vdl^l Zoomorphology (1983) 102:143-163 Zoomorphology <: Springer-Vcrlag 1983 Structure, Ultrastructure, and Function of the Terminal Organ of a Pelagosphera Larva (Sipuncula) Edward E. Ruppert' and Mary E. Rice^ I Departmenl of Zoology, Clemson University, Clcmson, SC 29631, USA - Departmenl of Inverlebrale Zoology, National Museum of Natural History Smithsonian Institution, Washington, D.C. 20560, USA Summary. The terminal organ, a structure enabling pelagosphera larvae o? bipuncula to form temporary attachments to substrata, was examined behaviorally and with light and electron microscopy for larvae of Golfin- gia misakiana, collected from the Florida Current. The terminal organ appears as a retractile rounded knob with a short neck joining the poste- rior extremity of the trunk. It can attach larvae directly to substratum or can secrete a tether-like mucus strand about which the organism moves. In unattached larvae, the terminal organ is often placed in the mouth. The terminal organ of a 5.5 day old larva consists of 29 cells- 8 epidermal, 3 mucus, 2 tension-bearing, 5 sensory, 10 retractor muscles and 1 unknown cell. The mucus cells are presumed to release the adhesive material while the microvilli on the tension-bearing cells, with their dense cores of microWaments, bear the strain. The latter are joined directly to the retractor muscles which originate on the dorsal body wall near the anus. Two of the sensory cells terminate within the cuticle flanking the adhesive pore and are assumed to be cuticle strain receptors Three sensory cells terminate in cilia that extend posteriorly from the pore These may function in substratum evaluation prior to temporary attach- ment, or settlement preceding metamorphosis. The terminal organ is compared to adhesive organs in other soft-bodied metazoans and al- though It approximates the structure found in some rotifers, it is consid- ered to be independently evolved within the Sipuncula. The terminal organ can be understood as an adaptation in young larvae for protective attachment and facilitation of feeding whereas, in older larvae, it may only function in substrate evaluation prior to settlement. A. Introduction The planktomc larval stage in the phylum Sipuncula is known as the pela- gosphera. The term was first used by Mingazinni (1905) who, in describing Offprint requi'sis to: M.K. Rice 0720-213X/83/0102/0143/$04.20 r, E Riippcil and M.E. Rice 144 an oceanic larva from the South Pacitlc Ocean, believed it to be a planktomc adult sipunculan. His mistake was soon realized and, more recent y, the term pelagosphera has been used to designate the planktomc larval slage of sipunculans that succeeds the trochophore and is characterized by a prominent band of mctatrochal ciha utilized by the larva as ^* 'o^om?;o;>' organ (Rice 1967, 1973, 1975, 1978, 1981; Hall and Schcltema 1975, Schel- tema 1975- Scheltema and Hall 1975). Most pelagosphera larvae arc plank- totrophic, although a few are known to be lecithotrophic, having a pelagic or benthi-pelagic stage of short duration (Gerould 1906; Akesson 1958, 1961 a- Rice 1967 1975). Whether planktotrophic or lecithotrophic, a char- acterislic feature of nearly all pelagosphera larvae is a terminal o?'g'ii' com- monly used for temporary attachment to a substratum. In larvae of differen species and in larvae of the same species, but differing ages, <-hc terminal organ may show varying degrees of prominence and use. When highly devel- oped the terminal organ functions for temporary attachment ol the arva to a substratum. When relatively small, the terminal organ is rarely, U ever, used for attachment and has been presumed to be primarily sensory in function (Hatschek 1883). ^ ,^ ,h,. Only a few previous reports on sipunculan larvae have referred to the histology or structure of the terminal organ and these studies have been on hxr^Ie with relatively minute organs. In his studies of the development of Sipunculus nudus. Hatschek (1883) described the terminal organ ot this larva as an epithelial thickening, usually invaginated, which was provided with tufts of sensory cilia. Akesson (1961b) described the histology of 4 oceanic pelagospheras from preserved plankton collections of the Galathea expedition off Natal, noting the presence of an invaginated terminal organ, comprised of stratified glandular epithelium, which he compared to a similar but more complex terminal organ in adults of the genera Xenosiphon and Sipunculus. Secretory cells have also been reported in the terminal organs of the young lecithotrophic larva of Golfingia elongata (Akesson 961 a) and the young planktotrophic larva o? PImscolosoma ugassizi {R\CQ \)Ii). In planktotrophic larvae of the open ocean the degree of development of the terminal organ ranges from a prominent organ by which the larva may attach when a substratum is contacted to a small, relatively insigmhcant structure usually retracted and not observed to serve as an attachment or^^an These larvae arc found in the surface waters of all of the major east-west currents in the North and South Atlantic oceans. They may live in the larval stage as long as 8 months (Schcltema and Hall 1975) attaimng sizes from one-half to several mm in length. Little is known of the adult affinities of most of these larvae, and in only one instance is there intorma- lion on both the young and older pelagosphera larval stage of a single species This species is tentatively identified as Golfingia nnsakuma (Ikcda 1904) and has been reared in the laboratory from oceanic larvae to sexualy mature adults (Rice 1978, 1981). From spawnings of these adults early larval stages have been obtained and studied (see Materials and Methods). The terminal organ of the young larva of Goljingia misakuma has been Sipunculan Terminal Or 145 found to be highly developed as a structure for attachment whereas in the older oceanic larva it is proportionately small and rarely extended In this paper we have chosen to examine in detail the behavioral mor- phological, and ultrastructural features of the terminal organ of a youo'^ pelagosphera of Golfingia misakkina of 4 to 5 days after fertilization In addition, comparative observations will be noted on the older oceanic larva ot the same species, but of unknown age. By this multi-i^iceted investigation we hope to learn more about the function of the organ, particularly at the earlier stage in development, and to clarify the role of the organ in the biology of the pelagosphera larva. The ultrastructural observations also will be compared with those on adhesive organs of other marine inverte- brates and a mechanism suggested for larval attachment and release This IS the first intensive study of the terminal organ of sipunculans and we anticipate that it will serve as a basis for subsequent comparative studies ot the structure and significance of this organ in the phylum Sipuncula. B. Materials and Methods For this study young pelagosphera larvae were obtained from spawnings of adults reared n the laboratory from oceanic larvae collected in the Florida Current. The oceanic larvae u>?n ^f",'"" previously and lenlativcly idenlificd as Goljhf^ia misakUma by Rice (1978 n8l)^harlier reports on this larval type are those ofHali and Scheltcma (1975) who designated a as Type C and Hacker (IS98) who nan,cd il Baccaria oliva. The adult reared from larvae in he laboratory dlllers in some morphological and developmental characters from lield- collected ammals (Rice 1981), although the significance of these differences remains to be IxkSfmAY "'''"' '"'' '^"''^"''' "'" "''' '''?'''"' '?' '"" '^'-'^'?"^'"^-'' ^"^ ^"^f'?f^'^' nmaklana Oceanic larvae vvere collected by plankton tows off the central east coast of Florida ?0 to 25 miles oUshore Irom Fort Pierce in the Florida Oiirrent, a component of the C}ulf Stream system. Tows of 0 to 15 mm duration were made with a '/, m net of 125 ^m mesh. Plankton hril'^"''','r't "'??'" ''fnif "'?"'"''^ '''^'??'^ ''"?^'"'^ ""^'?'-^ ^"'-^?^ ""d i"^"ced to metamorphose by the method ol Rice (1978, 1981). After metamorphosis animals were maintained in fine sediment less than 102 urn in diameter in a recirculating system or in dishes in which the months''"' '" ?'' ^'"''' " '"""''? ^'^"''' "''""'?''?'' '""' "'''?'"'^^ "" '^ ^" 1" For spawning, adults were removed from the sediment and placed in fingerbowls Spawn- ings o eggs and spcm, obtained from October through March, were transferred to tall covered petri dishes of 500 ml capacity and maintained at room temperature (about ?5 C) Three to our da3'S after fertilization the young pelagosphera larvae, resulting from metamorphosis UoTnsm ''?/?'?', ?;'"'"' ""'""*^ "'""" ^^'' '" '-??"'^entrations ranging from about 100 to 500 per dish. Usually sea water was allowed to stand in these dishes 2 to ^ davs pnor to lran.s(cr so thai a thin biological film covered the bottom. Yount' larvae were (Id an algal mi.xture, including Phaeodactylum sp., Chlorella auwirophica. Dwmliclla salina ho- chiysi.s };alhami. and Asicrionella i^lacialis. Specimens were prepared for transmission electron microscopy by relaxation in 10% etha- nol in sea water, fixation in 2.5% glutaraldehyde buffered with Millonig's phosphate buffer at room temperature, postfixation in phosphate-bufTered 2% osmium tetroxide, dehydration m an alcohol and propylene oxide or acetone series, and embedment in Fpon 81? Serial thin sections were obtained for 5 larvae (4.5 and 5.5 days old) and were collected on Formvar himed sjot grids For scanning electron microscopy specimens were fixed in 2.5% phosphate- bufteix-d glutaraldehyde, dehydrated ?, alcohol and acetone, dried in a Dent?n critical point drier (CO,) and coated with gold paladium in a Technics Hummer 1 sputter-coaler E E Ruppert and M.E. Rice 146 ? n C. Results /. Morphological and Behavioral Observations of Live Larvae In developmental stages reared in the laboratory, we observe the first ap- pearance of the terminal organ at the premetamorphosis stage of the trocho- phore three days following fertilization. Arismg as an evagination of the posterior body wall, it appears as a bulge or protuberance ot the posterior extremity of the trunk. By 4^2 days, at the time of metamorphosis of the trochophore to the pelagosphera larva, the terminal organ is well- formed, resembling a posterior rounded knob connected to the body by a more narrow, elongate neck (Figs. 1-4). At this time larvae begin to attach by the terminal organ to the bottom of the culture dishes, although the maiority continues to swim for one or more days, some moving throughout the culture dish and others swimming near the bottom. Within a few days, usually by 6 days of age, most of the larvae are firmly attached. The shape of the terminal organ is changeable in living larvae, depending on the degree of extension and retracdon of the body. Although usually knob-shaped with an elongate neck, it may be stretched into a cylindrical or club-shaped organ. The neck of the organ has considerable fiexibihty and extensibility, allowing the larva to move about its point of attachment in all directions. Whether attached or free, the terminal organ can be re- tracted into the trunk of the larva either entirely or in part. On maximal retraction the posterior body wall confinuous with the neck of the organ is also withdrawn into the trunk, leaving a prominent depression in the posterior body in the region of the retracted organ. In a larva with retracted Fia ( Photomicrograph of living pch.gosphcra larvae M days after fertihzalion. Spawnings aivmg ri e to these larvae and to those in figs. 2, 3, and 4 were Iront adults reared m the f?bontt?y ironr oceanic larvae collected in the Florida Current and tentafvely tdent,hed T?:?an?ia ,nisak,ana. The larvae are attached by their terminal organs (w) to debns. Note that the terminal organ of the middle larva is retracted Fig 2 Scanning electron micrograph of a pelagosphera larva 11 days after fertilization. Ventro- lateral view, ml, metatrochal band of cilia; lo. terminai organ Fig. 3. Scanning electron micrograph of the terminal organ of a pelagosphera larva II days after fertilization Fie 4 Photomicrograph of terminal organ of a living pelagosphera larva 6'/, days after fertil- Stom Note attachment strand [as) that serves to attach the young larva to a substratum Fig 5. Scanning electron micrograph of larva of unknown age collected from ?'='^^"1^ Pl""'\\??; o he Florida Current. This larval type when reared in the laboratory g.ves nse to an adu ?enlati4ly identilled as Golfingia misakiamr The arrow ponrts to the termmal organ vvh.ch rela ively snt II in comparison with that of the younger larva of the s.tn.e speces shown !n Fif 2 L contrast to the smooth cuticle of the young larva, that of the olde, ocean.c larvals covered with prominent papillae, ml. metatrochal band ol cilia Fig 6 A higher magnification of the posterior region of the larva in F,g. 5 shows the terminal organ (?IITOW) and its relation to the surrounding cuticular papillae Sipiinculan Terminal Organ 147 F H Rupncrt and M.!'. Rice 148 terminal organ, the posterior body is shortened and rounded whereas in a larva with extruded terminal organ, the body is attenuated, the region immediately anterior to the terminal organ being frequently t.gh ly con- stricted Retraction of the terminal organ is dependent on the contraction of a pair of posterior retractor muscles (described below); extension, on the other hand, is apparently brought about by the contraction of the muscu- lature of the posterior body wall, forcing the extrusion of the organ. Larvae have been observed in the laboratory to exhibit a variety ol behavioral activities. An attached larva can stretch out at any angle from its point of attachment. It can extend straight upward perpendicular to the bottom; in this position we believe the larva can feed on suspended matter in the water column by directing particles into the mouth through eihary activity of the ventral head and metatroehal regions. The larva can also bend its body in a C-shape so that the ventral surface of the head is apphed to the bottom of the dish. In this latter position larvae have been observed to feed by grazing on diatoms and bottom detritus in the vicinity of their attachments. Larvae are capable of turning in complete circles about their points of attachment. Although usually attached, larvae en release themselves and swim freely through the water Most commonly free larvae remain near the bottom of the culture dish. Hei-e they are ob- served to move along either parallel to the bottom or with ventral head applied to the bottom and posterior end upward, apparently feeding on bo tom detritus. Unattached larvae may remain relatively quiescent on the bottom Frequently they are observed to curl the body so that the terminal organ is placed in or near the mouth; the function of this behavior, charac- teristic of all sipunculan larvae, is not understood. In attempting to determine the manner of larval attachment, we have removed larvae from culture dishes along with debris to which they arc attached for examination under the compound microscope. In some in- stances we have observed a thin strand of material, designated as the at ach- ment strand which joins the tip of the terminal organ to the available Substratum Fig. 4). The strand is llexible, bending or folding upon itsell as the larva moves in different directions, or extending out straight as the larva moves away from its attachment. It functions as a tether limiting the distance the larva is able to move. There is no evidence that the strand is elastic The attachment strand is difficult to detect because of its mini, e dimensions and because it is often obscured by surrounding debris. We have not been able to determine, therefore, whether this is ^c usual or only an occasional mode of attachment of the terminal organ. That the terminal organ produces an adhesive secretion is evident from the adherent de?is commonly surrounding the organ. Algal cells, fecal matter, and other extraneous material accumulate around the terminal organ and Irequtntly remain attached to it even after its release from the substratum. The irmness of attachment varies considerably among larvae in a culture dish. Some are readily dislodged by suction of a pipette fitted with ^' ?'^bber bulb whereas others can be dislodged only by scraping off he him o deb s to which they are attached. When suction is applied to those most tightly Sipunciikin Terminal OrL'an 149 attached, they retract the anterior body into the trunk and spin like a top on their strongly cemented points of attachment. Once attached, larvae remain in one position for long periods Patient observation of cultures under the dissecting microscope has revealed an occasional release and reattachment of a few individual larvae Larvae ob- served to release have been in a feeding position bent over with ventral head apphed to substratum. Al release the attachment is abruptly disrupted and the terminal organ directed upward away from the substratum The arva then swims away, usually remaining close to the substratum and ircquently moves along with ventral head on the substratum, apparently leeding as it progresses. Prior to reattachment the larva ceases swimming remaining quiescent on the bottom. On one occasion attachment occurred alter tiie larva had moved around in a circle with terminal organ extended toward the bottom and head directed upward. If the larva is not movin" actively around its point of attachment, it is not easy to determine whether attachment has taken place. It has been assumed thai attachment has oc- curred il the larva is not dislodged by jarring of the dish or by currents ol water directed toward it by means of a pipette. Incidental ob.servations in the laboratory indicate that successful attach- ment IS dependent on availability of an appropriate substratum. Larvae vvill not attach to sand grains or fine sediments, and when introduced to dishes containing such substrata, they continue to swim for a few days then die. The larvae settle readily, however, in glass culture dishes in which sea water has been allowed to stand for a day or two and a thin film ot undetermined origin has developed on the bottom of the dish. //. Eli'cinm Microscopic Observations I. Gcn-cral Organization of the Terminal Organ and Trunk (?\g?, 7 8) When viewed with the transmission electron microscope, the terminal' organ of the 5.5 day larva appears as a prominent, bulb-shaped outgrowth of the posterior body wall (Fig. 7). It lacks, however, both the peritoneal epitheli- um and a continuation of the coelom, i.e. it is a compact organ The distal pore of the terminal organ is a circular interruption of the outer two layers of the overlying cuticle (4 5 pm diam) that receives the necks of the various gland and sensory cells associated with the organ. The bulbous part of the terminal organ joins the trunk by a narrow stem through which pass neuntes and 10 well-developed retractor muscle cells (Fig 8) The body wall of the trunk of the 5.5 day larva resembles that of the terminal organ except that the euficle is thicker (0.5 1.8 ^m Fig 21) and the musculature is not as strongly developed. Apart from a well-developed sphinct^er muscle associated with the hindgut, the body wall muscles consist only of outer cu'cular muscle cells (40 50 nm thick in transverse section situated just below the epidermis) and the eoelomic epithelium, or peritone- um, which lorms the longitudinal body wall musculature (0.6-4 0 urn thick m non-nuclear regions. Fig. 21). Myofilamenls were not observed in perito- 150 K.n. Rupperl and M.H. Rice Sipuncuhin Terminal Oraan 151 ncal cells of older larvae where a layer of longitudinal body wall muscles IS present (unpublished observations). 2. Cellular Organization of the Terminal Organ (Figs. 7 30). The bulbous portion ol the terminal organ is composed of 29 cells which include- 8 epidermal cells, 3 mucus cells, 2 tension-bearing cells, 5 sensory cells I internal cell of unknown function and 10 retractor muscle cells We will offer a short description of each of the ceil types below. The pore of the terminal organ is a distally situated dimple that is formed by the microvillar borders of the necks of 2 mucus gland cells (in the S 5 day larva), 3 sensory cells and the apical borders of the 2 tcnsion-bearmg cells (Figs. 7, 27 29). Seven cells, therefore, arc involved in the formation and functioning of the adhesive pore of the terminal organ in the 5 5 dav larva. 3. Epidermal Cells (Figs. 8 10, 25, 26). The epidermis of the terminal organ forms a thin (0.9 1.4 ^m) covering over the organ; it secretes the cuticle and elaborates microvilli that extend through the cuticle (Fig 21) The mi- croviUi are occasionally branched. The epidermal cells arc intcrjoined by apical zonulae adhaerentes followed basally by septate desmosomes The cuticle of the terminal organ of the 5.5 day larva is thin (0.3 -0 7 |im) and consists of 3 layers. The innermost part of the cuticle is formed by a basal zone of fibrous material that resembles surface coat material (Fi-^ 21) and IS quite unlike the basal cuticular zone of older larvae and adults''that possess a meshwork of collagen fibers (unpublished observations). Apieally there is a dense layer resembling a membrane (10 nm thick; Figs. 9, 12) and an outermost cpicuticular layer of electron dense material (40 70 nm thick; oceamc larva, see Fig. 19), Microvilli from the epidermal cells pass through the cuticle at intervals, forming slightly bulbous tips with radiating fibers of surface coat material extending into the external environment. The Golgi apparatus of each cell (as many as 3 Golgi bodies/cell) is situated in the perinuclear region near the apical plasmalemma. Small (170 nm) dense vesicles arc produced by the Golgi apparatus that may fuse immediately below the plasmalemma to form larger (400 600 nm) vesicles containing loosely consolidated material of varying densities. These vesicles are in close association with the apical plasmalemmata of all the epidermal cells but we have not observed any vesicles releasing their contents into the developing cuticle. Similar vesicles are present throughout the trunk epidermis (Fig. 21). It is also possible that these 400 600 nm vesicles may be involved in uptake rather than release of material but we have no experi- Figs. 7-8. TEMs of pclagosphcni larvae. Fig. 7. SagiUal TtiM seelion of a 4 S day larva Nolc aeeclcraled dirferenlialion of ihe terminal organ eompared with the trunk body wall' I'lg. 8. Iransverse TF.M seetion of a planklonic pclagosphera of undetermined age U is partly relraeted mlo the trunk, co. coelom: cu. culicje; <.,n. extracellular matrix; cp epidermis- J gul; ?,f;. mucus gland; m, neurile bundles; /;., peritoneum; po. pore of terminal organ'.,/ retractor muscles; i?, terminal organ 152 K.F? Ruppcrt and M.E. Rice 0.5p m > mg Sipunculan Tcrmin;il Organ mental data to resolve the issue. The epidermal cells were also noted to contam RER, mitochondria and an occasional lysosomc (1.4 |.im). 4. Mucus Glaml Cells (Figs. 10, 12, 16, 27). The terminal organ contains 3 secretory cells that produce granules we identify structurally as mucus (maximum diameter 1.2 ^im, Figs. 10, 12). The cells arc large (10.5-11 7 ^m long) and irregularly shaped with a nucleus situated nearest the proximal end of each cell and a short neck extending distally to the pore of the terminal organ (2 cells only). The cytoplasm of the cells is packed with mucus vesicles that are released at the pore through an apical collar (0 7 \im diameter) o? 12 microvilli (0.8 l.Ojim long each, Fig. 12). The microvilli do not appear to have cores of microfilaments. One mucus cell is situated vcntrally in the left half of the terminal organ and opens ventrolaterally into the pore. The second, and slightly larger, mucus cell is dorsally situated trom the midlinc to right side of the terminal organ and opens dorsally into the pore. The third is situated medially in the right half of the organ. A gland neck extending to the pore of the terminal organ was not found in any of our sections associated with this third mucus cell. We assume that the neck appears later in development. Each mucus cell contains perinu- clear RER and at least one Golgi body between the nucleus and the mucus granules. Additionally, the nuclei of both mucus cells contain distinctively electron-dense chromatin, a fact that distinguishes them immediately from nuclei of other cells in the terminal organ in our electron micrographs (Fig. 10). Two of the mucus cells were found to be innervated basally (Fig. 27). 5. Tension-Bearing Cells (Figs. 9, 11, 16, 28). Two cells form the primary tension-bearing elements of the terminal organ pore (Figs. 9, 11, 16). The perikarya of both cells are situated immediately interna? to the pore one to the Iclt of the pore and slightly ventral in position, and one to the right ot the pore and slightly dorsal in position. The apical surfaces of these cells iorm most of the surface area of the pore. Other cells involved with the pore contribute only narrow necks. Many microvilli (170 nm long) with dense cores of microfilaments issue from apical membranes of these cells (Figs. 11, 16). The basal surfaces of both tension-bearing cells are deeply invaginated just below the pore region to permit the insertion of retractor muscles on a point immediately internal to the pore itself (Fig. 11). As a result of this invagination, the distance between the apical and basal membranes of the Figs. 9-12. SagitUil Tl^M scclions of terminal organs. Fig. 9. Pore, tension-bearing eells and retractor muscle cell integration or4.5 day larva. Fig. 10. Mucns cells of ^,5 day larva Fig 11 Detail of tension-bearing cell o( 4.5 day larva. Fig. 12. Mucus ?land neck of 5 S day la'rva' C^omparc the microvilli wuh cores otmicroniaments on the tension-bearina cell with the micro- villi lacking such cores surrounding the gland neck. ac. tension-bearing cells- a, cuticle- en epidermis; lui. hcmidcsmosomes: /;;/; microfilamcnls.- m;.^. mucus eranulcs; ?w microvilli- no' pore ol terminal organ: ?-/;;, retractor muscles ' U.E. Ruppcrl and M.E. Rice 154 tension-bearing cells below the pore is as litlle as 1.0 pm. Two muscle cells extend into the basal invagination of each cell and Form a scries of junctional complexes with the basal plasmalemmala ol^ the tension-bearing cells across the extracellular matrix. These junctions resemble hcmidesmosomcs ?Fie 111 Bundles of microfilaments extend apically from the basal junc- tional complexes on the tension-bearing cells to the cores of the microviUi protruding into the pore of the organ (Figs. 11,16). The apical cytoplasm of these tension-bearing cells is moderately dense with 350-880 nm vesicles containing uniformly electron-opaque material (Figs "9 11 16 18). These vesicles are again RER/Golgi-derived. The dorsal- most'cc'll contained two Golgi bodies and the more ventral cell, one. The innervation of these cells has not been determined. 6 Sensorv Cells (Figs. 12-18, 29, 30). Five sensory cells occur within the terminal organ of the 5.5 day larva. Three open intx, the pore ot the organ and two are associated with the cuticle llanking the pore on the left and right sides. Four of the 5 cells are secondary sensory cells and 1 is a primary '''" Threc^lask-shaped cells are situated in the bulb of the terminal organ one dorsal and on the left side of the organ, one medial in PosiUon, and one ventral on the right side of the organ (Figs. 16 18, 29). AH 3 possess basal nuclei and are innervated by basal synaptic boutons. Each cell pos- sesses a long narrow neck that terminates in a collar ol microvilh at the terminal organ pore. A single cilium extends mto the pore fi-om the microvi 1- lar collars of both the ventral and medial cells (Figs. 16, 17). A ciliun is absent from the dorsal cell but a diplosome is present immediately internal to the microvillar collar of this cell and its presence m^ '"P?cate that cilio- genesis has not yet occurred (Fig. 16). All 3 cells produce RER/Golgi-de- r'ved vesicles with electron-dense material mostly 200 300 nm m diameter but some as large as 600 am have been noted. These vesicles closely resemble those produced by the tension-bearing cells discussed above. Two additional sensory cells occur in the terminal organ and each pos- sesses a dendritic ending associated with the basal layer of the cuticle. These endings do not extend into the pore of the terminal organ but are immcdiate- Iv lateral to it Each cell possesses a single Golgi body situated adjacent to the dendritic side of the nucleus that packages umform electron-dense material into small vesicles ranging in size from 350 nm to one large vesicle 1 iim in diameter. Many of the vesicles are electron-lucent, a fact that may indicate inadequate fixation of the material in some o? the vesicles. One of these cells, a secondary sensory cell, is situated on the right side of the terminal organ pore (Figs. 12-14, 30). It terminates apically in a cilium that passes dorsally through the basal cuticular layer from a position to iust below the pore to a position just above the pore. This mtracuticular cilium vertically flanks the right side of the terminal organ ''"'The second presumed proprioceptive cell, a primary sensory cell, is situ- ated dorsally and medially within the terminal organ (Fig 30). A long den- drite extends from the perikaryon to the right side of the pore where it "^\?-TV?i,v Sipuncuk?n Terminal Organ 155 0.5pm Figs. 13-18. Sensory eclls of terminal organ of 5.5 day larva. Sagillal TBM sections of the organ. Figs. 13-15. Serial thin sections showing disposition of intraculicular sensorv eilium. Fig. 16. Two pore-associated sensory cells. Note foreign material adhering to tcnsioti'-bearing cell microvilli. Fig. 17. Detail of third pore-associated sensory cell. Fig. 18. Secretion granules m porc-assoeiated sensory cell, ac, tension-bearing cell; hh. basal body: ci. eilium: cu cuticle- Jm. lorcign material; m/, microniaments: mi;, mucus granules; po. pore of terminal organ' nil. retractor muscles; sc, sensory cell tcrmmatcs in a narrow neck (600 nm) containing a basal body that protrudes into the basal cuticular layer. We assume that this basal body will induce ciliogenesis at a later stage of development. 7. Inierna/ Cell (Fig. 26). One internal cell, of unclear function, is found ventrally in the terminal organ bulb immediately below the epidermis in the 5.5 day larva. 156 E.H. Ruppcrl and M.E. Rice Sipunculan Terminal Organ 8. Retractor Muscle Cells (F,gs. 7-10, 22, 23, 28). Ten spindle-shaped retrac- tor muscle cells enter the terminal organ and form iiinctions with 16 of the 18 non-muscle cells. The retractor muscle cells extend from the terminal organ through the dorsal part of the coelom to insert on the body wall on each side of the hindgut (length 70 um in a relaxed, fixed specimen- maximum diameter 1.5 1.75 um). All the pcrikarya of the retractor muscle cells are situated m the coelom (Fig. 7). The retractor muscles appear smooth (non-striated) when examined by light-microscopy. Electron microscopy also indicates that the muscles are smooth. Unhkc typical invertebrate smooth muscle however where thin lilaments anchor to sarcomeral dense bodies and to sites on the sarcolemma (Reuter 1977), Z-rods arc present in these muscles, as is more typical of the various forms ot obliquely-striated muscle (Knapp 1978). The Z-rods radiate ccntripetally from their origins on the sarcolemma of the cells which arc oval to circular in transverse section (Figs. 23, 24). The Z-rods do not however, show the regular ordering that occurs commonly in obliquely- striated muscle of many annelids (Wissocq and Boilly 1977). These muscles therefore, seem to be somewhat intermediate between typical invertebrate obliquely-striated and smooth types. Thick (30-50nm) and thin (5 6 nm) myofilaments are distributed throughout the sarcoplasm in the non-nuclear parts of the cells (Figs 23 24). The ratio of thin to thick filaments is approximately 12 14-1 The lengths of the tapered thick filaments were difficult to determine because they appear to wmd in and out of the plane of our longitudinal sections I he ength of one filament measured at least 3 um, but wc arc uncertain whether this represents its total length. The retractor muscle cells of the 5.5 day larva possess an excitation- contraction coupling system that consists only of subsarcolemmal cistcrnae ot sarcoplasmic rcticulum coupled with the sarcolemma to form so-called peripheral couplings (Oliphant and Cloncy 1972). The retractor muscles o an older, oceanic larva of G. misakiana showed T-shaped invaginations ot the sarcolemma (Fig. 23) that make diadic contacts with the SR These retractor muscle cells were 6-7 f.im in diameter. The observed rate of retraction of the terminal organ in the 5 5 day larva was 220 ^m/s (one observafion). 9. Innervation of the Terminal Organ (Figs. 2, 15). Three subepidermal bundles of 8 10 neuntes each enter the neck and bulb of the terminal organ Kigs 19-24 TEMs of pelag?sphcra larvae. Kig. 19. Kpicmieular lllamenls ofplanklonic pcla- gosphcra ol undelermmed age. Fig. 20. Junclional eomplex between pcriloneal eells ol ihc larva in [?ig. 19. T he peritoneal cells arc non-myocpilhelial at this stage. Fig. 21. Sagittal section through the posterior trunk and terminal organ of 5.5 day ian-a. No^e that what i^ designated as the peritoneal lining ,s the longitudinal trunk musculature at this stai;e of development hig.22 Sagittal section ol coelomic part ot retractor muscle of terminal organ in 5 5 day larva. Hg. 23. Transverse section ot a terminal organ retractor muscle in planktonic larva ol undetcrmmed age. F.g. 24. Transverse section of a longitudinal trunk muscle of larva in lig. 2.1 Note the ingrowth of the sarcolemma in these older larvae, co. coelom- n, cuticle- ?7'. epidermis; nv. neuntes; /?^ peritoneum: m. retractor muscles; sr. sarcoplasmic rcticulum' /i>, terminal organ; rj-,/-rods 158 E.F.. Ruppcrt and M.E. Ric Fips I'^IO Sa"itlal reconstruction of a cellular anatomy ol' the tcrniinal o.-gan of 5^5 day hfv;?Y?m scr al TEM sections. Most nerve and muscle cells om.tled. F.gs. 25, 26. hptdcrn al ; Zie^tl cell (Fig. 26). Gaps Medicate m,ssn.g data. Fig. 27. Three mua. ^an^ ce . Fig. 28. Two tension-bearing cells. Fig. 29. Three porc-assocated senso.y cells. F,g. 30. Two cuticle-associated sensory cells from the trunk (Fig. 8). One bundle is ventrally situated and continuous with the ventral nerve cord. The other two bundles arc lateral to dorsolatenil in position, one on the right side of the organ, the other on the left. The neurites from the left and right bundles extend into the trunk and then curve ventrally where they probably join the ventral nerve cord ftom the right and left sides of the body. D. Discussion /. Funclional Integration of Cells in the Terminal Organ as an Organ of Larval Adhesion Although all cells in the terminal organ, except the muscle cells, arc ectoder- mal derivatives, it is apparent that only the 8 epidermal cells ftmction to maintain the integrity of the organ and to secrete the cuticle (Figs. 25, 26). Sipunculan Terminal Organ The mucus cells are implicaled as the adhesive producing cells because ol their considerable volume of vesicles, the relation of the gland necks to the pore of the terminal organ, and the observed breakdown of vesicles into the pore (Fig. 12). Mucus is known to possess adhesive properties (Hui- dity, surface activity, cohesion: Dahlquist 1977; Tyler 1976- Riegcr and Tyler 1979). We presume that the observed attachment strand (pp 6) sec- reted by these larvae is a mucus thread. That attached larvae can spin while anchored on their attachment strands suggests that either intermolecu- lar shearing occurs within the strand or, more likely, that the strand simply twists as the animal rotates. The tension-bearing structures appear to be the tension-bearing cells and perhaps part of the surrounding cuticle. This interpretation seems rea- sonable because the apical surfaces of the anchor-cells constitute most of the pore surface and their microvilli contain bundles of microfilaments that extend basally to join directly to the retractor muscles via junctions across the extracellular matrix. The two sensory cells with intracuticular cilia would seem to function as cuticular strain receptors, a possibility supported not only by the general position of the cilium of one cell and the basal body of the other but by the orthogonal orientation of the cilium of one cell and the basal body ot the other within the cuticle (Fig. 30), This orientation of receptoral cilia flanking the pore could allow the animals to monitor cuticular strain in any direction. As discussed earlier in the paper, the retractor muscles are responsible lor retraction of the terminal organ into the trunk of the animal flexion of the terminal organ and for bearing tension when the larva is attached to a substratum. The specific function, or functions, of the 3 remaining sensory cells associated with the pore of the terminal organ are uncertain. Because larvae are selective in the substrata to which they will attach, it is possible that these receptors evaluate the suitability of substrata for adhesion. Alternative- ly, the spatial relationship of their cilia to the pore of the terminal organ suggests that they may be incorporated into the attachment strand as it IS secreted and could serve therefore to monitor strain in the tether The most difficult aspect of function to determine is the mode of release of the terminal organ from the substratum. On this point our data do not allow us to decide whether this event occurs mechanically or chemically or both. Mechanical release might be accomplished by muscular contraction' as in a sudden ?exion of the terminal organ, or swimming movements that could tear the larva from its point of attachment. Evidence for the possibility of chemical release is the observation of numerous secretory vesi- cles in the anchor cells and in the 3 pore-associated sensory cells. These could possibly function as chemical rcleascrs. //. Comparison of the Terminal Organ with Adhesive Organs in Lower Metazou Discrete adhesive organs relying on mucus secreting cells are documented in Acocla and Nemertodermaiida among the Turbcllaria, in Ncmerlea, 160 K.E. Ruppcrt and M.E. Rice Gnalhostomulida, and perhaps some Nematoda and Rotifera (see Rieger and Tyler 1979 for review). Of these, the terminal organ ot the pelagosphera is generally similar to the rotifer adhesive organ [Philodina Dickson and Mercer 1966). Both share the following characteristics: (1) possession ol retractile, posteriorly-situated organs. (2) Mucoid material >s released from multiple symmetrically oriented glands at distinct pores. (3) Retractor mus- cle bands extend from the trunk into the adhesive organ to form junctions with the tension-bearing cells. (4) Variants of obliquely-striated muscle are nresent The adhesive organ of Philodina, however, differs in its organization ivom the pelagosphera organ in the following ways: (1) the organ terminates in two toes (2) There arc 12 gland cells with a longitudinal array of microtu- bules immediately below the plasmalemma of each cell neck and the necks do not terminate in a collar of microvilli. (3) The epidermis lacks an ex racel- lular cuticle and microvilli. The cells instead produce a dense apical layer of filaments resembling a terminal web. (4) The anchor cells show microyi - lar-likc processes but the cores contain a fiber resembling a ciliary rootlet. The processes may therefore be modified cilia rather than true microvilli. (5) A layer of circular muscle is present proximally in the organ. This com- parison indicates that although the roliferan and sipunculan atl^chment organs share similariUes at a general level of comparison, they are decided y dissimilar on closer examination. In the absence of any known designs to link the disparate organizations of these two organs, we conclude that there is a low probability of homology between the two, i.e. it appears likely that sipunculans and rotifers independently evolved their adhesive organs The other major design of adhesive organs in lower Metazoa is the duo-gland adhesive organ described originally by Tyler (1976). These organs consist minimally of two gland cell types, a viscid gland secreting the adhe- sive substance and a releasing gland secreting the chemical releaser. Al- though we cannot rule out the simultaneous occurrence of a duo-gland mechanism with a mucus adhering mechanism in the terminal organ system, we consider this unlikely for the following reasons: (1) there are no estab- lished examples of the simultaneous occurrence of these two systems in a single adhesive organ in any lower metazoan. (2) We have not seen evi- dence of release of any vesicles except those of the mucus glands of the nelagosphera. (3) On the basis of size, the vesicles observed in the anchoring cells and in the three pore-associated sensory cells qualify as vi.scid granules but^we have not found any cells that contain vesicles ^hj^^ ^?X'^M^rHn (0 1-0 2 urn) to previously described releasing granules (Tyler 1976, Mai tin 1978) We conclude, on the basis of current comparative data, that the terminal organ adhesive system of the pelagosphera larva is a design unique to the Sipuncula that shares similarities to adhesive organs of other meta- zoans only at the most general levels of comparison. ///. Significance of the Terminal Organ in Larval Biology and Life Histories of Sipunculans The terminal organ plays an important role in larval biology by enchancing the efficiency of larval feeding and enabling the exploitation o? dif?ercnt Sipunculan Terminai Organ sources of food. By attaching to Ihe substratum, the larva is able to feed both on the bottom surrounding its point of attachment as well as from the overlymg water column. While feeding on the bottom, the larval body IS bent so that the ventral head and mouth region arc applied to the substra- tum, thus enabhng the animal to graze on detrital material, algal coverings and bacterial films. The area of grazing in this attached position is limited by the extensibility of the larval body; however, the temporary nature of the attachment allows the larva to release itself and move to new and differ- ent feeding areas. When feeding from the water column, the larva maintains an upright position from its point of attachment with the ciliated mouth region directed upward to feed by a ciliary mucus mechanism on suspended particulate matter. A more direct contribution to the feeding process may be the provision of a source of food through the secretory activities of the terminal organ Surrounding the attachment of the organ there frequently accumulates a considerable amount of adherent debris held together by mucus secretions and consisting of particulate matter, feces, and bacteria. The mucus secre- tions, while entrapping particulate matter, also provide a medium for bacte- rial growth. A characteristic behavioral pattern of all sipunculan larvae IS the placement of the terminal organ in the mouth, and, although the signincancc of this behavior has not been documented, one result could be the ingestion of some of the adherent matter. Other possible functions of this behavior include a transfer of secretions (J?gerstcn 1963) or commu- nication between head and terminal regions as a means of testing the substra- tum prior to either attachment or settlement. Sensory structures alluded to earlier in this discussion may function in (lie selectivity of a substratum for attachment in younger larvae That some selectivity does occur is evidenced by the failure of young larvae to attach to sand grains or other granular material. Also young larvae are less likely to attach to the bottom of a clean glass dish than to one which has accumulated a thin film presumably of microorganismal origin For older larvae the sensory complex of the terminal organ could conceivably be utilized in the selection of a site for settlement, burrowing and subsequent metamorphosis. The changes in the terminal organ o? Golfingia misakiana in prominence and function during the course of larval developmental history reflect eco- logical adaptations and behavioral modifications that are important in a consideration of the overall life histories pattern of species. The young larval stage of Golfingia misakiana is well adapted for a benthic existence The strong attachment to the substratum, afforded by the terminal organ pre- vents It from being swept away by currents and allows it to feed from the bottom and the overlying water. As the larva grows the terminal organ IS proportionately reduced relative to body size and its function in attach- ment IS lost. At this stage the larva could then be transported by currents into the surface waters where it is commonly found in the major currents of the open ocean. Whereas the younger stage is adapted for feeding and growth, the older larva, with its well developed metatrochal band of swim- ming cilia, is well adapted for dispersal (Rice 1981). Although the attach- H ri. Ruppcrt and M.E. Ri? 162 ment of the terminal organ is lost m the older larva, it is speculated that the sensory function is retained possibly for substratum teslmg m prepara- tion for settlement. An evaluation must await further ultraslruclural studies of the organ in older larvae. 4ckno^,tcchmcn,s. This study was carried out at tlie Smitlisonjan Marine Station at Link P ,r Fort Picree Florida ^^m. The authors acknowledge Juhanne Pn-amo lor her able ?s'slane? in plc^atlll of ,Batcr,als lor electron mteroseopy and lor the product.on ol the scanning electron micrographs. References ?kesson B (1958) A study of the nervous system of the Sipcunculoidcae with some remarks on the development of two species Phascolion momhi Montagu and Goljmiim munaa Kele,- stein Unders?kningar over ?resund 38:1-249 ?ke n B (1961a) The development of Gol?mia dongala Kelerstem (S.puncuhdea) wUh loZ remarks on the development of neurosecrelory cells in s.puneuhds. Ark Zool Akcsson B (1961 b) Some observations on pelagosphera larvae. G?lalhea Rep 5:7^7 Dahluuist CA (1977) Adhesion an interdisciplinary science. Interdise Sc. Rev 2.14?-15U oSson MR Mercer Fll (1966) Fine structure of the peda] gland ol Philo?u,a roseola (Rot,- ?e,S- iTl'I^SlSS on'u^'embryology of the Sipuncul.dae, 11, The development of P/it,v((>/<)s')(>Ki. Zool.lahrb Abt AnatOntogTiere23-.77 162 Iiacto V (189X) Die pelag.schen Polychaeten - und Achaetenlarvcn der Plankton-ExpedU.on. Fr"cb Planklon-Expcd Humboldt-Stiftung 2; 1-50 Hall' IR. Schellema RS (1975) Comparative morphology ol open-ocean pclagosphaera. In Proe Intern Symp Biol Sipuncula and Echiura, Vol 1. Naueno Delo Press, Belgrade pp 18.,- Matl!hek B (1883) IJeber Entwicklung von Sipmuulus nudus. Arb Zool Insl Univ Wien Zool Slat Tiiest 5:61 140 ? , ,, ? -r , OA i ?7 Ikeda I (1904) The Gephyrea of .lapan. .1 Coll Sci Imperial Univ Tokyo 20.1-87 Jiigersten G (1963) On the morphology and behaviour of Pelai^osphaeru larvae (Sipuneuloidea). Kni;:'M?;?m neiroSiscular system. In: Physiology of annelids. Mill PJ (ed). Aca- demie Press New York, pp 161-206 M-.rtin GG (1978) The duo-gland adhesive system of the archiannelids ProuMus and Sacco- cinus and the turbellarian Monocrlis. Zoomorpholog.e 91 .63 -75 Minga/inni P (1905) Un Gefireo pel?gico: Fdagosphera Aloy.sn n. gen. n. sp. Rend Atad Oliph'mtrw Cloney RA (1972) The ascidian myocardium: sarcoplasmic retieulum and excita- tion-contraction coupling. Z Zeilforsch Mikrosk Anat 129:395 412 Renter M (1977) Ultrastructure of the stylet protractor muscle in Oyratnx hemaphrodaus (Tvirbellaria.Rhabdocoela). Acta Zool 58:179 184 ? ?? r- ir ? Rice MF (1967) A comparative study of the development of Phascolosonw agcssizii. ,o^wi,u, ?,!?.i;