers separately in ex- plants. It is then found that entodermal pieces round up into ciliated balls, which may remain alive and rotating for several days, but which do not regenerate (Fig. 5, £). Pieces of ectoderm, on the other hand, round up, and within a period of 7 to 11 days regenerate small, com- plete polyps (Fig. 5, F, G, and //). The evidence indicates that ento- derm is irreversibly differentiated as entoderm, whereas ectoderm is labile and dependent for its differentiation. (The reversibility of the ectoderm is presumably due in whole or in part to the presence in ecto- derm of so-called interstitial or restitutional cells.) Whether or not the entoderm is regionally differentiated has not been directly deter- mined. There is, however, indirect evidence that it is. When ectoderm is isolated alone it rounds up and within a few days regenerates fairly well balanced polyps, the number and size of the tentacles of which depend on the size of the piece (Fig. 5, G and PI}. The form of the regenerate has little relation to the region of the poly]) from which the ectoderm is taken. (Ectodermal explants have not been made from the oral disc, tentacles, or base.) Pieces from the stalk region, however, are somewhat slower regenerating than pieces from the body. When ectoderm and entoderm are isolated together, the re- generates reveal very distinct regional differences, as has already been described. These observations indicate that regional differentiation 112 FRANCIS GILCHRIST exists primarily in the entoderm, and that the ectoderm is largely de- pendent upon the entoderm for its regional characters. LOCOMOTION Experiments on the Origination of Stolons In the preceding section we noted that the pedal stolons of the scyphistomas typically form from the upper portions of the stalk, and that as they grow outward they tend to point upward away from the base. They also have a tendency to form on the side of the stalk away FIG. 8. Pedal stolons formed by half polyps. A. A polyp has been cut ap- proximately into halves. B. The piece which retains the base forms a typical pedal stolon from the side opposite the cut. C. The piece which lacks the base forms a short stolon from near the lower end. D. \ half polyp in locomotion toward the left. E. Later stage of the same polyp. from the remnants of older stolons. These facts suggest that the base and the older stolons have an inhibiting influence upon the formation of new stolons. To tesl this assumption two experiments were performed: ( 1 ) Twenty polyps were selected which showed neither buds nor Ions, and to which no injury had been done. A bit of the calcareous -uhstrate adhering to the base was taken as evidence that the base was intact . Krom ten of these polyps the base was then cut away. Twenty- four hours later about e«|iial numbers of the operated and the unop- eratcd polyps (five and six respectively) had produced stolons. The rate of stolon formation is therefore not greatlv, if at all, influenced bv BUDDING AND LOCOMOTION IN SCYPHISTOMAS the removal of the base. But the stolons of the two groups were char- acteristically different. In all but one instance, the operated animals produced stolons which grew downward from the lower end of the re- maining stalk ; while the control group produced long, typical pedal stolons which grew upward and outward from the side of the stalk. The experiment was repeated with similar results. (2) Twenty polyps were again selected as before, but were cut 8 FIG. 9. Operations on pedal stolons. A. A section of the stalk including a beginning stolon. B, C, and D. The stolon elongates, attaches, and contracts, and in due time regenerates. E. An isolated stolon which has developed a branch stolon at its base. F. The original stolon has attached and contracted ; but the branch stolon remains unattached and elongated. G. A pedal stolon was crushed near its proximal end and immediately contracted in the region of injury (tempo- rary contraction). H. Later the tip attached and permanent contraction took place. /. An isolated stolon was cut into two pieces. / and K. The distal piece remained elongated until its tip attached, when it contracted. L and M. The proximal half immediately rounded into a ball which later adhered lightly to the bottom of the dish. longitudinally into halves, except that in each instance the cut passed just to one side of the base (Fig. 8, A). Forty-eight hours later essen- tially the same numbers of each group had produced stolons (15 of those with base; 17 of those without). In most cases those without a base had produced short, downward-growing stolons (Fig. 8, C) ; while all those with a base had formed upward and outward-growing stolons (Fig. 8, B). Thus again the presence of a base influenced the place of origin and the direction of growth of the stolons. 114 FRANCIS GILCHRIST A further observation in connection with this latter experiment is important : The stolons always grew from the uninjured side of the half polyp. This was not only true of the first stolon, but of several suc- ceeding stolons as well (Fig. 8, D and E) ; with the result that the partial polyps locomoted in one direction for a time; namely, in the di- rection away from the injured side. These observations may be summarized by saying that (1) the rate at which stolons are formed is little if at all influenced by the presence or absence of a base, or by wounding ; but (2) the position and direction of outgrowth of each new stolon is definitely affected. Xew stolons commonly grow from that portion of the stalk which is farthest from the base, and on the side opposite to the remains of older stolons. New stolons also grow from the side opposite a lateral wound. ls on the Elongation of Stolons A pedal stolon begins as a blunt cone, first of entoderm, then of entoderm and ectoderm as well. \Yithin the space of a few hours it elongates until it has become a long, thin, tendril-like process with a -olid core of entoderm. What processes are involved? We have al- ready observed that elongation is a process of morphallaxis, in which the growing stolon draws to itself material from adjacent regions of the stalk. Experiment confirms this observation, and also indicates the primary role of the tip in the process of outgrowth. (1) In several experiments beginning stolons were cut away along with a greater or less amount of the stalk (Fig. 9, A). In most cases the stolon continued to elongate, and in some instances (provided the base was not included) the entire fragment became transformed into a thin stolon of nearly uniform cross-section (Fig. 9, B). From this we conclude that the stolon is a " self -differentiating " system, physiologi- cally independent of the polyp proper, and capable of exerting some measure of control over stalk material immediately surrounding the ba-r of the stolon. (2) In further experiments the tips of elongated stolons were crushed by pinching them with forceps. The stolons immediately con- tracted; and within the space of a few hours they usually had become completely withdrawn into the material of the polyp stalk. In some instances, however, when the injury was not too great, the tip was able to reor^ani/e it>elf, or a new tip just proximal to the injured tip was formed. When this occurred the stolons re-elongated. Again, when the tip of a stolon was lightly stained with the vital dye, Nile blue sul- phate, it remained extended: but when the staining was heavy, the stolon contracted. In several instances after such a contraction a new BUDDING AND LOCOMOTION IN SCYPH1STOMAS 115 tip developed just proximal to the stained region, and the stolon again elongated. We conclude from these experiments that the outgrowth of a pedal stolon is a self-determined morphallactic process, independent of the polyp proper, but dependent upon the presence and activity of a free and uninjured tip. Experiments on /lie Contraction of Pedal Stolons The contraction (or retraction) of a stolon normally takes place immediately after its tip has attached to some solid object. If attach- ment fails to occur, the stolon may remain elongated for two or more days without contracting. The following experiments show that physio- logical continuity with an unattached tip is necessary if stolons are to remain uncontracted. ( 1 ) It was regularly observed that when stolons are cut away from the polyp proper, they remain elongated until and unless the tips adhere to the substrate. The stimulation which results from injury at the proximal end of the stolon at most produces a slight, temporary contrac- tion of the ectodermal sheath. When the tip adheres to the substrate, however, the isolated stolon immediately and permanently contracts into a ball (Fig. 9, C). Control is obviously from the tip. A week or so later the ball will have regenerated a tiny whole polyp with mouth and circle of tentacles (Fig. 9, D}. (2) In many instances isolated stolons developed a second tip, usually near their proximal end. When this occurred a branch stolon grew out (Fig. 9, £) under the controlling influence of the second tip. It was then noted that when one of the tips attached to the substrate, only that portion which was under the control of that tip contracted. The part under the control of the free tip remained elongated (Fig. 9,F). (3) When isolated stolons are cut into two or three pieces, the distal pieces with their free tips invariably remain elongated until and unless attachment occurs (Fig. 9, 7, J, and K). (The few exceptions were presumably due to injury to the tip.) The proximal and inter- mediate pieces, on the other hand, invariably round up into balls (Fig. 9, I, L, and M). Occasionally after rounding up, such a piece may develop a tip and then re-elongate ; but in every case observed the time interval was long (at least twenty-four hours), and the phenomenon was clearly one of regeneration. (4) When an elongated stolon is crushed by a pair of forceps at some point between the tip and its base, a wave of contraction spreads both directions from the point of injury. The contraction which 116 FRANCIS GILCHRIST moves toward the tip is temporary (it is apparently of ectoderm only), and indeed it never quite reaches the tip (Fig. 9, G). In a short time this distal portion is again elongated. The contraction which moves proximally, on the contrary, is permanent. (It presumably is a tem- porary ectodermal contraction followed by complete ectodermal and entodermal retraction.) A day later the proximal portion, in the case of isolated stolons, is rounded into a ball. In the case of stolons which have not been cut away from the polyp, the portion proximal to the cut becomes completely resorbed into the stalk, while the distal part may dangle as a loose appendage connected with the polyp by a thin filament of ectoderm only. In Fig. 9, // such a piece has attached to the sub- strate and contracted. (5) Ligatures of fine thread were tied about elongated stolons, breaking the continuity of the entodermal core but not of the ectodermal sheath. It was then observed that the portions of the stolons proximal to the ligatures promptly contracted, and in many instances succeeded in drawing the distal portion through the ligature and into the wall of the polyp stalk. Any portion of the stolon which remained distal to the ligature, however, continued elongated. These several experiments plainly indicate that the control within the stolon is a one-way control, that it is of the tip over more proximal levels, and that physiological continuity, especially of the entoderm, is essential if the stolon is to remain elongated. It is further noted that neither the stimulus of wounding, nor the effects of healing of the wound, have a permanent effect upon the stolon in producing contrac- tion. Only the attachment of the tip, or injury to the tip, or the removal of the tip is effective. It has been emphasized that the presence of a free tip is necessary if the stolon is to remain elongated. However, if the cut be clean and very close to the end of the stolon, so that a part of specialized ectoderm of the tip remains as a part of the stolon, the stolon may in some in- stances remain elongated and in due time attach and contract normally. It thus appears that the functional "tip" of the stolon is the terminal glandular region of the ectoderm. DISCUSSION General Formative Principles The polyps which we have just described come as near to being the plastic "candle flames" and 'whirl pools" which Thomas Henry Huxley and others have discussed as any objects of animate nature, certainly as anv objects among the Metazoa. However, they are n< BUDDING AND LOCOMOTION IN SCYPHISTOMAS 117 without a considerable degree of stability of structure. The upper body possesses a definite morphology which is relatively constant. The lower body and stalk, although plastic and changing, nevertheless per- form according to certain general principles. Still, it cannot be said that these principles are always rigidly obeyed. As a first principle we note that although the outgrowths which take place from the lower body and stalk possess one or the other or both of two potentialities; namely, the power to form hydranths and the power to form stolons ; yet these two potentialities are not realized to the same extent and in the same way at the different levels of the poly]) proper. (a) The power to form a hydranth may be realized at any level below the polyp's hydranth. In the lower body and upper stalk, how- ever, there is a sort of opposition between the parent hydranth and the bud hydranth, which leads to the bud promptly pinching itself away from the parent (fig-type of bud), and retaining its connection with the parent by only a thin strand of ectoderm. When the differentiation of a bud hydranth is delayed (stolonic buds), the pinching away from the parent is also delayed. In the lower stalk region no such opposition exists, and the buds which are formed here (hydra-type buds) long remain attached to the parent by a broad, fleshy union. Intermediate conditions are to be found at intermediate levels. The experiments upon the regeneration of fragments show that the power to form hydranths is greatest in the material immediately below the circle of tentacles ; yet buds do not normally form here. It is ob- vious that the hydranth of a polyp inhibits the formation of a bud and so of a second hydranth in the very region where the tendency to form hydranths in explants is strongest. (b) The potentiality of stolon production is most pronounced in the upper portion of the stalk. It is here that the more typical stolons are produced; and it is here also that the buds (fig-type) are most certain to form stolons. In the lower stalk region there is an opposition or incompatibility between the polyp proper and the production <»l stolons, with the result that but few stolons are produced, and these are not typical stolons, but short processes which attach quickly. The buds of this region (hydra-type buds), moreover, are mostly without stolons. Why does this incompatibility exist? A first hypothesis is that the base and the older stolons exert some sort of inhibitory influence upon the production of new stolons. Against this view, however, is the fact that the removal of the base has little if any influence on the rate at which new stolons form. A second possible hypothesis is that the entoderm in the region where new stolons form is in a sense " younger ' 118 FRANCIS GILCHRIST entoderm, that is, it has not been involved in stolon-forming activity for a longer period of time. Tn brief, the location and structure of the several types of buds and stolons illustrate a principle of opposition and toleration: opposition between parent hydranth and the bud hydranth, and between the base and the stolon; toleration, however, of hydranth for stolon, and of base for buds. Buds and stolons develop most frequently at inter- mediate levels because the material of these intermediate levels is less differentiated either hydranth- ward or base-ward than that of the upper and lower levels. The Role of the Germ Layers A second principle to be noted concerns the parts played by the two germ layers in budding and locomotion. The entoderm appears to play the independent role. Its cells rearrange themselves in characteristic fashion as the- bud develops or as the stolon forms and elongates. The ectoderm appears to be more passive, responding perhaps to some sort of inductive stimulus of the entoderm, especially where the ectoderm is in close contact with the entoderm; or adapting itself to the form as- sumed by the entoderm. Except in the tentacles, at the oral lips, and at the tips of stolons little specialization of the structure of the ectoderm is to be observed. There is experimental evidence that the entoderm is the principal -eat of invisible differentiation. \Yhen small pieces of ectoderm from the body or stalk are isolated they constitute themselves into whole and fairly well-balanced polyps, and they do so quite irrespective of whether they are from upper or lower levels. Similar pieces of entoderm show a higher degree of differentiation by not regenerating. Pieces which are composed of both ectoderm and entoderm, however, regenerate, but show strong regional differences, both qualitative and quantitative. Buds Why do bud- arise? There is no answer more satisfactory than that given by Child — that they arise as a result of physiological isola- tion when the region which becomes the bud escapes from the dominat- ing and inhibiting activity of the parental hydranth. They do not arise near the upper end because the inhibitory activity of the parental hy- dranth i-. strongest here. They do not arise at the lower end because the hydrantb- forming power is weakest here. They arise at an inter- mediate level. The pinching away of the bud from the parent's body accompanies the differentiation of the bud hydranth. BUDDING AND LOCOMOTION IN SCYPHISTOMAS 119 An important feature to note is that the pinching away of the hud is at first complete only so far as the entoderm is concerned. The hud may retain its connection with the parent hy a strand of ectoderm for a considerable time. The morphogenetic correlations seem to take place within the entoderm. Pedal Stolons Origination. — The problem of the origin and behavior of pedal stolons is of special interest. The polyps produce pedal stolons (also buds) one after another more or less continuously. Isolated pieces sooner or later produce stolons. Why then do stolons originate? The answer is not " polarity," for the stolons usually form at intermediate levels, and in explants they sometimes arise very close to the hydranth. It seems as though something accumulates in less differentiated regions which leads to the production of stolons. The earliest sign of a pedal stolon is a cone of entoderm on the side of the stalk, which pushes outwardly, displacing the intervening meso- glcea, and comes into close contact with the surface ectoderm. As soon as this occurs the ectoderm responds by becoming the '" tip " of an outgrowing stolon. One is inclined to interpret the process as an " induction " comparable to that by which, in amphibian development, the chorda-mesoderm of the primary archenteric roof induces overlying ectoderm to become the neural plate. However, critical experiments upon this phase of stolon development have not yet been performed. Elongation. — As soon as a specialized tip is present the stolon elon- gates ; and it does so by drawing to itself material from the stalk. Hav- ing elongated, it continues in this state, so long as the tip is free, so long as it is uninjured, and so long as physiological continuity remains be- tween the tip and the remaider of the stolon. Something which the tip does produces elongation. What may be the nature of this action? A first suggestion is that the tip dominates in the physiological manner which Child has frequently described. The pedal stolon is then to be compared with a bud; its tip is the apex of the bud. According to Child a bud originates when some region on the side of the parent's body becomes so increased in its rate of physiological activity that it becomes " physiologically isolated " from the individuating forces of the remainder of the animal. The region then begins to grow away from the parent and to differentiate into a new individual. Child (1929) was able to produce such a bud experimentally in the hydroid Cory- iiiorpha. a form which normally never produces buds in nature, by merely wounding the side of the stalk of the polyp. The first sugges- 120 FRANCIS GILCHRIST tion, then, is that the outgrowth of a pedal stolon is an example of physiological isolation and comparable to budding. The tip of the stolon is physiologically dominant, the remainder of die stolon sub- ordinate. Several observations indicate against this hypothesis and emphasize the contrast rather than the similarity between the origin of buds and the origin of pedal stolons. (1) The tip of a stolon never differentiates into the apex of a polyp; that is. into a hydranth. Indeed, the opposite occurs. When a stolon is cut away and attaches by its tip, it is very apt to regenerate a small hydranth ; but in this case the apex of the hydranth is always formed at the proximal end of the stolon. In fig-type buds, similarly, the physiologically dominant apex of the bud becomes the oral region of a hydranth ; but the stolon is an outgrowth from the side of the bud of a different physiological nature. (2) A hydranth may be produced by the stimulation of wounding, but stimulation or wounding never produces stolons. Again, the oppo- site is the case. When a polvp is divided longitudinally, the stolons always form from the uninjured surface of the pieces. Moreover, stimulation or injury to the tip of an elongated stolon results in the immediate retraction of the stolon. (3) The development of a hydranth seems to depend upon some general and quantitative property of the cells which produce it. such as a high rate of metabolism; but the development of a pedal stolon depends upon the presence of a definite specialized region which we have called the tip. So long as a part of this specialized tip is present. elongation continues; but if the specialized tip is removed, contraction takes place. A second lupothesis to account for the control which the tip of a pedal stolon exerts over the rest of the stolon is that the tip is an organ of internal secretion and produces a hormone, which, diffusing into the entoderm. cause* the latter to organize itself as the central core of a stolon. This hypothesis recalls the activity of the growth hormone ot plants, auxin. It has been found that in plants the growing tips of most stems, petioles, flower stalks, and coleoptiles produce a substance which, diffusing from the tip into the regions immediately below, induce the cells of the latter to elongate' in a longitudinal direction (see Went, l'M5 ). May it not be that something comparable to this takes place in the pedal stolon? Mav we not .suppose that the specialized glandular cells of the ectodermal "tip" of the stolon act as an organ of internal secretion and synthesize a "growth hormone" which diffusing into the entoderm causes the cells of the latter to arrange themselves into a BUDDING AND LOCOMOTION IN SCYPH1STOMAS 121 solid core? May we suppose that this internal secretion of the hormone continues until the specialized cells come into contact with a solid surface (or are injured) and discharge their cementing substance to the out- side? Internal secretion then ceases, the hormone supply is cut off, and the elongated stolon is retracted. This hormonal hypothesis seems to cover the facts. However, be- fore it can he more than a mere working hypothesis, it will be necessary to know more concerning the structure and interrelations of the germ layers at the tip of the stolon; and grafting experiments will have to he performed. If we may provisionally accept this hormonal hypothesis, then it follows that within the polyp we have two different physiological types of control : (a) Within the hydranth we have dominance and subordina- tion of the sort which Child has described. In relation to a gradient- field which is established, the localization and differentiation of the zones of the hydranth is accomplished. Dominance of this sort is not due to a specific influence emanating from the functional apex, but rather to the establishment of a labile configuration of forces with the apex as its center (see Gilchrist, 1937). (b) We have secondly, by hypothesis, hormonal control within the pedal stolon. The early differentiation of one small region as a specialized organ of internal secretion results in a definite arrangement of the cells in the surrounding area, and a stolon is produced. Both controlling regions, the lips of the mouth and the tip of the pedal stolon may be termed apices or " centers of organization " ; but the apex of the hydranth dominates because of its high rate of physio- logical activity and its influence on the configuration of forces in the gradient-field; while the apex of the stolon controls because of the ac- tivity of a specific product of its metabolic activity. The first type of control is quantitative and dynamic ; the second qualitative and chemical. The first, to employ Child's (1921) terminology, is transmissive ; the second is transportive. Moreover, the differentiation of the hydranth and the development of the stolon are different sorts of processes. The first is a relatively irreversible process in which cells become specialized as oral lips, peristome, or tentacles. The formation of a stolon, on the contrary, is entirely reversible. It is a temporary arrangement of the nature of a morphallaxis, which persists only so long as the hormonal control emanating from the tip persists. Contraction. — We have described two different types of contraction, which we may designate as temporary and permanent contraction. Temporary contraction takes place when a stolon is mechanically stimu- lated, as by pricking it with a needle or pinching it with forceps. The R> JtS^fAS 122 FRANCIS GILCHRIST wave of response is then seen to move in both directions from the point of injury, although the wave which moves distally may never reach the tip. Temporary contraction of this sort is never complete, and is ap- parently of the ectoderm onlv. Within the space of an hour or so the stolon is again fully elongated. Permanent contraction, or better, retraction, is of an entirely dif- ferent nature. It takes place normally when the tip of a stolon at- taches to the substrate. Experimentally it may be produced by injuring or removing the tip, or indeed by merely breaking the physiological continuity by means of a ligature. The first evidence of such contrac- tion is seen in the movement of the cells of the entodermal core of the stolon toward the stalk of the polyp. Permanent contraction is best thought of as a negative process; a sort of undoing of the factors which produce elongation. In terms of the hypothesis of a growth hormone, permanent contraction results from the disappearance of the hormone. The ectoderm is active in physiological (that is, temporary) contrac- tion. Certainly in the stolon, the entodermal core appears to be pas- sively compressed or thrown into a spiral when the stolon is stimulated. The entoderm, on the other hand, appears to play the more active role in morphogenetic (permanent) contraction. It plays the primary part also in the plastic straightening into an upright position of the stalk of a polyj) after it has been bent down as a result of stolon attachment. In these morphogenetic changes of form the ectoderm appears to be passive, or at least to do little else than contract and thus supply tension. SUMMARY The processes of budding and locomotion have been studied in a scyphoxoan polyp, presumably the scyphistomas of Aurclia. It has been found (1) that the buds which develop near the hydranth of the polyp are quick to pinch away from the polyp (except for a connecting strand of ectoderm) and to migrate away by means of pedal stolons; whereas the buds which develop near the polyp's base may long remain attached. This indicates an opposition between the hydranth of the polyp and the bud. (2) The stolons which develop at the upper end of the stalk of the p«lyp are quite commonly organs of locomotion as well as of attach- ment. They elongate, attach by their tips, contract, and so draw the polyp forward. The new stolon becomes the new base of the polvp, while (be i ild base becomes drawn out and finally breaks from its attach- ment. The fact that new stolons form awav trom the base and away BUDDING AND LOCOMOTION IN SCYPHISTOMAS 123 from the older stolons indicates an opposition between base and young stolons. (3) The entodcrm apparently plays the primary role in the forma- tion and differentiation of buds, and in the formation, contraction, and final resorption of pedal stolons. Entodermal cells of the stalk of the polyp rearrange themselves into a solid core as the stolon elongates ; they actively migrate back into the stalk as the stolon contracts. (4) Small pieces taken from any region of the polyp show some power of regeneration. In general, there is a decline in hydranth- forming potentiality from the upper to the lower end. Pieces from the hydranth (oral disc and circle of tentacles) are irreversibly deter- mined to form parts of a hydranth. Pieces from the body or stalk may regenerate whole polyps, although the relative size of the regenerated hydranth decreases as the base is approached. Base-forming tendency is strongest at the lower end. The power to form pedal stolons is greatest in pieces from the upper stalk. (5) Small pieces of ectoderm only may round up and regenerate whole polyps. Pieces of entoderm round up but do not regenerate. Pieces including both ectoderm and entoderm regenerate in a manner typical of the region from which they are taken. The entoderm is thus the seat of irreversible invisible differentiation (chemo-differentiation). (6) Various operations were performed on pedal stolons and upon the polyps which produce them; such as fragmentations of the polyp, and injury, isolation, fragmentation, and ligation of the stolons. The results indicate (a) that new stolons commonly grow from portions of the polyp's stalk farthest from the base and opposite the remains of older stolons, also on the side opposite a lateral wound; (b) that the outgrowth of a stolon is a self-determined morphallactic process inde- pendent of the polyp proper but dependent upon the presence and ac- tivity of a free and uninjured " tip " ; (c) that contraction immediately follows in a part of a stolon when the physiological continuity between the part and the tip is interrupted. (7) It is suggested that the formation of the specialized ectodermal tip of a stolon is the result of an induction originating in underlying entoderm ; and that the tip having thus originated acts as an organ of internal secretion in producing a " growth hormone." The hormone diffusing into the entoderm causes the entodermal cells to arrange them- selves as the solid core of a stolon. The secretion of the hormone ceases when the specialized cells of the tip come into contact with the substrate and discharge externally. 124 FRANCIS GILCHRIST LITERATURE CITED CHILD, C. M.. 1921. The origin and development of the nervous system. Univer- sity of Chicago Press. CHILD, C. M., 1929. Lateral grafts and incisions as organizers in the hydroid, Corymorpha. Physiol. Zool.. 2: 342. DALYELL, JOHN* GRAHAM, 1834. (On the propagation of Scottish zoophytes.) Edinburgh Nezv Philos. Joiini., 17: 411. DALYELL, JOHN GRAHAM, 1836. Farther illustrations of the propagation of Scot- tish zoophytes. Ibid.. 21: 88. GILCHRIST, F. G., 1937. The hydmin ilubia (1935). See Plates I and 1 1. Doflein ( 1'MSj li-ured the division sphere of Ainccba protcus and Levy (1924) MITOSIS IN AMCEBA PROTEUS 129 presented outline sketches showing the various changes in the division sphere during fission in this species. The most accurate drawings showing this process are those of Chalklcy and Daniel (1933). The " hyaline area " indicating the position of the nucleus during early division phases mentioned by Chalkley and Daniel could usually be seen under the binocular dissecting microscope. The present authors found, as in Amoeba dubia, cytoplasmic currents and general internal activity occur, although from external view the animal seems to be completely inactive. Observations on the living division spheres under the highest magnification possible did not reveal the dividing nucleus. Various procedures were used in preparing the living division spheres for study. In none of the preparations, including those made by the agar method kindly suggested by Dr. Chalkley, could any of the early stages in nuclear division be seen. At the time of cytoplasmic division, nuclei could, however, be readily observed. As shown in Plates I and II the cytoplasmic division may follow one of two types both of which also may occur in Amoeba dubia. The first type of division is shown in Plate I. Here in Figs. 1-5 are shown slight progressive changes in a living division sphere prior to elonga- tion. A slight elongation can be noticed in Fig. 5. In Fig. 6 projec- tions indicate the beginning of formation of the daughter amoebae. These projections grow, pushing out in opposite directions and the remaining portion of the division sphere becomes drawn out into a cylindrical form as in Fig. 7. The connecting cylindrical part gradually becomes thinner and the daughter cells have numerous short, blunt pseudopodia. As in Amoeba dubia, cytoplasmic currents can be observed to flow, first in one direction and then in the other in this connecting strand. The pulling out process proceeds rapidly (Figs. 8-11) until, just before the connecting strand breaks, all motion of cytoplasm within it ceases. Each broken end snaps back towrard its respective daughter amoeba becoming broader and thicker as it is withdrawn. It is interest- ing to note that this type of division although most frequently met with in Amoeba dubia occurred less frequently in Anweba protcus than the second type. In the second type of cytoplasmic division no long connecting strand can ever be seen. The stages preceding either type of division are similar up to the point (Fig. 6) where indications of the forming daughter amcebse become obvious. The two daughter amoebae flow out very rapidly forming very large coarse pseudopodia (Figs. 12 and 13). The connection between them as shown in Fig. 14 is very short as compared with that in Fig. 11. From this point separation occurs v/ithin a few seconds leaving the daughter amoebae lying close together. 130 J. A. DAWSON, W. R. KESSLER AND J. K. SILBERSTEIN In some instances it is difficult to detect the complete division until after fixation (Fig. 18). Our results agree closely with those of Chalkley and Daniel (1933) in respect to the duration of the process of fission in Amoeba protcus. In general it has heen found that the correlations between pseudopodial width and stage of nuclear division as described by the above-mentioned writers are roughly correct. In our experience, however, wide diver- gences may exist between pseudopodial configuration and the stage of nuclear division. Division spheres which had already begun to elongate showed, upon fixation and staining, early stages of division. The pseudopodia are coarser in the later stages of fission. Tn this work, however, it has been found impossible to isolate any desired stage of nuclear division with the 97 per cent accuracy claimed by Chalkley and Daniel (1933). THE PROPHASE I The resting nucleus of Amoeba protcus which is entering on the prophase becomes considerably swollen so that when observed in sur- face view, although still discoid in shape, the slight, biconcave depressions have now pushed out. The nuclear membrane shows clearly, being somewhat thinner than in the vegetative condition. The peripheral granules still stain fairly strongly and have now decreased in number. They are much less uniform in size and more widely separated. At the same time there is forming, in the central portion of the nucleus (endo- some) a mass of numerous small, deeply-staining, rod-like granules which are all definitely less than 0.5 /j. in length (Figs. 43, 44, 45 and 60). These granules have a plate-like arrangement. Spindle fibers are lacking at this stage. According to Chalkley and Daniel " the chromatin has left the membrane and is evidencing a tendency to aggregate in a zone . . . the chromatin appears completely EXPLANATION OF PLATE I Photomicrographs. All figures were photographed using a Leitz compound microscope and Leitz apochromatic lenses. Number 3 objective and 15 X oculars were used in all cases. All figures were made from living amoebae in process of division. Magnification approximately 90 diameters. FIGS. 1-5. Amoeba protcus Pallas (Leidy). Showing successive stages in division spheres prior to elongation. Compare with Figs. 17 and 18. FIG. 6. Division sphere showing beginning formation of daughter amcebie. FIG. 7. Slightly later. FIGS. 8-11. Showing cytoplasmic bridge connecting daughter amoeba; in suc- cessive stages. Type I division. See text. MITOSIS IN AMCEBA PROTEUS ..,-, 132 J. A. DAYYSOX, W. R. KESSLER AND J. K. SILBERSTEIX withdrawn from the membrane . . . tending to form an irregular ring." Later (1936) Chalkley retracts this statement saying that the peripheral granules play no part in the formation of the equatorial plate. In our preparations no evidence can he found that the peripheral granules leave their position close to the membrane and migrate to a central portion of the nucleus. On the contrary, they gradually fade out of the picture and take no part in the formation of the plate. In preparations of this stage very definite evidence is present to show that the granules which are to form the plate are at no time arranged in a ring hut are in the form of a disk. THE MKTAPHASE Beginning with the definitely formed metaphase plate as shown in Figs. 30-35, 46 and 61 the following condition is found. The nuclear membrane is still present. This is especially clear from our sections. Fibers are now seen extending from the plate to the nuclear membrane. At this sta-e. due to the shape of the nucleus and the delicate character of the membrane at the polar regions of the spindle, multipolar ap- pearances have been observed. This effect is, we believe, due solely to fixation artefacts. The plate is composed of a very large number of individual, deeply-staining small chromatin granules. It is approxi- mately circular in outline, the diameter being about 25 //,. The granules now are ± 0.3 /u, in diameter and are quite uniform in size. In a few preparations, especially in sectioned material, larger achromatic granules may be observed. These lie outside of the plate among the spindle fibers. It is believed that these bodies are remnants of the former peripheral granules which have not yet disintegrated and that they take no part in the mitotic process (Figs. 35 and 46). The dividing chromatin of the metaphase plate in Amoeba protcus EXPLANATION OF PLATE II Photomicrographs. Fi.mires 12-15 from living annelne in process of division. •lil'icatiiin approximately W diameters. FK;S. \2 15. .lntn-l>. Total mount. Division sphere in pn>pha->e. I leidenhain's heina- lin. Y 7n. l-'i'.. 17. Total mount. I)i\i>ion sphere in late anaphase. Safranin. X 400. I-'n;. IS. Total mount, (\toplasinic division, Type- II. Xote peripheral po- sition of nneli'i. 1 I.-idenhain's hematoxylin. X 4011. MITOSIS IN AM. 37. 47. 48 and <>2). The nuclear membrane, although delicate, is still present and may be traced com- pletely around the figure, especially in sectioned material (Figs. 36 and 62). The spindle libers extending between the two separating plates can he clearly seen at this stage. They are extremely numerous. The polar liber- now definitely terminate at a point, giving a bipolar spindle. The chromatin granules comprising the plates are similar in size and -taining capacity to those of the preceding stage. The larger achro- matic granules now occur less frequently but a few may be observed in sectioned material (Figs. 3<>, 37, 47 and 48). As the anaphase progresses the daughter plates become more widely M-parated (Figs. 38, 39, 40, 49 and 50). The nuclear membrane is still present. It is now very delicate and cannot be seen in rdl total mounts. In sections, however, its presence may still be detected. In EXPLANATION OF PLATE III Photomicrographs. All magnifications X 1,200. All figures from sectioned amu-ha1. Sections (> p thick. All fixed in Flemming except Figs. 27-29 which were fixed in Schaudinn's. All stained with Hcidenhain's hematoxylin except Figs. 28 and 29 (Feulgcn) and 30-34 (sat'ranin). FIGS. 19-21. Vegetative nucleus serially sectioned through surface. Note deeply stained peripheral granules and in Fig. 20 endosome showing throughout. I- n;s. 22-26. Vegetative nucleus serially sectioned at right angles to long axis. Fie. 27. Vegetative nucleus showing endosome as fixation artefact. IMCS. 28 and 29. Vegetative nucleus. Fculgen. In hoih the peripheral granules stain deeply. Fixation as in Fig. 27. IMCS. 30-34. Serial section of metaphase plate. Note spindle fibers and unclear membrane. Compare with Fig. 4<>. IMC. 35. Early metaphase plate. Xotc achromatic grannies and nuclear membrane. !MI,~. 3'p and 37. Successive \iews, early anaphase. Note nuclear membrane, divided plate, bipolar spindle and achromatic granules. Compare with Figs 47 and FJC-. 38-40. Si-rial sections through mid-anapbase. N'ote nuclear membrane • specially at polar regions, spindle fibers and achromatic grannies. Compare with Figs. 49 and Fii.-. 41 and 4_'. Telopha-r, side and surface view- from same ainu-ha. Note "parachute" in I-'i-. -II and convex type of plate-. \"o achromatic granules present. In Fig. 42 note nud<-ar membrane and line granulation. Compare with Fig. 54. MIT< )SIS IN \.\I<1 I! \ I'l't i I I I S 19 22 23 24 - . 21 •,: ' 34 T'tr* 36 * - * • If 37 ^ ' it 01 4t * . 25 26 -• 38 - 40 42 PLATE III 136 J. A. DAYYSOX, W. R. KESSLER AND J. K. SILBERSTEIN tnid-anaphase the spindle fibers are numerous and clear as are also the polar fibers. An excellent idea of the nature of the daughter plates in this stage may he had from a three-quarter view as shown in Figs. 51 and 63. The size and disposition of the chromatin granules in the plates are clearly shown. The plates move farther apart and their planes still remain in most cases parallel. Each gradually becomes saucer-shaped with the convex surfaces facing each other (Fig. 52). The polar spindle fibers are still present but the libers between the plates have practically disap- peared. At this stage a few achromatic granules may still be seen in some preparations. In a similar stage in Aincrba dubia the achromatic granule- are relatively numerous. The nuclear membrane so clearly visible in Ania-ba dubia at this stage becomes difficult to follow in Anuvba prolcns. In our prepara- tions the nuclear membrane is present, surrounding the polar regions EXPLANATION OF PLATE IV Photomicrographs. All magnifications X 1,200. All figures from total mounts except Figs. 55-59 which are from sections, 6 M thick. Figures 43-54 fixed in Carnoy-Lebrun. With the exception of Fig. 58 all remaining figures from amceba; fixed in Flemming. All stained with Heidenhain's hematoxylin. FIGS. 43-45. Nuclei in prophase. Note peripheral granules and cloud-like forming plate. Compare with Fig. 60. FIG. 46. Metaphase plate. Note achromatic granules. Compare with Figs. 30-35 and Fig. 61. FIGS. 47 and 48. Early anaphase. Two views of same figure at different focus. Note separating plates, spindle fibers and achromatic granules. Compare with Figs. 36, 37 and 62. FIGS. 49 and 50. Mid-anaphase. Two views of same figure at different focus. Note spindle fibers. Compare with Figs. 38-40. Fi<;. 51. Mid-anaphase. Three-quarter view showing granulation of plates. Compare with Fig. 63. FIG. 52. Late anaphase. Note curving of plates, polar fibers and achromatic granules. Compare with Fig. 17. FIG. 53. Early telophase. Note greater curvature of plates and achromatic granules. Compare with F'ig. <>4. IK;. 54. Later tel<>ph;iM- showing one nucleus in side and one in surface view. Xote character of grannies and forming "parachute." Compare with slightly lati-r condition in Figs. 41, 4J and 65. FK.S. 55-59. Showing reconstructing nuclei. Figure 55 one-half hour after division. Xote small size of peripheral granules and vacuolated endosome. Figure 57. Two hours after division. Note typical nuclear shape. Peripheral granules increased in si/e. More granular. FlG. 58. Three hours after division. Xote increase in sixe of peripheral granules and those of tin- en-losomc. Compare with Fig. 68. Sectioned in surface view. FIG. 5°. Five hours after division. Xote similarity to vegetative nucleus. MITOSIS IN AMCEBA PROTEUS 137 • • + « / •! >;•••-> V ••- »: * .• 43 • ^. I V tf 't 44 45 I 46 47 48 53 52 49 55 -. 50 .••-. •'. 56 .1 . . <>. 54 PLATE IV 59 138 J. A. DAWSOX, \V. K. KESSLER AX1) J. K. SILBERSTEIX of the separating plates. It is difficult to make out with certainty on the inner surfaces of the plates. Some evidence of invagination or pinching off of the membrane in the equatorial zone of the spindle has been obtained. \Ve believe that the membrane does not break down but divides in this manner just after mid-anaphase and thus comes to encircle each daughter nucleus. THE TELOPHASE The telophase stage is cliaracUTi/i-d by the further migration of the daughter plates (nuclei). At this time the division sphere has definitely elongated and the daughter plates (nuclei) are found near its poles. An early telophase is shown in Figs. 53 and 64. The daughter plates are now thicker, somewhat more convexlv curved, and consequently have a shorter diameter. Measurements show the diameter now to be about 11 to 12 p. The nuclear membrane appears much more dis- tinct. The polar spindle fibers form a conical cap on the concave surfaci- of the plate. Xo sign of interzonal fibers can be seen. At the same stage in the division of Amoeba dubla such fibers are charac- teristically present. The disappearance of the achromatic granules is practically complete at this time. Later stages of the telophase are associated with cytoplasmic divi- sion of the animal. Typical later telophases are shown in Figs. 18, 41. 42, 54 and 65. The nucleus may now best be described by calling it a " parachute." The polar spindle libers represent the shrouds of the parachute and the plate, the dome. Such structure is most apparent in sections one of which is shown in Fig. 42. In surface view (Figs. 42 and 54) one can observe that tin- granules originating from the plate EXPLANATION OF PI, ATI-: V All drawings made with camera lucida. X 1, /"<><). FIG. 60. Same as Fig. 43. X'uclcus in prnphasc showing peripheral granules rly defined and forming plan. FlG. fil. Similar to Fig. 33. Median section through metapliase nucleus showing dividing plate-,, spindle fibers and nuclear membrane. FII;. f>2. Same as Fin. Reconstructing nucleus ten minutes after division. Note uniform distribution of granules and delicate nuclear membrane. FIG. ''7. l\i -i -misti'iicting nucleus one and a half hours after division. FIG. 68. Reconstructing nucleus three hours after division. XTote presence of peripheral granules. MITOSIS IN AMCEBA PROTEUS 139 — , , V -• f. •»•—.*» . «,«*'•• "V" X > *SF • «' 61 •'' L^ ^^^^111.'1 •n ^•x.i;i'^i;^y .\ , «6 / . ; ..• .». ^1 : w fJB PLATE V 140 J. A. DAWSON, W. R. KESSLER AND J. K. SILBERSTEIN are extremely fine and numerous. They do not stain so intensely. The diameter of the granules is unchanged. The diameter of the nucleus in surface view varies from 12 to \5 p. whereas the measure- ment of the short axis of the nucleus is approximately 4 to 5 p At this stage but one type of staining granule is present and the nucleus is obviously much smaller than in the adult vegetative condition. RECONSTRUCTING NUCLEUS Although at the time of cytoplasmic division the nucleus has a definite membrane, it can by no means be considered completely recon- stituted as Chalkley and Daniel (1933) state. We have studied the reconstituting nucleus both in total mounts and in sections from the time of division to over 5 hours thereafter. At the moment of division there is no evidence of the characteristic biconcave, discoidal shape (Fig. 18) but it resembles a thin saucer. Ten minutes after division it becomes slightly biconcave and measures approximately 20 p. in the long diameter and 6/1 in the short diameter. The former plate still occupies its central position in the nucleus. In surface view it is seen to be composed of numerous, extremely small (less than 0.3 /A in diameter) granules which stain less intensely than those of the telophase. During the interval between cytoplasmic separation and ten minutes after division, peripheral granules first make their appearance. These, in the beginning, are indistinguishable in size and staining capacity from the granules of the plate. From 10 to 30 minutes after division the nucleus becomes slightly larger being approximately 24 p, in the long axis at 30 minutes. The peripheral granules also increase slightly in size now measuring from 0.5 p. to 0.7 p. in diameter. They appear somewhat irregularly placed and seem to have fibrillar connections. The central region which marks the former position of the plate becomes slightly granular, vacuolated and stains more faintly (Figs. 55 and 66). At the end of one hour after division the nucleus is approaching normal vegetative size, measuring through the long axis approximately 30 p. and through the short axis 13 p.. The final definitive shape of the vegetative nucleus has been nearly if not entirely assumed at this time (Figs. 56 and 67). The peripheral granules are slightly larger meas- uring about 8 p. in diameter. The central portion of the nucleus, the endosome, is now vacuolated, has a reticular appearance and stains more heavily. It is definitely granular and extends throughout the entire nucleus. MITOSIS IN AMCEBA PROTEUS 141 Other than a slight increase in size, little change can be noted in the two-hour reconstructing nucleus (Fig. 57). Three hours after division both the peripheral and some of the central granules have increased in size (Fig. 68). The picture is not essentially different four hours after division. The peripheral granules, however, show a perceptible increase in size, being 0.9 /x in diameter. The central mass (endosome) is definitely granular and extends throughout the nucleus. It stains evenly and fairly heavily in sectioned material. About five hours after division (Fig. 59) at room temperature (in this case 30° C.) it is difficult to distinguish the reconstituting nucleus from the vegetative condition. DISCUSSION From the preceding account of mitosis in Amoeba protons it is obvi- ous that the process differs markedly from that in Amoeba dubia. It is also to a certain extent at variance with the accounts given by other workers with the same species. Descriptions of the vegetative nucleus of Amoeba proteus by former workers agree in general that a nuclear membrane is present and that a layer of peripheral granules underlies this membrane. In regard to the structure and function of the central portion of the nucleus or endosome considerable diversity of opinion exists. Thus Taylor (1923) describes an endosomal body which is clearly marked off from the outer region of the nucleus. Doflein (1918), although he does not specifically state a similar condition, has figured it as such. Our observations have led us to agree with Belar (1926) that such a sharply marked-off endosome is evidence of faulty fixation and our study of the structure of the vegetative nucleus of Amoeba proteus convinces us that the account given by Calkins (1898) is essentially correct. We are unable to confirm Schaeffer's statement (1926) that the vegetative nuclei of Amoeba proteus and Amoeba dubia belong in the same category. In Amoeba dubia no endosome as such could be seen but the staining granules were evenly distributed through- out the nucleus. We believe that the uniformly distributed mass of relatively large granules in the Amoeba dubia nucleus are homologous to the peripheral granules of the Amoeba proteus nucleus. Both disappear during mitosis and both are reformed " de novo " from the dividing chromatin. Doflein (1918), Belaf (1926) and Chalkley and Daniel (1933) all hold that the peripheral granules are the dividing chromatin and that the endosome supplies no chromatic elements. It is noteworthy that Chalkley in his recent brief account (1936) completely reverses his opin- 142 J. A. DAWSON, W. R. KESSLER AND J. K. SILBERSTEIN ion in this regard and states that " these granules give rise to, or con- tribute to, the formation of the spindle fibers and the pole caps." Our work shows that the peripheral granules begin to disintegrate during the prophase and do not migrate to the central portion of the nucleus where the metaphase plate is forming. The metaphase plate originates from the endosome. The multipolar metaphase figures shown by Carter (1912) and Doflein (1918) are, as can readily be seen from our photographs and figures, solely due to artefacts of fixation. It is noteworthy that such figures are shown from sections. We have found that the so-called '' multipolarity " is accentuated in sections through the outer portion. Chalkley and Daniel1 (1933) claim that the nuclear membrane dis- appears at metaphase. Although Carter (1912) figures a nuclear mem- brane surrounding the metaphase figure, Doflein (1918) questions this and states that the nuclear membrane is not present at this stage. On the contrary, we have found the nuclear membrane clearly defined as late as mid-anaphase and also present in the early telophase. Indica- tions of its presence in the intervening short period (at most five min- utes) may also be seen. We therefore believe that the nuclear mem- brane does not break dowrn during mitosis in Amoeba protcus. Our observations on the reconstructing nucleus in Amoeba protcns have led to a better understanding of the origin of the peripheral granules. These begin to reappear just after cytoplasmic division and gradually grow in size with the growth of the nucleus. The genesis of these granules is decidedly similar in Aniocha dubia and we are forced inescapably to the conclusion that they originate from the plate material, i.e., they are endosomal in origin. Binary fission in Amoeba protcns and Amoeba dubia represents, we believe, the method of reproduction. The conception of Doflein that multiple division is a regular occurrence in Amoeba protcns is, in our opinion, erroncu^. The critical review by Johnson (1930) concerning 1 In a more complete paper appearing in December, 1936, shortly before this paper went to press, Chalkley stresses again the absence of the nuclear membrane during metaphase and anaphase in the division of Amoeba protcus. Evidence is presented by us in this paper to show that the nuclear membrane is present during metaphase and early anaphase. We are in accord with Chalkley in finding that the dividing chromatin in Amoeba protcus originates in the karyosome (endosome), but it is difficult to follow his reasoning when he states that the polar caps, the spindle fillers and possibly the "new nuclear membrane" are formed in part from the peripheral granules. A study of his paper fails to reveal evidence for such a belief. It does seem entirely probable, however, that the so-called peripheral chromatin is derived from the karyosome. The fact that it gives the Feulgen re- action f"r :i short period during reconstruction while it is migrating from the karyosome to the periphery would clearly indicate that such is the case. MITOSIS IN AMCEBA PROTEUS 143 diverse theories of reproduction in the large free-living species of Amoeba and the recent work of Halsey (1936) serve to confirm this belief. SUMMARY 1. In Amoeba protcus the vegetative nucleus is a discoid, biconcave structure with a nuclear membrane, a layer of peripheral granules and a central endosomal mass. 2. During prophase the peripheral granules begin to disintegrate and the dividing chromatin originates from the endosome. 3. In metaphase the plate is fully formed and consists of numerous, small, deeply-staining chromatin granules. Spindle fibers, both inter- zonal and polar, are first apparent at this time. The interzonal fibers disappear at late anaphase. The polar fibers persist throughout the process. 4. The granules of the metaphase plate divide to form the two plates of the anaphase which continue to separate, becoming condensed and curved until the telophase stage. 5. The nuclear membrane is clearly present up to mid-anaphase and again at telophase. There is evidence that it persists throughout the entire process of mitosis. 6. The nucleus is fully reconstituted about five hours after cyto- plasmic division. The peripheral granules are formed from the plate mass and appear a few minutes after division. BIBLIOGRAPHY BELAR, KARL, 1926. Der Formwechsel der Protistenkerne. Ergcb. 11. Fortschr. der Zool., Bd. 6. CALKINS, GARY N., 1898. The phylogenetic significance of certain protozoan nuclei. Annals N. Y. Acad. Sci, 11: 379. CALKINS, GARY N., 1933. The Biology of the Protozoa. Lea and Febiger. Philadelphia. CARTER, LUCY A., 1912. Note on a case of mitotic division in Amoeba proteus Pall. Proc. Roy. Phys. Soc. Edin., 19: 54. CHALKLEY, H. W., 1930. Stock cultures of amoeba. Science, 71: 442. CHALKLEY, H. W., 1936a. The behavior in mitosis of the karyosome and so- called peripheral chromatin in Amceba proteus. Anat. Rec., 64: 104. CHALKLEY, H. W., \936b. The behavior of the karyosome and the "peripheral chromatin " during mitosis and interkinesis in Amoeba proteus with par- ticular reference to the morphologic distribution of nucleic acid as indi- cated by the Feulgen reaction. Jour. Morph., 60: 13. CHALKLEY, H. W., AND GEORGE E. DANIEL, 1933. The relation between the form of the living cell and nuclear phase of division in Amoeba proteus (Leidy). Physiol. Zool, 6: 592. DAWSON, J. A., 1928. The culture of large free-living Amoebae. Am. Nat., 62: 453. 144 j. A. DAWSOX, W. R. KESSLER AND J. K. SILBERSTEIN DAWSON, J. A., WALTER R. KESSLER AND JOSEPH K. SILBERSTEIN, 1935. Mitosis in Amoeba dubia. Biol. Bull.. 69: 447. DOBELL, C., 1914/). Cytological studies on three species of Amoeba. Arch, jilr Protist.. 34: 139. DOFLEIN, F., 1918. Die vegetative Fortpflanzung von Amoeba proteus Pall. Zoo/. Ans., 49: 257. HALSEY, H. RANDOLPH, 1936. The life cycle of Amoeba proteus (Pallas, Leidy) and of Amoeba dubia (Schaeffer). Jour. Expcr. Zool.. 74: 167. JOHNSON, PERCY L., 1930. Reproduction in Amoeba proteus. Arch, fur Protist., 71: 463. KIDDER, GEORGE W., 1934. Studies on the ciliates from fresh water mussels. IT. The nuclei of Conchophthirius anodontae Stein, C. curtus Engl., and C. magna Kidder, during binary fission. Biol. Bui!., 66: 286. LEIDY, JOSEPH, 1879. Fresh Water Rhizopods of North America. Washington. LEVY, JOSEPH, 1924. Studies on reproduction in Amoeba proteus. Genetics, 9: ' 1J4. SCHAEFFER, ASA A., 1916. Notes on the specific and other characters of Amoeba proteus Pallas (Leidy), A. discoides spec, nov., and A. dubia spec. nov. Arch, jiir Protist., 37: 204. SCHAEFFER, ASA A., 1926. Recent discoveries in the biology of Amoeba. Quart. Rev. Biol., 1: 95. TAYLOR, MONICA, 1923. Nuclear divisions in Amoeba proteus. Quart. Jour. Micros. Sci., 67: 39. Vol. LXXII, No. 2 April, 1937 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY REPRODUCTIVE CYCLES AND SUPERFETATION IN PCECILIID FISHES C. L. TURNER (Prom the Department of Zoology, Northwestern University) INTRODUCTION Reproductive cycles in the pceciliid fishes are of unusual interest for three reasons: (1) this family of fresh-water fishes arose in the tropics and for the most part has remained under tropical and sub-tropical con- ditions where the great seasonal variations in light and temperature found in temperate and arctic zones do not exist; (2) all the species in the family are ovoviviparous. The embryos have fairly large yolk sacs and are retained in the ovarian follicles after fertilization until birth, necessitating a respiratory exchange between parent and embryos. The ovary consequently serves the double function of providing the female gametes and of providing a site and the proper conditions for retaining growing embryos; (3) stiperfetation occurs in some of the genera. In some instances two broods of embryos of different ages are present in the ovaries and in the most extreme case there may be at one time as many as six small separate broods. Pceciliid fishes have been reared in aquaria for years and some of the general facts concerning size of broods, intervals between broods, cor- relation between size of brood and female parent, variation of brood size and intervals between broods at the different seasons of the year, are well known. Bits of reliable information concerning reproduction in a number of genera and species are found in most of the popular books on aquarium fishes. One species in particular, Ganibnsia affinis, has been introduced into new areas in the tropics as a mosquito eradicator and its life history and reproduction have been well studied. Such studies have given relatively little attention, however, to the developing ovocytes within the ovary and the relation between these groups of ovocytes and the broods. In order to study reproductive cycles effectively it has been found necessary to know the stages of development of the embryos and of the 145 146 C. L. TURNER ovocytes within the ovary and to correlate this information with the records of the reproductive cycles. Furthermore, it is necessary to have as much of this information as possible from single females. Conse- quently, as far as possible, it has been the procedure to secure the brood- bearing record of an individual and then kill the individual at a definite period after the bearing of a brood in order to examine the embryos and developing ovocytes in the ovary. By following this practice through- out the year data have been secured upon the following relations : ( 1 ) the total number of broods produced within a year; (2) the relative time interval between hn><>ds; (3) the relative size of broods; (4) the exact time of fertilization of a group of eggs as related to the birth of a previous brood ;md as related to younger ovocytes; (5) the period of time over which fertilization takes place; (6) the variation in stage of development within a single brood; (7) whether or not fertilization occurs so as to form a new brood before the former brood is voided (super fetation) ; (8) the relation of the interval between broods and of the size of broods to environmental factors. Any large collection of gravid females made in the field furnishes material for a study of the stages in the ovocytes as related to the age of the embryos being retained in the ovary. Superfetation if it exists can also be studied in this type of material. In the present .study the following species have been studied by the method of maintaining laboratory cultures and dating and dissecting the gravid females : Gainbusia a/finis, Gainbnsia holbrooki'i, Gainbusia pa- nitco, Mollicnisia splicnops, Mollienisia latipinna, Xiphophorus hcllcrii, Xiphophonis strigatus, Pseudoxiphophorus bimaculatus, Platypcecilus inaculatus, Platypcecilus rariatus, Lchistcs reticulatus, Liinia vittata, Heterandria forniosa, Micropcecilia pi eta, Micropcecilia parce, Poccilistcs pi euros p Hits and Quintana alrizona. The following have been studied by the use of large numbers of living specimens in the field: Xiplio- phonts hcllcrii , Xiphophorus striyatus and Pseudoxiphophorus b'nnacn- lalus, at Cordoba, Mexico (March, 1932) ; Pe present in the ovary at the beginning of the season." The writer is inclined to agree with Kuntz that different groups of ova reach maturity at different times and that each group is fertilized separately, for two reasons: first, all the young are born at practically the same time and none remain in the ovary ; second, within ten days after birth another lot of ova. approximately equal in number to the brood born previously, has developed and is ready for fertilization. It may be implied from the statement of Kuntz that he assumes Gambusia has an annual reproductive cycle in which all the ova for the annual cycle are elaborated at the same time. These interpretations concerning an annual cycle fail to take into account that the pcecili'ids arose under tropical conditions and that they have short, often-repeated cycles each similar to one complete annual cycle in a fish of the temperate zone. Under the alternating conditions of the temperate zone these short cycles are repeated as often as environmental conditions permit within one fa- vorable season and then become retarded or cease until favorable con- ditions are re-established. The ovarian cycle in Mollienisia latipinna and Mollienisia sphcnops resembles that of Gainbnsia. During the spring and summer broods are born approximately thirty days apart. Ova destined for a new brood reach a maximal diameter of about .7 mm. by the time the embryos of the earlier brood are ready to be voided — about twice the size of the cells in Gambusia at the same stage. However, they grow rapidly and reach a diameter of over 2 mm. by the time they are ready for fertilization. Apparently fertilization does not extend over so long a period in Mollienisia as in Gambusia and the embryos, just after fertilization, do not show the wide range in stage of development characteristic of Gambusia. Pseudoxiphophorus bimaculatns has an ovarian cycle very similar to that of Mollienisia latipinna and Mollienisia splienops. However, the interval between broods is longer, being thirty-five to forty days during May, June and July. In an ovary containing a brood ready for birth the next younger group of cells has reached a diameter of only .7 or .9 mm. in diameter. After the brood has been evacuated, however, the cells grow rapidly to a large size, about 2.4 mm., before fertilization. Lebistes Type. — The rhythm of brood production in this type is very similar to that in the Gainbnsia type. During late spring and summer broods are produced approximately every thirty days. A rec- 150 C. L. TURNER ord of brood production by ten females of Lebistcs rcticnlatus is shown in Table I. The data are taken from unpublished records of Dildine and the writer. It will be noted that the most regular production of broods and the shortest interval between broods occur in the late spring and early summer months. In late fall and winter months the interval be- tween broods is nearly as long as in the late spring when broods are produced from twenty-two to thirty days apart. The specimens were kept under fairly constant conditions as regards food, temperature and salt and gas content of the water. Xo artificial lighting was employed, however, and the exposure of the specimens to light was therefore con- TABLE I Showing dates of broods produced by ten specimens of Lebistes reticulatus . Numbers under months indicate day on which brood is produced. Intervals between broods are expressed in days. Specimen Number January Interval February Interval 1 3 Interval z« C. < Interval a % Interval V z. 3 >— i Interval jj, ~5 i — » Interval 4-» (A 3 SC 3 < Interval September Interval October Interval November Interval December > h. <• 26 5 21 39 29 22 20 26 16 31 16 6 1 <.< 4 27 1 34 4 7 22 28 19 23 13 8 18 9 40 9 60 28 39 8 28 6 27 2 20 10 30 2 trolled by the M-axmal waxing and waning of daylight. The increase and decrease in brood production paralleled the increase and decrease in daylight. To what extent increased exposure to light may serve as a stimulating agent to reproduction, possibly through the medium of a -/mail-stimulating hormone from the pituitary, remains to be worked out experimentally. In Xiphophorus and Platypcecilus the brood-bearing record is very similar to that in Lebistcs with a tendency to a slightly longer period be- tween broods. The same increase in the interval between broods dur- ing the winter months has also been noted. Conditions in the ovary during various phases of the reproductive cycle could be determined only by using the period in which a brood is REPRODUCTIVE CYCLES IN POECILIID EISHES 151 voided as a first stage and then obtaining regular stages a few days apart up to the time for the voiding of the next brood. Such a series was obtained during the height of the reproductive season. The find- ings in a number of cases are, included here to indicate conditions within the ovary. (1) A specimen of Lcbistcs was sacrificed immediately after it had produced a brood of nine young. No young were found in the ovary and it was evident that all of the brood had been born during a three or four-hour period. In the ovary were found six ova approxi- mately 1.4 mm. in diameter, orange-yellow in color and somewhat trans- parent. Four additional eggs similar in structure and color but vary- ing in diameter from 0.8 to 1.0 mm. apparently belonged to this group. Presumably all would have been fertilized to form the next brood. An- other group of eight cells represented a still younger set which would have been fertilized still later. These younger cells could be distin- guished from the older group of cells by their smaller size and differ- ence in structure. The younger group contained cells white in color, opaque and less than 0.5 mm. in diameter with more variation in size than in the older group of cells. It should be emphasized that there were definite gaps in size between the older and the younger group of cells and also between this younger group and a still younger mass of cells in various stages of development. (2) A second specimen was sacrificed and the ovary examined eight days after a brood had been born. Fertilization had recently taken place in a group of ova about equal in number to the brood born eight days before. Some of the fertilized eggs had reached an early blastula stage, some were in early segmentation stages and two had apparently not yet been fertilized. The eggs in the group as a whole varied in diameter from 1.2 to 2.0 mm. One of the larger eggs had not yet been fertilized while several of the smaller ones of the group had been fertilized. Therefore, size, except within the limits indicated, is not an index of the fertilizability of an egg. A second group of cells, about equal in number to the group undergoing fertilization, was opaque and slightly yellow in color. These cells varied in diameter from 0.4 to 0.7 mm. in diameter. (3) The ovary of a specimen killed twenty-one days after the birth of its brood of six contained a new brood of seven, the individuals of which were 5 mm. long. Seven transparent yellow eggs, varying in diameter from .6 to .9 mm., represented the next oldest group. Two were slightly opaque, indicating that they were slightly younger than the others Fifteen cells, the largest of which was about .25 mm. and the smallest less than .1 mm. in diameter, were white and opaque or had transparent centers and opaque white layers at the periphery. The development of cells in the ovary up to the time that they be- 152 C. L. TURNER come mature and ready for fertilization mav he summarized as follows : 1. A large number of transparent cells less than 0.1 mm. in diameter are ready to he drawn upon for differentiation. 2. During the first month some of these cells begin to differentiate and develop a white opaque layer at the periphery. A number of the cells reach a diameter of about 0.5 mm. and become white and opaque throughout. At the same time an older group of eggs and also a de- veloping brood is pre>ent in the ovary. 3. During the second month the oldest members of this group of developing cells grow rapidly, become orange-yellow and translucent and reach a diameter of from 0.8 to 1.4 mm. by the end of the month. The number of cells undergoing this development is approximately equal to the number of individuals in a brood. 4. During the third month these orange-yellow translucent eggs com- prise the oldest group in the ovary, the embryos of the previous brood having been born. During the first ten days of the month they grow to a maximal size of 2 mm., some being as small as 1.4 mm. in diameter, and at the end of this ten-day period all are fertilized. Variations in the >tages of the young embryos indicate that fertilization does not take place in all simultaneously. Fertilization of all members of the group is accomplished within two days. During the remaining twenty days of the month the embryos develop and at the conclusion of the month they are voided. In the meantime younger waves of cells are passing through stages of development already described. A comparison between the Lcl'istcs and the Gitnibusia types indi- cates: (1) that, at the time of the birth of a brood of embryos, the ovocytes of the next younger group is much larger and better differ- entiated in Lcbistcs than in (,'ainbnsia ; (2) that the period of time elapsing between the birth of a group of embryos and the fertilization ot the next group of ova is shorter in the Lcbistcs type than in the Gambusia type. I );ita obtained in Xiphophorus and Platypcecilus were strikingly similar to those obtained in Lcbistcs. The adults of Platypcecilus and Xiphophorus are larger than those of Lcbistcs but ova of the same age are equal in size. However, in Lcbistcs fewer eggs come to maturity at one time and the broods contain fewer individuals. Other species whirh have cycles like that of Lcbistes or intermediate between the l.i-bistcs and the Gambusia type are Pcecilia vivipara, Micropcecilia picta, Micropcecilia parcc, Brachyrhaphis cpiscopi and Liinia rittutn. Others not yet studied will undoubtedly be added to the list. Quint ana atrizona Type. — In this type there is a very short interval between the birth of a brood and tin- fertilization of the ova of the fol- RKI'RODUCTIYK CYCLES IN PCECILHD FISHES 153 lowing brood. In a female killed and examined twenty hours after the birth of a brood, the next group of cells in the ovary was found to be fully developed. These ova, approximately equal in number to the brood just born, were 1 .H to 2.2 mm. in diameter. Some had apparently been fertilized a few hours before. Of this group — nineteen in all- seven had reached an advanced segmentation stage, nine were in earlier segmentation stages and three had not yet been fertilized. It was ap- parent that fertilization had taken place in most of the ova immediately after the evacuation of the previous brood and that fertilization in the remaining three cells would have been complete within a few hours more. Differentiation and growth of the ova next in line for fertiliza- tion had reached a final stage by the time the embryos of the previous brood were born and the interval between the two events was reduced almost to the vanishing point. TABLE II Illustrating production of broods in Pwcilistes pleurospilus No. of young Interval in days Date in brood between broods November 20 2 December 16 2 26 January 8 2 23 February 1 2 23 February 21 3 20 March 15 3 22 April 3 7 19 April 25 7 22 May 10 8 16 May 24 9 14 June 5 9 12 If the ova should develop still more rapidly and reach their final stage of differentiation and growth before the older brood of embryos w-as evacuated a condition would exist in which it would be possible for fertilization to occur before the older brood had been evacuated and for two broods to exist in the ovary at the same time. Such a condition of superfetation does occur in the next species of poecilikls to be con- sidered. Reproductive Cycles Involving Superfcetation Poscilistcs plcurospihis and P&ciliopsis in fans. — Specimens of Pceci- listcs pleurospilus were observed from November to May in order to secure records of the broods produced. Single specimens were then sacrificed at different intervals after the production of a brood to de- termine the state of development of embryos and ova within the ovary. The record of broods produced is shown in Table II. The record 154 C. L. TURNER shows that: (1 ) the interval between broods was longest in the winter months and that it was becoming shorter in the early spring months; (2) the broods were becoming larger as the spring advanced; (3) in general, also, the broods were smaller than those of poeciliid fishes in which superfetation does not occur: (4) during the spring months the interval between broods is much shorter than in Lcbistes. Stove (1935) finds the interval between broods in this species to be ten to twelve days. This is apparentlv the record of specimens repro- ducing in May or June. A female of Pcccilistcs pleurospilus, killed two days after the last brood of eight young had been born, contained two broods of unborn embryos in the ovary. The older brood consisted of nine embryos which were well developrd but far from the stage in which they would be born. The eyes were well developed and pigmcnted. and pigment cells were present over the brain, down the mid-dorsal line and along the lateral line. All embryos were in approximately the same stage of development. The second brood contained nine very young embryos, with only a few somites developed. A group of twelve cells about .5 mm. in diameter, together with many smaller undifferentiated cells, were also present in the ovary. Another female, sacrificed twelve days after she had produced a brood, contained an ovary in which two broods were well separated in age and also a group of developing ova nearly 2 mm. in diameter. An abundance of preserved material of Pcsciliopsis infans was avail- able and forty-two ovaries were dissected. In every ovary two broods of embryos were found, together with a group of cells approximately equal in number to the embryos in a brood, well in advance of the de- velopment of a mass of smaller undifferemiated cells. Embryos of a single brood varied somewhat more in their degree of development than was the case in Pircilistcs pleurospilus but the least advanced member of the older brood was always separated by a considerable gap from the most advanced member of the younger brood. The condition found by Jlenn (1916) in Pseudopcecilia friu and Diphyacantha cliacocnsis indicates that the ovarian cycle of these forms is probably like that of Pwciliopsis infans and Pcecilistcs pleurospilus. Henn dissected five ovaries and found in each a brood of large embryos, a second brood of small embryos with about the- same number of indi- viduals as in the first group and a small number ot developing ova. In the first two of these species, at least, the sequence of events for development of ova and embryos would be as follows: (1 ) a group ot ova grow and differentiate until they reach a diameter of about 2 mm. At this point an older brood of embryos, A, is at the point of being REPRODUCTIVE CYCLES IN PCECILHD FISHES 155 born and a much younger brood, />', is also present in the ovary; C2) the ova are fertilized, forming brood C", and the- older brood, A, is born at practically the same time; (3) broods B and C develop for a period of 10 to 22 days and during this time another group of younger ova develop to the maximal size; (4) brood B is born, brood C is retained in the ovary and the eggs of the new group are fertilized to form brood D. Superfetation, as observed in the species mentioned above, is ap- parently caused by the earlier development of the younger ova, al- lowing earlier fertilization. A sequence might be arranged from the species described so far, illustrating regular stages through which super- fetation might arise. (1) In Gambusia and Mollicuisia the ova re- main quite small until the birth of the brood already contained in the ovary ; they may not reach a state in which they can be fertilized for about two weeks. (2) In Lcbistes the developing ova are much larger and better developed when the brood in the ovary is born and this con- tinued growth brings them to the point of fertilization a few days earlier. (3) In Quintana atrizona the ova have reached a still more advanced stage at the birth of the brood in the ovary ; they have reached their maximal size and complete differentiation and fertilization takes place immediately. (4) The ova in Posciliopsis in fans and Pcecilistes pleurospilus attain their full size and complete differentiation long be- fore the brood in the ovary is born. Fertilization of the ova takes place and the resulting embryos are retained in the ovary as a second brood. Heterandria fonnosa. — This species has small broods with short intervals between the broods. The records shown in Table III are those of two young females that have just attained sexual maturity. The most important features in the records are: (1) the broods are very small, one or two embryos being produced at a time during the winter months and as many as five or six at one time during May and June; (2) the interval between broods is long during the winter months, thirty-three to forty-one days. It be- comes shorter with the approach of spring, reaching the shortest interval —three to eight days — in May and June. Seal (1911) made a record of broods produced by larger specimens during the months of July and August. He found that the broods containing from four to sixteen embryos were born at intervals of from four to nine days. A female which had produced a brood of nine was sacrificed and examined on May thirteenth. The ovary was well filled with embryos and when these were dissected out of their follicles and arranged in groups it was found that six levels of development could be distin- guished. There were nine embryos at the oldest stage (brood A} and these would presumably have been born about seven days later. 156 C. L. TURNER external parts of the embryos were well developed except the caudal fin, and the yolk sac was almost completely absorbed. They were about 5 nun. in length (snout to base of caudal fin). Might embryos formed TABLE III Brood production in Heterandria formosa specimen No. 1 No. of young Interval in clays Date in brood between broods December 12 1 January 24 February 9 2 16 March l". 1(> March 17. .. 16 March 29 12 April 6 2 8 April 14 > 8 April 26 2 U April 30 3 5 May 4 2 4 May 12 5 8 May 20 3 May 26 2 <> June 2 3 JuneS 6 June 14 i 6 June 17 5 June 20 4 June 28 * 8 July 1 1 July 11 2 10 July 25 1 14 Specimen No. 2 No. of youiiK Interval in days Dad in brood between broods December 10 2 January 21 2 41 February 7 2 17 February 28 2 21 March 6 1 6 March 17 3 11 April 4 2 18 April 7 2 3 April 20 2 13 April 26 3 6 April 30 1 4 May 9 3 9 May 12 2 3 May 20 2 8 May 26 6 6 May 31 4 5 brood /.'. The members of this brood were considerably younger than those forming brood .-/. There were eight nnhryos in brood C, nine in brood I), eight in brood II and at least six in brood /•'. It is possible that REPRODUCTIVE CYCLES IN PCECILIID FISHES 157 a still younger brood was present. The largest unfertilized cells were about .7 mm. in diameter and it is evident that fertilization would occur when the cells had reached this size. The only peculiarity in the history of the ova up to the time of fer- tilization lies in the fact that they are very much smaller at fertilization than the ova of the other species described. This fact would indicate that the larger quantity of yolk elaborated in the ova of the others is not required in Heterandria fonnosa. If the larger amount of yolk is not .formed and yet the embryos go on to the more advanced stage while being retained in the ovary, it must be assumed that the follicle cells surrounding the developing embryos are responsible for furnishing the food materials. The expanded yolk sac, containing little yolk, is in a position to absorb the food materials, for it completely surrounds the embryos in the earlier stages, becomes very vascular, and is in contact with the walls of the follicle almost up to the time of birth. Both viviparity and superfetation in this species have been carried to an extreme degree and it is quite likely that there will be found in some small pcecili'id fish a stage intermediate between that of Heterandria fonnosa and that represented by Pceciliopsis infans and Pcecilistcs plcitrospilus. It is reported by Stoye (1935) that the mode of reproduction in Priapella bonita and in Phalloptychus jannarins is similar to that in Heterandria fonnosa, one or two young being born at a time at intervals of a few days. It is also reported by Stoye that in Micropcecilia branneri the young are born one or two at a time every few days. This observation would tentatively class this species with Heterandria fonnosa in its mode of reproduction. A most interesting situation is presented in this case since it has been stated by Stoye and verified by the writer that two other species in this genus (M. pares and M. plcta) produce broods of thirty to forty embryos at intervals essentially like those of the Lebistes type. Unfortunately no breeding specimens of Micropcecilia branneri have been available and gravid ovaries have not been examined. SUMMARY A graphic summary of the relations of growing ovocytes, mature ova, the fertilization process, the broods of embryos retained in the ovary and the time of birth is offered in Table IV. Each group represents the situation within a single ovary at the height of the reproductive season and at the moment in which a brood is born. The term brood is used to indicate a group of growing and differentiating ovocytes of approxi- mately equal development up to the time of fertilization and also the 158 C. L. TURNKR embryos produced l>y the Fertilization of these ova up to the time for liirth. Xo quantitative relations are implied concerning comparative .size of ovocytes or embryos in different species nor of comparative lengths of time that embryos are retained in the ovary. A study of Table IV leaves one with the impression that the extreme type of superfetation in Heterandria formosa might have arisen from the simpler type as illustrated in Poccilistcs plewrospttus. The argument also appears to be defensible that the simpler type of superfetation found in Poccilistcs might have arisen from a condition like that illus- trated by Quint ana atrhona in which a new group of ova are read}' for TABLE IV {Brood Gembusia 1 Brood Lebistes (Brood Xiphophorus 4 Brood Platypoecilus JBrood ("Brood Quintana 4 Brood atrizona (Brood Poecilistes fBrood pleurospilus 4 Brood Poeciliopsis [Brood infans Brood Brood Brood Heterandria < Brood forraosa Brood R vr»r»n -} Growing ^ Fertilization — ^Retained — Ovocytes Embryos A -> Birth 4 B-. . ^ 7 A \ f . .. •» A R \ 7 P \ A (Supecfoetation) 7 \ 13 "^ f ^ fSuoerfoetetLoriJ x. \> J D.% 7 F ^ r, . k, fertilization when a brood is burn but in which fertilization is delayed for some hours. All of this implies that the extreme cases of super- fetation should be considered as the specialized and derived types while those without superfetatinn are the simpler. Since the PcecilikUe un- doubtedly arose from an oviparous cyprinodonl such an assumption does not seem to be overdrawn. That group of the P parallel to the seasonal waves produced in such forms as Coitus and 1'crca. It takes two years to produce a batch of mature ovocytes in /V;v equivalent to the long cycle of Coitus and 1'crca instead of a long-drawn-out annual cycle, as suggested by Harney and Anson and Kuntz. rests upon the question of whether there is synapsis in re- sidual gonial cells on a large scale to produce the entire season's potential ova or synapsis of only enough residual gonia to form the ova for a single brood. If synapsis occurred on the necessary large scale and only once a year, followed by the maturing of small groups of ovocytes, an annual cycle such as occurs in Cot/us and Pcrca might be postulated. However, during the height of the breeding season small waves of synapsis of gonia occur at regular intervals and new ovocytes are being added at the lower levels as the ones at the upper levels reach maturity and are fertilized. Sl'M MARY 1. All species oi the Pceciliidse are ovo-viviparous. At the height of the breeding season broods are produced at regu- lar intervals varying from about forty-live days in some species to five or .six days in others. RKPRODUCTIVE CYCLES IN PCECILIID FISHES 163 3. In laboratory-reared specimens and in specimens breeding in temperate zones the shortest interval bet ween broods occurs in spring and early summer and the longest interval during fall and winter. 4. In the reproductive cycle of Mollienisia and some other species, the oldest group of ovocytes remaining in the ovary just after a brood has been extruded is very small. Growth is rapid and the ova are ready for fertilization about eight days after the voiding of the last brood of embryos. 5. In Lcbistcs and other species with similar reproductive cycles, the oldest group of ovocytes remaining in the ovary just after a brood has been voided are much larger than in Ganibitsia and fertilization occurs a few days earlier than in Gainbusia. 6. In Quintana atrisona the oldest group of ova remaining in the ovary are fully developed and ready for fertilization at the time of birth of the previous brood of embryos. 7. Superfetation occurs in Pcecilistes pleurospilus and several other species. Fertilization of the oldest group of ova takes place before the voiding of the brood already in the ovary and two broods at different levels of development are found in the ovary. 8. In Heterandrla formosa and possibly other species there is an extreme development of superfetation with six or more small broods at different levels of development occurring in the ovary at one time at the height of the breeding season. 9. Older groups of ova and embryos retained in the ovary retard the development of younger groups of cells and a balance is maintained be- tween this retarding influence and the agent, assumed to be a follicle- stimulating hormone from the pituitary, which forces the cells to grow from the lower to the upper levels of development. 10. The short reproductive cycles of the Pceciliidse are the equivalent of the longer annual or biennial reproductive cycles of the fishes of the temperate zones. LITERATURE CITED BAILEY, R. J., 1933. The ovarian cycle in the viviparous teleost, Xiphophorus helleri. Biol. Bull., 64: 206. " BARNEY, R. L. AND B. J. ANSON, 1921a. Seasonal abundance of the mosquito de- stroying top-minnow, Gambusia affinis, especially in relation to male fre- quency. Ecology, 2: 53. BARNEY, R. L. AND B. J. ANSON, 1921?;. The seasonal abundance of the mosquito- destroying top-minnow, Gainbusia affinis, especially in relation to fecundity. Anat. Rcc., 22: 317. EIGENMANX, CARL H., 1906. The fresh -water fishes of South and Middle Amer- ica. Pop. Sci. Monthly, 68: 515. HANN, H. W., 1927. The history of the germ cells of Cottus bairdii Girard. Jour. Morph. and Physiol., 43: 427. ftp* 164 C. L. TURNER HENN, ARTHUR \\"., ](>\(>. IX. On various South American Pceciliid Fishes. Annals Carnegie Museum, 10: 93. HILDEBRAND. S. F., 1917. Notes on the life history of the minnows Gambusia affinis and Cyprinodon varicgatus. Kept. U. S. Comm. Fish, for 1917, App. VI : 3. HILDEBRAND, S. F., 1921. Top minnows in relation to malaria control, with notes on their habits and distribution. Pub. Health Bull., No. 114, U. S. Pub. Health Service. HILDEBRAND, S. F., 1925. A study of the top minnow, Gambusia holbrooki, in its relation to mosquito control. Pub. Health Bull., No. 153, U. S. Pub. Health Service. HUBBS, C. L., 1921. An ecological study of the life-history of the fresh-water atherine fish, Labidesthes sicculus. Ecology, 2: 262. KUNTZ, ALBERT, 1913. Notes on the habits, morphology of the reproductive organs, and embryology of the viviparous fish Gambusia affinis. Bull. U. S. Bur. Fish., 33:" 181. Pmuri'i, KKICH, 1'KIS. Fortplanzungsgeschichte dcr viviparen Teleosteer Glari- dichthys januarius und G. decem-maculatus in ihrem Einfluss auf Lebens- weise, makroskopische und mikroskopische Anatomic. Zool. Jahr. Abt. f. A>iat., 27: 1. SEAL, W. P., 1911. I'.reeding habits of the viviparous fishes Gambusia holbrookii and Hetcrandria formosa. Proc. Biol. Soc. Washington, 24: 91. SEAI i . AI.VIN, 1917. The mosquito fish, Gambusia affinis, in the Philippine Islands. Philippine Jour. Science, 12: 3. STOYE, F. H., 1935. Tropical Fishes for the Home. Sayville, New York. VAN OORDT, G. J., 1'^-S. Tin- duration of the life of the spermatozoa in the ferti- lized female of Xiphophorus hclleri Regan. Tijdschr. d. Ned. Dicrk. V ere en., 1:1. Xm.oTMSKY. X., 1901. Les moeurs du Girardinus decemmaculatus, poisson vivi- pare. Arch. Zool. E.rper. et Gen., 3: 9. THE EXPERIMENTAL DECOMPOSITK ).\: AND REGENERATION OF NITROGENOUS ORGANIC MATTER IN SEA WATER x THKODOR VON BRAND, NORRIS W. RAKESTRAW AND CHARLES E. RENN (l-rovi the Woods Hole Occanographic Institution, Woods Hole, Mass.) HISTORICAL Although the formation of nitrate from composted nitrogenous or- ganic materials has been long known and practically applied, the current conception that specific biological agencies bring about these natural processes in soils, fresh waters, and the sea was developed only in the last quarter of the nineteenth century. In his studies on acetic acid fermentation Pasteur (1862) anticipated the role of bacteria in the ulti- mate oxidation of ammonia to nitrate, but Schloesing and Miintz (1877 et seq.) first demonstrated the biological nature of the intermediaries. The demonstration was performed in soils ; Miintz, shortly afterward (1890), showed that ammonia was produced in soils by microbial break- down of proteinaceous organic matter. That the oxidation of ammonia to nitrate proceeded through nitrite was distinguished by Munro (1886), but not until Winogradsky (1891, 1892) performed his classic re- searches were the two oxidations recognized as separately determined by specific ammonia-oxidizing and nitrite-oxidizing bacteria. Following these researches Adeney (1895) demonstrated the com- plete sequence of nitrate regeneration in natural and sewage polluted water, establishing the steps by clear-cut quantitative methods. His procedures determined the basis for standard sewage analysis. It has been held by Brandt (1899, 1902) and many others that the cycle of nitrogen in the sea follows essentially the same phases, and, using this interpretation, a number of researches into the occurrence and distribution of specific marine organisms have been made. These in- vestigations have been reviewed and extended by Waksman, Hotchkiss and Carey (1933). There is considerable conflict in the literature on biological nitrification in the sea, but the conventional picture derived from studies of soils and fresh waters is nevertheless applied to oceano- graphic chemical data. 1 Contribution No. 129 of the Woods Hole Oceanographic Institution. 165 166 VON BRAND, RAKESTRAW AND RENN EXPERIMENTAL Because of the growing interest in the regeneration of nitrate in the sea it seemed highly desirable to make a quantitative investigation of the course of organic decomposition and of the simultaneous appearance of nitrogenous products in the \vater. To this end, a number of experi- ments were carried out in which a natural source of organic matter— plankton material from customary net-hauls — was allowed to rot and TABLE I Series I. Source of organic matter: mixed plankton tow. Micrograms of nitrogen per liter. Date Plank- ton In water Total in system J Dissolved organic N Bacteria in thousands per ml. Am- monia Nitrite Nitrate nitrite Nitrate Total in water 6-27 188 0 0 5 5 5 193 150 100 29 150 160 30 80 0 5 5 85 204 7-2 131 107 0.1 5 5 112 243 180 5 110 160 3.8 6 2 166 276 123 10 82 277 11 185 10 195 22 15 2.4 17 77 308 40 18 225 2.7 6 3 231 24 40 16 272 25 200 32 232 20 27 55 70 15 28 185 80 90 10 275 30 125 110 145 35 270 36 8- 1 60 180 215 35 275 5 11 150 275 125 286 7 .... 47 0 110 300 190 300 347 10 8 2.2 300 298 308 811*... Portion inoculated \\nli diatoms 1 9 .... 285 0 0 20 20 20 305 22 300 0 13 13 13 313 8-11*.. Control portion, not inoculated 12 0.5 <(,o 360 14 V1<> 17 0.6 330 330 19 37 0 360 360 360 397 200 21 0.8 330 330 22 0 * One portion inoculated with diatoms and placed in the light. + Does not include dissolved organic nitrogen. ORGANIC DECOMPOSITION AND REGENERATION 167 decompose in sea water. Observations were made to follow the dis- appearance of nitrogen in the decomposing material and its appearance in the water in the form of ammonia, nitrite and nitrate, in the effort to reproduce artificially the cycle of organic decomposition and eventual regeneration. If, by means of the soluble compounds resulting from the decomposition, the water could be rendered fertile and capable of TABLE II Series II. Source of organic matter: mixed plankton tow. Micrograms of nitrogen per liter. In water Bacteria Date Plank- Total in Dissolved in ton Am- Nitrite Nitrate + Nitrate Total in system J organic N thousands per ml. inonia nitrite water 7-12 400 18 0.2 8 8 26 426 420 82 14 345 176 15 >120 0.5 8 8 265 17 275 >120 0.2 6 6 270 20 202 550 2.3 8 6 558 760 80 23 580 24 16 43 25 600 27 627 27 60 80 20 28 650 110 150 40 800 25 29 90 30 550 240 300 60 850 940 8- 1 340 550 590 40 930 16 5 12 800 820 7 78 4 850 60 914 992 8 910 10 8 550 1000 450 1000 20 12 190 1100 910 1100 14 35 0.1 1050 1050 1100 25 17 0.1 1100 1100 18 133 (1250) 8-21*.... Portion inoculated with diatoms 27 305 28 1.0 31 840 7.7 320 312 (330) (1170) 9- 2 1010 0 25 25 (35) (1045) 3 470 8-21*.... Control portion, not inoculated 22 7 0.5 28 0.3 1050 1050 29 1000 31 0.5 * One portion inoculated with diatoms and placed in the light. t Does not include dissolved organic nitrogen. 168 VOX BRAND, RAKESTRAW AND RENN supporting a new organic growth, comparable to that previously decom- posed, the complete cycle would have been carried out. It was also hoped that periodic chemical analysis, by methods now available, might throw light on the sequence of the various steps in the cycle. The raw material was mixed plankton, collected in the usual way in a Xo. 20 net, quickly washed, and kept on ice for the few hours before it could be returned to the laboratory. This material was then sus- pended in 10-15 liters of fresh, filtered sea water and stored in the dark at a temperature of 20-25° C. Several different analyses were carried out on this material at the start and at intervals thereafter, the various fractions being determined as follows : 1. Total nitrogen contained in the suspended or participate matter. This is called " plankton nitrogen " in the following descriptions, but also includes bacterial nitrogen as well as that in any other form of suspended matter. This determination follows the procedure described by von Brand (1935) and consists essentially of precipitating the TABLE III Series IV. Source of organic matter: mixed plankton tow, strained through No. 8 bolting silk. Micrograms of nitrogen per liter. In water Bacteria Date Plank- ton Am- monia Nitrii*- Nitf.ltr + nitrite Nitrate Total in water * Total in system * in thou- sands per ml. 7-28 115 43 55 170 152 29 0.6 12 12 ?00 30. .. 83 110 305 8- 1 . 63 150 1.0 15 14 165 228 190 5 38 205 <1 10 10 215 253 35 7 10 40 190 200 0.6 10 15 10 15 200 215 255 16 14 240 0.5 20 20 25 17. . 41 17 17 20 13 21 35 '() 22. . 210 21 (250) 18 27 58 175 120 1 10 20 315 373 29 190 31 29 220 240 20 269 9- 2 25 225 225 (0) 250 5 59 8 260 14 14 310 330 20 344 (400) 27. .. 130 290 160 10-4. . 1 * Does not include dissolved organic nitrogen. ORGANIC DECOMPOSITION AND REGENERATION 169 suspended matter in a sample of the water by the addition of alkali. The precipitate drags down all suspended and colloidal matter, and after settling and centrifugation is separated and resuspended. The resulting small volume of suspension is then used for a determination of total nitrogen by the method of Krogh and Keys (1934) : fusion with KOH in a stream of hydrogen, followed by recovery of the ammonia. 2. Ammonia in the water, by a slight modification of the method of Krogh (1934), in which the original design of the still was changed somewhat, to make it more compact and to permit heating electrically. 3. Nitrite in the water by the well-known Griess-Ilosvay method. 4. Nitrate (including nitrite) in the water, by Harvey's reduced TABLE IV Series VI. Source of organic matter: washed, persisting culture of Nitzchia. Micrograms of nitrogen per liter. In water Date Plankton Ammonia Nitrite Nitrate + nitrite Nitrate Total in water * Total in system * 8-14 0.1 15 633 2 17 17 19 652 18 514 120 0.1 15 15 165 679 21 430 230 0.1 16 16 246 676 27 298 310 0.4 24 24 334 632 9-2 201 410 1.3 20 19 430 631 8 226 25 25 14 490 0.1 17 17 507 (700) 27 0.2 20 20 * Does not include dissolved organic nitrogen. strychnine method. Nitrate then calculated by subtracting the value of nitrite as obtained under 3. 5. Bacterial counts. The data from the chemical analyses are collected in Tables I to IV and some of them are also presented graphically in Figs. 1 and 2. Several different " series " were carried out: I, II and IV, in which the organic material was a mixed tow of both phyto- and zooplankton, of varying amount in each case; and VI, in which a persisting washed culture of diatoms was used as the source of organic matter. In Series II the sea water in which the plankton was suspended came from the surface of Woods Hole Harbor, which probably accounts for its higher content of dissolved organic nitrogen than that of the water in 170 VON BRAND, RAKESTRAW AND RENN the other series, which was obtained from the surface a fe\v miles to the south of Marthas Vineyard. CHEMICAL The rate of deconi] motion of the plankton varied in the different series, but after an interval of from 8 to 20 days decomposition stopped. At this time there was still a constant and not inconsiderable amount of nitrogen still remaining, cither in non-decomposable plankton residues, bacterial cells, or in other forms. The amount of this residual nitrogen was from 20 to 35 per cent of that originally determined in the plankton, or 7 to 10 per cent of that calculated by addition of the different soluble and insoluble nitrogen fractions. The rate of decomposition is greatest in the first few days. In Series II, IV and VI it dropped rather suddenly, but in Series I it fell gradually over a longer period of time. 5 DAYS" FIG. 1. Series I. The decomposition of nitrogenous organic matter in mixed plankton, showing the appearance of soluble nitrogen compounds in the water in which it is suspended. One portion inoculated with diatoms on the forty-fifth day. Ammonia appeared in the water rapidly and immediately at the be- ginning of the decomposition. Evidently the liberation of ammonia is involved in the very first steps in the degeneration of organic matter. It appears more slowly as time goes on, and reaches a maximum when the plankton decomposition ends. During this time there is no rise in the nitrite or nitrate in the water. The disappearance of ammonia is accompanied by the appearance of nitrite in the water, and this in turn is eventually oxidized to nitrate. The phase during which nitrite was present was 15-20 days in Series I and II, but extended to 40 days in Series IV. In any event, it is not until nitrite reaches its maximum and starts to disappear that nitrate increases. ORGANIC DECOMPOSITION AND REGENERATION 171 Eventually, however (after 45 days in Series I, 40 days in IT, and 65 days in IV), all the available nitrogen was oxidized to nitrate. At this point, in both Series I and IF, a portion of the water was placed in the light and inoculated with fresh diatoms, which grew rapidly, raising the " plankton nitrogen " again and lowering the nitrate to its original minimal value. In this way the complete cycle was carried out. The question of the existence of soluble nitrogen compounds inter- mediate between plankton material and ammonia could not be definitely answered, but there was no evidence of any such. Anything in the nature of colloidal cell-material, partially decomposed, must have been carried down in the precipitation and determined with the rest of the plankton nitrogen. In the transformation of ammonia to nitrite and nitrate it seems that we are dealing not with an equilibrium of simultaneous processes but with distinct, consecutive steps, each doubtless determined by its own set of conditions. £400- 40 45 50 55 60 65 FIG. 2. Series IV. The decomposition of nitrogenous organic matter in mixed plankton, showing the appearance of soluble nitrogen compounds in the water in which it is suspended. Plankton previously filtered through No. 8 bolting silk. A peculiar situation is revealed by the data from Series I, II and IV. The analyses indicate a larger amount of nitrogen regenerated in soluble forms in the water than that lost by the plankton during decomposition. In the tables the " total nitrogen in the system " has been calculated by addition of all the separate forms determined. This total shows in general a continuous increase throughout the decomposition, especially in Series II. That it is not observed in Series VI may be in some way connected with the fact that the organic material in this case consisted of diatoms. There are three possible explanations for this nitrogen discrepancy : 1. Participation of other forms of nitrogen in the cycle — most plausibly the dissolved organic nitrogen in the water. 172 VOX BRAND, RAKESTRAXV AND RENN 2. Systematic errors in cither tlic sampling procedure or the analyti- cal method, so that the values for plankton nitrogen are regularly too low. 3. Nitrogen fixation. Kach of these possibilities has been critically examined and at the present time none of them seems an altogether acceptable explanation. Determinations of the dissolved organic nitrogen in the water, by the method of Krogh and Keys, did not show any significant change during the experiments. Filtering out the larger plankton organisms in Series IV, to lessen the chance of irregular sampling, had no effect. On the other hand, it is true that the decomposition of very much more finely divided organic matter (diatoms) in Series VI yielded only an equiva- lent amount of ammonia, but this experiment had to be stopped before nitrification began, which is usually the time when the most pronounced difference between added and recovered nitrogen appears. According to currently accepted views of biological fixation of nitro- gen the third possibility seems doubtful. A further study of this nitrogen discrepancy is now under way. BACTERIOLOGICAL The regeneration of inorganic nitrogen from its various organic forms in plankton proceeds through four recognizable stages, each of which is incident to the development and activity of a special bacterial flora. Following the death of plankton organisms, brought about in these experiments by handling, unfavorable temperatures, lack of illu- mination, crowding, and inadequate nutrient supply, the more labile components of the dead cells undergo hydrolytic cleavage, simultaneously liberating a fraction of the amino nitrogen as ammonia. This break- down, it would appear from the rate at which ammonia appears and plankton nitrogen diminishes, is not analogous to the digestion of pro- tein by higher animals where individual protein complexes are first split into smaller molecules bearing amino nitrogen. A part of the decomposition may be autolytic, but it is evident from the tables, which show an attending rapid development-el various species of bacteria, that these organisms must play a large part. The absolute numbers are never very high — in fact, not greater than would be ex- pected in freshly taken, unaltered sea water stored for such intervals of time (Waksman and Renn, 1936; ZoHell and Anderson, 1936). The decline in plankton nitrogen is not so precipitous as the decrease in bacterial numbers. This may mean that the various fractions of plank- ton nitrogen arc not equally susceptible to bacterial attack. ORGANIC DECOMPOSITION AND REGENERATION 173 Liberation of ammonia from plankton and other nitrogenous organic matter is, bacteriologically speaking, uns|K'dali/<'(l. but is dependent upon the relative proportions of available nitrogen and carbon in the decom- posing materials — the materials used in these experiments are, as Red- field (1934) has shown, characteristically nitrogen-rich. Following the first burst of bacterial activity the ammonia concentra- tion reaches a high level capable of supporting an active nitrite- forming population. Unsuccessful attempts were made to follow the develop- ment of the specific bacterial flora responsible for the oxidation of am- monia to nitrite. It appears from the many investigations upon nitrite- forming bacteria in the sea, that the process is due to specific organisms of the Nitrosoinonas group analogous to the specialized nitrite-forming flora of soils (Waksman, Hotchkiss and Carey, 1933; Carey and Waks- man, 1934). The rate at which nitrites are developed here is roughly the same as that observed in cultural studies of the process/' Crude cul- tures prepared by inoculating sea water containing ammonium" salts with mud or sea water usually give positive tests for nitrite within ten days or two weeks (depending upon the quantity of inoculum), after which the complete oxidation of ammonia to nitrite proceeds rapidly. The final stage in nitrogen regeneration progresses more slowly, as is evident from Figs. 1 and 2. Nitrate formation is also effected by a specific bacterial flora, in this case, Nitrobacter, an efficient population of which seems to develop at a lower rate than either the non-specific am- monifying or specialized nitrite-forming bacteria. In culture experi- ments where crude inocula of nitrate-forming bacteria are added to sea water containing nitrite as the only sources of nitrogen, the period of incubation necessary to demonstrate measurable quantities of nitrate ranges from 30 to 90 days. Richer inocula bring about more rapid nitrate formation (Waksman, Hotchkiss and Carey, 1933). It has been assumed that the relatively late appearance of nitrate in cultures made up of raw sea water to which organic nitrogen in some form has been added is due to the toxic effect of ammonia on nitrate- forming bacteria — a condition that has been clearly demonstrated in ex- periments on soils. But this relation, if it exists in these experiments, is masked by the slow growth of nitrate- forming bacteria even under favorable conditions. Thus in Series I and II the phase of nitrate formation began promptly after the disappearance of ammonia; in Series IV, on the other hand, it did not begin until 15 days after the ammonia had disappeared. The regeneration experiments described here certainly are not identi- cal with natural conditions and may not be expected to parallel exactly the courses of the individual phases of nitrate formation in the sea. On 174 VOX BRAND, RAKESTRAW AND RENN the other hand, they offer a clear picture of their relations. Further, they validate the conventional conception of the nitrogen cycle as derived from the various studies of the individual processes, by bringing them together in a single summary experiment. SUMMARY In conclusion, this study has shown : 1. That it is possible to reproduce the complete cycle of nitrogen regeneration. 2. That the transformation of decomposing plankton (especially diatoms) into ammonia is very rapid, beginning as soon as the initial substance disappears from the body. The amounts of soluble nitroge- nous substances of higher molecular weight in these experiments can have been only very small. 3. The main stages in the decomposition arc : dead body — ammonia —nitrite — nitrate. 4. Under these conditions, at least, no toxic substances are formed which inhibit the flowering of diatoms. The rapid and abundant de- velopment of the latter showed the regeneration of the water to be complete. BIBLIOGRAPHY ADENEY, W. E., 1885. The course and nature of fermentative changes in natural and polluted waters, and in artificial solutions, as indicated by the com- position of the dissolved gases. Sci. Trans. Roy. Dublin Soc., 5 (2) : 539. BRANDT, K., 1899. Ueber den Stoffwechsel im Meere. JViss. Mecrcsnntcrs. Kid, N. F., 4: 213; 1902. Ibid., 6: 23. CAREY, C. L., AND S. A. WAKSMAX, 1934. The presence of nitrifying bacteria in deep seas. Science, 79: 349. COOPER, L. H. N., 1935. The rate of liberation of phosphate in sea water by the breakdown of plankton organisms. Jour. Mar. Biol. Ass., 20: 197. KROC.H, A., AND A. KEYS, 1934. Methods for the determination of dissolved or- ganic carbon and nitrogen in sea water. Biol. Bull., 67: 132. KROGH, A., 1934. A method for the determination of ammonia in water and air. Biol. Bull., 67: 126. MUNRO, J. M. H., 1886. The formation and destruction of nitrates and nitrites in artificial solutions and in river and well water. Jour. Client. Soc., 49: 632. MUNTZ, A., 1890. Sur la decomposition des engrais organiques dans le sol. Compt. Rend. Acad. Sci., 110: 1206. PASTEUR, L., 1862. £tudes sur les mycodermes. Role de ces plantes dans la fer- mentation acetique. Compt. Rend. Acad. Sci., 54: 265. REDFIELD, A. C., 1934. On the proportions of organic derivatives in sea water and their relation to the composition of plankton. James Johnston Me- morial Volume, p. 176. University of London. SCHLOESING, TH., AND A. MUNTZ, 1877 et seq. Sur la nitrification par les fer- ments organises. Compt. Rend. Acad. Sci., 84: 401; 85: 101S; 86: 892; 89: 891, 1074. ORGANIC DECOMPOSITION AND REGENERATION 175 VON BRAND, T., 1935. Methods for the determination of nitrogen and carbon in small amounts of plankton. Biol. Hull., 69: 221. WAKSMAN, S. A.. M. HOTCHKISS, AND C. L. CAREY, 1933. Marine bacteria and their role in the cycle of life in the sea. II. Bacteria concerned in the cycle of nitrogen in the sea. Biol. Bull., 65: 137. WAKSMAN, S. A., AND C. E. RENN, 1936. Decomposition of organic matter in sea water by bacteria. Biol. Bull., 70: 472. WINOGRADSKY, S., 1892. Contributions a la morphologic des organismes de la nitrification. Arch. Sci. Biol. St. rctcrsbonrg, 1: 87, 127. ZoBELL, C. E., AND D. Q. ANDERSON, 1936. Observations on the multiplication of bacteria in different volumes of stored sea water and the influence of oxygen tension and solid surfaces. Biol. Bull., 71: 324. STUDIES IN THE PIGMENTARY SYSTEM OF CRUSTACEA II. DIURNAL MOVKMKXTS OF THE RETINAL PIGMENTS OF BERMUDAN DECAPODS L. H. KLEINHOLZ (From the Bermuda Biological Station for Research, Inc., and the Biological Laboratories, Harvard University) * Diurnal rhythms in hehavior and physiological rhythms in certain organ systems of animals have long intrigued biologists and led to much interesting speculation as to the nature of the mechanisms involved in such activities. That the phenomenon is a general one is seen from the cases cited by Welsh (1930a). Such periodicity is most strikingly ex- hibiled in the pigmentary changes of various animals, being reported in shrimps by Gamble and Keeblc (1900). in brachyurans by Megusar (1912), in isopods by Menke (1911), in amphibians by Slome and Hogben (1929), in reptiles by Redfield (1918), and in cyclostomes by Young (1935). It is reasonable to expect that further study will re- veal the existence of such rhythms in additional forms. \Yelsh (1930o, 1935, 1936) has shown that diurnal changes in the migration of the retinal pigments occur in several crustaceans even though the animal- are maintained under constant conditions of illumination or of darkness. In all the cases mentioned hormonal factors are involved to a greater or lesser degree in controlling pigmentary activity. Since the present studies (Kleinholz, 1937) are concerned with chromatic responses in crustaceans, only this branch of the general subject of color changes will be discussed here. The control of the bodily changes in color in crustaceans has been shown by Perkins (1928) and by Koller (1928) to be maintained by an endocrine which is liberated into the circulatory system from within the eye-stalks. Hanstrom (1935) and his as- sociates (Sjogren, 1934; Carlson, 1936) have made histological studies of the eyes of many crustaceans, and have correlated the activity of the chromatophorotropic principle in eye-stalk extracts with the presence within the eye-stalk of a secretory structure which they call the blood gland. The mechanism that is involved in the control of retinal pig- 1 The experimental work was carried on in Bermuda and was made possible by a grant from the James F. Porter Fund of Harvard University. I am indebted tn Dr. J. F. G. Wheeler, Director of the Bermuda Biological Station, for placing the facilities of the Station at my disposal. 176 DIURNAL MOVEMENTS OF THE RETINAL PIGMENTS 177 incut migration in crustaceans is not yet completely known. Recent studies on Palccinonc/cs vnhjaris (Kleinholz, 1936) present evidence for a humoral control of the distal and the reflecting pigments in the retina, hut the proximal pigment appears not to be affected by the same eye- stalk extracts which activate the first two sets of pigments. It is pos- sible that one set of pigment cells is under nervous and the others under hormonal control. Before much progress can be made in analyzing the mechanisms by which diurnal rhythms are maintained under constant external condi- tions, it is advisable to study the phenomenon in a number of different species to determine which combinations of retinal pigments may be in- volved in such activities. This report on diurnal changes in the retinal pigments of several Bermudan crustaceans is offered to this end. MATERIALS AND METHODS The various crustaceans used in these experiments were obtained by dredging at several stations in the vicinity of the Biological Station. After a sufficient number of animals of the same species had been brought into the laboratory, the specimens were divided into two groups, one of which was placed in a white porcelain bowl illuminated by a 40- watt electric lamp at a distance of 18 inches, while the second group was placed in a container in the dark-room. At least 12 hours w^ere allowed for individuals to become adapted to light and to darkness. After a period appropriate for adaptation, specimens were removed for fixation of the retinal pigments. For purposes of description the term " day-light " eye is used to indicate the retina of a specimen that was kept constantly illuminated and which was fixed or examined in the daytime, while " night-light " eye represents the condition of the pig- ments in an illuminated retina that was fixed at night; conversely, " day-dark " and " night-dark " are used to designate those specimens maintained in constant darkness whose retinal pigments were fixed dur- ing the day and at night, respectively. When a specimen was taken from the light-adapted group, a similar specimen was removed from the container in the dark-room for fixation at the same time. Two methods of fixation for histological study \vere used. In one method, the animals were dropped into hot water (80° C.) for 10-20 seconds to fix the positions of the retinal pigments, and were then tran- ferred either to 5 per cent formalin or to a modified Bouin's solution containing 7 per cent acetic acid. In the second method the entire specimen, after the usual period of adaptation to darkness or to light, was dropped into a vial of Bouin's solution. The exoskeleton of most of the crustaceans contained large amounts of calcareous salts, but these 178 L. H. KLEIN HOLZ troublesome deposits were dissolved by the acetic acid in the fixative, so that sectioning the eyes was greatly facilitated. When the exoskeleton had been sufficiently softened by the fixative, the eye-stalks were excised and embedded in paraffin by a rapid dioxane treatment. After a preliminary rinsing of the excised eyes in water, the stalks were rolled over Filter paper to remove any excess moisture DL DD Fir,. 1. Entire eye-stalks of Eusicyonia, viewed by transmitted light through the low powers of tlir mirrosmpr and showing tin- positions of UK- distal retina! pigment. DL, from a day-light animal ; DD, from a day-dark specimen; NL, from a night-light individual ; ND, from a night-dark shrimp. and were then placed in full-strength dioxane over anhydrous calcium chloride. They were left in this fluid usually overnight (8-10 hours) and tin- next morning were placed directly into soft paraffin. The tis- sues were allowed to become infiltrated with the wax for about three hours, using two changes of paraffin, and were then embedded. No difficulty was encountered in cutting serial sections at 10 micra. Some DIURNAL MOVEMENTS OK THE RETINAL PIGMENTS 179 sections were subsequently stained with Delafield's hematoxylin and eosin, and others were mounted unstained. I am indebted to Mr. M. D. Burkenroad <>f tin- Bingham Oceano- graphic Foundation at Vale University, and to Dr. Fenner Cbace, Jr., of the Museum of Comparative Zoology at Harvard University for identifications of the collected crustaceans. The two macrurans studied were Eusicyonia a. sp. (to be described by Mr. Burkenroad) and Trachypeneopsis mobilispinis, while the brachyurans were Portunus anceps, Portunus dcpressifrons, Parlhciiopc scrrata, and Calappa flammea. Eusicyonia n. s[>. This new species of shrimp was taken from dredgings at 10-15 fathoms off Murray's Anchorage. Of the 12 specimens obtained for experimental study, 2 were males and the rest females. "While preliminary observations of the eyes of specimens kept under constant illumination showed that there was a change in position of the distal retinal pigment at night (Fig. 1), sectioning the eye was neces- sary to disclose the photomechanical changes of the remaining pigments. The eye of Ensicvonia is similar in structure to that of Palcemonetes as described by Parker (1897) and by Welsh (1930&). The positions of the three types of pigment cells are shown in Figs. 2-5. An unusual situation was discovered in studying a number of sec- tioned eye-stalks of this species. Of the 12 specimens at hand, the retinas of 4 showed complete absence of reflecting pigment. The pos- sibility that this might be a secondary sexual difference is eliminated by the fact that 2 of these shrimps were males and 2 females. The pos- sibility that the difference is a fixation artifact can be ruled out by com- paring eyes of specimens fixed by the two different methods. The re- flecting pigment in the integument of many of the lower vertebrates is guanin, and is commonly found there in the crystalline condition (Ewald and Krukenberg, 1882). So far as I know, however, no studies have been made on the chemical nature of the pigments in the crustacean retina. If the retinal reflecting pigment is guanin, it might be slightly soluble in the acid components of the Benin's solution. But the pig- ment was lacking only from those eyes which had been fixed with hot water and then preserved in formalin. This treatment does not seem to be sufficiently harsh to effect such changes. More-over, of 4 speci- mens of Trachypeneopsis which had been fixed by the same method, the eyes of 1 lacked this reflecting pigment, while those of a second specimen were slightly deficient (Fig. 15) ; the retinas of the other animals showed what may be called " typical " amounts. It seems more likely that the differences in amount of reflecting pigment are due to 180 L. H. KLEINHOLZ individual variations, possibly in the nucleoprotein metabolism of the animals, than to any fixation artifact. On studying sections of the eyes of Eusicyonia fixed under the four experimental conditions (l''ii>. 2-5, 10-13) the following situation is found in the retinal pigments. The granules of reflecting pigment un- dergo no apparent chunks in position in response to changes in light intensitv. The main mass of the reflecting pigment is located on the distal face of the basement membrane ; smaller amounts are found be- low the basement membrane and capping the distal ends of the distal retinal cells. The distal and the proximal pigments show readily dis- cernible photomechanical changes: Day-light eye A. The distal pigment cells are to be found in the typical light-adapted state, the cells having moved proximally and come to rest against the proximal retinular cells. />. The proximal pigment is also in the position typical for the light-adapted retina, much of it having migrated above the basement membrane to surround the rhabdomc. 1 VMM. A NATION FOR Pl.ATK I All figures are camera-lucida outlines to insure correct proportions, but the details have been drawn diagrammatically. DM, basement membrane; DD, day- dark retina; DL, day-light retina; DP, distal retinal pigment; ND, night-dark eye; NL, night-light retina; PP, proximal retinal pigment; RP , reflecting retinal pigment ; R, rhabdome. FIG. 2. Ommatidium from a day-light eye of Eusicyonia. The distal pigment lies against the proximal retinular cells, while much of the proximal pigment ha> migrated above the basement membrane to surround the rhabdome. FIG. 3. Ommatidium from a day-dark retina of Eusicyonia. The pigments are in the typical dark-adapted position, the distal pigment having migrated distally towards the cornea, and the proximal pigment having moved completely below the basement membrane. FIG. 4. Ommatidium from a night-light eye of Eusicyonia. The distal pig- ment is in the position characteristic for the dark-adapted retina while the proximal pigment is found above the basement membrane. FIG. 5. Ommatidium from a night-dark eye. Both pigments are in the posi- tions expected for a dark-adapted retina. FIG. 6. Ommatidium from a flay-light eye of I'urtninis anccps. The distal pig- ment is dispersed proximally toward the basement membrane, while much of the proximal pigment has migrated distally above the basement membrane. FIG. 7. Ommatidium from a day-dark retina of Portnmis. The positions of the pigments are the same as in the day-light eye, in spite of the fact that the ani- mals were kept in constant darkness. FK.. X. Ommatidium from a night-light eye of Portitmts. The positions of the retinal pigments are typical for what is expected in an illuminated eye. Fi<;. {). Ommatidium from a night-dark eye of the same species. The distal pigment has migrated distally, while the proximal pigment has moved proximally tin- h;iM-m<-iit membrane. --PP -RP •--BM -R H •BUZ^T ~ BM S--BM !--BM r-PP -DP •-BM -BM PLATE I 181 182 L. H. KLEINHOLZ Day-dark eye A. The distal pigment is in the position characteristic for the dark-adapted retina, forming a collar around the distal ends of the cones. Ii. The proximal pigment has moved entirely below the basement membrane. Night-light eye ./. The distal pigment, in spite of the fact that the retina has been under constant illumination, is in the position typical for the dark-adapted eye, at the distal ends of the cones. D. The proximal pigment, however, is in the position found in the usual light-adapted eye, much of it having moved above the basement mem- brane. Xight-dark eye A. The distal pigment is in the position characteristic for a dark-adapted retina. B. The proximal pigment is also in the position found in the dark-adapted eye, having migrated completely below the basement membrane. EXPLANATION OF PLATE II Figs. 10-13 represent retinas of Eusicyonia in the four experimental condi- tions; Figs. 14-17 are photographs of similar retinas of Trachypeneopsis inobili- s finis; Figs. 18-21 are photographs of the entire stalks of Portunus anccps. All the eyes are oriented so that the distal end of the retina is at the left and the proximal end to the right. D, distal pigment; P, proximal pigment; R, reflecting pigment. Fir,. 10. Day-light retina. The distal pigment has migrated proximally, the proximal pigment has moved above the basement membrane and lies directly to the right of the distal pigment, but the reflecting pigment is fixed. FIG. 11. Day-dark retina. The distal pigment is seen to the left in its distal position, while the proximal pigment has migrated completely below the basement membrane. FIG. 12. Night-light retina. The distal pigment is in the same position as in the day-dark retina, but the proximal pigment (which has been retouched in this photograph) is in the typical light-adapted position. FIG. 13. Night-dark retina. The distal and the proximal pigments are in the positions characteristic for a dark-adapted eye. FIGS. 14-17 are comparable retinas of Trachypeneopsis, arranged in the same sequence as Figs. 10-13 and showing similar movements of the retinal pigments. Fig. 15 shows the very small amount of reflecting pigment in this retina compared with the others of this series. FIG. 18. Day-light eye of Portunus. The distal pigment is the black band tn the left. Granules of this pigment are dispersed proximally, while granules of the proximal pigment are dispersed distally towards the left. FIG. 19. Day-dark eye. The positions of the two pigments are the same as in Fig. 18, in spite of the fact that the animals were maintained in constant darkness. FIG. 20. Night-light eye. The pigments arc in the positions characteristic for a light-adapted retina. FIG. 21. Night-dark eye. The granules of the distal pigment have migrated distally from the processes in which they had been dispersed, while the proximal pigment granules have moved completely below the basement membrane. The blackness of the processes is due to the stain, and not to any contained pigment. 13 17 i-i-D P ' t-D 21 .*r- 12 Atfl "•*> •y 'VD 20 P 11 15 •< -D P 10 PLATE II 183 184 L. H. KLKINHOLZ Trachypeneopsis mobilispinis Crustaceans of this species were dredged off sandy bottoms in Castle Harbor and in Bailey's Bay. Fight animals were available for study; three of these were males and the rest females. A particularly thick cxoskelcton and dense pigmentation obscured the eye-stalk so that tin- positions of tin- distal retinal cells of whole eyes could not be determined bv direct microscopic examination. Study of sectioned retinas, however, revealed a close similarity to the con- dition found in /:iisicyoniiJ. the chief difference being in the si/e rela- tionships of the ommatidial components. In this species, too, the reflecting pigment apparently undergoes no positional changes. In some specimens this pigment was completely absent, and in one only a trace of it was evident ( Fig. 15 ). The retinal responses to light and to darkness art- so similar to those in I:nsicyonia that the same diagrams of Plate I (Figs. 2-5) serve to illustrate the changes. The distal and the proximal pigments show the usual posi- tional changes in adaptation to light and to darkness. The resemblance to Eusicyonia in tin's respect is heightened by the fact that the persistent diurnal rhythm is found in the distal pigment cells, which, in night- light eyes, are in the position characteristic for a dark-adapted retina. Brachyurans I'our species ot brachyuran crustaceans were studied for the move- ments of the retinal pigments. Since the responses in all four were found to be identical, thev mav conveniently IK- described together. Kxi'i.. \\.\TION IIF Pi. ATI: 1 1 1 I'i.ys. 22 25 arc retinas of .''urtlicnopc scrnila and Ki.u». _'<> _"> represent those of Cal(tf>f>ii fltimmea. Tlie retinas are oriented with the distal end at the top of the photograph and tin- proximal end toward the bottom. /), distal pigment; /', proximal pigment. FIG. 22. Day-light retina. The distal pi.ument is dispersed proximally wliilc the proximal pigment granules, have miniated, for the most part, above the base- ment membra i n . Fin. 23. Day-dark retina. The positions of the two pi.mncnts are the same as in the day-light eye, in spite of constant darkness. FIG. 24. Night-light retina The pigments are in the positions typical for the light-adapted eye. FIG. 25. Night-dark retina. The granules of the distal pigment have retreated distally, while those of the proximal pigment have withdrawn completely from the processes to a position below the basement membrane. FIGS. 26-29 are photographs of letinas of (.'tilnf>fni arranged in the same se- quence as Fi'js. 22 25 and sbouiiiL; a similar persistent rhythm in the day-dark eye. DIURNAL MOVEMENTS OF THE RETINAL I'K.MKXTS 185 23 ' I 'l I I, ' I. ', •' ' I " ''•" '"" -4^>^**^-teK;S*^ :;..'/ y "f*y. ' D P ' ~\ 1 1 1 1 1 r tA**" *' ^"^ 1 1 D ' i i i P D i - 28 \i ^ryrj^v- ' * D '\ P 29 PLATE III 186 L. H. KLE1XHOLZ There are two types of black retinal pigments in these animals, a proximal and a distal set of cells; reflecting pigments seem to be com- pletely absent. Diagrammatic illustrations of ommatidia are shown in Figs. 6-9. The photomechanical changes of the retinal pigments in these brachyurans, although not as marked as those in the macrurans. are still evident : Day-light eye A. The distal pigment is dispersed proximally toward the basement mem- brane. B. Most of the proximal pigment has migrated above the basement mem- brane and surrounds the rhalxlmne. Day-dark eye A. The distal pigment remains in the position typical for a light-adapted retina. B. The proximal pigment also shows the persistent rhythm, remaining in the position found in a li-.'ht-adapted eye in spite of conditions of constant darkness. Night-light eye .-]. The distal pigment granules arc in the same position as in the day-light eye. B. The proximal pigment is also in the light-adapted position. Night-dark eye A. The distal pigment granules have retreated distally from the processes and lie in the cells \vlrch are located along the posterior region of the cones. B. The proximal pigment has moved completely helo\v the basement mem- brane so that the processes of these cells are entirely free of pigment granules. DISCUSSION There is still much ignorance as to the nature of tin- mechanisms in- volved in these rhythmic activities. 1 f the retinal pigments were uni- lorm in their periodic responses under constant conditions, the problem as to the mediating agents would be relatively simple. P.ut published ac- counts of the behavior of these pigments among the different crustaceans reveal perplexing heterogeneities in this respect. Examination of Table I will show the responses repm-ted for various crustaceans. In the first four species of macrurans. there is a uniformity ot re- sponse in that the same set of retinal pigment, the distal, shows per- sistent periodic- movements in the- same- direction under the1 same con- ditions of illumination. The next two species of shrimps an- unusual in that they show evidence of possessing a double rhythm; not only are distal and reflecting retinal pigments involved in the periodic migra- tion, but the periodicity occurs twice within a 24-hour cycle, once- during the daytime when the' animals are maintained in darkness, and once at night when the shrimps are kept under constant illumination. DIURNAL MOVEMENTS OF TIIK RETINAL I'K.MENTS 187 Up to this point, the various responses, although still slightly con- fusing, are open to interpretation on a hormonal basis as found in the case of Pahciiionctcs (Kleinholz, 1936). The pu//.ling situations are those where the distal pigments show a persistent rhythm, but the re- flecting pigments at the same time undergo only the normal changes in position characteristic for the particular condition of illumination. In Lcandcr tcinticoniis and in Lat rentes fncontin the situation is re- versed, the reflecting pigments showing the rhythm while the distal pig- ment cells respond only to differences in light intensity. If endocrine control of these two pigments is universally present in the crustaceans, TABLE I Crustacean Distal Reflecting Proximal Investigator Macrobrachium olfersii NL Welsh (1930a) Macrobrachium acanthurus Eusicyonia n sp NL NL Ab or F — Welsh (1930a) This paper Trachypeneopsis mobilispinis . . . Leander affinis NL DD and NL Ab. or F DD and NL — This paper Welsh (1935) Anchistioides antiguensis DD and NL DD and NL Welsh (1936) Portunus anceps DD Ab. DD This paper Portunus depressifrons DD Ab. DD This paper Parthenope serrata DD Ab. DD This paper Calappa flammea DD Ab. DD This paper Leander tenuicornis NL Welsh (1935) Latreutes fucorum NL NL Welsh (1935) Cambarus virilis DD Bennitt (1932a) Peneopsis goodei F F DD Welsh (1935) — , pigment shows normal photomechanical changes but shows no periodicity. Ab., pigment is absent from the retina. F, pigment is present, but undergoes no positional changes. NL, rhythm in a night-light eye, the pigment moving into a typical dark position. DD, rhythm in a day-dark eye, the pigment moving into the position characteris- tic for a light-adapted retina. such differences in response may possibly be due to threshold variations in reactivity to the same hormone, or. there may be two hormones in- volved, one for the distal pigment and one for the reflecting pigment. The mechanisms involved in the migration of the retinal pigments in the four brachyurans reported here is less open to analysis chiefly because so little is known about them. These crustaceans are the only ones thus far reported which show a pronounced periodicity in both the distal and in the proximal pigments at the same time. Cambarus and Peneopsis seem to fall into the same group in that the proximal pig- ment shows the rhythm. 188 L. H. KLEIXHOLZ The mediating agency for the migration of the proximal pigment is not known. In Pahcmonctcs the eye-stalk extracts which affect the distal and the reflecting pigments have no effect on the migration of the proximal pigment. Early workers on the migration of the proximal retinal pigment were of the view that this activity was under nervous control. But as Bennitt (1932/0 states, the main argument against this belief is that no efferent nerve fibers have been found supplying these cells, their only nervou.s connection apparently being afferent fibers going to the optic ganglia (Parker, 1895). From the results of his experi- ments on the interrelation between the eyes of crustaceans (with regard to proximal pigment migration) Bennitt (1932&) believes that an endo- crine control may be involved. It is evident from this discussion that the possibility of formulating a general theory for the phenomenon of persisting diurnal rhythm is still remote. The fact that those crustaceans which have been studied lack the uniformity of behavior that is sought by the experimentalist, is probably sufficient proof of the complexity of the mechanism. In our present state of knowledge it can only be hoped that when the retinal pigments of other crustaceans have been studied with regard to hor- monal control as in Palccmonetes, and the innervation of the proximal pigment cells studied by means of modern neurological techniques, a sufficient amount of information will have been gathered to attempt a general explanation of persistent periodicity in the retinal pigment mi- gration of crustaceans. SUMMARY 1. Persistence of a diurnal rhythm in the migration of the retinal pigments of several Bermudan crustaceans, in spite of constant condi- tions of illumination or of darkness, is reported. Two macrurans. a new species of Eusicyonia and Trachypeneopsis mobilispinis, show tin- distal retinal pigment in the position characteristic for the dark-adapted eye when the retinas of illuminated specimens are fixed at night. 2. Four species of brachyurans, Portnmts anccps, P. depresstfrons, Parthenope scrrata, and Calappa flanimca. show the distal and the prox- imal pigments in the typical light-adapted position when the specimens are kept in the dark-room and are fixed during the day. 3. The retinas of some of the macrurans showed deficiencies or ab- sence of the reflecting pigment. This is thought to be due to differences in metabolism rather than to fixation artifacts. 4. Hormones are considered as the possible mediating agency in- volved in the phenomenon of persisting diurnal retinal rhythms. DIURNAL MOVEMENTS OF THE RETINAL PIGMENTS 189 LITERATURE CITED BENNITT, R., 1932o. Diurnal rhythm in the proximal pigment cells of the crayfish retina. Physiol. Zoo!., 5: 65. BENNITT, R., 1932b. Physiological interrelationship in the eyes of decapod Crustacea. Physiol. Zool., 5: 49. CARLSON, S. PH., 1936. Color changes in brachyura crustaceans, especially in Uca pugilator. Kunt/l. Fysiograf. Sdllskap. Lund Forhand., 6: 1. EWALD, A., AND C. FR. W. KRUKENBERG, 1882. Ueber die Verbreitung des Guanin, besonders iiber sein Vorkommen in der Haut von Amphibien, Reptilien, und von Petromyzon fluviatilis. Untcrsucli. ous dcm Physiol. lust, dcr Univ. Heidelberg, 4: 253. GAMBLE, F. W., AND F. W. KEEBLE, 1900. Hippolyte varians : a study in color- change. Quart. Jour. Micros. Sci., 43: 589. HANSTROM, B., 1935. Preliminary report on the probable connection between the blood gland and the chromatophore activator in decapod crustaceans. Proc. Nat. Acad, Sci, 21: 584. KLEINHOLZ, L. H., 1936. Crustacean eye-stalk hormone and retinal pigment mi- gration. Biol. Bull, 70: 159. KLEINHOLZ, L. H., 1937. Studies in the pigmentary system of Crustacea. I. Color changes and diurnal rhythm in Ligia baudiniana. Biol. Bull., 72: 24. ROLLER, G., 1928. Versuche iiber die inkretorischen Yorgange beim Garneelen- farbwechsel. Zcitschr. vcrgl. Physiol., 8: 601. MEGUSAR, F., 1912. Experimente iiber den Farbwechsel der Crustaceen. Arch. Entwick.-mcch., 33: 462. MENKE, H., 1911. Periodische Bewegungen und ihr Zusammenhang mit Licht und Stoffwechsel. Arch. ges. Physiol., 140: 37. PARKER, G. H., 1895. The retina and optic ganglia in decapods, especially in Astacus. Mitt. Zool. Stat. Neapcl, 12: 1. PARKER, G. H., 1897. Photomechanical changes in the retinal pigment cells of Palaemonetes, and their relation to the central nervous system. Bull. Mus. Comp. Zool., 30: 273. PERKINS, E. B., 1928. Color changes in crustaceans, especially in Palsemonetes. Jour. Expcr. Zool., 50: 71. REDFIELD, A. C., 1918. The physiology of the melanophores of the horned toad Phrynosoma. Jour. Expcr. Zool., 26: 275. SJOGREN, S., 1934. Die Blutdriise und ihre Ausbildung bei den Decapoden. Zool. Jahrb. Abt. Anal., 58: 145. SLOME, D., AND L. HOGBEN, 1929. The time factor in the chromatic responses of Xenopus Isevis. Trans. Roy. Soc. South Africa, 17: 141. WELSH, J. H., 1930a. Diurnal rhythm of the distal pigment cells in the eyes of certain crustaceans. Proc, Nat. Acad. Sci.. 16: 386. WELSH, J. H., 1930^. The mechanics of migration of the distal pigment cells in the eyes of Palsemonetes. Jour. Expcr. Zool., 56: 459. WELSH, J. H., 1935. Further1 evidence of a diurnal rhythm in the movement of pigment cells in eyes of crustaceans. Biol. Bull., 68: 247. WELSH, J. H., 1936. Diurnal movements of the eye pigments of Anchistioides. Biol. Bull, 70: 217. YOUNG, J. Z., 1935. The photoreceptors of lampreys. II. The functions of the pineal complex. Jour. Expcr. Biol., 12: 254. BACTERIA AXI) THE IM loSI'HORUS CYCLE IX THE SEA1 CHARLES E. RENN (Prom the II' nods Hole Occanographic Institution, Woods Hole, and the Biolot/ic>tl Laboratories, Harvard University) Marine bacteria, growing and multiplying rapidly under favorable conditions, are potential accumulators of the sea's labile organic matter. Lohmann (1911) has estimated that they may bind as cell substance each year from six to three hundred times the annual production of slower growing nannoplankton feeding upon them. It is surprising, then, to discover, as Fischer (1894) and many others -nice have pointed out. that the bacterial population of the sea and large fresh- water lakes is very low — rarely exceeding a viable count of a hun- dred at the surface and extending through a lesser range in the depths i 1'iirge and Juday. 1922; Reus/er, 1933). Lohmann found the mass of bacterial cells in ->ea water to represent an insignificant fraction of the nannoplankton ; and it may be calculated, making generous allowance for non-viable cells, that the amount of phosphate, nitrate, or other nutrient substance bound bv bacteria is immeasurably small. Thus, a million cells per cubic centimeter, at least a thousand times the likely bacterial population, would be equivalent to less than 3 mgm. J'()4 per cubic meter (Buchanan and Fulmer, 1933. pp. (>(>-7l). Bacterial cell >ubstancc is notoriously rich in phosphorus, but it is apparent from their small numbers, that bacteria do not compete effectively for the sea's limited store of this element. Do these small numbers also indicate a low limit to their activity in the breakdown of debris and the regenera- tion of inorganic nutrient O If Fischer's observations regarding the short life of marine bacteria are valid ( he estimated their average viable span to be in the order of a halt hour), then Lohmann's emphasis on their role as converters of organized material rather than accumulators gives a sound picture of their relations in marine economy. Rate of growth, length of the viable pi-riod. and rapidity ot autolvsis after death become critical tactors in determining their value as regenerating agents. Several experiments make it seem plausible that short-lived bacteria an- active in phosphate regeneration, and oiler confirmation for the smaller, niter-sea^ dial phosphorus cycles indicated in the investigations < "iitrilnitioii No. 133 from the Woods Hole Oceanographic Institution. Contribution from the Bermuda Biological Station for Research, Inc. 190 BACTERIA AND THE PHOSPHORUS CYCLE IN THE SEA 191 of Harvey and co-workers (1935) on plankton production, and re- marked by Seiwell (1935). \Vben specimens of sea water arc stored in bottles at room temperature1 tbe bacterial population therein rises in t\vo or three days to a level much higher, from a thousand to a hundred thousand times, than that of natural waters. This sudden increase is followed by a precipitous drop in viable numbers to a persistent low count (Waksman and Renn, 1936). That such unrestricted develop- ment of bacteria may withdraw appreciable quantities of phosphate from • o \C-\J x O£ or .7 ^1 J O " > , / \ ' 28 UJ i- Jioo J _j UJ to (T UJ V \ 245 o: UJ Q. tL 80 / — -^_ \ PHOSPHATE ^^-^ > Q. J-5 ^» '' "^"^ v -~*~*~~~~' 20 en i • ~~^ Q ~ or \ z •^ UJ / \ ^ | 1 < — .4 Q- 60. / \ ———"""" 16 ^ Q / -\' OXYGEN O UJ O •' s' \ X §3 Z 40. j / \ h- 12 z O UJ C 1 / r ft CE . 2 ^ © / \ 8 UJ Z -T- / xx 1— UJ X ' 8)20- O / \ / \ BACTERIA ex. ( Li 0 X -/ Q. / 0 a — 0 TIME IN DAYS FIG. 1. Showing assimilation and regeneration of phosphate by bacteria in stored samples of sea water from Woods Hole Harbor — July 10, 1935. Phos- phates in triplicate, other determinations in duplicate. sea water is shown by the decreasing concentrations in unpreserved specimens taken for analysis during the first two or three days of storage. The first experiment was designed to relate these two behaviors- increase in bacterial numbers and assimilation of phosphate. Water from the mouth of Woods Hole Harbor was enriched with K,HPO4 to bring the phosphate content to about 100 mgm. per cubic meter, filtered 192 CHARLES E. REXN through quartz sand, and dispensed into sterile, standard, oxygen de- mand bottles. These were stored in the dark at room temperature for varying lengths of time after which phosphate, bacterial number, and dissolved oxygen determinations were made. Figure 1, incorporating the results, clearly shows the rapid decrease in phosphate during the first three days of storage as noted before. Simultaneously the bac- terial numbers and oxygen consumption reach their maximum. After three days the viable bacteria fall to a low level, and the oxygen uptake SCO rt DATS 15 MGM GLUCOSE PFR LITER 14 DAYS 30 MGM GLUCOSE PER LITER 0135 14 DAYS 60 MGM GLUCOSE PER LITER Fir,. 2. Showing assimilation and regeneration of phosphate in stored sea water from off Bermuda, January, 1936. All determinations in triplicate. slackens off. On the seventh day there are signs of phosphate regenera- tion. There is, then, an interval of two or three days in which 20-25 mgm. of phosphate per cubic meter is bound as non-viable bacterial cells. The count of viable cells has qualitative value only, indicating the phase of growth rather than actual numbers. Assimilated phosphate is held by a quantity of bacterial cells that is roughly proportional to tin- organic carbon oxidized during rapid growth — in this case an equivalent of 0.75 mgm. glucose per liter.) It has been pointed out that the length BACTERIA AND THE PHOSPHORUS CYCLE IN THE SEA 193 of time that phosphorus may be held by bacterial cells synthesized during decomposition of organic phosphorus compounds is significant in esti- mating their activity in phosphate regeneration. If this is long, then the process is retarded, but if, as this preliminary experiment suggests, the cells lyse within a short time, the turnover through their agency may be rapid. In January, 1936, there was an opportunity to study the complete cycle of bacterial assimilation and regeneration under particularly favor- able conditions, using the waters about Bermuda and the facilities at the Bermuda Station for Biological Research. The waters in this region are characteristically poor in phosphate, organic phosphorus, and avail- able carbon — components that had proven troublesome in attempts to extend the experiments in Woods Hole. (A number of determinations showed the total phosphorus in shallow bays between St. Georges and Shore Hills to be less than 40 mgm. PO4 per cubic meter.) Sea water, taken from the surface about three miles southeast of Castle Roads, was filtered through quartz sand (imported), and enriched with varying amounts of phosphate and glucose, and a uniform excess of nitrate and iron. This was dispensed into a series of three-liter bot- tles and stored in the dark at room temperatures of 19° to 21° C. At intervals specimens were taken for phosphate analyses. Results of the periodic inorganic phosphate analyses on the pre- pared, stored water are given in Fig. 2. Phosphate assimilation is complete in about five days, by which time it may be assumed bacterial growth has passed its maximum. Regeneration follows almost at once —rapid at first, and then leveling at fourteen days toward completion. Most of the phosphate is regenerated within this time, but a fraction, ranging from an eighth to a third, and greatest where relatively large quantities of glucose have been added, is still bound at the end of thirty- seven days when the experiment was terminated. It seems clear that bacterial cells do not bind phosphorus for more than a few days under storage conditions. If autolyzing bacterial cells liberate phosphorus as phosphate there always exists the possibility that diatoms and other phytoplankton may similarly bring about direct regeneration. Two attempts to demonstrate direct regeneration gave negative results, but it will be recognized from the methods followed that the mechanism is by no means disproved. Ten-day cultures of Nitschia grown in synthetic media were filtered and washed free of inorganic phosphate on a Seitz filter. The cells were suspended in 200 ml. of distilled water and washed on a second filter with 800 ml. of sea water, the lysate passing into a sterile aspirator 194 CHARLES E. REXX bottle fitted with tubes for periodic sampling. Phosphate analyses at intervals of 14. 20, and 48 hours indicated no increase whatever in in- organic phosphate content. A modification was arranged in which cultures of Xitzcliia were washed on the Seitz filter, removed with a thin layer of the filter pad, lysed in distilled water for two hours, suspended in ten volumes of sea water, and allowed to stand. After three and eight hours, respectively, specimens were centrifuged and analyses made. These yielded no measurable amount of phosphate. DISCUSSION AND SUMMARY Bacteria, despite their potentialities as rapidly multiplying organisms, do not compete with higher phytoplankton for phosphate. Many fac- tors may be suggested to account for their low numbers, but short life and brief viability period undoubtedly play a part. Under storage conditions where their rapid growth is favored thev assimilate significant quantities of phosphorus, but this is quickly regenerated following their death. This suggests that the numbers of viable cells in natural water is not a complete index of bacterial activity, and that the efficiency of the organisms as agents of regeneration is inversely proportional to their life span. The relations of the rapid growth phase, short period of maximum viable population, and the declining phase are not unique in marine bac- terial cultures but are characteristic of bacterial populations in general (Buchanan and Fulmer, 1933, pp. 69-71). The curve representing them is the resultant of many growth and death curves over much shorter intervals, so that the number of cells actually produced during three, five, seven or fourteen days is indeterminately greater than the population at any one time. Many other mechanisms for phosphate regeneration may be active in the sea. These experiments demonstrate one possible system, in- volving the agency of bacteria. They also offer an explanation for the insignificant amounts of bacterial cell substance in the nannoplankton. BIBLIOGRAPHY BiRf.E, E. A.. AND C. JUDAY V>22. The inland lakes of Wisconsin. The Plankton. I. IK quantity an-! chemical composition. II' is. Geol. Nat. Hist. SUIT. Hull. Sci. Ser" 64: 13. Hi < MA \.\\, K. E., AND I'".. I. EeLMER, l('.vv Physiology and Biochemistry of Bacteria, Vol. 1 (a) pp. 42-50, (/>) pp. 69-71. Williams and Wilkins, Baltimore. I ' HER, 1'.. IX'M. Die Bakterien des Meeres. Ergebniss der Plankton-Expedition del Huml)«.ldt Stiftung. p. 60. Kiel and Leipzig. BACTERIA AND THE PHOSPHORUS CYCLE IN THE SEA 195 HARVEY, H. \V. L., L. H. N. COOPER, M. Limn u, AND E. S. RUSSELL, 1935. Plank- ton production and its control. Jour. Mur. />/<>/. Ass'n, N. S.. 20: 407. LOHMANN, H., 1911. Ober das Nannoplankton und die Zentrifugierung klcinster Wasserprobcn zur Gewinnung di-ssi-lbi-n in lebendcm Zustande. Int. Rev. Ges. Hydrobiol. und Hydrogr., 4:1. SEIWELL, H. R., 1935. The annual organic production and nutrient phosphorus re- quirement in the tropical western north Atlantic. Jour, du Conscil Int. pour VExplor. dc la Mcr. 10: 20. REUSZER, H. W., 1933. Marine bacteria and their role in the cycle of life in the sea. III. Distribution of bacteria in the ocean waters and muds about Cape Cod. Biol. Bull., 65: 480. WAKSMAN, S. A., AND C. E. RENN, 1936. Decomposition of organic matter in the sea by bacteria. Biol. Bull, 70: 472. INDUCTION OF KXDOMIXIS IX PARAMECIUM AURELTA T. M. SONNEBORN (From the Department of Zoology, Johns Hopkins University) Attempts to induce endomixis in Parainccinin aurclia have in the past led to conflicting results, [olios (1916) and Young (1917) re- ported successful inductions; but Woodruff (1917) found that the time of occurrence of endomixis was but slightly modifiable by external con- ditions. Erdmann ( 1"20) shared Woodruff's views on the inner de- termination of endomixis and rejected the results of Jollos and Young as unsound. The present paper, in general agreement with Jollos. re- ports a reliable1 method of inducing endomixis and sets forth the quantitative difference's in response to this method at different periods of the interendomictic interval and in different stocks of the species. The procedure employed for inducing endomixis was that which has been described in earlier papers ( Sonncborn. 1CM6). The organisms were under culture in isolation lines under uniform conditions. To induce endomixis, the surplus animals available after the daily transfer of the lines to fresh culture fluid were collected into 10 drops (approxi- mately 625 cu. mm.) of fresh culture medium in a Columbia dish, and were kept without further change at 31° ± 1° C. liy culture in this manner endomixis was induced, as hereafter described. Two stocks of Parainccinin aurclia were studied with relation to the induction of endomixis: these \\cre the stock l\. dealt with in earlier papers by the author and associates, and Woodruff's long-lived stock, here designated \V. At various times over a period of three years. 11 groups of lines of stock I\ and X groups of stock \Y were investigated. Of the 19 groups studied. IS were started with one or a few ani- mals in the climax of endomixis. while one was started with 24 ex- conjugants. In each group begun with animals in endomixis, the few original animals were soon expanded into usually 12 isolation lines. In all groups, one or more individuals from each isolation line were stained daily, to determine the occurrence of endomixis. Any line that was found in endomixis was discarded, in order that the interval of time since the previous endomixis should be the same for all the individuals placed in any one of the small mass cultures. The mass cultures derived as above described from these isolation lines were observed at intervals not exceeding 12 hours, to determine 196 INDUCTION OF ENDOMIXIS IN PARAMECIUM 197 whether conjugation was in progress. Also, a random sample of 20 to 50 individuals of each culture was stained usually once a day to deter- mine whether endomixis was in progress. Stock R As set forth in an earlier paper (Sonneborn, 1936), mass cultures of stock R set up soon after endomixis or conjugation, in the way above described, soon pass into conjugation, so that such cultures are un- available for the study of the induction of endomixis. I>ut when the interval from the preceding nuclear reorganization (endomixis or conju- TABLE I Stock R. The percentages of daily isolation lines at 27.5° C. in endomixis on successive days of the interendomictic interval, compared with the percentages of animals in endomixis in mass cultures at 31° C. on corresponding days. In each case, the mass cultures had been derived from the isolation lines on the preceding day. Percentage of endomixis In mass cultures at 31°C. 0 0 0 37.6 58.7 39.7 53.0 93.4 98.3 97.5 97.5 100 97.0 100 100 100 gation) is greater, conjugation does not occur, but some of the indi- viduals pass into endomixis. The percentage of those thus undergoing endomixis was accurately determined, and compared with the percent- age occurring in the isolation cultures (still in progress) from which they were derived. In this way the effect of the mass culture condi- tions in inducing endomixis is numerically expressed. Typical results are given in Table I. From the twelfth until the twenty-fourth day after the previous endomixis, the percentage of indi- viduals in endomixis in the mass cultures is always much greater than that in the isolation cultures (until by the twenty-fourth day the last isolation line has passed into endomixis, bringing the comparison to a Day since initial In isolation lines endomixis at 27. 5° C. 9 0 10 0 11 0 12 .8-3 13 0 14 0 15 16.7 16 42.9 17 0 18 50.0 19 57.1 20 0 21 66.7 22 0 23 0 24. .100 198 T. M. SONNEBORN close). \Yith increasing time since the preceding endomixis, the per- centage of individuals in which endomixis was induced quickly rose from 37.6 per cent on the twelfth day to 98.3 per cmt on the seventeenth day, after which it remained at or near 100 per cent. Similar results, differing only in details, were obtained with the other groups of stock R, including the group begun with ex-con jugants. These results thus confirm the reports of Jollos (1916). Endomixis is induced by the treatment set forth and, within the range of intervals examined, it is induced less readily the shorter the interval since the preceding endomixis. This relation to the time since a previous endo- mixis is also shown in other ways. (1 ) It took a longer period in mass culture to induce endomixis in the earlier mass cultures derived from given isolation lines than in the later ones. Thus the mass cultures set up on the seventh to tenth days of a certain series yielded no endomictic animals until these cultures were two days old, whereas in the later cul- tures of the series endomixis was induced in one day. (2) The pro- duction of large percentages of endomixis in the early cultures required a type of treatment that was unnecessary in the later cultures. Thus, a culture set up on the tenth day after mdomixis yielded no more than 10 per cent endomixis on any of the four days it was examined, but a sub- culture from it set up on the third day and provided with fresh culture fluid gave 44.4 per cent endomixis on the following day. Similar re- sults were obtained with other cultures : a culture set up on the eleventh day after endomixis gave but 5.4 per cent endomixis on the second day; it was then divided into two parts one of which was provided with 5 drops of fresh culture fluid, the other not. On the next day, the former had 69.2 per cent endomixis, the latter 5.8 per cent. It appeared that the percentage of endomixis in these early cultures reached its maxi- mum at about the time when the food supply was giving out and that further increase in the percentage of endomixis could be obtained only by supplying additional food. 'I he relations observed were usually of the type just set forth, but very rarely exceptional results were obtained. In lines descended from one endomictic individual, mass cultures set up on the first, third, and iitth days after endomixis yielded at once another endomixis instead of conjugation, as would normally have been expected at this period. These lines also went into endomixis after 5 to 8 days in the isolation cultures. It is perhaps significant that in this case the initial nuclear reorgani/ation was highly abnormal: on the day following the stage of i»ni]>lete inacronnclear fragmentation there were neither numerous -mall fragments nor pale anlagcn of the new macronucleus, such as nor- mally occur; instead, several large, aggregated spheres of chromatin INDUCTION OF ENDOMIXIS IN PARAMECIUM 199 were present. Thereafter, normal-appearing macronuclei were char- acteristic of this line. Similar abnormal reorganizations have on sev- eral occasions been followed by extraordinarily short intervals, and ap- pear to be tlu- basis of the unusually quick recurrence of enclomixis in these cases. Stock IV1 In stock YV, Sonneborn and Cohen (1936) have shown that mass cultures set up in the way here set forth do not yield conjugants, and examination at not more than 12-hour intervals showed that conjugation did not occur in the present cultures. This made it possible to deter- mine from the start the percentages of individuals in endomixis. In Table II are given, for a typical group of lines of stock W, the percentages of individuals in endomixis in the mass cultures set up on successive days of the interendomictic interval and kept for various periods of time at 31° C. The table is to be read as follows: the cul- ture set up on the second day yielded no endomixis ; the culture set up on the third day yielded 8.6 per cent endomixis after five days at 31° C., i.e., 5 -(- 3 = 8 days after the initial endomixis and so on. Table II shows that endomixis was regularly induced every day beyond the sev- enth after the initial endomixis. In the isolation lines endomixis did not occur until the thirtieth day. In some of the other groups, the earliest induction of endomixis oc- curred sooner than in the typical group. In two groups endomixis was induced four days after the initial endomixis and in one group, three days. Usually, however, at first only small percentages of endomixis could be induced after 5 days of treatment. Later, comparable small percentages were induced in four days of treatment. Still later, treat- ments of four days yielded nearly 100 per cent endomixis; then 100 per cent endomixis was obtained in three days and finally in two days. In contrast to stock R, which yielded 100 per cent endomixis in one day in the later cultures of a series, stock W never yielded 100 per cent enclomixis in less than two days of treatment. In stock W, as in stock R, the percentage of induced endomixis increased with the time since the preceding endomixis. This is shown in Table II and also by a comparison of the percentages of endomixis induced on the same day in the cultures of several groups differing in time since the last endomixis. Thus, the mass cultures set up on a certain day yielded after two days 100 per cent in a group 24 days past endomixis, 97.4 per cent in a group 19 days past endomixis, and 8.2 per cent in a group 4 days past endomixis. In two-clay-old cultures 1 In the work with stock W, the author was greatly assisted by Dr. B. M. Cohen. 200 T. M. SOXXKBORX set up on another day there were 100 per cent in a group 23 days past endomixis. 30.8 per cent in a group 18 days past endomixis, and 0 per cent in a group 13 days past endomixis. TABLE II Stock \Y. The percentage of individuals in endomixis in mass cultures set up daily from a group of isolation lines. Cultures set up from the first to the twenty- eighth day after the initial endomixis and kept at 31° C. for 1 to 5 days. During the entire period there \vas no endomixis in the isolation lines from which the mass cultures were derived. Days from endomixis till mass cultures set up Percentage of endomixis in mass cultures Number of days since culture set up 1 2 3 4 5 2 0 0 0 0 — 3 — — — — 8.6 4 — — — — 29.7 5 — — — 6.5 — 6 — — — — 26.8 7 — — — 83.9 — 8 — — 65.6 97.1 — 9 — 32.3 69.1 97.8 — 10 0 36.7 100 — — 11 0 96.9 — — — 12 0 56.0 100 — — 13 0 60.6 100 — — 14 0 100 — — — 15 3.3 100 — — — 16 0 96.3 — — — 17 20.0 97.4 — — — 18 0 — 100 — — 19 — 92.9 — — — 20 0 100 — — — 21 — 100 — — — 22 23 10.5 100 85.3 1 — — 24 — 93.3 — — — 25 — 48.4 47.6 48.5 50.0 26 — 23.5 77.8 — — 27 — 88.9 — — — 28 — 96.9 — — — On the other hand, factors other than the time since the last nuclear reorganization influence the percentage of endomixis induced. There were many instances in which the percentage of endomixis was less at a later period after endomixis than earlier. For example, in Table II. there wa> '><>.'' per cent endomixis after two days treatment of the iiiltun- -it up mi the eleventh day after endomixis. and only 56.0 per INDUCTION OF KNDOM1XIS IN PARAMECIUM 201 cent in the culture set up a day later. Such irregularities were prob- ably due to differences in the environmental conditions at different times, as indicated by the fact that the same departure from the general trend was usually shown by all the cultures set up on a particular day. An example of the effect of one such environmental change was provided by the accidental rise of temperature to 35° C. in the box ordinarily kept at 31° C. In four groups of cultures exposed to this high tem- perature, the percentages of endomixis were 25.7 per cent, 32.9 per cent, 36.0 per cent, and 0 per cent, as compared with 79.8 per cent, 75.1 per cent, 96.7 per cent, and 43.9 per cent, respectively, for these same four groups on the two preceding days at 31° C. Thus, rise of temperature from 31° to 35° reduced the percentage of endomixis induced; al- though, as earlier shown, rise from 27.5° to 31° greatly increased the percentage of endomixis. There are doubtless other as yet unanalysed factors influencing the induction of endomixis. SUMMARY In both stocks R and W of P. aitrclia, endomixis can be induced by placing at 31° C. small mass cultures containing the surplus animals from isolation lines. In stock R, induction cultures set up soon after endomixis yielded conjugants, so that the induction of endomixis could not be studied quantitatively in these. Later cultures, however, showed increasing percentages of induction until 100 per cent was obtained after one day at 31° C. Earlier cultures which gave but low percent- ages of endomixis with this treatment could be induced to give larger percentages by subculturing or adding some fresh culture medium to the original induction culture. In a few exceptional cases, abnormal reorganization of the nuclear apparatus was soon followed by a normal endomictic reorganization in isolation lines, and by the induction of endomixis instead of conjugation in the induction cultures set up im- mediately after the abnormal reorganization. In stock W, the absence of conjugation made it possible to examine the percentages of endomixis induced at all periods of the interendo- mictic interval. During the first few clays after an endomixis, it re- quired prolonged treatment to induce even a low percentage of endo- mixis, but as the time since the last endomixis increased, higher and higher proportions of endomixis were induced with shorter and shorter treatments, until 100 per cent endomixis could be induced in two da\ >. Irregularities in this progression are probably due to environmental dif- ferences in treatment from day to day. Rise of temperature above 31° C. was shown to reduce the percentage of endomixis induced. Other conditions also probably play a similar role. 202 T. M. SONNEBORN Stocks R and \V differed in t\vo respects in regard to their reaction to the conditions favoring the occurrence of endomixis. (1) In stock \V, endomixis could regularly be induced 7 days after a preceding endo- mixis, and in some cases, as early as the third day ; but in stock R. the typical response at this period was conjugation, not endomixis. (2) Late in the series of induction cultures, 100 per cent endomixis could be induced with a treatment of one day in stock R. but treatment of two days was required in stock W. LITERATURE CITED ERDMAXX, R., 1920. Endomixis and size variations in pure bred lines of Para- mecium aurelia. Arch. f. Ent'^'.-mcch., 46: 85. JOLLOS, V., 1916. Die Fortpflanzung der Infusorien und die potentielle Unster- blichkeit dor Einzelligen. Biol. Ccntralbl., 36: 497. SoNNEHoKx, T. M., 1936. Factors determining conjugation in Paramecium aurelia. I. The cyclical factor : the recency of nuclear reorganization. Genetics, 21: 503. SONNEHORX, T. M. AND B. M. CoHEX, 1936. Factors determining conjugation in Paramecium aurelia. II. Genetic diversities between stocks or races. Genetics, 21: 515. F, L. L., 1917. The influence of general environmental conditions on the periodicity of endomixis in Paramecium aurelia. Biol. Bull.. 33: 437. ;, R. T., 1917. Experimental induction of endomixis in Paramecium aurelia. /d»r. Expcr. Zool., 24: 35. SOME EFFECTS OF DIET OX THE GASTRIC EPITHELIAL CELLS OF THE GRASSHOPPER, MELANOPLUS DIFFERENTIATES THOMAS1 CHARLES HODGE, 4TH (Prom t/ic B'wlo per cent that of the satisfactory. In view of the wide individual varia- tion, this figure is hardly significant. Gross morphology is not affected by this difference in diet. Number of Cells The number of cells per section on the satisfactory diet ranged from 365 to 775 (average 552.7). The number on the oat diet, from 340 to n4S i average 489.0). Thus the average number of cells per section DIET AND GASTRIC EPITHELIAL CELLS 205 for the satisfactory diet was 1.13 times that for the oat diet; or the oat diet, 88.5 per cent that of the satisfactory. A comparison of the number of cells per unit area is difficult. No two caeca are exactly the same shape nor are they provided with villi and crypts to exactly the same extent. However, a ratio between the average diameter of the caecum and the number of cells yields a rough comparative figure. For the satisfactory diet this ratio ranges from 1.34 to 2.38 for the various caeca, with an average of 1.79. For the oat diet the range is 1.36 to 2.02 with the average 1.82. Thus the aver- age ratio for the satisfactory diet is 98.4 per cent that of the oat diet; for the oat diet, 1.02 times that for the satisfactory. TABLE I Inclusions Percentage of Cells Ratio, Satisfactory: Ratio, Oat: per Section Cells with Inclusions Oat Diet Satisfactory Diet per Sec- Gran- ules Vacu- oles tion Gran- ules Vacu- oles Ei- ther Gran- ules Vacu- oles Both Gran- ules Vacu- oles Both Satisfactory diet Minimum 12 29 Average 35.8 43.5 552.7 6.5 7.9 14.3 1.35 2.93 1.86 Maximum 83 78 Oat diet Minimum 12 8 Average 23.5 13.2 489.0 4.8 2.7 7.7 .738 .342 .538 Maximum 30 30 Rate of Replacement of Cells Characteristic of insect mid-gut epithelium are the nidi from which new secreting cells for replacement proliferate by mitosis. The number of mitoses visible at any one time gives an indication of the rate of replacement. The satisfactory diet showed 4 to 8 mitoses per section, with an average of 6.4. The oat diet showed 8 to 14 with an average of 11.6. Thus the rate of replacement for the satisfactory diet was 55.2 per cent that of the oat diet; or the oat diet, 1.8 times that for the satisfactory. Secretory Products By the technique used here the secreting epithelial cells show two types of inclusions : darkly-staining granules and clear vacuoles. They 206 CHARLES HODGE, 4TH originate just distad to the tip of the nucleus and are progressively larger as they are found successively nearer the distal end of the cell. I in- terpret them as some of the gastric secretory products (Hodge, 1936). The numbers of these inclusions are given in Table I. Thus the satis- factory diet showed 1.35 times as many granules, 2.93 times as many vacuoles, or 1.86 times as many inclusions, as the oat diet. The oat diet showed 74 per cent, 34 per cent, and 54 per cent, respectively, as many as the satisfactory. Dimensions of Cells The average values in microns and cubic microns for the dimensions measured: length, width and volume of cell and nucleus, and width of striated border, are tabulated below. The average for all the cells on the satisfactory diet differed only very slightly from that for all the cells on the oat diet. The greatest variation is seen in nuclear volume and the length of the filaments of the. striated border. In general the cells and their nuclei on the satisfactory diet are slightly longer and more slender than those on the oat diet and the filaments of the striated border are shorter : Diet Cell Length ( ell Width Cell Volume Nuclear l.< ugth Nuclear Width Nuclear Volume Striated Border Niirli'n- phi-iiiiii- Ratio Satisfactory. . Oat ' . 61.6 58.3 12.2 12 6 9484.7 9450 7 16.9 16.3 8.1 8.6 580.7 663.0 * 5.0 5.6 1 : !(>..* 1 : 14.3 In view of the great variation between minimum and maximum val- ues for these various dimensions (see the more extended data in Table 11 ) these difterenco are hardly significant. More striking differences are observed in comparing the dimensions of cells located on the villus and crypt areas respectively. On the satisfactory diet the cells vary greatly in relation to their location on these two areas. In general on the villi the cells are shorter and wider, the striated borders narrower. and the nuclei more slender than in the crypts. On the oat diet the cells of the two areas are more similar. This may be summarised as in Table II. DISCUSSION The diet of oat leaves alone is undoubtedly unsatisfactory for this specie-,. Tliis has been shown by rate of growth ( I lodge, 1933) and DIET AND GASTRIC EPITHELIAL CELLS 207 by abnormalities in cytology of the germ cells (C. E. McClung, pcr- sonal communication, unpublished). Tins study shows that the rate of replacement of the cax*al epithelial cells on the oat diet is almost double that on the satisfactory. Replace- ment is more or less continuous in insect mid-gut epithelium. Appar- TABLE II Satisfactory Diet Oat Diet Villus Crypt Ratio Villas : Crypt Villus Crypt Ratio Villus : Crypt Cell length Minimum 29 55.0 95 6 12.4 18 2245 8891.4 26741 11 16.9 23 5 6.7 11 171.5 364.4 838.3 2 3.7 6 1 : 24.4 44 67.0 99 8 12.0 17 3213 10077.9 22289 11 16.9 26 6 9.0 14 367.6 796.9 2060.5 3 6.3 12 1 : 12.6 .8 : 1 .8 : 1 1.0 : 1 .9 : 1 1.0 : 1 .7 : 1 .5 : 1 .6 : 1 .5 : 1 40 56.7 88.1 6 13.2 21 2069 10012.7 23286 9 16.6 26 5 8.1 13 111.9 561.9 1372.2 3 4.9 8 1 : 17.8 37 59.8 84 6 12.0 18 2829 8902.3 23621 11 16.0 22 5 9.1 16 252.2 762.0 1658.9 4 6.4 11 1 : 11.7 .9 : 1 .9 : 1 1.1 : 1 1.1 : 1 1.0 : 1 .9 : 1 .7 : 1 .8 : 1 .7 : 1 Average Maximum Cell width Minimum Average Maximum Cell volume Minimum Average Maximum Nuclear length Minimum Average Maximum Nuclear width Alinimum . . Average Maximum Nuclear volume Minimum Average Maximum Striated border Minimum . ... Average Maximum Nucleo-plasmic ratio. . . . ently it is necessitated by the exhaustion of holocrine secretory cells, the gradual wearing out of merocrine cells, or other factors. On the oat diet the number of the large holocrine type of secretory inclusions demonstrated by the technique used here is less than on the satisfactory diet. If the greater replacement results from overstrenuous secretion, it must be of the merocrine type demonstrated by Woodruff (1933) and others. 208 CHARLES HODGK. 4TII On the satisfactory diet the more actively secreting cells of the villi arc conspicuously smaller than the crypt cells. From this one is led to the generalization that among these cells the smaller type is the secreting type. On the oat diet the cells, though shorter in length of nucleus and cell body, are yet for all other dimensions as large or larger than those from the satisfactory diet, and in each case show less difference between the villus and crypt areas. On the whole, therefore, the villus and crypt cells on the oat diet are more similar morphologically than the villus and crypt cells on the satisfactory diet. Furthermore, their general morphology approaches more nearly to that of the larger, less actively secreting crypt cells. If morphology be taken as a criterion of secretory function the oat cells are not secreting more strenuously, but less. Other factors that suggest themselves as reasons for wear and re- placement are the results of toxic influences, and inanition. Toxic ma- terials resulting from the digestion of the oat leaves have not been dem- onstrated directly, but the high mortality among grasshoppers on the • .at diet (Hodge, 1933) may be due to toxicity. The fact that the cells on the oat diet are provided with longer filaments of the striated border is also suggestive. Longer filaments might conceivably have a distinct protective function, especially since these can be demonstrated to be covered with a mucous layer in the living condition (Hodge, 1936). Inanition in other material (Miller, 1927, on rats; Sun, 1927, on mice) produces loss of digestive epithelium. This desquamation, in animals so dissimilar on the part of a tissue not normally so adapted, cannot be too closely analogized to the replacement in the grasshopper. Hut it may well be that the greater replacement on the oat diet mav be due to at least two factors — toxic effect of the diet and a partial inanition. The two types of large secretory inclusions demonstrated by the technique used here — namelv, granules and vacuoles — are shown in no- ticeably larger number on the satisfactory diet than on the oat diet. One and one-third times as many granules, and almost three times as many vacuoles were counted on the satisfactory diet as on the oat diet. The nature and function of these two types of inclusions are as yet uncertain. If they represent two distinct types of cn/ymc material, it would seem that the oat diet hinders the formation of the vacuolar type more than that of the granular type. However, in each case either ne dietary component necessary for their elaboration is relatively de- lia'eiit in the nutrient materials the grasshopper can obtain from the oat leaves or an injurious material either directly or indirectly obtained DIET AND GASTRIC EPITHKLTAL CELLS 209 from the same source inhibits their elaboration. This difference may he the result of a generalized partial inanition, but in other material in- anition leads to vacuolization rather than the reverse (d'Ancona, 1927; Tirelli, 1928; Miller, 1927). The difference in dimensions between the crypt and villus cells on the satisfactory diet is understandable from the fact that the villi are the more actively secretory (Woodruff, 1933). In the case of each di- mension studied except cell width and nuclear length the cells of the villi give much the smaller figure. The corresponding dimensions for the oat diet, while in most cases smaller for the villus than for the crypt cells, do not show such wide divergence. For some dimensions, in fact, the average value for the villus cells is greater than for the crypt cells. Thus on the oat diet the villus cells in size are more like the crypt cells which secrete less ; and from this morphological criterion are less differentiated for the normal function of villus cells, — namely, secretion. The slightly higher nucleo-plasmic ratio on the oat diet suggests an inanition. Such an increased nucleo-plasmic ratio has been reported in other material during inanition (d'Ancona, 1927, on eels ; Kremer, 1932. on frogs; Truszkowski, 1927, on dogs, and 1928. on frogs). A comparison of the relation of the nucleo-plasmic ratio in the crypt areas with that in the villi shows that on the satisfactory diet the ratio in the crypts is twice that in the villi, while on the oat diet it is only 1.5 times as great. This suggests again the conclusion borne out by the dimensions of the cells, that there is less difference between the types of activities of the cells in these two areas respectively on the oat diet than on the satisfactory. The difference in length of the filaments of the striated border is less readily explicable. The exact purpose of the striated border is not yet well understood. The secretory products are discharged through this border. This is effected by total rupture of the border or by a tem- porary separation of the filaments into clumps or pencils, depending on the nature of the secretory operation at the moment. The shortness of the filaments on the actively secreting villi on the satisfactory diet would seem to be an adaptation to this function, to permit easy passage of the secretion. If the striated border subserves any active part in secretion it would seem from the length on villus and crypt respectively on the satisfactory diet that shortness of filaments is in some way correlated with active secretion. From this point of view the greater length of the filaments on the oat diet, correlated with the lesser number of inclusions, points to decreased secretion on this diet. 210 CHARLES HODGE, 4TH The fact that on the oat diet the filaments on the villi are more nearly the length of those in the crypts suggests again a partial relin- (|uishment of the active secretory function of the villus. This area seems to have assumed a more generalized condition more nearly re- sembling the crypts. If the striated border serves more particularly as a protective sur- face, the increased length of the filaments on the oat diet would point to a greater need for .such protection. Some product of injurious na- ture from the wholly or partially digested oat leaves is indicated. The shorter filaments shown on the satisfactory diet would indicate that such products are not liberated in the digestion of the wheat or other satisfactory food plants. At least three factors are indicated by this study as contributing to the various effects of the unsatisfactory oat diet. Inanition is evidenced by the greater rate of replacement and nucleo-plasmic ratio on the oat diet. The retarded growth observed in rearing the animals (Hod^c. 1''33) is further evidence of this. Injury by or toxicity of some prod- uct of digestion or component of the oat leaves is also suggested by the increased rate of replacement. Perhaps the longer filaments of the striated border are also a response to some such factor, as a protective adaptation. The mortality observed in rearing the insects (Hodge, 1933) points toward this as well. Loss of secretory ability on the part of the cells on the oat diet, and especially in the villi, is evidenced by the lesser number of secretory inclusions, as well as by the approx- imation in dimensions of the cells of the villi to a size nearer that of the cells of the crypts. LITERATURE CITED D'ANCONA, U., 1927. Studi still' inanizionc. I. L'azione del lungo digiuno sulle cellule e sin tessuti. ./;;;. Jour. Anal.. 39: 135. HOWIE, C., 4fH, 1933. Growth and nutrition of Melanoplus differcntialis Thomas (Orthoptera: Acridid;i'). I. Growth on a satisfactory mixed diet and on diets of single food plants. Physiol. Zool., 6: 306. lldiM.i, C., 4m, 1936. The anatomy and histology of the alimentary tract of the vi asshoppcr, Melanoplus differmtialis Thomas. Jour. Morph., 59: 423. KRKMER. J., 1932. Die fortlaufmdm Veriindcrungrn drr Amphihicnlebcr im Hungerzustande. Zcitschr. Mikrosk. Anat., 28: 81. Mm. i i', S., 1927. Effect of inanition on the stomach and intestines of albino rats underfed from birth for various periods. Arch. Path. at\d Lab. Mcd., 3: 26. SMITH, A. II., 1929. The relation of diet to changes in tissues. Yale Jour. Biol. and Mcd., 1: 145. ., T. P., 1927. Histophysiological study of the epithelial changes in the small iiitt-stinc of the albino mouse after starvation and refeeding. Anat. Rec., 34: 341. DIET AND GASTRIC EPITHELIAL CELLS 211 TIRELLI, M., 1928. Modificazioni del condriona e del lacunoma nelle cellule in- testinali di " Gambusia holbrooki " durante le diverse fasi dell'attivita funzionale e durante il digiuno. Atti R. Accad. Nat. Lincei, 7: 255. TRUSZKOWSKI, R., 1927. Studies in puriue metabolism. IV. The nuclear-plasmic ratio in dogs in carbohydrate and protein feeding and in starvation. Biochem. Jour., 21: 1047. TRUS/KOWSKI, R., 1928. Studies in purine metabolism. V. The nuclear-plasmic ratio of frogs. Biochem. Jour., 22: 1°li., 55: 53. LICNOPHORA LYNGBYCOLA, A NEW SPECIES OF INFUSORIAX FROM WOODS HOLE K. FAURE-FREMIET (From the Marine lni>l<> 50 per cent tusi I and second (II) cleavage with dense and sparse populations of Arbacia eggs. No. of Experiments Method Mean Difference in Miiiuir- Statistical Probability I 11 I II 6 Paraffin 0.93 2.17 0.3* 0.129* 8 Glass slides L.53 2.2 0.068* 0.036 15 21 23 I landing drops Moist chamber -0.57 0.54 -0.24 2.7 0.215* 0.1 ()4* 0.84* 0.002 17 18 10 12-13 Do. 1 10 per cent Do. 90 per cent Do. connected drops 1.9 -0.03 2.03 3.16 2.17 2.95 0.0084 0.67* 0.0036 0.0002 0.0134 0.0014 True means and combined significance 0.88 2.23 0.053 0.0016 * Results taken alone are not statistically significant. (0.03 minutes, probability 0.67). In two other sits of experiments this time difference increased to 1.9 and 2.03 minutes earlier cleavage for the more crowded drops with statistical probabilities of 0.0084 and 0.0036 respectively. The fourth comparison made (line 4, Table II) gave positive results which, however, lack statistical significance for first cleavage although there is good significance for the second cleavage. The use of two connected drops, one with a den.se and the other with a sparse population <>f eg^s, sometimes permitted the observers to determine the time to 50 per cent cleavage in different regions with different densities. Two such determinations (in which Allee was observing and I -Ivans was recording) are summarized in Fig. 1. Averages of the time to 50 per cent first or seeond cleavage for the experiments suminai i/ed in Table II have lessened value because of the varieties of techniques used. The mean differences are based on individual paired experiments in any one of which the difference in density of the egg population was the only factor known to vary. If the true mean, considering the number of individual tests, is taken, there is a mean difference in time to 50 per cent first cleavage of 0.88 minutes SOME EFFECTS OF NUMBERS ON CLEAVAGE RATE 225 with tlu- eggs in tin- denser populations cleaving first. This has a sta- tistical probability ot 0.053 which is slightly above the conventional limit of statistical significance. l>\ .second cleavage, this difference has increased to a mean of 2.23 minutes and the statistical probability has increased to 0.0016. nrst Second % cleaved 55.92 85.83 too 56.17 86.58 100 5mm. B First Second % cleaved FIG. 1. Diagrams of the connected drops used in two different experiments. Each drop contained 20 cu. mm. The measurements give distances between points indicated by the double arrows. Figures below the diagrams, unless otherwise indicated, give time in minutes to the designated cleavage. The values just given include those with hanging drops where the relations are obviously different from those obtaining in the other types of experimentation. If the hanging-drop experiments are omitted, these true means become 1.17 and 2.68 minutes with "p" values of 0.016 and 0.0000 respectively. Since these hanging-drop experiments do not support the evidence from the rest of the work to date and since 226 \V. C. ALLEE AND G. EVANS we do not know why this is the case, it is probably better to include these in the summaries. At least this is the conservative procedure and is in line with the general practise in these experiments of weighting the results against the experimental trend whenever it was impossible to be absolutely impartial. Thus far in this account we have been dealing mainly with time rela- tions with regard to 50 per cent cleavage. It is profitable to examine the assembled data from another approach which may be indicated by the following question : 1 low many cases were there in which there was evidence of a shorter time elapsing between fertilization and 50 per cent cleavage in the denser population of eggs and in how many cases TABLE III Relations between first and second cleavages. Without With Hanging II. tuning Kviilence of Stimulation in Denser Population (D) Drops Drops 1 . I ) cleaved first in both cleavages 46 48 2. First cleavage a tie; D faster in second 24 27 3. D slower in I ; faster in 1 1 5 5 4. Sparse faster in I; second even. . . 5 5 Totals 80 85 Kvidence of Retardation in Denser Population 5. S cleaved first in both cleavages 5 8 6. First cleavage a tie; S faster in 1 1 0 5 7. D faster in I ; S faster in II . . 0 0 8. S slower in I ; second, even 1 2 6 15 9. No observed difference. 13 17 Totals 1') 32 did the opposite hold true? The answers are given in Table 111 prop- erly subdivided among the different possibilities. In the preparation of this table, unless there was a clear difference between the time to 50 per cent cleavage Mich that the difference could not be the result of error in observation, the case was recorded as showing no difference. Table III shows correctly that our data clearly indicate speedier cleavage in the more densely populated drops. These experiments taken together demonstrate fairly conclusively that there is a difference in time to 50 per cent ol first and second cleav- age for .Irhacia eggs in relatively sparse and in relatively dense popula- tion. The data collected arc clearly statistically significant on this point SOME EFFECTS OF NUMBERS ON CLEAVAGE RATE 227 for the second cleavage, less so for the first. It is necessary to inquire into the degree of crowding which will bring about this result. Observa- tions on this point were made from time to time during the progress of the experiments. Those obtained later in 1934, using the paired drop technique with one drop densely crowded with eggs closely associated or even actually directly connected by a narrow isthmus with a similar drop which contained but few eggs, are summarized in Table IV. TABLE IV Numbers present with relation to the crowding effect. Difference in Fifty Per Cent Probability Size Eggs No. Experi- Dense Sparse Cleavage Drops Fol- ments No. eggs No. eggs Cu. lowed Mm. Sparse I II I it 12-13 1600 + 800- 2.04 2.95 0.0036 0.0014 10-40 10-30 15 65-164 5-24 0.23 1.26 0.7 0.045 20 5-24 14 22-56 5-18 0.08 -0.3 0.38 0.426 20 5-18 In this table the difference in time to 50 per cent first and second cleavages are compared for crowded drops containing over 1,600 eggs as compared with that in more sparsely populated drops containing 800 eggs or less. In reality the listing of these hundreds of eggs in the sparser populations is hardly fair for in all cases the eggs watched in such drops were out of contact with each other while those in the densely populated drops were closely crowded. Under these conditions (which have already been discussed in some detail) the eggs in the more crowded drops cleaved significantly earlier than did those in the accompanying sparsely populated drops of equal size. When the more crowded drop contained but 65-164 in a 20 cu. mm. drop and the accompanying drop held 5-24, the difference in time to first cleavage was insignificant but the denser eggs cleaved slightly sooner at second cleavage and the difference of 1.26 minutes is just within the upper limit for statistical significance. When the population in the more dense drop is still further reduced and that in the sparsely populated drop remains about the same, the differ- ence disappears for both cleavages. Factors Known to Retard Cleavage The factors known to retard cleavage which may have operated in one or more of these experiments are : 1. Hypertonic sea water. 2. Lowered temperature. \V. C. ALLEE AND G. EVAXS 3. Contamination with coelomic fluid. 4. Contamination with fragmented eggs. 5. Increased amount of metabolic wastes in the water. Of these, there was no opportunity for differential contamination with ccelomic fluid in the different parts of a given test. The precau- tions taken against hypertonicity have already been outlined and while they were not always successful when differential treatment was un- avoidable as in the time of Betting up of the various drops, especial care was taken to load the experiment against the trend of experimental findings. As regards temperature differences, when one observer examined the one or two slides holding dense populations and the other handled the half dozen or more with sparser ones, there was the possibility of the former having a slightly higher temperature from the more steady con- tact with the fingers of the observer. This possibility was eliminated in the tests with the assembled moist chambers which constitute the ma- jority of the experiments reported here. Furthermore, these hitter ex- piTiments yielded greater and more consistent differences than were found under conditions that might have been suspect, hence there is no evidence that the observed differences are the result of an externally induced temperature differential. There is, however, a possibility that the high rate of oxidation of the larger number of newly fertilized eggs confined in a small space may produce a temperature differential sufficient to account for the observed results. No tests of this possibility have been made to date. The best evidence that can be cited in its favor is that a slight differential increase in temperature would produce the difference1 which we have found. The isolated eggs were definitely favored as regards the presence of fragmented or immature eggs. Allee found it psychologically difficult to select other than good eggs for the relative isolation of the sparse lots while in scooping up from 50 to 8,000 eggs for a more densely popu- lated drop, no such selection \\;is possible. TTence the sparsely popu- lated drops contained less debris both absolutely and proportionally than did those with the large number of eggs. Overcrowding. — In the most densely populated drops used in the regular experiments, eggs were present in about the proportion of 0.5 cc. of centrifugcd eggs (lowest speed with power centrifuge for one min- ute) to six cc. of water. In Syracuse watch-glasses with this concen- tration there was reduced cleavage. Tn drops of 10 or 20 cu. mm., the eggs are nearer to the surface of the drops and while they may be piled three or four deep in the center, cleavage goes almost as well as in the SOME EFFECTS OF NUMBERS ON CLEAVAGE RATE 229 finger-bowl controls. When the same concentration is placed in a watch- glass or finger-howl, i.e. in the same concentration when the eggs are all stirred up, the eggs settle to the bottom in a much more dense mass than aii}r we worked with on the slides and under these conditions they were definitely over-crowded. Mass Protection. — In an attempt to avoid the added evaporation incident to changing clouded covers, a microscope stage-cooling device was developed. This consisted of a brass quadrangle about the length of the microscope stage and somewhat narrower and about three centi- meters high. Glass tops and bottoms were sealed on with DeKotinsky cement. The upper glass plate had patches of etched lines to facilitate counting and accommodated two of the 3 cm. glass rings which were sealed to the glass with vaseline. Intake and drain tubes were soldered to the brass quadrangle and were attached to the sea water supply since this was the coolest water available. This was further cooled by run- ning it through a copper coil placed in a bucket of ice water. The drops were placed on the glass as usual except that during isolation, a covering watch-glass was used for a partial moist chamber. After covering the drops with the usual vaseline-sealed, glass cover, the water was turned through the chamber and the temperature fell. With the room tem- perature at 24°, in one instance, water emerging from the cooling device showed 12° ; it was usually held at from 17-19°. This device was dis- carded when it appeared that there was difficulty in proper cleaning of the surface to which the eggs were exposed and no data secured by this means have been included in the preceding tables. Some twelve experiments were tried using this gadget. The mean difference to 50 per cent cleavage for those that cleaved was 5.0 minutes for first and 5.62 minutes for second cleavage, with " p " values of 0.024 and 0.0052 respectively. Four accompanying experiments made in the assembled moist chambers already described had differences to 50 per cent first and second cleavages of 0.31 and 2.8 minutes respectively with the latter difference statistically significant. In two cases with the stage- cooling device, sparse populations of 6 and of 10 eggs failed to cleave and in two other experiments sparse lots of 24 and of 34 did not reach 1 5 per cent cleavage. The accompanying densely populated drops de- veloped to or beyond the blastula stage and in three of the four com- parisons just made, many in the more crowded lots were actively swim- ming after 24 hours. These data strongly suggest the presence of some toxic contaminating substance which was not completely removed by the methods used in washing these stage-cooling devices. Further tests showed that in connected drops gradients of resistance could be demon- strated which depended on the numbers present, the more eggs within 230 \y. C. ALLEE AXD G. EVANS the limits tested, the greater the percentage of cleavage and the further development would proceed before death. Such mass relations appear to be closely related to the mass protection from toxic materials which has been repeatedly demonstrated (cf. Allee, 1931, 1934). Could this have been the explanation of the more rapid cleavage oil- served in the denser populations? There is a suggestion that it may have been a factor in the paraffined slides, which, however, is not borne out by comparative studies unless there was some other toxic agent act- ing similarly in the remainder of the experiments. To suppose that there was mass protection from toxic materials in the other experiments would imply either some sort of toxic emanations from the glass itself or from some chemical previously in contact with it (Richards, 1936), or that traces of the vaseline and /or vacuum grease remained over from the washing and were poisonous, or that some of the soap from the mild sud> UM-d remained after the extensive and careful rinsing. There is no evidence for the presence of toxic materials from any of these sources all of which were considered as possible means of experimental error before this set of experiments was begun. Direct tests made by using water which had stood over masses of broken glass showed no differ- ence in development as compared with similar cultures in ordinary sea water. Direct tests for toxic effects from vaseline and from the vacuum grease used made by coating slides with these substances and placing drops with different numbers of eggs on them, yielded no evidence of contamination from this source. Hence we concluded (and later ex- periments to be reported in another paper justify the conclusion) that the differential results obtained are not produced by mass protection >uch as was demonstrated to be operating in the experiments made with the stage-cooling device. There is also internal evidence from the experimental results re- ported in Table II that something more is happening in these experi- ments than would be expected from the simplest of mass relations whether mass protection or otherwise. The data given there show that the mean time to 50 per cent first cleavage for all the experiments was 0.88 minutes (omitting hanging-drop experiments. 1.17 minutes) and to second cleavage was 2.23 minutes (omitting hanging drops, 2.68 min- utes). The mean time to 50 per cent first cleavage was 57 minutes and to the same stage- in second cleavage was 85 minutes. I f this were a case of simple mass relationships one would expect the acceleration of 0.88 minutes during a period of 57 minutes to continue at the same rate during the following 28 minutes to second cleavage. At this rate the- total acceleration would approximate 1.31 minutes, SOME EFFECTS OF NUMBERS ON CLEAVAGE RATE 231 which is not the case. Even when one corrects for the fact that for the first ten minutes after fertilization, on the average, the eggs which are to make up the isolated or sparse populations are in fairly dense lots, the expected acceleration on the basis of uniform, simple mass ac- tion would be 1.40 which even yet is far from the observed value. Even omitting the hanging-drop experiments, which we decided above not to do, and using the basis of calculation which will give highest re- sults, simple mass action would call for an acceleration by second cleav- age of 1.87 against the observed value of 2.68 minutes. These rela- tionships indicate that the observed phenomena are not based directly on the simplest sort of mass physiology. Other possible causal factors which deserve investigation include the effect, if any, of increased carbon dioxide, of supernatant water from eggs both before and after cleavage, and of mitogenetic rays. SUMMARY 1 . Other conditions being equal and under a variety of experimental conditions, eggs of Arbacia punctulata cleave more rapidly when in relatively dense as opposed to relatively sparse populations. The de- creased time to first cleavage in the dense populations was 0.88 minutes and to second cleavage was 2.23 minutes. The first difference is prob- ably not statistically significant ; the second value is clearly significant. 2. Among other conditions, these relations were observed when some thousands of eggs in a drop of 20 cu. mm. of sea water were connected by a narrow strait with a similar drop holding some few tens or even a few hundreds of eggs. 3. If the eggs were crowded together too densely, the time to first and second divisions was definitely retarded and the percentage of final cleavage was reduced. 4. When 22 to 56 eggs were placed in one drop of 20 cu. mm. con- nected by a strait with another containing 5 to 18 eggs, no difference in cleavage rate was observed. 5. The observed differences are not a result of differential tempera- tures externally imposed, differential hypertonicity or hypotonicity, con- taminations with ccelomic fluid or with fragmented eggs. 6. Such results may be obtained by mass protection from toxic ma- terials. There is, however, no indication that differences here reported were so caused. 7. Although no supporting evidence is presented here, the results may conceivably have been the result of differential temperatures pro- duced by the high rate of oxidation of the massed eggs in a small space, 232 \V. C. ALLEE AND G. EVANS by chemical emanations from the eggs, including carbon dioxide, by stimulation from mutual contact or by mitogenetic rays. Discussion of these problems is reserved for the present. 8. In addition to their intrinsic interest, the results provide another instance of physiological activities \\hich ])roceed more rapidly at an in- termediate optimum than when either too few or too many are present. REFERENCES ALLEE, W. C., 1931. Animal Aggregations. University of Chicago Press. ALLEE, W. C. 1934. liiol. Rev., 9: 1. FRANK, G. AND M. KTHM-INA, 1930. Arch. Untie, d. On/.. 121: 634. JUST, E. E., 1928. The <.',>llcctinf/ AY/. 3. RICHARDS, O. W., 1930. Physiol. Zobl, 9: 246. FEEDING RATE OF CALANUS FINMARCHICUS IN RELATION TO ENVIRONMENTAL CONDITIONS1 JOHN L. FULLER (From the Woods Hole. Oceanographic Institution and the Biological Laboratories, Clark University) OBJECTIVES The amount of food available to a plankton-feeding animal is deter- mined by the concentration of suitable food material in the water, and by the rate of the animal's feeding activity. The calanoid copepods, like many other zooplanktonts, are generally considered to feed by filtering out particles from a current of water generated by the animals (Cannon, 1928). By determining the number of food organisms re- moved by an animal from a suspension of known concentration, the volume of water which has been filtered clear of the organisms can be calculated and, if the chemical composition of the organisms be known, the amount of nutriment made available to the animal may be esti- mated. Preliminary measurements of the filtering rate of Calanus finmarchicus have been reported by Fuller and Clarke (1936). Further investigation of changes in the feeding rate induced by varying environ- mental factors of ecological importance was expected to yield informa- tion useful in the quantitative study of aquatic food cycles. Diatom concentration, light and temperature were expected to be variable factors in the sea and were chosen for study. METHODS In the experiments here reported the rate of feeding of the copepod Calanus finmarchicus, collected in Vineyard Sound near the whistle buoy, was measured in a suspension of the diatom Nitzschia closterium (Plymouth strain). It seems probable that Nitzschia is a suitable food for Calanus, as Crawshay (1913-15) kept individual copepods alive for as long as 80 days in persistent cultures of this diatom. Allen and Nelson (1910, p. 470) reared Calanus from eggs to copepodid stages in a mixed culture in which Nitzschia was predominant. Three stage V Calanus were placed in 15 cc. of sea water containing a known concentration of diatoms. Changes in concentration were followed for two to four days, counts being made in a hemacytometer. 1 Contribution No. 132 of the Woods Hole Occanographic Institution. 233 234 JOHN L. FULLER A series of counts on successive samples showed fair agreement with the theoretical Poisson distribution. The small excess variation found is attributed to the difficulty of shaking the diatoms to obtain even distribution without injuring the copepods. Enough diatoms were counted for each sample to L;ive a result statistically valid to between 5 and 10 per cent. Under the experimental conditions this is the maxi- mum precision obtainable. In these experiments a counting error of 10 per cent causes a discrepancy of about 30 per cent in the estimation of the filtering rate. Obviously only large variations can be satis- factorily studied. TABLE I Feeding rate at different diatom concentrations Cells per cc. Expt. No. Date begun Duration Wu &Nu Ci Ci hoars cc. 47 Aug. 18 1,875 1,140 42 1.32 2,100 48 Aug. 18 1,875 1,155 42 1.28 2,060 63 Ann. 23 19,000 15,400 38 .66 1 1 ,400 64 Aug. 23 19,000 13,100 38 1.17 16,100 14 July 9 64,700 27,000 81 1.30 55,000 15 July 9 64,700 27,700 81 1.25 55,000 71 Aug. 26 190,000 147,000 38 .81 136,000 72 Aug. 26 200,000 133,000 38 1.29 212,000 11 July 3 335,000 86,000 84 1.61 356,000 10 July 3 375,000 220,000 84 .43 221,000 8 June 29 390,000 130,000 84 1.32 371,000 7 June 2(> 410,000 230.000 84 .60 256,000 Average 1 .09 Calculation of the amount of water, Wx, swept free of diatoms by each copepod in .v hours was made by means of an equation derived as follows: (1) (2) (3) (4) dN = -dWC dC ---- dN IV dW VdCjC Wx = Tin d/C2. N represents the number of diatoms per copepod; V the volume of water per copepod; C\ and C« respectively the concentrations of diatoms at the beginning and end ol the period of .v hours. The number of diatoms eaten in x hours per copepod, AA^, is given FEEDING KATK OK CAI.ANUS AND ENVIRONMENT 235 by the equation: (5) &NZ = V(Ci - C2). Parallel control suspensions were always counted to determine whether the diatom numbers were varying independently. If a con- trol changed greatly during the experimental period, the results of the corresponding experimental series were discarded. If the change was small, the final concentration of diatoms in the control was substituted for C\ in equations (4) and (5). This substitution assumes that the diatoms in the experimental containers increased or decreased inde- pendently exactly as those in the control. This is, of course, only an approximation. Observations were also made on the production of fecal pellets. Since these pellets vary greatly in size no quantitative relationship could be obtained between the number of diatoms ingested and the number of pellets ejected. However, the formation of these excreta was a useful check on the reduction of diatom numbers observed directly. EFFECT OF CHANGING DIATOM CONCENTRATION Table I gives the results of experiments with different concentra- tions of diatoms. In the lowest concentration the diatoms were con- centrated by the Nielsen-von Brand (1934) method before counting. These experiments were carried out at 13° C., and the animals were shielded from direct sunlight. The values for Wu and N^ in the table represent means over a period of several days. Diatom con- centration often remains stationary for a day or more, thus indicating that feeding is not a continuous process. The maximum rate of filtra- tion is higher than these values. The highest rate observed was in Experiment 11 where over a period of 14 hours W equalled 2.87 cc. which would give a Wu value of 4.9 cc. There appears to be no cor- relation between the concentration of food and the filtering rate. A comparison of the filtering rates of animals collected at different times during the summer yields no greater differences than a comparison of animals collected at the same time and used in parallel experiments (e.g. Nos. 10 and 11). The indication is that Calanus, though it some- times does not feed actively, under otherwise constant conditions filters a definite volume of water per day, and obtains nutriment in direct proportion to the concentration of food particles. EFFECT OF LIGHT ON FEEDING RATE In certain experiments there appeared to be a diurnal feeding rhythm, removal of diatoms taking place most rapidly at night. This 236 JOHN L. FULLER is in accord with the observation of Marshall (1924), who notes that Calanus captured by tow-net hauls in the early morning had full gut?, while those caught later in the day often had no food in their ali- mentary tracts. Figure 1 represents the course of t\vo experiments (8 and 1 1 j which show this rhythm. It is not, however, an invariable KK;. 1. Diurnal feeding rhythm of Calanns. Stippled areas represent hours between 8 P.M. and 8 A.M. The scale of Kxperiment 15 i> adjusted by adding 0.5 to logarithms of cell numbers. Ordinates: Logarithm-cells per 0.1 en mm. Abcissae: Time in days. effect, as is shown by the course of Kxperiment 15, also shown in Fig. 1 . Marshall likewise found that during the summer feeding occurred during the day as well as at night. Kxperiments 26 and 32 differed from those reported above in being kept in darkness. Table II summarizes the results of these two experiments. FEEDING RATE OF CALANUS AND ENVIRONMENT 237 TABLE II Feeding rate in dark ( Vlls per cc. 0 Expt. No. Date begun Duration WH NM Ci C2 hours cc. 26 lulv 16 54,000 32,000 90 .27 45,000 48* .76* 32 July 30 61,000 20,000 65 1.8 76,000 * Animals actually fed only during last 48 hours of experiment. lated for this period. value calcu- The results do not differ significantly from the average given in Table I, although there are admittedly too few cases to permit a definite answer. An attempt was made to study the effect of con- tinuous light. The animals did not feed during the two days they were observed, but they were old stock in poor condition, and no reliance should be placed on this observation. Unfortunately it was impossible to collect more Calanus to make satisfactory experiments. EFFECT OF TEMPERATURE ON FEEDING RATE Experiments were carried out at 8° C. and at approximately 3° C. The latter temperature was not precisely controlled as the containers were placed in a refrigerator in general laboratory use. Both series were kept in the dark. Feeding is much slower at 3° but still goes on with the formation of small compact fecal pellets. The very low filtering rates observed in Experiments 75 to 79 may be explained by the fact that the animals used were collected late in the season when the Calanus population was receding rapidly, and had perhaps stopped eating. The filtering rate in Experiment 80, in which the same stock was used, is seen to be much slower than that in Experiments 55 and 56. The high value for Wu in the latter experiments seems to indicate that 8° C. is somewhat more favorable for feeding than 13°. Excluding Experiments 75 to 80, the average values of W2i at 3°, 8° and 13° are respectively, 0.35, 2.83 and 1.09 cc. ORGANIC NITROGEN CONTENT OF NITZSCHIA AND OF PLANKTON Through the kindness of Dr. Theodor von Brand, an analysis of the organic nitrogen content of Nitzschia was made by the method he has developed for small amounts of plankton (von Brand, 1935). Nitzschia contains 0.9 microgram of nitrogen per million cells. As- suming the carbon-nitrogen ratio in this diatom to be equal to that 238 JOHN L. FULLER generally found in marine phytoplankton, approximately 8:1, \ve have 7.2 micrograms of carbon per million cells. Analysis of ihe fecal pellets of Calanus fed on Nitzschia showed that 0.55 microgram of nitrogen appeared in the feces for each million cells ingested. Roughly, about half the organic nitrogen was retained by the copepods. Marshall, Nicholls and Orr (1935) have calculated from measure- ments of oxygen consumption the daily nutritive requirement of a stage Y Calanus in summer as equal to 13 micrograms of carbohydrate. Assuming all the nitrogen uf Nitzschia to be in protein, and the excess carbon to be in fat, one may calculate the number of cells which would supply an equal amount of energy. If the protein is assumed to contain 16 per cent nitrogen and 50 per cent carbon; the fat to contain 75 per cent carbon. \\e have: TABLE III Effect of temperalure on feeding rate Cells per cc. Expt. No. Temp. Date begun Dura- tion Wtt Ntt Ci C2 °c. hours cc. 41 3° Aug. 18 263,000 215,000 90 .38 64,000 43 3° Aug. 18 263,000 223,000 90 .32 53,000 55 8° Aug. 22 85,500 30,000 42 2.88 159,000 56 8° Aug. 22 85,500 31,000 42 2.78 156,000 80 8° Aug. 31 24,000 14,000 65 .98 18,500 (75, 76) Aug. 31 85 .00 0 (79) 8° 0.9 X 0.16 --- 5.62 7 protein containing 2.8 7 C. 7.2 -- 2.8 4.4 7 Cin fat 4.4 X 0.75 == 5.87 7 fat. IVoteins and carbohydrates have equal energy values per unit of weight. Fat has 2.25 .is great an energy content as carbohydrate. Calculating fat and protein in terms of carbohydrate, we obtain: 5.87 ; : 2.25 = : 13.2 5.6 18.8 7 -weight of carbohydrate equivalent in energy content to a million Nitzschia cells. X 1,000,000 == 690,000 number of Nitzschia cells which would lo.o . . i f j contain theoretical tOOd require- ment. FEEDING RATE OF CALANUS AND ENVIRONMENT 239 The maximum amount of water filtered per day (W-n) in any experiment was 2.88 cc. A concentration of 240,000 cells per cc. would be required if Calanus is to obtain its theoretical food require- ment from this volume of water, even if all the organic matter is utilized. Probably this value should be doubled since only half the nitrogenous organic matter appears to be assimilated. In any case this is a much higher concentration of phytoplankton than ever occurs in nature, but since the size of phytoplankton cells varies so greatly, comparisons based on cell numbers are worthless. It is, however, possible to determine the organic nitrogen content of the particulate matter in sea water, and to compare this with the food requirements of Calanus. Table IV summarizes the results of two sets of analyses by Dr. von Brand of the particulate matter suspended in Vineyard Sound water. All macroscopically visible organisms were removed TABLE IV Organic N in particulate matter in Vineyard Sound water Date Depth Temperature N meters °C. per liter micrograms July 2 0 17.8 15 15 13.5 27 30 9.1 13 August 14 0 20.0 38 15 18.0 42 30 12.0 19 from the August 14 samples before analysis. This was not done for the July 2 samples. The maximum value found was 427N per liter, a nitrogen concen- tration equal to that of a Nitzschia culture containing 46,700 cells per cc. This is so far above the average number of diatoms per cc. in Vineyard Sound — the figure is probably close to 100 — that, even allowing for the fact that one large cell is the equivalent of many Nitzschia, it appears probable that the major portion of the particulate matter at this station was not in the form of diatoms. The remainder, detritus or other organisms, is presumably of great significance as a source of food. Even in this region of high nitrogen content 2.88 cc. of Vineyard Sound water contained less than one-fifth the theoretical food require- ment, and in all probability not over one-tenth this amount since half the nitrogenous matter is believed inassimilable, and some of the particles were probably too large or too small to be captured. Yet Calanus survives and grows in this region. Two explanations are 240 JOHN L. FULLER possible: (1) Calanus in nature requires less food than that calculated from laboratory experiments on oxygen consumption, or (2), Calanus in nature filters a greater volume of water and thus obtains more than the estimated amount of food. NUTRITIVE RK fu] moults 1 VatllS in moult 1 leaths not in moult lVrr<-ntaK<' n-arhini; moult Average survival No. Stage days m $00,000 2 III 0 2 0 63.2 8.9 12 IV 1 8 3 5 V 0 1 5 B2 150,000 11 \\ 1 3 7 33.3 9.7* 4 V 0 1 3 B3 30,000 14 IV (1 5 9 38.9 9.5 4 V 0 2 2 B4 15,000 8 IV 0 3 5 41.1 10.4* 11 V 0 5 5 B5 Of 10 IV 0 1 9 18.7 11.0* 6 V 0 2 3 B6 Harbor 4 IV 0 3 1 60.0 7.2 water 16 V 0 9 7 * In each of these experiments one stage V still living after 19 days, t All attempts to moult near end of experiment. Examination >hc>\\< •n experiments on the keeping of plankton ani- mals under artificial conditions. Jour. Mar. Biol. Ass'n, N.S., 10: 555. FULLER, J. L., AND G. I.. CI.AKKI-:, 1936. Further experiments on the feeding of Calanus finmarchicus. Biol. Bull., 70: 308. LOWXOES, A. G., 1935. The swimming and feeding of certain calanoid copepods. Proc. Zool. Sor. London, p. 687. MARSHALL, S., 1924. The food of Calanus finmarchicus during 1923. Jour. Mar. Biol., Ass'n, 13: 473. MARSHALL, S. M., A. G. NICHOLLS, AND A. P. ORR, 1935. On the biology of Calanus finmarchicus. Fart VI. Oxygen consumption in relation to environ- mental conditions. Jour. Mar. Biol. Ass'n, 20: 1. NIELSEN, E. S., AND T. VON BRAND, 1934. Quantitative Zentrifugenmethoden zur rianktonbestimmung. Conseil Int. pour I'Explor. dc la iner. Rapf>. et Proces verbaux, 89: 99. VON BRAND, T., 1935. Methods for the determination of nitrogen and carbon in small amounts of plankton. Biol. Bull., 69: 221. CYST FORMATION IN THE GLOMERULAR TUFTS OF CERTAIN FISH KIDNEYS ALLAN L. GRAFFLIN 1 (From the Department of Anatomy, Harvard Medical School) In an earlier study (Grafflin, 1933), the paradox, in an old specimen of daddy sculpin (Myoxocephalus scorpins), of a kidney showing many glomeruli anatomically but no glomerular function physiologically was satisfactorily explained. The glomeruli had been rendered incompetent by degenerative changes affecting both the vascular tufts and the neck segments, so that it was practically impossible to find a single glomerulus which on anatomical grounds could be considered functional. In young fish of the same species there was found adequate anatomical basis for the varying, but low, glomerular function which could be demonstrated physiologically. However, considerable degeneration was already pres- ent in the youngest specimens examined, and these changes became steadily more prominent with increasing age (as judged by weight). For a complete discussion of the glomerular changes noted, the original paper should be consulted. Some glomeruli, relatively quite infrequent, exhibited cystic cavities in their vascular tufts, and this is the particular problem which concerns us here. " Probably the most interesting glomeruli in this material are those showing what may be called a central cystic degeneration. In some instances the tuft shows a more or less spherical and very well-delimited cavity which is either entirely free from coagulum or shows it in only small amounts. These clear spaces show a wide variation in size. ... In other tufts the cyst shows rather poor delimitation and considerable amounts of coagulum." (Grafflin, 1933, p. 65.) At that time such cyst formation had not been observed in the glo- meruli of any other species (fish or higher vertebrates). Although the specimens of M. scorpius showed rather numerous parasites in the kid- ney, all of the evidence indicated quite clearly that the parasitism played no role whatsoever in the observed glomerular changes. Under the cir- cumstances, one would reasonably be led to the conclusion that the cysts represented one manifestation of the generalized process of glomerular degeneration. However, the following statement was made : ' The 1 Fellow of the John Simon Guggenheim Memorial Foundation (1934). The specimens of Ophlchthys, Crenilabnis and Cori'ina were collected at the Stazione Zoologica, Naples, Italy. I wish to thank Professor R. Dohrn for his many kind- nesses while I was a guest in his laboratory. 247 248 ALLAN L. GRAFFLIN sharp delimitation of these central cysts in some cases suggests that we may even be dealing here with a malformation of the tuft " (p. 66). In the course of the last several years instances of cyst formation have been found in the glomerular tufts of seven additional species, two of them lungfishes, four of them marine teleosts, and one of them an arid- living reptile (the horned toad). None of these species exhibits the generalized glomerular degeneration characteristic of M. scorpins. Fur- ther discussion of the problem will be deferred until after a description of the findings. LUNGFISHES Protoptcrus (Cthiopicus (African lung fish} In one of the available specimens (No. 21) two striking instances of cyst formation were observed, of widely different sizes and both ex- hibiting a delicate coagulum (Figs. 1 and 2). In another specimen (No. 34) two cysts were likewise observed (Fig. 3). Though closely adjacent, they are in different glomerular tufts, which are located, how- ever, in the same glomerular cluster. In these latter instances the cyst contains much heavier coagulum, including what appears to be cellular debris, and in addition some well-formed cellular elements, presumably inwandering leucocytes. All four of these cysts are sharply delimited, and are lined by a flattened, endothelium-like epithelium (particularly well shown in Fig. 3). Immediately outside of this epithelial lining there is a well-defined basement membrane, which stands out quite clearly in Fig. 3. In comparison with non-cystic glomeruli, there is no increase in cellularity of the tuft, and no detectable abnormality of EXPLANATION FOR PLATE I FIGS. 1 and 2. A large and a small cyst in glomerular tufts of Protopterus (Cthiopicus. Iron hematoxylin and orange G. FIG. 3. Two cysts, side by side, in adjacent glomerular tufts of Protoptcrus (cthiopicus. Heidenhain-azan. X 410. FIG. 4. Parasite in glomerular tuft of Protoptcrus (cthiopicus. Iron hema- toxylin and orange G. FIG. 5. Small cyst in glomerular tuft of Lepidosiren parado.va. Iron hema- toxylin and orange G. FIGS. 6, 7 and 8. Cysts in glomerular tufts of Myoxoccphalus octodccim- spinosus. Heidenhain-azan. FIG. 9. Glomerular tuft of Myoxoccphalns octodccimspinosus, showing exten- sive region of degenerative change, interpreted as prohablc precursor of cyst forma- tion. Heidenhain-azan. All sections 5 M. All microphotographs at X 350 except Fig. 3. All cysts, ex- cept the small one to the left of Fig. 3, photographed at point of greatest cross- sectional area. CYST FORMATION IX ( ,!.( )M KkI'LAR TUFTS 249 \ PLATE I 250 ALLAN L. GRAFFLIN the afferent and efferent vessels or of the peripheral capillaries. The intracapsular spaces are free from coagulum. the ciliated neck segments are normal, and the first portions of the proximal convoluted segments are entirely comparable with those of adjacent nephrons showing no glo- merular abnormality. In Specimen Xo. 21 the presence of a parasite in the glomerular tuft is a not infrequent finding. These parasites always exhibit a characteristic structure, which is shown in Fig. 4. Careful study of the material has failed to unearth any positive evidence that the parasites may ultimately he associated with the appearance of cysts such as those described above, but for the pre>cnt such an association cannot be denied and must IK- leit an open question. - Lepidosiren paradoxa (South American lung fish) In one of the specimens available- a single small cyst was observed ( Fig. 5). It is essentially similar in structure to the cyst shown in Fig. _'. It differs in having a 1< -v> rigidly spherical outline and in containing. in addition to coagulum. a few formed cellular elements. The general statements made above concerning the cvstic glomeruli of Protopterus likewise apply to this instance in I.cpidosircu. Although parasitic re- mains are present in the kidnev. they have- never been observed in the Ldomeruli, and it is concluded that they have nothing to do with {In- formation of the cyst in question. K\l'I.. \\ATKiN FOR Pi. ATI-. II Fn.s. In. 11, 12, l.i and 1-1. Five instances »i cyst formation in glnmerular tufts of Opliichlln's inihrrbis. S n. FlG. 15. Xunnal glnmcrular tuft nf Ophichthys iinhcrhis. without cyst forma- tion, showing large intracan-ulnr space frequently observed in tin- present material. SM. FlGS. l'i and 17. Isolated instances of cyst fnnnatinn in .nlonu-nilar tufts of Crcmlabnis pn; •' ( l-'i.u. 1'p) and Corrimi /m/ni ( I;i». 17). 5 M. l-'n.s. IS and \(>. '\'\\« instances <>\ cyst formation in (jlomcrular tufts of the horned toad, Phrynoxoiini corinttum. In, a. All scrtioiis stainrd with hciuatoxyliii and rosin ; all microphotographs at X 350 ; all cysts idiototiraiilu-d at point i.f t;ri-ati>t cross-sectional area. 2 These si>i'iniMi- were n.lKrtrd in Africa \<\ I'roU-ssor llonu-r \Y. Smith. For the sake of completeness the following data aiv .uivrn. Specimen No. 2\ : col- lected in July, 1928; kept alive in dry estivation from XoNemher 1. 1()28, for 427 days; replaced in water for ten days and killed, as il seemed aliout to die alter the appearance of superficial infection. Spi-cimni Xo. -i4 : collected in July, 1()2S; kept in water until January 10. l('.i(>. lieiug fed regularly; accidentally killed by exposure to cold. From a study of the available material, there is no evidence that a period of estivation has any bearing whatsoever upon the formation of cysts. CYST FORMATIOX IX < ,!.( )M Kk U.AR TUFTS 251 % u * 13 J ri 14 PLATE II 252 ALLAN* L. (iRAKFLIX MAR INK TELEOSTS Myoxocephalus octodecimspinosus (sculpin i As reported in detail elsewhere (Grafrlin, 1937), the glomeruli df this species exhibit a wide variation in size and vascularity, but are in no sense the seat of generalized degenerative processes such as char- acterize the glomeruli of the closely related daddy sculpin (.17. scorf'hts}. Nevertheless, glomernlar cysts have- been found to occur in all of the six specimens carefully examined. In five ot these specimens such cysts are present in only relatively small numbers. In the sixth, in one restricted portion of the kidnev, cystic glomeruli comprise almost one- third of the total glomeruli. Xo reason for their large number in this region is apparent. Three example-, of cvstie glomeruli. to show the range in size encountered, are given in Kig-. <*. 7 and S. The cysts are sharply delimited, are lined bv a markedly flattened epithelium I the nuclei of which are particularly apparent in Kig. 6), and contain coag- uluin of varying density. As in the lunglishes. the intracapMilar spaces are free from coagulum. and the afferent and efferent vessels and the peripheral capillaries are apparently normal, as are the associated neck and proximal convoluted segments. A careful search for some lead as to the mechanism by which these cysts might arise yielded only tin- isolated glomerulus shown in Kig. (), which is nevertheless very striking. While the lower part of the tuft retains essentially normal glomcrnlar organization, the bulk of the tuft has lost it completely, and exhibits for the- most part a coarse, strmgv meshwork containing numerous naked nuclei. It seems reasonable to suppose that the degenerative process, which is here so clearly apparent, might eventually result in the forma- tion of a large cyst of the type shown, for example, in Kig. S. Small unicellular parasites regularly occur in the sculpin kidnev. However, since they are never found in the glomeruli. it is concluded that they play no role in the formation of the cysts. Ophichth \s iinbct'bis In material from the caudal kidney o! the single specimen available for study (Naples — weight 28 grams), live instances of cyst formation have been observed (Kigs. 10 14). In all cases the ciliated neck seg- ment is fully patent and entirely normal in appearance, and the intra- capsular space shows no trace ot coagulum. That the relatively large size of the intracapsular space (sec particularly Fig. 14) bears no essen- tial relationship to the presence ot the cysts is clear from an examina- tion of the normal glomeruli of this specimen, in which this space is frequently nnusuallv large (Kig. 15). The cysts v:irv in si/e. but are CYST FORMATION IN GLOMERULAR TUFTS 253 all sharply delimited and are lined by a flattened epithelium. They contain a considerable amount of coagulum, which has withdrawn to one side of the cavity, presumably in the course of fixation. The capil- laries persisting in the outer rim of the tuft are normal in appearance and fully patent, and the visceral epithelium is identical in thickness and appearance with that of the normal glomeruli. The afferent and efferent vessels are likewise normal. The cysts tend to be spherical in shape, but in one case (Fig. 12) the cyst wall has collapsed, perhaps in the course of fixation. The glomerulus shown in Fig. 12 is to be compared . with one previously described in the daddy sculpin (Grafflin, 1933, Figs. /and 15). In three instances (Figs. 11, 12 and 14) the ciliated neck can be readily traced into the first portion of the proximal convoluted segment, which is entirely normal in appearance. In another case (Fig. 10) the sections do not include the transition. In the fifth instance (Fig. 13) the ciliated neck, approximately lOO/x in length, passes into a curious segment with flattened cuboidal epithelium, which shows neither cilia nor brush border. The cytoplasm is scanty and lightly eosinophilic ; the nuclei are closely packed but show no signs of degeneration. This segment persists for about 250 /JL, at which point it shows a transition to the normal epithelium of the first portion of the proximal convoluted segment. Crenilabrus pavo In material from the caudal kidney of the single specimen available for study (Naples — weight 120 grams), a single instance of cyst forma- tion was observed (Fig. 16). The cyst is spherical and sharply delim- ited, is lined by a flattened epithelium, and contains a moderate amount of coagulum. The peripheral capillaries and the afferent and efferent vessels are entirely normal in appearance, and the visceral epithelium is unthickened. There is a very delicate coagulum in the intracapsular space. The ciliated neck segment, which is very short, is fully patent, and the associated first portion of the proximal convoluted segment shows no abnormality. Corvina nlgra The solitary instance of cyst formation observed in this species (single specimen, caudal kidney; Naples — 575 grams) is illustrated in Fig. 17. The cyst is essentially spherical, is lined by a flattened epi- thelium, and exhibits a rather coarse, stringy coagulum. A few naked nuclei, fairly well preserved, are found scattered at random through the 254 ALLAN L. GRAFFLIN coagulum. The peripheral capillaries, few in number, contain normal red cells, and the afferent and efferent vessels show no detectable ab- normality. The glomerular membrane is in many places appreciably thickened. The neck of the tubule is patent, though the lumen is quite small, and the first portion of the proximal convoluted segment is en- tirely normal in appearance.3 HORNED TOAD ( PHRYXOSOMA CORNUTUM) In surveying the available sections of the kidney of the horned toad, which is an arid-living reptile, two instances of cyst formation, entirely comparable with those observed in the fishes, were found (Figs. 18 and 19). Only a portion of the glomerulus shown in Fig. 18 is present in the sections, and both the afferent and efferent vessels and the neck segment are absent. The cyst is spherical and well-delimited, is lined by a flattened epithelium, and contains a considerable amount of coag- ulum. In the two sections adjacent to the one photographed there is present a dense, irregular, deeply basophilic mass of debris, which oc- cupies perhaps one-quarter of the total cross-sectional area of the cyst. The peripheral capillary loops contain normal red cells, and show no thickening of the glomerular membrane as compared with the normal. There is a distinct coagulum in the intracapsular space. This glo- merulus has formerly been briefly noted by Vilter ( 1935, p. 383). The appearance of the cyst in Fig. 19 is deceptive, due to the manner in which the coagulum has been precipitated. Actually the cyst is sharply delimited from the surrounding tissue, and is lined by flattened epi- thelium. The glomerular membrane is not thickened, and the peripheral capillary loops contain normal blood cells. The afferent and efferent vessels seem entirely normal, and the neck segment is fully patent. DISCUSSION In seeking for an explanation for the formation of the cysts de- scribed above, we are led to the following considerations: (\) All of the evidence indicates quite clearly that the presence of parasites in the kidnev has no bearing upon the formation of cysts, except in the case of one specimen of Protopterns (No. 21). In this specimen, the tendency of the parasites to locate in the glomerular tuft is suggestive. The parasite might become walled off. and. with the sub- ;i In tliis specimen one small, degenerate avascular tuft was observed which showed a central cavity containing basophilic debris. The peripheral rim of tissue was hyalinized and almost structureless, still containing scattered nuclei and nu- clear fragments. This cavity is in no sense typical of the cysts discussed here, and will not be further considered. CYST FORMATION IN GLOMERULAR TUFTS 255 sequent evacuation, or degeneration and absorption, of the organism, cystic cavities of the type observed might persist. However, there is no direct evidence in the material at hand that such is the case. From the available evidence, it is concluded that the glomerular cysts are formed predominantly, or entirely, on some basis other than parasitism of the glomerular tuft. If parasitism can play a causative role, it is a com- pletely separate process and of minor importance for the present problem. (2) It seems almost certain that at least some, of the cysts are formed as the result of a degenerative process in the glomerular tuft. In favor of this view are some of the pictures observed in the kidney of Myoxocephalus scorplns (Grafflin, 1933) and the striking glomerulus observed in Myoxocephalus octodecimspinosus (Fig. 9). The two instances of cyst formation in glomerular tufts of the horned toad are particularly interesting. In the first place, this is the only species above the fishes in which such cysts have yet been recorded. In the second place, the reptilian glomerulus usually exhibits a central, avascular, cellular core, which, according to Regaud and Policard (1903) and Cordier (1928), is made up of connective tissue. Such a core is constantly present in the glomeruli of the horned toad (Marshall and Smith, 1930; Vilter, 1935). In the two glomeruli illustrated in Figs. 18 and 19, the cystic cavities occupy the region of the typical central cellular core, and replace it to such an extent that no characteristic por- tion of the core is any longer recognizable. One is led to wonder whether the cysts might not have arisen as the result of degeneration of the central avascular area. The irregular mass described above for the larger cyst (Fig. 18) is opaque and amorphous, and has all the appearance of calcified debris ; it might be construed as the remains of the original core. Let us now examine the available fish material in the light of these considerations for the horned toad. The glomerular tufts of Cre- nilabnis tend to be somewhat cellular, and one occasionally finds a cen- tral avascular core. The tufts of Corvina tend to be quite cellular, and it is not infrequent to find a typical central cellular core, entirely comparable with that seen in the horned toad and pigeon (see below). In Myoxocephalus scorpius (Grafflin, 1933) many tufts show a mark- edly cellular center, which may be entirely avascular. In an earlier study (Grafflin. 1929) it was shown that the relatively few glomerular structures present in the kidney of the adult goosefish (Lophins pis- cat orius) have lost all connection with renal tubules. The important fact for the present problem is that eight out of thirty-one of these 256 ALLAN* L. GRAFFLIN " pseudoglomeruli " which were carefully studied showed a central de- generation of the glomerular tuft. In seven of them the center of the tuft was hyaline, eosinophilic and entirely avascular, and showed a few scattered nuclei and nuclear fragments; in two of these seven this cen- tral area was vacuolated in addition.4 In all of these tufts the central hyaline area was very sharply delimited from the peripheral tissue. If our well-delimited cysts are to be interpreted in terms of degen- eration and liquefaction of a central avascular portion of the tuft, cer- tainly it is just as conceivable that such a process could occur in all of the fishes described above as in the horned toad. In favor of this in- terpretation are the amorphous mass in one of the cysts of the horned toad and the scattered nuclei in the cyst of Corvlna. However, ranged against such an interpretation are the following facts. ( 1 i The two cysts observed in the horned toad are isolated in- stance's. whereas one might reasonably expect them to be numerous on tin's basis. (2) In the pigeon, whose glomeruli likewise exhibit a cel- lular avascular core, an exteiiM've search of abundant material failed to reveal a single instance of cyst formation (Yilter, 1935). (3) Though many glomeruli of the goosefish show marked degeneration of the cen- tral part of the tuft, no instance of cyst formation in such a tuft lias yet been observed. (4) In the specimens of Ophichthys, Lepidosiren and Protop/rrus, the glomerular tufts are well vascularized, and no ac- cumulations of cells have been observed which in any way suggest a cellular avascular core. Similarly in the sculpin (M. octodecimspino- sus) , in the usual range of glomerular size, no well-defined central core has yet been observed. (5) Some of the cysts are very small, and it seems perfectly clear that as we see them they are at their maximum size. It is hardly conceivable, in view of the findings in the pigeon, that an avascular region of the si/.e represented bv these small cysts would undergo degeneration. (6) The cysts are in general spherical, and give everv indication of having contained fluid under pressure. If we were dealing merely with a degeneration of the central portion of the tuft, one would more logically expect collapse of the tuft rather than distension of the type observed. In summary, while some of the glomerular cysts are apparently formed on the basis of a degenerative process in the glomerular tuft. 4 The eighth tuft was very small and atrophic, ami sho\\ed a central degenera- tion to the point of cavity formation, the cavity containing granular debris. The picture is not at all typical of the cysts discussed here, and will not be further In the leuenil to Fig. 7 of this earlier paper there is an obvious error. It is dear from the illustration that no ciliated neck segment is present, and that the intracapsular space open, directly into a segment whose cells exhibit the brush l.'ir) the occlusion at both ends of a portion of one of the glomerular capillaries. On either basis one could readily under- stand (1) the flattened epithelial lining of the cyst; (2) the subsequent enlargement of the cyst without, at the same time, any encroachment upon or collapse of the surrounding capillaries; and (3) the presence within the cyst of coagulum, which would represent simply a seepage of plasma proteins into the completely closed cavity. Also, such a mode of formation would be consistent with the small size of some of the cysts and the failure to find, in adult animals, more than occasional sug- gestive intermediate stages in cyst formation. SUMMARY Well-delimited glomerular cysts have been observed in the kidneys of the following species of fishes: Protopterus athiopicus, Lcpidosiren parado.ra, Myoxocephalus sc orpins, Myoxocephalus octodecimspinosus, Ophichthys imberbis, Crenildbrus pavo, Corvina nigra; and in the horned toad, Phrynosoma cornutum. It is concluded that these cysts are probably formed in two ways: (1) on the basis of a degenerative process in the glomerular tuft; (2) as the result of a malformation of the glomerulus in embryological development. LITERATURE CITED CORDIER, R., 1928. Etudes histophysiologiques sur le lube urinaire des reptiles. Arch, dc liiol.. 38: 109. GRAFFLIN, A. L., 1929. The pseudoglomeruli of the kidney of Lophius piscatorius. Am. Jour. Anat., 44: 441. GRAFFLIN, A. L., 1933. Glomerular degeneration in the kidney of the daddy sculpin (Myoxocephalus scorpius). Anat. Rcc., 57: 59. GRAFFLIN, A. L., 1937. The structure of the nephron in the sculpin, Myoxocephalus octodecimspinosus. Anat. Rec., 68: in press. AIARSHALL, E. K., JR., AND H. W. SMITH, 1930. The glomerular development of the vertebrate kidney in relation to habitat. Biol. Bull., 59: 135. REGAUD, C. AND A. POLICARD, 1903. Recherches sur la structure du rein de quelques Ophidiens. Arch. d'Anat. Micr., 6: 191. VILTER, R. W., 1935. The morphology and development of the mctanephric glomerulus in the pigeon. Anat. Rcc., 63: 371. Vol. LXXII, No. 3 June, 1937 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY m^ '/go5** &/»•••** THE PRODUCTION OF INTERMEDIATE-WINGED APHIDS WITH SPECIAL REFERENCE TO THE PROBLEM OF EMBRYONIC DETERMINATION l A. FRANKLIN SHULL (From the University of Michigan, Ann Arbor, Michigan) Aphids which are intermediate in structure between the partheno- genetic and the gamic type have already been studied in relation to the order of embryonic segregation (Shull, 1930, 1931, 1933). More abundant and more eariy produced at will are those intermediates which are structurally between the winged and wingless parthenogenetic types. They offer the s?.ne opportunity to examine certain features of the mechanism of development as do any other intermediates which can be artificially produced. Partly to obtain large numbers of these winged-wingless intermedi- ates for analysis, and partly in the hope that the method of obtaining them would throw some light on their nature, extensive experiments in which the environmental agents known or believed to modify the aphid cycle were combined in a variety of ways were carried out over a period of years. This paper describes the experiments, which involved a total of 530,133 aphids, of which 9,152 were intermediate-winged. The Stocks of Aphids Two independent strains of aphids of the species Macrosiphum solanifolii have been used. One was collected near Ann Arbor in the year 1923 and has been maintained parthenogenetically ever since. This strain was clone A of a former paper (Shull, 1932) until it experienced a " mutation ' in the fall of 1929, in which its characteristics were greatly altered, becoming clone A' of that paper. The other strain was collected in Ann Arbor in 1931 ; to date none of the results from this line has been published. 1 Contribution from the Department of Zoology, University of Michigan. This investigation is part of a series which has been aided over a period of years by grants from the American Academy of Arts and Sciences, the National Research Council, and the Elizabeth Thompson Science Fund, as well as from the Faculty Research Fund of the University of Michigan. 259 260 A. FRANKLIN SHULL Since it has been shown that conditions affecting the aphids at least a generation in advance of the beginning of an experiment may modify the results of the experiment (Shull, 1935), four different stocks of each of the two strains were continuously reared. One was kept at a temperature of 24°, one at 14°, one at alternating temperatures (24° for 8 hours, 14° for 16 hours daily), and one at room temperature which fluctuated considerably. All were in continuous light which was not necessarily of uniform intensity. The Method of the Experiments In general a complete set of experiments would proceed as follows. From each of the eight stocks mentioned above a number of aphids, both winged and wingless, were drawn shortly before (or in a few cases just after) they became adult. Some of each type of female were placed at a temperature of 24°, others at 14°. In each of these temperatures some were kept in continuous light, others in intermittent light (8 hours light, 16 hours darkness). In addition, some aphids fnun each of the stocks except the one in room conditions were brought to room temperature for one generation prior to the beginning of the experiment ; their offspring, having grown up at room temperature, were then reared in each of the various combinations of conditions just out- lined. It was never possible to do all of these things at once. However, all experiments involving one stock, and sometimes those from two of the stocks, could be done simultaneously. Since fairly consistent results were obtained from each treatment, it is unlikely that any serious error lias been introduced by performing the various parts of an experiment at different times. Effect of Continuous and Intermittent Lie/lit in the 1923 Strain The results of these experiments can be presented only as total num- bers of individuals of the several kinds produced under each combina- tion of conditions. This is best done in tabular form. In branching tables, the contrasting conditions placed nearest the data are the ones whose effect is best shown, since pairs of adjoining numbers represent the; contrast. Other contrasts are between numbers at some distance from one another, and are less readily seen. It will be necessary, there- fore, to arrange the data in several ways to show the significant results. In Table I the last column preceding the data (the fourth column) includes the two contrasted light conditions, namely, continuous and in- termittent (8 hours of light, 16 hours of darkness). It is therefore the .fleet of light which is best shown by this table. To save space the INTERMEDIATES AND EMBRYONIC DETERMINATION 261 TABLE I The percentages of winged and intermediate-winged offspring produced by both wingless and winged aphids of the 1923 strain in various combinations of light and temperature both during and before the experiments. Stock parents came from Tempera- ture of parents pre- ceding experi- ment Tempera- ture of parents during experi- ment Light on parents during experi- ment Offspring From wingless parents From winged parents Total number of off- spring Percent- age winged Percent- age inter- mediate- winged Total number of off- spring Per- centage winged Percent- age inter- mediate- winged 24° 24° 24° Cont. 8-16 7,640 5,832 50.7 32.5 0.26 3.23 12,816 6,777 59.7 37.0 0.37 4.01 14° Cont. 8-16 7,835 5,592 46.5 16.6 0.18 5.17 17,056 17,335 42.4 19.2 0.28 6.35 Room 24° Cont. 8-16 2,522 1,650 47.4 38.4 0.12 3.03 2,261 1,737 40.0 28.6 0.04 4.38 14° Cont. 8-16 2,387 2,445 62.7 15.4 0.08 4.05 2,301 1,647 32.2 9.0 0.22 10.75 14° 14° 24° Cont. 8-16 6,550 5,240 49.4 47.4 0.11 5.42 8,309 7,085 33.6 20.5 0.13 2.87 14° Cont. 8-16 7,182 5,307 55.0 20.4 0.07 8.95 6,806 4,084 21.3 11.5 0.13 11.75 Room 24° Cont. 8-16 880 846 51.1 39.7 0.11 5.52 3,329 3,326 42.4 29.8 0.24 2.56 14° Cont. 8-16 1,238 1,017 50.8 18.6 0.00 7.37 3,630 3,169 29.6 10.4 0.08 9.09 Alt. Alt. 24° Cont. 8-16 5,581 5,547 49.3 39.5 0.13 2.94 10,505 9,205 55.3 43.7 0.39 4.20 14° Cont. 8-16 7,257 5,702 35.8 13.0 0.12 7.87 12,786 8,975 39.6 12.9 0.13 10.92 Room 24° Cont. 8-16 2,902 2,709 44.9 35.4 0.21 3.03 4,094 3,545 35.8 25.3 0.07 2.03 14° Cont. 8-16 3,225 2,760 44.6 17.8 0.09 8.48 4,650 3,190 34.1 7.4 0.04 7.62 Room Room 24° Cont. 8-16 4,415 3,912 46.8 41.1 0.18 2.94 8,015 6,821 45.8 30.3 0.29 4.31 14° Cont. 8-16 4,694 3,049 43.8 13.9 0.17 5.77 9,268 7,263 44.8 10.8 0.21 9.54 262 A. FRANKLIN SHULL actual numbers of individuals of the several kinds are not given; only their percentages of the total, which are essential to quick comparison, are indicated. Any one who wishes to know the numbers can readily calculate them from the total. Though this paper is concerned only with intermediate individuals, the percentages of normal winged aphids is also given for the sake of a comparison to be made on a later page. From Table I it is clear that more intermediate-winged aphids of this strain were produced in intermittent light than in continuous. This is true for both winkles-, and winged parents, for every stock, and for all temperatures both before and during the experiments. That is. in every combination of other agents, intermittent light yielded more in- termediates than did continuous light. Moreover, the differences arc large ones ; in no instance is the ratio le.ss than 10:1, and in many sets of conditions the intermediates were 20 or even 50 times as numerous in intermittent as in continuous light. Evidently continuity or discon- tinuity ot light is of major importance in the- production of intermedi- ate- in the H23 strain. I- 1 feet of Current Temperature in the 1923 Strain The results of the experiments with the 1()23 strain are rearranged in Table II so as to put the temperatures used during the experiment (24° and 14°) in the fourth column, next to the data, thereby contrast- ing most sharplv the effects of these current temperatures. The last column of this table, relating to those reared in intermittent light, shows the most striking contrast. In every combination of other conditions, more intermediate-winged offspring were produced at 14° than at 24°. The differences are fairly large, the ratios ranging roughly from 1.3: 1 up to nearly 4:1. These differences are obtained, however, only in intermittent light. In continuous light (column 7), the differences are small and mostly in the opposite direction. That is. more intermediates are produced at 24° than at 14° (one exception). Though all the numbers are small, the fact that nearly all the differences are of the same sign can hardly mean other than that there is a .significant reversal of the effect of tem- perature in continuous light as compared with the temperature effect in intermittent light. Wingless and U'int/ed J'arenls in the 1()-J Strain P.y rearranging the experiments with the two types ot parents at the right of the branching table, as is done in Table III. whatever difference there is in the tendency of wingless and winged parents to produce in- i' i mediate-\\ inged offspring is best shown. INTERMEDIATES AND EMBRYONIC DETERMINATION 263 TABLE II The data of Table I rearranged to show the contrast between 24° and 14° tempera- ture in the production of intermediate-winged aphids by the 1923 strain. Nature of par- ents Stock parents came from Tempera- ture of parents pre- ceding experi- ment Tempera- ture of parents during experi- ment Offspring Produced in continuous light Produced in 8-hr, light, 16-hr, darkness Total number of off- spring Percent- age winged Percent- age inter- mediate- winged Total number of off- spring Percent- age winged Percent- age inter- mediate- wingcd C/l en — SD c £ 24° 24° 24° 14° 7,640 7,835 50.7 46.5 0.26 0.18 5,832 5,592 32.5 16.6 3.23 5.17 Room 24° 14° 2,522 2,387 47.4 62.7 0.12 0.08 1,650 2,445 38.4 15.4 3.03 4.05 14° 14° 24° 14° 6,550 7,182 49.4 55.0 0.11 0.07 5,240 5,307 47.4 20.4 5.42 8.95 Room 24° 14° 880 1,238 51.1 50.8 0.11 0.00 846 1,017 39.7 18.6 5.52 7.37 Alt. Alt. 24° 14° 5,581 7,257 49.3 35.8 0.13 0.12 5,547 5,702 39.5 13.0 2.94 7.87 Room 24° 14° 2,902 3,225 44.9 44.6 0.21 0.09 2,709 2,760 35.4 17.8 3.03 8.48 Room Room 24° 14° 4,415 4,694 46.8 43.8 0.18 0.17 3,912 3,049 41.1 13.9 2.94 5.77 •a OJ M C > 24° 24° 24° 14° 12,816 17,056 59.7 42.4 0.37 0.28 6,777 17,335 37.0 19.2 4.01 6.35 Room 24° 14° 2,261 2,301 40.0 32.2 0.04 0.22 1,737 1,647 28.6 9.0 4.38 10.75 14° 14° 24° 14° 8,309 6,806 33.6 21.3 0.13 0.13 7,085 4,084 20.5 11.5 2.87 11.75 Room 24° 14° 3,329 3,630 42.4 29.6 0.24 0.08 3,326 3,169 29.8 10.4 2.56 9.09 Alt. Alt. 24° 14° 10,505 12,786 55.3 39.6 0.39 0.13 9,205 8,975 43.7 12.9 4.20 10.92 Room 24° 14° 4,094 4,650 35.8 34.1 0.07 0.04 3,545 3,190 25.3 7.4 2.03 7.62 Room Room 24° 14° 8,015 9,268 45.8 44.8 0.29 0.21 6,821 7,263 30.3 10.8 4.31 9.54 264 A. FRANKLIN SHULL This difference, though sometimes striking, is not everywhere of the same sign. In 21 of the 28 sets of conditions the winged parents pro- duced the more intermediate offspring, and most of the differences are of some size. Of the 7 exceptions to this rule, 6 are from parents which were reared at room temperature throughout their immature stages, suggesting that perhaps the peculiarly variable conditions thus introduced were favorable to intermediate-winged offspring. However, the seventh exception (which is the seventh pair in the left side of Table III) was exhibited by parents not reared at room temperature; and the difference shown here is one of the more striking ones in the table. It seems impossible to bring the wingless and winged aphids under any general rule concerning their tendency to produce intermediate-winged offspring. Effect of Antecedent Temperatures in 1923 Strain Two other rearrangements of the results of the experiments would be required to show most directly the effects of the temperatures ap- plied to the aphids prior to the beginning of the experiments. One table would show any difference in the number of intermediate offspring produced by aphids taken from the several temperature stocks, that is, the stocks reared at 24°, 14°, alternating, and room temperatures re- spectively. The other rearrangement would show whether bringing the parents to room temperature through all their immature stages before using them in an experiment affected the number of their intermediate- winged offspring. The differences which these tables would show are not regular enough to warrant their presentation in that form. A state- ment of them, however, should be made. Keeping the experimental parents at room temperature through their immature stages, when that represented a change from previous con- ditions (as it would for aphids taken from the 24°, 14°, and alternating- temperature stocks), caused them to produce fewer intermediate-winged offspring in 16 out of 24 combinations of other conditions. Of the 8 which showed the opposite effects, 6 were reared at 24° during the ex- periment, the other 2 were reared at 14°. Among these S likewise were aphids from each of the four temperature stocks. Again it is impos- sible to state a general rule describing the effect of a one-generation transfer to room temperature in advance of an experiment. Still less regular is the difference between the various temperature stocks, in their tendency to produce intermediate-winged offspring. Aphids from each stock yielded more intermediates than did those from all the other stocks, in from one to three of the combinations of other • •onditions. This is true regardless of whether the aphids were brought INTERMEDIATES AND EMBRYONIC DETERMINATION 265 to room temperature for one generation before the experiment or not. Aphids changed from the alternating-temperature stock to room tem- perature yielded more than did those reared continuously at room tem- TABLE III Differences between wingless and winged parents of the 1923 strain in their production of intermediate-winged offspring under various conditions, (ws == wing- less, wd = winged.) cn cn tf3 -*-» M bo 4-1 bC fl ^ c a a C ^j C a a 1.1 •3S m m a <-i IU a E GJ b Qc ,r t 5J og 3 came rents kl a C/J its 0 o. v 71 M its .a i— £* rt v O 1) cn tn C +s 0> C g 1-1 w aj a M O 00 a t-l "*-. o o. tf> o a oo tc a 51 Temperature preceding ex OJ ely related. This arrangement shows that aphids taken fnnn the room tempera- ture stock yielded more intermediate-winged offspring than did those of any of the stocks kept at other temperatures (24°, 14°, alternating), provided the stock conditions were maintained up to the very beginning of the experiment. This is true of both wingless and winged parents, and of every combination of other conditions, though one of the differ- ence^ is SO -light a> not to appear as a difference in the table. Changing the parents to room temperature from one of the other temperatures (24°, 14°, or alternating j yielded on the whole more in- termediate-winged offspring than did continuing the stock temperature up to the beginning of the experiments, if wingless parents were used, but fewer intermediates if winged parents were used. There are, how- ever, three exceptions out of twelve to each part of that statement. Moreover, one of the exceptions is one of the only two very striking differences to which attention was called in the preceding section. These exceptions, which do not themselves possess any one characteristic in common, weaken the generalization that seems otberwi.se warranted. Effect of Temperature and of Xalitrc of Parents on 1'nnluctlon of Intermediates in 1931 Strain The remaining factors produce too little or too variable an effect on intermediacy to warrant rearrangement of the data in tables to show those effects clearl\-. The following statements of these irregular effects can be verified from the tables already given, by comparing per- centages not adjacent to one another. The nature of the parents has no regular influence on the number of intermediates. Winged parents produced more intermediate offspring in 17 combinations of external conditions, but fewer in 1 1 combinations. Eight of the latter group of 11 came from 14' and alternating tempera- ture stocks and in addition were at room temperature for one genera- tion before the experiment started. Current temperature had no consistent effect, since aphids at 14° produced more intermediates than did those at 24° in 13 combinations. but fewer in 14. and an e<|iial number in 1. Xo single factor is common to all of the 13. nor to all of the- 14. INTERMEDIATES AND EMBRYONIC DETERMINATION 271 Relation of Intermediacy to //7m/ I'roditclion in tlic 1V31 Si rain As was stated earlier, the only situation in which the 1931 clone pro- duced very many intermediate-winged individuals was intermittent light and a current temperature of 14°, applied to parents which came from the 24° stock but which were reared at room temperature throughout their own immature stages (eighteenth line, Table IV). These conditions are not exceptionally conducive to production of normal-winged aphids. Winged individuals were, it is true, more abundant (47 per cent) under these conditions than in the experiments as a whole (34 per cent) in the 1931 strain. But other combinations of conditions resembling this one yielded higher wing production as often as lower. Thus, changing only the current light from intermittent to continuous served to reduce wing production 13 per cent; and changing only the current temperature from 14° to 24° reduced wing production 19 per cent. But omitting only the change to room temperature during the immature stages of the parents in the experiments increased wing production 29 per cent; and taking the parents from the 14° or the alternating temperature stock instead of the 24° stock increased wing production 38 per cent and 39 per cent respectively. None of these percentages appears directly in any of the tables, since they are based on the combined output of both wingless and winged parents. It may also be pointed out that the set of conditions yielding the most intermediate individuals produced fewer winged ones than any other set of conditions that included a current temperature of 14°, but more than any set involving 24° temperature during the experiments. On the whole, therefore, the greatest number of intermediate aphids would seem to be produced under circumstances which favored medium wing production in this strain. Less connection between wings and intermediate wings is to be seen in a comparison of lines 17 and 18 in Table IV. The two very large percentages of intermediates are shown in line 18. The conditions re- sulting in the aphids of line 17 differ from those producing the aphids of line 18 only in that they included continuous light instead of inter- mittent light. This one change reduced the proportion of winged aphids only moderately, but reduced the intermediates very greatly. Con- tinuous light reduces the number of winged aphids in some other clones (clone A, for example, Shull, 1932), and in general does so for the 1931 strain used in these experiments. That response is well shown in Table V, where the percentages of winged individuals are consistently lower in the upper half of the table than in the lower. But in that 2/2 A. FRANKLIN SHULL upper half, the percentages in the ninth line stand out as far higher than any others. Continuous light had little of its usual effect on the num- ber of winged offspring in this particular combination of other agents, but reduced the intermediate-winged individuals to as low a proportion as the average of all others in continuous light (upper half of Table V). It is possible, therefore, to change the production of intermediates with- out greatly changing the production of normal-winged aphids. I f the relation of intermediates to winged aphids is to be studied by the correlation method, the data of the 1931 strain must be broken up into three parts. Since intermittent light favors both wings and inter- mediate wings, strung correlation between the two would be shown by the data as a whole, but this would indicate nothing more than Table IV does. Current temperature also has a marked effect on wing produc- tion in intermittent light, though not in continuous light. Since calcula- tion of correlation coefficients is justified only when the two variable qualities are distributed unimodally, and is most significant when the distribution is fairly symmetrical, separation of the data into three groups is necessary to avoid wide gaps and extreme skewness. The three groups are (1 ) those aphids reared in intermittent light and at 24°, (2) those reared in intermittent light at 14°, and (3) those reared in continuous light without distinction of temperature. From two of these have been omitted two percentages (winged aphids in one, intermediates in the other) because they were so high as to create the discontinuity which grouping the data was designed to prevent. These high percent- ages must have some special reasons with which correlation studies are not fitted to deal. The three coefficients of correlation between per- centage of winged and percentage of intermediate- winged are, respec- tively, .30 ± .24, -- .70 db .15, and .03 ± .20. Two are insignificant, the third strongly negative. It seems clear, then-fore, that whatever is inducing wing develop- ment in these more unified segments of the experiments is not par- ticularly conducive to production of intermediates — is, indeed, some- what opposed to it. Conceivable ScJicines of Stimulation The general method of explaining intermediates proposed by Gold- schmidt (1923 ct al.) particularly for intersexes is perhaps applicable to aphids. It involves the assumption that something (a hormone) re- sponsible for the development of structural characters is present in dif- ferent concentrations at different times, usually in increasing concentra- tions with increasing age during the critical period of development. INTERMEDIATES AND EMBRYONIC DETERMINATION 273 Since the sexes start development under the influence of different groups of genes, and since they typically possess in the end quite differ- ent structures, it was natural to assume two hormones, one for each sex. Development of the individual began with one of these hormones in the ascendancy, but somewhere in the course of differentiation it was over- taken by the other hormone, the moment of passing being the " turning point." Structures determined before that point were like those of one sex, characters determined after that point were like those of the other sex. It would be possible to make the same assumptions regarding inter- mediates between winged and wingless aphicls. These aphids, however, are presumably genetically alike. Moreover, the difference between them lies almost entirely in the possession by the winged type of a group of characters which are wanting in the wingless. It is therefore simpler to represent the differentiating characters as dependent upon the concen- tration of a single substance in the critical period of development. No serious flaw would be created in the concept of the fundamental nature of the intermediates even if this representation should prove to be in error. It is assumed for the purpose of this discussion that this one hor- mone varies in concentration during development in the general manner portrayed by Goldschmidt's curves, which were originally suggested, no doubt, by the course of other biological processes in developing organ- isms. In proper concentrations it will be regarded as a stimulus to the development of the differential characters of the winged type, though it could as well be considered an inhibitor. The unsegmented egg will be represented as possessing only a minimal quantity of this hormone, though it is conceivable that the egg might have a higher concentration than any later stage. Within these self-imposed limits we may imagine the influence which various situations would have on the frequency and nature of intermediates. Differentiating Characters. — The winged female aphid differs from the wingless in having wings, three ocelli on the head, wing muscles in the thorax, and extra sensoria on the third segment of the antennae (4-6 in the wingless, 15-18 in the winged). There are color differences, the third segment of the antennae being blackish and the thorax brownish in the winged type, but these are subject to considerable variation and are not included in this discussion. It is the four structural characters named which are assumed to develop if a certain concentration of the stimulating hormone is attained early enough. Sharply Defined Single Thresholds. — The line of demarcation be- tween wingless and winged could be a sharply denned level of concentra- 274 A. FRANKLIN SHULL tion of the hormone, such that any deficit, however slight, would leave the differentiating structures wholly absent, while any excess, however small, over this level would cause those structures to be fully developed. The simplest case would be that in which the same concentration of the hormone would stimulate the development of all four structures — wings, ocelli, wing muscles, and sensoria. The time in development at which these structures have their fate (their presence or absence i decided could likewise be a sharply defined moment, such that if the stimulus were applied in advance of that time the structures would develop, while if the stimulating level of the hor- mone were attained even only slightly after that time the organs would be entirely lacking in the adult. The simplest case is again that in which the time of determination is the same moment for all four of the struc- ture-. This .situation is represented in Fig. 1. .-/, in which time during de- velopment is expressed by horizontal distances and concentration of the hormone by vertical distances. The curves indicate the rate of accumu- lation of the hormone during development. Any curve passing to the left of and above the dot in the middle of the chart would be that of a winged adult : any curve below and to the right, that of a wingless adult. Intermediate individuals would be practically excluded by the rela- tions here1 described, since very seldom would a curve of concentration pass exactly through the intersection of stimulation level and time of determination. I'iffcrciit '/'lines a f Determination. — The preceding situation could be modified by having different times of embryonic determination for the four structures, but the same- concentration of the hormone as a stimulus for all. This condition is represented in Fig. 1, />. The order of de- termination of the four differentiating features is arbitrarily assumed purely for purposes of illustration. All curves of hormone concentra- tion rising rapidly enough to pass the stimulating level in advance (to the left) of all the.se times of determination would be those of winged aphids ; while all curves, reaching that level to the right of (after) all the times of determination would be those of wingless individuals. Curves between these limits would lead to mosaics. The second curve in H would be that of an aphid having ocelli, wing muscles and extra antennal sensoria fully developed, but lacking wings entirely. The third curve would pertain to one lacking wings and ocelli altogether, but having wing muscles and extra sensoria fully present. The fourth curve would belong to an aphid having extra antennal sensoria to the lull num- ber, but lacking all the other marks of a winged individual. It is proposed in this paper to use the term mosaic tor a patchwork "i typical parts, the word intermediate for an individual in which one INTERMEDIATES AND EMBRYONIC DETERMINATION 275 ur more structure's are partially developed. '' Typical parts " means, in wingless aphicls, the absence of any degree of the marks of a winged aphid. A mosaic thus would have one or more of the features of the winged aphid completely developed, hut lack the others entirely. One individual could he both mosaic and intermediate, by having certain structures either fully developed or wanting, other structures partially developed. The conditions described in /> would lead to a certain number of mosaics, but rarely to intermediates. How many mosaics there were would depend partly on the lengths of the intervals between the times of determination. Different Levels of Stimulation. — The several features of a winged aphid could be stimulated to develop by different concentrations of the hormone. If there were a single time of determination for all four of these features, the general situation would be as shown in Fig. 1, C. Without further explanation it should be clear that the second and third curves are those of mosaics, the second possessing ocelli, wing muscles and extra sensoria, the third having only wing muscles and sensoria. Intermediates would be rare, and the number of mosaics would depend on the extent of the differences among the levels of stimulation. Range of Stimulation. — It seems likely that a certain concentration of the hormone could not act suddenly to determine a complete struc- ture. More probably a given concentration, if applied early enough, serves to stimulate a slight development of the structure, while a greater concentration would cause a greater development of it. There would thus be a level (concentration) of beginning differentiation, and a level of complete differentiation. If this range of stimulation were identical for all four of the features of the winged aphid, and if there were a sharp time of determination likewise common to all four of them, the conditions would be represented by Fig. 1, D. Under these conditions there could be some intermediates, but no mosaics. Any individual whose hormone concentration rose above the level of beginning differ- entiation before the time of determination, but did not rise as high as the level of complete differentiation until after that time, would be an intermediate. All four of the distinguishing structures would be only partially developed. They might not be developed to equal degrees, since they might respond at different rates to increasing quantities of the hormone. The second curve in D would belong to an aphid having wings, ocelli, wing muscles, and antennal sensoria highly, but not com- pletely, developed, the third curve to one in which these structures were only slightly differentiated. How many intermediates were produced would depend in part on 276 A. FRANKLIN SHULL Wd Ws Stimu- lation level for all Wd Ws Ws Stimu- lation level for oil Time of determination for all Period of determination for all Wd Wd • : ' M Wd Times of determination Wd Stimu- lation level forall Stimu- lation levels Time of determination for all Wd VMs F ComPldiff. for Begin.J all Period of determination forall G WdMM MO Ws Wd Comp.) diff. for Begin.J oil Times of determination Wd Ws Time of determination for all H Wd M*l M + ) Ws Stimu- lation levels Ws Period of determination forall Fir;. 1. Curves representing hormone concentration in relation to production of intermediate-winged individuals in aphids, tinder various assumptions concerning times or periods of determination and levels or ranges of stimulation during em- bryonic development. /, intermediate; M, mosaic; Oc, ocelli; Sn, scnsoria ; Wd, winged; Wg, wings; Wm, wing muscles; Ws, wingless. INTERMEDIATES AND EMBRYONIC DETERMINATION 277 the difference between the concentration which would begin and that which would complete the differentiation. Period of Determination. — It is not necessary that a structure have its fate irrevocably fixed at one moment in development. That deter- mination may be spread over a period of time. If the hormone rises to a stimulating concentration just before that period ends, the structure may be supposed to be slightly developed. If the hormone reaches a stimulating level soon after the beginning of the period of determina- tion, the structure would be highly but not completely developed. In Fig. 1, E, it is assumed that there is one period of determination for all structures and one sharp level of stimulation for all of them. The above situation would lead to the production of some inter- mediates, but no mosaics. The middle curve of E would be that of an aphid having perhaps half-developed wings, ocelli, wing muscles, and extra sensoria. How many intermediates there were would depend on the length of the period of determination. Range of Stimulation and Period of Determination. — There may conceivably be a range of stimulation levels (as in D) and also a period of determination (as in E). The simplest such situation is one in which all four differentiating structures would be stimulated by the same range of concentrations of the hormone and be determined during the same period of time. It is represented in Fig. 1, P. Any curve of hormone concentration which passed through the rectangle formed by the inter- section of the range of stimulation and the period of determination would be that of an intermediate. All the differentiating structures of such an individual would be partially developed, though not necessarily to an equal extent. There would be no mosaics. How many inter- mediates were produced would depend partly on the size of the rectangle —the extent of the range of stimulation and the length of the period of determination. Range of Stimulation unth Different Times of Determination. — If there be one range of stimulating concentrations of the hormone com- mon to all four of the structures, and sharply defined times of deter- mination which are different for all four, the conditions are those por- trayed in Fig. 1, G. Any curve cutting one of the heavy vertical lines would belong to an individual intermediate for the structure whose time of determination and range of stimulation that heavy line represents. The same individual would be mosaic for any structures whose heavy line its curve did not cut. Thus, the second curve in G would be that of an aphid lacking wings, having almost complete ocelli, and having fully developed wing muscles and sensoria. The third curve would be- long to an individual lacking wings and ocelli, but having partially de- veloped wing muscles and fully developed sensoria. A. FRANKLIN SHULL Individuals not typically winged or wingless, produced under the above conditions, might well he both mosaic and intermediate. Some of them would perhaps be mosaic alone, especially if the range of stimula- tion were narrow and the times of determination were widely spaced. Most of them would be intermediate only, if the range of stimulation were broad and the intervals between times of determination were very short. Common Period of Determination with Different Levels of Stimula- tion.— Another possibility is that all structure's have the same period of determination but different sharply defined levels of stimulation, as in Fig. 1. II. An individual would be intermediate for a given character if its hormone curve cuts the heavy line representing its period of deter- mination and level of stimulation. It would be mosaic for (either pos- sess or lack) any structure whose corresponding heavy line it did not cut. Intermediates would prevail if the period of determination were long and the stimulation levels close together. Mosaics would be more com- mon if the period of determination were short and the levels of stimula- tion far apart. Many atypical individuals would doubtless be both mosaic and intermediate1. One Level of Stimulation with Distinct Periods of Determination.— A single sharp level of stimulation for all distinguishing structures may IK- combined with periods of determination which are different for the different characters. If these periods do not overlap, the situation is as pictured in Fig. 2. /. The second curve in / is that of an individual with partially developed wings and the other three features of a winged aphid fully formed. The third curve is that of a mosaic having wing muscles and sensoria fully developed but wholly lacking wings and ocelli. Intel-mediates would prevail among atypical individuals if peri- ods of determination were long and close together; mosaics would lead in numbers if periods of determination were short and separated by wide intervals. Single Time of Determination with Separate Ranges of Stimula- tion.— If all the differentiating structures were determined at the same moment, but each had its own range' of stimulation which did not over- lap any of the others, the situation would be that represented by Fig. 2, /. The second curve of this chart, cutting one of the heavy lines rep- resenting time of determination and range of stimulation but missing the others, represents a mosaic and intermediate individual. The third curve, missing all the heavy lines, is that of a mosaic only. Intermedi- ates would be favored, among atypical aphids, by wide ranges of stimu- lation close together; mosaics would be favored by narrow ranges widely INTERMEDIATES AND EMBRYONIC DETERMINATION 279 Wd M Stimu- lation level for all Wd M, Stimu- lation ranges Periods of determination Periods of determination. Wd Wd Ws Stimu- lation ranges Time of determination for all Ws .Ws ComP-ldiff. for Begln.JaU Periods of determination N Wd Wd Ws Ws Comp diff Begin diff. Overlap, stimu- lation ranges QUO £ Periods of determination Wd M + I Wd Ws M + l Ws Stimu- lation ranges Overlapping periods of determination Wd Wd Ws M+I Stimu- lation ranges Period of determination for all Wd Ws Ws :)Cornp] overlap. •Miff. stimu- Overlapping periods of determination FIG. 2. Continuation of Fig. 1, representing increasing substitution of ranges for levels of stimulation, and of periods for times of determination, and increasing overlapping in place of distinct ranges and periods. 280 A. FRANKLIN SHULL One Range of Stimulation with Distinct Periods of Determina- tion.— If stimulation occupied a range common to all structures, and each organ had a different period of determination without overlapping, Fig. 2, A', would represent the conditions. The three curves in the middle are those of both mosaics and intermediates, since each curve cuts one or more of the heavy rectangles and misses one or more others. If the rectangles were narrow and high and close together laterally (that is, if periods of determination were short and not far apart while the threshold of complete differentiation were far above the threshold of be- ginning differentiation), most atypical individuals would be intermediate only. If periods of determination were short and widely separated, and the range of stimulation were narrow, mosaics would predominate. Common Period of Determination with Distinct Ranges of Stimula- tion.— The situation indicated by this title is portrayed in Fig. 2, L. Individuals would be intermediate for those characters whose rectangles their hormone curves traverse, mosaic for those whose rectangles are missed altogether. If the common period of determination were long, and the ranges of stimulation were narrow and close together, most atypical aphids might well be intermediate only. In most other situa- tions of this general kind, combined mosaics and intermediates would be more likely. Distinct Periods of Determination and Separate Ranges of Stimula- tion.— In Fig. 2, M, is shown the combination of different periods of determination and different ranges of stimulation, without any over- lapping of either periods or ranges. Many possibilities for either few or many intermediates, few or many mosaics, are presented by this situation. If the structures having early determination have low stimulation ranges, and those determined late have high stimulation ranges, so that the rectangles are arranged along a line rising obliquely to the right (as is more or less true in M), and if the curve of concen- tration of the hormone follows approximately this same course, the curves would cut most of the determination-stimulation rectangles if they cut any of them, and mostly intermediates would result. An atypical individual might easily be intermediate for all the distinguish- ing characters. If, however, structures determined early had high ranges of stimulation, and those determined late had low ranges of stimulation, so that in the figure the rectangles of determination and stimulation were arranged along a line sloping downward to the right, cutting the curve of hormone concentration nearly at a right angle, mosaics would prevail if the rectangles were small, and mosaic-inter- mediates would be most common if the rectangles were large. Many other relations between the period of determination and range of stimu- INTERMEDIATES AND EMBRYONIC DETERMINATION 281 lation could exist, such that the rectangles would he distributed irregu- larly in relation to the hormone curve. Each of these sets of relations would result in a particular frequency of mosaics or intermediates or both. Distinct Periods of Determination with Overlapping Ranges of Stimulation. — The opportunity for pure mosaics would be reduced by having the ranges of stimulation overlap. This condition along with distinct periods of determination, is shown in Fig. 2, N . Unless the overlapping of ranges were very slight, it would he difficult for the hor- mone curve to pass between rectangles. While curves near the winged or the wingless extreme might miss some of the rectangles, they could scarcely miss all of them and still belong to atypical individuals. Overlapping Periods of Determination with Distinct Ranges of Stimulation. — Equally difficult, with the preceding situation, for the production of pure mosaics would be the condition in which the periods of determination overlap while the ranges of stimulation are separate. In Fig. 2, O, where this situation is represented, there might appear to be a greater chance for one of the atypical curves to miss all the rectangles than in N, but that is only because the periods of determina- tion have been assigned a smaller spread than was given to the ranges of stimulation in N. Most atypical aphids produced under these con- ditions would have to be intermediate with respect to some characters, and could easily be intermediate for all their distinguishing marks. Overlapping of Both Periods of Determination and Ranges of Stimulation. — Still more favorable to intermediates, as compared to mosaics, than any of the preceding arrangements is that in which both the periods of determination and the ranges of stimulation overlap, as in Fig. 2, P. If the limits of the several periods and the several ranges did not differ very greatly, there would be few atypical curves that did not traverse several of the determination-stimulation rectangles, and many would traverse all of the rectangles. Intermediates would thus predominate over mosaics, though individuals both intermediate and mosaic could be fairly common. The above are not all of the conceivable combinations of stimulation and determination. They will suffice, however, to illustrate the possi- bilities farther than there is any present possibility of forming judg- ments from them. Nature of the Production of Atypical Forms Since the factors which govern morphogenesis are as yet only very imperfectly understood, the possibilities outlined in the preceding sec- A. FRANKLIN SHULL tion may all be taken into account. What information, if any, concern- ing them is furnished by the methods of producing the intermediates? Analysis of the characters of the atypical aphids has only begun, and it is too early to attempt to draw from them the conclusions to which complete study may lead. Already, however, it is clear that most if not all of them are intermediate in one or more respects. To what extent they may also be mosaic is not at present known, but very few of them are pure mosaics. This fact should mean that the first few situations represented in the curves of Fig. 1 an- less likely to exist than are the later ones or those of Fig. 2. Since sharply defined times of determination and levels of hor- mone stimulation would lead to mosaics, the facts so far ascertained favor the existence of a spread in one or both of these elements. Either there must he a difference between the concentration of the stimulating hormone which will lead to slight development of a structure and the concentration which will complete it, or the amount of development must depend in part on how long the effective stimulation is applied, or both of these statements must be true. Only a more complete knowledge of individual intermediates will indicate how much spread is required. I'ntil some physiological basis of embryonic determination2 in the aphids is discovered, either a range of stimulating concentrations or a gradual process of determination may be postulated to explain inter- mediate individuals. Which is the more likely, or are both probable? One fact reported in this paper which appears to bear on this ques- tion is the large difference in the capacity of the two strains used to pro- duce intermediates. The 1931 line produced less than one-tenth of the number of intermediates produced by the 1923 strain. Also, this ca- pacity to produce intermediates must be subject to large modifications, since the high capacity of the 1923 line (clone A'. Shull. 1932) arose suddenly by a " mutation " from a stock (clone A ) which was produc- ing apparently no more atypical individuals than the 1931 strain now produces. Which of the possible bases of intermediacy is more likely to exhibit large differences in different strains, or more likely to be sub- ject to considerable change? 2 The word " determination " is used in this paper despite its obvious defects. Since different events incurring at different times may be necessary antecedents of a morphogcnctic change, a structure could not be said to be actually determined until the last such event had taken place. A morphological feature may be deter- mined, as far as one agent is concerned, later than it is determined in relation to another agent. Were it possible, in the aphid experiments, to designate a single agent in relation to which determination was being considered, a better term that would not be too cumbersome could be devised. But the tables presented in this paper indicate that a complex set of circumstances is involved. Any substitute for the word determination that would be an improvement upon it would be unwieldy. Biologists are not likely to be misled, in this connection, by the briefer term. INTERMEDIATES AND EMBRYONIC DETERMINATION 283 Intermediary could be made more frequent by increasing the length of the period or periods of determination, or by increasing the range of stimulation, so that more curves of concentration of the hormone would pass through them. It could also be made more abundant by bunching the curves of concentration ; that is, by creating a set of conditions which would cause larger numbers of individuals to develop their stimu- lating hormone at approximately the same rate, which would have to be a rate that would make the curves of concentration traverse the lines or rectangles representing periods of determination or ranges of stimula- tion or both. This concentration of the hormone curves in a narrow band would have to occur in one strain of aphids, but not in another subjected to the same combination of light and temperature, or any other combination yet tested. •/ The second of these postulates, the clustering of the rates of hor- mone development, involves an increase in the measure of control ex- erted by the animals over the process of development. A strain of aphids in which more individuals developed in the same way would be exercising more regulation than would a line in which hormone increase was more nearly random. Now, the 1923 strain, which is the one in which such clustering of rates of hormone development would have to occur, does not regulate wing production nearly as precisely as does the 1931 line. Wing development responds to light conditions much more definitely in the latter than in the former, as a glance at the respective tables shows. The fact that the effect of intermittent light is reversed in the two strains is not at issue ; the significant point is that the effect is much more precise in the 1931 strain than in that of 1923. It seems scarcely likely, therefore, that the control over rate of hormone concen- tration would be greater in the older line than in the newer one. If this conclusion is justified, there is left the extent of the period of determination or of the range of stimulating concentrations of the hor- mone, or both, to account for intermediates. Nothing in the experi- ments here reported appears to provide a choice as between these two possibilities. Ho\vever, embryonic development is known to consist in some animals, at least in part, of a chain of events in which some or all of the events provide the stimuli leading to later events. To whatever extent this is true, it would be difficult to imagine any great extension of the periods of determination of structures. Were many of these to be extended, and were each to wait upon the culmination of some preceding event, embryonic development would be correspondingly prolonged. Nothing is known that would render it probable that the embryos which become intermediate or mosaic adults take longer to develop, though the numerous ovarioles would make it possible for that to happen and for 284 A. FRANKLIN SHULL the young aphids to be borne out of their regular turns based on the time of starting cleavage. On the whole, it would seem unlikely that em- bryonic development is retarded greatly by the factors leading to in- tcrmediacy. The concentration of the hormone, on the contrary, is subject to no such known limitation. So far as anyone knows, the level of stimula- tion of completed differentiation could be farther above the level of be- ginning differentiation in one line of aphids than in another. So far as known there is no chain of events involved in the concentration of the hormone, such that one structure cannot begin to be stimulated to de- velop before the stimulation of another is completed. It is suggested ,-is more probable, therefore, that intermediacy in these aphids depends upon a gap between the thresholds of stimulation of beginning and (if complete differentiation, rather than upon periods of determination or clustering of the rates of hormone development. How this conclusion is related to the differences in the number of intermediate individuals induced by environmental conditions is difficult to ascertain. If the general concept that hormone concentration and range of stimulation levels combine to determine the winged type is cor- rect, higher proportions of intermediates could be produced at the cx- prnse of winged aphids by lowering the hormone curves (reducing the rate of accumulation of the hormone) or at the expense of wingless aphids by elevating the hormone curves — all on the assumption, made to simplify the discussion, that the hormone is a stimulant rather than an inhibitor of the development of the characters of the winged females. The more striking changes in the number of intermediates produced by light and temperature in the 1923 strain were in general of a sign op- posite to that of the change in the number of typical winged females produced by the same conditions. The suggestion contained in this comparison is that intermediates were being produced at the expense of winged adults. The reduction of the winged ones was in general, how- ever, several times as iM'eat as the increase of the intermediates. In other words, many winged aphids were being converted into wingless ones as well as some into intermediates. This fact is not opposed to the supposition that hormone accumulation was being retarded ; but since there were many wingless in most environmental conditions, a mere lowering of hormone curves would not necessarily leave any more of them in the intermediate zone. It would seem necessary to suppose that, in the lowering of the hormone concentrations, there was cither an accumulation of the curves in the intermediate /one, or a widening of the ranges of stimulation so that more curves would intersect them. INTERMEDIATES AND EMBRYONIC DETERMINATION One reason for not assuming an accumulation of hormone curves in the intermediate regions has already been pointed out. It is worthy of note that in the 1931 strain, the only set of condi- tions which induced many intermediates served to increase, rather than diminish, the number of winged individuals (from wingless parents) or had practically no effect on the number of winged ones (from winged parents). If the explanations discussed above were to have general validity, it would be necessary to suppose that the intermediates from wingless parents were being produced at the expense of wingless ones, which might be due in part to a general lifting of the hormone curves. This contrast in the behavior of the two strains recalls their different responses to intermittent light in the production of wings ; intermittent light produces fewer winged aphids in the 1923 strain, but more winged ones in the 1931 strain. There may be some common basis for these two differences. What reversal would be required in the hormone scheme to account for these contrasts it is impossible to say with any high degree of probability. Whether intermediates are in any way dependent on the direction of some physiological change (that is, winged to wingless, or wingless to winged) can hardly be judged from the experimental methods of pro- ducing them. Analysis of the characters of the intermediates will be needed for an answer to this question. SUMMARY Two strains of aphids were subjected to various combinations of temperature and light. In one strain (collected 1923) distinctly more intermediate-winged females were produced in intermittent than in con- tinuous light, in every combination of other conditions. Among those reared in intermittent light, more intermediates were produced at a tem- perature of 14° than at 24°, in every combination of other conditions. Those reared in continuous light, however, produced usually fewer in- termediates at 14° than at 24°, and the differences were small. In gen- eral, winged parents produced more intermediate offspring than did wingless parents, but there was much irregularity of this relation. Greatest production of intermediates and greatest production of typical winged females did not occur at the same time; there was, indeed, a tendency for the two frequencies to be opposed to one another. Other relations in this strain were rather indefinite. In the second strain (col- lected 1931) more intermediates were usually produced in intermittent than in continuous light, but the differences were small. More inter- mediate offspring were produced by parents drawn from a stock kept at 286 A. FRANKLIN SHULL room temperature than by parents taken from stocks kept continuously at 14° or 24°, or from a stock regularly alternated daily between 14 and 24°, though again most of the differences were small. The greatest (and the only large) number of intermediates in the 1931 strain were produced by taking aphids from a stock kept at 24°, changing them for one generation to room temperature, then putting this latter generation, when adult, at 14". In general, in this strain the greatest frequency of intermediates occurred when the frequency of typical winged aphids was medium. Other relations in the 1931 strain were not striking. If it be as>umed that production of the characteristics of winged aphids be dependent on suitable concentrations of a hormone which in- creases in amount (luring development (other assumptions could be made), intermediacy or mosaicism might result from various combina- tions of times or periods of determination, or of levels or ranges of concentration of the stimulating hormone, or both. Ranges of stimu- lating concentrations are less subject to known limits than are periods of determination, and are held to be the more likely, inasmuch as strains of aphids differed greatly in their tendency to produce intermediates. [ntermediacy, as contrasted with mosaicism, is favored by extended periods of determination and ranges of stimulation, especially if there is overlapping of these periods and ranges for the several structures. In part, the intermediates may replace winged aphids in the 1923 strain, and wingless ones in the 1931 strains. LITERATURE CITKD GOLDSCHMIDT, R., 1923. The Mechanism and Physiology of Sex Determination. (Translated hy R. Dakin.) Alcthnen and Co., London. Snri.i., A. F., 1930. Order of embryonic determination of the differential features of gamic and parthenogenetic aphids. Zcitsclir. iinl. .l!>st. it. I'crcrb., 57: 92. SHULL, A. F., 1931. Order of embryonic segregation in intermediate aphids not reversed by low temperature. Am. Nat., 65: 469. SHULL, A. F., 1932. Clonal diftYienres and clonal changes in the aphid Macro- siphum solanifolii. Am. Nat., 66: 385. SHULL, A. F., 1933. The time of embryonic segregation in aphids as determined from intermediate types. l'r<>c. Nat. Acad. Sci., 19: \(>4. SHULL, A. F., 1935. Combinations of current and antecedent conditions in rela- tion to wing-production of aphids. Hiol. Bull., 68: 35. THE LOPING OF LAND-SNAILS G. H. PARKER (Prom the Biological Laboratories, Harvard University) A little more than three decades ago Carlson (1905) published a very circumstantial account of the locomotion of a California!! land- snail, Helix dupetitliouarsi, which may he described as a lope. The animal by appropriate muscular contractions lifts its head well above the substrate, projects it forward much in advance of its former position, and brings it down again on the surface over which it is progressing. Thus the anterior third of the snail forms an arch of which the head is one base and the mid-region of the foot the other, the part between being lifted well off the ground and representing the region of principal muscular action. This arch makes up one step in the snail's forward locomotion and passes over its foot and body in an antero-posterior di- rection to disappear at the tail. According to Carlson as many as three such arches may be seen progressing as waves over the snail at once. This lope results in unusually rapid progression and is employed by the animal in place of its ordinary £low creeping as a means, for instance, of escaping from an enemy. In 1911 I called attention to this remarkable form of progression as described by Carlson and expressed the opinion, that in the lope the large waves were probably combined with the waves for slow locomo- tion, a point not touched on by Carlson. I also pointed out that the lop- ing waves took what has been designated as a retrograde course, that is, they ran from anterior to posterior, whereas in helices generally the ordinary locomotor waves are direct, from posterior to anterior. To clear up the question of the possibility in snails of a double set of waves, one opposite in direction to the other, and to elucidate further matters connected with so exceptional a scheme of progression, called for an examination of living specimens of Helix dupetithouarsi which at that time was impossible for me. During a period of work at the Kerckhoff Laboratories of the Cali- fornia Institute of Technology in the spring of 1936, I noticed in the early morning on the brick walks of the campus interrupted snail trails of the kind figured by Carlson (1905) for this loping species. I im- mediately made a search for the snails and soon found living specimens in the act of making broken trails. The loping waves on these snails 287 G. H. PARKER ran from anterior to posterior as described by Carlson. The maximum number of them seen on the foot at once was three, the foot itself being about 4.5 cm. long. \Yhen the snail was in full action the arches of the body were so high that light from the early morning sun could IK- seen shining under them. These arches at times tended to become obliter- ated toward the hind end oi" tin- foot and in some instances they ceased after having passed over onlv the anterior half of the pedal surface, but as a rule- they continued with full vigor all the way to the hind end. As a result of these differences the trails made by the snail varied from a succession of is, dated spots through a series of connected bead-like spots to a continuous trail of uniform width. Such variations were to be noted in different parts of the single trail of a given individual, show- ing that from time to time the snail changed its type of locomotion. A number of these snails were brought into the laboratory for fur- ther and closer study. Here they were made to creep upon large plate- of glass where thev could be induced to exhibit the same loping loco- motion that they had shown on the damp brick paths. A single loping wave was found to pass over the foot from anterior to posterior, a dis- tance of about 4.5 cm., in from eight to ten seconds. The snail while loping covered about ten centimeters in four minutes. \Yith the snail creeping on a glass plate it was possible to observe the foot from below and to record its action. In ordinary creeping the foot was covered by a succession of small transverse waves, about six in all, which passed from posterior to anterior and coursed over the length of the foot in from five to six seconds each. These both in direction and in general character agreed with the type of locoinotor wave long since described for the majority of helices. When tin- snail began to lope, the large waves were easily followed from below the glass plate. Though the direction which they took over the foot was the opposite of that taken by the smaller waves, the two systems ran simultaneously and without interference. Thus the underside ot the loot showed at once a system of small pedal waves running from behind forward, di- rect waves, and another of large loping waves running from anterior to posterior, retrograde waves. As a result of this double system the animal moved forward rapidly. It is hardly necessary to remark that the direction in which any such set of waves runs is independent ot the direction in which the part of the foot that is concerned with the loco- motor step moves. In both sets of waves tin- portion of the foot that is helpful in locomotion moves anteriorly; in the loping waves this loco- motor action begins at the anterior end of the snail's body and toot and proceeds posteriorly and in the ordinary creeping waves it begins in the posterior part of the foot and proceeds anteriorly. The musculature LOPING OF LAND-SNAILS involved in the two systems of locomotion must, of course, he distinct; that for loping must include the body musculature directly dorsal and lateral to the foot and that for ordinary creeping the musculature im- mediately in the foot itself and next its creeping sole. Of the scores of instances of loping that 1 have observed from the underside of the glass plate, I have never seen the two systems of waves in any other relation than that just described. Further. I have never seen a snail in motion that has not exhibited the system of small waves. These are an invariable accompaniment of locomotion. When loping is undertaken, it is always superimposed on the smaller wave-system though without disturbance. Loping never occurs except after the snail has begun ordinary creeping. Thus the question left open by Carlson as to the relation of these two types of locomotion receives its answer. I preserved the shells of the snails with which I worked in Pasadena and on my return to Cambridge, Massachusetts, I referred them to Mr. \Y. j. Clench of the Museum of Comparative Zoology for identification. Much to my surprise he informed me that they were the shells of Helix aspcrsa, a common European snail which was known to have been in- troduced into California some forty years ago and specifically into the region about Los Angeles. On looking up its habits I found in Taylor's monograph of British land and freshwater mollusks (1914) a brief statement with figures attributed to L. E. Adams showing that the Eng- lish representatives of this species were known to lope. This discovery led me to suspect that possibly Carlson in his original investigation had really worked on Helix aspcrsa and not on Helix dupetithouarsi. However, through the kindness of Mr. Clench I re- ceived a number of living Helix (now Epiphragmorpha) dupetithouarsi from Cypress Point, Monterey County, California, and I had the oppor- tunity of testing these snails as I had done with Helix aspersa. They loped as H. aspcrsa did and I could confirm all the statement made by Carlson for H. dupetithouarsi. Moreover, this species, like H. aspcrsa, always showed the ordinary locomotor waves when it crept and never loped except when these waves were present. In both species, then, loping appears to be a form of locomotion literally superimposed on ordinary creeping and a means of accelerating progression beyond that attainable by ordinary methods. BIBLIOGRAPHY CARLSON, A. J., 1905. The physiology of locomotion in gastcropods. Biol. Dull., 8: 85-92. PARKER, G. H., 1911. The mechanism of locomotion in gastropods. Jour. Morpk., 22: 155-170. TAYLOR, J. W., 1914. Monograph of the Land and Freshwater Mollusca of the British Isles. Leeds. Vol. 3. THE PHYSIOLOGY < >F DIGESTION IX PLANKTON CRUSTACEA II. FURTHER STUDIES ox THE DIGESTIVE EN/YMES OF (A) DAPHNIA AM) I 'ol.YI'HKM CS ; (B) DlAPToMt S AND CALANUS AKTIIL'R D. IIASI.HK (/•>.»>/ the Li»iiit>li>reviously (|unted by the author (1935). in addition to Gellis and Clarke ( 1935) and 1 Vhn ( 1('30), that plankton Crustacea feed on participate food and derive no nourishment from dissolved organic matter, or at least ahsorb so little of it that dissolved food stuffs account for only a small part of their mctaholic needs. This circumstance' makes it necessary that these or- ganisms possess an eiixyme system adequate for complete digestion of paniculate food. ( )ur first studies established the presence of a protcin- ase, amylase and lipasc in Ihiplinia. This paper extends these studio to the peptidascs of Ihtphnia and to the proteinascs. amylases and lipases of Polyphemus, a gymnomerous cladoceran. and to the copepods Diaptomus and Calunns. Certain ex])eriments to study the beha\'ior of Daphnia proteinase at low temperature an- also included. PKITAK. \TION OF EXTRACTS Daphnia pitlc.r were netted from Lake Monona. \\'iscoiisin. in May 1935 when a pure culture was obtainable in lars^c quantities. A pure culture of l'iil\'^liciinis s]>. \vas obtained from Crystal Lake-, Wisconsin, in August 1('.U. /iui^/iniiiis sp. was found fairly abundant in a ho^ lake near Tnnit Lake, \\ i\con>in, in July, 1935. Calanus jiiiinarcliicns was collected from l)ii//ard> I'.ay, Massachusetts, through the facilities of the Woods Mole ( Jci-ano^raphic Institution. The glycerol extracts were made from the acetone-dried and pe- troleum ether-extracted whole organisms after grinding. The water extract-- of Daphnia were j^repared from the fresh, ground organisms and after filtration preserved with toluol. 290 DIGESTION IN PLANKTON CRUSTACEA. II 291 PEPTIDASES OF DAPHNIA Our criterion of the presence of dipeptidase was the splitting1 of leucylglycine, as pointed out by Waldschmidt-Leitz (1931). Some con- tention exists as to what was called aminopolypeptidase (Johnson et al, 1936), but for convenience this enzyme was considered present when the substrate leucyldiglycine was split, and carboxypolypeptidase was considered present when the substrate chloracetyltyrosine was attacked. Titrations were made by the Linderstrp'm-Lang (1927) method in 90 per cent acetone, using N/10 alcoholic HC1 and naphthyl red as indicator. No attempt was made to purify or separate the peptidases. The crude extract was used in all cases. TABLE I Hydrolysis of Peptides by Crude Extract Enzyme Substrate PH Hydrolysis* of one linkage in cc. of HC1 3 hrs. 12 hrs. 19 hrs. Dipeptidase Aminopolypeptidase Carboxypolypeptidase ^/-leucylglycine l" Daplmia and hog proteinase at 20° and 8° C. The results are expressed in milligrams of trichloracetic acid soluble N in 20 cc. of filtrate. Digestion time Tniiiici.it ure 20° C. 8°C. I >,i]ilmia Hog 1 >,i]ihlli;i Hog hours 1 2 5 13 24 0.2S 0.81 2.56 4. IN 6.01 6.41 10.78 16.7S 0.00 0.29 0.93 1.78 2.22 3.23 5.28 8.90 One gram of dried, defatted h»g pancreas was mixed with 100 cc. of 50 per cen i glvcerol. After standing at room temperature for 24 hours the mixture was tillered through silk bolting cloth. Ten cc. of this filtrate were added to 125 cc. suspension of casein in water (10 gm./lOO cc. H..O) and kept at 20° C. Twenty-five-cc. samples were removed at intervals and precipitated with 35 cc. of 10 per rent trichloracetic acid. The suspended protein wa> filtered off and the total nitrogen (Kjeldahl) was determined in duplicate on 20 cc. of the filtrate. A similar digest was set up at 8° C. and the same procedure carried out. The hath consisted of running Lake- Mendota water whose tem- perature was 8° C. at the time of the experiment. DIGESTION IN PLANKTON CRUSTACEA. II 293 Glycerol extracts of dried, defatted, ground DapJinia were prepared by extracting 1.0 gm. of fhiplinia with 100 cc. of 50 per cent glycerul. This extract was treated in the same manner as described above for hog pancreas. After 2 hours digestion (Table II), the hog pancreas produced 48 per cent more nitrogen at 20° than it did at 8° C. DapJinia proteinase produced 36 per cent more nitrogen at 20° than at 8° C. In 5 hours the respective percentages were 53 and 36. From these data it appears that Daphnia proteinase behaves, in general, like that of the hog and follows the van't Hoff law, so that it possesses no super-activity at a low tem- perature. 20 FIG. 1 FIG. 2 FIG. 1. Hydrolysis of gelatin at pH 7.2 by a proteinase of Polyphemus. The ordinate represents decreasing viscosity (AV) ; the abscissa is the time in minutes. FIG. 2. Hydrolysis of starch by an amylase of Polyphemus. The ordinate is the number of cc. of N/40 Na.S.Os, or the amount of N/40 reducing groups ; the abscissa is the time in hours of incubation at 37° C. SOME ENZYMES OF POLYPHEMUS Proteinase. — A glycerol extract was made from comminuted, dried and defatted Polyphemus. One part was diluted to 0.025 per cent for use in proteinase determination and another to 1 per cent for amylase and lipase determination. To 0.5 cc. of 0.025 per cent glycerol extract was added 5 cc. of 1.5 per cent Sargent's gelatin ; the mixture was incu- bated at 34° C. and the viscosity periodically measured in an Qgtwald viscosimeter by the method of Northrop (1922). The proteinase was designated as a tryptic enzyme, for it was highly active at pH 7.2, al- most inactive at pH 4.35 and completely inactive at pH 3.0. The curve in Fig. 1 indicates the nature of the enzyme activity on gelatin at pH 7.2. 294 ARTHUR D. HASLER Amylasc. — Polyphemus extract (1 per cent) was also capable of hydrolyzing starch. Twenty-five cc. of 3 per cent starch at pH 7.2, 25 cc. of water and 5 cc. of 1 per cent glycerol extract were incubated at 37° C. At intervals 10-cc. samples were withdrawn and the reducing groups determined iodometrically according to Baker and Hulton (1920). The results are shown in Fig. 2. Lipasc. — Four per cent tributyrin was attacked by an csterase of the extract. Fifty cc. of 4 per cent tributyrin emulsified with sodium glyco- cholatc and 5 cc. of 1 per cent glycerol extract were placed at 37° C. Ten-cc. aliquots were withdrawn and titrated with N/20 NaOH. The titration figures can be read from Fig. 3. FIG. 3 FIG. 4 FIG. 3. Hydrolysis of tributyrin by lipase of Polyphemus. The ordinatc rep- resents the number of cc. of N/20 NaOH; the abscissa is the time in hours. FIG. 4. Hydrolysis of gelatin at pH 7.48 by proteinases of Diaptoinus (D) and Calainis finmarchicus (C). The ordinate represents the decreasing viscosity of the gelatin (AV) ; the abscissa is the time in minutr>. SOME DIGESTIVE EN/.YNTES OF DIAPTOMI s AXD CALAXUS The purpose of this study was to analyse extracts of a marine and a fresh water coprpnd for enzymes which attack proteins, carbohydrates and fats. An attempt was made with Calanus to corroborate the results of Bond (1934) by the use of different extraction and assay methods. Clarke (1934 i lias adequately discussed the problem of marine cope- pod nutrition and has (1935) presented good evidence that their food, as in Daphnia, is of a participate nature. The work of Bond represents a study of some of the enzymes of the marine copepod Calanus \\ninarchi- cus. Alcoholic extracts that he made were able to hydrolyze gelatin, starch and ethvlbutvrate. DIGESTION IN PLANKTON CRUSTACEA. 11 Materials 295 The marine copepod Calamis jiiiinurcliicits and the fresh \vater cope- pod Diaptomus sp. were collected at the places mentioned in an earlier division of this paper. Methods of extraction were identical with those described ahove for hog and Daphnia. Examination of the Extract Proteinase. — A proteinase that hydro ly zed gelatin was demonstrated in the extract of Calamis and Diaptomus by the viscosity method. On the basis of the Daphnia proteinase unit (1 unit: 20AV/20, Hasler, o FIG. 5 FIG. 6 FIG. 5. The pH activity curve of Calanns proteinase on gelatin. The ordi- nate represents decreasing viscosity of the gelatin (AV) ; the abscissa represents PH. FIG. 6. Hydrolysis of starch by amylases of Diaptomus (D) and Calamis (C). The ordinate represents the number of cc. of N/40 Na2SnO3, or the amount of the N/40 reducing groups ; the abscissa is the time in hours of incubation at 37° C. 1935), 0.5 cc. of 1 per cent glycerol extract of dried, defatted Diaptomus contained 1.5 units. A similar amount of Calanus extract had a content of 1.7 units. One per cent extracts of Daphnia had 1.5 units. The amounts of enzyme in these extracts appear to be of the same magnitude. The curves in Fig. 4 show that digestion took place within the first few minutes of enzyme activity. Sufficient extract was available from Calamis to construct a pH activity curve. Figure 5 shows the enzyme to be of definite tryptic type with maximum activity between pH 7-8. Inactivity resulted when the substrate pH was reduced to pH 3.5. Medium activity was evident 296 ARTHUR D. HASLER between pH 4-6 which indicated the presence of a katheptic or auto- lytic enzyme of the tissues. The alcoholic extracts of Calanus made by Bond (1934) were most active at pH 8.0-8.49. The secondary opti- mum of his extracts at pH 3.6-4.0 were unconfirmed in this work. The secondary optimum re>ulting from the autolytic enzymes was pi I 5.6. The discrepancy <>f pi I optima may be due to variations in extracting technique, for Mansour-Bek (1932) found that crude proteinase extract of Maja squinado had a pH optimum on gelatin of 6.0; the optimum for purified en/vim- was pH 8.1. The pH of the buffered substrates used in this experiment was determined with the aid of the glass electrode. Am\lasc. — Both Diaptomus and Calanus extracts contained active amylase which hydrolyzed starch. The iodometric titration previously 0.5 IMC. 7. 1 1 ydntlyMs <>f trilmtyrin by lipases of I )iti/^!/>iinis (/') and Calantts (C). The nnlinak' indicate the miniluT uf cc. of N/20 NaOH ; the abscissa is the time in hours. describee] was used as a criterion of starch hydrolysis. The amounts of N/40 Xa.,S,( ).. present in 10 cc. of digest can be read from Fig. 6. These figures are equivalent to the amount of N/40 reducing sugars lib- erated by hydroKsis. In both instances 5 cc. of 1 per cent glycerol extract were added to 25 cc. of 3 per cent starch at pll 7.2 and 25 cc. of water and incubated at 37° C. At intervals 10-cc. samples were with- drawn and the reducing groups titrated. Lipasc. — Trilmtyrin was attacked by an esterase present in both Di- aptomus and Calanus extracts. Fifty cc. of 4 per cent tributyrin were emulsified with sodium glycocholate to which were added 5 cc. of 1 per cent extract of Diaptomus; extracts of Calanus were examined by the same procedure. At intervals 10-cc. aliquots were withdrawn and ti- trated with X, 20 XaOII. The results are shown in Fig. 7. DIGESTION IN PLANKTON CRUSTACEA. II 297 Experiments in all instances were duplicated and simultaneously run with adequate controls. DISCUSSION Biochemical analyses presented in Table I, and by the author in a previous paper (1935), demonstrate that Daphnia possesses a system of proteolytic enzymes which can attack a protein (in this case gelatin and casein) and cleave it to amino acids. This bears out the contention that plankton crustaceans can utilize particulate organic matter. The experiments on temperature effect on enzyme activity show that the Daphnia and hog enzymes follow, in general, the van't Hoff rule that a reaction approximately doubles for every 10° C. increase in tempera- ture. It suffices to conclude from these data, that a low temperature does reduce the entire metabolism of the organism. More data should be collected on this subject, for Wiersma (1928) held that Astacus amy- lase was strongly active at 0° C., while Yonge (1926) found that en- zymes of Ostrea were inactive at this temperature. Another cladoceran (Polyphemus) and a copepod (Dlaptonius) have been added to the list of plankton Crustacea which can be definitely said to possess functional digestive enzymes capable of attacking the three important food stuffs. It is quite clear from the review of Kriiger (1933) and Yonge (1931) that these entomostracans together with Daphnia do not differ from the enzyme types of the larger Crus- tacea such as Astacus and Maja. Although this wrork concludes the long series of experiments designed to solve the problem of nutrition in plankton Crustacea as challenged by Putter, there remains an untouched field in enzyme chemistry of Crus- tacea, namely the kinetics and purification of these enzymes. A study of the kinetics of purified enzymes has been made by Mansour-Bek (1932) on the decapod, Maja squinado. She brought to light some interesting similarities between vertebrate and invertebrate enzymes, e.g. purified proteinase was activated by mammalian enterokinase. A continuation of her stimulating work would be an excellent contribution. Difficulties may be expected, however, in working with entomostracans, for the present purification methods require large amounts of extract to carry out proper elution. Appreciation is expressed to Professor H. C. Bradley for furnishing laboratory facilities and for gladly proffering biochemical advice ; also to Professor C. Juday for biological suggestions and personal interest in the problem. 298 ARTHUR D. HASLER SUMMARY 1. Three peptidases (dipeptidase, aininopolvpepticla.se and carboxy- polypeptidase) were found in water extracts of Daphnia. The presence of these and a proteinase (author, 1935) demonstrate that Daphnia can completely utilize a protein. 2. Proteinases of Daphnia were no more active than those of the hog pancreas at a temperature of 8° C. Both enzyme systems follow, in general, the van't Hoff rule. 3. Proteinasi's. amylases and lipases were found in glycerol extracts of Polyphemus and Diaptomus. All proteinases were of the tryptic type with an optimum activity at pll 7-8. They simulate Daphnia en- zymes and show that plankton Crustacea have complete enzyme mecha- nisms for digestion of participate food. LITERATURE CITED BAKER, J. L., AND H. F. E. HULTON, 1920. The ioclometric estimation of sugars. I! indicia. Jour., 14: 754. BOND, R. M., 1934. Digestive enzymes of the pelagic copepod. Calanus fin- marcliicus. Biol. Bull., 67: 461. CLARKE, G. L., 1934. The role of copepods in the economy of the sea. Fifth Pacific Science Congress, Vancouver, B. C. AS. 5., 2017. CLARKE, G. L., AND S. S. GELLIS, 1935. The nutrition of copepods in relation to the food-cycle of the sea. Biol. Bull., 68: 231. DEHN, v. M., 1930. Untersuchungen iiber die Yerdauung hei Daphnien. Zeitsdir. vergl. Physlol, 13: 334. GELLIS, S. S., AND G. L. CLARKE, 1935. Organic matter in dissolved and in col- loidal form as food for Daphnia magna. 1'hysiul. ZooL. 8: 127. HASLER, A. D., 1935. The physiology of digestion of plankton Crustacea. T. Some digestive enzymes of Daphnia. Biol. Bull., 68: 207. JOHNSON, M. J., G. H., AND W. II. I'KTERSON, 1936. Tin- magnesium-activated leucyl-peptidase of animal erepsin. Jour. Biol. Chan.. 116: 515. KRUGER, P., 1933. Vergleichender Fermentstoffwechsel der niederen Tiere. Erg. 1'hysiol.. 35: 538. LiNDF.RSTR0M-LAN<;, K., 1927. Volumetric determination <>f umino nitrogen. Comfit, rend. trav. lab. Carlsberg., 17: 4. MANSOCK-I'.KK, J. J., 1932. Die Proteolytischen Kn/yme von Maja squinado Latr. Zcitschr. vergl. Physiol.. 17: 154. NORTHROP, J. II., AND 1\. G. HussEY, 1922. A method for the quantitative deter- mination of trypsin and pepsin. Jour. (/<•«. Physiol., 5: 353. WALDSCHMIDT-LEITZ, E., 1931. The mode of action and differentiation of proteo- lytic enzymes. Physiol. Re?'.. 11: 358. WIERSMA, C. A. G., AND R. VAN DER VEEN, 1928. Die Kolilehydratverdauuiig bei Astacus fluviatilis. Zeitsdir. vergl. Physiol.. 7: 269. Yo\<,K, C. M., 1926. Structure and physiology of the organs of feeding and di- gestion in Ostrea cdulis. Jour. Mar. Biol. Ass'n., 14: 295. YONOE, C. M., 1931. Digestive processes in marine invertebrates and fishes. Jour. Conscil. Int. ILvflor. dc la Mcr., 6: 175. SYMMETRY AND REGULATION IN MNEMIOPSIS LEIDYI. AGASSIZ1 B. R. COONFIELD (From the Department of Biology, Brooklyn College, and the Marine I'ioloyical Laboratory, \\'oods Hole, Mass.') The ctenophorcs, a group in which Mncniiopsis is included, are usu- ally referred to as biradially symmetrical animals. By some they are believed to have primarily a radial symmetry. There is of course the idea of bilateral symmetry expressed in the term biradial. As one sees this animal swimming freely in water it does seem to possess a radial symmetry. This impression is received because its organs are distrib- uted about a central, longitudinal axis. But a close examination of this animal reveals that most of its organs are paired and are located on two opposite sides of the central, single organ, the stomodeum (Figs. 1 and 11). A method other than and including the study of the anatomy of this animal has been used by me in getting information about its sym- metry. This information was obtained during a series of experiments which had been planned to show other characters of this animal. The one feature which has been demonstrated repeatedly during experimenta- tion on Mnemiopsis is the power to regulate itself. This power to regu- late its body following injury either in nature or caused by experiments seems to have some bearing on symmetry. I propose, therefore, to show in this report that Alnciniopsis possesses a particular type of symmetry, bilateral symmetry without dorso-ventrality, as well as the power to regulate itself, and that in this animal symmetry and regulation are asso- ciated with each other. EXPERIMENTS The operations were performed on the specimens and each was kept in a separate finger bowl which was partly immersed in running water. Observations, drawings, and photographs were made at frequent inter- vals. Only a very few of the experimental animals failed to survive. The changes observed during the reorganization following the operations were recorded in detail. In a great many instances these changes were observed quite easily with only the aid of glasses of low magnification. The drawings were made with the aid of a camera lucida while the pho- tographs were taken with a Leica camera with an extension tube outfit. 1 Contribution No. 19 from the Department of Biology, Brooklyn College. 299 300 B. R. COONFIELD Although the experiments listed and described in this report seem to be quite varied they can he grouped under two headings; experimental cuttings, and experimental graftings. EXPERIMENTAL CUTTINCS The animals were1 cut across at four levels of the body and only the oral pieces were retained for observation. By this cutting at different levels some of the oral pieces lost only the very apical tip of the animal AE AE Fi<;. 1. This is a diagram representing the shape and symmetry of the body of Mnemiopsis. The plates and sonic of the smaller canals arc not sh<>\vn. . //:, adesophagcal row; .//'. apical organ; ./'/'. adtentacular row; /•". infnndib- uluni ; T, tentacular canal. The stomodeum is shown in the center of the drawing with the ]iara-ja>tic canals lying parallel to and near it. while others had removed from them the part of the body apical to the bases of the auricles. These experiments were done in accordance1 with an objective somewhat different from the one included in this report. The details of these experiments will hi1 reported later. However, as the pieces of the animal were passing through the phases of regeneration several results were pertinent to the problem considered herein. The oral pieces usually regenerated the lost portions of the body per- fectly in all details. Hv this regeneration the regular connections of canals and rows were soon established and the single, apical organ was SYMMETRY AND REGULATION IN MNEMIOPSIS 301 formed (Figs. 9, 15 and 17). During regeneration the two adjacent adtentacular rows and a single adesophageal row on either side of these connected to each other. As this was taking place the tip of the stomo- deum formed a single hull) (Fig. 18), which immediately divided to form two hulhs (Fig. 4), and each of these became connected with a set of ST FIG. 2. An adesophageal view of Mnemiopsis showing a stomodeal graft after the second day. FIG. 3. This drawing shows an apical organ grafted to the adtentacular re- gion and about mid-way between the ends of the body. The drawing was made on the fourth day after the operation. FIG. 4. This shows the usual features at the apical end of the body during regeneration following the removal of an apical part of the animal. The two bulbs can be seen lying at the upper end of the stomodeum. This drawing was made on the fourth day. AE, adesophageal row; AT, adtentacular row; B, stomodeal bulb; G, graft; P, paragastric canal ; 57', stomodeum ; T ' , tentacular canal. four canals of the plate rows. The two bulbs became smaller until they assumed the size of a tube or canal and became the radial canals. At about this time also a single, apical organ formed at the tip of the sto- modeum between the two bulbs. The apical plate of this organ was regularly laid down on a line between each of the two adjacent pairs of adesophageal rows (Fig. 9). 302 R. R. COOXFIELD Certain irregularities were observed on the regenerating oral pieces following the cutting at each of the four levels. These consisted of a failure of the eight rows to connect as a unit resulting in the formation of two apical organs < Figs. 7 and 19), a failure of the usual row con- nections with or without an apical organ (Figs. 10, 14 and 16), and a splitting of the tip of the stomodeum resulting in a separation of the opposing pair> of ade>ophageal rows (Fig. 8). Here a single apical organ regenerated between each pair of rows and the adtentacular rows 6. FIG. 5. This shows, the result of ^raftint;- two apical pieces together. Each piece represents les: ili.ui half of an animal. The para.uastric and tentacular canals are not shoun. This drawing was made mi the sixth day. Fi<;. (). This sh»\vs the result of yraftm.i; together two equal and longitudinal pieces of different animals. I'.ach piece contained slightly more than one half of the liody. The drawing was m.ide on the third day. .//., adeMipha.ijcal row; ./'/'. adtentacular row; M, mouth; ST, stomodeum. did not regenerate. Although a few of these irregularities occurred on the oral piece following the cutting at each of the four levels of the hody. a majority were formed hy the oral pieces produced by a cross cut at the liases of the auricles (Fig. 11 ). The two regenerated apical or- gans of a single, oral piece were always at first oriented at right angles to the n-gular one ( Kigs. 7 and 8). Within a few days, however, these organs began to turn and finally they assumed the regular direction of orientatu >\\. SYMMETRY AND REGULATION IN MNEMIOPSIS 303 EXPERIMENTAL GRAFTINGS Several types of graftings were done and since they all have hearing on the problem of regulation and reorganization they are listed and described here under the one heading. A few of these experiments wrere of an exploratory nature hut their results were definite enough to make them significant. Each type of grafting is described here in a single paragraph. 1. Four mid-pieces of separate animals were grafted together with their regular orientation maintained, and the apical and oral ends of a fifth animal were grafted to the apical and oral regions respectively of the fused mid-pieces. By this method of grafting, an animal consisting of the mid-pieces of four animals with the apical and oral ends of a fifth animal wras formed. All of the pieces fused at the cut regions and all canals and the stomodeum of each fused in the regular manner. Even the coordination of the plates as well as the feeding reactions at the one mouth were similar to these processes of a normal animal. 2. The apical end of an animal was grafted to the mid-adesophageal surface of another animal. This end fused to the host successfully. Soon after the fusion at the edge of the graft was complete a mouth broke through at one adtentacular region. The canals and rows of the host and the graft fused with each other and the plate movement con- tinued in the regular manner of coordination and direction in each por- tion of the experimental animal. 3. Two large portions obtained by cutting two animals lengthwise at a slight angle and to one side of the stomodeum were grafted to each other (Figs. 12 and 13). In some the two large portions of animals of an equal size \vere fused together (Fig. 13) while in other experiments the portions were from animals of unequal size (Fig. 12). There was no apparent difference in the results of the two types of experiments. In all cases the two pieces fused within three days and the corresponding rows fused with each other. This fusion was near the ends of the rows and since two adjacent rows on a normal animal connect to each other there was no irregularity in this fusion. The plate coordination in each piece was quite regular and independent of the other piece. 4. Two pieces were obtained by making a longitudinal cut near the stomodeum and parallel to it through the entire length of the animal. The larger of these two pieces was grafted to another one of similar size and origin. In this graft the orientation of both pieces was un- changed. Fusion took place within three days and the graft continued to live as a single animal. After the operation this organism possessed two apical organs, a single stomodeum, a double number of tentacular 304 B. R. COONFIELD canals, a 1>ifu routed paragastric canal, and a single month (Fig. 6). The behavior of this animal was similar to that of any normal one. 5. The two small portions of the animals which were cut in the ex- periments described in the section above were grafted together. Fusion took place and within six days a single apical organ regenerated and the fused pieces formed a normal animal. 6. The apical pieces of two animals were grafted together at their oral regions so as to form an animal having two opposing apical ends. These two pieces fused and healed very quickly. Although fusion took place, each piece maintained its identity by forming a mouth, by co- ordinating its own plate movement, and by regenerating auricles and lobes (Fig. 5). 7. Two halves obtained by cutting animals longitudinally through the adtentacular plane were grafted with their orientation reversed. The oral end of one piece was in contact with the apical end of the other piece. Fusion took place immediately and within six days the two halves began to rotate on each other, bringing the similar ends of each piece near each other. This rotation was not carried very far, however, before the specimens died. 8. The apical organ of one animal was grafted to the surface of another animal at various levels of the body. The graft contained some of the infundibulum and the bases of the ad radial canals. Fusion took place immediately and the short pieces of the adradial canals connected to each other at first and later they connected to the adtentacular canals EXPLANATION OF FIGURES 7-11 FIG. 7. This shows the regenerating end of an oral piece. The oral piece was obtained by making a cross cut through the body at the level of the auricles. Drawn on the fourth day. FIG. 8. A drawing of a regenerating oral piece obtained as in Fig. 7. Here the two sets of adcsophageal canals are connected to the stomodeum by a separate infundibulum. The adtentacular canals failed to regenerate. Drawn on the fourth day. FIG. 9. This shows the regular method of regeneration following the re- moval of an apical piece. This specimen had been cut across at the level of the auricles. The eight rows are formed and the single apical organ is shown in its regular orientation. Drawn on the fourth clay. FIG. 10. This specimen was obtained as in Fig. 9. The result of regeneration failed to bring the canals and rows to the regular type of connection. Drawn mi the fourth day. FIG. 11. This is a diagram to show the distribution of the canals and other organs in an oral piece obtained by cutting the animal across at the bases of the auricles. AR. adesophageal rows; /•", infundibulum; L, lobe; P, paragastric canal; ST , stomodeum ; T, tentacular canal. SYMMETRY AND Kl<:< It'LATIOX IN MNKMIOPSIS 305 AE 11. of the host (Fig. 3). The graft always became oriented in a regular manner after its canals joined those of the host regardless of its previous orientation. Later the graft joined the stomodeum of the host (Fig. 3). 9. A small portion of the mid-region of the stomodeum of an animal was grafted to the surface of another animal between its adtentacular rows. Very soon after the healing was complete the stomodeum of the host curved out toward the graft. This bending out toward the graft took place before there was any mechanical connection between the graft 306 B. R. COOXFIELD and the stomodeum. Eventually the graft and the stomodeuin fused and in addition usually the tentacular canal of the host became connected to the graft (Fig. 2). The graft formed an opening at the surface of the host and thus provided an additional mouth. DISCUSSION In considering the problem of symmetry in Mnemiopsis it is natural to observe first the anatomy of this animal. The organs are equally distributed along the opposite sides of the main, central, longitudinal axis of its body (Figs. 1 and 11 ). There are four plate rows, two auricles, one paragastric canal, one tentacular canal, and one lobe on each side of the stomodeuin. The apical organ and the mouth are at the two poles MI' tlu- stomodeum and they lie in a plane at right angles to it. Then as far as the organs, with the exception of the apical organ and the mouth, are concerned they are distributed in the body bilaterally about the sto- modeum. Therefore this distribution of organs indicates a bilateral symmetry without a dorso-ventfality. Does Mnemiopsis have any features other than its anatomy which might give some information about its symmetry? In answering this question I shall consider certain observations made on experiments in this animal. The outstanding feature of this animal during regeneration EXPLANATION OF PLATE I FIG. 12. This shows two longitudinal pieces from animals of different sizes fused together. This fusion was complete by the end of four days. FIG. 13. This is a photograph of the fused two longitudinal pieces from ani- mals of equal sizes. The fusion was complete by the end of the fourth day. FIG. 14. An apical view of an oral piece obtained by cutting an animal across through its infundihulum. This shows the irregular row- connection and two apical organs have regenerated. Photographed on the third day. IK;. 15. An apical view of an oral piece obtained as shown in Fig. 14. This shows the regular row connection and but a single apical organ has regenerated. Photographed on the third day. FIG. 16. An apical view of an oral piece obtained by cutting an animal across at the auricular region. This shows the irregular row connection and two apical organs have regenerated. Photographed on the third day. FIG. 17. An apical view of an oral piece obtained as is shown in Fig. 16. This shows the usual row connection with but one apical organ regenerated. Photographed on the third day. FIG. 18. This show's the formation of a single bull) at the apical end of the stomodeum. This follows the cutting off of the apical portion of the animal. The rows which arc not continuous at the top of the photograph were broken during the handling of the specimen. FIG. 19. An apical view of an oral piece obtained by making a cross cut at the region of the auricles. This shows the usual connections of the rows and two apical organs have been formed during regeneration. Photographed on the third clay. PLATE I 308 B. R. COOXFIELD is its ability to regulate its own reorganization following any type of disturbance to its body that I have made. This power to regulate has been remarkably demonstrated during the process of regeneration. \Yhen this animal has been cut into halves, thirds, or fourths either by cross cuts or by longitudinal cuts the removed parts were regenerated perfectly with but few exceptions (Coonfield, 1936a). Following the cutting of its body across at any level above the mouth the procedure of regeneration showed a bilateral arrangement with but few exceptions. The apical end of the stomodeum formed a bulb (Fig. IS), then this bulb split in two (Fig. 4). Later these bulbs thinned out and formed the two radial canals. I loth the tentacular and the paragastric canals established their regular connections with these two bulbs (Fig. 4) and as these bulbs disappeared as such these canals became connected to the base of the infundibulum ( Fig. 2). In the meantime the apical end of the stomodeum became- elongated and formed the infundibulum with a single apical organ formed on the end of this structure-. During the regeneration of the apical organ the plate rows were regular in their reformation. Two adjacent adtentacular rows and a single adesopha- geal row on either side of them became connected to each other at first (Fig. 9). Then as the bulbs became the radial canals the plate rows established their usual connection to the infundibulum by way of the radial canals. The- regular organization was complete alter the apical organ had regenerated. How can the irregularities in regeneration be accounted tor even though they are relatively few in number? These irregularities have appeared following the cutting of the animal into halves, thirds, and fourths (Coonfield, 1936«). A very few of the regenerating pieces failed to reform the lost plate rows. A few irregularities an- given in the present report. These are concerned with the failure of the rows to become connected in the regular manner ( Figs. 10, 14 and 16). or with the regeneration of t\\o apical organs instead of a .single one (Figs. 7. 8, 14, 16 and 19), or with a considerable splitting of the stomodeum to form two infundibula (Fig. 8). Fven though I cannot offer a satisfac- tory reason for the- appearance of these irregularities 1 believe they are significant in supporting the view of bilateral symmetry. For in all except three or four that failed to reorganize the plate rows in the regu- lar manner the reforming organs were distributed according to bilat- erality. So then the regulation of the reforming organs in the body of Mnemiopsis whether they proceeded in the regular manner or exhibited some irregularities supports the view of bilateral svmmetry. Information having a direct bearing on the problem ot .symmetry and regulation in addition to that shown by a study of the anatomy of Muc- SYMMETRY AND REGULATION IN MNEMIOPSIS 309 sis and observations on regeneration in tin's animal was obtained by performing experiments involving grafts. Tliis animal showed a re- markable ability to regulate itself both according to structure and physi- ology in response to this type of experiment. The ability of this animal to organize the grafts in experiments wherein a whole body was formed by fusing the mid-pieces of four animals with the apical and oral pieces of a fifth animal, the noninterruption of the regular activities of both the graft and the host following the fusing of an apical piece at right angles to the surface of the host, and also wherein a large portion of two ani- mals were fused together (Figs. 12 and 13) show the ability of this ani- mal to regulate itself. The reorganization following the fusion of two longitudinal parts of the body (Fig. 6) shows that the animal almost succeeded in organizing a normal body. This opinion is confirmed by noting that the two parts of the stomodeum fused as one and with but one mouth, the two paragastric canals on each side of the stomodeum united all along except near the infundibulum, and the actions of this animal were similar to those of a normal one. In the experiments whereby two apical pieces were fused to each other at their oral regions the ability of each piece to regulate itself although being fused to an- other similar one was definitely shown. Further evidence supporting the view of regulation in a physiological manner was obtained by observing polarity in Mnemiopsis following ex- periments involving grafts. Polarity has been observed previously in this animal (Coonfield, 1934 and 1936/?). The data contained in the present report show that when two halves of an animal were grafted together with their orientation reversed each piece attempted to rotate so as to bring the two similar regions together. Also when an apical organ or a piece of the stomodeum was grafted to the surface of an animal these grafts always became oriented and established connections to the plate rows or the other canals, and to the stomodeum (Fig. 3). It is interesting to note that the apical organ became oriented only after connections to the canals had been established and both the canals and the stomodeum were attracted to the stomodeal grafts immediately after each had become healed in the body of the host. The apical grafts per- sisted long enough to show that their form and function were unchanged by their newr location. Also the stomodeal grafts which were taken from the side of the stomodeum not only continued to function as a stomodeum but in addition each formed a mouth. In summing up the data obtained from observing the various types of grafts which were car- ried out on Mnemiopsis, it seems to me that these data show that regula- tion is demonstrated by them in this animal. Since regulation in this 310 B. R. COONFIELD animal can be associated \vith symmetry, it seems that these experiments support the view of bilaterality without dorso-ventrality. CONCLUSIONS 1. Mnemiopsis poss( ss< - ;i bilateral symmetry without a dorso-ven- trality. This i> >li<>\vn by the anatomy of this animal and by its method of regulation t"wing a disturbance to its body. 2. Symmc'try and regulation are two closely associated features of Mnemiopsis. 3. Regulation is demonstrated in Mnevniopsis by the following re- sults. (/. Reorganization during regeneration which follows the cutting of its body. /'. Reorganization during the fusion, orientation, and connections of grafts. LITERATURE CITED • Coo.NKiEi.n, B. R., 1934. Coordination and movement of the swimming plates of Mncmiopsis leidyi. ALSIS-O/. Biol. Hull.. 66: 10. COONFIELD, B. R., 1936(7. Regeneration in Mnemiopsis leidyi, Agassiz. Biol. Bull., 71:421. COONFIELD, B. R., 193(i/i. Apical dominance and polarity in Mnc-minpM^ leidyi, Agassiz. Biol. Bull., 70: 460. ON THE SIGNIFICANCE OF THE POLAR SPOT IN RIPE UNFERTILIZED AND IN FERTILIZED ASCIDIAN EGGS A. DALCQ (Brussels) AND G. VANDEBROEK (Ghent) Cohen and Berrill (1936) have given an interesting account of observations on eggs of Ascidiella aspcrsa and Phallusia rnainillata. After the jelly was digested by the gastric juice of the crab Munida, the eggs were stained in toto with a vital dye (particularly Nile-blue sulfate) and subsequent events studied with special reference to changes in the form of the germ and in the appearance of the vital dye within the egg. These are described in the ripe egg, from the time of fertilization until the formation of the polar bodies, and during cleavage, gastrulation, and formation of the tadpole. In the ripe unfertilized egg they have observed the " polar pit " (" tache polaire " of Ascidiella — Dalcq, 1932) though they question whether this cortical differentiation marks the position of the maturation spindle and identifies the location of the animal pole. Their doubt is not based on continuous observation of the region of the polar pit from which the dye entirely disappears before the elevation of the polar body, nor have they been able to determine directly whether the polar body is actually formed in the polar pit or not. They are influenced, however, by certain cases in which foreign particles adhere to the cortical layer of naked eggs in the vicinity of the polar pit. On measuring the distance between such a particle and the pit, and, later, between the same particle and the polar body, they found considerable variation in the distances. In two cases, drawings of which are given in their paper, the distance was increased; the same was true of "several" other eggs. In one case the distance did not change. While Cohen and Berrill consider the possibility of a displacement of the marking particle by some move- ment of the cortical layer, they discard such an explanation on the following grounds. During the maturation period the egg shows, grossly, only a flattening and an elongation, while the deformation dur- ing cleavage, though much greater, is unaccompanied by any change in the distance between the polar body and the adhering particle. Hence they do not believe it possible that the earlier deformations of the egg could be responsible for the displacement of the small body fixed to the cortex. They feel justified in stating, therefore: " Since Dalcq accepted the site of the polar pit as being identical with that of the polar body and 311 312 A. DALCQ AND G. VANDEBROEK used it to orient the eggs for cutting, it follows that some doubt is thrown on the validity of his conclusions regarding the localization of presumptive germinal regions in the fertilized egg of Ascidiclla aspcrsa, although, of course, not on the existence of such presumptive regions " (p. 84). They do not state what, in their opinion, the significance of the polar pit is. On the basis of experiments cited below we have good reason to believe that their conclusion is not justified. It is unfortunate that Cohen and Berrill have discarded the possible explanation of the behavior of the attached particles in their experiments on the basis of cortical movements. Such movements are common in eggs and may be easily identified by means of definitely localized vitally stained areas which may be followed through ensuing stages. This method was used by one of us (G. V.) at Roscoff * during the summer of 1935 with outstanding success. \Ye shall mention hen- only the re- sults necessary to dispel the doubt thrown by Cohen and Herrill on the significance of the polar "pit," perhaps better called the polar "spot." In addition, we shall describe briefiy an anomaly observed by the senior author (A. I).) which gives further proof of the existence of a matura- tion spindle immediately under the polar spot. The ripe egg of Ascidiella - contains the first maturation spindle, as stated by Cohen and Berrill, and not the second, as previously stated by Dalcq on the basis of the examination of unfertili/.ed eggs and the ap- pearance of the spindle in sectioned preparations. On continuous ob- servation two polar bodies may be seen to be successively extruded by the egg after fertili/ation. The jelly of the e^g was removed with the aid of mounted needles (watchmakers' forceps can also be used), some dozens of naked eggs easily being prepared in but a few minutes. The method is not tedious. In tlu- discussion of merogony experiments Dalcq has already called attention to the fact that after insemination of naked eggs, the penetra- tion of the sperniato/oon into the egg may be delayed by as much as 1 \Ve \vish to express our gratitude to Professor Dr. Ch. Perez and Dr. G. Teissier for tin exi > llcnt working opportunities accorded both of us during several stays at the " Labonitoiiv Lacaze-Duthiers." - Because of some confusion as to the identification of the species used, it is necessary to state that our experiments were made on the Ascidiclla which is found in abundance in the great pool of the Roscoff Laboratory, and which was used by Dalcq and by Tung in 1931 and 1932, and called by them Ascidiella aspcrsa. Berrill (1928) suggests that, because of the small size of the animal and the fact that the eggs do not float, the Roscoff species should be called Ascidiclla scahra. the name of Ascidiclla aspcrsa being reserved for the great Ascidiclla, with floating eggs, common in Plymouth (and in the vicinity of Boulogne). Dalcq and Tung, following this suggestion, used the name Ascidiella scahra in their subsequent publications. SIGNIFICANCE OF POLAR SPOT IN ASCIDIAN EGGS 313 half an hour. Yandebroek has more recently discovered, in the course of experiments which will he reported in detail elsewhere, that the mo- ment of sperm penetration can he recognized hy a series of character- istic deformations of the egg which persist for ahout two minutes. The egg then returns to the spherical condition until the time of first polar hody formation. As is noted helow, the first and second polar bodies are formed 6 and 20 minutes respectively after the penetration of the spermatozoon. The actual times vary with temperature but the above are those observed in the present experiments. In dealing with eggs in CM. M M CJJ. M? CM. FIGS. 1 AND 2. Three views of eggs subjected to local vital staining: (a) ripe egg; (b) after the formation of the first polocyte ; (c) during the formation of the yellow crescent and hyaline plasma zone. Colored territories stippled in deep black, when superficial invaginated parts (Figs. 16 and c~) marked with ooo; p.s.: polar spot; pb1: first polocyte; f>b1 + ~: the two polocytes ; M.: mesoblastic side; C.N.: chordoneural side; c.m.: mesoblastic region (yellow crescent). The two eggs are viewed from their left sides. the brief period preceding maturation it is especially important to recog- nize the delay mentioned above and to identify the moment of sperm penetration. This, unfortunately, Cohen and Berrill have failed to do. In the experiments reported below, localized areas of denuded ripe unfertilized eggs were vitally stained with the dye ' Brillant Cresyl.' The stained areas have a diameter of from 50 to 75 /j. (ca) and persist, perfectly localized, during the hours immediately following fertilization. If the egg is not fertilized the dye rapidly diffuses. Three experiments which demonstrate the significance of the polar spot will be described. 314 A. DALCQ AND G. VANDEBROEK In the first case (Figs. !'. c' : viewed from the mesoblastic side. Other ab- breviations as in Figs. 1 and 2. creased their distance from the polar end by about 30° (Figs. \b, 2b, 36, and 3//). They maintain these positions in spite of the considerable deformations which the egg undergoes during the period of expulsion of the two polar bodies. In the ease1 illustrated by Fig. 1, the presence of the mark enables us to determine that the- first polar body is actually formed in the polar spot about (> minutes after fertilization. The sec- ond polar body is formed next to the first about 14 minutes later (Figs. Ib and If). It appears, there-fore, that the polar spot is also the matura- tion site. In the post-maturation phases during which the female pronucleus migrates toward the center of the egg to meet the male pronucleus and SIGNIFICANCE OF POLAR SPOT IN ASCIDIAN EGGS 315 the yellow crescent is gradually formed, the behavior of the marks in the vegetative pole is very interesting. This is to be described in detail elsewhere (by G. V.) but is suggested by the mark (3 of case 1. In the present consideration we are directly concerned with marks on the ani- mal half of the egg and these consistently show a movement toward the animal pole which results in their return to their original positions (Figs. \c, 2c, 3c, and 3c'). They show no further change during the cleavages which follow. The appearance of the yellow crescent and the adjacent hyaline protoplasm which contains the fusion nucleus allows one to determine definitely the orientation of the egg. Using these landmarks, it was found that in the egg of Fig. 1, the mark a was in the plane of sym- metry, just above the chordoneural material, while the mark ft lay under the future yellow crescent, i.e. on the mesoblastic side of the germ. In the second egg the mark had been put between the polar spot and the presumptive mesoblastic crescent, nearly in the plane of symmetry. In the third case, on the other hand, the mark was located in a region to one side, in the left half of the germ. In these three typical experi- ments, therefore, the chief regions around the polar spot have also been explored. To summarize briefly, it has been found that, after fertilization, the cortical layer of the ascidian egg undergoes definite displacements. During the initial period of active deformations, the cortical layer sur- rounding the animal pole moves toward the equator. Later, when the conjugation of the pronuclei takes place, the same material is again shifted nearer to the animal pole. In conclusion, it may be stated with- out the least hesitation that the polar spot indicates the site of formation of the polar bodies, and is the result of the presence of the first matura- tion spindle just below the cortex. In view of these conclusions the question now arises as to how the data of Cohen and Berrill are to be interpreted. It seems clear that the variations they record in the distance between the adhering particle and the animal pole result from the activity of the cortical film ; the cases in which the distance did not change may be explained on the assumption that the eggs were first observed somewhat too late. This explanation is not entirely satisfactory, however, for they appear to have observed some of the eggs continuously and yet fail to record any secondary shift of the materials towards the animal pole. In this connection we should like to make several suggestions inasmuch as it must be admitted that observations made by means of adhering particles are less reliable than those based on local vital staining of parts of the egg itself. The possi- bility of a displacement of the particle during the later deformations of 316 A. DALCQ AND G. VANDEBROEK the egg should he examined. It is possible that the particle might eventually penetrate the cortical film and adhere directly to the under- lying granular cytoplasm. For the latter form, the morphological value of the polar spot is now established beyond any doubt and it would appear that Dalcq was justi- fied in basing hi.s merogony experiments on this indication of the animal- vegetative axis of the i-^i;. Sectioning of the ripe ascidian egg in various planes relative 1" this primary axis and subsequent fertilization FIG. 4. KYo >iistruetii >n of a lixrd abnormal .Iscidicllii e.yi;'. Fixation, Allen; Sections, 6M; Staining. Holland*.-. IVi^peetivc vieu a! 45 (method of I.ison). a, b, c: the three maturation spindle>. The eosinophil Mibeortical material is stippled. of both pieces seems to be at present the >ole method of analysing the distribution of the potencies at that Mage. Kurther experiments have been carried on by the senior author during re-cent years and a short account of the results has been presented (Dalcq, 1935). This has led to a furtluT eytological study of the structure of the ripe egg, and to tin- discovery of a preexisting bilateral symmetry indicated by the arrange- ment of eosinophil subcorlical granules (the mate-rial of the future yel- low crescent) in the lonn ot a large crescent which is more developed on one side of the primary axis. This important feature of the organ- ization of the ripe Ascidiclla egg together with the variations in the structure of merogonic twin embryos, whose anatomy has been thor- SIGNIFICANCE OF POLAR SPOT IN ASCIDIAN EGGS 317 oughly studied in a quantitative way, with several graphic reconstruc- tions, will be fully described in another place. In relation to the problem of the significance of the polar spot con- sidered here, it seems important to draw attention to an anomaly which has to be borne in mind when performing merogony experiments. Among numerous sections of ripe eggs, the senior author has observed, four or rive times, eggs which appear to possess two polar spots. Such eggs were at first discarded without further consideration, but later, a similar case was fixed and sectioned. The slide shows that the egg con- tains no less than three maturation spindles. In Fig. 4 may be seen a graphical reconstruction of the egg by Lison's method (1936). The three spindles lie in the region free from the eosinophil granules which nearly cover the vegetative hemisphere. The three spindles are small, of approximately equal size, and each supports some chromosomes. Owing to the direction of the plane of sectioning, which was more or less parallel to the primary axis of the egg, two of the spindles (a and b} lie under the part of the cortex toward the observer of the drawing; the third c, is situated on the opposite side, not so far from b as it ap- pears to be on the drawing. The infrequent occurrence of such eggs is of some importance to the investigator who is performing merogony experiments and throws some light on the significance of the polar spot. It is clear that the existence of a secondary spot corresponds here to a supplementary spindle. Why twro and not three spots were seen may be explained in either of two ways; either one of the spots escaped the eye of the ob- server, or one of the spindles did not adhere sufficiently to the cortex to be seen /'// vivo. No positive information has been obtained concerning the origin of such anomalies. It seems probable that the spindle mate- rial became divided when the rupture of the germinal vesicle took place. SUMMARY Ripe Ascidiclla eggs have been subjected to local vital staining, chiefly in the regions surrounding the animal pole. Fertilization is im- mediately followed by a shift of the cortical layer towards the equator; when the yellow crescent appears, the material returns to its original position. The continuous observation of the ripe and the fertilized egg shows, contra Cohen and Berrill, that the polar spot marks the place where polocytes will be extruded. In addition, attention is called to cases where there are two or three maturation spindles in the same egg. 318 A. DALCQ AND G. VAXDEBROEK LITERATURE CITED BERRILL, N. J., 1928. The identification and validity of certain species of ascidians. Jour. Mar. B'wl. Assn., 15: 159. COHEN, A., AND N. J. BERRII.I., 1936. The early development of ascidian eggs. Biol. Bull., 70: 78. DALCQ, A., 1932. fitude des localisations germinates dans 1'oeuf vierge d'Ascidie par des experiences de merogonie. Arch. d'Anat. Micros., 28: 223. DALCQ, A., 1935a. L'organisati»u (k- 1'oeuf chez les Chordes. Gauthier-Yillars a Paris. DALCQ, A., 19356. La regulation dans le germe et son interpretation. Coinpt. rend. Sac. dc Biol., 119: 1421. LISON, L., 1936. Une method^ in>uvelle dc reconstruction graphique perspective. Bull. d'Histol. J /'/>/., 13: 357. TUNG, Ti-CHow, 1932. Experiences de coloration vitale sur 1'oeuf d'Ascidiella aspersa. Arch, de Biol., 43: 451. TUNG, Ti-Cnow, 1934. Kecherches sur les potentialites des blastomeres chez Ascidiella M.'ahra. Experieiia-s de translocation, de combinaison et d'isole- ment de blastomeres. Arch, tl'.lnat. Micros., 30: 381. REPRODUCTIVE SYSTEM AND COPULATION I> AMPHI- SCOLOPS LANGERHANSI (TURHKLLARIA ACOELA) LIBBIE H. HYMAN (From the Bermuda Biological Station far Research and the American Museum of Natural Plistory) The Turbellaria Acoela constitute one of the most interesting in- vertebrate groups because they are the lowest of the bilateral animals (if the Mesozoa be excepted). Their simplicity of structure is in all probability primitive and not the consequence of degradation. Among their primitive characters are the absence of a definite digestive ento- clerm, the presence of a statocyst, the netlike nervous system with ra- dially arranged principal nerve strands, the scattered sex cells, the ab- sence of female ducts, and the simple construction of the copulatory mechanism. In the Acoela the entoderm has apparently not yet differen- tiated from the general mesentodermal cells. In the most primitive genus, Ncmertoderma, a definite brain is lacking, the nervous system consists of a subepidermal plexus, and the reproductive system is limited to the sex cells, as ducts and copulatory apparatus are entirely wanting (Steinbock, 1931). The extremely complicated reproductive system characteristic of the Platyhelminthes has evidently arisen within the phylum itself, starting from conditions in the Acoela, where this sys- tem may not have advanced beyond the stage seen in ccelenterates, as in Ncmertoderma. The first step in the development of a reproductive system among the Acoela consists in the formation of more or less defi- nite paired channels in the parenchyma for the transport of the sperm. These somewhat vague vasa deferentia unite to a common terminal duct in whose wall a penis gradually evolves, first as a muscular thickening of the wall, then in most Acoela as a conical penis composed of parenchyma, muscle fibers, and in some cases gland cells also. Female ducts are ab- sent throughout the group but in the majority there is a seminal re- ceptacle, the seminal bursa, in the form of a rounded parenchymatous organ which may or may not be connected to the exterior either by its own pore or by way of the male duct and pore. This seminal bursa has no direct connection with the eggs but possesses a variable number of projecting exits, composed of a hard material arranged in rings. These exits were termed "mouthpieces" by Mark (1892) and his name has been adopted by German workers in the form Mundstuck, but they so 319 320 LIBBIE H. HYMAN much suggest the nozzle of a hose that T shall call them nozzles. In some Acoela there are numerous seminal bursae each with one nozzle. It is thus evident that most of the Acoela possess a copulatory mechanism in the form of a protrusible penis papilla and a seminal re- ceptacle but to my knowledge copulation has never been recorded for the group. The only account in the literature which I have been able to find is that of Gardiner (1898), for Polycharus candalus. Gardiner noticed that restless individuals often mounted on quiescent ones and remained in this position for a brief period. Sections of the quiescent worms showed sperm on the surface and in the parenchyma and Gardiner therefore concluded that in this species fertilization occurs by means of hypodermic impregnation. While hypodermic impregnation is probably of wide occurrence in the Acoela it is to be presumed that a true copulation also takes place since the necessary mechanism is pres- ent, as von Graff has pointed out (1908). During a stay at the Bermuda Biological Station in the summer of 1935 I had ample opportunity to witness the breeding habits of an acoel present in the aquaria there. After careful study of living intact and pressed specimens and serial section>. I have identified this animal as Amphiscolops langcrhansi (Graff) 1882. Specimens agree in all de- tails with von Graff's 1904 description. This species has flourished and bred for some time in the aquaria at the Bermuda station but its source is unknown. Dr. Wheeler, the Director, informed me that the animal had not been found in nature at Bermuda although I collected a very closely related form from the sea- weeds near the station. The species has apparently been introduced into the aquaria from some outside source. It is recorded by von Graff from Madeira, the Canary Islands, and the Mediterranean. Amphiscolops langcrhansi (Fig. 1) is a small worm, reaching 4-5 mm. in length, and of elongated form with a bilobed posterior end. Near the anterior end is a conspicuous statocyst and two small eyes. The brown color is due to numerous zoo.\antliell;e situated in the pe- ripheral part of the parenchyma. A study of their role in the life of the animal has recently been made by M. V. Welsh (1936). In ad- dition there occur just under the epidermis areas formed of an appar- ently crystalline material, called concremcnt crystals by von Graff; they are dark by transmitted, opaque white by reflected light and confer a pattern on the animal. This pattern differs considerably in different individuals but its general arrangement is indicated in Fig. 1. The mouth is central and near the posterior end can be noticed the rounded mass of the penis. Although the reproductive system of .-I. langcrhansi has been ade- 'fef !:• -;t ,i&|fc ;ifesi- >©:> iow».^ ^//••/.fev^V" " \' -,o\ t^f^/^/.. -.!,/;; ^J^)\) FIG. 1. Amphiscolops lainjcrhansi, from life, showing the pattern of concrc- ment granules. FIG. 2. General view of the reproductive system in a living pressed specimen. FIG. 3. Penis and seminal hursa in sagittal sectinn. FIG. 4. Beginning of copulation, rolled-up stage. FIG. 5. Attitude during copulation. FIG. 6. Copulating pair in motion, right-hand individual advancing and drag- ging left-hand one with its dorsal surface against the glass. 322 LIBBIE H. HYMAN quately described and figured by von Graff (1904), it is necessary to review its main features in order to understand the copulatory process. Its general appearance as seen in a live worm pressed under a cover- glass is shown in Fig. 2. The ovaries consist of strands of eggs in a paired lateral location. Behind and unconnected with the eggs lies the median rounded seminal bursa, or sperm receptacle. This hears sev- eral anteriorly directed nox/.les. The number of these is variable and increases with age so that small ones occur along with large ones in the larger specimens. Yon Graff records finding 6-1 1 nozzles and I have found 2-8, usually 4 large ones and one or more small ones. These nozzles consist of a hard yellow material, commonly spoken of as chitin- ous ; but no proof exists of the occurrence of chitin in the Platyhcl- minthes and they arc pmhably cuticular in nature, that is, composed of a scleroprotein allied to keratin. The nozzles are cross-striated and ap- pear to consist of a series of superimposed rings. They are pointed towards the ovaries and serve to discharge sperm against the eggs. The end of the nozzle buried in the tissue of the bursa connects with a rounded sac filled with sperm. The bursa opens below by the female genital pore. Immediately behind the bursa lies the large rounded mass of the penis filled with gland cells. To either side a somewhat indefi- nite channel, the vas deferens. containing masses of sperm, approaches the penis, passes around to its posterior side and there opens into the penis base. The testes consist of cell groups scattered throughout the lateral parenchyma. For an adequate understanding of bursa and penis, sections are necessary. This region of tin- animal is illustrated in Fig. 5, a median sagittal section. The seminal bursa is composed of loose reticular fibers in the meshes of which parenchyma cells occur. It has no definite cavity but there are many small spaces and .several larger spaces each occupied by a dense mass of sperm. From each such sperm mass a nozzle extends .'interiorly. Posteriorly and ventrally the bursa opens into the female genital pore bv wav of a small imagination which is not lined by a definite epithelium. The penis considerably resembles that of triclads but is of a looser, less muscular construction. It consists of a loose mass of parenchyma! and gland cells, not well differentiated from each other, through which run radiating muscle fibers. These muscle fibers join the general sub- epidermal muscle layers. As in triclads, the penis is divisible1 into an internal mass, the penis bulb, continuous with the general parenchyma. and the projecting papilla. This papilla is continuous anteriorly with the parenchyma of the seminal bursa; it is free only posteriorly, where it is separated by a deep cleft from the succeeding part of the body. COPULATION IN AMPHISCOLOPS LANGERHANSI The body wall behind this cleft forms an anterior projection which un- expectedly has been found to play an important role in copulation. I shall therefore call it the sperm guide. Unlike the rest of the body surface, which has no definite epithelium (the epidermal cells are " eingesenkt " as the Germans say) the penis lumen and cleft separating penis papilla from the sperm guide are lined by a more or less definite epithelium. The cavity in which the penis papilla lies opens ventrally by the male pore which is so close behind the female pore that the two may be considered with von Graff to form a common shallow space. In one aquarium at Bermuda which contained five or six hundred specimens of Amphiscolops, copulation was observed on numerous oc- casions. At all daylight hours from 6 A.M. to 5 or 6 P.M., one or two to a dozen pairs in copulation could be seen on the wall of the aquarium facing the light. Apparently no copulations occur after dark and usu- ally the number decreased after 4 P.M. As the number of sexually ripe individuals in the tank was estimated at less than two hundred and as fifty or more copulating pairs were counted daily, it is evident than any one animal must copulate repeatedly. To test this, five copulating pairs wrere isolated into a finger bowl. The following day, two egg masses were found in the bowl and one pair was again in copulation. Copulation occurs in the following manner. When two individuals wandering on the aquarium wall happen to come in contact they gener- ally give each other quick little touches resembling nips with the an- terior end. This behavior commonly occurs whether copulation ensues or not. Only the larger specimens copulate. Often a large specimen was seen attempting copulation with a smaller one which declined the invitation. If both individuals are ripe for copulation, they, following a short period of the nips just mentioned, suddenly roll up into a ball (Fig. 4), with one individual on the outside. The posterior end of this individual is so curved that its ventral surface is in contact with the ventral surface of the tail end of the other copulant. After a very short interval, less than thirty seconds, the two worms unroll and are seen to be in firm connection. The upper individual simply flattens its ventral surface against the glass and becomes quiescent. The other one is compelled to twist about until its ventral surface touches the glass. Both then become motionless with ventral surfaces against the glass and heads pointed more or less directly away from each other (Fig. 5). If disturbed they will crawl about, one animal with ventral surface against the substratum dragging the other, dorsal side next the substratum, after it (Fig. 6). The union is thus very firm, and the copulatory act is prolonged. In a number of undisturbed pairs which were timed con- nection lasted from forty to fifty-five minutes, usually about fifty 324 LIBBIE H. HYMAN minutes. The animals then quickly separate, each gives a comical little shake, as if settling its viscera into place, and proceeds about its business. The process of copulation ronnot be understood by a mere inspec- tion of copulating pairs. Indeed, study of these with a high-power hand-lens was puzzling rather than informative for the location of the Fi<;. 7. Section through a copulating pair, reconstructed from several sections, showing nianiHT cking f sperm, directed hy the sperm jjuidi-, tlnuini; into thu si-minal bursa of each c<>pulant. In all ti.mirr- : 1. mmitli, 2, ovary, 3, nozzles of suninal hursa, 4, seminal bursa, 5, femali Ln-nital pore, 6, vas . l.\ ^pcrm mass in bursa, 14, penis bulb, 15, pmi> jiapilla, Id, sperm guide, 17, cutaneous glands, 18, gland cells nf ]n-ni> bull'. connection inditatcd that the penis papilla \\a> not inserted into the seminal bursa. The natural assumption, of course, is that the penis papilla dischai'^i s directly into the bursa and a statement to this effect is made by Biv^lau (1933, p. 1 IS). Inspection of Fig. 3 shows, how- ever, that the seminal bursa has no definite lumen and that the female COPULATION IN AMPHISCOLOPS LANGERHANSI 325 genital pore is too small and shallow to receive the penis. Fortunately it is easily possible to fix pairs in copulation and several such have been studied in serial sections. A sagittal section of a copulating pair is drawn in Fig. 7. It is seen that not only the penis papilla but also the sperm guide is protruded and that the penes are interlocked in a dovetailed fashion such that the lumen of each penis grasps the posterior part of the penis of the partner. This leaves the anterior part and sperm guide of each copulant directed towards the bursa of the other. The sperm issue in a stream from the penis lumen and pass over the anterior edge of the partner's penis to- wards the latter's bursa. They are prevented from escaping into the sea-water by the sperm guide which directs them towards the bursa. They enter the bursa and pass into the sacs connected with the nozzles. Study of serial sections of the bursse of copulating animals has shown that the sperm easily penetrate throughout the bursa by way of spaces and so the one stream of sperm reaches all of the sperm sacs in the bursa. It is evident that some glandular secretion accompanies the sperm for the lining epithelium of the proximal part of the penis lumen contains secretion granules and a secretion staining with haematoxylin and thus probably albuminous in nature also is found in the penis lumen. The copulation is thus seen to be mutual as in the Platyhelminth.es in general, each worm fertilizing the other. The manner of copulation is, however, unique, consisting of an interlocking of the penes. The only comparable case that occurs to me is that of the slugs where in many species the penes twine about each other in copulation and deposit a sperm mass upon each other. The eggs are laid at night, probably after midnight, in flat gelatinous cakes stuck to the aquarium glass or other objects. The material in which the eggs are imbedded is probably secreted by the very numerous cutaneous glands. It is evident from anatomical considerations that the eggs cannot be laid by way of the female genital pore. They probably issue through the mouth but unfortunately the process was not ob- served. The number of eggs in the egg mass was counted in 44 cases and found to range from 3 to 16, with the majority containing 6 to 9 eggs. The eggs hatch on the fifth day into completely formed little worms which, however, do not possess any zooxanthellse. Early in the morning, about 6 A.M., the eggs laid during the night are found for the most part in 16- to 32-cell stages but a few earlier cleavages can generally be discovered. The early cleavages were watched in a number of eggs and did not seem to accord with the standard ac- count, derived from the work of Bresslau on Convoluta (1909, 1933). The first cleavage was sometimes equal, sometimes unequal. A three- 326 LIBBIE H. HYMAN cell stage was common and the four-cell stage showed no definite arrange- ment into two micromeres and two macromeres. The cleavage of the four-cell stage was not followed satisfactorily owing to the sudden shifts in position occurring at the moment of cleavage. The time re- maining after the animals began to lay eggs was too brief to permit any extended study of the cleavage, but a number of egg masses were pre- served and may be studied at some future time. For the privileges of a stay at the Bermuda Biological Station I am greatly indebted to the President, Professor E. G. Conklin, and the Director, Dr. J. F. G. Wheeler. SUMMARY In the aoK-lous turbellarian Amphiscolops lanycrhansi, copulation occurs by the interlocking of the penes. The sperm pass over the an- terior surface of the partner's penis into the seminal bursa of the latter. The body region immediately behind the penis is also protruded in copulation and acts as a sperm guide directing the sperm into the partner's bursa and preventing their dissemination into the water. REFERENCES BRESSLAU, E., 1909. Die Entwicklung dor Acoelen. Vcrhand. dcutsch. Zool. Gescll., 19:314. BRESSLAU, E., 1933. Turhellaria. Kiikenthal-Krumbach Handbuch dcr Zoologie, Vol. II. GARDINER, E. G., 1898. The growth of the ovum, formation of the polar bodies, and tlic fertilization in Polychoerus caudatus. Jour. Morph., 15: 73. GRAFF, L. v., 1904. Marine Turbcllarien Orotavas und der Kiisten Europas. I. Einleitung und Acoela. Zcitschr. wiss. Zool., 78: 190. GRAFF, L. v., 1908. Acoela. Bronn's Klassen und ( )rdnungen des Tierreichs, vol. I V, Abt. 1 c. MARK, E. L., 1892. Polychoerus caudatus nov. gen. et nov. sp. rcstschr. f. Leuckart. p. 298. ' STEINBOCK, O., 1931. Ergebnisse einer von E. Keisinger und O. Steinbock mit Hilfe des Rask-0rsted Fonds durchgefuhrten Keise in Gronland 1926. 2. Nemertoderma bathycola nov. gen. nov. spec. / 'idcnsk. Mcdd. fra Dansk. Nat. Hist. For en., 90: 47. WELSH, M. F., 1936. Oxygen production by zooxanthellae in a Bermudan turbel- larian. Biol. Bull., 70: 282. A NEW GENUS OF HYDROID AND ITS METHOD OF ASEXUAL REPRODUCTION SAMUEL STOCKTON MILES (From the Mount Desert Island Biohc/ical Laboratory) A tubularian hyclroicl (Plate I) was found in the waters of French- man's Bay off Salsbury Cove, Mt. Desert Island, which apparently is a hitherto undescribed genus. It is a solitary form dredged off mud bottom at about 40 to 60 feet along with Corymorpha pcndnla and Acaidis primarius. Many specimens were found. Its peculiar interest lies in an unusual mode of asexual reproduction. The name Dahlgrcnclla farcta, nov. gen. et nov. sp., is proposed for this form. The generic name DaJilgrcnclla is suggested to express my indebtedness to Professor Ulric Dahlgren of Princeton University for his kind assistance and advice in this work, as well as for his method of collecting mud-living hydroids.1 The specific name farcta is descrip- tive of the method of asexual reproduction described above, being the Latin word for " stuffed." The ovoid bodies resemble a string of sausages, the word for which in Latin is " farcimen '' from the same root as farcta. Doctor Willie W. Smith of New York University was particularly helpful in the preparation of the manuscript for publication. I also wish to express my thanks for the facilities given me at the Mount Desert Island Biological Laboratory. Genus Dahlgrcnclla Characteristics: Hydroid stage (1) solitary. (2) single whorl of oral tentacles. (3) whorl of basal tentacles ringed with nematocysts. (4) clusters of medusae buds just oral to basal tentacles. (5) asexual reproductive bodies found in perisarc posterior to foot. 1 Netting is slowly dragged over the bottom. A piece 2 by 4 feet is a con- venient size with a one-half inch mesh. The forward edge is fastened along a board 2 feet by 2 inches to which is lashed a 2-foot length of one-inch copper pipe. 327 SAMUEL STOCKTON MILKS --B -B PLATF. I. l>tililv 50 mm. glass slides in stack- ing finger-bowls. The frustules will anchor the hydranth to the slide overnight so that they may be taken from one finger-bowl to another for changing the water and may be placed in IVtri dishes for examining under the microscope without greatly disturbing them. P>y this means the attachment and movements of the more elongate frustules may be studied. Clusters of medusae (Plate III) and medusae buds are located above A NEW GENUS OF IIYDROII) 331 the ring of basal tentacles. The young medusa; are about one mm. long when first set free. A single short tentacle densely covered with nematocysts is present. The four tentacle bulbs around the rim of the ovoid bell contain red pigment. In addition to those of the tentacle bulbs and tentacle, nematocysts are sparsely scattered over the entire bell. The ovoid bodies (Plate IV) enveloped in the hydrorhiza are PLATE IV. Dahlgrenella farcta, development of asexual reproductive bodies (orientation not shown). Body constricting off from foot, 1; body after constric- tion (which may be followed by subsequent division), 2; body dividing, 3; body after division (which may be followed by further division), 4; developing hydroid with first appearance of tentacles and f rustules, 5 ; older hydroid with appearance of perisarcal region, 6 ; older hydroid with well differentiated tentacles, 7. The hydroid usually penetrates the enveloping hydrorhiza between the stages 5 and 6. formed by fission of the foot and in some cases redi vision of the parts constricted off from it. Rudimentary tentacles and frustules appear in SAMUEL STOCKTON MILES the most posterior bodies first and development proceeds until the young hydroid is formed. As many as twenty bodies, each destined to develop into a new individual, have been observed in a single hydrorhiza. It may be of some significance in the interpretation of this mode of reproduction that the young individuals appear to develop oriented op- positely from the parent. The hypostomata of the young in most cases point posteriorly in the hydrorhiza. Since two individuals pointed in the same direction may develop from a single Ixxly constricting equa- torially, it is evident that the orientation is not dependent upon the loca- tion of the constriction. The possibility of the presence of an axial gradient should be investigated. However, the bodies become motile early and possibly reverse their positions by their movements. This possibility seems remote since one can observe in the position and struc- ture of the sheath no conditions which might cause the bodies to point uniformly in a direction in which they had not developed. It is extremely difficult to determine the orientation of the bodies for several reasons. First, the hydrorhiza is often rendered opaque by extraneous matter which has adhered to it. Second, it is impossible to distinguish oral from aboral end until frustules or tentacle buds appear. Third, soon after the appearance of frustules and tentacles the young hydroid penetrates the hydrorhiza thus allowing mechanical influences to disturb its orientation. Therefore this can be determined only under good conditions and only for a short period in the development of the bodies. However, in a series of 17 specimens kept on glass slides over a period of 4 days, 25 bodies were observed which apparently pointed in the opposite direction from the parent while 2 seemed to point in the same direction. Of 7 other bodies whose orientation was more ques- tionable 6 seemed opposite from the parent and 1 the same. In only 1 case was there evidence that the body had reversed its position. At the end of the 4 days in the 17 specimens there were 133 bodies of which 13 appeared during the last night. The taxonomy of this form is fairly evident. The arrangement of tentacles in oral and basal whorls with medusa clusters located just above the former clearly indicates its relationship to the genera of the family Corymorphidae. It differs from Corymorpha and Hybocodon in having but one whorl of oral tentacles. This characteristic it shares with Ectopleura. The presence of a single marginal tentacle in the medusa is likewise indicative of its place with the Corymorphidce. In this it is similar to Hybocodon and Coryinorflnt but differs from Jlcto- plcitra. The arrangement of tentacles as well as the distribution of A NEW GENUS OF HYDROID 333 nematocysts upon them is very similar to Trichorhisa brunnea which is generally classified in the family Pennariidae as well as to Hetcro- siephanus annulicornis. The formation of a new genns for this hydroicl seems justified not only because of oral tentacles, size, and form of hydrorhiza but espe- cially because of the unusual method of asexual reproduction. An at- tempt to identify the medusa stage with any previously described form was also fruitless. THE RELATION BETWEEN FOOD, THE RATE OF LOCOMOTION AND REPRODUCTION IN THE MARINE AMCEBA, ELABELLULA MIKA1 DWIGHT LUCIAN HOPKINS (Department of Zoology, Duke University, and the Bermuda Biological Station for Research) INTRODUCTION It is known that the rate of locomotion in amoeba? varies with many of the factors in the environment, such as the osmotic pressure (Hop- kins, 1928), the hydrogen ion concentration (' Pantin. 1923; Hopkins, 1928; Mast and Prosser, 1''32; Pitts and Mast, 1933 and l'>34), tin- concentration of salts (Pantin, 1926a and b ; Hopkins. 1929; and Pitts and Mast, 1933 and 1934), the nature of the substratum (Hopkins, 1929), and temperature (Schwitalla, 1924 and 1925; Pantin, 1924; Mast and Prosser, 1932, and Hopkins, 1937). It has been shown, however, that within wide limits the rate of locomotion is not affected by the oxy- gen tension (Hulpieu, 1930; Pantin, 193();r per sq.mm. of bottom + sq.mm. of surface film. Concentrations of bacteria X/2 A'/4 .V/8 A'/ 10 y Culture I 25.0 9.2 9.0 10.0 8.2 Culture II 18.0 28.8 12.6 7.9 5.6 Average 21.5 19.0 10.8 8.9 6.9 bacteria, was obtained by averaging the results of five minutes locomo- tion for each of five amu-b;e. The results arc presented in Table III. From the results of Table III it is observed that the- rate of locomo- tion of these amu-bae has not been influenced by the variations in kind and concentration of the bacteria on which thev have been fed. The Rate of Locomotion of Flabcllula mini in Artificial Sea U'afer at 21.6 ± 0.2° C. and its Constancy U'hcn .III Pood Particles are Absent from the Medium By a food particle is meant here any particle which the amreba will ingest. Since, as has been shown, the rate of locomotion of PI ab ell ul a mira is within wide limits independent ot the amount and kind of food it has had available, it seemed worth while to determine accurately the rate under definite and reproduceable conditions. Also the question arose as to whether the rate of locomotion in the absence of external stimulation is variable or constant. Schwitalla (1()24) has maintained FOOD AND RATE OF LOCOMOTION IN 339 that locomotion in Amccba protons is never constant but varies in a rhythmical fashion. Hahnert (1932) and Pitts (1933), however, have shown that Schwitalla's rhythms were due to the fact that in measuring the rate he did not take into consideration the changes in form of the amoeba; while observing the rate. They maintain that in Amoeba proteus when the form is monopodal the rate of locomotion is constant. In the experiment now to be described, the object was to remove from the environment of the amoebae, as much as possible, all chances of stimulation from the environment. The main and most frequent stimulation from the exterior to which they react is to particles which they take in as food. They engulf not only bacteria but in addition yeast, carbon particles, dust, etc. In order to eliminate interference to locomotion due to such particles the following procedure was followed. TABLE III Showing the relation between the kind and concentration of bacteria and the rate of locomotion. X is concentration of bacteria in the stock suspension, and X/2 the dilution. Rate in micra/minutes in artificial sea water. Temperature 21.6 ± 0.2° C. Bacterium No. i 2 4 5 7 8 X 28.0 27.0 30.0 28.0 32.0 30.0 X/2 32.0 28.0 30.0 26.0 29.0 28.0 Average 30.0 27.5 30.0 27.0 30.5 29.0 Average for all experiments 29.0 The Pyrex glass dishes, or locomotion dishes, were washed in sulfuric acid-potassium bichromate cleaning solution, rinsed several times in distilled water, dried, polished with a clean cloth and placed on a sheet of platinum and heated to a red heat for about ten minutes. In this way the surface of the dish wras freed of all organic particles. The artificial sea water was filtered through three layers of No. 50 Whatman filter paper, after the filter paper had been washed thoroughly by run- ning through it several hundred cubic centimeters of artificial sea water. The amcebae were washed free of bacteria by removing them singly from the culture with a sterile capillary pipette and passing them through five changes of the filtered artificial sea water. They were finally placed in filtered artificial sea water in the cleansed Pyrex glass dish. The dish was then covered with a thoroughly cleansed cover-slip to keep out dust from the air. The dish containing the amcebae was placed in the observation chamber of the constant temperature apparatus 340 DWIGHT LUCIAN HOPKINS (Hopkins, 1937) at 21.6 ± 0.2° C. Cultural temperature for the amoebae used was from 18° to 22° C. Thus no adaptation to tempera- ture (Hopkins, 1937) was necessary before locomotion could become constant. One hour, however, was allowed for the amoebae to recover from the effects of handling and the small change in light intensity as- sociated with the artificial light used for the observations. The en- vironment thus established was remarkably free of external stimulating agents. The medium, however, still contained a few bacteria and small particles which could be seen under low-power of the microscope. These were very infrequent such that they could be detected before the amoeba? under observation contacted them. The performance of each of three amoebae was observed. The observation on each amoeba was continued until it contacted and engulfed a visible particle which re- sulted in a temporary cessation of locomotion. During the periods of TABLE IV Showing the performance of three anm-bie in the absence of food particles. Temperature 21.6° C. AiiHtba Distance in mm. traveled by amcrli.i- in e.irli of a number of consecutive minutes X 118 Rate micra/min. A 4, 4,3, 4,3, 4 4 *i *i 4,4, 4, 4, 3, 4, 4, 4,5, 5, 4, 4, 4, 4, 4, 4, 4, 5, 3, 4,4, 4 4 'i ^> 4,4, 4, 5, 4. 4, 4, 4,4, 4, 4, 4, 3, 4, 5, 4, 4 4, 4 4 4 4 4 4 4 33.9 B 3, 4,4, 5,4, 4,4, 4,4, 3,3, 3, 3, 3, 4,4, 4, 5. 31.9 C 4, 4,4, 4,4, 4,4, 4,4, 4,4, 4, 4, 4, 4,4, 4, 4,4,4,4,4,4. 33.9 Average 33.23 observation these amu-ha- contacted a few particles which they did not engulf, and no interruption in the locomotion occurred. Amoeba A traveled 55 minutes before it engulfed a particle and ceased temporarily to move; II IS minutes, and C 25 minutes. The minute to minute per- formance of each of these three ani. On the physiology of amoeboid movement. IV. The ac- tion of magnesium. Brit. Jour, lixpcr. Biol., 3: 297. PANTIN, C. F. A., 1930<7t On the physiology of amoeboid movement. V. Aiuerobic movement, /'roc. Roy. Soc., 105: 538. FOOD AND RATE OF LOCOMOTION IN AMCEB.E 343 PANTIN, C. F. A., 1930/>. On the physiology of amoeboid movement. VI. The ac- tion of oxygen. Proc. Roy. Soc., 105: 556. PITTS, R. F., 1933. The relation between the rate of locomotion and form in Amoeba proteus. Biol. Bull., 64: 418. PITTS, R. F., AND S. O. MAST, 1933. The relation between inorganic salt con- centration, hydrogen ion concentration and physiological processes in Amoeba proteus. I. Rate of locomotion, gel/sol ratio, and hydrogen ion concentration in balanced salt solutions. Jour. Cell, and Com par. Pli\siol., 3:449. PITTS, R. F., AND S. O. MAST, 1934. II. Rate of locomotion, gel/scl ratio, and hydrogen-ion concentration in solutions of single salts. Jour. Cell, and Compar. Physiol., 4: 237. RICE, N. E., 1935. The nutrition of Flabellula mini Schaeffer. Arch. f. Protist., 85: 350. SCHWITALLA, ALPHONSE M., 1924. The influence of temperature on the rate of locomotion in Amoeba. I. Locomotion at diverse constant temperatures. Jour. Morph., 39: 465. SCHWITALLA, ALPHONSE M., 1925. The influence of temperature on the rate of locomotion in Amoeba. II. The rate of locomotion in amoeba at different temperatures. Jour. Morph., 41 : 45. THE CHROM. \TOPHOROTROPIC HORMONE OF THE CRUSTACEA: STANDARDIZATION, PROPERTIES AND PHYSIOLOGY OF THE EYE-STALK GLANDS 1 A. A. ABRAMOWITZ (From the Marine Biological Laboratory, Woods Hole, Mass., and the Biological Laboratories, Harvard University} STANDARDIZATION The need for a method of determining quantitatively the amount of hormone in a given extract of a gland is obvious in chemical or quantita- tive physiological work dealing with endocrine systems. Although the existence of such a system for the pigmentary effectors of crustaceans has been known for almost ten years, no reliable method for the stand- ardization of the eye-stalk hormone has been devised. Navcz and Kropp (1934), and Kropp and Crozier (1934) suggested that the eye- stalk hormone may be standardized by the Avcna coleoptile method since they found that the eye-stalk extract accelerates the growth rate of Avcna coleoptiles. These authors employed water extracts of the eye-stalks of Palcevnonetes. Since many other substance's in addition to the chromatophore hormone are present in a water extract of the eye- stalks, it is not certain that the growth effects on plants arc due actually to the chromatophore hormone. On the basis of biochemical evidence, Carlson (1936) suggested that the growth reactions of plants to the eye-stalk extract was due to the presence of some substance other than the chromatophore hormone. If this is true, it would then appear that Kropp and Cro/ier have not been measuring the chromatophore hormone but ]>o,~ibly some other substance present in a water extract of the eye- stalks. Consequently, it was felt that the most reliable way of assaying the chromatophore hormone would be to measure the response of the tissue (the chromatophorcs) which this hormone normally affects.2 A con- venient laboratory animal on which a method of assay may be devised is the fiddler crab, (lea. In the vicinity of Woods Hole, two species, U. pugilator and U. pnyna.v, are abundantly present. Both these spe- 1 Aided in part by a grant from tlic Rockefeller Foundation, administered by l'~. L. Hisaxv. 2Cf. liurn, J. II., 1930. I'hys. AVr., 10: 146. 344 CHROMATOPHOROTROPIC HORMONE OF CRUSTACEA 345 cies contain black, red, white and yellow integumentary chromatophores. Only occasionally, however, are red chromatophores observed in U. piigna.v. The extreme color change of U. pugiiax ranges from jet black to pale yellow (Figs. 3, 4) and of U. f>uyilator from dark brown to cream white (Figs. 5, 6). The melanophores are the chief agents of this change although the movements of the other pigments are instru- mental in producing the end result. Megasur (1912) demonstrated that extirpation of the eye-stalks of Uca resulted in the production of the pale phase of its color change due to melanophore contraction. Carlson (1935, 1936) confirmed this observation and interpreted it as an effect specifically due to the loss of the endocrine glands contained in the eye-stalks. An extract of the extirpated eye-stalks when injected into a pale (blinded) animal produced the dark phase resulting from melanophore expansion. Carlson showed further that if 0.1 cc. of a solution containing the extract of 1 eye-stalk of Uca in 1 cc. of water was injected into a blinded 3 crab, the melanophores became expanded within 2 hours, remained expanded for about 4 hours, and finally con- tracted so that the animal resumed the pallor previous to injection. This is specifically a hormonal reaction. There appears to be no other way whereby the melanophores of a blinded Uca may be induced to expand. Preliminary Experiments The statements of Megasur and Carlson concerning the effects of blinding and of injection of eye-stalk extracts were first confirmed (Figs. 3, 4, 5, 6). The typical response of a blinded animal following the injection of an extract of its eye-stalks in a dose equivalent to 1/20 of a single stalk may be conveniently divided into four phases : (a) The first perceptible response. This response is evidenced by the beginning of melanophore expansion and occurs from 15-20 minutes after the injection has been made. (b) The attainment of the maximal effect. After the first perceptible response has occurred, the melano- phores become maximally expanded within one hour. (c) Duration of melanophore expansion. Following the attainment of the maximal effect, the melano- phores remain fully expanded for about 31/4 hours. (liore expansion varies as a function of concen- tration. The times for the completion of the other three phases of the response are for all practical purposes constants, independent of con- centration provided the doses employed are greater than 0.003 E.S. In order to measure the response of the animals to a given concentration, it is therefore necessary to determine only the time of injection and the time at which the animals become again pale. The difference in these two readings is obviously the duration of the response plus 1.75 hours, 1 The letter M.S. \\ill lie used throughout this paper as an abbreviation for the extract equivalent to a given number of eye-stalks. For example. 1 K.S. indicates that the extract is equivalent to that obtainable from one eye-stalk; 0.5 E.S. de- notes an extract equivalent to half th»t obtained from one eye-stalk, etc. CHROMATOPHOROTROPIC HORMONE OF CRUSTACEA 347 which may be used as a constant equal to the sum of the times of the other three phases of the response. Variables in the Method of Assay There are several important variables which must be controlled com- pletely if such a method of assay is to be used. (1) Physiological Uniformity of the Responsive Tissue. — The black chromatophores were used as the sole criterion in obtaining readings since they are the chief instruments in determining the appearance of the animal with respect to paleness or darkness. In blinded specimens of Uca the pigment in all melanophores of the appendages and body is uniformly and extremely concentrated, and remains so indefinitely re- gardless of environmental conditions which would provoke expansion in a normal animal. Thus it appears certain that the responsive tissue in all test animals is in a physiologically uniform state. TABLE I Size of eye-stalk in relation to hormone concentration Average Size of Eye-stalk Injected Dose No. Tested Weight of Test Average Response Animals grams hours Small (animal weight, 1.7 grams) 0.025 E.S./cc. 8 3.5 3.75 Medium (animal weight, 2.8 grams) 0.025 E.S./cc. 8 3.6 4.09 Large (animal weight, 5.1 grams) 0.025 E.S./cc. 8 3.5 4.88 (2) Concentration of the Injected Hormone. — This variable is easily controlled since the method of preparing the extract unless other- wise stated was maintained constant. The eye-stalks were cut off and ground in a small mortar in the amount of sea water necessary for any particular concentration. This extract was brought to a boil so that a coagulum (presumably the tissue proteins) formed. The solution was filtered, and the filtrate made up by the addition of unboiled sea water to the desired concentration. About 0.6 cc. of the sea water was usu- ally lost during the process of boiling and filtration. This resulted in a slightly hypertonic extract, but this factor is immaterial for the purpose of the experiment (cf. Carlson, 1936). The solution was allowed to cool to room temperature, usually about one-half hour elapsing, and the injection made. While the concentration of the hormone in terms of numbers of eye-stalks can be controlled quite accurately, the size of the eye-stalks used in making an extract of a certain concentration could not be controlled so easily. Larger eye-stalks may contain large glands and hence more hormone, as shown by Table I. To control this factor, A. A. ABRAMOWITZ the extracts were prepared from eye-stalks of animals of a definite and uniform size. (3) Sice of the Test Animal. — This variable proved to be a very dis- concerting factor at the beginning of the work. It was reasonable to suppose that large animals would show a smaller response to a given dosage than small animals because they would contain a greater amount of the responsive tisane, and because the hormone would be subjected to greater dilution. That this is plausible is shown by Table II. Consequently, in order to avoid the use of a curve for the calibration of response as a function of the size of the test animals, assay experi- ments were always carried out on animals of uniform weight. (4) Individual I'ariation. — This variable is probably the most sig- nificant of tin isc aln-adv mentioned because it is not easily controlled. Since it is almost impo»ible to carry out all experiments on the same animal, the best method would be to use sufficient numbers of animals TABLE II Size of test animal in relation to response Hormone Concentration No. <>f Animals Tested Average Weight Average Response grams hours 1.0 E.S./cc... 10 4.36 4.13 l.OE.S./cc... 6 3. -86 5.56 1.0 E.S./cc 10 1.55 7.83 and to treat the data statistically. In performing an assay, from 15 to 20 animals of the same size and weight were injected with the same volume and concentration of hormone and the average response taken. The standard deviation of the arithmetic mean was calculated and de- terminations made to see if the differences in the response at different concentrations were really significant. (5) Volume of Injected Dosage: Place of Injection. — Several ex- periments made to determine the most suitable volume of hormone to be injected led me to adopt finally 0.05 cc. as the standard injection vol- ume. Carlson used 0.1 cc. but while this is satisfactory for U. pugnu.r and very large specimens of U. pugilator, it seemed to be too great a dose for the size of crabs employed in most of my experiments. Tin- most convenient region for the injection of the hormone into the body spaces was found after many trials on different regions of the body to be through the soft tissue forming the joint between the coxipodite and the protopidite of tin- walking legs. This method allows for speed and CHROMATOPHOROTROPIC HORMONE OF CRUSTACEA 349 accuracy of injection for, using a 27-gaugc hypodermic needle, one can inject 20 crabs within 2 minutes. Method Finally Adopted Upon the basis of these preliminary experiments, the standardization of the eye-stalk hormone was completed using blinded Uca pugilator 5 as a test animal. One hundred and fifty such specimens, whose eye-stalks were extirpated two days previously, were arranged in ten groups of 15 animals each. All animals were of uniform size and weight (2.46 ± .01 gram). A stock solution of hormone was made by extracting the ex- tirpated eye-stalks as previously described. A series of 10 concentra- tions ranging from 5 E.S./cc. to 0.03 E.S./cc. was prepared. Five- hundredths of a cc. of these concentrations was used as the injection volume in all cases. Each of the 15 animals in a particular group re- ceived 0.05 cc. of a particular concentration, so that each of the ten groups of animals was injected with but one of the ten various known concentrations. The difference between the time of injection and time at which the animals became pale again was noted. The results of this experiment are shown by Fig. 1. The experiment was repeated five times, and although the points obtained did not fit the curve as well as those of Fig. 1 do, it was generally true that the magnitude of the re- sponse was exponentially proportional to concentration. It can be seen from Fig. 1 that the relation between the complete response of the melanophores and the concentration of hormone is least sensitive over the range of concentrations above 1.0 E.S./cc. and most sensitive over the range from 0.06 E.S./cc. to 1.0 E.S./cc. This sensi- tive range is re-plotted in Fig. 2, which shows that there is a linear relationship between the duration of melanophore expansion and con- centration of hormone. Figure 2 was therefore used to establish the Uca unit. Two possi- bilities are open. One can choose an arbitrary response near the mid- point of this curve and designate this value as one Uca unit. Assays would therefore be made by repeated dilutions of an unknown until the resulting response equals that chosen to represent one Uca unit. A second possible method would be to inject 0.05 cc. of an unknown con- centration, obtain the average response and read off from Fig. 1 the concentration in terms of Uca E.S./cc. which gives the same response. The latter procedure \vas adopted because both amount to the same thing provided the unknown concentration falls within the sensitive 5 Uca pugilator was substituted for U. pugnax as the standard test animal be- cause it is much more uniform in weight and size and because the rate of mor- tality and the frequency of autotomy are much lower than in U. pugnax. 350 A. A. ABRAMOWITZ range." The response (4.9 hours) produced by 0.05 cc. of an extract of 1 E.S. of r. pugilator (2.46 grams) per cc. of solution was designated as one Uca unit. HORMONE CONTEXT IN THE EYE-STALKS OF VARIOUS CRUSTACEANS In estimating the hormone content in the eye-stalks of various crustaceans, the following method was used: a number of eye-stalks 5 3. The sensitivity of the curve would be increased Vff\ + a-.. by obtaining more points along this range, and by using extremely large numbers of animals to cut down the size of the standard error. CHROMATOPIIOROTROPIC HORMONE OF CRUSTACEA 351 to 1 eye-stalk.7 Five-hundredths of a cc. of the solution was injected into each of 15 animals and the average of the responses taken. The concentration in terms of U. pugilator E.S./cc. which produces the same average response was read off from Fig. 1. The hormone con- tent of one eye-stalk of various crustaceans as compared with one eye- stalk of U. pugilator (1 unit) is given in Table III. The extracts were always made during the day and from animals which had been main- tained under illumination (daylight). 3 - U. ffi « I 2 0.062 0.125 0.25 0.5 Concentration Eye-stalks/cc. 1.0 FIG. 2. Curve showing relationship between the duration of melanophore ex- pansion and concentration of hormone (values replotted from Fig. 1). Theoreti- cally, the curve should go through the 0 point. EXTRACTION AND PURIFICATION Although ten years have elapsed since the eye-stalk hormone was discovered, very few chemical properties of the hormone are known. Perkins showed that the hormone was soluble in water, and that it was resistant to boiling. Carlson (1936) and Abramowitz (1936a, c) dem- onstrated that it was soluble in alcohol but insoluble in ether. Carlson also added that the hormone was not destroyed when treated for short periods of time with acid or alkali. 7 Unfortunately this proved to be a mistake because a few readings obtained in assaying the potency of the eye-stalk extracts of certain crustaceans were not within the sensitive range, and therefore cannot be considered as accurate readings. 352 A. A. ABRAMOWITZ Since a method for standardizing the hormone has been devised, an attempt was made to purify the substance. One thousand eye-stalks of Uca pugilator were dried and pulverized. The total activity of the dry powder was 1000 L'ca units, or 0.6 Uca units per ing. of dry powder. This material was extracted several times with small volumes of light petroleum ether in order to remove a red carotenoid pigment which was found to contaminate the subsequent fractions. The ether solution was washed with small amounts of distilled water. The water layer was then added to tin- residue insoluble in ether and the ether layer being only slightly active was discarded. The residue which was insoluble in ether was then extracted three times with distilled water. The water solution was boiled and filtered, and the filtrate dried in a current of TABLE III Hormone content of the eye-stalks of various crustaceans Averapc Weight of Animals Species Uca pugilalor Units grams 2.46 1 K.S. / \'ii pugilator 1.0 3.3 11.0 309.0 1 E.S. 1 E.S. 1 E.S. Uca pugnux Pagnrus pollicaris Illiniums americunus 1.0 1.25 1.20 2.0 1 Head IIif)f>a talpoida 0.25 0.38 219.0 1 E.S. 1 E.S. Crago borealis Carcinus niwnas 0.25 1.25 80.0 1 K.S. Cancer irroratus 5.0 0.4 1 E.S. Palcemoneles vulgar is 0.36 15.0 1 E.S. Uca minax 1.5 127.0 1 E.S. Li hi n id dnbia 4.0 warm air. The dried material was washed with small amounts of chloroform to remove traces of the red pigment, and the chloroform washings being only slightly active wen- discarded. The dried filtrate was then extracted with 95 per cent ethaiml, and the alcohol solution was centrifuged. The alcohol-soluble fraction was decanted and dried. Both the alcohol-soluble and alcohol-insoluble- fractions were active, tin- latter biin- much less so than the former, and hence was discarded. The material soluble in (>5 per cent ethanol was then dissolved in hot absolute alcohol and precipitated by the addition of ether. The activity of the material soluble in absolute alcohol was approximately two Uca units per mg. Further attempts to precipitate the material proved un- satisfactory because of the exceedingly small yield. The loss in total CHROMATOPHOROTROPIC HORMONE OF CRUSTACEA 353 activity was about 60 per cent. The material which is soluble in absolute alcohol is devoid of pigment and is apparently protein-free.8 PROPERTIES OF THE HORMONE The eye-stalk hormone is readily soluble in water, but not completely soluble in ethanol or methanol. It is only slightly soluble in acetone, and insoluble in organic solvents such as benzin, chloroform, or ether. The hormone is thermostable, but is destroyed by oxidation. It does not decompose when boiled with HC1 or NaOH in a 1 per cent solution for short periods of time. If the hormone is boiled for 2 hours wyith NaOH, the activity is completely destroyed. The hormone adsorbs easily to various substances present in crude extracts of the eye-stalks. If the eye-stalks are extracted with benzin, and the benzin-soluble ma- terial (chiefly pigment) and the benzin-insoluble material tested in equivalent doses, only the benzin-insoluble is found to be active. How- ever, if the benzin-soluble material is concentrated and then tested, some activity will be found. The same phenomenon was observed when working up sea-water extracts of the eye-stalks of various crustaceans. These extracts were dried, and the dried material extracted with 95 per cent ethanol. Two volumes of acetone were then added to the alcohol solution to precipitate the salts, which settle out in crystalline form. The salt crystals were also found to be slightly active. The observations indicate that the hormone adsorbs very readily. The hormone may be kept in a water solution in the refrigerator for some time without ap- preciable loss of activity. However, it is destroyed slowly when kept in a water solution at room temperature. PHYSIOLOGY OF THE EYE-STALK GLANDS One of the basic problems in endocrine research is to determine the factors which affect the production of a hormone and the mechanism which regulates its release into the circulation. The standardization of the eye-stalk hormone has made it possible to investigate the physiology of the eye-stalk glands, for the amount of hormone in the stalks of ani- mals maintained under special conditions can now be determined. The presence or absence of the hormone in the circulation can be determined quite easily by observing the states of the chief chromatophores with the aid of a microscope. This can be illustrated by the pigmentary reactions of Uca and Palamonetes, the two animals chosen for this investigation. Perkins (1928) and Brown (1933) have shown that the contraction of 8 The alcohol-soluble material gave negative results when tested with Millon's reagent and with the Xanthaproteic test. 354 A. A. ABRAMOWITZ the red and yellow chromatophores of Pahcmonctcs is due to the pres- ence of the hormone in the blood stream, and that the expansion of these chromatophores results from the absence of the hormone-. The situation is just reversed in I'ca for in the brachyurans the eye-stalk hormone expands the melanophores and the erythrophores when it is circulating through the animal, while its absence results in the contrac- tion of these chromatophores. Palamonetes Fortv .specimens of Palceiuotictcs vnhjaris of uniform size and weight were separated into four equal groups. One group was placed in a black ve»el. another in a white, a third group in a yellow and a fourth in a blue. The vessels were then placed under tin- illumination of two 75-watt electric bulbs. The animals were supplied with a con- TABLE IV Hormone content in the eye-stalks of Palaunonetes under various environmental conditions Xunihcr nl" Eye-stalks Tested Condition Average Weight of Animals Uca Units/E.S. of Palaemonetes grams 20 I'lack-adapted 0.25 0.46 20 White-adapted 0.29 0.50 20 Yellow-adapted 0.30 0.4Q 20 Blue-adapted 0.25 0.47 16 Darkness (day) 0.28 0.25 tinuous current of fresh sea water, and left undisturbed for S hours. A fifth group of 8 animals were placed in total darkness for a day. At the end of S hours, the eye-stalks of each group under illumination were extracted with sea water and each of the 4 extracts assayed to deter- mine the amount of hormone present. The eye-stalks of the animals placed in total darkness were extracted in darkness on the following day. The results of this investigation are listed in Table IV. This experiment was repeated three times and the same result was obtained in each case. The amount of hormone present in the eye-stalks of animals kept under illumination \vas twice that obtained t rom ani- mals maintained in darkness. This is confirmatory of Kropp and Crozier's finding that stalk extracts of animals kept in darkness did not depress the growth rate of Lupinns as much as extracts made from ani- mals exposed to light. It is also in agreement with the results of Klein- CHROMATOPHOROTROPIC HORMONE OF CRUSTACEA 355 holz who found that extracts from animals kept in the darkroom pro- duced weaker responses in the retinal pigments than extracts of animals under illumination. Equally significant is the finding that the amount of hormone in the eye-stalks of animals showing a continuous release of the hormone (white-adapted group) was the same as that of animals showing no, or a subminimal release of the hormone into the circulation (black-adapted group). This situation becomes understandable when the functional cycle of an endocrine gland is considered. The normal physiology of an endo- crine gland consists chiefly in synthesizing and storing a hormone, and finally releasing it into the circulation. These three processes may con- ceivably be controlled separately, or may be controlled uniformly or in part by the same mechanism. In Palccmonctcs, as in some other crus- taceans, the endocrine glands (Blutdriise of Hanstrom, 1934) of the eye-stalk are innervated from the cerebral ganglia. The functional in- nervation of a gland would therefore afford a simple mechanism for the release of the hormone into the circulation. The results listed in Table IV can be readily explained upon the basis of relative rates of hormone synthesis and hormone release. In darkness, there is no release of the hormone into the circulation as indi- cated by the inactive position of the distal pigment cells of the retina, and by the expansion of the red integumentary chromatophores (Brown, 1935£).° The rate of synthesis of the hormone must also be decidedly reduced as evidenced by the assays of the eye-stalks. In the presence of light, however, hormone synthesis is greatly increased, and this effect is produced regardless of the background over which the animals are kept. Hormone synthesis, therefore, is due to incident light and is not pri- marily dependent on reflected light. This accelerating effect of incident light may therefore be termed the primary effect. The release of the hormone, howrever, is definitely brought about by light reflected from backgrounds such as yellow or white. This is shown by the contraction of the red and yellow chromatophores of the integument and by the inward migration of the distal cells of the retina under such conditions. The release of the hormone into the circulation is maximal. However, in the black-adapted animals there is no release or only a sub-minimal release of the hormone into the circulation as 9 The condition of the red pigment in Palamonetcs maintained in darkness does not seem to be definitely settled. Perkins (1928) stated that the red pigment was concentrated in complete darkness. Brown repeats this statement (1933) but later (19356) is of the opinion that the red pigment is expanded in shrimps kept in darkness. 356 A. A. ABRAMOWITZ indicated by the expanded state of the red chromatophores.10 Yet under these two conditions, the amount of hormone in the eye-stalks is the same. This situation is understandable if we consider that the white back- ground reflex produces not only a maximal release of the- hormone (and acceleration of synthesis due to the primary effect) but also a con- comitant increase in the rate of synthesis so that there is a balance be- tween a high rate of synthesis and a high rate of release. This balance must be attained since Pahcmonetcs retains a transparent hue for months TABLE V Hormone content during diurnal rhythm Condition Uca pugilalor Uca pugnax Phase Number of Eye-stalks Extracted Uca Units/E.S. Phase Number of Eye-stalks Extracted / ,,; Units/E.S. Retina extirpated 1 week previously Dav, light Black Black Pale Inter.* 10 8 26 10 0.98 0.94 1.30 0.96 Black Black Inter. Pale 10 10 18 8 0.90 1.61 1.00 0.91 Dav, dark \ieht. litrht Night, dark . . . Normal animals Dav. light Black Black Pale Inter. 30 10 10 10 1.0 2.0 1.0 0.90 Black Black Inter. Pale 10 10 18 10 0.95 0.97 0.93 0.99 Dav, dark Night, light Night, flark 1 eye-stalk removed 3 days previously Dav, liirht Black Pale 16 16 0.95 0.95 Nitrht, lieht "The abbreviation "inter." is used to designate the intermediate condition in coloration between black and pale. if kept continually over an illuminated white background. If the white background reflex effected only a maximal release, one would reasonably expect that over a long period of white background-adaptation, the level 10 The hormone is probably released in amounts and at a rate which consti- tutes a sub-minimal stimulus for the integumentary chromatophores. Some release must occur because Kleinholz (1935) writes that the retinal pigments of Palcc- monetes are under control of the eye-stalk hormone, and in animals kept on a black background the retinal pigments are in their active state. Hence, if the same hor- mone affects both the retinal cells and the integumentary chromatophores, it must follow that the amount of hormone released in illuminated black-adapted animals is minimal for the eye-pigments but sub-minimal for the body pigments. CHROMATOPHOROTROPIC HORMONE OF CRUSTACEA 357 of synthesis would be outstripped by the rate of release. The glands would therefore become exhausted and the animals would become dark due to chromatophore expansion. This apparently never occurs. In the black-adapted specimens, the release of the hormone is proceeding slowly and consequently, a balance is also obtained between a lower rate of synthesis and this lower rate of release. The fact that the amount of hormone in the eye-stalks is the same in both white-adapted and black-adapted specimens is due simply to the primary effect of incident light. This interpretation of the physiology of the eye-stalk glands may be summarized as follows : Incident light (as opposed to light reflected from backgrounds) induces an acceleration of hormone synthesis but exerts only a sub-minimal release of the hormone into the circulation. This effect of light may be called the primary effect, and occurs in specimens placed on any background provided overhead illumination is present. The white background response, however, is due to the combination of the primary effect of incident light and of the effect of reflected light, which is to produce a maximal release of the hormone and a concomitant increase in the rate of synthesis. Uca Uca differs from Palccuwnctes in that it undergoes a periodic change in color and that background adaptation is lacking. Both Uca pugilator and pugna.v show a periodicity of color change. The diurnal rhythm of Uca pugna.v was described by Megasur (1912). The animals are black by day, pale by night, and this daily cycle repeats itself regardless of background or of light intensity. I have confirmed these statements and have extended them to Uca pugilator. When the eye-stalks are re- moved the rhythm is permanently abolished (at least until regeneration of the stalks takes place) and the animals remain pale regardless of background or light intensity. Periodicity is therefore controlled by a rhythmical release of the hormone into the circulation, the release oc- curring every 12 hours. Several animals (Uca pugna.v) were maintained upon an illuminated white background from Oct. 20-Nov. 14, 1935. During the day they were jet black in color, but at 5 :30 P.M. they began to pale so that at 7:00 P.M. all the animals were pale (lemon yellow in color). At about 8 A.M. on the following morning the animals turned black again. The same was true of specimens maintained from Oct. 20-Nov. 14, 1935, in total darkness. Background adaptation in Uca seems therefore to be lacking. Periodicity is the chief factor in its color change. .OX A. A. ABRAMOWITZ Coni])lete extirpation of both stalks leads to permanent destruction of the rhythm. Extirpation of but one eye-stalk does not impair the periodicity. One eye-stalk is therefore sufficient for the continuance of normal chromatic activity. It" the retinal portion of both eye-stalks is cut off cleanly by a sharp scalpel, the rhythm is likewise interrupted. PLATE I All photographs are from life and one -hall' life size. FIG. 3. Two specimens of Uca piHjna.v photographed on a black background. Above, a specimen whose eye-stalks had been extirpated a day previously; and below, a normal animal in daylight. FIG. 4. Same as Fig. 4, but photographed upon a white background. Fir;. 5. Two specimens of Uca f>u(/ihitor. Above, a blinded animal 2 hours after injection of an extract of its extirpated eye-stalks. Below, an uninjected blinded specimen. Frc. 6. Same as Fig. 6, but photographed upon a white background. Such animal^ when maintained in the- light on a white background remain black during both day and night, for 3 or 4 days. Similarly prepared animals maintained in total darkness remained black during both day and night for 3 days. After this period of lime the rhythm appeared again but was quite erratic for a day or two, after which it appeared to CHROMATOPHOROTROPIC HORMONE OF CRUSTACEA 359 be normal again. During the past summer, I repeated some of these experiments with large numbers of animals (Uca putjilator). In con- stant light the periodic change in color occurred at 10-11 P.M. and at 9 A.M. Moreover, only about 70 per cent of the animals showed peri- odic changes. The hormonal content of the stalks of selected animals under various phases of their diurnal rhythm was determined, and the results given in Table V. Animals of equal weights were used. The average weight of Uca pngilator was 2.5 grams, of U. f>nt/na.r 3.5 grams. The results listed in Table V show that the amount of hormone pres- ent in the eye-stalks is the same whether the animals are in the pale phase of their rhythm or whether they are in the dark phase. There are some variations in the quantity of hormone extracted under different states, but these are statistically insignificant. As in Pahcuionctcs, the quantity of hormone extracted from the eye-stalks is the same in condi- tions during which the hormone is being continually secreted or is absent from the blood stream. This situation may be similarly inter- preted. The diurnal rhythm of Uca is therefore an external expression of a diurnal release of the hormone into the circulation. During the 12 hours of release, the rate of production of the hormone may be in- creased to keep pace with its secretion into the blood. The mechanism which controls this diurnal release differs from the release mechanism of Palcrinonetes. The latter is definitely brought about by environmental factors (white background reflex), while in Uca, the diurnal release proceeds regardless of environmental factors. Fur- thermore, the theoretical increase in rate of synthesis is dependent on light in Palcrinonetes, and is independent of light in Uca, since the di- urnal rhythm proceeds in total darkness. The common feature shared by both these forms is that the release is probably controlled nervously. This is not as certain in Uca as it is in Palcrinonetes, but the fact that the Blutdriise in Uca are innervated and the observation that their nor- mal physiology may be temporarily disturbed by destruction of the reti- nal portion of the eye-stalks seem to point towards the presence of a functional innervation. At any rate, it is difficult to imagine that the diurnal rhythm is an intrinsic property of the gland cells, for an inherent diurnal synchronism of all the cells in a gland would be somewhat sur- prising. Furthermore, if this conception is true, one would reasonably expect to find variations in the content of hormone during various phases of the rhythm since the glands would secrete during the day and syn- thesize during the night. The data in Table V show that this does not occur. It would be more satisfying to regard the blood glands as being regulated by a diurnal flow of nerve impulses. In either case, the ori- 360 A. A. ABRAMOWITZ gin of the diurnal rhythm is unknown, for one is now left without an explanation of the cause of a diurnal flow of nerve impulses. The inter- mediary step in the cycle is dear, however, for this is merely a diurnal release of the hormone into the circulation. The change in body colora- tion is but the external expression of this diurnal release. A final experiment was performed to determine if a compensation of either hormone synthesis or release is made by the intact eye-stalk when its fellow is removed. One eye-stalk was removed from each of 32 specimens of I Y, p. 320, but see also ]). 328) is correct in noting that the more usual condition of the red and yellow pigments of animals maintained in darkness was slight dis- persion, and that a long sojourn (2-3 weeks) in darkness resulted in the same condition of these pigments as that occurring when animals were adapted to a red background (cf. p. 319, Brown, 1935/?), there is little reason for postulating separate autocoids for the body and retinal pig- ments. Threshold differences would account for the various responses. If the earlier observation of Perkins (1928) that the red and yellowr pigments are contracted in darkness is correct, the existence of separate hormones would be clearly indicated (because the chief body pigments would react completely independently of the retinal pigments in dark- ness and on an illuminated black background), and the entire problem would be greatly simplified. Secretion and synthesis of the retinal hor- mone would take place only under the action of light ; in darkness, re- lease would be abolished and synthesis reduced. For the chief body pigments, it would have to be assumed that the gland produces and re- leases the hormone continuously, that synthesis is accelerated by incident light regardless of background, and that release is inhibited by a black background. The latter process seems theoretically difficult, for a white background reflects while a black background absorbs light. Finally, the stimulus for release or for inhibition of release may depend on the excitation of certain portions of the retina by a particular ratio of inci- dent to reflected light. All these possibilities are well worth investiga- tion. SUMMARY A method for the standardization of the crustacean eye-stalk hor- mone on the blinded fiddler crab, Uca, has been described. The Uca unit has been defined as the amount of hormone contained in 1 cc. of solution, 0.05 cc. of which when injected into each of 15 specimens of Uca pugilator blinded 2 days previously produces a response whose average duration is about 5.0 hours. The response is measured as the amount of time intervening between the injection of the hormone and the time at which the animals again become pale, an interval during which the melanophores expand, remain expanded for some time, and finally contract. The hormone content in the eye-stalks of various crustaceans was determined. A method for the extraction and puri- fication of the hormone has been described, and some chemical and physical properties of the hormone have been listed. 364 A. A. ABRAMOWITZ The amount of hormone extracted from the eye-stalks of Palcr- inonetes is the same regardless of whether the hormone is secreted con- tinuously into the circulation, or whether it is continuously absent, conditions which arc brought about by illuminated white and black surroundings reaper lively. In darkness, there is no release of the hormone into the blood, and a very low content of hormone in the eye-stalks, approximately half that obtained from the stalks of illu- minated animals. It is postulated that light, regardless of back- ground, causes an acceleration in hormone synthesis, and that light depending on certain backgrounds such as white, causes a maximal re- lease of the hormone into the circulation with a concomitant increase in rate of production of the hormone. The diurnal color rhythm of Uca is an external expression of a diurnal release of the hormone into the circulation. Both release and synthesis arc' independent of environ- mental condition.^, and it is suggested that they are controlled by a di- urnal discharge of nerve impulses from the C.X.S. This discharge, during the day, would exert a 12-hour release of the hormone with a concomitant increased rate in production, the absence of the discharge during night would cut off release- and slow down rate of synthesis. LITER. VTl'UE CITED ABRAMOWIT/, A. A., 1935. Color changes in cancroid crabs of P.ennuda. Proc. Nut. .lead. Sci., 21: 677. AI:K \MO\\ n/. A. A., 193'n;. Action of crustacean e\e-stalk extract on melano- phores of hypophysectomized fishes, amphibians, and reptiles. 1'roc. Soc. Exper. Biol. />. The action of mUrmedin on crustacean rnelanophores and of the crustacean hormone on elasmobranch melanophores. F> Nat. Acad. Sci.. 22: 521. ABRAMOWITZ, A. A., 1936c. The chromatophorotropic hormones. Anal. Rcc., 67: Supp. No. I, 108. BROWN, F. A., JR., 1933. The controlling mechanism of chromatophores in Pala-monetes. Proc. Nat. Acad. Sci. 19: 3J7. BROWN, F. A., JR., 1935a. Control of pigment migration within the chromato- phores of I'a1;emoiietes vulgaris. Jour. 7:.r/v;-. 7.«ol., 71: 1. |',KO\V\, !•'. \.. |u., 1935/'. Color changes in I'ahemonetes. Jour. Morph., 57: 317. i !ARLSON, SVI.N I'M.. 1''35. The color changes in Uca pugilator. 1'roc. Nat. Acad. ,21: :4'>. CAKI.MIN, SVEN PH., 1936. Color changes in Hraclnura crustaceans, especially in 'Uca pugilator. K'ini. I'lii-i'., 3: 265. KLEINHOLZ, I.. II.. 1935. Crustacean eye-stalk hormone and retinal pigment mi- gration. li',,,1. Hull.. 70: 159. KROPP, B., ANIJ W. J. CROZIF.R, 1934. The production of the crustacean chromato- phore activator. J'roc. Nat. .-lead. Sci., 20: 453. CHROMATOPHOROTROPIC HORMONE OF CRUSTACEA 365 MEGASUR, F., 1912. Experimcnte iihcr den Farbwechsel dcr Crustaceen. Arch. Entu'ick.-wcch., 33: 462. NAVEZ, A. E., AND 15. KKOPP, 1934. The growth-promoting action of crustacean eye-stalk extract. />/<>/. />'»//., 67: 250. PERKINS, E. B., 1928. Color changes in crustaceans, especially in Palaemonetes. Jour. E.vpcr. ZooL, 50: 71. YOUNG, J. Z., 1935. The photoreceptors of lampreys. II. The functions of the pineal complex. Jour. E.vpcr. Biol., 12: 254. CALCIUM RKIMVTIOX AX I) TIM-: PROLONGATION OF LI KM IX 'I' III-: I«: in horizontal rows percentage disintegrated, lower figures number of eggs counted. Age in Hours In Sea \Vater In 2/3 Ca In 1/3 Ca In 1/6 Ca In 1/12 Ca 3.7 2.4 0 1.4 0 16} 108 125 105 146 132 14.3 9.4 2.5 0.7 0.4 22 112 139 121 139 241 48 48 4.4 4.6 0 39$ 115 159 113 152 84 85 4.9 0.6 0 48 134 112 143 177 100 100 68 10 0.95 66 1 114 121 210 100 74 17 90 i 133 147 100 82-100 115 218 (Sclnvlitrr. l'M6), it was found that exposure for several hours to Ca-free media produced no noticeable deterioration of Arbacia eggs and they were- able to cleave when fertilized in sea water. On the other band, 25 minutes in isosmotic GaCL resulted in the disintegration of 40 per cent ( _M out of 58 counted). After one hour and nine minutes none survived. Cleavage was reduced to zero after 15 minutes and this effect could not be reversed by an equal period of washing. The lethal influence of pure CaCL, not shown with sodium and potassium chlorides within the time indicated above, led to the work in CALCIUM REDUCTION AND LONGEVITY 369 which calcium concentration was varied as described under methods. To illustrate the nature of the observations the data of one experiment are presented in Tables I and II. The eggs were obtained at 2:13 P.M. August 4 from an urchin collected and brought to the laboratory at 11 :10 A.M. Ninety-eight per cent cleaved on test. After washing in two changes of 250 cc. of sea water, and concentrating into 50 to 60 cc., a 10-cc. sample was placed in each of the experimental solutions at 4 :30 P.M. An equal sample in sea water served as a control. In Table I the top figure at each age is the percentage disintegrated, i.e., visibly in bad condition. The lower figure is the number of eggs upon which the percentage is based. Table II represents the percentage TABLE II Effect of dilution with Ca-free medium on prolongation of fertilizability in surviving eggs. Upper figures percentage cleaved, lower figures number of eggs counted. Age in Hours at Fertilization In Sea Water In 2/3 Ca In 1/3 Ca In 1/6 Ca In 1/12 Ca 93 94 81 89 87 17 86 96 102 105 99 71 50 69 61 60 22J 61 102 96 105 114 65 50 67 59 80 40| 62 58 116 78 84 78 67 58 48| 101 104 127 22 62 84 67 41 109 102 2 45 91 51 89 which cleaved as a fraction of those which appeared to be in good con- dition. The lower figures are numbers of eggs upon which the per- centages are based. These data exclude disintegrated eggs and show physiological condition in the survivors, insofar as fertilizability and cleavage measure such condition. Table III is derived from Tables I and II, by the product of percentage of cleavage and percentage of live eggs, and is appended to give an indication of the extension of fer- tilizable life in terms of the total population at the beginning. 370 VICTOR SCHECHTER The photographs in Plate I are taken from the above-described experiment. Photographs 1 to 4 are of eggs aged 39^> hours in sea water, % calcium, %, and ^o calcium. They show progressively fewer disintegrated with greater lowering of the normal calcium concentra- tion. Photographs 5 to 8 are, respectively, samples of the same, fer- tilized at 40% hours of age. Photographs 9 and 10 contrast the condi- tion at 66% hours of age of unfertilized eggs in sea water (all of these having disintegrated long before) and in %2 calcium. The latter, fer- tilized when 67 hours old, show a very high percentage of cleavage (Photograph 11) and even when 91 hours old (Photograph 12) the cleavage percentage is high and early cleavage, at least, is normal. TABLE III Effect of dilution with Ca-free medium. Surviving percentage of total popula- tion able to cleave. Age in Hours at Fertilization In Sea Water In 2/3 Ca In 1/3 Ca In 1/6 Ca In 1/12 Ca 17 89 91.8 81 87.7 87 22^ 61 45.3 67.3 60.5 60 40 ; 33.8 27 64 56.3 80 48 * 74 66.6 58 67 7.05 56 83 91 0.52 37.4 Seven experiments qf the kind just described were performed. In three the eggs of two or three females were combined, giving a total of eight and in four others single lots of eggs were used. The results of these twelve are presented in Fig. 1. Calcium concentration, in terms of that normal to sea water as unity, is plotted on the ordinate axis and time for survival of 50 per cent of the population at various concentrations on the abscissa. Full data support all points except at ^o calcium where measurements are available only from 5 lot> of eggs. It is apparent from the graph that the- span of life increased steadily upon dilution with Ca-free medium. From 30 hours in sea water, longevity rose 300 per cent to 90 hours at %o the normal calcium (and other changes in this solution to be brought out later). Four experiments were carried out with a constant fraction of sea water, calcium variation being obtained by adding CaCU to the artificial medium as described under methods. In this way the pH was constant and differential dilution of any possibly toxic constituents of sea water was also avoided. The results are expressed in Fig. 2, and for compari- son the range from 0.88 to 0.167 calcium is shown on Fig. 1 by a dotted CALCIUM REDUCTION AND LONGEVITY 371 W*"«V.' :-Jas S£ • « 6 ^T»" 4f *« "k" 4* »*" *.* f» • • j » ^«C % ' • *-Cf ?*fif*y -^ tr tvvri ^%:7 - *y* S^^S^^l 12 EXPLANATION OF PLATE I PHOTOGRAPHS 1-4. Eggs aged for 391/{> hours in normal calcium (sea water), %, Vs, and Vis calcium respectively. PHOTOGRAPHS 5-8. Cleavage in eggs fertilized after 40';4 hours in above solu- tions, respectively. PHOTOGRAPH 9. Eggs aged 66H hours in sea water. PHOTOGRAPH 10. Eggs aged 6614 hours in ^2 calcium. PHOTOGRAPH 11. Eggs aged in Vi2 calcium, fertilized when 67 hours old. PHOTOGRAPH 12. Eggs aged in Vr2 calcium, fertilized when 91 hours old. 372 VICTOR SCHECHTER line. In Fig. 2 the larger dots represent the average of observations on four lots of eggs and the smaller dots those of two. The pH for all points was 7.6 except that of sea water which was 8.3. In Fig. 3 are given the averaged results of two experiments with y$ sea water plus •"','; dextn»>e and difiVivnt amounts of isosmotic CaCL added. The measured pH varied at random from 6.7 to 7.0. 0.1)1) - •2 0.33 rt 4-1 g 3 £ _3 go.67 1 .00 30 40 50 60 70 SO 90 Hours for Fifty Per Cent to Survive FIG. 1. F.tTeet "i" dilution <>i" ^-a water with isosuiotic Ca-free salt M>1uti"ii on the longevity of uufertili/.ed . lr/>, /,•/,/ e^ti cells. Unity on the onlinate represents undiluted sea water, pll i> S.3 rxcrpt ~.<> in ',; sea water and 7.4 in ^o. DISCUSSION AND ( 'o.\( i.rsiox The results leave no dotiht. I liclirvr. that drcrraM1 in calcium pro- longs the life of the .Ir/niciu egg. \\"ith a >tep-by-.ste]> i-liminatioii of other variables this phenomenon becomes increasinglv clear. Divergence between l'ii;>>. 1 and 2 in the ran^e ot calcium conci'ii- tration close to that of >cja water and the fact that the control group falls off the curve on Fig. 2 mav hi- accounted for on the basis of pll. Smith and ('lo\\c> (1(>24) have already shown that the optimal point for longevity is belo\\- the pi I of sea water. I'.ut even with constant pH. as in Fig. 2, the effect of decreased calcium is evident. The steepne-- CALCIUM REDUCTION AND LONGEVITY 373 of the curve between calcium concentrations 1 and 2.6 may mean that in this range pH becomes the limiting factor for longevity, until the 0.00 - 1.00 03 c o U >--•, 3 a U 2.00 30 40 50 60 70 Hours for Fifty Per Cent to Survive 80 Fig. 2. Effect of calcium reduction on longevity of Arbacia eggs in % sea water + % Ca-free solution. pH constant at 7.6. Calcium varied by addition of 0.29 molar CaClo. Point at unity on the ordinate represents sea water control, pH 8.3. lethal action of calcium is again sufficient to bend it toward the left. It may be that these two factors are not mutually exclusive but that acidification acts through the mechanism of calcium by a reduction in 374 VICTOR SCHECHTER the base-binding power of cellular materials as they approach their isoelectric points. The use of a non-electrolyte in the experiments of Fig. 3 excludes the possibility that increase in sodium, potassium and magnesium, which was colligative with decrease in calcium in the other experiments, is re- o.oo - l-°° 2 c u r J - _ 2.00 1 1 1 1 _'ii 30 40 50 60 70 Hours for Fifty Per Cent to Survive 80 FIG. 3. Effect of calcium concentration on longevity of Arbacia eggs in % sea water + % 0.95 molal dextrose. pH ± 6.9. Calcium varied by addition of isos- motic < ;i< '!,. sponsible f<>r the results. Here there was no electrolyte variable but calcium ; and it may perhaps be significant that there appears to be a most nearly straight-line relationship. Actually, however, the curve may be S-shaprd, but unlike that of Fig. 2, swinging quickly toward the ordinate axis in the absence of ions antagonistic to calcii CALCIUM REDUCTION AND LONGEVITY 375 The influence of dextrose is, of course, a disturbing factor. In pure dextrose the permeability of the egg rises, but 0.00005 M CaCl., is suffi- cient to keep it at sea water level (McCutcheon and Lucke, 1928). The lowest concentration of CaCL in my experiments with dextrose is fifty times this amount. Here, however, calcium is antagonized by sodium and potassium (in ^ calcium solution). At 0.34 concentration there is an excess of 0.002 M free calcium ; and at this point, with permeability change presumably ruled out, life duration is 350 per cent greater than at 2.6x calcium, as seen from the graph. It must be kept in mind, however, that the permeability studies were made on fresh material ex- posed to the experimental solutions for short periods of time. The use- fulness of the results in this discussion assumes transferability to my experimental conditions involving long exposure. As this writing is in progress an advance abstract has appeared in which Whitaker (1936) reports that 1 per cent dextrose and 1 per cent alcohol in sea water, each prolongs the fertilizable life of Urechis eggs. Although it is premature to comment in the absence of a complete ac- count, may I point out (Heilbrunn, 1934) that with low free calcium (which is probably the condition in the interior of an egg) 1 per cent alcohol (and 1 per cent ether) prevents the surface precipitation re- action, in which calcium is significantly involved. Also, Gray (1926) found that sugar and alcohols act like calcium in their stabilizing effect on the matrix of Mytilus gill tissue. It may, therefore, be that the ac- tion of alcohol and of sugar on the Urechis egg is effected through a change in the calcium equilibrium. If calcium is a factor in determining longevity in Arbacia egg cells, we might expect a similar mechanism within the body of the sea urchin, where the cells may live for a prolonged period. Analysis of the coelomic fluid 2 of female sea urchins showed a calcium content of 0.395 mg./cc. against 0.41 mg./cc. in sea water. On the basis of the data of this paper the difference, of about 31/2 per cent, is insufficient to account for longevity of eggs within the animal. However, calcium content in the gonad may be lower than that of the ccelomic fluid. No direct evi- dence is available but Heilbrunn (1928, p. 148) has reported that eggs taken from the ovary are much more resistant to cytolysis in isosmotic CaCL than if they are first equilibrated with the salts of sea water. Therefore, regardless of the negative results from ccelomic fluid analy- sis, a calcium mechanism for longevity of ovarian eggs is a possibility. Finally, because of my interest in bioelectric phenomena (Schechter, 1934), may I point out that electrical potentials are correlated with cal- - The analysis was made by Daniel Mazia of the University of Pennsylvania to whom I wish to extend mv thanks. 3/6 VICTOR SCHECHTER cium changes within the range of these experiments (Dan. 1936). Electrophoretic potential rises from • - 30 to - - 20 millivolts as cal- cium concentration is reduced. SUMMARY 1. At ±21° C. in sea water 50 per cent of the eggs of Arbacia pniictulata remained undisintcgrated for about thirty hours. 2. With decrease in calcium to 1-J1, that found in sea water and a drop in pH from 8.3 to 7.4, 50 per cent of the eggs survived for ninety hours. 3. With pH constant at 7.6. 50 per cent of the eggs survived for 53 hours in an artificial solution with normal calcium concentration and for 90 hours in yr^ calcium. 4. With other electrolytes kept constant by the use of dextrose solu- tion 50 per cent of the eggs survived for 20 hours in 2.6x the normal calcium and for more than SO hours in ^o concentration. REFERENCES CITED DAX, K., 1936. ElectrokimnV Minlio of marine ova. I'hysiol. Zool., 9: 43. GRAY, J., 1926. The propcrtR-- »t~ an intercellular matrix and its relation to elcc- troht. ~. Brit. Jour. /:.r/vr. Biol.. 3: 167. HEILBRUXX, L. V., 1928. Colloid Chemistry of Protoplasm. I'.erlin. HEILBRUXX. L. V., 1^34. The effect of anesthetics on the surface precipitation re- tion. Bi,>l. Hull. 66: 2M. LlLLlE, R. S.. 1912. Certain means by which starfish e.utis naturally resistant to fertilization may be rendered normal and the physiological conditions of this action. Biol nidi. 22: 328. LOF.B, J., 1913. Further experiments on natural death and prolongation of life in the egg. Jour. /f.r/vr. Zool., 15: 2(11. LOEB, J., AND J. H. NORTHROP, 1917. \Yhat determines the duration of life in Metazoa? Proc. Xat. Acad. Sci., 3: 3N2. McCuTCHEOx, M.. AMI I'.. l.riKK. 1'^'S. I'liTect of certain electrolytes and non- electrolytes on permeability of living cells tk- sea water. Biol. Bull.. 52: 161. SCHECHTER, V.. 1934. Electrical control of rhi/nid formation in the red alga, Griffithsia bornetiana. Jour. Gen. I'liysii'l.. 18: 1. SCHECHTER, V., 1936. Comparative hypotouk- cytolysis of several types of in- vertebrate egg cells and the influence of age. Biol. Bull.. 71: 410. SMITH, H. \V., AND G. H. A. CLOWES, 1924. Tin- influence of hydrogen ion con- centration on unfertilized Arbacia, AMi-ria^ and Chsetopterus eggs. Biol. Bull.. 47: 304. WHITAKER. D. M., 1936. Extension of the fcrtili/abk- life of unfertilized Urechis eggs by alcohol and by dextrose. Wistar Inst. Bibl. Service, No. 284, September 15. WOODWARD, A. E., 1918. Studies on the physiological significance of certain precipitates from the egg secretions of Arbacia and Asterias. Jour. Ex per. Zodl., 26: 459. THE EFFECT OF ULTRA-VIOLET LIGHT UPON EARLY DEVELOPMENT IN EGGS OF URECHIS CAUPO 1 H. Y. CHASE (From the Department of Zoology, Hoivard University, Washington, D. C.) INTRODUCTION During the summer of 1936 while at the Hopkins Marine Station, Pacific Grove, California, the writer observed the effect of ultra-violet radiation upon the early development in eggs of UrccJiis caupo, Fisher and MacGinitie (1928a, 1928b). The eggs of this marine worm are immature when shed, hence maturation ensues after fertilization. The present paper gives results which show the effect of ultra-violet radia- tion upon maturation and first cleavage, together with the effect upon the rate of these early stages. The study does not aim at a physical analy- sis of radiation effects, hut rather, at the qualitative effects of radiation upon maturation and first cleavage. MATERIAL AND METHODS The worms were collected at Elkhorn Slough, Monterey Bay Re- gion, California. To insure the use of gametes in best condition the usual practice was to collect animals as often as tide conditions were favorable. Experimental animals wTere kept no longer than three weeks. They gave a yield of eggs and sperm of high fertilization capacity throughout the entire period of the work. Quantities of gametes were removed from the worms (males and females were kept in separate aquaria) when needed. Eggs were placed immediately into dishes con- taining 250 cc. sea water and dry sperm were kept in covered Syracuse watch glasses. Irradiation was by means of an Analytic Model Quartz Lamp, Hanovia, which operated on a 110-volt circuit, alternating current, 60 cycles, 5 amperes. All handling of eggs and sperm was done at a controlled temperature of 20° C. (plus or minus 2.5° C). For regulation of the temperature of the sea water into which the eggs were placed several water-baths were used. Within each bath a thin-walled, flat-bottomed, glass dish 1 This work has been supported in part by a grant from the Committee on Radiation, National Research Council. 377 H. Y. CHASE was supported. This dish held the eggs and adjacent to them was placed the bulb of a calibrated thermometer. Conditions were standardized as much as possible. The ultra-violet lamp was allowed to burn for 10 minutes before exposures were made. During this time traces of ozone were removed and the light reached its maximum intensity. Radiations were made at a distance of 30 cm. from the source of the light and the unfertilized eggs were given non- lethal exposures to the full spectrum of the mercury arc, for periods of 30, 35, 40, 45. 50. 55. and 60 seconds. An experiment consisted of placing eggs whose controls gave 99 per cent to 100 per cent polar body extrusion and 97 per cent to 100 per cent first cleavage, in a thin layer on the bottom of a dish filled with sea water to a depth of 2.5 mm. This dish was placed in the constant tem- perature bath under the ultra-violet light and the eggs were irradiated. The dish was then removed and placed in a constant temperature bath fitted to the microscope, where the eggs were observed for five minutes for activation (development of fertilization membranes) by the radia- tion. When no activation was found the eggs were inseminated with a standard sperm suspension (one drop of dry sperm from a capillary pipette to five cc. sea water) and allowed to develop. The eggs were examined later for the extrusion of polar bodies and for first cleavage. Counts were made of the number of eggs in 100 which extruded first and second polar bodies, and of the number of eg^.s in 10'.) which showed first cleavage. These counts were repeated 19 times for each group of irradiated eggs. Twelve sets of experiments were made so that the data given below are for 24,000 eggs at each exposure. The time-lapse was measured also, from the time of the insemina- tion of the eggs to the time of the beginning of the particular stage con- sidered. The time-lapse measurements for polar body extrusion were made on the reacting half of 10 eggs in the field of the microscope at the time of observation and the criterion was that moment when the polar body was sufficiently extruded to be identified unmistakably at the periphery of the egg. In cases where eggs were irradiated for 50. 55, or 60 seconds it was not possible to measure1 the- time-lapse for the re- acting half of 10 eggs because of the small number of polar bodies which could be found at the periphery of the eggs. I therefore meas- ured the time-lapse for individual eggs and considered the average of the measurements made in the 20 counts as the time-lapse measurement. The criterion for first cleavage was that moment when the egg was elongated unmistakably at or in one or more axes (according to the regularity of cleavage) and the cleavage furrow could be seen as a thin. shining line which was especially clearly defined with the type of ULTRA-VIOLET LIGHT AND DEVELOPMENT IN URECHIS EGGS 379 illumination used. The time-lapse was measured for the reacting half of 10 eggs in the field of the microscope, hut at exposures of 50, 55, or 60 seconds when there were very few cleaving eggs it was necessary to measure the time-lapse for individual eggs and to take the average of the measurements made in the 20 counts as the time-lapse from in- semination to first cleavage. • THE EXPERIMENTS The extrusion of polar bodies and the first cleavage of the egg, to- gether with the time after insemination at which these phenomena occur, are affected when the egg is irradiated before fertilization. The action TABLE I Percentage of polar body extrusion and first cleavage in eggs of Urechis caupo exposed to ultra-violet light for various lengths of time at a distance of 30 centimeters. Length of Exposure (seconds) First Polar Bodies Second Polar Bodies First Cleavage 30 91.3 88.9 73.1 35 78.4 73.9 57.7 40 63.8 55.1 40.7 45 31.4 18.5 14.0 50 9.5 2.8 5.2 55 4.7 1.4 4.2 60 2.3 0.68 2.5 All controls showed 99 per cent to 100 per cent polar body extrusion and 97 per cent to 100 per cent first cleavage. The percentage is based upon the average of 12 experiments in each of which 2,000 eggs were counted. of ultra-violet light upon these early stages in development is reported as the percentage of extrusion of polar bodies and cleavage, and the effect upon the rate of reaction is reported as the percentage increase in the time-lapse for each stage. The results are presented. Percentage Extrusion of Polar Bodies and Percentage Cleavage in Irradiated Eggs From the results shown in Table I it is apparent that extrusion of polar bodies and the first cleavage of the egg are suppressed when eggs are irradiated with ultra-violet light. Different exposures varying from 30 to 60 seconds, with an increase of five seconds for successive dura- tions of exposure, produce increasing suppression of the developmental stages until the longer periods of exposure are reached when there is a practically total suppression. Average percentages for all irradiations are presented for each duration of exposure which show unmistakably 380 H. Y. CHASE that the suppression of polar body extrusion and of first cleavage varies directly with the length of exposure. Since the data are based upon observations and counts of large numbers of eggs (the average per- centages are for 24,000 eggs) the evidence is fairly conclusive. These results on the suppression of polar body extrusion confirm observations TABLE II • Effect of ultra-violet light upon the time-lapse from insemination to polar body extrusion and to first cleavage in eggs of Urechis caupo exposed for various lengths of time at a distance of 30 cm. Percentage increase in time-lapse equals time-lapse in radiated eggs — time-lapse in control time-lapse in control X 100. Stage I. ninth of Exposure • .Mil:-) Time-lapse in Radiated Eggs (minutes) Time-lapse in Control (minutes) Per Cent Increase in Time-lapse 30 30.0 29.8 0.67 35 31.2 28.5 9.47 40 32.1 28.3 13.43 First polar bodies 45 50 34.8 38.3 28.6 27.5 21.68 39.27 55 41.3 28.3 45.94 60 42.3 28.6 47.90 30 44.3 42.3 4.73 35 46.4 42.6 8.92 Second polar bodies 40 45 47.3 52.9 42.3 43.8 1 1 .82 20.7S 50 55.6 42.1 32.07 55 57.7 42.7 35.10 60 59.7 42.2 40.76 30 69.2 67.4 2.67 35 70.1 67.2 4.31 40 70.3 68.0 3.38 First cleavage 45 50 75.2 82.5 68.9 69.2 9.14 19.22 55 84.0 68.7 22.27 60 89.2 68.5 30.22 Each time-lapse measurement repn •-> -m - iln- .1 venire time from insemination to the particular stage and is based upon il.ita for thr reacting half of 24,000 eggs observed at each exposure save 50, 55, and 60 xromls. At these exposures the average time-lapse was taken for the number of ri^s in 24,000 which showed tin- particular stage. of Just (1933) for eggs of \crcis liinbata. Knmi the protocols of this observer the evidence shows that radiation effects were so pronounced that essentially a total suppression of polar body extrusion was caused when eggs were irradiated for 60 .second^ at a distance of 25.5 cm. from the lamp. Just reported similar results when eggs were given longer exposures at different distances from the lamp. ULTRA-VIOLET LIGHT AND DEVELOPMENT IN URECHIS EGGS 381 Percentage of Increase in Time-lapse front Insemination to Polar Body Extrusion and Pirst Cleavage The data in Table II show the effect of ultra-violet radiation upon the rate of the early developmental stages in eggs of Urechis caupo. Time-lapse measurements for eggs irradiated at different exposures were compared with similar measurements in controls and the relation be- tween the time-lapse from insemination to a particular stage in both the experimental and the control eggs was given as the percentage increase in time-lapse. The data show that wherever a developmental stage was affected by radiation the reaction time of the egg (time-lapse from in- semination to the stage) was retarded. At the comparatively short lengths of exposure the percentage increase in the time-lapse was small. Each increase in length of exposure caused an increase in the percentage increase in time-lapse. The steady increase in the time-lapse from in- semination of the egg to a particular stage as indicated in Table II may be regarded as evidence that irradiation of eggs not only affects certain stages in development but also affects the reaction rate of the eggs. DISCUSSION Suppression of polar bodies and of cleavage by such agencies as ex- posure to extremes of the viable range of temperature for eggs-, ' treat- ment with hypotonic sea water, or subjection to the action of narcotizing substances, before or after activation, has been reported by many in- vestigators. The data reported here suggest ultra-violet light as a most effective agent for suppression of early developmental stages. The exact nature of the action of ultra-violet radiation is not defi- nitely known. An important factor appears to be the extent of the absorption of radiant energy by the inner protoplasm of the egg and by the superficial layer, each of which is a site of complex reactions which underlie morphological changes and developmental processes. Evi- dence presented by Redfield and Bright (1921) and Just (1933) shows an alteration of the initial changes involved in the cortical reaction in the egg of Nereis fertilized after exposure to ultra-violet light. Another type of evidence is given by Tchahotine (1921a), who correlated local centers of injury produced in the peripheral layer of sea urchin eggs with permeability changes. Tchahotine (1921&) further pointed out the probability of the coagulation of the colloids of the superficial layer of the irradiated sea-urchin egg. While eggs of Urechis caupo, unlike eggs of Nereis limbata, extrude no jelly following fertilization and show no visible alteration of the superficial layer other than the separation of a tough, pellicle-like membrane (see Chase, 1935) from the vitellus of II. Y. CHASE the egg which develops into the fertilization membrane, it is possible that certain changes take place in the egg cortex which are affected by radiation as in the case of other species of eggs. A serious alteration of the physical and chemical properties of the peripheral layer of the eggs conceivably may be a factor in the suppression of polar body ex- trusion in eggs which are radiated before they are fertilized. On the other hand, the developmental stages which have been studied are closely related to various phenomena which occur deep within the egg. Of these the viscosity changes are the most widely studied. \Yhile no attempt was made to observe the effect of radiation on such changes in eggs of Urcchis caupo, evidences of the effect of ultra-violet radiation upon viscosity of protoplasm of other eggs is reported. Of these investigations the most significant for this particular problem are the observations mi the egg of Ascaris, which falls in the same category with the egg of I 'rcchis with respect to the time at which the egg may be fertilized. Schleip (1923), in the course of observations on the effect of ultra-violet radiation on morphological components of Ascaris eggs, reported increased viscosity. Similarly, Ruppert (1924) centrifuged radiated egg> in his studies on the effect of ultra-violet light upon dif- ferent stages in the development of eggs of Ascaris and his results indi- cate increased viscosity. Such changes alone may be associated with marked inhibition of phenomena which underlie maturation and cleav- age processes and the resulting suppression of these stages in develop- ment. A i vtological study of the irradiated eggs is being made and it may give significant evidence on the effect of ultra-violet light upon morpho- logical changes in the eggs. SUM MA UN' 1. When unfertilized eggs of I'rcchis caupo are exposed to ultra- violet light for different lengths of time, then fertilized, polar body ex- trusion and first cleavage are suppressed. In addition to the suppres- sion of these >tam-> radiation causes an increase in the reaction rate of the egg which bears a direct relation to the length of exposure. 2. Possible factors in the suppression of these stages in maturation and first cleavage are the alteration or probable injury of the egg cortex and accompanying changes in its chemical and physical properties, and viscosity changes in the egg endoplasin which inhibit internal phe- nomena. The writer wishes to express his gratitude to Dr. A. C. Giese and Mr. E. W. Lowrance, School of Hiulogical Sciences, Stanford University, and Professor ULTRA-VIOLET LIGHT AND DEVELOPMENT IN URECHIS EGGS 383 I -Yank Coleman, Department of Physics, Howard University, for technical advice and assistance. He is also indebted to Dr. \V. K. Fisher and the staff at Hopkins Marine Station for their many courtesies. BIBLIOGRAPHY CHASE, H. Y., 1935. The origin and nature of the fertilization membrane in various marine ova. Rial. Bid!.. 69: 159. FISHER, W. K., AND G. E. MACGINITIE, 1928<;. A new echiuroid worm from California. Ann. and Mag. Nat. Hist., Ser. 10, 1: 199. FISHER, W. K., AND G. E. MACGINITIE, 19286. The natural history of an echiuroid worm. Ann. and Mag. Nat. Hist., Ser. 10, 1: 204. JUST, E. E., 1933. Observations on effects of ultra-violet rays upon living eggs of Nereis limbata exposed before insemination. Arch. Entw.-Mcch.. 130: 495. REDFIELD, A. C., AND E. M. BRIGHT, 1921. The physiological changes produced by radium rays and ultra-violet light in the egg of Nereis. Jour. Phvsiol., 55: 61. RUPPERT, W., 1924. Empfindlichkeitsanderungen des Ascariseies auf verschiedenen Stadien der Entwicklung gegeniiber der Einwirkung ultravioletter Strahlen. Zcitschr. f. u'iss. Zool., 123: 103. SCHLEIP, W., 1923. Die Wirkung des ultravioletten Lichtes auf die morpho- logischen Bestandteile des Ascariseies. Arch. f. Zcllforsch.. 17: 289. TCHAHOTINE, S., 1921a. Sur le mecanisme de 1'action des rayons ultra-violets sur la cellule. Ann. dc I' lust. Pasteur, 35: 321. TCHAHOTINE, S., 19216. Les changements de la permeabilite de 1'oeuf d'Oursin localises experimentalement. Compt. rend. Soc. dc Biol., 84: 464. THE ELECTRICAL CHARGE ON NUCLEAR CONSTITUENTS (SALIVARY GLAND CELLS OF SCIARA COPROPHILA) LEON CHURNEY AND HERBERT M. KLFJX 1 (From the Department of Zoology, University of Pennsylvania) In spite of the importance of electric charge in the behavior of col- loidal materials, very little work has been done to determine either the sign or the magnitude of the charge on the particles of the cell nucleus. It is possible by cataphoretir experiments to determine at least the sign of the charge on the various structural elements of the nucleus. At the present time there is almost no direct data on the- sign of the electric charge on animal chromosomes or on the nuclei of living animal cells. Botta is one of the few investigators who has attacked the problem. He reports that the chromosomes of embryonic chick heart cells are nega- tively charged during mitosis, but that there is little or no effect of the electric current on the chromatin during interkinesis. von Lehotzky states that he was unable to obtain results in his experiments on Protozoa and on frog blood cells. It is possible to obtain definite information concerning the sign of the charge on giant chromosomes such as are found in the salivary glands of fly larvae. The glands of the dipteran Sciara coprophila are especially favorable material for cataphoretic studies. The glands were dissected out of the larvae and studied in the Ringer's solution used by Belaf for grasshopper spermatocytes. The isotonicity of this solution with the salivary gland cells was determined by comparing glands in body fluid with those in the Ringer's solution. The electrical set-up used in these studies consisted of a non- polarizable system of Cu-CuSO, electrodes with agar bridges. The lower extremities of two tubes, 6 mm. in diameter, were filled with a 3 per cent agar gel made up in Ringer's solution ; above this gel was placed a saturated solution of CuSO,. into which the ends of the copper wires carrying the current were introduced. A double-throw switch in the circuit permitted the reversal of the direction of current flow. A gland, dissected out in Ringer's solution, was placed in a drop of this medium on a glass slide, and a cover-glass was supported above it. The agar tips of the electrodes were placed in contact with the fluid, at oppo- 1 We arc pleased to acknowledge tlic aid of Dr. L. V. Heilbrunn of the Uni- versity of Pennsylvania during the course of this investigation. We arc also grateful to Dr. C. W. Mctz of Johns Hopkins University for furnishing us with cultures of Sciara copropliila. 384 KLKLTKICAL CHARGE ON NUCLEAR CONSTITUENTS 385 site ends of the slide. By varying the amount of fluid between the electrodes the current strength was varied from approximately 5 to 25 milliamperes. The strength of the current passing through the nucleus was not. of course, determined. In the course of these experiments, one of the difficulties which pre- sented itself was that of ascertaining whether or not the cell was alive. We believe that the nuclei and the chromosomes contained within them were alive, at least during their first responses to the passage of the current. In any case, all subsequent reactions were essentially the same, even in cases where there was reason to suspect that the cell was dead. After the current had passed through the cell for a short time, appar- ently irreversible cytosomal coagulation set in, indicating the death of the cell. In cases of extreme coagulation, very little, if any, response to the passage of the current was noted on the part of either the nuclei as a whole or the chromosomes. In order that the chromosomes migrate cataphoretically, it is essential that the electric current pass through the gland cell and through tbe nucleus. Conditions must, therefore, favor such a passage of current. If the gland is placed in a large quantity of salt solution, practically all the electric current will pass around the cells. However, as the quantity of surrounding fluid, or its conductance, is decreased, the current will begin to flow through the cells. In the case of the salivary glands, the position of the gland in relation to the direction of current flow is of primary importance. If the long, spindle-shaped gland is parallel to the lines of current flow, the current will tend to flow along the central canal of the gland. On the other hand, if the gland is perpendicular to the lines of current flow, there is much more tendency for the current to pass through the cells, partly because of the fact that in this position the canal does not act as a short circuit, but also because of the fact that the path of the current leads through only two cells, and not through a long line of them. The first apparent effects of the current flow on the nucleus were usually an increase in the nuclear volume, and an increase in the refrac- tivity of the chromosomes. The nucleus as a whole, i.e., the nuclear membrane, moved toward the cathode (positive charge), whereas the chromosomes contained within it moved toward the anode (negative charge). These two opposing movements are simultaneous (see Fig. 4). When the direction of current flow was reversed, both the nuclei and the chromosomes responded instantaneously to the change by a cor- responding reversal of their migration (see Figs. 1, 2 and 3). This re- versal of migration was obtainable several times before the death of the cell and the subsequent coagulation prevented any response. When the circuit was broken, the nuclei and the chromosomes tended to return to LEON CHURXEY AND HERBERT M. KLEIX /' < 4- . > » f ' ' • » tt • - KXIM.. \\.\TIO\ ( )]' IM.ATI-: I 4 Photomicrographs <>f tin fixed and un>t;nm-d ci-lK nf tin.1 salivary .ulamK of Sciara coprnphila larv;c, showinjr the (lit it of the pa^sa^e nf electrical currents. The anode i^ indicated, in each case, hy a + >iiiii. I;i,mires 1, 2. and 3 are the same cells under different experimental conditions. 1'iuure 4 shows cells of another preparation. FH;. 1. Chromosomes migrating toward the anode, nuclear membrane toward the cathode. Note the crenation of the nuclear meinhrane as it is pressed against the cytosomal granules, l-'ight milliamperes. I'u;. 2. Same, direction of the current reversed. ( "hromosomcs migrating toward anode. Eight milliamperes. Same, direction of current again reversed, establishing same polarity as in 1. Eight milliamperes. EH;. 4. Another preparation. The nucleus as a whole is moving toward the cathode; its former position in the cytosome is indicated by the clear area toward the anode. The chromosomes are shown oriented toward the anode. Ten milli- amperes. ELECTRICAL CHARGE ON NUCLEAR CONSTITUENTS 387 their original positions in the cell. This was due, presumably, to a polar- izing current in the reverse direction. The movement toward the cathode and the simultaneous movement of the chromosomes toward the anode brought about distortions and abnormalities of the nuclear membrane (see Fig. 1). The chromo- somes, in their movement toward the positive pole, often exerted suffi- cient pressure against the nuclear membrane to bring about an elonga- tion of the nucleus in the direction of the current flow. On the other hand, the opposing force with which the nucleus as a whole moved to- ward the cathode was at times great enough to carry the chromosomes with it toward this pole (see Fig. 4). In such cases, however, the chromosomes always remained on the anodal side of the nucleus. The nuclei, in their movement toward the cathode, tended to resume their spherical shape until the opposing motion of the chromosomes made this impossible, as described above, or until their movement toward the cathode was prevented by the densely packed granules of the cytosome. In the latter case, the nuclei often flattened themselves against these granules, while the anodal side of the nuclear membrane continued to flow inward (toward the cathode), carrying the chromosomes with it. This pressure of the membrane against the chromosomes often brought about a crenation of the membrane. Upon the breaking of the circuit, the nuclei tended to return to their former positions in the cells, and in doing so, tended to resume their spherical shape. In cells which were mechanically crushed, nuclei often remained in- tact and floated partially or entirely free of the cytoplasm. Nuclei which were altogether free to move through the medium migrated to- ward the cathode. Since these nuclei were approximately midway be- tween the slide and the cover-glass, their migration to the negative pole could not be due to the water current of endosmosis that flows in the opposite direction through the center of the chamber. The chromo- somes, in these cases, were attracted to the anodal side of the nuclei, but with a force generally insufficient to distort the nuclei or to deter their rapid motion toward the cathode. Nuclei partially free from the cyto- plasm, but adhering to the latter, displayed the characteristic reactions described in preceding paragraphs. The observations indicate that the charge on the chromosomes is negative. This is in accord with various studies on plant chromatin (Pentemalli, McClenclon, Hardy, Meier, Zeidler, von Lehotzky). Most of the observations of these investigators were made on material fixed after the exposure of the material to electric current, and such a proce- dure is perhaps open to question. However, von Lehotzky studied liv- ing plant cells. The observation that the nucleus moves toward the cathode is new. LEON CHURNEY AND HERBERT M. KLEIN Perhaps other observers failed to note such a movement because of the fact that for the most part their observations were made on material fixed presumably after the creation of flow of the current. In our own studies we observed that ;t^ M>I>II as the circuit was broken the nuclei immediately tended to return to their original positions. The positive charge on the nucleus as a whole is perhaps conditioned by a positive charge on the cytoplasmic colloids. Just as a quartz particle or a blood cell takes its charge from the colloids in which it is immersed and which arc adsorbed on its surface, so the nucleus may owe its positive charm- to the charge of protoplasmic colloids (in this connection compare Heilbrunn, 1928). The fact that isolated nuclei are positively charged does not necessarily invalidate this concept. These experiments give us an indication of the complex electrical character of the cell components, in contrast to the excessive simplicity revealed in the cataphoresis of quartz particles or other inanimate sys- tems. We have a living cell with a negative charge at the surface, per- haps a positive charge in the cytosome, a positive charge on the nucleus imbedded in it, and a negative charge within the nucleus (on the chro- mosomes). In conclusion, this paper has attempted to stress the fact that the nucleus as a whole is positively charged, while the chromosomes within it are negatively charged. The complexity of the living system with respect to the electrical charge of the cell components has been contrasted with the apparent simplicity of inanimate systems. A more quanti- tative study of these phenomena is planned. BIBLIOGRAPHY BELAK, KARL, 1929. Bcitrage zur Kausalanalyse appreciation to Professor W. R. Coe for his advice and criticism throughout the course of this work. MATI.KIAL AND MI.I nous V. mercenaries is found along the Atlantic Coast from the Gulf of St. Lawrence to the coast of Texas. In the past few decades this species has become established on the Pacific Coast where it was transplanted with the eastern ov.ster. The material gent-rally used in this work was collected in Long Island Sound, near Mil ford. Connecticut. Young clams were obtained at regular intervals from a single clam bed and prepared for microscopic examination. A large number of small clams was also taken from concrete tide-filling tanks with a capacity of several thousand gallons, at the U. S. Fisheries F.inlogical Laboratory in Milford, Connecticut, where environmental conditions, such as temperature and salinity of the water were observed and recorded. In the preparation of the mate- rial for study, standard methods of cytology and histology were em- ployed. In the present study all the observations on living material were checked by histologically prepared material in order to determine- ac- curately the characteristics of the constituent cells of the gonads. The method employed by Spiirck (1925), which consisted of boring a small hole through the shell of an animal and removing a piece of the gonad for examination, is not .satisfactory, as very misleading conclusions may be reached by assuming the condition of the entire reproductive system from the examination of a single small sample of the gonad. Similarly, the method of Orton ( 1('27). consisting largely of the examination of the living tissue to note the condition of the gonad with regard to the production of sex elements, is alsn upen to criticism because of the im- pi i.ssibilitv of distinguishing many important details under such con- ditions. AGK AND SIXK A review of the literature fails to disclose any important contribu- tions on the subject of sexuality of / '. mercenaria in general and on the development nt" the primary gonads in particular. The- only work is that of Melding (1912), where he states that the average hard-shelled clam is capable of spawning when in its third summer, for sexual products PRIMARY GONAD AND SEXUAL PHASES IN VENUS 391 could not be found at an earlier age. The sexes were found to be sep- arate, each animal presumably remaining either male or female all its life. He gives no histological data, and it is apparent from his work that his conclusions are based only on macroscopic examination of living clams and upon his observations of their spawning activities. It was impossible to obtain for this study small clams whose shell- length was less than 4 mm. long. Therefore the smallest animals stud- ied were 4-5 mm. long. These were found at Trumbull Beach, Long- Island Sound, during September and October. As the age of studied animals is of importance, several attempts were made to secure clams of a known age. This was accomplished by collecting seed clams and by keeping them under observation. There are two possible conclusions as to the age of small, 5-7 mm. clams, collected for this study during October-December ; namely, that they set in the year they were collected, in which case their age would have been only a few months, or that they were about 14-18 months old, setting the preceding year. Belding (1912), studying V . mercenaria along the coast of Massachusetts over a period of five years, found that the rate of growth of the clam is largely determined by its environment, and that, as a rule, the growth in any bed is fairly uniform. In his experiments at Monomoy Point, he found that the average size of a 14-months-old clam (collected in October) was 25.59 mm. Judging by the description of the Monomoy Point experiment, the environmental conditions there closely resemble those of Trumbull Beach, at which place the young clams were obtained for the present work. The similarity of conditions of the two places makes it logical to assume that the rate of growth of clams at Trumbull Beach is more or less identical to that of clams of Monomoy Point. Thus, clams 5—7 mm. long, collected in October— December, presumably were of that, and not of the preceding year set. This conclusion is supported by the observation of Belding (1912) that young clams im- mediately after setting showed an average gain of 3.4 mm. per month, which would make a length of about 10 mm. by the end of November. In the present paper all the remarks concerning the size of animals refer to the length of the shell. DEVELOPMENT OF PRIMARY BISEXUAL GONAD Gonad tissue first appearing in the young animal consists of a very thin layer of cells between the muscular body wall and the stomach. Gonads of the juvenile phase first appear, not near the muscular body walls, but a short distance from them at the level of the heart or slightly below it. The follicles of the juvenile gonad are at first composed of a single layer of germinal epithelium cells. These cells of elongated type 392 VICTOR L. LOOSANOFF are rather irregular in shape and size and possess very large, deeply- staining nuclei. During the early stages there is virtually no lumen spg.2 pg Fir.. 1. Primitive bisexual gonad of a yun^ clam soon after formation of the lumen. 7, indifferent cell; cp.c., follicular cells; DC., young ovocyte; spy.l. pri- mary spermatogonia; -v/>r.2), and spermatids (sf>t.) ; f.c., follicular cells. Nutritive cells (n.c.) are seen near the follicular wall. formed, the walls nf the follicles being almost in contact with each other. As the animal grows, the follicles of the primary gonad begin PRIMARY GONAD AND SEXUAL PHASES IN VENUS 393 to ramify through the loose connective tissue. A few weeks later the germinal epithelium begins to differentiate into gonia and at this time the lumen in the follicle is formed (Fig. 1). Rapid proliferation and specialization of cells follows. In small animals, 5-12 mm. long, the gonad does not extend very far in the ventral direction, seldom reaching below the line correspond- ing to the middle part of the stomach. The follicles of the gonad are not very numerous, and usually only six to eight of them can be found in one cross-section of the entire animal. In cross-sections the gonad follicles of such animals already show the lumen, which is usually oval or round in shape with a greatest diameter of 50-100 microns. Care- ful study of sections reveals that different gonad follicles of the same animal exhibit a widely different degree of development and bisexuality. Some of the follicles may consist of only a few indifferent cells while in others the germinal epithelium has already differentiated into male and female cells. All degrees of such processes can be observed. In some cases the follicle contains a few gonia along its wall and a mass of spermatogenic cells in early stages of development in the lumen (Fig. 2). In other, more advanced, cases spermatozoa are already formed, some- times occupying the largest part of the lumen, while a few small ovo- cytes showing mitochondrial filaments of yolk nuclei are found along the walls of the follicle (Fig. 3). Gonads at such stages are distinctly bisexual and because the proliferation of spermatogenic cells is very rapid such young gonads acquire a strong male appearance (Loosanoff, 1936o). In some cases the appearance of the follicles looks as though the spermatozoa were already discharged. For instance, a large sample of animals 5-7 mm. long was collected on Trumbull Beach in October, 1934. Upon examination of the prepared material it was observed that some of the follicles had already discharged spermatozoa. The lumen of such follicles was large and empty while along its walls young ovo- cytes and indifferent cells were present. As a rule, few phagocytes could be found inside the lumen (Fig. 4). With the purpose in view of establishing the age of those animals, very careful examination of their shells was made under a dissecting microscope. In not a single case was an indication of the winter ring found. The outer surfaces showed an embryonic shell and 7-13 growth lines formed at approxi- mately equal intervals (Fig. 5). Thus the studies of the shells indi- cated that the animals of that sample were only a few months old. If the evidence that V . rncrccnaria produces and discharges gametes during the first summer of its life is reliable, then it resembles in this respect many of the other bivalves whose sexual development is well 394 VICTOR L. LOOSANOFF sp.g-1 spg-1 f.c n-c n.c. oc. I;K;. 3. Primary bisexual gonad nf predominantly male character. Sperma- tozoa are pn-M-nt. occupying the center of the lumen. A few ovocytes (oc.) are lying along the follicular wall. /. indifferent cells; spt/.l. primary spermatogonia ; spg.2, secondary spermatogonia; spc.l. primary ^IH -\ niatocytes ; sf>c.2, secondary spermatocytcs ; .?/>/., spermatids in different stages of spcrmiotcleosis ; spz., sperma- tozoa; n.c., nutritive cells; f.c., follicle cells. Size 7-12 mm. IMG. 4. Gonad follicle of a young clam (7 mm.) after discharge of the first crop of spermatozoa. /, indifferent cells; oc.. young ovocytes; spy.l, primary .spermatogonia; n.c., phagocyte; f.c., follicle cells. PRIMARY GONAD AND SEXUAL PHASES IN VENUS 395 known (Nelson, 1922; Coe, 1936r). It is apparent, however, that there cannot be a definite rule applied to all Pelecypoda mollusks as to the time when the}' first form functional gametes. As a matter of fact, considerable differences are often found in closely related species and in different localities, as Sparck (1925) has so clearly demonstrated in 0. edulis. O. litrida is capable of passing through three sexual phases dur- ing the first year of development in warm waters (Coe, 1932), while O. rirg'mica of the northern part of the eastern coast of the United States does not form any ripe gametes until the second year (Needier, 1932). In V. mcrccnana, ripe spermatozoa are formed within 3-5 months after setting. Among several hundred young animals studied, approximately 98 per cent of them passed through such a protandric male phase. The remaining 2 per cent appeared to develop into females without passing through a functional male phase. FIG. 5. Shells of the young clams, age about 3 to 4 months, collected in Oc- tober-November, showing the absence of the winter ring. All of these were al- ready provided with gonads containing mature spermatozoa. Natural size. CONDITION DURING FIRST WINTER According to the writer's observations, the shell-growth of clams in Long Island Sound ceases in November. Small (6-10 mm.) clams ex- amined at that time differ very little from those studied in October. Again gonad follicles in various stages of development are found. In many follicles spermatogenesis continues and cells of all stages of de- velopment, including spermatozoa, are present (Fig. 6). In other cases spermatozoa have already been discharged and the distended lumen is virtually empty (Fig. 4). Often numerous phagocytes are seen invad- ing the lumen or attached to the walls of the follicle. Those presumably ingest the residual spermatic cells. In a few instances the gonad tissue has already ramified ventrally below the stomach and individual follicles 396 VICTOR L. LOOSANOFF can be seen confined to the space between the muscular body wall and tbe digestive gland. By the middle part of December the temperature of the water of Long Island Sound reaches the point at which the period of hibernation for this species begins (Loosanoff, 1936/>). Examination of young ani- mals collected during that period, which extends until the middle part of April, or until the water temperature rises above 5.0° C, reveals that B9-1 FIG. 6. Bisexual primary gonad »i' ;i \< >ung individual about 16 weeks of age showing the cells of both sexes hut predominantly <>i" male type. spi/.l. primary spermatogonia ; spg.2. secondary spermatogonia ; .v/v./. primary spermatocytes; spc.2, secondary spermatocytes; .?/>/., spiTiuatids ; spc.. spermatozoa; oc., large ovocytes showing mitochondria! bodies and yolk nncK i ; f.c., follicle cells. changes of only minor importance occur in the primary gonads. There is apparent neither increase in the nunibrr of cells constituting the gonad, nor increase in the size of the follicles. Spermatogenesis is virtually discontinued and cells of the intermediate sta^e-; are few in number. The spermatozoa, however, are retained in large numbers in many of the follicles. Unless carefully examined, many of the animals at this stage may be mistaken for true males, but the presence of young ovocytes showing distinct mitochondria] bodies and yolk nuclei indicates the actual PRIMARY GONAD AND SEXUAL PHASES IN VENUS 397 bisexual character of the gonads. In other cases the bisexuality of an individual is more easily noticed because of the large number of young ovocytes present (Fig. 6). A few pycnotic cells arc sometimes seen. Hibernation continues until the middle of April. The lowest tem- perature of the year is reached in February or the first part of March. After the middle of March the water temperature begins to rise very slowly and reaches 5.0° C. by the middle of April. CONDITION IN SPRING AND SUMMER By the time the water temperature reaches 7.0-8.0° C., which usu- ally takes place at the end of April, some minor changes in the gonads of young animals begin to occur. They are manifested by resumption of spermatogenesis in some of the individuals. The process, at that time, is very slow, and newly-formed cells, mostly spermatocytes of the first and second order, are few in number. The majority of animals still have gonads in the state resembling that of the hibernation period. As soon as the water temperature advances to a 10° C. mark, usually by the middle of May, more pronounced changes take place in the gonads. Active spermatogenesis is resumed in the follicles of many animals. Numerous gonia, primary and secondary spermatocytes are formed but few spermatids are seen. In many follicles spermatozoa retained since the preceding autumn are present. At this time a slight extension of the follicle is already noted. Phagocy tic-nutritive cells begin to surround the follicles in which gametogenic activities are going on. Gonads con- tinue to retain their bisexual nature because many small ovocytes meas- uring 20—30 microns can always be found in all follicles, even those showing a distinct preponderance of spermatogenic cells. Young ovo- cytes are, as a rule, found lying in contact with the follicular wall. Their large nuclei make them easily distinguishable from other cells. Phagocytic-nutritive cells are often found in large numbers along the outer walls of the gonad follicles and a few penetrate into their lumens (Fig. 7). These cells, measuring up to about 12-14 microns and having a nucleus 4—4.5 microns in diameter, occur as a rule near the follicular wall. Their numbers appear to vary with the seasons, the greatest number being found during the active stages of gametogenesis. They are found during all stages of development from indifferent gonad to mature stages. Often some of them project through the follicular walls into the lumen, while others invade it. In clams such cells appear to perform functions of twro types. First, their intimate contact with the gonad during the most active stages of gametogenesis indicates that they contribute certain substances necessary for the developing cells. Thus, their nourishing function may be assumed. Second, their pres- ;-i- VICTOR L. LOOSANOFF ence inside the gonad follicles immediately after spawning suggests their purely phagocytic role of removing partially cytolysed, degenerated and residual cells. As the season advances and the water temperature gradually in- creases, reaching about 15.0° C. by the end of June, rapid prolifera- tion of sex cells progresses in a parallel manner. Gonads of young animals examined at that time of the year show that spermatogenesis proceeds very actively and that spermatogenetic cells in all stages of development are rilling the follicles, often occupying the entire space of the lumen. It often happens that various follicles of the same animal nc. FIG. 7. Gonad of the young clam (1.0 cm.) Mirrounded by the nutritive pliagocytic cells (n.c.). Other letters as in Fig. 6. Late April-May. show quite different stages of development. Frequently one of the fol- licles possesses a large number of spermatozoa, while another follicle, situated next to it, has most of its cells in early stages of spermatogene- sis. Nutritive cells are quite numerous, sometimes entirely surrounding the follicle (Fig. 7). The clams are feeding and growing rapidly, many of them showing an increase in total length of shell of 3 mm. since the end of the hibernation period. On examining the clams passing through active gametogenesis, one's attention is immediately drawn to the large numbers of nutritive cells surrounding the outer walls of the follicles. In the follicles themselves PRIMARY GONAD AND SEXUAL PHASES IN VENUS 399 great activities are noticed. They are quite distended and filled with spermatogenetic cells. At this time mature spermatozoa are already more numerous than cells of other stages. Spermatocytes in all stages of development and spermatids in spermioteleosis are found. Regard- less of such spermatogenetic activities and the distinctly male character of gonads, young ovocytes are always found in small numbers along the wall of the follicles. There is a rapid extension of the branching sys- tem of follicles, but this extension is directed chiefly in the posterior di- rection and not ventrally ; consequently very few follicles extend below the level of the stomach. As has been mentioned before, some of the animals examined dur- ing the winter had follicles virtually devoid of all except indifferent cells. Such animals apparently discharged their spermatozoa in the fall. In June and July those animals also exhibit gametogenetic activities and spermatozoa are rapidly formed. It is very interesting to note, how- ever, that in such cases the follicles usually contain more female cells than the follicles of the animals which retained spermatozoa throughout the winter. The variation in the proportion of cells characteristic of the opposite sexes found in clams of the same age strongly indicates that genetic factors and probably the effect of environment influence the production of sex cells. SPAWNING In August the water over the natural clam beds reaches a tempera- ture high enough to induce the spawning of clams. If young clams, which by this time have reached the size of 2.0-3.0 cm. are examined, their gonads reveal the fact that the spawning process is not a phenome- non of short duration but lasts for some time. Different follicles of the gonad of the same animal contain sex cells in various quantities. Some follicles have already discharged their contents while in other follicles spermatogenesis is in progress and the lumen is still filled with mature spermatozoa. Apparently, several days, or perhaps weeks, are required for young clams to complete their spawning. Furthermore, by exam- ining a sample consisting of many animals, the conclusion can be reached that there is considerable difference in the spawning behavior of indi- viduals because some of the animals collected at the same time and from the same place have their sex products completely discharged, while others still retain their spawn in various quantities. This indicates that the entire population of a certain bed does not begin spawning at exactly the same time, and that the spawning season of clams extends for a considerable period of time. In many follicles, which during the spawning period are partly or 400 VICTOR L. LOOSANOFF fully freed of spermatogcnic cells, young ovocytes, many of them in the spireme stage, are growing along the walls. In some instances groups of ovocytes occupy considerable portions of follicles. A few nutritive cells may be observed in the vicinity of follicles whose contents are al- ready discharged. Branching gonad tubules begin at this time to extend ventrally, occupying the space between the body walls and the digestive gland. During the months of August and September the animals grow very rapidly, reaching the size of 2.5 to 3.2 cm. Many animals have grown as much as 2.0 cm. since the end of the hibernation period. Simul- taneously with the growth of the animals, ramification of gonad follicles proceeds. The follicles spread in all directions and envelop the stomach and intestines as well as the spaces in connective tissue between the body wall and the digestive gland. TRANSFORMATION TO DEFINITIVE MALES AND FEMALES In September, after the spermatozoa have been discharged, two types of individuals become distinguishable as definitive males and females. In the males a second or third period of spermatogenesis begins in the autumn and continues at a reduced rate throughout the winter. Mature spermatozoa are retained in the follicles all this time. Many young ovo- cytes and numerous indifferent cells are also present. In the spring, with the increase of water temperature, renewed rapid branching of follicles takes place and the gonads then begin to acquire the typical male character of the adult. Spawning follows later in the summer. In animals destined to become females radical changes take place in the gonads upon the completion of the initial male phase. After the dis- charge of spermatozoa the follicles are left in a distended state. Their lumens are large and empty (Fig. 8) with the exception of a few pyc- notic bodies consisting of cytolysed spermatogenic cells. An irregular layer of indifferent cells, ovogonia and primitive ovorytes measuring up to 20-25 microns, and a few spcrmatogonia remain for a time along the walls of the follicles. Nutritive cells lie in close contact with the gonad walls and occasionally can be found inside the lumens of the follicles. No significant changes occur in the gonads of future females during the hibernation period, for during rill this time the follicles remain distended, round or oval in outline, with virtuallv emptv lumens. There is little indication of ovogenetic activities. In the middle of May, regardless of the fact that the water temperature is 9.7° C. (i.e., about four degrees above hibernation mark), and that the animaN already had been feeding for approximately one month, their gonads remain in a semidormant PRIMARY GONAD AND SEXUAL PHASES IN VENUS 401 FIG. 8. The gonad of the animal destined to become a female after the com- pletion of the initial male phase. /, indifferent cells ; ov., ovogonia ; oc., young ovocytes ; pyc., pycnotic cells ; f.c., follicle cells. FIG. 9. Early ovogenetic activities in the young female clam producing its first crop of eggs ; ov., ovogonia ; oc., ovocytes in various stages of development ; f.c., follicle cells. Age 18-19 months. Middle of May. 402 VICTOR L. LOOSANOFF state. Only in rare instances are synaptic activities in young ovocytes noticed. In some follicles, however, pycnosis and phagocytosis of sper- matogenetic cells left over from the last autumn proceed very rapidly. In clams producing their first crop of ova, slow ovogenetic activities become apparent in the middle of May and active- ovogenesis begins in June when the water temperature approaches 15° C. Within a few days after such a temperature has been reached the appearance of the gonads undergoes a marked change. Young ova in various stages of development, some of them 42-45 microns in size, grow from the walls of the follicles into the lumen (Fig. 9). The growth and proliferation of ova are very rapid and the animals formerly functioning as males have now reached the stage of functional females. Not all the follicles begin to produce ova at the same time. Frequently, in the same animal some of the follicles contain large, rapidly growing ova, while a few others are still in an apparently dormant state showing many indifferent cells along their walls, ovogonia and very minute ovocytes. The first crop of eggs produced by the young female is usually small and in this respect they can always be distinguished from fully adult individuals, because in the latter case the follicles contain many more ova. After spawning, the gonads of both sexes may contain some cells characteristic of the opposite sex. thus retaining their bisexual character. It is especially noticeable in the case of males, where- comparatively large ovocytes are easily distinguished from male cells. FUNCTIONAL HERMAI-HRODITISM I 'elseneer (1926) has shown that hermaphroditism is quite common among mollusks. Among IVlecypoda numerous cases of partial and true hermaphroditism have been reported. As has been stated in the chapter on the development of the primary gonads of }'. incrccnarla, many grades of bisexuality occur in that animal during the first two years of its life. The examination of the gonads of adult males also reveals, in almost every instance, the presence of small ovocytes .sonu-wherc along the walls of the follicles. In all such eases, since the cells of both sexes are present in the gonads of the same individual, the term "' partial hermaphroditism" may be applied. However, under functional or true ];K.. 10. Functional hermaphroditism in Venus. Portions of two adjacent follicles of the same individual containing cells of opposite sexes, spc.l. primary sperm; i' ; sfc.2. secondary spermatocytes ; spt., spermatids; spz., spermato- /»;i', oc., laii'e ii\ncytcs showing mitochondria! bodies and yolk nuclei. Fi<;. 11. I'ortidii of hermaphroditic gonad showing the ripe ova surrounded l>y spermatogenetic cells. s />permatogonia are much less numerous than in the animals examined in October and the early part of November. Many spermatocyto are seen in the synaptic stage or in the process of division. Practically the same condition exists in De- cember. In all cases a large number of pycnotic cells are found scat- tered among the healthy ones and the phagocytic cells are actively en- gaged in clearing them from the follicles. The middle part of December may lie considered as the- threshold of the winter hibernating period of V . iiicrccnaria of Long Island Sound (Loosanoff, 1936a). The- water temperature by that time decreases to a critical point of about 5.0° C. and the animal, confronted by an un- favorable environment, undergoes important physiological modifications which induce hibernation. But if the gonad sections prepared from ani- mals collected during the early part of the hibernation period are com- pared with those collected one or two months previous to its onset, little difference can be noticed in their character. Throughout January and February spermatogenetic activities are continued at a reduced rate. Animals collected in February possess large sex -lands, enveloping the stomach, liver, intestine, and penetrating into the foot. The fol- licles are fully extended and filled with spermatogenetic cells of which spermatozoa are the mo>t numerous. Kxcluding a brief post-spawning period, the spermaries of adult males of / '. /;/<•/•<•<•//). The question of physiological ripeness of spermatozoa found in clams at all other times, except pre- spawning and spawning periods, still remains open because no eggs ca- pable of fertilization could be obtained for such an experiment in late fall, winter and spring. The month of March may be regarded as the time of the year when the spermatogenetic activities of the clam are the least conspicuous. The spermaries of animals collected during that month contain mainly sper- matogonia and spermatozoa, while the cells of the intermediate stages are present only in small numbers. In the middle of April the water temperature passes above the critical hibernation point for clams. Nevertheless, as far as production of gametes is concerned, the animals do not respond noticeably to the favorable change in environment. The gonads remain quiescent, showing little spermatogenetic activity until the middle of May, when the water temperature begins to fluctuate be- tween 10.0° and 13.5° C. As soon as the temperature reaches 15.0° C., which in Long Island Sound takes place early in June, vigorous sperma- togenesis is again initiated and production of spermatozoa proceeds at an extremely rapid rate. The presence of a large number of spermatids in the follicles is very characteristic of this part of the year. Simul- taneously with increase in wrater temperature further increase in activi- ties of spermaries is noted. In June, July and during the first part of August many animals are found with ripe gonads. In several cases it was possible to induce such animals to spawn by a gradual increase of the water temperature. Among animals living in very shallow water or on tidal flats, individuals with partly discharged gonads may be found early in July, but the largest part of the clam population spawns in Au- gust when the curve of water temperature reaches its culmination point. FEMALES During the immediate post-spawning period, after the gametes have been discharged, the sexes of J7. mercenaria can still be distinguished easily. A description of the post-spawning condition observed in the clam spermaries has already been given. The ovaries in the post- spawning stage are characterized by the presence of a few unspawned, but apparently ripe, ova and by a large number of ovogonia and very young ovocytes. The presence and normal appearance of these ovocytes during the stage when sex reversal may occur in mollusks, points to- ward the conclusion that in V . mercenaria the change of sex from female to male seldom takes place. 410 VICTOR L. LOOSAXOFF J, •> i **v- i f *y. * «C ^ ^M;? «*!<£ TII • ^3 ^ •' *a £s I;IL'. 1. l-'cin.-ilc yonad immediately ai'tcr parti;il s|ia\\ nin.u. I'nllii-les with un- >l>a\viu-er. 'I'he ovarian t'nllielc-, arc siill in a con- tracted state. Manx \nnn.u OVOCytCS air growing. A few nn-])a\vncil ova still remain. X .inn. l-'h.. ,v l-'i-malc' .m mad in October. X 300. l;n;. -4. l-'i-malr .umiad in N'i ivcmher, showing the rajiid growth of young ovo- es. > SEASONAL GONADAL CM \\CKS OF ADULT CLAMS 411 The number of large unspu\vned ova found in the ovaries of recently spawned clams varies greatly with individuals. In some cases almost the entire crop of ripe ova is discharged by the time spawning activities end, while in the majority of cases, a comparatively large number is re- tained. The difference between the follicles of the same animal, as to the number of unspawned ova left, is often very marked. Almost im- mediately after spawning the ovarian follicles begin to contract. This contraction does not proceed very far, however, and they are left in a semicontracted state with unspawned ova still retained in the lumen (Fig. 1). An almost continuous layer of ovogonia, indifferent cells, and young ovocytes in various stages of development and growth is often distinguished along the inner walls of the ovarian follicles. The spawning of an individual clam is not completed in one attempt but is extended for a certain period of time, depending upon the indi- vidual and ecological peculiarities. Female clams with partly discharged sexual products were found on our experimental bed throughout August and September. Their ovaries contained, in various proportions, the cells of old and new crops. Towards the end of September only a few- ripe ova were found. Apparently a delayed but nevertheless normal discharge of eggs takes place this late in the season. The discharge of the ripe eggs evidently removes the factors inhibit- ing the production and growth of a new crop of ovocytes. As soon as the ovaries are freed of the bulk of their ripe eggs, ovogenetic activities are commenced. At first a few young ovocytes are seen growing and protruding towards the ovarian cavity (Fig. 2). The new ovocytes at this time vary markedly in size. The average size is approximately 17-20 microns but some of the youngest ones measure only 8-9 microns, and the largest, which are less numerous than others, reach the size of 35-40 microns. In almost every case studied the ovarian follicles were found to contain a few very large ovocytes measuring 55-60 microns. Such ovocytes, many of which were still connected by their egg stalks to the ovarian walls, represented the cells of the old crop. Early in October a considerable increase in the number of new ovo- cytes is observed. Many of these ovocytes are still of a rather small size as compared with fully grown eggs, a few of which are left in the ovaries (Fig. 3). At this time the average size of a newr crop of ovo- cytes is about twenty-five microns. It should be remembered, however, that individual differences are great. It is often found that in a sample composed of a large number of animals a few individuals may have their gonads in either a greatly retarded or an advanced stage, as com- pared with the conditions observed in others of the same samples. This 412 VICTOR L. LOOSANOFF is sometimes very confusing and unless the sample studied is composed of a large number of animals, erroneous conclusions may be reached. During the latter part of October and in November the growth of young ovocytes is very rapid. Some cells of the- new crop have already reached the size of a mature egg but most of the cells are between 33 and 55 microns in size. There are still many minute ovocytes making their appearance but their number is considerably smaller than it was in the early part of October (Fig. 4). Simultaneously with the growth of the ovocytes, the follicles expand and often proliferation and formation of new follicles in the surrounding connective' tissue can be observed. The ynung sex cells of >uch follicles are much smaller in size than those of the old ones. In X<>\ ember the discharge and absorption of all unspent eggs are quite completed. In only exceptional cases can such cells lie found at that time. Orton (1933; found that in most cases the unspent eggs of 0. cdulis are extruded from the gonad within a few days after the principal >pawning, while a small percentage of individuals may retain the re- maining ova for as long as two months and then discharge them in the usual way. In the case of /". iiicrccnariu the fate of some unspawned ova rs probably similar to the fate of those in O. cdulis. Studies of the gonads of V . incrccnaria in post-spawning stages reveal that the number of undischarged eggs in the ovaries gradually diminishes as the season pr< s. Some of the unspent ova found in the ovaries of clams during September— November are normal in appearance, having no indi- cation of any decomposing processes g"ing on. The fact that few if any phagocytes are observed in lumens of the- ovarian follicles during the post-spawning period al>o adds weight to the conclusion that some of the ova unspent during the spawning season are extruded later in the normal manner. Such extrusion is completed in November. Regardless of the fact that the temperature of the water over the clam bed decrea.se> very rapidly during November, and in December falls below the point at which hibernation of clams begins, the growth and development of gametes pn.eeeds normally. Ovaries of clams col- lected late in I )ec< inber already contain large numbers of ova which have a virtually mature appearance ( |-'ig. 5). During that part of the sea- son ovarian follicles contain relatively but very few of the minute ovo- cytes which were so characteristic for the gmiads of early fall. It is quite clear that the young ovocytes which appeared in the follicles after completion of spawning have grown by now \<> a large si/.e. and that since the appearance of the first and numerous group of ovocytes in September and < >rlXAD \l. CIIANU-.S ( i|- \IH I.T ( I. \MS 413 protrude into the- lumen occupying much of the available space in the follicles. A few pycnotic cells, some of them ovocytes in very early stages of growth, may often be noticed. It is very strange that during this time as well as the entire post-spawning period, when the ovogenetic activities of clams are at their height, very few nutritive cells are present near the follicles containing growing ovocytes. In the development of FIG. 5. Female gonad in December, containing large number of ova of mature appearance. X 300. FIG. 6. Female gonad in May, containing ripe ova. X 300. male gametes the role of such cells is of considerable importance since it is apparent that they participate in providing nutrition for the developing sex cells. In female clams ovogonia and ovocytes continue their growth mainly by extracting nourishment from the mesenchyme and blood ves- sels between the follicles. The nutritive cells, so important in male gonads, are much less numerous and apparently of lesser significance in the metabolic cvcle of the ovarv. 414 VICTOR L. LOOSANOFF During the winter months, January-March, few changes occur in clam ovaries. The slow growth of young ovocytes continues but at a much reduced rate. There are few if any new young ovocytes begin- ning to develop from ovogonia during this time ot the year. Large ovocytes, constituting the majority of the cells of the ovaries, measure up to 66-68 microns, i.e., equaling the size of the mature cells. Occa- sionally pycnotic cells may be noticed in one of the follicles, but as a rule there is little evidence of pycnosis. I hiring April and early May the growth of the undersized ovocyte-, continues. 15 y the middle of May the ovaries acquire a fully ripe appearance with approximately of normal oval shape. During the second part of Mav some of the eggs of the clam are mature not only morphologically ( Kig. (>\ but also physiologically. On May 25. 1'J.vl. four large female clams were taken from the bed and brought to tin- laboratory where they were placed in an aquarium. The temperature of the water over the beds was 14.0° C. and that of the aquarium water 26.0° C. After two hours of exposure to such a tem- perature two femaUvs began spawning. Discharged eggs were fertilized by the addition of sperm suspension and they began to develop in the normal way. Thus, during the end of May. June and July, sexually mature and ripe clams merely wait for the water temperature to reach the point at which the proper .stimuli inducing spawning reaction are produced. The general spawning of the clam population of the Charles Island beds occurs most frequently either late in July or early in August, and continues into September. DISCUSSION Studies of the changes occurring in the spcrmarics of adult F. incr- ccnaria reveal that the post-spawning period, when the gonads an- devoid of all but undifferentiated cells, is o! verv brie! duration. This is prob- ably the only period when the change of sex from male to female can take place. Production of a new crop of .spermatozoa begins very soon after the completion of spawning. The main portion of spermatozoa for the next vcar's crop i> produced during the autumn. Mature sper- matozoa in small numbers may be found in the spermarics of a few clams as earlv as the end of September. In \ovember the spermatozoa are alreadv the most numerous cells m the gonad lolbcles. which at this time acquire an appearance of ripeness. Spermatogenesis continues at a reduced rate' throughout the winter but is practically at a standstill during the earlv spring. The second manifestation of very rapid SEASONAL GONADAL CHANGES OF ADULT CLAMS 415 spermatogenetic activities takes place in June, \vlicn the water tempera- ture reaches 15.0° C. The discharge of spermatozoa follows when the critical spawning temperature of 23-25° C. is reached. Summarizing the observations of the processes occurring in the ovaries of V. mercenaries, it may be stated that the production of a new crop of eggs begins immediately after spawning. The chief growing period of the young ovocytes is in the autumn. In December and Janu- ary the gonads of many female clams already present a ripe appearance. During the winter and spring only a few new ovocytes begin their de- velopment ; cells produced in the autumn constitute the greatest portion of the next year's crop. Some ova, undischarged during the spawning season proper, may remain in the ovaries for a considerable length of time, and are finally discharged in the normal way. Many of the ova ripen approximately two and a half months before the water of the clam beds reaches the critical spawning temperature. Compared with the sexual cycles of other pelecypods, it is rather unusual that in V . mcrccnaria the development of gametes occurs during the season when the water temperature rapidly decreases and the eco- logical conditions for gonad development appear to be less favorable than in the spring and early summer. According to Coe (1932), low temperature during the winter may delay the completion of the sexual phase in 0. lurida for several months. A phase may be inaugurated in October and completed in April, while in the summer the same phase might be completed in a few weeks. The same author states that in the autumn ovulation in 0. lurida is often inhibited. Judging by this it is probable that after the completion of the last phase of the year the ani- mal does not display very active gametogenesis in the following autumn and winter, and that the main production of the gametes of the next sexual phase takes place in the spring when the water temperature be- gins to increase. In O. gigos (Amemiya, 1928) the production of gametes occurs chiefly in the spring. According to observations of the writer, adult individuals of 0. virginica of Long Island Sound enter into the resting stage soon after completion of spawning and remain sexually inactive until the next spring. In O. connnercialis (Roughley, 1933) the gonads are found in the resting stage throughout the winter and gametogenesis commences in the spring. Probably the same condition will be found in many other mollusks. In V. mercenaries, on the other hand, the most active gametogenesis and the production of the year's crop of sex cells occurs in the autumn and early winter. Since ovo- genesis continues throughout fall and winter there is no actual period of recuperation which is commonly observed in other allied forms 416 VICTOR L. LOOSANOFF living in approximately the same geographical areas and subjected to similar ecological conditions. Retention of morphologically ripe sperm and ova throughout the greatest part of the year is another interesting feature of the sexual behavior of }'. incrccnarla. BIBLIOGRAPHY AMEMIYA, I., 1928. A preliminary note on the sexuality of a dioecious oyster (O. gigas Thunbcrg). Jap. Jour. ZooL. 2: 99. AMEMIYA, I., 1929. On the sex-change of the Japanese common oyster, Ostrea gigas Thunhcrg. Proc., Imp. Acad., 5: 2S4. COE, WESLEY R., 1932. Development of the gonud.s and the sequence of the sexual phases in the California oyster (Ostrea lurida). Hull. Scripts Just. Oceanoyr., Tech. Scries, 3: 119. LOOSAXOFF, VICTOR L., 1936a. Temperature and hibernation of hard shell clam (Venus merccnaria). Fisheries Service Hull., No. 252, p. 4. LOOSANOFF. VICTOR L., 1936b. Sexual phases in the quohog. Science. 83: 287. NELSON, T. C, 1928. Relation of spawning of the oyster to temperature. Ecology, 9: 145. ORTON, J. H., 1933. Observations and experiments on sex-change in the European oyster (O. edulis). Part III. < 'n the fate of the unspanned ova. Part IV. On the change from male to female. Jour. Mar. ttiol. Assn.. 19: 1. ROUGHLEY, T. C., 1933. The life history of the Australian oyster (Ostrea com- mcrcialis). Proc. Linn. Soc. New South ll'alcs, 58: 279. THE RATE OF WATER PROPULSION BY THE CALIFORNIA MUSSEL DENIS L. FOX, H. U. SVERDRUP AND JOHN P. CUNNINGHAM * (From the Scripts Institution of Oceanography of the 'University of California, La Julia, ( '(///'/iiniKi) Numerous investigators have appfoached by means of various ex- perimental techniques the problem of the volume of water which may pass in a given time through the mantle cavity of lamellibranch mollusks. The establishment of an approximate, or even a minimum figure for the volume of water pumped per day by any of the many marine plankton- feeding animals would be of interest to physiologists who may be con- cerned with nutritional, respiratory, or excretory activities of the animal itself, to planktologists whose problems deal with the numerous factors which influence the numbers and distribution of microscopic plants and animals that may be consumed by the filtering animals, and to ocean- ographers who seek information regarding biological factors concerned in modifications of the physical and chemical character of both water and bottom in various regions of the sea. In many instances, those who have contributed to our knowledge of the water-filtering and feeding activities of lamellibranchs have been primarily interested in the sanitary and other technical aspects of cultivating oysters and mussels for human consumption. Thus Viallanes (1892) presented approximate figures for the relative rates of filtration of water by French oysters, Portuguese oysters, and mussels. He placed the animals in separate crystallizing dishes upon the bottom of a tank supplied with flowing sea water, keep- ing initially empty dishes of the same size and design alongside of the ex- perimental ones, in the same tank. At the end of a specified time inter- val the material which had been filtered from the water by the animals, and either swallowed and later voided as fecal material, or else rejected as pseudofeces directly from gill and mantle surfaces, was collected, dried, and weighed. The weight of filtered material was obtained by sub- tracting that of the detri-tus which had settled out of the water by gravity into the control dishes. Viallanes performed similar experiments more critically by dispersing a known quantity of dry clay in a given volume of water, drying and weighing the quantities of clay deposited by the different bivalves after 24 hours. From his first experiment, wherein 1 Chemist, Federal Works Progress Administration Project No. 691 ; Cali- fornia District No. 12. 417 418 FOX, SVERDRUP, AND CUNNINGHAM the suspended material naturally present in the sea water was the only available particulate matter, he calculated that his animals (of age = = 18 months) filtered water at the following relative rates : French oyster, 1.0; Portuguese oyster, 5.5; mussel, 3.0. The clay experiments, how- ever, showed very different results, which were left undiscussed. Table I shows briefly his results. Yiallanes found that the- mucus which was secreted by the mollusks and in which the precipitated clay was incorporated, weighed only 4 per cent of the weight of the whole mass. Ranson (1926) cites the work of Yiallanes and reports briefly his studies of the mechanism of filtration. Both investigators emphasize the importance of the filtering function of lamellibranch plankton- and detritus-feeders to ostreiculturists and to oceanographic science. TABLE I Filtration and deposition of suspended clay by Yiallanes' lamellibranchs. Animal Initial quantity of dry clay Weight of clay di-|H»ited in 24 hours Minimum volume of water filtered per hr. (our calculation) Mussel (18 months) Uriim/liter 0.0546 grams 1.768 liters 1 35 Portuguese oyster (18 months). . . French oyster (18 months) 0.0546 0.0546 1.075 0 199 0.82 0015 Criticisms of Yiallanes' experiments which must be borne in mind are: (\ ) in his first experiments with naturally-occurring particulate matter, the portion of the material which had been actually swallowed, the undigested residue of which was finally voided as fecal material, must have undergone changes of various kinds and degrees depending upon the relative rates and processes of digestion in the different species, and have subsequently vitiated the interpretations that were based merely upon relative weights; (2) in both of the experiments, the quantities of detritus or clay which had been removed from the water but still remained in the animals' alimentary tracts were not weighed or considered. Since in each case such small total weights were being measured, this item might have constituted a considerable, though by no means predictable or constant fraction of the total figure. Galtsoff (1926, 19280, 1928ft) measured the rate of flow of water through the gill chamber of the American oyster, and calculated the work done by the propelling cilia. He used two methods, one a direct one in which the water issuing from the exhalant chamber was collected in a measuring vessel, the other in which the rate of advancement of a RATE OF WATER PROPULSION BY MUSSEL 419 stream of water flowing from the exhalant chamber through a glass tube was measured by stop-watch. Both methods involved the inser- tion of a small glass rod between the opened valves to prevent their closing, and placing into the gill cavity a rubber tube, to carry off the discharged water, packing other open spaces around the tube with cotton. Galtsoff (I928b) records a maximum figure of about 3.9 liters per hour for the water intake of a single healthy adult oyster three to four inches in length at a temperature of 25° C. ; the temperature at which the high- est rate of flow occurred was found to be between 25° and 30° C. He points out, and shows by collected data the fact that considerable varia- tions exist in the rate of flow produced by individual oysters. His average temperature-flow rate curve, taken from data collected in the study of many individuals, shows, at temperatures close to 30° C., a maximal rate of about 2.4 liters per hour while at 20° C. the average rate lies close to 2 liters per hour. At temperatures between 24 and 27.9° C., filtration took place at the rate of from 2.5 to 2.9 liters per hour. He points out that the filtering action of the oyster is dependent upon two mutually independent functions, namely the beating of the cilia and the opening and closing of the valves of the shell, the process occurring only when the valves are apart and the cilia beating. His studies showed that the oyster keeps its shell open for an average time of about seventeen hours out of twenty-four. (June to October.) Numerous attempts have been made to estimate the rate of flow of water through the gill chambers of lamellibranchs on the basis of plank- ton counts in the water itself, and in the animals' stomachs. The results are, however, not satisfactory because, as Galtsoff (1928&) points out, there are daily and seasonal variations in plankton numbers in the water ; not all of the diatoms filtered out by the gills are ingested but may be rejected instead. Such experimental work should, however, if carefully controlled, afford minimum figures. It would seem for various reasons probable that most methods used to date might be expected to yield re- sults that are lower than the true figure for the volume of water filtered per day. Various investigators (see also Galtsoff, 1928b) have obtained widely different results for the rate of water-pumping by plankton feeders. Collateral data regarding temperature are, however, not always avail- able. Briefly, the general conclusions may be 'recorded as shown in Table II. The inconsistencies which appear from an examination of these figures serve to emphasize the wide variations according to the method used, and the unreliability of the plankton count method. Galtsoff's studies convinced him that as many as 18.9 per cent of phytoplankton 420 FOX, SVERDRUP, AXD CUNNINGHAM such as the diatoms Chcctoccros and Rltizosolatia, and the dinoflagellates Peridinium occanicitiu and Ccnitinin might escape being caught while passing through the gills, and that from 50 to 89 per cent of bacteria present in the water passed the gills <>f his oysters. Work at this Insti- tution, however, shows that the mussel is quite success ful in removing bacteria from water. Thus ZoBell and Landon (1937) demonstrated that the mussel removed about 99.9 per cent of added bacteria from sea water. (See also Fox et al., 1936.) Without doubt, GaltsofFs direct methods are the most accurate for TAIM.K II Filtration of water by plankton-feeders. Investigator Animal Method Avcniuc rate of filtration (liters per hour per animal) Grave i!905)... Oyster Plankton counts in water 0.167 and stomachs Moore (1913). Oyster Plankton count- in \\.itn ca. 1.25 and stomachs Allen (1914). Freshwater Rubber tube parked into 1.4 mussel exhalant chamber- Wells (1916). . . Oyster- Plankton counts in water 7.5 and .-tomach- Nelson (1921). . Oyster Rubber apron dividing in- 5.7 (but see text below) halant and ex ha la nt chambers Galtsoff (1926, 1928a, 4) Oyster Rubber tube packed into 2.5-2.9 at 24-26.9° C. exhalant clramber Dodgson (1928) Mussel ( l< .iriiri; of sir-pensions 2.0 (minimum, at 17° C.) (Mytilus edulis) Parker (1914). Sponge • ilas> tube tied into oscu- 3.2 (Spino- lum • sella) I)amas (1935).. Cardium Clearing of mud suspen- 0.1 sions ^ the measurement of the rate of flow of water through the mantle cavity under the conditions imposed upon the oyster. Xeedless to say, a question arises in one's mind regarding the possible influences that block- ing the valves apart with a glass rod, inserting a rubber tube into the ex- current chamber, and packing the openings in the mantle surrounding the inserted tube with wads of cotton might exert upon the normal feed- ing and filtering behavior of the animal. Nelson (1935) refers to earlier efforts to measure the water filtered by an oyster, and to the fact that very diverse results wen- obtained, lie wril ; The introduction of a tube into the cloacal chamber inter- RATE OF WATER PROPULSION BY MUSSEL 421 feres with normal operation of the branchial hearts described by Hop- kins and may disturb the visceral ganglion." He adds, " Also, in Ostrea virginica much of the water from the right demibranch leaves by an asymmetric chamber on the right side and separate from the cloacal chamber." Using a modification of the rubber apron of Moore (1908), he claims to have measured all the water passed by an oyster, without interfering with its normal activities. He reports the amazing value of 26 liters of water per hour, passed by an oyster 11.5 X 8.9 cm. in size, at the optimum temperature of 30° C. This is about tenfold the average values found by Galtsoff for the oyster and by ourselves for the mussel. Fresh oyster sperm were found by Nelson to increase markedly the rate of water propulsion by male oysters, while in females no response was observed unless spawning occurred, in which case the filtration-rate was temporarily reduced. Parker (1914) reports on the strength and volume of water currents produced by sponges. He measured the average height to which the excurrent stream of water might reach when glass tubes of appropriate size were tied securely into an osculum. By measuring the rate at which carmine particles, etc., were carried out of a glass tube of known dimen- sions, Parker concluded that the sponge Spinosclla discharged water from its oscula at a rate of about 4.5 cc. in five seconds, or about 78 liters per day. A colony of Spinosclla having as many as twenty oscula might, he concluded, strain in a day more than 415 gallons of water. Damas (1935) studied the activities of plankton-feeders, especially Cardium and other lamellibranch mollusks, with reference to their role in the deposition of marine muds. He calculated, on the basis of quanti- tative observations of the extraordinary rate of deposition of mud pel- lets by Cardium, that 1,000 such individuals produce, on an average, a layer of mud of 0.45 meter in thickness per square meter per year, or 1,250,000 cubic meters of mud per year in the 250 hectares (1 hectare - 10,000 sq. m. ; 2.471 acres) colonized by Cardium in the roadstead of Zeebrugge. The question of the natural filtration rate has been approached by Dodgson (1928) and by ourselves under conditions which would seem to simulate more closely those of nature. Dodgson and co-workers, working with the bay mussel, Mytllns cdulls, at Conway, Wales, pre- pared turbid suspensions of different substances such as flour, clay, fine silt, and even ordinary muddy river water, in sea water. On the basis of many experiments, Dodgson claims that " end-points," i.e. the time at which the formerly cloudy solutions become quite limpid, could be deter- mined without difficulty, the last traces of turbidity disappearing almost suddenly. At 17° C., mussels placed in turbid suspensions of fine silt, 42 FOX, SVERDRUP, AND CUNNINGHAM mud, clay, or flour filtered the water to clarity at estimated minimum rates varying between 1.9 and 2.6 liters per hour per mussel. His av- erage figure of about 2 liters per hour per mussel is expressed as being probably far less than the actual quantity, since, in order to remove quan- titatively a suspended substance from a given mass of water, the animal must, because of the constant mixing of the filtered with the unfiltered water, pass some of the water through its mantle cavity many times over, even should the water issuing from the cxhalant siphon be com- pletely cleared. EXPERIMENTAL In our work with the California sea mussel, Mytilits califoniianus, which unlike the At. cdidis that inhabits bays, estuaries and river mouths, attaches itself to rocks and pilings near open, unprotected shores, we used some refinements of Dodgson's general method. We considered that serial measurements of a finely divided substance remaining sus- pended in a given mass of water in which mussels were immersed should provide approximate data regarding the rate of filtration of the water by the animals. We have not overlooked the fact that individual animals may vary considerably in their relative rates of propelling water through their gill chambers ; we have also considered the possibility that the rate of filter- ing by the animal may depend to a considerable extent upon such factors as (a) particle size, (b) concentration and (c) chemical nature of the chosen suspended material (i.e. whether of nutritional, inert, or injurious character), and the possible influence of such properties upon mucus secretion, ciliary motion, and frequency of closure of the valves. For convenience in analyzing at intervals remaining suspended mate- rial, and in order to duplicate to some extent the conditions of nature, use was made of a calcareous marine- mud, whitish to light grey in color, from Bird Key Harbor, Florida. In preliminary experiments ~ use was made of material which, according to Dr. K. M. Thorp, had passed through a 0.48-mm. mesh screen. Its introduction into water in which mussels were immersed had no perceptible influence upon the animals, which remained with valves apart and continued to filter water. Mus- sels when kept out of water for a short interval, then placed in a very turbid suspension of the mud, opened their valves without delay, and began "pumping" water. Figures 1-4 show a series of photographs illustrative of the rather striking clarification of turbid water in a rela- tively short time. Into each of two graduated cylinders, each containing - These were undertaken in the early summer of 1935 by the senior author in collaboration with Dr. Roderick Craig of the Division of Entomology and Para- sitology, University of California. RATE OF WATER PROPULSION I'-V MUSSEL 423 Fie. 1. FIG. 2. FIG. 3. FIG. 4. FIGS. 1-4. Removal of mud from suspension by a mussel. (Photographs by Dr. R. Craig.) 424 FOX. SYKRDRUP. AND CUXXIXCHAM 2.000 cc. of sea water, were placed 20 grams of the calcareous mud which was mixed with the water by several inversions of the temporarily stoppered containers. After the coarsest particles had settled out within a few minutes, the containers were placed in a window to afford good lighting, and a large mussel of 17 cm. length, and weighing about 485 grams, was placed in the one on the left side. The first photograph shows an identical degree of turbidity in the two suspensions at the be- ginning, as noted in the faint Tyndall beam from the edges of the cards strapped at the rear outer wall of each cylinder. In half an hour ( Fig. 2). the turbidity in the mussel-containing jar had been greatly decreased. and the picture shows the printed letters upon the card, easilv legible through the water. In 50 minutes (Fig. 3 I the turbidity in the left con- tainer was all but gone, and in the last of the series ( 1 hour and 45 minutes) we see a perfectly clear solution. Xote particularly the serial appearances of the string by means of which the animal had been low- ered slowly into the container. The temperature of the- collected sea water was initially about 20° to 21 c C '.. and could have increased by not more than a degree or two. Inspection of the sediment at the bottom of containers on the day following a filtration experiment revealed three types of precipitated mud: (1) very fine. homogeneous powder which had titled out by [ .ivity; (2) pile- of pseudofeces in the form of amorphous, stringy masses of material which had been filtered out by the mu»el. incorpo- rated in mucus, and expectorated from the mantle at the excurrent siphon: the extrusion of this material could be observed continuously from the beginning of the experiment: (3i true characteristic feces in the form of short, discrete. Hattish straps composed of the mud which had been ingested by the mussel. The intricate arrangements possessed bv lamellibranchs for filtering the fine detritus and small organisms from water have been described and discussed extensively by other authors ( Moore. l('i)5: Allen, 1914: Kellogg, 1915: Dodgson. 1928; Yoiige. 1'LN; and others cited especially by the latter). \Yc know from the experiments reported in this paper and from previous ones (Fox et al.. 1936) that the material filtered by the mussel from water may be in part swallowed (even if inert and nutritionally useless) and in part expectorated in mucus-laden strings ur pseudofeces from the edges of the mantle, especially if the water contains a great amount of suspended matter. \Ye \\ere not in this particular study interested primarily in the relative quantities that were swallowed or rejected; we were interested in the rate of diminution of suspended material, and made preliminary measurements" in order to learn (1) how :; In tin- l.itr summer of 1935 by the senior author jointly with Mr. Rae Sduvenck. from the Department of Chemistry, Sacramento Junior College. RATE OK WATER PROPULSION BY MUSSEL 425 rapidly water may be propelled through the gill chamber of the mussel during such metabolic activities as feeding, respiration and excretion, and (2) how rapidly suspended organisms or other material may be re- moved from water, to be ultimately deposited in an altered state upon the bottom of the sea. The results of these preliminary experiments were, however, not en- tirely satisfactory. Their difficulty of interpretation was doubtless due to the fact that we were at that time unaware of all the precautions which had to be taken in order to obtain results which could be readily analyzed. The experiments were therefore repeated in the fall of 1936 (D. L. F. and J. P. C.), with the introduction of several refinements in tech- nique and materials ; these experiments will here be dealt with more fully. In the first of them the same calcareous mud was employed, but of a much finer grade, having passed (according to Dr. Thorp) through a 0.086 mm. mesh screen. The dry mud was shown by analysis to con- tain 1.66 per cent moisture, 90 per cent CaCO3, and 6 per cent SiO2 (by difference). In later experiments, we used pure CaCO3 instead, and found that it served equally well. (See below.) In our experiments 8,000 cc. of fresh sea water, to which were added initially 32 grams of mud (4 grams per liter of water), were placed in each of a series of large battery jars. The suspensions were allowed to stand for about one hour, to allow time for larger aggregates of mud particles to settle to the bottom. The mussels were handled carefully throughout the work ; they were placed on the table for a time with the valve openings vertical to the surface, in order to allow the water within the mantle cavity to drain out when the animals opened their shells ; this not only prevented changes in the volume of water introduced into the containers, but prob- ably rendered the animals sufficiently " thirsty " to insure their com- mencing activities almost directly they were immersed in the suspen- sions ; the prompt opening of their valves after immersion could be recognized by bubbles of displaced air rising to the surface, and often also by watching the animals which lay close to the glass walls of the jars. A moderate stream of air was introduced through a glass tube reaching to the bottom of each jar, sufficient to insure constant homog- enous mixing without disturbing the animals or stirring the heavier material which was deposited on the bottom by ordinary settling or by the animals. Although the animals kept their valves apart during the course of the experiments, it was observed, by following the course of suspended particles, that the currents of water flowing into the incurrent and out of the excurrent siphon were not of continuous intensity, but occurred 426 FOX, SVERDRUP, AND CUNNINGHAM at apparently rhythmical intervals, the intensity of the stream alter- nately increasing to a maximum, then gradually diminishing to a mini- 2O 40 6O 80 100 Time in Mm u 140 160 .0 - 3.0 - 2.0- 4.0- 2.O FIG. 5. Removal of suspended CaCO3 by mussels ; first series. First Scries: Large mussels — _' animals. Lengths 130 mm. and 110 mm. (average = 120 mm.). Total wri.uht 333 grams (average == 166.5 grams) each. Medium mussels — (M,) 4 animals. Lengths 97, 100, 100 and 103 mm. (av- erage = Inn mm. i. Total \\eiuht 302.5 grams (average = 75.6 grams) ch. (M,.) 4 animals. I. m^ths 95, 95, 105 and 105 mm. (aver- - 100 mm.). Total weight ~>(>. 4 grams (average = 74.8 grams) each. Small mussels— (S.) 9 animals. Li-ngths 70. 7S. <\7, 73, 76, 72, 75, 74 and 65 mm. ( average = 68.9 mm.). Total weight 308.4 Drains (average = 34.4 grams) each. (S-) 9 animals. Lengths 78, 69, 79, 75, 82, 79, 60, 62 and 82 mm. (average = 74 mm.). Total weight 360.0 grams (average = 40 grams) each. I « n-cc. samples taken every 15 minutes. Temperature range 22.85° to 23.4° C. mum. This could be observed particularly well in the large mussels. Mucu>-ladrii strands of the filtered mud were expelled at a slow but RATE OF WATER PROPULSION IJY MUSSEL 427 nearly constant rate from the edges of the mantle at the exhalant open- ing. These pseudofeees either fell rapidly to the bottom, or, if attached to an air bubble or two, rose to the top. They did not break down and become redispersed. Samples were removed at stated intervals (10 or 15 minutes) from a uniform place in the center of the container and at about the- mid-depth point, with calibrated 10-ml. pipettes, and introduced into stoppered ves- sels for analysis. These samples were acidified before washing them into containers for analysis, in order to insure that none of the colloidal CaCO.{ material remained adsorbed to the walls of the vessels. The microchemi.cal method of Kirk and Moberg (1933) for the analysis of 40 60 BO IOO I2O 140 160 Time m Minuter 4.0- FIG. 6. Removal of suspended CaCO3 by mussels ; second series. Second Scries: Six mussels used. Medium (M). Lengths 98, 108, 100, 103, 105, 107 mm. (average == 103.5 mm.). Total weight 611.5 grams (average == 102 grams). Ten-cc. samples taken. Temperature range 20.8° to 21.35° C. calcium in blood or sea water was employed. The calcium is precipi- tated as the oxalate, washed, and titrated in acid solution with standard potassium permanganate. By this method the amount of calcium is determined with an accuracy of ± 4 mg. /liter. The concentration of dissolved calcium in the ordinary sea water used in these experiments was determined by the same method. Controls (i.e. jars containing identical quantities of water and sus- pended mud, with similar moderate aeration but without mussels) were always carried out, parallel with the experiments, and samples were taken from these at the same intervals. Two series of experiments (Numbers 1 and 2) were undertaken using the calcareous marine mud ; the results are shown in Figs. 5 and 6. 428 FOX, SYERDRUP, AND CUNNINGHAM The data for these figures were collected and originally tabulated in exactly the way shown in Table III. Two series of experiments (Numbers 3 and 4) were also performed with the use of pure, finely divided CaCO:1 ; the results are recorded in Figs. 7 and 8 (from data tabulated as in Table III. which shows data from which Fig. 8 was derived). Freshly collected specimens were 4O 60 BO 100 Time in Minu tes 140 160 6.0- 4.0- 3.0 - 2.0- FIG. 7. Removal of su.si>rnded CaCO;! by mns-rK ; third scries. Third Scries: Very large mussels. E,. Two animals. Lengths 17'' ;m grams total) in each tank. Temperature range 18°-20° C. used in these experiments. Placed carefully in the containers, they soon opened their valves, and showed no objection whatever when the car- bonate was added, but continued to filter the water in a normal manner. In one experiment (see Fig. 7) two pairs of very large mussels, col- lected at an exceptionally low tide-, were employed. These individuals were selected from the catch on the basis of fair uniformity of si/.e and their readiness to open their valves and pmpd water without showing RATE OF WATER PROPULSION BY MUSSEL 429 cessation of activity or other disturbed responses to slight mechanical stimuli such as stirring the water in which they were immersed, or tapping upon the container. Four jars were set up, each containing 6 liters of sea water. Two selected mussels were placed in each of two jars while the other jars were left as controls. Air was passed at a uniform rate through each jar of water as before. After all four of the animals had parted their valves, the additional two liters of water were added to each jar (to make up the total of 8 liters), the suspended CaCO:! being added with the water in the operation. 20 40 60 140 160 -6.0 -5.0 -4.0 -3.0 -2 .0 Time in Minutes FIG. 8. The " zero points " of the two control jars (without mussels), which agree very closely, were obtained by taking samples immediately after adding the CaCO3 in the regular manner and are taken to serve as initial points for the duplicate experiments as well, since identical quantities of the salt had been added to the same volumes of wrater under the same conditions, and previous experiments had shown good checks in the initial quantities of suspended material under such conditions. Table III shows data relating to duplicate experiments on (1) medium-sized mussels (M4 and Mn) and (2) small mussels (S.f and S4) all run simultaneously along with a control (C). Four mussels were placed in each jar, and the experiments were conducted in a manner 430 FOX, SVERDRL'P, AND CUNNINGHAM identical to the previous set. Again, tin- initial value for suspended CaCCX, in the control jar was taken to represent a reliable figure for that of the experimental jars. In these experiments wherein CaCXX was used, it is noted that al- though the initial quantity of CaCO:! suspended in the water exceeds by an average of some ten-fold the quantity suspended when the calcareous marine mud was employed in the earlier experiments, nevertheless the TABLE III Removal of suspended CaCOs by mussels — fourth series. Medium mussels: M4 — 4 animals. Lengths 105, 100, 102 and 109 mm. (Average = 104 mm. each). Total weight 342.5 grams (Average = 85.6 grain- each). M6 — 4 animals. Lengths 106, 104, 102 and 108 mm. (Average = = 105 mm. each). Total weight 337.5 grams (Average := 84.4 grams each). Small mu-M'l-: S, 4 animals. Lengths 78, 79, 77 and 74 mm. (Average = 77 mm. each). Total weight 139 grams (Average = 34.75 grams). 84; — 4 animals. Lengths 75, 81, 74 and 77 mm. (Average = 77 mm. each). Total weight 156 grams (Average = 39 grams each). Ten-cc. samples taken. Temperature range 19°-20.3° C. CaC< ). u-> >1. Two grams per liter (Sixteen grams total in each tank). Time Minutes Control C Medium Mi M<-ilium MJ Small 83 Small 84 Ca Ca Ca Ca Ca grams/liter grams ,'litfr grams/liter grants /liter grams /liter 1:15 P.M... 0 1.2356 — — • — — 2:10 P.M.. 15 1 . 1 1 85 .9542 .9314 1.0142 1.1000 2:25 P.M. 30 1.0171 .7828 .7271 .8385 .9542 2:40 P.M. 45 .9585 .6714 .6142 .7571 .8642 2:55 P.M.. 60 .8928 .6057 .5457 .6771 .8214 3:10 P.M.... 75 .8685 .5543 .5114 .6371 .7642 3:25 P.M.. 90 .8242 .5185 .4728 .6000 .7200 3:40 P.M. 105 .7828 .4871 .462 S .5500 .6814 3:55 P.M. 120 .7528 .4771 .4443 .5343 .6428 4:10 P.M. 135 .7300 .4585 .4457 .5143 .6085 4:25 P.M. 150 .7028 .4457 .4385 .4957 .5671 4:40 P.M. 165 .6685 .4414 .4243 .4771 .5314 4:55 P.M. 180 .6500 .4343 .4257 .4685 .5143 rate of removal of suspended material by iiiu^rK of from 69 to 105 mm. in length is similar in both series of experiments. (See Table IV.) The total calcium content represents the sum of the amount of cal- cium which is present in solution in the sea water and the added amount which is present in suspended material. The decrease of the calcium content with time is due to: 1. Gravitational settling of suspended material. 2. Removal (filtration) of suspended matt-rial by the mussels. In the control experiments, which were undertaken without mussels in the vessel, the decrease was clue to gravitational settling only, and a RATE OF WATER PROPULSION BY MUSSEL 431 general law can be found for the rate of this settling. During the ex- periments with mussels in the vessel, the decrease was due partly to settling and partly to removal of the material from suspension by the mussels ; knowing the former, the rate of removal of the calcium by the mussels can be determined by difference. In all cases the mussels de- creased the calcium content to a minimum value of about 430 ing. Ca per liter, whereas a determination of the contents in water filtered through paper gave about 440 mg. Ca per liter. Additional unpublished experi- ments have shown that a small amount of the calcium in the water, present probably in a very finely suspended state, could be filtered by the mussels but was not filtered out by paper alone, previous to analysis. It TABLE IV Liters of water passed per mussel per hour. Size of mussel Series No. Experi- Number m in and value ment of Value of b liters of a No. Mean Mean mussels per hour weight length grams mm. (I) 0.39 1 166.5 120 2 0.93 2.2 2 75.6 100 4 1.81 2.8 3 74.8 100 4 1.82 2.9 4 34.3 69 9 2.80 2.1 5 40.0 74 9 2.57 1.9 (2) 0.70 6 101.9 103.5 6 2.59 2.5 (3) 0.54 7 415 176.5 2 5.05 18.1 8 368 180 2 1.36 3.3 (4) 0.41 9 85.6 104 4 1.51 2.2 10 84.4 105 4 1.85 2.9 11 34.8 77 4 0.98 1.1 12a 39.0 77 4 0.65 0.5 12b 39.0 77 4 1.02 1.2 appears therefore justifiable to assume that the content in solution was about 430 mg. Ca per liter and that the suspended amount was equal to the observed value minus 430 ing. Ca per liter. The value 430 may be in error by about 1 mg. /liter whereby a slight uncertainty is intro- duced as to the amount of suspended calcium, and this uncertainty exerts a corresponding slight influence upon the further treatment of the data, since this has to be based upon the suspended amount of calcium and not upon the total amount. By plotting the results from the control experiment one obtains a curve of a form which suggests that in the absence of mussels the amount of suspended calcium can be expressed as an exponential tune- 432 FOX, SVERDRUP, AND CUNNINGHAM tion of time. We assume, therefore, that the amount which is precipi- tated in unit time is proportional to the total amount which is suspended. If we have in the vessel M liters of water containing p ing. of suspended calcium per liter, the amount precipitated in unit time will he uMp, where a is a factor of proportionality. The amount precipitated in tin- time dt will he : (a M p dt) Concerning the action of the mussel, we will assume that in unit time each mussel pumps m liters of water through its system ami removes all suspended calcium from this (juantity of water. The amount removed by one mussel in unit time will then he /;/ p, and if we have n mussels the amount of suspended calcium removed in the time dt will he: (n m p dt) The total amount removed owing to hoth ordinary precipitation and to the action of the mussels. representing the total decrease of the sus- pended amount of calcium, is therefore : (1) d(M p) -- - (nm + aM) p dt where the minus sign indicates a decrease. Therefore; dp f (2) (nm . \ = e (-u+a)' = ,-bt (3) p = p0e V.vT -- p0e The logarithmic decrement a can be- determined from the control ex- periment and the logarithmic decrement /> from the experiments with mussels. The amount of water which passes through one mussel in unit time is then : (4) m = M— — n I f our assumptions are correct, each experiment must reveal a linear relationship between the natural logarithm of f> (the amount of sus- pended calcium in mg. liter) and the time. Furthermore, we must ob- tain nearly the same values of />„. the amount of suspended calcium at tin- beginning of the experiment, since m each case the initial conditions were as similar as possible. RATE OF WATER PROPULSION BY MUSSEL 433 In order to examine this question, the values of /> have been com- puted by subtracting 430 mg./liter from the observed values of the total calcium content. All values of /> smaller than 5 mg./liter have been omitted (except in one instance when a value of 4.3 mg./liter has been retained) since, owing to the slight uncertainty involved in the micro method for determining the amount in solution (430 mg./liter), small values of /> have no significance. In the case of the control experiments values only from the first three hours have been considered, although /> remained greater than 5 mg./liter. In Figs. 5, 6, 7, and 8, the values of /;/ />, as derived from the four series of experiments, have been plotted against time. The straight lines represent lines of regression which have been computed by the method of least squares. It is seen that all points fall so near to the lines that the deviations may be considered accidental. Accidental devi- ations must arise, since the samples taken from the middle of the jar cannot be expected to show exactly the average contents of suspended calcium, and since errors of determination are present. In the case of the small mussels (S4), in series No. 4, a break in the line appears to occur after 130 minutes, indicating that the rate of propulsion of water through the mussels suddenly changed. From the control experiments, we obtain various values of our con- stant a, using one hour as the unit of time (see Table IV). The varia- tion of the values of a may be due to differences in the velocity of the stream of air which was passed through the water in order to insure homogeneous mixing, or, to a lesser extent, to differences in the effect of the salt water in coagulating the suspended material. In all experi- ments, howevec, the velocity of the air stream was adjusted so that the rate of bubbling through control jars and experimental jars was as closely identical as estimation would permit. From the experiments with mussels we find the values of b which are shown in Table IV and the values of m (in liters per hour) which have been computed by means of formula (4), introducing M (8 liters), the number of mussels (»), and the value of a which was found by simultaneous control experiment. The fact that in all cases we find a linear relationship between In p and time appears to furnish strong evidence for the correctness of our assumptions, but from our equations it is evident that we should obtain a linear relationship on other assumptions than these which have been introduced. We could assume that the mussels remove only a constant fraction of the suspended calcium when //; liters pass through the mus- sel. The amount removed would then be ciun p, where c must be sup- 434 FOX, SYERDRUP. AND CUNNINGHAM posed to be a constant factor smaller than 1.0. It appears, however, improhahlc that a mussel should remove, say. 50 ing. /liter when the content was 100 mg. liter and in the same interval of time. 2 mg. /liter when the content was 4 mg. liter. A change in the ratio of the re- moved calcium from say 80 per cent of a high content to 100 per cent of a small content is. however, possible and would not be detected. If such a change takes place, our values of ;» are conservative, being some- what low. Another conceivable possibility which may be mentioned is that the amount of water passing through the mussel chamber might de- crease with time and the fraction of calcium removed might increase (or vice versa). However, the product nip must, since /;; /> is a linear fraction of time, remain nearly constant, and this means that the amount of water passing through the mussel chamber and the fraction of cal- cium removed would have to change in opposite directions at the same rate. It is difficult to conceive a mechanism which would work in such a manner. The simplest explanation of our results appears to be that our original assumptions are correct, vi/... that a constant amount of water passes through the mussel chamber in unit time and that virtually all suspended (and colloidal) calcium in this amount of water is re- moved. \Yc have furthermore supposed that the water passing through the mussels has the average calcium content of the water in the vessel. In order to satisfy this condition, the water must be stirred, since the mus- j sels lying on the bottom of the vessel remove practically all suspended calcium from the water passing through them. The stirring must be adapted to the number of mussels; the greater this number the more rapid must the stirring be. The effect of insufficient stirring upon the observations would be that /;/ p would no longer be a linear function of time, but would decrease more and more slowly as the experiment ad- vanced. The reason for this is that the amount of suspended calcium removed by the mussels would no longer be proportional to [>, but since the bottom layer, where the mussels lie. would be depleted of calcium more rapidly, it would be proportional to some function of [> which de- creases with time. \Yc can introduce ;/;;/ p c ' ' --• In pn - at - ~: (1 •- e~ct) RATE OF WATER PROPULSION BY MUSSEL At great values of t, we obtain 435 n m In p ---- In po - - -- at, meaning that the decrease in the calcium content approaches the value which is due to gravitational settling only. The coefficient c which has been introduced depends upon the rate with which water passes through the mussels and upon the rate of stirring, and a determination of this coefficient is hardly possible. A computation therefore cannot be based upon the results of such experiments. During the experiments which are dealt with here, the water was stirred by passing through a constant stream of air, and the agreement between observed and computed values shows that the stirring was sufficient. Two preliminary experiments which Fox and Schwenck undertook in the summer of 1935 with two very large mussels weighing 485 and TABLE V Rates of water propulsion by mussels. Weight of mussels Length of mussels Total Rate of water propulsion Number ex- no. periments mussels Range Average Range Average used Range Average grams grams mm. mm. liters/hr. Hlers/hr. 337-515 431 174-182 178 4 6 1.8-18.1 6.4 75-166 93 95-130 102 6 24 2.2- 2.9 2.6 34- 40 37 60- 82 74 4 26 0.5- 2.1 1.4 515 grams show a wider scattering of the observed values, but they indi- cate a linear relationship between /;; p and t, giving m equal to 1.8 and 2.4 liters per hour respectively. Other preliminary experiments by Fox and Schwenck, in the summer of 1935, gave less consistent values. In these experiments a large num- ber of small mussels were used and In p decreased rapidly at the begin- ning, but slowly at the end of the experiments, probably because the stirring was insufficient. The results of all experiments, including the preliminary ones by Fox and Schwenck, are summarized in Table V. The very large mus- sels show an enormous range in their rate of water propulsion, perhaps owing to individual differences or perhaps because they work intermit- tently. The value obtained from one of the experiments, 18.1 liters per hour, is quite enormous, but it is undoubtedly correct since it is based on good observations during one full hour, and it shows that in certain cir- 436 FOX, SVERDRUP, AND CUNNINGHAM cumstanccs large mussels can pump great quantities of water through their systems. The average value for the large mussels, 6.4 liters per hour, cannot he given any weight owing to the wide range of the single values and the small number of experiments. The medium-sized mus- sels appear to be more consistent in their behavior, the range of the single values is small and the average value, 2.6 liters per hour, can therefore be considered nearly correct. The small mussels show again a wider range of their rate of water propulsion and while the average value is uncertain, it undoubtedly lies below that of the larger animals. It would be expected that both the weight and the volume (capacity of the gill chamber) of the mussel should be proportional to some fairly constant power of one of the linear dimensions, say the length of the shell, and that weight should show a linear relationship to capacity. The capacity (in ml. ) was determined by allowing mussels whose gill cham- bers had been emptied of sea water to refill their cavities on immersion in a known volume of sea water, then causing the animals to close their valves, removing them from the vessel, and measuring the residual water in the container. Plotting the average lengths of a scries of different sized animals (in mm.) against their capacities (in ml.) showed a steeply rising curve of an exponential character ; length plotted similarly against weight of animals (in grams) emptied of water showed a curve of simi- lar character (see also Galtsoff, 1931. who obtained a similar relationship in the Hawaiian pearl oyster) ; finally the relationship between weight and capacity was shown to be a linear one. An extensive series of experiments performed upon mussels of dif- ferent known capacities, weights, and linear dimensions would perhaps disclose an interesting relationship between any of these attributes (say the volume), and the rate of propulsion of water by each size of " pump- ing system." SUMMARY A method is described for determining tin- approximate average rate at which the California mussel, Mytilus calif or nianus, propels water through its gill chambers. The method consists of analyses at frequent intervals of the amount of calcium remaining in suspension (as CaCO:i) in a given volume of continually stirred water containing the mussels, which remove the sus- pended material as they pass the water through their chambers. Mathematical treatment and interpretation of the data obtained, sup- port the conclusions that (1) virtually all of the suspended matter is re- moved as the water passes over the- mucous surfaces of the gills and mantle of the mussel; (2) the mussel propels the water rhythmically RATE OF WATER PROPULSION BY MUSSEL 437 through its filtering system at a rate which on an avc-rage is constant, varying according to the size and perhaps also according to other physi- ological attributes of the animals. In medium-sized animals (of 95 to 130 mm. length) the propulsion rate may vary between extreme values of 2.2 and 2.9 liters per hour, and has an average value of approximately 2.6 liters per hour. ACKNOWLEDGMENTS Grateful acknowledgment is expressed to Dr. Roderick Craig, and to Mr. Rae Schwenck, each of whom assisted materially in the pre- liminary experiments; to Dr. E. M. Thorp who supplied us with the screened marine muds ; and to several members of the Federal Works Progress Administration on this Project who rendered help in the col- lection and preparation of organisms, technical laboratory assistance, drafting, and typing. LITERATURE ALLEN, W. R., 1914. The food and feeding habits of freshwater mussels. BioL Bull., 27: 127. DAMAS, D., 1935. Le role des organismes dans la formation des vases marines. Ann. dc la Soc. Gcologiquc dc Belgique, 58: 143. DODGSON, R. W., 1928. Report on Mussel Purification. Ministry of Agriculture and Fisheries. Fisheries Investigations, Series II, Vol. 10, No. 1. Fox, D. L., ET AL., 1936. The habitat and food of the California sea mussel. Bull. Scripts hist, of Oceanography, Tech. Scr., 4: 1. GALTSOFF, P. S., 1926. New methods to measure the rate of flow produced by the gills of oyster and other molluscs. Sci., 63: 233. GALTSOFF, P. S., 1928a. The effect of temperature on the mechanical activity of the gills of the oyster (Ostrea virginica Gm.). Jour. Gen. Ph\siol., 11: 415. GALTSOFF, P. S., \928b. Experimental study of the function of the oyster gills and its bearing on the problems of oyster culture and sanitary control of the oyster industry. Bull. Bur. Fish., 44: 1. GALTSOFF, P. S., 1931. The weight-length relationship of the shells of the Hawai- ian pearl oyster, Pinctada sp. Am. Nat., 65: 423. GRAVE, C, 1905. Investigations for the promotion of the oyster industry of North Carolina. Report, U. S. Comm. Fish. 1903 (1905) : 249. KELLOGG, J. L., 1915. Ciliary mechanisms of lamellibranchs with descriptions of anatomy. Jour. Morph., 26: 625. KIRK, P. L. AND E. G. MOBERG, 1933. Microdetermination of calcium in sea water. Ind. and Eng. Chem., Analyt. Edition, 5: 95. MOORE, H. F., 1905. Anatomy, embryology, and growth of the oyster. Report, U. S. Comm. Fish, 1903 (1905) : 317. MOORE, H. F., 1908. Volumetric studies of the food and feeding of oysters. Bull. Bur. Fish., 28 (Part 2) : 1297. MOORE, H. F., 1913a. Condition and extent of the natural oyster beds and barren bottoms of Mississippi east of Biloxi. Rep. U. S. Comm. of Fisheries, 1911 (1913). Bur. of Fish. Document No. 774. MOORE, H. F., I9l3b. Condition and extent of the natural oyster beds and barren bottoms of Mississippi Sound, Ala. Ibid. Bur. of Fish. Document No. 769. 438 FOX, SVERDRUP, AND CUNNINGHAM NELSON, T. C, 1921. Report of the Department of Biology of the New Jersey Agricultural College Experiment Station for the year ending June 30, 1920 (1921), p. 317. NELSON, T. C., 1935. Water filtration by the oyster and a new hormone effect thereon. Anat. Rec., 64 (Suppl. 1): 68. PARKER, G. H., 1914. On the strength and the volume of the water currents pro- duced by sponges. Jour. E.vpcr. Zoo/., 16: 443. RANSOM, G., 1926. Le filtration de 1'eau par les Lamellibranch.es et ses con- sequences. Bull, dc rinstitiite Oceanographique, No. 469, p. 1. VIALLANES, H., 1892. Recherches sur la filtration de 1'eau par les Mollusques et applications a I'Ostreiculture et a I'Oceanographie. Compt. rend. Acad. Sci., 114 (2): 1386. WELLS, W. F., 1916. Artificial purification of oysters. A report of experiments upon the purification of polluted oysters by placing them in water to which calcium hypochlorite has been added. Reprint No. 351, Publ. Health Re- ports, Vol. 31, No. 28, U. S. Publ. Health Service, p. 1848. Washington (cited by Galtsoff, 1928ft). YONGE, C. M.. 1928. Feeding mechanisms in the invertebrates. Biol. Rev.. 3: 21. ZoBELL, C. E., AND W. A. LANDON, 1937. The bacterial nutrition of the California mussel. Proc. Soc. E.rpcr. Biol. and Mcd. (in press). INDEX A BRAMOWITZ, ALEXANDER A. The chromatophorotropic hormone of Crustacea, 344. Acanthephyridae, eyes of, 57. Accelerator substances, evidence for pro- duction by ultraviolet radiation of Limulus muscle, 75. ALLEE, W. C., AND GERTRUDE EVANS. Some effects of numbers present on the rate of cleavage and early de- velopment in Arbacia, 217. Amoeba proteus, mitosis in, 125. Amphiscolops langerhansi, reproductive system and copulation, 319. Arbacia, cleavage rate, as affected by numbers present, 217. Arbacia punctulata, prolongation of life in egg cells and calcium reduction, 366. AREY, LESLIE B. Observations on two types of respiration in Onchidium, 41. Aurelia, budding and locomotion in scyphistomas of, 99. "DACTERIA, and phosphorus cycle in ' the sea, 190. BEADLE, G. W. See Clancy and Beadle, 47. VON BRAND, THEODOR. Observations upon the nitrogen of the particulate matter in the sea, 1. — , - — , NORRIS W. RAKESTRAW AND CHARLES E. RENN. The ex- perimental decomposition and re- generation of nitrogenous organic matter in sea water, 165. Budding and locomotion in scyphistomas of Aurelia, 99. /^ALANUS finmarchicus, feeding rate of, in relation to environmental con- ditions, 233. Calcium reduction and prolongation of life in egg cells of Arbacia punctu- lata, 366. CAMERON, JOHN ANDREW. The mitotic rate in tadpole skin after repeated injury, 37. Caterpillars, gustation and olfaction in, 7. Centrifuging, behavior of maturation spindles in polar fragments of Ilya- nassa eggs after, 88. CHACE, F. A., JR. See Welsh and Chace, 57. CHASE, HYMAN Y. The effect of ultra- violet light upon early development in eggs of Urechis caupo, 377. Chromatophorotropic hormone of Crus- tacea, 344. CHURNEY, LEON, AND HERBERT M. KLEIN. The electrical charge on nuclear constituents (salivary gland cells of Sciara coprophila), 384. Clam, adult, seasonal gonadal changes, 406. — , development of primary gonad in, 389. CLANCY, C. W., AND G. W. BEADLE. Ovary transplants in Drosophila melanogaster: studies of the char- acters singed, fused, and female- sterile, 47. Cleavage, rate in Arbacia, effects of numbers present, 217. Color-changes, and chromatophorotropic hormone of Crustacea, 344. — , and diurnal rhythm in Ligia baudiniana, 24. COONFIELD, B. R. Symmetry and regu- lation in Mnemiopsis leidyi, Agassiz, 299. CUNNINGHAM, JOHN P. See Fox, Sverd- rup and Cunningham, 417. Cyst formation in glomerular tufts of certain fish kidneys, 247. T~}ALCQ, A., AND G. VANDEBROECK. On the significance of the polar spot in ripe unfertilized and in fertilized ascidian eggs, 311. DAWSON, J. A., WALTER R. KESSLER AND JOSEPH K. SILBERSTEIN. Mitosis in Amoeba proteus, 125. Decomposition, experimental, nitrogen- ous organic matter in sea water, 165. 439 440 INDEX Depressor substances, production by ul- traviolet radiation of Limulus muscle, 75. DETHIER, VINCENT G. Gustation and olfaction in lepidopterous larvae, 7. Digestion, enzymes of, in Daphnia and Polyphemus, Diaptomus and Cal- anus, 290. Drosophila melanogaster, ovary trans- plants in, 47. "PLECTRICAL charge on nuclear con- ' stituents of salivary gland cells, 384. Embryonic determination, production of intermediate-winged aphids and, 259. Endomixis, induction of, in Paramecium aurelia, 196. Enzymes, digestive, in Daphnia and Polyphemus, Diaptomus and Cal- anus, 290. EVANS, GERTRUDE. See Allee and Ev- ans, 217. Eyes, of Acanthephyrida?, 57. pAURE-FREMIET, E. Licnophora lyngbycola, a new species of infu- sorian from Woods Hole, 212. Feeding rate of Calanus finmarchicus in relation to environmental conditions, 233. Food, effects on gastric epithelial cells of grasshopper, Mclanoplus differen- tialis Thomas, 203. — , rate of locomotion and reproduction in marine Amoeba, 334. Fox, DENIS L., H. U. s\ i KIIKIT AND JOHN P. CUNNINC.II \M. The rate of water propulsion by the California mussel, 417. FULLER, JOHN I.. I-Ycding rate of Calanus finmarchicus in relation to environmental conditions, 233. QASTRIC epithelial cells, effects of diet, in grasshopper, 203. GILCHRIST, FRANCIS G. Hudding and locomotion in the scyphistomas of Aurelia, 99. Gonad, changes of, seasonal, in adult clam, 406. , primary, development in Venus mercenaria Linmeus, 389. Gorgodera amplicava, development of, in final host, 80. GRAFFLIN, ALLAN L. Cyst formation in the glomerular tufts of certain fish kidneys, 247. Grasshopper, effects of diet on gastric epithelial cells of, 203. Gustation and olfaction in lepidopterous larvae, 7. GUTTMAN, S. A. Evidence for the pro- duction of accelerator and depressor substances by ultraviolet radiation of Limulus muscle, 75. TJASLER, ARTHUR D. The physiol- ogy of digestion in plankton Crus- tacea. II. Further studies on t In- digestive enzymes of (A) Daphnia and Polyphemus. (B) Diaptomus and Calanus, 290. HODGE, CHARLES, 4TH. Some effects of diet on the gastric epithelial cells of the grasshopper, Melanoplus differ- entialis Thomas, 203. HOPKINS, DWIGHT LUCIAN. The rela- tion between food, the rate of loco- motion and reproduction in the ma- rine Amtrba, Flabellula mira, 334. Hormone, chromatophorotropic, of Crus- tacea, 344. Hydroid, new genus and method of asexual reproduction, 327. HYMAN, LIBBIE II . Reproductive sys- tem and copulation in Amphiscolops langerhansi (Turbellaria Aca-la), 319. TLYANASSA eggs, behavior of matura- tion spindles in polar fragments of, after centrifuging, 88. Injury, mitotic rate in tadpole skin after, 37. Intermediate-winged aphids and embry- onic determination, 259. J^- ESSLER, WALTER R. See Dawson, Kessler and Silberstein, 125. Kit! neys, fish, cyst formation in glomeru- lar tufts of, 247. KLEIN, HERHERT M. See Churney and Klein, 384. KLEINHOLZ, L. II. Studies in the pig- mentary system of C'rustacea. I. Color changes and diurnal rhythm in Ligia baudiniana, 24. — , — . — . Studies in the pigmentary system of Crustacea. II. Diurnal INDEX 441 movements of the retinal pigments of Bermudan decapods, 176. T ICNOPHORA lyngbycola, new spe- cies of infusorian from Woods Hole, 212. Ligia baudiniana, color changes and diurnal rhythm, 24. Locomotion and budding in scyphistomas of Aurelia, 99. — , rate of, reproduction, and food in marine Amoeba, 334. LOOSANOFF, VICTOR L. Development of the primary gonad in Venus mer- cenaria Linnaeus, 389. — , — , — . Seasonal gonadal changes of adult clams, Venus mer- cenaria (L.), 406. Loping of land-snails, 287. V/f ATURATION spindles, behavior of, in polar fragments of eggs of Ilyan- assa obtained by centrifuging, 88. MILES, SAMUEL STOCKTON. A new genus of hydroid and its method of asexual reproduction, 327. Mitosis in Amoeba proteus, 125. — , rate of, in tadpole skin after re- peated injury, 37. Mnemiopsis, symmetry and regulation, 299. MORGAN, T. H. The behavior of the maturation spindles in polar frag- ments of eggs of Ilyanassa obtained by centrifuging, 88. Muscle, Limulus, production of accelera- tor and depressor substances by ul- traviolet radiation, 75. •MTFROGEN of particulate matter in the sea, 1. Nitrogenous organic matter in sea water, experimental decomposition and re- generation, 165. Nuclear constituents, electrical charge on, in salivary gland cells, 384. QDLAUG, THERON O. Notes on the development of Gorgodera ampli- cava in the final host, 80. Olfaction in lepidopterous larvae, 7. Onchidium, two types of respiration in, 41. Ovary transplants in Drosophila melano- gaster, 47. pARAMECIUM aurelia, induction of endomixis, 196. PARKER, G. H. The loping of land- snails, 287. Phosphorus cycle and bacteria in the sea, 190. Pigments, retinal, diurnal movements of, Bermudan decapods, 176. Polar spot, significance of, in ripe unfer- tilized and in fertilized ascidian eggs, 311. 13 AKESTRAW, NORRIS W. See von Brand, Rakestraw and Renn, 165. Regulation and symmetry in Mnemiopsis leidyi, 299. RENN, CHARLES E. Bacteria and the phosphorus cycle in the sea, 190. — , - — . See von Brand, Rake- straw and Renn, 165. Reproduction, asexual, in new genus of hydroid, 327. — , cycles of, and superfetation in poeci- liid fishes, 145. — , relation between food, locomotion rate and, in Amoeba, 334. — , system of, and copulation in Amphiscolops langerhansi, 319. Respiration, two types, in Onchidium, 41. Rhythm, diurnal, and color changes in Ligia baudiniana, 24. C ALIVARY gland cells, electrical charge on nuclear constituents, 384. SCHECHTER, VICTOR. Calcium reduc- tion and the prolongation of life in the egg cells of Arbacia punctulata, 366. Sea water, nitrogenous organic matter in, experimental decomposition and re- generation of, 165. SHULL, A. FRANKLIN. The production of intermediate-winged aphids with special reference to the problem of embryonic determination, 259. SILBERSTEIN, JOSEPH K. See Dawson, Kessler and Silberstein, 125. Snails, land, loping of, 287. SONNEBORN, T. M. Induction of endo- mixis in Paramecium aurelia, 196. Superfetation and reproductive cycle in poeciliid fishes, 145. SVERDRUP, H. U. See Fox, Sverdrup and Cunningham, 417. 442 IMM.X Symmetry and regulation in Mnemiop- sis leidyi, 299. BURNER, C. L. Reproductive cycles and superfetation in poeciliid fishes, 145. {JLTRAVIOLET light, effect on early development in eggs of Urechis caupo, 377. - radiation, production of accelerator and depressor substances, in Limulus muscle, 75. Urechis caupo, eggs of, effect of ultra- violet light, 377. yANDEBROECK, G. See Dalcq and Vandebroeck, 311. \A7ATER propulsion, rate of, by Cali- fornia mussel, 417. WELSH, J. 1L, AND F. A. CHACE, JR. Eyes of deep sea crustaceans. I. Acantht-phyridae, 57. Volume LXXII Number 1 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. G. CONKLIN, Princeton University FRANK R. LlLLIE, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University W. M. WHEELER, Harvard University E. E. JUST, Howard University EDMUND B. WILSON, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor FEBRUARY, 1937 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, i'a. Agent for Great Britain : Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between June 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the remainder of the year. INSTRUCTIONS TO AUTHORS Preparation of Manuscript. In addition to the text matter, manuscripts should include a running page head of not more than thirty-five letters. Footnotes, tables, and legends for figures should be typed on separate sheets. Preparation of Figures. The dimensions of the printed page (4%x7 inches) should be borne in mind in preparing figures for publication. Draw- ings and photographs, as well as any lettering upon them, should be large enough to remain clear and legible upon reduction to page size. Illustrations should be planned for sufficient reduction to permit legends to be set below them. In so far as possible, explanatory matter should be included in the legends, not lettered on the figures. Statements of magnification should take into account the amount of reduction necessary. Figures will be reproduced as line cuts or halftones. Figures intended for reproduction as line cuts should be drawn in India ink on white paper or blue-lined coordinate paper. Blue ink will not show in reproduction, so that all guide lines, letters, etc. must be in India ink. Figures intended for reproduction as halftone plates ^hould be grouped with as little waste space as possible. Drawings and li-ttering for halftone plates should be made directly on heavy Bristol board, nut pasted on, as the outlines of pasted letters or drawings appear in the n -production unless removed by an expensive process. Methods of repro- duction nut regularly employed by the I5iological Bulk-tin will be used only at the author's expense. The originals of illustrations will not In. returned except by special request. I 'irt'ctions for Mailinii. Manuscripts and illustrations should be packed Hat between stiff cardboards. Large charts and graphs may be rolled and in a mailing tube. Authors will be furnished, free of charge, one hundred re- prints without covers. Additional copies may be obtained at cost. /'rue/. Page proof will be furnished only upon special request. When eferences are made in the text, the material referred to should be marked clearly on the galley proof in order that the proper page numbers may be supplied. Kiitered Mctolii-r Id, 1';|)2, at Lancaster. Pa., as second-class matter umler Aet of Congress of July 16, 1894. LANCASTER PRESS, Inc. LANCASTER, PA. THE EXPERIENCE we have gained from printing some sixty educational publica- tions has fitted us to meet the standards of customers who demand the best. We shall be happy to have workers at the MARINE BIOLOGICAL LABORATORY write for estimates on journals or monographs. Our prices are moderate. A Perfect Illustration Or the lack of it, may make or mar a scientific paper. For 65 years we have specialized in making reproductions by the Helio- type process of the most delicate pencil and wash drawings and photo- graphs; and by the I leliochrome proc- ess of paintings and drawings in color . A series of drawings reproduced by us appear on p. 139 of this journal. Ask the editor to whom you submit your next paper to secure our esti- mates for the reproduction of your illustrations. The Heliotype Corporation Est. 1872 172 Green St., Jamaica Plain, Boston, Mass. IMPORTANT NOTICE TO SUBSCRIBERS IBRARIES and individuals desiring to complete ets or runs of the BIOLOGICAL BULLETIN will be able to obtain certain numbers and volumes at reduced prices. In order to equalize stock, a number of special offers are being made, including reduced prices on a limited number of copies of the Index to the first sixty volumes of the Bulletin. Prices will be furnished on request. Please address communications to: SECRETARY, THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, Massachusetts BIOLOGICAL MATERIALS For a number of years the Supply Department has been furnishing Living Marine Materials. From experience it has been found that during the period between the first of November and the end of February these live animals and plants can be shipped with the most success. This service has proven to be of great value to both instructor and student, particularly to those who are located at some distance from the sea coast. Set I Live Marine Material for five-gallon Aquarium $ 5.00 Set II Live Marine Material for ten-gallon Aquarium 10.00 We assume responsibility for shipments going as far west as the Missis- sippi and as far south as Georgia, and guarantee to deliver the animals in good condition. We are overstocked with preserved horseshoe crabs (Limulus polyphemus), Nereis, and Ensis, and are glad of this opportunity to offer these and a few additional specimens at the following reduced prices : Limulus, 2Y-2. to 3% inch breadth of carapace . .$1.50 per doz. Nereis, 5-7 inch average 1.00 " " Nereis, extra large, 12" and over 1.50 " " Metridium, small, for dissection 1.00 " " Ensis, 8-10 inch specimens 1.00 " " Drosophila cultures (list of stocks on request) . 1.00 each LOBSTERS During the summer months local fishermen bring in hundreds of Lobsters at a time for preservation. It is impossible in handling so many of these specimens to keep all of them intact, and often an " arm," or claw, becomes detached. This does not lessen the value of the Lobster for classroom dissection. At present we have a large number of such Lobsters on hand, both plain preserved and injected, and we would be pleased to quote greatly reduced rates on same upon request. SUPPLY DEPARTMENT Est. 1890 MARINE BIOLOGICAL LABORATORY WOODS HOLE, MASS. CONTENTS Page VON BRAND, THEODOR Observations upon the Nitrogen of the Particulate Matter in the Sea 1 DETHIER, VINCENT G. Gustation and Olfaction in Lepidopterous Larvae KLEINHOLZ, L. H. Studies in the Pigmentary System of Crustacea. I. Color changes and diurnal rhythm in Ligia baudiniana 24 CAMERON, JOHN ANDREW The Mitotic Rate in Tadpole Skin after Repeated Injury. ... 37 AREY, LESLIE B. Observations on Two Types of Respiration in Onchidium. . . 41 CLANCY, C. W., AND G. W. BEADLE Ovary Transplants in Drosophila melanogaster : Studies of the Characters Singed, Fused, and Female-Sterile 47 WELSH, J. H., AND F. A. CHACE, JR. Eyes of Deep Sea Crustaceans. I. Acanthephyridae 57 GUTTMAN, S. A. Evidence for the Production of Accelerator and Depressor Substances by Ultraviolet Radiation of Limulus Muscle 75 ODLAUG, THERON O. Notes on the Development of Gorgodera amplicava in the Final Host 80 MORGAN, T. H. The Behavior of the Maturation Spindles in Polar Fragments of Eggs of Ilyanassa Obtained by Centrifuging 88 GILCHRIST, FRANCIS G. Budding and Locomotion in the Scyphistomas of Aurelia ... 99 DAWSON, J. A., WALTER R. KESSLER AND JOSEPH K. SILBERSTEIN Mitosis in Amoeba proteus 125 Volume LXXII Number 2 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, E. G. CONKLIN, Princeton University E. N. HARVEY, Princeton University SELIG HECHT, Columbia University LEIGH HOADLEY, Harvard University M. H. JACOBS, University of Pennsylvania H. S. JENNINGS, Johns Hopkins University E. E. JUST, Howard University Columbia University FRANK R. LILLIE, University of Chicago CARL R. MOORE, University of Chicago GEORGE T. MOORE, Missouri Botanical Garden T. H. MORGAN, California Institute of Technology G. H. PARKER, Harvard University W. M. WHEELER, Harvard University EDMUND B. WILSON, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor •*••<*• t APRIL, 1937 firs Printed and Issued by LANCASTER PRESS, Inc. PRINCE &, LEMON STS. LANCASTER, PA. THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between June 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the remainder of the year. INSTRUCTIONS TO AUTHORS Preparation of Manuscript. In addition to the text matter, manuscripts should include a running page head of not more than thirty-five letters. Footnotes, tables, and legends for figures should be typed on separate sheets. Preparation of Figures. The dimensions of the printed page (4% x 7 inches) should be borne in mind in preparing figures for publication. Draw- ings and photographs, as well as any lettering upon them, should be large enough to remain clear and legible upon reduction to page size. Illustrations should be planned for sufficient reduction to permit legends to be set below them. In so far as possible, explanatory matter should be included in the legends, not lettered on the figures. Statements of magnification should take into account the amount of reduction necessary. Figures will be reproduced as line cuts or halftones. Figures intended for reproduction as line cuts should be drawn in India ink on white paper or blue-lined coordinate paper. I'.liu- ink will not show in reproduction, so that all guide lines, letters, etc. must be in India ink. Figures intended for reproduction as halftone plates .should be grouped with as little waste space as possible. Drawings and Irtu -ring for halftone plates should be made directly on heavy Bristol board, not pasted on, as the outlines of pasted letters or drawings appear in the reproduction unless removed by an expensive process. Methods of repro- duction not regularly employed by the Biological Bulletin will be used only at the author's expense. The originals of illustrations will not be returned cxcc-pt by special request. lUrci lions for Mailing. Manuscripts and illustrations should be packed fiat lift\vc(.'ii stiff cardboards. Large charts and graphs may be rolled and -cut in a mailing tube. .'\\-printx. Authors will be furnished, free of charge, one hundred re- prints without covers. Additional copies may be obtained at cost. I'roof. 1'a-i- proof will be furnished only upon special request. When ••ferences are made in the text, the material referred to should be inarkdl clearly on the galley proof in order that the proper page numbers may be supplied. Kntrred October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. LANCASTER PRESS, Inc. LANCASTER, PA. & THE EXPERIENCE we have gained from printing some sixty educational publica- tions has fitted us to meet the standards of customers who demand the best. We shall be happy to have workers at the MARINE BIOLOGICAL LABORATORY write for estimates on journals or monographs. Our prices are moderate. A Perfect Illustration Or the lack of it, may make or mar a scientific paper. For 65 years we have specialized in making reproductions by the Helio- type process of the most delicate pencil and wash drawings and photo- graphs; and by the Heliochrome proc- ess of paintings and drawings in color. Ask the editor to whom you submit your next paper to secure our esti- mates for the reproduction of your illustrations. The Heliotype Corporation Est. 1872 172 Green St., Jamaica Plain, Boston, Mass. IMPORTANT NOTICE TO SUBSCRIBERS IBRARIES and individuals desiring to complete sets or runs of the BIOLOGICAL BULLETIN will be able to obtain certain numbers and volumes at reduced prices. In order to equalize stock, a number of special offers are being made, including reduced prices on a limited number of copies of the Index to the first sixty volumes of the Bulletin. Prices will be furnished on request. Please address communications to: SECRETARY, THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, Massachusetts BIOLOGICAL MATERIALS Since 1890 the Supply Department of the Marine Biological Laboratory has been furnishing- both living and preserved specimens to schools and col- leges. It is the desire of the Laboratory to con- tinue this service in an efficient and satisfactory manner, and your cooperation is earnestly desired. All our materials are freshly collected each sea- son and are carefully prepared by men of long experience. Our current catalogue lists an excellent supply of specimens for summer school use. We guarantee all our materials to give absolute satisfaction. Catalogues are furnished on request SUPPLY DEPARTMENT Est. 1890 MARINE BIOLOGICAL LABORATORY WOODS HOLE, MASS. CONTENTS Page TURNER, C. L. Reproductive Cycles and Superfetation in Poeciliid Fishes . . 145 VON BRAND, THEODOR, NORRIS W. RAKESTRAW AND CHARLES E. RENN The Experimental Decomposition and Regeneration of Ni- trogenous Organic Matter in Sea Water 165 KLEINHOLZ, L. H. Studies in the Pigmentary System of Crustacea. II. Diurnal movements of the retinal pigments of Bermudan decapods. . 176 RENN, CHARLES E. Bacteria and the Phosphorus Cycle in the Sea 190 SONNEBORN, T. M. Induction of Endomixis in Paramecium aurelia ... 196 HODGE, CHARLES, 4TH Some Effects of Diet on the Gastric Epithelial Cells of the Grasshopper, Melanoplus differentialis Thomas 203 FAURE-FREMIET, E. Licnophora lyngbycola, a New Species of Infusorian from Woods Hole 212 ALLEE, W. C., AND GERTRUDE EVANS Some Effects of Numbers Present on the Rate of Cleavage and Early Development in Arbacia 217 FULLER, JOHN L. Feeding Rate of Calanus finmarchicus in relation to Environ- mental Conditions 233 GRAFFLIN, ALLAN L. Cyst Formation in the Glomerular Tufts of Certain Fish Kidneys. 247 Volume LXXII Number 3 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. G. CONKLIN, Princeton University E. N. HARVEY, Princeton University SELIG HECHT, Columbia University LEIGH HOADLEY, Harvard University M. H. JACOBS, University of Pennsylvania H. S. JENNINGS, Johns Hopkins University E. E. JUST, Howard University FRANK R. LILLIE, University of Chicago CARL R. MOORE, University of Chicago GEORGE T. MOORE, Missouri Botanical Garden T. H. MORGAN, California Institute of Technology G. H. PARKER, Harvard University EDMUND B. WILSON, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor JUNE, 1937 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain : Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between June 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the remainder of the year. INSTRUCTIONS TO AUTHORS Preparation of Manuscript. In addition to the text matter, manuscripts should include a running page head of not more than thirty-five letters. Footnotes, tables, and legends for figures should be typed on separate sheets. Preparation of Figures. The dimensions of the printed page (4% x 7 inches) should be borne in mind in preparing figures for publication. Draw- ings and photographs, as well as any lettering upon them, should be large enough to remain clear and legible upon reduction to page size. Illustrations should be planned for sufficient reduction to permit legends to be set below them. In so far as possible, explanatory matter should be included in the legends, not lettered on the figures. Statements of magnification should take into account the amount of reduction necessary. Figures will be reproduced as line cuts or halftones. Figures intended for reproduction as line cuts should be drawn in India ink on white paper or blue-lined coordinate paper. Blue ink will not show in reproduction, so that all guide lines, letters, etc. must be in India ink. Figures intended for reproduction as halftone plates should be grouped with as little waste space as possible. Drawings and lettering for halftone plates should be made directly on heavy Bristol board, not pasted on, as the outlines of pasted letters or drawings appear in the reproduction unless removed by an expensive process. Methods of repro- duction not regularly employed by the Biological Bulletin will be used only at the author's expense. The originals of illustrations will not be returned except by special request. 1 Directions for Mailing. Manuscripts and illustrations should be packed flat between stiff cardboards. Large charts and graphs may be rolled and sent in a mailing tube. Reprints. Authors will be furnished, free of charge, one hundred re- prints without covers. Additional copies may be obtained at cost. Proof. Page proof will be furnished only upon special request. When cross-references are made in the text, the material referred to should be marked clearly on the galley proof in order that the proper page numbers may lie supplied. Knti-ml October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. LANCASTER PRESS, Inc. LANCASTER, PA. THE EXPERIENCE we have gained from printing some sixty educational publica- tions has fitted us to meet the standards of customers who demand the best. We shall be happy to have workers at the MARINE BIOLOGICAL LABORATORY write for estimates on journals or monographs. Our prices are moderate. A Perfect Illustration Or the lack of it, may make or mar a scientific paper. For 65 years we have specialized in making reproductions by the Helio- type process of the most delicate pencil and wash drawings and photo- graphs; and by the Heliochromc proc- ess of pa int ings and drawings in color. Ask the editor to whom you submit your next paper to secure our esti- mates for the reproduction of your illustrations. The Heliotype Corporation Est. 1872 172 Green St., Jamaica Plain, Boston, Mass. New Junior Model- VAN SLYKE MANOMETRIC BLOOD GAS APPARATUS This Junior Model apparatus meets the demand of suffi- cient accuracy and easy manipulation at loiv cost. The extraction chamber is the same as that supplied with the motor driven Research Model. It is guaranteed to be calibrated to ±0.003 ml. at 0.5 ml. and ±0.005 ml. at 2.0 ml. This new model is conveniently shaken by hand. The omission of the water jacket and the use of the meter stick to read the pressures on the manometer introduce an error not exceeding 0.2 vol. per cent. DESIGNED FOR STUDENT INSTRUCTION REASONABLY PRICED The low price of the Junior Model Van Slyke Manometric Apparatus makes possible its more general use in student instruction and clinical laboratories. At slight additional cost, this apparatus can be supplied with the millimeter scale etched on the glass manometer tube, and the latter fused to the apparatus. for Literature and Price List MACALASTER-BICKNELL COMPANY Cambridge, Mass. New Haven, Conn. Makers of Scientific Glass-ware MARINE BIOLOGICAL LABORATORY WOODS HOLE, MASS. SUPPLY DEPARTMENT BIOLOGICAL MATERIALS Plain Preserved and Injected Specimens Botany Specimens and Mounts Protozoan Cultures Drosophila Cultures Microscope Slides and Museum Preparations Our catalogues are sent promptly upon request, and we are also very glad to quote prices on any materials our customers might desire. We carry a number of speci- mens which are occasionally needed but which we do not list in our catalogue as there is not much demand for them. All materials are guaranteed to give absolute satisfaction. CONTENTS Page SHULL, A. FRANKLIN The Production of Intermediate-winged Aphids with Special Reference to the Problem of Embryonic Determination .... 259 PARKER, G. H. The Loping of Land-snails 287 HASLER, ARTHUR D. The Physiology of Digestion in Plankton Crustacea. II. Further studies on the digestive enzymes of (A) Daphnia and Polyphemus. (B) Diaptomus and Calanus 290 COONFIELD, B. R. Symmetry and Regulation in Mnemiopsis leidyi, Agassiz. . . . 299 DALCQ, A., AND G. VANDEBROECK On the Significance of the Polar Spot in Ripe Unfertilized and in Fertilized Ascidian Eggs 311 HYMAN, LIBBIE H. Reproductive System and Copulation in Amphiscolops langer- hansi (Turbellaria Acoela) 319 MILES, SAMUEL STOCKTON A New Genus of Hydroid and its Method of Asexual Re- production 327 HOPKINS, DWIGHT LUCIAN The Relation between Food, the Rate of Locomotion and Re- production in the Marine Amoeba, Flabellula mira 334 ABRAMOWITZ, ALEXANDER A. The Chromatophorotropic Hormone of the Crustacea 344 SCHECHTER, VICTOR Calcium Reduction and the Prolongation of Life in the Egg Cells of Arbacia punctulata 366 CHASE, HYMAN Y. The Effect of Ultra-violet Light upon Early Development in Eggs of Urechis caupo 377 CHURNEY, LEON AND HERBERT M. KLEIN The Electrical Charge on Nuclear Constituents (Sali ary Gland Cells of Sciara coprophila) 384 LOOSANOFF, VICTOR L. Development of the Primary Gonad and Sexual Phases in Venus mercenaria Linnaeus 389 LOOSANOFF, VICTOR L. Seasonal Gonadal Changes of Adult Clams, Venus mercen- aria ( L.) 406 FCX, DENIS L., H. U. SVERDRUP AND JOHN P. CUNNINGHAM The Rate of Water Propulsion by the California Mussel. . . . 417 MBl, WHOI LIBKAKY QIH 17IP E m