Morphology, Behavior, and Histogenesis of the Pelagosphera

Morphology, Behavior, and

Histogenesis of the Pelagosphera

Larva of Phascolosoma agassizii

(Sipuncula)

MARY E. RICE

m

SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY • NUMBER 132

SERIAL P U B L I C A T I O N S OF T H E S M I T H S O N I A N INSTITUTION

The emphasis upon publications as a means of diffusing knowledge was expressedby the first Secretary of the Smithsonian Institution. In his formal plan for the Insti-tution, Joseph Henry articulated a program that included the following statement:"It is proposed to publish a series of reports, giving an account of the new discoveriesin science, and of the changes made from year to year in all branches of knowledge."This keynote of basic research has been adhered to over the years in the issuanceof thousands of titles in serial publications under the Smithsonian imprint, com-mencing with Smithsonian Contributions to Knowledge in 1848 and continuing withthe following active series:

Smithsonian Annals of Flight

Smithsonian Contributions to Anthropology

Smithsonian Contributions to Astrophysics

Smithsonian Contributions to Botany

Smithsonian Contributions to the Earth Sciences

Smithsonian Contributions to Paleobiology

Smithsonian Contributions to Zoology

Smithsonian Studies in History and Technology

In these series, the Institution publishes original articles and monographs dealingwith the research and collections of its several museums and offices and of professionalcolleagues at other institutions of learning. These papers report newly acquired facts,synoptic interpretations of data, or original theory in specialized fields. These pub-lications are distributed by mailing lists to libraries, laboratories, and other interestedinstitutions and specialists throughout the world. Individual copies may be obtainedfrom the Smithsonian Institution Press as long as stocks are available.

S. DILLON RIPLEY

SecretarySmithsonian Institution

INTERNATIONAL BOOK YEAR • 1972

S M I T H S O N I A N C O N T R I B U T I O N S T O Z O O L O G Y • N U M B E R 1 3 2

Morphology, Behavior, andHistogenesis of the Pelagosphera

Larva of Phascolosoma agassizii(Sipuncula)

Mary E. Rice

SMITHSONIAN INSTITUTION PRESS

City of Washington

1973

ABSTRACT

Rice, Mary E. Morphology, Behavior, and Histogenesis of the PelagospheraLarva of Phascolosoma agassizii (Sipuncula). Smithsonian Contributions toZoology, number 132, 51 pages, 14 plates, 1973.—Morphology, behavior, andhistology are described for the pelagosphera larva of Phascolosoma agassiziiKeferstein. Observations are reported on histology, origin, and development ofthe following larval structures: body wall, coelom and coelomic elements, re-tractor muscles, digestive tract, buccal organ, lip gland, nervous system, nephridia,larval glands and terminal organ. Functional significance of certain structures,particularly organs associated with the mouth, is considered. Moreover, availableinformation on developmental histology of other sipunculans is reviewed andintra- as well as inter-phyletic comparisons are discussed. An account is givenof the general morphology of the larva and a few observations are noted onlarval locomotion and feeding behavior in the laboratory.

Official publication date is handstamped in a limited number of initial copiesand is recorded in the Institution's annual report, Smithsonian Year.

SERIES COVER DESIGN: The coral Montastrea cavernosa (Linnaeus) .

Library of Congress Cataloging in Publication DataRice, Mary E.Morphology, behavior, and histogenesis of the pelagosphera larva of Phascolosoma agassizii

(Sipuncula).(Smithsonian contributions to zoology, no. 132).Bibliography: p.1. Phascolosoma agassizii. 2. Larvae—Worms. I. Title. II. Series: Smithsonian Institution.

Smithsonian contributions to zoology, no. 132.QL1.S54 no. 132 [QL391.G5] 591'.O8s [595M7] 72-8959

For sale by the Superintendent of Documents, U.S. Government Printing OfficeWashington, D.C., 20402 - Price $1.25 (paper cover)

Contents

Page

Introduction 1Acknowledgments 2

Materials and Methods 2

Review of Early Development 3

Morphology and Behavior of the Larva 3

Larval Histogenesis 5Body Wall . 5Coelom and Coelomic Elements 7Retractor Muscles 7Digestive Tract 9Mouth and Associated Structures . 11Nervous System 14Nephridia 16Larval Glands 17Terminal Organ . 20

Comparisons with Other Phyla 21

Summary 22

Literature Cited 22

in

Morphology, Behavior, andHistogenesis of the Pelagosphera

Larva of Phascolosoma agassizii(Sipuncula)

Mary E. Rice

Introduction

As reported in an earlier paper (Rice, 1967), thedevelopment of Phascolosoma agassizii Keferstein isindirect with a pelagic, lecithotrophic trochophorefollowed by a prolonged pelagic, planktotrophicpelagosphera stage.1 The trochophore is character-ized by an apical tuft, a prominent equatorial bandof ciliated prototroch cells, mesodermal bands oneither side of a closed gut and a stomodaeum closedto the exterior by an overlying egg envelope. Meta-morphosis of the trochophore results in a secondlarval form, designated as a pelagosphera, in whichthe metatrochal ciliary band has replaced the pro-totroch as the chief locomotory organ, the mouthand anus have opened to complete the gut, the coe-lom has expanded into a spacious cavity and aterminal attachment organ has formed. In thispaper observations are reported on the morphologyand behavior of the pelagosphera larva of Phasco-losoma agassizii, bred from known adults andreared in the laboratory. The histology and de-

Mary E. Rice, Department of Invertebrate Zoology, NationalMuseum of Natural History, Smithsonian Institution,

Washington, D.C. 20560.

1A pelagosphera has been defined as a sipunculan larva,resulting from the metamorphosis of a trochophore, whichswims by means of a prominent ciliated metatroch and inwhich the prototroch either has been lost or has undergonea marked reduction (Rice, 1967:164).

velopment of the major organs and tissues of thelarva are described, and intra- and interphyleticcomparisons are discussed.

Sipunculan larvae were recognized as early as1850 when Max Mueller, working at Trieste, de-scribed and illustrated a larva which he presumedto belong to the genus Phascolosoma. A year later,in 1851, A. Krohn criticized this work, asserting thatthe larva depicted was not that of Phascolosoma,but rather belonged to Sipunculus nudus. In 1883,Hatschek published a treatise on the developmentof Sipunculus n.udus, describing in detail early de-velopment as well as the morphology and meta-morphosis of the larval stages. His material wascollected in the plankton at Messina, Sicily. Inter-est in larvae from oceanic plankton was elicited byMingazinni's (1905) description of a form whichhe erroneously identified as an adult, creating anew genus and species, Pelagosphaera aloysii. Thedescription was made from only one preserved andcontracted specimen which had been collected inthe Pacific between New Zealand and New Cale-donia at a depth of approximately 500 meters bythe planktonic expedition of the Italian Naval ship,Liguria. Working with material from the same ex-pedition, but from waters off India and Ceylon,Senna, in 1906, described similar forms. He con-cluded that they were larval forms and thus cor-rected Mingazinni's interpretation. Spengel (1907),in an independent investigation, also exposed

1

SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

Mingazinni's error. Comparing the larvae withthose of Sipunculus nudus in the Bay of Naples,he suggested that the structures which Mingazinniclaimed to be gonads might be instead glandularappendages of the esophagus. Thus Mingazinni'smistake was soon realized, but the name that hecreated, pelagosphera, remains well entrenched inthe literature and is today used to designate thelarval form of sipunculans which succeeds the stageof trochophore (Rice, 1967).

Other reports of planktonic larvae have beenmade by Heath (1910), Dawydoff (1930), Stephen(1941), Fisher (1947), Akesson (1961a), Damas(1962), Jagersten (1963), Murina (1965), Scheltemaand Hall (1965), and Hall and Scheltema (1966).Only two of these authors have attempted to assignadult specific affinities to the planktonic larvae.Fisher (1947) identified a larva collected off CapeHatteras as Sipunculus polymyotus on the basis ofthe number of longitudinal muscle bands. Murina(1965) designated larvae from the Gulf of Adenas Sipunculus aequabliblis and from the NorthwestPacific as Sipunculus norvegicus. Most of the ear-lier reports were based on preserved and contractedspecimens and it is only in some of the more recentstudies on living planktonic larvae (Jagersten,1963; Murina, 1965; Hall and Scheltema, 1966)that larval morphology, particularly of the headregion, has been correctly interpreted.

Four developmental patterns have been recog-nized within the phylum Sipuncula, only one ofwhich includes the long-lived planktotrophic pela-gosphera larvae of oceanic waters (Rice, 1967). Thefour categories of development are defined as fol-lows: (1) direct development with no pelagic stages,(2) indirect development with one pelagic larvalstage, the lecithotrophic trochophore, (3) indirectdevelopment with two pelagic larval stages, a leci-thotrophic trochophore and a short-lived lecitho-trophic pelagosphera, (4) indirect developmentwith a lecithotrophic trochophore and a long-livedplanktotrophic pelagosphera larva. Included in thislast category are the development of Phascolosomaagassizii, Sipunculus nudus, and unidentified spe-cies which have oceanic planktonic larvae.

Previous studies of sipunculan developmentwhich include significant accounts of histogenesishave been reported for 6 species, representing all 4developmental categories. The species are Golfingiaminuta, category 1, with direct development (Akes-

son, 1958), Phascolopsis gouldi and Phascolionstrombi in category 2 (Gerould, 1907; Akesson,1958), and Golfingia elongata and G. vulgaris incategory 3 (Akesson, 1961b; Gerould, 1907). Incategory 4 histogenesis has been described for onlyone species, Sipunculus nudus; developmental fea-tures of this planktotrophic species have been foundto differ markedly in several respects from the spe-cies in the first three categories with lecithotrophicdevelopment. The present account of histogenesisof the pelagosphera larva of Phascolosoma agassiziiprovides additional information on planktotrophicdevelopment in sipunculans and serves to broadenthe basis for comparative evaluation of develop-mental features within the phylum and for inter-pretation of affinities with other groups.

ACKNOWLEDGMENTS

This study represents part of a dissertation sub-mitted to the University of Washington in partialfulfillment of the requirements for the degree ofDoctor of Philosophy. The author wishes to ex-press her appreciation to Dr. Robert L. Fernald,Director of the Friday Harbor Laboratories, for hisadvice and encouragement and to thank him foruse of the facilities at the Friday Harpor Labora-tories. I am also deeply grateful to Dr. Paul Illgof the University of Washington for his many help-ful and valuable suggestions throughout the courseof this investigation. Support was provided, inpart, by a Public Health Service fellowship 5—Fl—GM-16, 344-03 from the National Institute of Gen-eral Medical Science.

Materials and Methods

Adult specimens of Phascolosoma agassizii werecollected from the San Juan Archipelago on thenorthwestern coast of the state of Washington andfrom Vancouver Island, British Columbia. Theywere found in intertidal habitats either under rocksor in a coarse gravel mixed with sand. Collectionswere made just previous to or during the breedingseason which lasts from the middle of June to lateAugust or early September (Rice, 1967). Animalswere maintained at a temperature of 10° to 12° Cin glass dishes in which the sea water was changedonce or twice daily. When spawning occurred, thefertilized eggs were removed and placed in tall

NUMBER 132

covered glass petri dishes in which the embryos andlarvae were reared. Dishes with developmentalstages were maintained at 10° to 12° C on the seawater table or in a refrigerated incubator; waterwas changed daily for early embryonic stages, butless frequently for later larval stages. Food for thelarvae was supplied by additions of unfiltered seawater, supplemented by mixed diatoms and dino-flagellates collected from tidal pools.

For histological studies, unless noted otherwisein the text, larvae were fixed in ice-cold 2 percentosmium tetroxide, buffered with 0.2 M S-collidineat a pH of 7.5 (Bennett and Luft, 1959) and em-bedded in Epon according to the method of Luft(1961). One or two micron sections were cut with

glass knives on a Porter-Blum ultramicrotome, af-fixed to glass slides and stained with Richardson'sstain (equal parts of 1 percent azure II and 1 per-cent methylene blue made up in 1 percent sodiumborate) for examination with the light microscope(Richardson et al., 1960). Additional stains forspecific structures are mentioned in the text. Beforefixation the larvae were relaxed in a mixture ofsaturated aqueous chloretone and sea water (1:2)(Clement and Cather, 1957).

Review of Early Development

The egg of Phascolosoma agassizii, ovoid inshape, but somewhat flattened sagittally, is rela-tively low in yolk content and undergoes spiral,unequal cleavage to give rise to a freely swimmingblastula with modified blastocoel (Rice, 1967). Gas-trulation occurs chiefly by epiboly, although in-vagination plays a minor role, forming a narrowarchenteron (Plate la,b). Mesodermal bands, de-rived from two large mesodermal teloblasts, aresituated on either side of the endoderm cells whichare distinguished by their small green cytoplasmicgranules. At the site of the closed blastopore, thestomodaeum is formed by an invagination of rap-idly dividing ectodermal stomatoblasts. From therosette cells in the center of the apical plate, thecilia of the apical tuft extend through the poresof the overlying egg envelope and the embryo, atits equator, is encircled by two equatorial rows ofciliated prototroch cells (Plate \a-d). A remnant ofthe blastocoel persists on the inner side of the pro-totroch cells as the "prototrochal cavity" into which

yolk granules from the prototroch are dischargedduring early development (Plates \c,d, 2a-d).

The trochopore stage extends from 2i/£ to 9 or10 days of age (Plates lc,d, 2a,b). In the early tro-chophore the essential size and shape of the eggenvelope are retained, but at approximately 5 daysof age gradual changes begin to take place. Thisperiod of gradual change in the trochophore, ex-tending over 4 or 5 days and leading up to themore rapid changes of trochophoral metamorpho-sis, is referred to here as premetamorphosis (Plates2a-d, 3a). During premetamorphosis alterations oc-cur in size and shape of the trochophore, a narrowcoelomic cavity is formed by the splitting of meso-dermal bands, and the differentiation of larval or-gans such as gut and retractor muscles is initiated.Metamorphosis of the trochophore into the pela-gosphera larva occurs at the age of 9 or 10 daysand lasts for a period of 12 hours or less at 10° C(8 days at 12° C). It is marked by a rapid increasein length (30 percent), expansion of the coelom,opening of mouth and anus, reduction of the pro-totroch, formation of the metatroch, and emergenceof the terminal organ (Plate Sac).

In the following report on morphology, behavior,and histogenesis of the pelagosphera larva of Phas-colosoma agassizii, descriptions, unless otherwise in-dicated, will refer to the larva as it appears at 20days of age. Larvae have been maintained in thelaboratory up to an age of 7 months, with relativelyfew observed changes other than an increase insize, a change in relative proportions, and a modi-fication of cuticular patterns. Duration of larvallife under natural conditions is unknown. Trans-formation to the juvenile worm has not been ob-served and all efforts to collect larvae from theplankton for comparative studies have been unsuc-cessful.

Morphology and Behavior of the Larva

The body of the pelagosphera larva of Phascolo-soma agassizii is marked by four distinct regions:head, thorax, elongated trunk, and terminal organ(Plates 5c, 4a-f, 5). The average length of the lar-val body at 20 days of age is between 400 and 500microns; by stretching or contracting, a given larvamay increase or decrease this length as much as 50percent. The average maximum width is approxi-mately 100 microns.


The head of the larva is covered by cilia on theventral surface and divided into two ventral lobesby a median ciliated groove which leads into themouth. The groove begins as a wide channel atthe tip of the head and becomes narrower as it ap-proaches the mouth. Below the mouth is a ciliatedlobe (Plates 4c,d, 5, 8b) which will be referred toas the "lip" in conformity with Jagersten's (1963)terminology for pelagosphera larvae. The lip usu-ally projects out at right angles from the ventralsurface of the head, but when the head is with-drawn into the trunk, the lip is pulled in parallelto the head. When the larva is feeding on the sub-stratum this lip may be pushed back so that it isflattened out against the substratum in the sameplane as the ventral surface of the head. The dorsalsurface of the head lacks ciliation with the excep-tion of the prototrochal cilia which are now re-duced in size and number. Located in a dorso-lateral position about half of the distance downthe length of the head are a pair of eyespots, hem-ispherical in shape and of a deep orange pigmenta-tion (Plates 4b, \\b,d). The entire head may bewithdrawn into the trunk when the larva is inac-tive or when it is startled by some stimulus suchas a movement of the surrounding medium.

Not only may the head be withdrawn into thetrunk, but also the region immediately posterior tothe head, known as the "thorax," may be with-drawn when introversion is complete. In the ter-minology of Jagersten (1963) the head plus thethorax make up the introvert. Hence the thoraxmay be defined as that part of the larval body ad-joining the head which is retractable into theposterior portion of the larva. It includes the meta-troch and is characterized by a thinner cuticle thanthat of the trunk.

The trunk includes that portion of the larvalbody posterior to the thorax with the exception ofthe terminal organ which is considered to be a dis-tinct region of the body. When extended the trunkis wedge-shaped, being broadest anteriorly and be-coming increasingly narrow toward the posteriorend. It is at the same time curved so that thedorsal surface is slightly convex and the ventral sur-face concave (Plates 3c, 4a,c,f, 5). The anus isclearly visible on the dorsal surface of the trunk,usually about halfway along its length, but its rela-tive position varies according to the degree of bodyextension (Plates 4a,c, 5). Other openings, not so

easily observed, occur in the proximity of the «malopening. Slightly anterior to the anus in ventro-lateral positions are the two nephridiopores. Dor-solaterally, on either side of the anus, are theopenings to the pair of large sacciform glands(Plates 4b, 14a-c). Three pairs of dorsolateral papil-lae, each papilla bearing a single bristle, appearbetween 10 and 20 days. The most anterior pairis in the midtrunk region at the level of the stomachand the other two pairs are located in the posteriorthird of the trunk (Plate \4d). When the larvahas reached the age of 28 days the cuticle of thetrunk is covered over its entire surface with nu-merous small papillae (Plates 4c,e, 5, 7b-d, IQd,13d). These cuticular papillae must not be con-fused with the epidermal papillae described above.The cuticular papillae are responsible, in Jager-sten's (1963) terms, for the "rough" appearance ofthe body surface. On the basis of the appearanceof body surface, Jagersten divided the pelagospherawhich he examined into two species: "rough" and"smooth."

From the posterior end of the trunk there ex-tends a narrow appendage, usually 50 microns inlength, but extensible to a length of 170 microns.This appendage is known as the terminal organ(Plates 3r, 4a,c,e,f, I4e,f). Such an organ has beendescribed by Hatschek (1883) in the larva of Sipun-culus nudus and has been noted by Jagersten (1963)in some unidentified pelagosphera. The terminalorgan can be retracted into the posterior end of thetrunk and this is the usual position when the larvais swimming or when it is quiescent. The organmanifests adhesive properties and its primary func-tion seems to be the attachment of the larva to thesubstratum.

The larva of Phascolosoma agassizii begins tofeed as soon as it completes metamorphosis. At thisstage it is chiefly a bottom feeder, although it isalso able to feed on small planktonic organisms.Most commonly the larva attaches by the terminalorgan to some debris on the bottom of the con-tainer and from this point of attachment it maystretch out great distances, either parallel to thebottom or upward at any angle. One frequent pat-tern of feeding is for the larva to stretch out alongthe bottom and then, with mouth applied to thesubstratum, to pull the head backward toward thepoint of attachment; the body is arched as the headprogresses toward the terminal organ and the

NUMBER 132

mouth takes in diatoms and various sorts of debrisas it moves along the bottom. The larva may alsoassume an arched position as it sweeps around theattached terminal organ in a complete circle. Theciliated ventral surface of the head is flattened uponthe bottom, the lip bent backward, while the larvagrazes over the substratum and clears the debrisfrom the area surrounding it. From time to timethe larva may release itself from its posterior at-tachment and swim with the ventral surface of thehead applied to the substratum and the terminalorgan pointing upward. In this manner it can covermore territory than would be possible in a fixedposition, or this may be only a means of finding anew location for settling. The larva may also moveabout by swimming along the bottom in an archedposition with the terminal organ dragging alongbehind, unattached; or it may crawl along in themanner of an inch worm. At other times it mayswim freely through the water, propelled by theactivity of the metatrochal cilia, or it may remainquiescent on the bottom, unattached with the intro-vert retracted. The larva appears to feed for themost part when it is attached with the introvertextended. Food intake is apparently accomplishedby ciliary activity on the ventral surface of thehead and by the occasional eversion of the buccalorgan. Possible functions of the buccal organ willbe discussed in more detail in the treatment of thehistogenesis of this organ (pp. 11-13). Occasionallythe larva bends both the head and the posterior endin a ventral direction so that the lip and terminalorgan touch, or else the terminal organ is pushedpast the lip to one side of the head. Whatever maybe accomplished by this activity is not obvious; infact, the activity may be abnormal since it is mostlikely to occur when there is no substratum onwhich the larva may settle. It is possible also thata secretion is transferred from the posterior sacci-form glands or from the terminal organ to the lip,or from the lip to the terminal organ. Jagersten(1963) reported that the terminal organ was pushedinto the mouth, but in Phascolosoma agassizii thiswas not observed.

The metatrochal cilia function mainly in locomo-tion. When the larva is attached, these cilia aid inthe movement of the anterior end around the pointof attachment, but they do not beat continuously.When the introvert is retracted the metatrochalcilia usually cease to beat.

Larval Histogenesis

The description of histogenesis in each organ orstructure in the larva of Phascolosoma agassizii willbe preceded by an account of its histology and fol-lowed by a comparative review of the informationavailable on the histology and development of thestructure in other sipunculans. The larval struc-tures to be described are the following: body wall,coelom and coelomic elements, retractor muscles,digestive tract, buccal organ, lip gland, nervoussystem, nephridia, larval glands, and terminal or-gan. In addition to the descriptive considerations,functional significance of many of the organs willbe discussed.

BODY WALL

Cuticle.—The cuticle of the larva is distinctivein each region of the body; hence the four regionswill be discussed separately.

In the head there are two distinct types of cuti-cle: the dorsal cuticle differs from the ventral andhas a separate origin from it (Plates 5b,c, lid). Thedorsal cuticle of the head is very thin and no struc-ture can be resolved within it. It is derived fromthe inner layer of the vitelline envelope, the outerlayers being sloughed off soon after metamorphosis.The dorsal cuticle stains very lightly with Richard-son's stain as does the ventral cuticle and both re-act positively to the periodic acid-Schiff stain. Theventral cuticle, penetrated by numerous cilia, isslightly thicker than the dorsal and is divided intotwo layers by a darkly staining central striation. Itoriginates from the lining of the stomodaeum andbears no affinity to the original vitelline envelope.

The cuticle of the thorax is structurally the sameas that of the trunk, but it is much thinner (Plate5). There is no sharp demarcation between thetwo, but rather a gradual transition. This cuticle,composed of two layers, is derived from the vitellineenvelope (Plate 13b). The inner layer, wider andmore lightly staining, represents the combined in-ner and middle layers of the vitelline envelope andthe outer layer of the cuticle is derived from theouter layer of the envelope. The inner layer givesa faint positive reaction to the periodic acid-Schiffstain and the outer layer is characterized by meta-chromasia.

Between 3 and 4 weeks of age small papillae form


in the cuticle of the trunk (Plates \c,e, 5, Ib-d,13d, He). These are filled at their base by the sub-stance of the inner layer, but within the apex,staining bright blue with Richardson's stain, a newsubstance makes its appearance. The papillae arecovered externally by the outer cuticular layer.

The cuticle of the terminal organ is identical tothat of the trunk and continuous with it. It differsonly in that it is much thinner.

The fate of the egg envelope and the formationof the cuticle are not uniform in all species ofsipunculans. Gerould (1907) observed the sheddingof the egg envelope in Phascolopsis gouldi andGolfingia vulgaris at the time of metamorphosis.The envelope ruptured at the postoral circlet ofcilia and the entire posterior portion was immedi-ately lost. Anteriorly the envelope remained at-tached longer but was eventually sloughed off bythe repeated introversions of the head. A fine gran-ular cuticle beneath the egg envelope became thelarval cuticle after the shedding of the envelope.Selenka (1875) described a transformation of theegg envelope into the larval cuticle in Golfingiaelongata. In the light of his own observationsGerould doubted the validity of Selenka's descrip-tion, but Akesson in 1961(b) confirmed Selenka'soriginal report. Akesson (1958) also had reportedthat the egg envelope in Phascolion strombi andGolfingia minuta was distended to form the larvalcuticle. Beneath the larval cuticle he observed asecond narrow cuticle, presumably the definitivecuticle of the juvenile form. The egg envelope, al-though distended, did not become elastic and whenthe larva contracted the envelope bulged out fromthe body. Bulges in the larval cuticle of Phasco-losoma agassizii have been seen only in abnormallarvae or in larvae that have been subjected torelaxants and fixatives. The egg envelope in P.agassizii is transformed into an elastic larval cuticleand no second cuticle has been observed.

In Sipunculus nudus the egg envelope is cast offat the initial metamorphosis along with the rem-nants of the prototroch cells (Hatschek, 1883). Inthis species the relationship of the egg envelope tothe embryo is quite different from that in othersipunculans (Gerould, 1903). The trochoblastsspread around the inside of the egg envelope form-ing a complete envelope or "serosa" surroundingthe developing embryo. The cells degenerate and

are shed with the egg envelope at the beginning ofthe pelagic larval stage.

Epidermis.—In the larva the epidermal cells aremarked by their vacuolar cytoplasm and large nu-clei with homogeneous, darkly staining nucleoplasmand prominent nucleoli (Plates 3c, 5,7c). Very finecytoplasmic extensions of the epidermal cells pene-trate into the cuticle.

The epidermal cells of the trunk are derivedlargely from the descendants of the second cell. Inthe early stages these cells are filled with yolk, butby the end of the trochophore stage much of theyolk has disappeared. After the formation of thecoelom, large intercellular spaces appear in the epi-dermis between the points of the epidermal attach-ment to the peritoneum (Plate 3a). As the cellsexpand into these spaces, their cytoplasm becomesintensively vacuolated (Plate 36). At the metamor-phosis of the trochophore to the pelagosphera theepidermis is stretched out into a much thinnerlayer (Plate 3c).

The epidermal glands will be considered in alater section on larval glands (pp. 18-19).

Muscles of the Body Wall.—The muscles of thebody wall are small and inconspicuous and are noteasily identified in any stage that has been studied(Plate lb,c). In the larva the longitudinal musclesappear in sagittal section as long, very fine, darklystaining fibers on the outer side of the peritoneumand the circular muscles are represented by smalldarkly staining dots, apparently attached to theouter side of the longitudinal muscles. The adultarrangement of longitudinal muscles into bundlesof fibers is not discernible in the larva.

Immediately posterior to the metatroch there isa broad circular sphincter which increases inbreadth with age (Plate 76). In a larva of 2 monthsthis sphincter may reach from the metatroch to thelevel of the stomach. At fixation it often contracts,pushing the stomach into an abnormal anteriorposition. The normal function of the sphincter isto close the anterior opening of the body after thehead has been retracted. In sagittal section thesphincter is seen as a number of relatively largespherical fibers well separated from each other andlocated to the inside of the epidermal cells.

Nothing can be said concerning the origin ofthe muscles of the body wall. In prelarval develop-ment small nuclei, resembling those of muscle, areseen scattered among the epidermal cells, but it has

NUMBER 152

not been possible to demonstrate whether thesegive rise to the circular and longitudinal muscles.

There is little agreement among other authorsas to the derivation of the muscles of the bodywall. Gerould (1907) reported the origin of thecircular muscles in Golfingia vulgaris and Phascol-opsis gouldi to be ectomesodermal, but he does notmention the source of the longitudinal muscles.Hatschek (1883), on the other hand, believed thatall of the body wall musculature in Sipunculusnudus was derived from endomesoderm. In hisstudy of Golfingia minuta and Phascolion strombiAkesson (1958) found that the circular muscleswere formed from ectomesoderm and the longi-tudinal muscles from endomesoderm.

Peritoneum.—The peritoneum lines the coelomof the larva, covering the inner side of the bodywall and continuing around the esophagus andrectum to surround the gut. The cytoplasm of eachperitoneal cell extends out from the nucleus as athin membrane which joins that of neighboringcells to form the peritoneal lining. The nucleiof the peritoneum are distinguished from those ofthe epidermis by their pronounced chromatin gran-ules and their lightly staining nucleoplasm.

The peritoneum is derived from the lateralbands of mesoderm which in turn arise from theearly mesodermal teloblasts (Plate 16). The lateralbands spread around the embryo and split intoinner and outer layers, splanchnic and somatic,respectively, to form the coelom (Plates 2b-d, la).The splanchnic mesoderm surrounds the gut andthe somatic forms the inner lining of the body wall.At first the nuclei of the peritoneum are very closetogether (Plate 3a), but with the elongation ofthe body they are pulled apart and the cytoplasmis stretched out into a thin mesothelium (Plate 3c).

The ciliated peritoneal cells of the adult havenot been identified in the larva.

coelomic cells which are moved about by the nearlycontinuous contraction and expansion of the larvalbody. There are two types of coelomocytes (Platesla-d, 14c). One is spindle-shaped with occasionalvariations in form that suggest amoeboid activity.The cytoplasm of this cell type stains darkly andin older larval stages it is packed with basophilicgranules. The second cell type is round and flatand the cytoplasm in younger larvae stains lightlywith relatively few inclusions. At two months thecytoplasm of the second cell type is characterizedby clusters of granules and vacuoles and the form-erly central nucleus is in an eccentric position(Plate 7c).

In the premetamorphic larva, cells resemblingthe second cell type appear to be budding off fromthe splanchnic peritoneum into the newly formedcoelom (Plate 7a). These coelomocytes increase innumber with age, but there is no evidence for theirlater derivation or for the derivation of the spindle-shaped cells. The round, flat cells of the larva aresimilar to and may be identical with the red cellsof the adult, whereas the spindle cells may representthe small amoebocyte of the adult. There are 3types of adult coelomocytes, however, that havenot been recognized in the larva. They are thelarge amoebocytes with large granules, the vesicles,and the urns.

The coelomocytes described by Gerould (1907,pi. 9: fig. 80b, c) in Golfingia vulgaris and Phascol-opsis gouldi are very similar to the round flat cellsof Phascolosoma agassizii. Although he presentedno evidence for their origin, Gerould suggestedthat they were probably derived from the peri-toneum in the posterior coelom. In addition tothese coelomocytes, Gerould also found within thecoelom 19 cells with large nuclei and scant cyto-plasm, which he believed to be the remains of thediscarded prototroch cells.

COELOM AND COELOMIC ELEMENTS

The expansive coelom of the larva extends with-out interruption from the most anterior part of thehead into the terminal organ and it is open to theexterior only by way of the nephridial organs.The coelom is formed between 6 and 7 days by asplitting of the splanchnic and somatic layers ofmesoderm.

Within the coelom there are freely suspended

RETRACTOR MUSCLES

Spanning the coelom lengthwise are four re-tractor muscles, two dorsal and two ventral, whichfunction to withdraw the head into the trunk. Inthe early larva up until the age of one month eachmuscle is composed of eight fibers and each fiberis a single cell with only one nucleus. A cross-sectionof the muscle reveals a circular arrangement offibers joined together by a thin peritoneal mesen-

8 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

tery enclosing a central cavity (Plates 12a, 13b).The fibers of each of the dorsal retractors begin

in a dorsolateral position near the base of thebrain, attached in a straight row to the peritoneallining (Plate 13a). The most median fibers arisemore anteriorly than the outermost which ariseslightly posterior to the level of the mouth. Atthe level of the ventral lip the fibers and theirsurrounding peritoneum pinch off from the peri-toneal lining forming a circle with a central space(Plate 13fr). The muscle then extends posteriorlythrough the coelom to the first posterior bristlewhere it again attaches to the peritoneal lining ofthe dorsal body wall, fanning out into a row offibers. Just below the bristle the two outermostfibers insert into the body wall (Plate 14d). Twomore fibers are inserted at the level of the posteriorglands and the last four fibers in the region of theanus.

The fibers of the ventral retractors arise fromthe body wall on either side of the mouth near theemergence of the circumesophageal connectivesfrom the brain (Plate 13a). The four most anteriorfibers lie between the epidermis and the connectivesand it is only posterior to the connectives that thefibers come into contact with the peritoneum. Thefour anterior fibers are joined by four additionalfibers in the region of the buccal organ and theeight fibers plus the interconnecting peritoneumseparate from the peritoneal lining, forming acircular muscle similar to that of the dorsal re-tractor (Plate 136). The ventral retractors passposteriorly between the buccal organ and the lipglands to the level of the first bristle where each ofthe ventral retractors gives off two fibers, one closeto the ventral nerve cord and another more later-ally. The circle of fibers then flattens out andbecomes continuous with the ventral peritoneallining. At the level of the nephridia two more fibersare inserted (Plate 13c) and the four remainingfibers in each of the ventral retractors insert farposteriorly at the level of the second bristle, thusexceeding by far the dorsal retractors in length.

The nucleus of each muscle fiber is located an-teriorly along the length of the fiber and all of thenuclei of a muscle are relatively close together(Plate 6rf). In both dorsal and ventral retractorsthe nuclei are found at the level of the loweresophagus or upper stomach, although the ventralnuclei may extend slightly farther posteriorly.

The retractor muscles of Phascolosoma agassiziiappear to be derived from ectoderm, but it has notbeen possible in this study to identify with certaintythe rudiments in very early stages. At the time ofcoelom formation groups of cells with nuclei whichresemble those of the later muscles are foundbulging into the coelom in the position of the futureventral retractors (Plate 6a, b). These cells arelocated at the anterior end of the newly formingposttrochal coelom and appear to be connected bycytoplasmic extensions to the ectodermal cells ofthe stomodaeum and the ventrolateral apical plate,just above the prototroch. In the position of thefuture dorsal retractors the early coelom is narrowand a cluster of cells corresponding to that in theventral region has not been identified; however,a few cells of similar cytology have been recognizedjust anterior to the coelom. Within a period of 24hours after the ventral cell clusters first are seen,both ventral and dorsal retractors make their ap-pearance as long fibers with central nuclei whichextend from the lateral ectoderm of the apicalplate through the coelom to the posterior somaticmesoderm (Plate 6c). The ventral retractors withtheir very thick fibers are much larger when firstformed than the dorsal retractors with their rela-tively thin and tenuous fibers. Both ventral anddorsal retractors are intimately associated withperitoneal cells, the nuclei of which, althoughusually smaller and rounder, are not always easilydistinguished from those of the muscle.

The above observations are interpreted to signifythat the retractor muscles originate from four areasof ectomesoderm located in dorsolateral and ven-trolateral positions in the apical plate. Theserudiments grow downward, retaining their cyto-plasmic connections in the head, and push into thecoelom taking with them an investing layer ofperitoneum. Long cytoplasmic processes or fibers areextended through the coelom and attach posteriorlyto the outer layer of coelomic mesoderm.

If the observations of this study are complete andif the interpretations are correct, then the forma-tion of retractor muscles in Phascolosoma agassiziidiffers from that observed in studies of othersipunculans. Gerould (1907) in his study of Gol-fingia vulgaris and Phascolopsis gouldi stated thatthe muscles were apparent in the early trochophoreprior to the division of the coelomomesoblast, ex-tending from the sides of the apical plate, from

NUMBER 152

which they were derived, posteriorly to the regionof the postoral circlet. He suggested that the musclespassed through the loose layer of mesoderm at thetime of metamorphosis.

Hatschek (1883) proposed that in Sipunculusnudus the retractors were derived from somaticmesoderm. He was unable, however, to trace theirearly development and recognized them only afterthey were already formed within the coelom.

Confirming Gerould's observations, Akesson(1958, 1961a) found in three species of sipunculans,

Phascolion stro?nbi, Golfingia minuta, and Gol-fingia elongata, that the retractor muscles were de-rived from ectomesoderm and that they wereformed outside of the mesoderm in very early stagesbefore the formation of the coelom. In Phascolionstrombi and Golfingia minuta the development ofthe muscles was reported in detail and it wasobserved that each retractor had two rudiments,an anterior and a posterior, which grew towardeach other and fused to form the completed muscle.This manner of development could not occur inPhascolosoma agassizii where each fiber of the larvalmuscle is a single cell which extends the entirelength of the muscle.

The reported incongruities demand furtherstudies on Phascolosoma agassizii and a reinvesti-gation of Sipunculus nudus before conclusions canbe drawn as to the mode of development of theretractor muscles of sipunculans and the specificvariations.

The larval retractor muscles are retained as theretractor muscles of the adult sipunculan. In Phas-colosoma agassizii it has been demonstrated thatthe muscle fibers increase in number with the ageof the larva. In the larva of 28 days each muscle iscomposed of only eight fibers arranged as a tubewith a central cavity, whereas at 2 months thenumber of fibers has greatly increased, completelyfilling the central cavity (Plate 13d).

DIGESTIVE TRACT

The gut of the larva is regionalized into four dis-tinct morphological and cytological portions: thelong, broad esophagus with posterior sphincter, theexpanded and bulbous stomach, the thin, loopedintestine, and the short rectum.

Esophagus.—The epithelial lining of the esopha-gus is composed of cuboidal cells with dense, long

cilia which form a central swirl in the lumen(Plates 1b, 13d). The epithelium is lined with aninner cuticle which is continuous with that of theventral head and stains brightly with periodicacid-Schiff. The nuclei are relatively large and thecytoplasm appears to be homogeneous with smalllipid droplets scattered through it. On the outerside of the epithelium is a thin layer of musclefibers covered by peritoneum. The muscle layerincreases in thickness at the junction with thestomach, forming a tight sphincter (Plate 76).When the head is retracted into the body theesophagus is pushed into a lateral position.

The esophageal lining is formed from the stom-odaeal invagination and is therefore ectodermal inderivation. The invagination begins at approxi-mately 60 hours and by the end of the trochophorestage it has elongated in an anterodorsal directionand the cells have developed cilia (Plate Id). As thestomodaeal invagination sinks inward the anteriorend of the stomach moves into a more dorsal posi-tion. At the beginning of the premetamorphosisstage (6i/2 days) the invagination opens into thecavity of the stomach and, as elongation proceeds,the stomach is moved into a position directlyposterior to the esophagus (Plate 2b).

Stomach.—The stomach is very large and bulbousin shape and its maximum width is nearly equalto that of the coelom (Plates 36, 5). Its diametermay exceed that of the mid-esophagus two or threetimes and that of the esophageal sphincter overlour times. The lining of the stomach is constructedof a single layer of cuboidal epithelium enclosingan expansive central cavity and, in contrast to theesophagus and intestine, there is no obvious inter-vening muscle between the epithelium and thesurrounding peritoneum (Plates Ib-d, 10c, 126).The apical border of the epithelial cells is striatedas well as weakly ciliated, the cilia being much lessconcentrated and shorter than those of the esopha-gus (Plates 76, 126, c). Within the cytoplasm ofthe epithelial cells there are numerous lipid drop-lets of varying sizes, all larger than those of theesophagus. In larvae older than one month, thelipid droplets coalesce to form one huge sphereof lipid in each cell (Plate 7c, d). The remainderof the cytoplasm is highly vacuolated, giving theappearance of a reticulum. The epithelial nucleihave a large eccentric nucleolus and the nucleo-plasm, with prominent chromatin granules, stains


more lightly than that of the esophageal cells. Al-though the cuboidal epithelial cell is by far themost common, there is a second cell type that israther sparsely scattered throughout the epitheliumand is most readily observed in the older stages.This is a smaller cell, pyramidal in shape and withhomogeneously staining basophilic cytoplasm.

The larval stomach is derived from endodermand is the first portion of the gut to be differenti-ated. It can be distinguished at two and one-halfdays when characteristically small, green cytoplasmicgranules appear in the endoderm and a cavity, thefuture lumen of the stomach, appears in the midstof the endoderm cells. The cavity elongates as thecells of the posterior portion of the stomach under-go division. In the premetamorphosis stage the cellsof the posterior stomach give rise to the rudimentof the intestine which makes its appearance as asmall round knob at the end of the stomach (Plate2b).

In the early trochophore the green granules in-crease slightly in size and, along with the relativelydensely concentrated yolk granules, mark the cellsof the stomach as distinct from all others (Platela). In the late trochophore some of the yolkgranules coalesce and, when the coelom is formed,these large clusters of yolk are cast off into it.During premetamorphosis the remainder of theyolk disappears, its fate undetermined. The greengranules now vary in size; some are quite largeand are recognizable as lipid droplets. After meta-morphosis the small granules disappear and thelipid droplets increase in size and decrease innumber until there is in each cell only one verylarge sphere (Plate 7c, d). From this evidence it isassumed that the green granules described in theearly stages were in actuality small lipid inclusions.Whether these lipid inclusions are responsible forthe green coloration in the living larva or whetherthere is some other constituent that has not beendefined cytologically is not known. The two are,however, linked together by circumstantial evi-dence in that they both appear at approximatelythe same stage of development. The esophagus, al-though not colored in the living larva, does containsome lipid, but it is not present in the same con-centration as in the stomach and intestine, bothof which are marked by green coloration.

At the time of metamorphosis the cavity of thestomach increases in size and the bulbous shape is

assumed. In older larvae (2 months or more) thestomach enlarges ami elongates to form a longcylinder which may extend more than half thelength of the body. The fate of the larval stomachis unknown. Although it occupies a place of greatprominence in the larva, it has no morphologicalcounterpart in the adult. Since only those sipun-culan larvae with a prolonged pelagic life, such asSipitnculiis nudus (Hatschek, 1883) and other un-identified pelagosphera (Akesson, 1961a; Damas,1962), possess a well-defined stomach, it is likely thatthe stomach is of functional significance to a pelagicmode of existence.

Intestine.—The stomach narrows abruptly intothe thin intestine which in the larva up to age of60 days is looped but not coiled. There is a descend-ing portion which extends anteriorly toward thedorsal anus. As it leaves the stomach the intestinallining is composed of a relatively thick, ciliatedcuboidal epithelium with a densely basophilic cyto-plasm. Within a short distance the wall becomesexceedingly thin, the cells are flattened and thecytoplasm vacuolated (Plates 12rf, 13c). The in-testinal epithelium is ciliated and there are smalllipid granules scattered throughout its length. Thewall of the intestine is completed by a thin musclelayer and a peritoneal covering.

The intestine originates from endoderm whichgrows out as an extension from the posterior endof the stomach in the late trochophore stage (Plate2b). Growing dorsally it attaches to the dorso-posterior ectoderm in the position where the futureanus will form (Plate 106). At metamorphosis thereis an elongation of the body posterior to this pointof attachment, extending the coelom far belowthe anal region and as the intestine continues togrow it loops downward into the posterior coelom.From the time of its formation there are lipiddroplets dispersed throughout the cytoplasm of theintestine, but they never become as large or asconcentrated as those of the stomach.

Rectum.—The larval rectum is a short tubesurrounded by a thin layer of muscle and an outercovering of peritoneum (Plates 5, 10c). The epi-thelium of the rectum is very much thicker thanthat of the preceding intestine and the cells areheavily ciliated. The cytoplasm is basophilic withno visible inclusions other than yolk granules inthe early stages, and the inner border of the cellsstains strongly with periodic acid-Schiff, as does

NUMBER 1S2 II

that of the esophageal cells. The rectum continuesinto the anus and at the anal opening the larvalcuticle and its associated epidermis turn inward tomeet the rectal epithelium (Plates 5, 10b, c). Sur-rounding the anus there is a thick sphincter muscle.

The rectum is formed during premetamorphosisfrom a dorsoposterior ectodermal proliferation.The rudiment of the rectum is composed of only afew cells which are soon joined by the growingintestine (Plate 10b). A ciliated cavity forms amongthe ectodermal cells and before metamorphosis thisopens into the intestinal lumen. A slight depressionappears in the cuticle opposite the rectal rudimentand at this point the cuticle ruptures during meta-morphosis, forming the anus and thus opening therectum to the exterior.

In all species of sipunculans that have beenstudied the esophagus is formed from a stomodaealinvagination, the rectum from a cluster of ecto-dermal cells referred to as the proctodaeum, andthe remainder of the gut from the entomeres ofthe blastoporal region. In the different species ofsipunculans there is a great difference in the timeat which a lumen appears within the gut and thetime at which the gut is completed. Golfingia min-uta, with a large egg rich in yolk, remains lecitho-trophic for 2 months and Phascolion strombi forone month (Akesson, 1958). In Golfingia vulgarisand Phascolopsis gonldi a lumen appears withinthe archenteron at 5 or 6 days, but they do notbegin to feed until the second or third week ofage (Gerould, 1907). The gut of Sipunculus nudusis functional at 3 days (Hatschek, 1883) and that ofPhascolosoma agassizii at 9 or 10 days.

MOUTH AND ASSOCIATED STRUCTURES

The term mouth will be used in a broad senseto include the anterior ciliated channel of theventral head, the ventral, posterior lip, and thebuccal groove through which the buccal organmay be protruded.

The ciliated channel divides the ventral surfaceof the head into two lobes and continues into theesophagus, forming the dorsal and lateral bound-aries of the esophageal opening. The ventralboundary of the opening is formed by the unionof two nonciliated folds. The channel of the ventralsurface of the head is derived from the roof of thestomodaeum at metamorphosis.

Posteriorly the mouth is expanded into the ven-tral lip (Plates 5, 8b), a lobe which usually extendsout perpendicularly from the head, but which maybe withdrawn along with the head into the trunkor may, when the larva is feeding, be flattened outagainst the substratum in a 180° angle from thehead. The lip is partially bifurcated by a deepciliated groove which begins near the distal endof the lip and extends to a ciliated pore near theproximal end of the lip. The lip glands, to bedescribed later, open into this pore. Distally theciliation of the lip is limited to the central grooveand a few lateral cilia, whereas that portion of thelip proximal to the pore is evenly ciliated. Whenthe mouth opens at metamorphosis, the floor of thestomodaeum is extended outward to form the sur-face of the lip (Plates 3b,c, 5, 9d). The cuticle sur-rounding the lower portion of the mouth isstretched out with the stomodaeal extension to formthe sides of the ventral lip.

Buccal Organ.—At the base of the lip is a long,transverse slit, the width of the mouth, which ex-tends downward as a nonciliated infolding forsome distance along the ventral side of the esopha-gus (Plates 8«, b, We). The slit, which will betermed the buccal groove is expanded in the centerof its posterior extremity and ventrally it may showone or more folds. The entire invagination is linedby a very thick cuticle which is stained by periodicacid-Schiff. On the dorsal side, the cells of theepithelium are very thin and lie in close associationwith the esophagus along much of its length. Onthe ventral side of the infolding and extendingaround under the posterior end is the buccal organ;when the infolding is everted the buccal organ pro-trudes to the exterior.

The buccal organ hangs down into the coelomof the larva as a large muscular sac within whichthere are two main cavities: posterior-dorsal andanterior-ventral; the latter may be divided again,making a small third cavity. The organ is circum-scribed by a sheet of muscle which attaches ven-trally to the proximal portion of the lip and, loop-ing down around the posterior extremity of thebuccal groove, attaches dorsally to the muscle layerof the esophagus (Plates 3f, 5, 8a, b, \\b, c). Theepithelium of the organ lines the ventral side of thebuccal groove and, with its overlying cuticle, is thefirst portion to reach the exterior when the organis protruded. The epithelial cells are very elongate


with thin necks leading to rounded tips whichbulge out into the cavity of the organ. The parti-tion that divides the cavity into chambers is formedby an extension of epithelial cells across the cavity,connecting to the outer muscle wall. The cytoplasmof these cells stains lightly and appears to befibrous (Plate 86).

The formation of the buccal organ is not entirelyunderstood. The outer muscular layer is formedfrom cells which grow downward from the sides ofthe anterior esophagus and posterior stomodaeumand join beneath in a solid mass (Plate 9c,d). Thecells are first recognized at the beginning of the pre-metamorphosis stage and by the end of metamor-phosis they have spread out as a single layer toform a muscular sac which hangs down from theesophagus. It has not been possible to discernwhether the cells originate from the ectoderm ofthe apical plate above the esophagus or whetherthey are derived from the mesoderm along thesides of the esophagus. The buccal groove has notbeen observed until the day preceding metamorpho-sis. At this time it extends downward as an invagi-nation from the floor of the stomodaeum, but incontrast to the stomodaeal lining, it has a heavycuticle and lacks ciliation (Plate 9d).

The buccal organ was first described in 1883 byHatschek in the larva of Sipunculus nudus. Hats-chek referred to this organ as the "Schlundkopf," aword which can be translated as the upper pharynx,esophagus, or buccal mass. The structure wasdescribed as a fold in the esophageal wall whichextended posteriorly as a groove, then curvedupward with the two layers of the fold tightly ap-posed to surround a round vesicle (Hatschek, 1883,pi. 5: fig. 57a). The lumen of the vesicle was tra-versed by spindle-shaped cells. The vesicle wasreported to arise from mesoderm and the fold froman ectodermal evagination of the esophagus. InPhascolosoma agassizii the fold does not continuearound the organ, but instead the organ is enclosedby a muscular sac.

In a study of pelagosphera collected by the DanaExpedition, Damas (1962:13—14) again describedthis organ and called it the "machoire." Damas'representation of the organ agrees in its essentialaspects with the findings in P. agassizii. He de-scribed a transverse groove in the posterior mouthwith a thick cuticle, covered ventrally by a uniqueepithelium of tall, cylindrical cells. Beneath the

epithelium was an oblong transverse mass of turgidtissue composed of muscle fibers and within thismass there were numerous cavities. The histologicaldetails of the turgid tissue as presented by Damasmay not be entirely reliable since his specimenswere fixed aboard ship in diluted formalin 30 yearsprior to his investigation.

Jagersten (1963) observed the organ in livingpelagosphera and referred to it as the "pharyngealbulb." Although he noted that it was in a positionat the base of the lip, he gave no information onthe structure of the organ.

The term buccal organ has been used in thepresent study because it does not imply a molluscanhomology as does "buccal mass" (Hatschek, 1883),nor does it attribute a speculative function as isimplicit in the term "machoire" (Damas, 1962).Jagersten's term, "pharyngeal bulb," is not appro-priate, since there is no differentiated pharynx inthe larva and, even if there were, this term wouldconnote a muscular expansion in the pharyngealtube, rather than a saccular evagination.

Investigators have pondered the possible func-tions of the buccal organ, but thus far no one hasprovided a conclusive demonstration of its signi-ficance in larval life. Hatschek's (1883:39) onlycomment on the function was "Die Function diesessogenannten Schlundkopfes ist mir ziemlich rath-selhaft geblieben." Although he did not observe theeversion of the organ, Hatschek noted that pres-sure on an overlying coverslip resulted in the ex-trusion of both the esophagus and the "Schlund-kopf." Damas (1962) did not examine living larvae,but he proposed that the organ might act as arocker, pushing food accumulated in front of itback into the esophagus, Jagersten (1963) observedin the living pelagosphera a "kneading" action ofthe buccal organ and he supposed that such an ac-tion might serve to break up large pieces of foodinto smaller pieces; however, he did not give evi-dence to support this supposition.

In the present study a cinematographic record ofa larva of Phascolosoma agassizii was made as itmoved along a filamentous diatom, continuallyeverting the buccal organ. The diatom was of toogreat a length to be ingested intact by the larva.The mouth of the larva was applied to the diatom,the body arched, and the terminal organ attachedto the substratum. At each eversion of the buccalorgan the head of the larva moved forward. After

NUMBER 1S2 13

a series of eversions the diatom strand was bentinward, appearing to be directed into the esopha-gus. In a final eversion the strand was pushed outof the mouth and the larva abandoned the diatom.From these observations it would appear that thebuccal organ is not only able to push food into theesophagus as proposed by Damas (1962), but alsoit can reject objects, pushing them away from theesophagus. The latter action might be of value iflarge and extraneous objects adhere to the pre-sumably sticky secretion of the lip gland. The ex-trusion of the buccal organ, moreover, blocks theesophageal opening and, if living organisms werebeing ingested, this motion might prevent theirescape. Another function, possibly performed bythe continual eversion of the buccal organ againsta substratum, could be the loosening of smalldiatoms and debris which would then be sweptinto the digestive tract by the directed ciliary move-ment. Although a function in the feeding processseems to- be the primary significance of the buccalorgan, it may also play some role in locomotion.The fact that the head moves forward when thebuccal organ is applied to the substratum suggeststhis possibility. On the other hand, the head canalso move forward by means of ciliary activityalone. A possible sensory function cannot be elim-inated even though the heavy cuticle of the organmakes such a function seem improbable. An in-nervation has not been demonstrated, but the con-nectives from the brain are very closely associatedwith the organ at either end of the buccal groove(Plate 12a).

Lip Gland.—In the larva of Phascolosoma agas-sizii the lip gland, composed of four pendulouslobes, hangs down from the inner surface of the lipinto the anterior coelom (Plates 36, c, 5, 8a, 6).The two median lobes of the gland are approxi-mately 50 microns in length and are slightly longerthan the two lateral lobes. Each lobe, suspendedfrom the lip by a long neck, ends in a bulbousexpansion. At their point of attachment to the lip,the necks of the lobes surround the lip pore (Plate8/;). Because of this arrangement it is inferred thatthe glands discharge a secretion into the pore.

In larvae up to an age of 20 days each lobe ofthe lip gland is composed of a single large cell, butin later larvae the number of cells increases so thata single lobe may consist of several cells. Each cellis characterized by a large nucleus with an ex-

ceptionally large nucleolus and a nucleoplasm thatstains deep blue with Richardson's stain. Thecytoplasm is highly vacuolated and penetrated bya fine reticular network (Plate 86). Large patchesof darkly staining material are scattered throughthe cell, sometimes concentrated at the peripheryor around the nucleus, and from these patchesthick strands of similar material radiate into thesurrounding cytoplasm. In larvae that have beenembedded in methacrylate the gland gives a nega-tive reaction to periodic acid-Schiff and alcian blue,but with safranin O small areas within the glandstain metachromatically.

The lip gland is derived from ectoderm. At thebeginning of the premetamorphosis stage, fourlarge cells are distinguishable among the cells ofthe ventral wall of the stomodaeum by their largesize, basophilic cytoplasm, and large nuclei eachwith a single prominent nucleolus (Plate 9a). Asthe cells increase in size the cytoplasm becomesprogressively vacuolated; by the time of metamor-phosis the characteristic cytology of the larval glandhas been assumed (Plates 36, c, 9b). At their firstappearance the cells are widely scattered and atthis time the lip pore cannot be identified. Theprocess by which contact is established between thegland cells and the lip pore is not understood.

The lip glands were first described by Max Muel-ler (1850) in a larva which he presumed to belongto the genus Phascolosoma. The glands were des-ignated as the unpaired organ and described as abilobed body opening into a ciliated excretory ductbeneath the introvert. The significance of the organwas not clear to Mueller, but he suggested that itwas either a secretory gland or a gonad. Later, in1883, Hatschek depicted the gland as the "Anhangs-druse" in the larva of Sipunculus nudus and re-ported it to be an esophageal invagination, the cellsat the blind end of which grew toward the coelomto form the glandular masses. The gland was foundto be composed of two glandular appendages, eachwith tall prismatic cells that ended at the lumenof a ciliated excretory duct. The duct opened nearthe posterior wall of the mouth. Mingazinni (1905)figured the glands in a pelagosphera larva that hadbeen collected at a depth of 500 meters in the SouthPacific between New Caledonia and New Zealand.Unaware of Hatschek's work, he believed the glandsto be gonads and therefore considered the animalto be an adult sipunculan, creating a new genus


and species, Pelagosphaera aloysii. Senna (1906) ina study of other planktonic sipunculan larvae,recognized the glands to be the esophageal evagin-ations described by Hatschek. In pelagosphera lar-vae from Monterey Bay, California, Heath (1910)found similar structures and identified them aslarval glands, rather than gonads. In his examina-tion of 1,900 pelagosphera from the Dana Expedi-tion, Damas (1962) reported a typical bilobed glandsuspended in the coelomic cavity from the end ofa ciliated canal. The canal opened to the exterioron the ventral surface of the anterior introvert. Thecells of the gland were arranged in the shape ofa fan around a central cavity, a continuation of theciliated canal. Akesson's (1961a) description of theglandular organ of four pelagosphera collected bythe Galathea Expedition agrees essentially with thatof Damas. The term lip gland was proposed byJagersten (1963) who was the first to observe thatthe duct of the gland opened onto a ventral lobeof the mouth which he designated as the lip. Jager-sten's observations have been confirmed here in thelarva of Phascolosoma agassizii; thus his term, lipgland, has been adopted in this report.

The structure of the lip gland of P. agassiziidiffers from that described in the literature in thatit is divided into four appendages, all of whichconverge at a pore on the base of the lip. There isthus no long "excretory duct" as found in otherpelagosphera and there is no central cavity withinthe gland. Rather than numerous cells around acentral cavity, in the early larva of P. agassizii thereare only four cells in the gland, each lobe con-sisting of a single cell suspended from the regionof the pore.

There is very little information available on thenature of the glandular secretion. Akesson (1961a)reported a basophilic staining with the Azan stainand Damas (1962) demonstrated a weak reactionwith periodic acid-Schiff. In the gland of Phascolo-soma agassizii there was no reaction with periodicacid-Schiff, but scattered patches of cytoplasm mani-fested an intense basophilia with Richardson's stain.Areas of the gland were stained a bright orangewith safranin 0, a reaction characteristic of mucus(Li I lie, 1951:285). The results are confusing, how-ever, in that there was no manifestation of meta-chromasia with Richardson's stain and there wasno reaction to alcian blue. Further work is neces-sary before the nature of the secretory product can

be denned. If it should be mucus, then such asecretion might function in a ciliary-mucus patternof feeding or in adhering the ventral surface of thehead to the substratum as the larva moves along insearch of food.

Both the buccal organ and the lip gland arelarval organs with no known fate in the adult. Akes-son (1961a) hypothesized that the lip gland ishomologous to the ventral sensory organ of theadult in the genera Phascolosoma, Siphonosoma,and Aspidosiphon. The ventral sensory organ hasbeen described as a tubular organ which opens onthe ventral midline beneath the collar and termin-ates at the anterior end of the ventral nerve cord.In some species the termination is expanded into acavity with a central statocyst (Akesson, 1958). Theventral sensory organ in the adult of Phascolosomaagassizii lias been found in the present study to bea simple, narrow, nonciliated tube ending blindlyat the anterior end of the ventral nerve cord at thepoint of junction of the circumesophageal connec-tives. In the larva the position of the lip pore issimilar: the connectives from the brain approximateeach other on either side of the lip pore and con-tinue along the sides of the lip groove. The trans-formation of the larva to the adult form has notbeen observed in this study; hence it has not beendemonstrated whether the lip pore and lip grooveof the larva are in any way related to the sensoryorgan of the adult.

NERVOUS SYSTEM

The central nervous system of the larva of Phas-colosoma agassizii consists of a cerebral ganglion,circumesophageal connectives, and an unsegmentedventral nerve cord.

The cerebral ganglion or brain is located dorsallyin the head in close association with the overlyingepidermis (Plate 1 la, b). Localized in the mostdorsal portion of the brain are small nuclei withdense chromatin granules and a lightly stainingnucleoplasm. Ventral to the nuclei is a mass ofcytoplasmic processes, the neuropile of the brain(Plate Mb). Two pairs of nerves originating fromthe neuropile have been identified: one pair ex-tends to the dorsoanterior surface of the head andanother pair to the esophagus. Embedded laterallynear the surface of the head in the posterior portionof the brain is a pair of eyes, each composed of

NUMBER 132 15

a hemispherical pigment cup which opens in ananterolateral direction (Plate lib, d). The opticnerve has not been identified.

The circumesophageal connectives leave thebrain near the level of the eyes and extend ventrallyto either side of the ciliated channel of the mouth(Plates 12a, 13a). They pass posteriorly along thechannel, then laterally around the buccal grooveand, finally, ventrally on either side of the lip poreand lip groove. Beyond this point they become verythin and their union has not been observed inyounger larvae. The connectives are characterizedby a few fibers and many small nuclei, similar tothose of the brain. Usually intimately associatedwith the epidermis or esophagus, the circumesoph-ageal connectives are not well-defined structures.

The circumesophageal connectives are joined atthe distal end of the lip by the ventral nerve cord(Plates 3b, c, 5, 8b). Up until the age of one monththe nerve cord is double throughout most of itslength. In the region of the thorax it is very slender,composed of two longitudinal rows of nuclei joinedby a transverse, slightly thickened mesentery whichprobably represents the neuropile. The two rowsof nuclei are attached by mesenteries to the peri-toneum of the body wall and the lip glands. Thenerve cord is thickened in the trunk region whereit lies within the epidermal layer of the body wallas two longitudinal cords (Plates \2b-d, 13c). Theneuropile of each cord is dorsal or dorsomedial tothe nuclei or in some regions the fibers of theneuropile may be scattered among the nuclei. Theseparation between the two neuropiles is not alwaysdistinct in the most anterior regions, but posteriorlythe neuropiles are clearly separate.

At the age of two months in most larvae thenerve cord has united. In the thorax it is bilobed,the two lobes joined at their bases with nuclei ina lateral position on the outer side of each lobe(Plate 13<i). In the trunk the two longitudinalcords have fused to form a single cord, suspendedby a thickened peritoneum. The nuclei of the cordare ventral and the neuropile dorsal as in the adult.Posteriorly the nerve cord becomes extremely thinand it has not been possible to trace it to the pos-terior end of the trunk.

The central nervous system is ectodermal inorigin. The rudiments of the brain are formedfrom the lateral cells of the apical plate by aninward proliferation in the early trochophore and

by the end of the trochophore stage the brain rudi-ments are well developed (Plate lla). The cells ofthe brain surround the rosette cells and extenddownward to the prototroch, forming the sides ofthe apical cavity. Beneath the cavity the lateralbrain rudiments are united by a central commis-sure, the future neuropile. The nuclei of the brainrudiments are small and ellipsoidal, measuring ap-proximately 4x5 microns, with prominent chro-matin granules and very little surroundingcytoplasm. When the mouth opens at metamor-phosis the ventral portion of the brain is pushedinto a more dorsal position, the apical cavity isobliterated, the rosette cells degenerate, and thetwo lateral lobes of the brain are united.

The pigment granules of the eye appear late inthe trochophore stage in a dorsolateral cell oneither side of the apical plate. The two cells arelocated directly above the prototroch at the ex-ternal surface beneath the egg membrane. Duringthe premetamorphosis stage the granules increasein number and assume the form of a cup with theopening directed posterolaterally. At metamor-phosis the opening of the cup is turned in ananterolateral direction, retaining its position nearthe surface.

The circumesophageal connectives appear tooriginate from the brain, growing down on eitherside of the esophagus. By the time of metamorphosisthe connectives are complete.

The nerve cord is proliferated inward from theventral ectoderm posterior to the stomodaeum. Atthe end of the trochophore stage a few scatteredcells, very similar in cytology to those of the brain,make their appearance in this position. In the pre-metamorphosis stage at the time of post-trochalelongation the ventral nerve cord is clearly visibleas two loosely arranged rows of small nuclei nearthe ventral surface of the developing embryo. Thetwo rows are separated posteriorly by epidermalcells, but in the region beneath the stomodaeum thecells are more scattered and there is no distinctseparation. The neuropile has not been identified inprelarval stages.

In all other studies on sipunculans the ventralnerve cord has been reported to arise from a medianunpaired thickening of the ectoderm of the trunk,beginning in the anterior trunk region and pro-gressing posteriorly (Hatschek, 1883; Gerould,1907; Akesson, 1958, 1961a). In Phascolosoma agas-


sizii the development of the ventral nerve cordbegins as a ventral ectodermal proliferation ofscattered cells in the anterior trunk region, but itdiffers from other sipunculans in that it progressesposteriorly as two separate longitudinal cords. Bythe time the larva has reached the age of 2 months,the two cords have united.

Gerould (1907) reported a transitory metamerismof the nerve cord and the mesoblasts into two tofour segments. Later in letters to W. K. Fisher andG. Pickford, he repudiated this report, explainingthe apparent metamerism to be the result of larvalcontraction (Hyman, 1959:657). Neither Hatscheknor Akesson found any evidence for segmentationin the ventral nerve cord, nor was any such evi-dence found in Phascolosoma agassizii.

Both Gerould (1907) and Hatschek (1883)asserted that the circumesophageal connectives wereformed by a bifurcation and forward growth ofthe anterior end of the ventral nerve cord. Akesson(1958) was unable to determine whether the con-nectives originate from the brain or the ventralnerve cord. Because in Phascolosoma agassizii theconnectives appear before the nerve cord is welldeveloped, it is suggested that they arise from thebrain.

Eye spots similar to those of Phascolosoma agassiziihave been described in the embryos of Golfingiavulgaris, Phascolopsis gouldi (Gerould, 1907), andGolfingia elongata (Selenka, 1875). The eye spotsas figured by Gerould are of the inverse type, unlikethe everse adult eye which is embedded within thebrain at the inner end of an ocular tube. In thelarva of Phascolosoma agassizii it has not beenpossible to demonstrate whether the larval eyes areinverse or whether they are rudiments of the adulteyes. Akesson (1958) observed rudimentary oculartubes in the larvae of Phascolion strombi andGolfingia minuta, but he did not find pigmentedeye spots in these species. Two pairs of eye spots,one large and one small, have been reported in thelarva of Sipunculus nudus (Hatschek, 1883) and inunidentified pelagosphera (Senna, 1906; Jagersten,1963). In Sipunculus nudus one pair of eye spotsappeared in the early gastrula and a second in theolder larva. Akesson (1958) hypothesized that thefirst pair might be the inverse type and the secondthe everse eye characteristic of the adult.

NEPHRIDIA

In the larva of Phascolosoma agassizii there is asingle pair of nephridia, each member of the pairwith a U-shaped ciliated canal opening at one endto the exterior by way of an epidermal pore and atthe other end to the coelom through a ciliatedfunnel (Plate lOd). The external pores are situatedin a ventrolateral position on either side of theventral nerve cord, slightly anterior to the level ofthe anus. From their point of attachment in theregion of the pore the nephridia are suspendedfreely within the coelom, usually extending pos-teriorly, although they may be pushed anteriorlyby the movement of the gut or the introversion ofthe head (Plate 4b). At the age of one month thenephridia measure approximately 40 microns inlength.

The arm ol the nephridium that is associatedwith the funnel is narrower in diameter than thearm that terminates at the external pore, thedifference being more marked in the region of theexternal openings (Plate Id). The nephridium isfour cells in width, with two cells surrounding thefunnel canal and two cells around the pore canal.The number of cells along the length of thenephridium increases with its growth. The cells ofthe pore arm are characterized by green granules,often enclosed within large vacuoles. The cytoplasmof both the pore and funnel arms is densely con-centrated with mitochondria and the nuclei arelarge with finely dispersed chromatin and prominentnucleoli (Plate Id). The nuclei of the cells of thefunnel are small and resemble those of theperitoneum that surrounds the remainder of thenephridium.

The rudiments of the nephridia have beenidentified in the premetamorphosis stage at thetime of coelom formation (Plates Id, 10a). Anephridial rudiment consists of two enlarged cellslocated ventrally within the mesoderm bands inthe middle of the post-trochal region of the de-veloping embryo. The distinctive nuclei are largeand lobate with a lightly staining nucleoplasm andone or two prominent nucleoli. In the premeta-morphosis stage the somatic and splanchnic bandsof mesoderm split anteriorly and posteriorly to thenephridial rudiment to form the coelom, leavingthe nephridial cells as a bridge connectingendoderm and ectoderm (Plates 2d, 3b, 10a, b). At

NUMBER 132 17

the time of metamorphosis the nephridial cells areseparated from the gut, but retain their connectionwith the body wall as they are overgrown by cellsof the somatic peritoneum (Plate 10c). It appearsprobable that the more interior nephridial cellgives rise to the funnel arm of the nephridium andthe more exterior cell to the pore arm. A prolifera-tion of small cells within the epidermis joins theouter nephridial cell to form the external pore,and cells resembling those of the peritoneum sur-round the nephridium to form the funnel. Thepores and the funnels were not observed untilseveral days after metamorphosis.

There is no protonephridium apparent at anytime during the development of Phascolosomaagassizii and the larval nephridium, as describedabove, undoubtedly becomes the definitive nephri-dium of the adult. The exact origin of the larvalnephridium has not been conclusively determined.The position of the nephridial cells in the meso-derm bands and the relationship of the developingnephridium to the coelom suggest that the nephri-dium of Phascolosoma agassizii is mesodermal inorigin, growing out from the coelom toward theperiphery. The source of the nephridial cells,however, has not been traced and the possibility oftheir migration inward from the ectoderm cannotbe excluded. Moreover, the exact derivation of thecells of the coelomic funnel is not known, and it ispossible that the funnel is formed from coelomicepithelium as a separate structure and at a laterstage than the tubular portion of the nephridium.

In the literature on sipunculan developmentthere are discrepancies concerning the mode offormation of the nephridium. Hatschek (1883)reported a purely mesodermal origin of thenephridium of Sipunculus nudus. He was able torecognize two nephroblasts within the mesodermat a very early stage by their distinctive yellowpigmentation. The division of each yellow cellwas found to give rise to a U-shaped chain of cellswith a central lumen opening at one end throughthe epidermis and at the other into the coelom. Theanlage of the coelomic funnel he reported to becomposed of cells of the peritoneal lining. Thereported origin of the nephridia of Sipunculusnudus from cells in the mesoderm bands is thussimilar to that observed for Phascolosoma agassizii.In the species studied by Akesson (1958) andGerould (1907), the nephridium was not observed

at its inception because of large amounts of coelomicyolk. Both authors assumed, nevertheless, that thenephridium proper was formed as an ectodermalinvagination and that the funnel arose from peri-toneum. Akesson'-s assumption of an ectodermalparticipation in nephridial formation was based onthe morphology of the larval nephridium inGolfingia minuta. The central cavity of thenephridium in this larva is lined with a cuticle thatis continuous with that covering the body wall.Akesson supposes, therefore, that the cells sur-rounding the cavity are ectodermal and that themore peripheral cells and the cells of the funnelare derived from mesoderm.

The observations of Akesson and Gerould, al-though incomplete, suggest that the nephridia ofthe species they studied are composite structures,conforming to the concept of mixonephridia byGoodrich (1945). Observations on Phascolosomaagassizii and Sipunculus nudus, on the other hand,do not rule out the possibility that the nephridialstructure may be a coelomoduct, derived entirelyfrom entomesoderm. In view of the informationavailable to him from the studies of Hatschek(1883) and Gerould (1907), Goodrich (1945)

assumed the sipunculan nephridium to be amixonephridium. Further evidence on the exactsource of the "nephroblasts," the derivation of thecoelomic funnels, and the comparative cytology ofthe cellular components is now necessary to clarifythe nature of the nephridia of the Sipuncula.

LARVAL GLANDS

The glands in the larva of Phascolosoma agassiziiare designated as the lip glands, the posteriorsacciform glands, the epidermal glands, and theglands of the head. The lip glands have beendiscussed previously as an associated structure ofthe mouth.

Posterior Sacciform Glands.—The posterior sac-ciform glands are thus termed because of theirshape and their position in the posterior half ofthe larva. There are two glands, one on either sideof the anus in a dorsolateral position (Plate 46).The glands are cellular sacs, composed of a singlelayer of cells surrounding a central cavity that opensto the exterior through a wide aperture in the bodywall (Plate 146, c). In the larva of one month of

18 SMITHSONIAN CONTRIBUTIONS TO ZOOLOCY

age each gland projects into the coelom to a lengthof approximately SO microns.

The cells of the gland are of two different types:one presumed to be secretory and the other sup-portive. The nucleus of the secretory type is round,approximately 4.5 microns in diameter, with ahomogeneous, darkly staining nucleoplasm and avery large, eccentric nucleous (Plate 146). Thereis only one such cell in each sacciform gland andit is located at the rounded extremity of the gland.The cytoplasm of this cell is thicker than that ofthe other cells of the gland and it is intenselybasophilic. The nucleus of the second cell type,the supportive cell, is smaller, measuring 2x3microns, with fine chroma tin granules. Many cellsof this type form the wall of the gland surroundingthe central cavity. The cytoplasm of the cells, alsobasophilic, becomes increasingly thin as the centralcavity enlarges and the gland increases in size. Thecentral cavity is filled with tightly packed globules,the secretory product of the gland. These globulesare stained a light blue with Richardson's stain, butthey are not stained by alcian blue or periodic acid-Schiff. Hence, the chemical nature of the secretionhas not been determined and the possible functionof the sacciform glands in larval life remainsenigmatic.

The posterior sacciform glands develop fromectoderm late in the premetamorphosis stage. Anectodermal cell on either side of the rectal rudimentincreases in size; the cytoplasm of the cell becomesbasophilic and the nucleus moves into a positiontoward the coelom (Plate 14a). In that part of thecytoplasm nearest the body wall, small globules,similar to those later found in the central cavityof the gland, make their appearance. This cell isidentical to the cell at the extremity of the maturegland, and it is presumed to be the secretory ele-ment of the gland. Smaller ectodermal cells areproliferated around the base of the larger cell, andas the gland grows these cells are extended to. formthe walls of the gland, surrounding the centralcavity. The aperture through the body wall iscompleted a few days after metamorphosis.

Glands such as the posterior sacciform glands inthe larva of Phascolosoma agassizzi have not beendescribed in any other species of sipunculans.

Epidermal Glands.—There are three pairs ofepidermal glands in the larva at 20 days of age(Plate I4d). Located in dorsolateral positions on

the trunk, one pair is near the middle of the bodyand two additional pairs are relatively closetogether in a more posterior region. In livinglarvae each gland appears as a small epidermalpapilla from which a very fine bristle, 5 to 10microns in length, protrudes to the exterior. Insectioned material the glands are found to bepyriform in shape, the more pointed end leading toa narrow canal which passes through the larvalcuticle. The rounded base of the gland, covered byperitoneum, projects for a distance of approxi-mately 10 microns into the coelom. The gland iscomprised of a cluster of small epidermal cells sur-rounding a diminutive cavity which opens into thecuticular canal. Within the cavity is a homogeneoussecretory material, staining light blue with Richard-son's stain, and from the base of the cavity thebristle extends outward through the canal. Incontrast to the other epidermal cells, the cytoplasmof the gland cells is scant and the nuclei arecharacterized by fine chromatin granules and anabsence of nucleoli.

The epidermal glands are derived from ectodermalproliferations, first evident late in the premetamor-phosis stage. The most anterior pair of glands, theearliest to appear, is fully developed with extendedbristles at the end of metamorphosis. By 17 daysbristles are also apparent on the two posterior pairsof glands. All of the bristles have vanished in larvaeolder than one month and many additionalepidermal glands have formed without bristles(Plate 13d).

Epidermal glands similar to those in Phascolosomaagassizii have been described in all studies ofsipunculan larvae; however, associated bristles havebeen reported only in Golfingia elongata (Selenka,1875; Akesson, 1961b) and Phascolion strombi(Akesson, 1958). In Phascolion strombi, studied byAkesson (1958) at Kristineberg, the bristles werenot arranged in pairs, but rather in six circlets,each circlet with four bristles. Golfingia elongata,also from Kristineberg, showed the same arrange-ment (Akesson, 1961b), but larvae of the samespecies at Roscoff entirely lacked bristles (Gerould,1907; Akesson, 1961b) and those at Nice possessedonly three pairs (Selenka, 1875). In an effort toexplain this variation among different populationsof the same species, Akesson (1961b) pointed outthat Golfingia elongata at Kristineberg lived atdepths of 30 meters, whereas those at Roscoff in-

NUMBER 132 19

habited the intertidal zone. Such a correlationbetween depth of habitat and presence of bristleswas not found in the present study in which larvaefrom an intertidal population of Phascolosomaagassizzi were observed to possess three pairs ofbristles.

Akesson (1958) speculated on the functionalsignificance of the glands and bristles, suggestingthat the secretion of the gland in Phascolion strombimight aid in dissolving the holes which the larvaeinhabited in gastropod or scaphopod shells, or thatit might have a protective effect against enemies.Although he was unable to demonstrate a sensoryelement in the gland, Akesson proposed that thebristles might be tactile organs. In the study ofPhascolosoma agassizii neither sensory cells norsensory nerves could be demonstrated in the glands;however, larval nerves are very difficult to traceand sensory cells might not have been recognized,therefore it cannot be concluded that a sensoryelement does not exist. Epidermal organs in theadult of this species have a sensory function as wellas secretory, and it is not unlikely that the organsof the larva might also have a dual function.

Glands of the Head.—There are many differentglands in the head of the larva of Phascolosomaagassizzi. Some of the glands secrete a materialthat stains with periodic acid-Schiff, in othersneurosecretory granules can be demonstrated, andin still others the secretion has not been charac-terized.

Opening on the ventrolateral surface of the headare two pairs of long, thin glands, approximately7 microns in width, packed with small globules thatstain very intensely with periodic acid-Schiff (Platellrf [g1, g-]). One pair opens directly above thelateral extremities of the buccal groove, passingposteriorly along either side of the esophagus for adistance of 40 microns. A second pair, somewhatshorter, opens more anteriorly and extends into thehead coelom. Also staining with periodic acid-Schiffis a third pair of smaller, club-shaped glands, ap-proximately 15 microns in length, that opensfarther anteriorly on the ventral surface of thehead (Plate \\d [g3]). These glands are filled withloosely packed globules, smaller than those of thelong, thin glands. The secretion of these threepairs of glands is discharged onto the ventral sur-face of the head and may aid in the trapping of

food particles which would then be swept by ciliaryaction into the esophagus.

Neurosecretory granules have been demonstratedin cells in the brain and in cells along the sides ofthe esophagus. After fixation in osmium tetroxideand embedding in methacrylate the larvae weresectioned at 1 micron and stained in paraldehydefuchsin by the method of Clark (1955). Prominentneurosecretory granules were stained a very deepred. In larvae embedded in Epon and stained withRichardson's stain, very large cells in the brain andalong the esophagus were assumed, because of theirposition and cytology, to be the neurosecretorycells of the methacrylate-embedded material. Thecytoplasm of these large cells was intenselybasophilic with vacuolar inclusions and the nucleuscontained a prominent nucleolus Plates 8b, lid[g5])-

There are other glands with an unidentifiedsecretion that are apparently unicellular and opento the exterior. The secretion, foamy in appear-ance, does not stain with Richardson's stain,periodic acid-Schiff, or paraldehyde fuchsin. Onepair opens on the ventral surface of the head be-tween the apertures of the long, thin glands de-scribed above, and the others open on the anteriorhead along the ventral boundary of the brain (Platellr, d[g*]).

The origin and fate of the glands of the larvalhead in Phascolosoma agassizii are not known. Inthe larvae of Phascolion strombi, Golfingia minuta,and Golfingia elongata, Akesson (1958, 1961b)described a larval cerebral organ, derived from therosette cells, with a secretion characterized in thefirst two species by basophilic, argentaffine stainingproperties. The more proximal cells of the organare eventually incorporated into the brain, whereasthe remainder form the cerebral organ of the adult.The cells that are incorporated into the brain arenot secretorily active in the adults of these species;however, Akesson (1961b) points out that cells inthe same position in the brain of the adultSipunculus nudus are neurosecretory and he as-sumes, because of their identical position, that theyare homologous to the incorporated cells of otherspecies. From these observations and assumptions,Akesson concludes that primitive epidermal secre-tory cells transform into neurosecretory cells withinthe brain. Since Akesson did not examine the larvaof Sipunculus nudus and Hatschek (1883), in his


study of the larva of this species, did not describea larval cerebral organ, further observations on thelarva are necessary before Akesson's conclusion canbe confirmed. In the larva of Phascolosoma agassiziino distinct cerebral organ was found, but neuro-secretory cells, of uncertain origin, were alreadypresent within the brain of the larva. Whetherother glandular cells later transform into additionalneurosecretory cells has not been determined in thepresent study.

TERMINAL ORGAN

The terminal organ is a retractable appendagelocated at the posterior end of the larva (Plates5c, 4a,c,e,f, \-ie,f). Usually a bulb-shaped organ, 25microns at its widest dimension and 50 microns inlength, the terminal organ may be elongated to alength of 170 microns. The function of the organis the attachment of the larva to the substratum.When the larva is swimming or lying unattachedon the substratum the organ may be retractedwithin the body.

Withdrawal of the terminal organ into theposterior portion of the trunk is accomplished bythe contraction of two retractor muscles. Thesemuscles originate within the terminal organ andinsert dorsally on either side of the dorsal midlineat the level of the middle pair of epidermal glands(Plate 14/). Each retractor muscle is composed of

six fibers and differs from the retractor muscles ofthe head, each of which has eight fibers.

The terminal organ is covered by a thin cuticle,a continuation of the cuticle of the trunk. Thecuticle at the posterior tip of the organ is inter-rupted by an aperture into which several glandcells open (Plate 14e,/). Gland cells, vacuolar inappearance, are not stained by Richardson's stain,periodic acid-Schiff, or alcian blue. Even thoughthe secretion has not been characterized by itsstaining properties, it is assumed that it is adhesivein nature because of the attachment of the larvato the substratum by means of the terminal organ.

The terminal organ and its retractor musclesform from an ectodermal proliferation, first ob-served during the premetamorphosis stage in aventral position near the posterior tip of the body.An invagination appears in the egg membraneopposite the ectodermal proliferation and at thetime of metamorphosis the ectodermal thickening

is pushed out through the invagination, takingwith it the surrounding cuticle, to form the termi-nal organ (Plate 14e). The cuticle at the posteriortip of the organ is ruptured, thus creating theaperture through which the secretion is discharged.At the end of metamorphosis the larvae settle onthe substratum, an indication that the gland cellshave become active in secretion.

A terminal organ has been reported in otherbenthopelagic or pelagic sipunculan larvae. Inthe larva of Sipunculus nudus Hatschek (1883)described an ectodermal thickening to which bristleswere attached. The thickening was connected in-ternally to the ventral nerve cord by thin fibers andhe supposed it to be a tactile organ. Usually re-tracted, the organ was only rarely observed to beextended. Akesson (1961a) described a similarorgan in unidentified, preserved pelagospheralarvae and noted the similarity to the terminalorgan of adult species of Sipunculus. In the larvaeof an intertidal population of Golfingia elongata,Akesson (1961b) found a terminal glandular organ,but in a deep-water population of the same speciesat Kristineberg he did not find such an organ.Larvae in the former population were observed toadhere to the substratum by means of the organ,but no mention was made of retractility. In morethan half of the species of pelagosphera thatJagersten (1963) examined, he observed a retractile"tail," considered to be identical with the terminalorgan described by Akesson (1961a). Jagerstenreported an odd behavior in which the "tail" wasplaced into the mouth possibly, he speculated, todischarge a secretion. In Phascolosoma agassizii asimilar behavior was observed whereby the bodywas bent so that the terminal organ touched the lipor else was pushed past the head (p. 5). Thisaction was most frequently noted when there wasno substratum on which the larva could settle, andwhat role, if any, it served in normal behavior wasnot determined. The primary function of theterminal organ in this species is attachment of thelarva to the substratum. It would be of interest toascertain whether the pelagosphera studied byJagersten ever use the terminal organ under naturalconditions to attach to a substratum, or to somedebris floating among the plankton. If not, thenthe behavior that Jagersten reports might be ex-plained as an adaptation to planktonic life and the

NUMBER 132 21

terminal organ in such forms may serve an entirelydifferent function from that of attachment.

Comparisons with Other Phyla

The phylum Sipuncula with spiral cleavage anda trochophore larva is, by definition, a member ofthe Protostomia. Close affinities with both annelidsand mollusks are apparent from developmentalstudies and much of this evidence has been re-viewed in detail by Hatschek (1883), Gerould(1907), Akesson (1958), and Clark (1969).

Comparisons of cuticle formation reveal that inmany polychaetous annelids, as in sipunculans, theegg envelope may undergo a transformation intothe larval cuticle. Such a transformation is foundin the sipunculan species Golfingia elongata,Phascolion strombi, Golfingia minuta (Akesson,1958; 1961b), Phascolosoma agassizii, Golfingiapugettensis, and Themiste pyroides (Rice, 1967)and in numerous polychaetes, some of which areClymenella torquata, Arenicola cristata (Wilson,1892), Pomatoceros triqueter (Segrove, 1941), Nereisdiversicolor (Dales, 1950), Diopatra cuprea andAutolytus fasciatus (Allen, 1959; 1964).

A larval structure, the metatrochal band of cilia,common to all pelagosphera larvae of sipunculans,is homologous in sipunculans and polychaetes, asindicated by the cell lineage studies of Gerould(1907). In contrast to that of most polychaetes,however, the metatroch of the pelagosphera is theprimary locomotory organ and at the time of itsformation, or soon thereafter, the prototroch islost or diminished. In Sipunculus nudus the proto-troch is unique in that it completely envelops theembryo. This structure has been compared byGerould (1907) to a similarly modified prototrochthat constitutes the velum of some aplacophoransand primitive lamellibranchs.

A protonephridium, found in the trochophoresof the many polychaetes and some mollusks, islacking in sipunculans. The larval nephridium ofthe sipunculan, a U-shaped tubular structure open-ing at one end to the exterior and at the other tothe coelom, becomes the nephridium of the adultwith no essential change. Its precise derivation hasyet to be determined, but studies of Phascolopsisgouldi, Golfingia vulgaris (Gerould, 1907) andGolfingia minuta (Akesson, 1958) suggest that it isa composite structure, formed from ectoderm

and coelomic epithelium. In Sipunculus nudus(Hatschek, 1883), however, and in Phascolosomaagassizii a purely mesodermal origin is suggested,but not proved. Goodrich (1945) considered thesipunculan nephridium to be a mixonephridium, astructure characteristic of most polychaetes but notof mollusks.

The nervous system develops similarly insipunculans and annelids and is comprised in bothof a cerebral ganglion, circumesophageal con-nectives, and a ventral nerve cord. The nerve corddevelops as a ventral ectodermal thickening whichis double in annelids, but in most sipunculans ismedian and single. An exception is the speciesPhascolosoma agassizii in which the nerve cord ispaired at its inception, although later it is unitedinto a single cord. Gerould (1907) considered theunpaired ventral nerve cord to be a primitivefeature, relating the sipunculans to primitiveannelids such as Polygordius. Akesson (1958) tookan opposite view, pointing out that among theturbellarians, annelids, and mollusks the orthogonalnervous system of turbellarians should be con-sidered the primitive one and that the nerve cord ofsipunculans therefore was not a primitive butrather a derived feature. The rudimentary, pairednerve cord in Phascolosoma agassizii is interpretedin this study as a reminiscence of a double nervecord that probably appeared in the development ofthe ancestors of extant sipunculans, relating thesipunculans more closely to the annelidan stem.

The anus of the sipunculan larva does not ariseat the posterior region of the blastopore as in mostpolychaetes, but rather it arises at the site of adorsal ectodermal proliferation, the rectal rudiment.In a few polychaetes, such as Hydroides (Shearer,1911), the anus develops similarly.

A striking and important difference from theannelids is the complete lack of segmentation of thenerve cord and the mesoderm. Although Gerouldin 1907 described a temporary division of the nervecord and mesoderm of Phascolopsis gouldi into twoto four segments he later retracted this report(Hyman, 1959:657). No other investigator has notedany indication of mesodermal or nervous segmenta-tion in sipunculan development. The arrangementof epidermal bristles into several circlets in thelarvae of Golfingia elongata and Phascolion strombiis regarded by Akesson (1958, 1961a) as evidence ofa transitory metamerism. The papillae from which


the bristles arise, however, are ectodermal and thereis no involvement of the mesoderm. The bristles, asecretion of the papillae, are readily dissolved byfixatives and are in no way homologous to thechaetae of polychaetes.

Jagersten (1963) noted the likeness of the lip ofthe pelagosphera to the foot of a gastropod larva.Previously Gerould (1907) had suggested a possiblehomology of the lip gland (glandular organ) andthe buccal organ ("Schlundkopf") to two similarlyplaced invaginations in the veligers of chitons: thepedal gland and radular sac. These organs, how-ever, have not been observed in the lecithotrophicpelagosphera which are assumed to represent a moreprimitive development (Rice, 1967); hence, itseems probable that they are adaptations of theplanktotrophic larva for feeding activity.

Summary

The pelagosphera larva of Phascolosoma agassizii,as observed from living specimens reared in thelaboratory, is characterized by a ventrally ciliatedand bilobed head, a thorax with prominent meta-troch which, along with the head can be withdrawninto the body, an elongated, wedge-shaped trunkwhich in older larvae is marked by characteristicallypapillated cuticle, and a posterior terminal organ.The larva attaches to the substratum by the termi-nal organ and in this position it feeds on bottom-dwelling diatoms and debris. It is able to releaseitself from the substratum and swim freely throughthe water by means of the metatrochal circlet ofcilia. Larvae were not observed to transform intothe adult form.

In a histological study of the structure, origin,and development of larval organs, the followingobservations were made: The stomodaeum isformed at the site of the closed blastopore and theproctodaeum arises later as a new formation in adorsal position. The coelom results from thesplitting of the mesoderm layers and the retractormuscles, originating from ectomesoderm, make theirappearance soon alter the formation of the coelom.At the time of metamorphosis of the trochophoreinto the pelagosphera larva the egg envelope istransformed into the elastic, extensible cuticle of thelarva. The nephridium arises from two conspicuouscells within the mesoderm bands. The ventralnerve cord, derived from trunk ectoderm, is double

at its inception, but by the age of 2 months the twolongitudinal cords have united. There are threepairs of epidermal glands with bristles and, inaddition, there is one pair of posterior sacciformglands opening to the exterior on either side of theanus. Two organs of the mouth, the buccal organand the lip gland, previously described only in thelarva of Sipunculus nudus and in unidentifiedpelagosphera larvae, are present in the pelagospheralarva of Phascolosoma agassizii.

Finally, the histogenesis of selected structures insipunculans shows many similarities to the poly-chaetous annelids and, to a lesser extent, to themollusks.

Literature Cited

Akesson, B.

1958. A study of the Nervous System of the Sipunculoi-deae with Some Remarks on the Development ofthe Two Species Phascolion strombi Montagu andColfingia ininuta Keferstein. Undersdkningar overOresund, 38:1-249. 105 figures. Lund: C. W. K.Gleerup.

1961a. Some Observations on Pelagosphaera Larva. Gala-thea Report, 5:7-17, figures 1-11.

1961b. The Development of Golfingia elongata Keferstein(Sipunculidea) with Some Remarks on the De-velopment of Neurosecretory Cells in Sipunculids.Arkiv for Zoologi, series 2, 13 (2) :511-531, figures1-11.

Allen, M. J.

1959. Embryological Development of the PolychaetousAnnelid, Diopatra cuprea (Bosc). Biological Bulle-tin, 116:339-361, figures 1-30.

1964. Embryological Development of the Syllid, Auto-lytus fasciatus (Bosc) (Class Polychaeta) . BiologicalBulletin, 127:187-205, figures 1-45.

Bennett, H. S., and J. Luft1959. S-collidine as a Basis for Buffering Fixatives.

Journal of Cell Biology, 6:113-117.

Clark, R. B.

1955. The Posterior Lobes of the Brain of Nephthys andthe Mucus Glands of the Prostomium. QuarterlyJournal of Microscopical Science, 96:545-565.

1969. Systematics and Phytogeny: Annelida, Echiura,Sipuncula. Chapter 1, pages 1-68, in Marcel Florkinand Bradley T. Scheer, editors, Chemical Zoology,Volume IV. New York and London: Academic Press.

Clement. A. C, and J. N. Cather

1957. A Tcchnic for Preparing Whole Mounts of VeligerLarvae, Biological Bulletin, 113:340.

Dales, R. P.1950. The Reproduction and Larval Development of

Xereis diversicolor (O. F. Miiller). Journal of the

NUMBER 132 23

Marine Biological Association of the United King-dom, 29:321-360.

Damas, H.

1962. La collection de Pelagosphaera du "Dana." DanaReport, 59:1-22, 10 figures.

Dawydoff, C. N.1930. Quelques observations sur Pelagosphaera, larve de

sipunculide des cotes d'Annam. Bulletin de laSocie'te de Zoologique de France, 55:88-90.

Fisher, W. K.

1947. New Genera and Species of Echiuroid and Sipun-culoid Worms. Proceedings of the United StatesNational Museum, 97 (3218):351-372.

Gerould, J. H.

1903. Studies on the Embryology of the Sipunculidae, I:The Embryonal Envelope and its Homologue.Mark Anniversary Volume, pages 439-452, 1 plate.New York.

1907. Studies on the Embryology of the Sipunculidae, II:The Development of Phascolosoma. ZoologischeJahrbu'cher, Abteilung t'iir Anatomie und Onto-gonie der Thiere, 23:77-162, plates 1-11.

Goodrich, E. S.1945. The Study of the Nephridia and Genital Ducts since

1895. Quarterly Journal of Microscopical Science,86 (342-344): 113-392, figures 1-100.

Hall, J. R., and R. S. Scheltema1966. Morphology of North Atlantic Sipunculid Larvae,

(Abstract). American Zoologist, 6(3):338.

Hatschek, B.1883. Ueber Entwicklung von Sipunculus nudus. Arbeiten

aus dem Zoologischen Institute der Universitat-Wien und der Zoologischen Station in Triest, 5:61-140, plates 1-6.

Heath, H.

1910. Pelagosphaera, a Larval Gephyrean. Biological Bul-letin, 18:281-284, figure 1.

Hyman, L. H.1959. The Invertebrates: Smaller Coelomate Groups.

Volume 5. New York: McGraw-Hill.Jagersten, G.

1963. On the Morphology and Behaviour of Pelago-sphaera Larvae (Sipunculoidea). Zoologiska BidragFran Uppsala, 36:27-35, figures 1-2.

Krohn, A.1851. Ueber die Larve des Sipunculus nudus, nebst vor-

ausgeschickten Bemerkungen iiber die Sexualver-haltnisse der Sipunculiden. Archiv fur Anatomie,Physiologie, und wissenschlaftliche Medizin, 368-379. 1 plate.

Lillie, R. D.1954. HistopatUologic Technic and Practical Histochem-

istry. 501 pages. New York: The Blakiston Com-pany, Inc.

Luft, L. H.1961. Improvement in Epoxy Resin Embedding Methods.

Journal oj Biophysical and Biochemical Cytology,9:409-414.

Mingazinni, P.

1905. Un Gefireo pelagico: Pelagosphaera Aloysii n. gen.,n. sp. Rendiconti delle sedute solemn delta R. Ac-cademia Nacionale dei Lincei, 14:713-720, figures1-2.

Mueller, Max

1850. Ueber eine den Sipunculiden verwandte Wurmlarve.Archiv fur Anatomie, Physiologie, und wissen-schlaftliche Medizin, 438-451, 1 plate.

Murina, V. V.

1965. Some Data on the Structure of Pelagospheres—Sipunculid Larvae. Zoologicheskii Zhurnal [inRussian], 44:1610-1619, figures 1-8.

Rice, M. E.

1967. A Comparative Study of the Development ofPhascolosoma agassizii, Golfingia pugettensis, andThemiste pyroides with a Discussion of Develop-mental Patterns in the Sipuncula. Ophelia, 4:143-171, figures 1-4.

Richardson, K. C, L. Jarett, and E. H. Finke

1960. Embedding in Epoxy Resins for Ultrathin Section-ing in Electron Microscopy. Stain Technology, 35:313-323.

Scheltema, R. S., and J. R. Hall

1965. Trans-oceanic Transport of Sipunculid Larvae Be-longing to the Genus Phascolosoma. (Abstract).American Zoologist, 5 (2) :216.

Scgrove, F.

1941. The Development of the Serpulid Pomatoceros

triqueter L. Quarterly Journal of MicroscopicalScience, 56:543-590.

Selenka, E.

1875. Eifurchung und Larvenbildung von Phascolosoma

elongatum (Kef.) Zeitschrift fsir wissenschlaftlicheZoologie, 25:442-450, 2 plates.

Senna, A.

1906. Sulla struttura di alcune larve (Pelagosphaera) diSipunculidi. Pubblicazioni del (R.) Instituto diStudio superiori practici e di Perfezionamento inFirenze; Sezione de Scienze Fisichee Naturale, 50-78.

Shearer, C.

1911. Development and Structure of Trochophore ofHydroides uncinarus (Eupomatus). QuarterlyJournal of Microscopical Science, 56:543-590.

.Spcngel, J. W.

1907. Eine verkannte Sipunculus-I^T\e. Zoologischen An-zeiger, 31:97-99.

Stephen, A. C.

1941. The Echiuridae, Sipunculidae and Priapulidae Col-lected by the Ships of the "Discovery" Committeeduring the Years 1926 to 1937. Discovery Report,21:237-260, 2 plates.

Wilson, E. B.

1892. The Cell Lineage of Sereis. Journal of Morphology,6:361-480.

2 4 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

PLATE 1

Early developmental stages of Phascolosoma agassizii: blastula, gastrula, trochophore. Epon, onemicron, Richardson's stain.

a. Sagittal section of a 24-hour embryo showing rosette cells (r) with apical cilia, modifiedblastocoel (b), prototroch cells (p) and three-layered egg envelope (en) perforated by finepore canals.

ft. Sagittal section of a 48-hour embryo showing rosette cells (r) , blastocoel (b), blastopore(bp), mesentoblast (ms), and archenteron (ar).

c. Sagittal section of trochophore showing gut cavity (gc) , prototrochal cavity (pc), prototrochcells (p) and posterior cavity (psc). Note yolk granules in the prototrochal cavity. 4 days(10-C).

d. Oblique sagittal section of trochophore. 51/2 days (10°C). Note apical cavity (ac), proto-trochal cavity (pc) and stomodaeum (st). The prototroch cells (p) have lost most of theiryolk granules at this stage. Prototrochal cilia are clearly shown in this photograph.

[ac —apical cavity, ar=archenteron, b=blastocoel, bp = blastopore, en —egg envelope, gc=gutcavity, ms^mesentoblasts, p = prototroch cell, pc = prototrochal cavity, psc=posterior cavity,r—rosette cell, st = stomodaeum.]

NUMBER 132 25


PLATE 2

Late trochophorc and piemetamorphosis stages of Phascolosoma agassizii. Epon, one micron,Richardson's stain.

a. Frontal section of late trochophoie showing rosette cells (r) with apical cilia, apical cavity(ac), prototrochal cavity (pc), and protonoch cells (p). The prototroch cells and rosette

cells have lost their yolk granules. 6 days (10°C).b. Frontal section of prcmetamorphosis stage. 7 days (10°C). Note apical cavity (ac), proto-

trochal cavity (pc) , coelom (cm) , and the peritoneal cells lining the coeloni. Three regionsof the gut can be identified: esophagus (es), stomach (stm) and intestine (in).

r. Frontal section of premetamorphosis stage showing cells of the ventral retractor muscles(vrm) traversing the coelom (cm) . 7i/2 days (10°C).

d. Frontal section of premetamorphosis stage through nephridial rudiment (ne). Posterior tothe nephridial rudiment the coelom (cm) is apparent as a narrow slit between splanchnicand somatic mesodermal layers. ~\/2 days (IO°C) .

[ac—apical cavity, cm —coelom, es=z esophagus, in = intestine, ne —nephridial rudiment, p =prototroch cells, pc = prototrochal cavity, pe=peritoneum, r —rosette cells, stm = stomach, vrm— ventral retractor muscles.]

NUMBER 132 27


PLATE 3

Metamorphosis from late trochophore stage to pelagosphera larva of Phascolosoma agassizii.Epon. One micron. Richardson's stain.

a. Sagittal section of premetamorphosis stage. 6 days (12°C). Note the intercellular spaces in theepidermis. The egg envelope (en) posterior to the prototroch (p) has begun to lose itsporosity and lamellation.

b. Parasagittal section of the beginning metamorphosis stage. 8 days (12°C). The egg envelope(en) has ruptured in the region of the mouth and the stomodaeal lining has everted to formthe ciliated ventral surface of the head (vh).

c Sagittal section of recently metamorphosed larva. 9 days (12°C). The coelom (cm) hasexpanded, the metatroch (mt) has become the functional locomotory organ, the terminalattachment organ (to) has emerged, the gut (evident here only in part) has been completed,and the egg envelope (en) has undergone a transformation into the larval cuticle (cu).

[bo—buccal organ, br—brain, cm = coelom, cu—cuticle, en —egg envelope, ep=epidermis,es=esophagus, in=intestine, lg=lip gland, lprrlip pore, mo=;mouth, mt = metatroch, ne— nephridium, np=neuropile, p = prototroch, pc = prototrochal cavity, pel = prototrochal cilia,pe = peritoneum, psg=posterior sacciform gland, stm = stomach, to = terminal organ, vh=ven-tral ciliated head, vnc=ventral nerve cord, vnn —ventral retractor muscle.]

NUMBER 132 29

pct-

cu

to-


PLATE 4

Photographs of pelagosphera larvae of Phascolosoma agassizii: living larvae and whole mounts.

a. Eleven days. Whole mount, fixed in 10 percent buffered formalin, mounted in euparol.No stain. Lateral view.

b. Four weeks. Fixed in osmium vapor (2 minutes) after relaxation in a 1:2 solution of satu-rated chloretone in sea water (3 minutes). Dorsal view.

c. Four weeks. Living larvae. Lateral view. Metatroch is retracted.d. Ten weeks. Living larva. Lateral view of head. Metatroch is retracted.e. Ten weeks. Living larva. Lateral view of posterior terminal organ (to). Note papillated

cuticle (cu)./. Ten weeks. Living larva. Lateral view of entire larva with retracted lip and metatroch.

Note change from 4-week stage (Ac) in relative proportions of body.

[a —anus, cu = cuticle, e —eye, in—intestine, l = lip, mt — metatroch, ne=nephridium, psg=pos-terior sacciform gland, stm—stomach, tozrterminal organ, vh —ventral ciliated head.]

NUMBER 132 31


PLATE 5

Sagittal section of larva of Phascolosoma agassizii. 28 days. The terminal organ, present at thisstage, is not shown in this section. Epon. One micron. Richardson's stain.

[a = anus, bo=:buccal organ, co=coelomocytes, cm = coelom, cu = cuticle, es=esophagus, epr=epidermis, in = intestine, lg=lip gland, lp^lip pore, mo=mouth, mt = metatroch, ne=:nephridium, re—rectum, stm —stomach, vhzrventral ciliated head, vnc=ventral nervecord.]

NUMBER 132 S3

vh

mt

—mo

-ep


PLATE 6

Sections showing formation of ventral retractor muscles and coelom in Phascolosoma agassizii.Epon. One micron. Richardson's stain.

a. Frontal section of an early premetamorphosis stage showing the cells of the ventral retractormuscle (vrm) growing into the newly formed coelom (cm). 6i/2 days (10°C).

b. Frontal section of premetamorphosis stage, slightly later than above. 7 days (10°C). Cellsof the ventral retractor muscle (vrm) have invaded the coelom (cm).

c. Frontal section of premetamorphosis stage showing completely formed ventral retractormuscle (vrm) traversing the coelom (cm). 7i/fe days (10°C).

d. Oblique frontal section of metamorphosed larva. 11 days (10°C). The coelom (cm) hasexpanded and the cells of the ventral retractor muscles (vrm) have elongated.

[cm —coelom, es = esophagus, mt = metatroch, pc=prototrochal cavity, vrm = ventral retractormuscle.]

NUMBER 132 S5


PLATE 7

Sections showing gut, coelomocytes and cuticle in prcmetamorphosis and larval stages of Phas-colosoma agassizii. Epon. One micron. Richardson's stain.

a. Frontal section of premetamorphosis stage. 7i/2 days (10°C). Note coelomocyte (co) whichappears to be attached to the splanchnic peritoneum. The stomach (stm), in contrast to theesophagus, is marked by a dense concentration of darkly staining granules of varying sizes.

b. Sagittal section of 28-day larva showing spindle-shaped coelomocytes (co), heavily ciliatedesophagus (es), weakly ciliated stomach (stm), and circular muscles of the postmetatrochalsphincter (crm).

c. Sagittal section of 43-day larva. Both round and spindle-shaped coelomocytes (co) areevident as well as the longitudinal musculature of the body wall and the papillated cuticle.

d. Cross-section of 57-day larva through the stomach (stm) and nephridium (ne). Each cellof the wall of the stomach contains one large darkly staining granule or globule. In thenephridium the ciliated canals of the two arms are apparent: the upper is associated withthe coelomic funnel and the lower terminates at the external pore.

[cm—coelom, co=coelomocyte, crm=circular muscle, cu = cuticle, en=egg envelope, ep=epi-dermis, es=;esophagus, in = intestine, lm = longitudinal muscle, mt = metatroch, ne = nephrid-ium, pc=prototrochal cavity, pe = peritoneum, stm = stomach.]

NUMBER 152 37


PLATE 8

Sections of larval organs of the mouth. Larvae of Phascolosotna agassizii. Epon. One micron.Richardson's stain.

a. Frontal section of 60-day larva showing buccal organ (bo) with buccal groove (bg) and lipglands (lg).

b. Sagittal section through lip and lip pore (lp) and associated organs of the lip of a 28-daylarva. The arrow points to a neurosecretory cell overlying the esophagus.

[bg=buccal groove, bo —buccal organ, es=esophagus, lg —lip gland, lp = lip pore, mt = meta-troch, vnc=ventral nerve cord.]

NUMBER 132 39


PLATE 9

Sections showing formation of organs of the mouth. Phascolosoma agassizii. Epon. One micron.Richardson's stain.

a. Frontal section of premetamorphosis stage. 7 days (10°C). Note the two large cells of thefuture lip gland (lg) and the stomodaeum (st).

b. Frontal section of premetamorphosis stage showing lip gland (lg) with cytoplasmic vacuolesand inclusions and stomodaeum (st). 9i/2 days (10°C).

c. Frontal section of premetamorphosis stage showing buccal organ (bo) and stomodaeum(st). 9 days (10°C).

d. Sagittal section of premetamorphosis stage. 9i/2 days (10°C). Note stomodaeum (st), buccalgroove (bg), buccal organ (bo), lip gland (lg), and stomach (stm).

[bg—buccal gland, bo z= buccal organ, lg=lip gland, st=stomodaeum, stm —stomach.]

NUMBER 132 41


PLATE 10

Sections showing formation and structure of nephridium and anus of Phascolosoma agassizii.Epon. One micron. Richardson's stain.

a. Frontal section of premetamorphosis stage. 7i/2 days (10°C). The nephridial rudiment(ne), consisting of two enlarged cells, is located within the mesodermal band which hassplit anterior and posterior to the developing nephridium to form the coelom (cm).

b. Sagittal section of premetamorphosis stage showing rectal rudiment (re), intestine (in),stomach (stm) and nephridial rudiment (ne). Note a slight invagination of the egg envelopeindicated by the arrow opposite the rectal rudiment; the envelope later will rupture at thesite of the invagination to form the anus. 8i/2 days (I0°C).

c. Sagittal section of recently metamorphosed larva. The anal opening (a) has been formed,completing the intestine (in), and the nephridial rudiment (ne) has enlarged. 10 days(10'C).

d. Sagittal section of 28-day larva showing the well-developed nephridium (ne). The arrowpoints to the ciliated funnel opening into the coelom.

[a—anus, cm=coelom, cu=cuticle, in=intestine, ne=nephridium, pc—prototrochal cavity, re=rectum, stm—stomach.]

NUMBER 132 43


PLATE 11

Sections of premetamorphosis stage and pelagosphera larvae of Phascolosoma agassizii showingdevelopment of the brain and the structure of the buccal organ and glands of the head. Epon.One micron. Richardson's stain.

a. Frontal section of premetamorphosis stage through the region of the developing brain (br)and neuropile (np). 7i/2 days (10°C).

b. Frontal section of 60-day larva showing brain (br), neuropile (np), eye (e), esophagus(es), and buccal organ (bo).

c. Frontal section of 60-day larva. Head glands, designated by the arrows, open to the exterior.They are the same as that labeled g4 in the sagittal section pictured in lid.

d. Sagittal section of 28-day larva. Five different glands of the head are apparent in thissection: gl, g2, and g3 are PAS-positive; g4 are the "foamy" glands with unidentified secre-tion; g5 have been characterized as neurosecretory cells.

[bg=buccal groove, bo=buccal organ, br=brain, drm=dorsal retractor muscle, e=eye, es=esophagus, gl-5=glands of the head, np=neuropile, vrm=ventral retractor muscle.]

NUMBER 138 45


PLATE 12

Cross-sections of a recently metamorphosed pelagosphera larva of Phascolosoma agassizii. 10days (10°C). Epon. One micron. Richardsons' stain.

a. Cross-section through region of the lip.b. Cross-section through region of the mid-stomach. Double nature of the ventral nerve cord

(vnc) is clearly apparent.c. Cross-section through anus and nephridium.d. Cross-section of posterior body through intestine.

[a = anus, bg=buccal groove, cec=circumesophageal connective, drm=: dorsal retractor muscle,es=esophagus, in = intestine, ne=nephridium, stm —stomach, vnc —ventral nerve cord.]

NUMBER 132 47


PLATE 13

Cross-sections of pelagosphera larvae of Phascolosoma agassizii. Epon. One micron. Richard-son's stain.

a. Cross-section of a 22-day larva through the region of the head. Note that in this region thedorsal retractor muscles (drm) are attached to the peritoneum of the body wall.

b. Cross-section of a 22-day larva through the metatroch (mt) and bilobed lip (1). Dorsaland ventral retractor muscles (drm, vrm) are seen suspended in the coelom as circles offibers.

c. Cross-section of a 22-day larva through the region of the nephridia (ne). Some of thefibers of the ventral retractor muscles (vrm) insert into the body wall in this section.

d. Cross-section of a 57-day larva through the region of the contracted postmetatrochal sphinc-ter. At this age the epidermal glands (epg) are well developed, the retractor muscles (drm,vrm) have increased in thickness and number of fibers and the nerve cord is bilobed.

[cec=zcircumesophageal connectives, cu=cuticle, drm = dorsal retractor muscle, epg=epidermalgland, es=:esophagus, in = intestine, l = lip, lgmlip gland, mtr=metatroch, ne —nephridium,vnc=ventral nerve cord, vrm —ventral retractor muscle.]

NUMBER 132 49

E•lHk . ••cec *•' IP

a


PLATE 14

Sections of premetamorphosis and larval stages of Phascolosoma agassizii. Stages in the forma-tion of the posterior sacciform glands and the terminal organ are illustrated. Epon. Onemicron. Richardson's stain.

a. Sagittal section of beginning metamorphosis stage showing posterior sacciform gland (psg).The nucleus of the gland cell is distinguished by the large eccentric nucleolus.

b. Oblique cross-section of the distal end of the posterior sacciform gland (psg) of a 28-daylarva through the secretory cell with its characteristic nucleus. The large, unstained gran-ules of the central cavity of the gland are of unidentified composition.

c. Sagittal section of posterior sacciform gland (psg) in a 43-day larva. Filled with amorphoussecretory material, the central cavity of the gland is surrounded laterally by supportive cellsand distally by the secretory cell. The cavity is continuous with the exterior by way of theproximal canal which penetrates the body wall.

d. Sagittal section of 22-day larva showing anterior epidermal gland (epg) with bristle (brs).e. Sagittal section of beginning metamorphosis stage showing the terminal organ (to) in an

early stage of formation. Note the gland cells (gl) which open to the exterior through theaperture at the distal end of the terminal organ.

/. Sagittal section of the terminal organ (to) of a 22-day larva. This section shows the aper-ture at the end of the terminal organ and the retractor muscles of the terminal organ (trm).

[brs = bristle, co—coelomocyte, cu=cuticle, epg = epidermal gland, gl —gland of terminal organ,ne=znephridium, pe= peritoneum, psg=posterior sacciform gland, stm=stomach, to = terminalorgan, trm —retractor muscle of terminal organ, vrm — ventral retractor muscle.]

NUMBER 132 51

U. S. GOVERNMENT PRINTING OFFICEl 1»7S 4«4-3SS/«O

Publication in Smithsonian Contributions to Zoology

Manuscripts for serial publications are accepted by the Smithsonian Institution Press, sub-ject to substantive review, only through departments of the various Smithsonian museums. Non-Smithsonian authors should address inquiries to the appropriate department. If submission isinvited, the following format requirements of the Press will govern the preparation of copy.

Copy must be typewritten, double-spaced, on one side of standard white bond paper, withW'2," top and left margins, submitted in ribbon copy with a carbon or duplicate, and accompa-nied by the original artwork. Duplicate copies of all material, including illustrations, should beretained by the author. There may be several paragraphs to a page, but each page should beginwith a new paragraph. Number consecutively all pages, including title page, abstract, text, litera-ture cited, legends, and tables. The minimum length is 30 pages, including typescript and illus-trations.

The title should be complete and clear for easy indexing by abstracting services. Taxonomictitles will carry a final line indicating the higher categories to which the taxon is referable:"(Hymenoptera: Sphecidae)." Include an abstract as an introductory part of the text. Identifythe author on the first page of text with an unnumbered footnote that includes his professionalmailing address. A table of contents is optional. An index, if required, may be supplied by theauthor when he returns page proof.

Two headings are used: (1) text heads (boldface in print) for major sections and chap-ters and (2) paragraph sideheads (caps and small caps in print) for subdivisions. Furtherheadings may be worked out with the editor.

In taxonomic keys, number only the first item of each couplet; if there is only one couplet,omit the number. For easy reference, number also the taxa and their corresponding headingsthroughout the text; do not incorporate page references in the key.

In synonymy, use the short form (taxon, author, date:page) with a full reference at theend of the paper under "Literature Cited." Begin each taxon at the left margin with subse-quent lines indented about three spaces. Within an entry, use a period-dash (.—) to separateeach reference. Enclose with square brackets any annotation in, or at the end of, the entry.For references within the text, use the author-date system: "(Jones, 1910)" and "Jones(1910)." If the reference is expanded, abbreviate the data: "Jones (1910:122, pi. 20: fig. 1)."

Simple tabulations in the text (e.g., columns of data) may carry headings or not, but theyshould not contain rules. Formal tables must be submitted as pages separate from the text, andeach table, no matter how large, should be pasted up as a single sheet of copy.

Use the metric system instead of, or in addition to, the English system.Illustrations (line drawings, maps, photographs, shaded drawings) can be intermixed

throughout the printed text. They will be termed Figures and should be numbered consecu-tively; however, if a group of figures is treated as a single figure, the components should beindicated by lowercase italic letters on the illustration, in the legend, and in text references:"Figure 9b." If illustrations (usually tone photographs) are printed separately from the text asfull pages on a different stock of paper, they will be termed Plates, and individual componentsshould be lettered (Plate 9b) but may be numbered (Plate 9: figure 2). Never combine thenumbering system of text illustrations with that of plate illustrations. Submit all legends onpages separate from the text and not attached to the artwork. An instruction booklet for thepreparation of illustrations is available from the Press on request.

In the bibliography (usually called "Literature Cited"), spell out book, journal, andarticle titles, using initial caps with all words except minor terms such as "and, of, the." Forcapitalization of titles in foreign languages, follow the national practice of each language.Underscore (for italics) book and journal titles. Use the colon-parentheses system for volume,number, and page citations: "10(2) :5-9." Spell out such words as "figures," "plates," "pages."

For free copies of his own paper, a Smithsonian author should indicate his requirementson "Form 36" (submitted to the Press with the manuscript). A non-Smithsonian author willreceive 50 free copies; order forms for quantities above this amount with instructions for pay-ment will be supplied when page proof is forwarded.

i