FAU Institutional Repository

Home , Molgula, Molgula oculata, Molgulidae

http://purl.fcla.edu/fau/fauir

This paper was submitted by the faculty of FAU’s Harbor Branch Oceanographic Institute.

Notice: ©1988 Marine Biological Laboratory. The final published version of this manuscript is available at http://www.biolbull.org/. This article may be cited as: Young, C. M., Gowan, R. F., Dalby, J., Jr., Pennachetti, C. A., & Gagliardi, D. (1988). Distributional Consequences of Adhesive Eggs and Anural Development in the Ascidian Molgula pacifica (Huntsman, 1912). The Biological Bulletin, 174(1), 39‐46. Reference: Biol. Bull. 174: 39â€”46.(February, 1988)

Distributional Consequences of Adhesive Eggs and Anural Development in the Ascidian Molgula pacifica (Huntsman, 1912)

CRAIG M. YOUNG', RICHARD F. GOWAN2, JAMES DALBY JR.3, CATHERINE A. PENNACHETTI2, AND DAVID GAGLIARDI2 Bamfield Marine Station, Bamfield, British Columbia, Canada VOR JBO

Abstract. Molgula pacifica (Huntsman) is a recently Molgulidae (Lacaze-Duthiers, 1874; Damas, 1902; Ber rediscovered ascidian that occupies shallow subtidal rill, 193 1) and Styelidae (Millar, 1954, 1962), though rocks on wave-swept coasts ofBritish Columbia. Individ within the Subphylum Urochordata, tailed larvae are uals occur most abundantly at sites with intermediate ex also lacking in the development ofmost thaliaceans (Ber posure at or near 4 m depth. On a scale of centimeters, rill, 1950). Anural development has been previously ob they are highly aggregated. Molgula pacifica is hermaph served in 10 molgulid species (Berrill, 1931), eight of roditic, self-fertile, and oviparous. Embryos develop on which live on sandy or muddy bottoms. Only two uro the bottom without passing through a typical tadpole dele species, Molgula oculata and M. occidentalis, in stage. Each ofthe egg follicle cells contains a single large habit soft bottoms. Conversely, on hard substratum, 13 adhesive vacuole that occupies most ofthe cell volume. of the 15 molgulid species with known developmental Shortly after spawning these vacuoles rupture, causing modes demonstrate normal urodele development (Ber the follicle cells to secrete a sticky mucus coat that ad sill, 1931; Whittaker, 1979; Torrence and Cloney, 1981). heres the egg to the substratum. Juveniles hatch and Both known anural styelids, Pelenaia corrugata (Millar, move away from the chorion using epidermal ampullae, 1954) and Polycarpa tinctor(Millar, 1962), live on sandy as reported for other anural molgulids. Adhesive eggs substrata. Berrill (193 1) reasoned that urodele develop may be an adaptation that permits anural development ment is an ancestral condition that gave rise to anural in high-energy hard-bottom habitats. Egg adhesion may development among sand-dwelling species because lar also explain the small-scale distribution ofthe species. val swimming and habitat selection have little value where the substratum is flat and homogeneous. By exten Introduction sion ofthis argument, he suggested that the few attached

â€œ¿.. . it is very probable that(anural development) is to be anural species represent reinvasions by sand-dwelling correlated with the outstanding peculiarity ofthe family, forms of the ancestral hard-bottom habitat (Berrill, namely, its unattached sand-flat habitat and its adapta 1931). Whittaker's (1979) discovery ofvestigial tail mus tion to such an existence.â€•(Berrill,1931) cle acetylcholinesterase in anural species provides mdi Most ascidians pass through a â€œ¿urodeleâ€•or tadpole rect support for these ideas. Although we might expect larval stage. Suppression of the tailed tadpole (termed that anural species colonizing rocky substrata would â€œ¿anuraldevelopmentâ€•) is known only in the families somehow compensate for their inability to swim, no compensatory features have been described. Received23July 1986;accepted28October 1987. Huntsman (1912) described Molgula pacifica from a 1 Division ofMarine Science, Harbor Branch Oceanographic Institu single specimen collected at Ucluelet, British Columbia. ton, 5600 Old Dixie Hwy., Ft. Pierce, FL 34946 (author to whom cor Besides a redescription by Van Name (1945) based on respondenceshouldbe addressed). this same specimen, nothing more is known about the 2 Department of Biology, University of Victoria, Victoria, British Columbia, Canada V8W 2Y2. biology of this species. We recently discovered a large 3 Department ofBiological Science, Florida State University, Talla subtidal population of Molgula pacijica in Barkley hassee, FL 32306. Sound, British Columbia, approximately 35 km from the

39 40 C. M. YOUNG ET AL.

Figure 1. Map ofthe study region on the southeastern edge ofBarkley Sound, British Columbia, Can ada. Arrows indicate subtidal sites that were systematically surveyed for Molgula pacifica populations. Circles attached to the tails ofthe arrows give qualitative density values as follows: closed circle: abundant; open circle:absent;half-shadedcircle:presentin lownumbers. type locality. In this paper, we describe the general char At an intermediate scale, quantitative data were taken acteristics of this species' habitat, give a quantitative at the Blowhole site on the southern shore of Barkley analysis ofdistributional patterns on several scales, pres Sound, where Molgula pacj/ica occurred abundantly. ent a general description of its anural development, and Within each oftwo long surge channels (called Blowhole report an unusual developmental adaptation that may sites 1and 2 hereafter), we counted all individuals in five permit exploitation of high-energy subtidal habitats. A randomly positioned 50 X 50 cm quadrats at each of redescription of the species will be presented elsewhere three depths (6 m, 4 m, 2 m). Within each depth, three (Pennechetti et al., in prep.). habitat types were surveyed: (1) gentle slopes or horizon tal surfaces in channel bottoms, (2) vertical surfaces on Materials and Methods the sides ofchannels, and (3) horizontal or sloping ridges or plateaus between surge channels. The data for each Distributionalsurveys site were analyzed by 2-way ANOVA in which both fac We studied the distribution of Molgula pacifica at tors (depth and habitat) were fixed. three scales. Large-scale qualitative surveys were made at Small-scale (within-habitat) distribution was quanti 16 sites along a 10 km exposure gradient in Barkley fled from underwater photographs taken ofthe rock sur Sound, British Columbia. At each site, at least 4 divers face. At each of 4 sites (3 at Blowhole, 1 at Execution began at 10 m depth and worked up the slope, stopping Rock), we photographed a 6 X 6 grid of contiguous rec to search more carefully at 6 m, 4 m, and 2 m depths. No tangular quadrats on rock walls or steep slopes. Each quantitative data were taken on these survey dives. Field quadrat encompassed an area 15 X 22 cm. Transparen notes consisted of observations on sizes of individuals, cies were projected and counted by at least three individ animal and plant associates, depth ofoccurrence, surface uals before the data were compared with expected ran angles ofoccupied rocks, and wave surge. dom (poisson) distributions using goodness of fit tests. DISTRIBUTION AND ANURAL DEVELOPMENT IN MOLGULA 41 80 dissection, then macerated through 253 @smnitex mono Site 1 Ridge filament mesh. Other individuals released gametes spon Wall taneously in seawater tables. Following fertilization, ex Bottom cess sperm were removed by repeated rinsing with fresh 60 seawater. Cultures were maintained at lOÂ°â€”12Â°Cin a shallow, flow-through seawater table. We attempted to induce spawning by light shock by incubating eight mdi viduals in darkness for 12.0 h, then exposing them to the 40@ subdued light ofthe laboratory. Follicle cells ofdissected eggs did not secrete adhesive

C,. spontaneously. We induced holocrine secretion for his tological study by placing the eggs in hypotonic seawater E 20 (1 part seawater, 2 parts distilled water). Eggs treated in Ct) this way appeared identical to eggs spawned naturally in the laboratory. Embryos were fixed in Torrence's fixative for 2.5 h, > followed by three 20-minute rinses in Torrence's buffer 2m 4m 6m (Torrence and Cloney, 1981). Embryos were fixed, C rinsed, and dehydrated on ice, then brought to room 80 temperature during the first change ofabso!ute isopropa >@ Site 2 no!. They were embedded in Luft's Epon 812 (Luft, 1961) and sectioned on a Reichert OMU2 ultramicro U) C tome using glass knives. Thin sections were stained with a) 60 uranyl acetate, post-stained with lead citrate, and exam 0 med at 60 kV using a Phillips EM 300 electron micro scope. To aid in the recognition ofmucus, l-@imsections were stained with an aqueous solution of0.25% toluidine 40 blue in 0.5% sodium borate (Humason, 1967; R. Burke, pers. comm.). Living embryos and juveniles were observed and pho tographed on Leitz or Zeiss photomicroscopes using 20 phase contrast, bright field, or Nomarski optics.

Results Distribution 2m 4m 6m The distribution ofMolgula pacifica showedclearpat terns at three scales. Figure 1shows the large-scale distri Depth bution, as determined by qualitative surveys, within the southern portion of Barkley Sound and the Deer Island Figure 2. Densities ofMolgulapacifica at three depth/habitat com group. The mouth of the sound experiences the rough binations in two surge channel systems at the Blowhole site. Error bars represent one standard deviation on each side ofthe mean. Data analy surfand surge conditions ofthe open Pacific coast. Wave sis is presented in Table!. action decreases regularly along the southern shore of the sound because of protection from the Deer Islands and Broken Islands. The conditions at any given site in the Other species of solitary ascidians were also counted in Deer group depend on the distance of the site from the one grid (Blowhole grid 1)to determine iftheir densities mouth ofthe sound, protection by other islands, and di correlated with those ofMolgula pacifica. rection ofexposure. Individuals ofM. paafica were not found at either the most exposed sites or the most pro Embryology and histology tected sites. The sites at which they were most abundant Individuals of Molgula pacifica were collected by had moderate exposures with constant, year-round scuba from the Blowhole site in Trevor Channel, British surge. This pattern is apparent along at least three expo Columbia, during July 1985, and placed in seawater ta sure gradients (Fig. 1). Along the southeastern shore of bles at the Bamfield Marine Station within one hour of Barkley Sound abundances went from zero at Seapool collection. Gonads were removed from some adults by rocks to high density at Execution Rock and Blowhole, 42 C. M. YOUNG ET AL. Table! distributions differed significantly from expected ran Analyses(channelwall,channelbottom,ofvariance testingthe importanceofhabitat dom distributions. Numerous quadrats with zero values m)on ridgebetweenchannels)anddepth (2,4,6 and quadrats on the upper tails of the observed fre emsatdensitiesof Molgula pacificain twochannelsyst quency distributions (densities as high as 727.2 animals the2)SourceBlowholesite(Fig. per m2 in individual quadrats) indicate aggregated spatial distributions. To visualize the nature of the clumps, we ofvariation FPSite d.f. SS MS depict the raw data as three-dimensional plots (Fig. 4). Each grid encompassed a rectangular area 1.188 m2. 1:Depth Gumps were gene@ailyon the scale of several quadrats 23.550.000Habitat 2 2.763 1.381 (i.e., tensofcentimeters). At Execution Rock, on agently 2.220.123DepthXhabitat2 0.260 0.130 6.330.001Error 4 1.485 0.371 sloping rock face, there was a sharp density gradient that â€”¿â€”Site 36 2.112 0.058 went from a region with few animals in one corner to a single large aggregation that occupied most of the re 2:Depth 29.030.000Habitat 2 2.048 1.024 mainder ofthe plot. 1.300.284DepthXhabitat2 0.092 0.046 To determine if the aggregations of Molgula pacijica 14.810.000Error 4 2.089 0.522 occupied the same sites as aggregations ofascidians with 36 1.270 0.035 â€”¿â€” tailed tadpoles, we computed correlation coefficients in which M. pac@ticaabundances were paired with abun dances ofthree other species occurring at Blowhole grid then dropped to zero at Bamfield Inlet and Dixon Island. 1. Two of the species, Pyura haustor and Chelyosoma On the leeward side of the Deer group, where the expo productum, arestrongly gregariousaslarvae,whereasthe sure gradient is more abrupt, intermediate numbers of third, Cnemidocarpa finmarkiensis, settles randomly animals were found at Edward King Island, but none were found on the southern tip ofDiana Island or on the Eastern Shore ofHelby Island. Finally, on the windward G â€”¿7.73 side ofthe Deer group, no animals were found on an ex p > 0.05 posed northern tip of Edward King Island, intermediate numbers were present on Seppings Island, high numbers were found on Diana Island, and abundances dropped to zero on the northwestern sides of Wizard Islet and Fleming Island. â€”¿ The two parallel surge channels surveyed at the Blow

hole site (Fig. 2) had different overall densities of Mol G â€”¿20.70 P < 0.01 gulapacifica, so the distributions were analyzed with sep 0 V 0 arate two-way analyses of variance (Table I). The same a general pattern was apparent at both sites. The main fac 0

tor of depth and the depth X habitat interaction (where B$owho4eG.id 3

@ habitat refers to channel bottoms, sides, or the ridges be 7 @ tween channels) were significant in both cases. At site 1, 8 G â€”¿21.75 C 5 8, P < 0.01 about 6 times as many M. pacifica occurred at 4 m depth 3 4 ! 3 than at either 2 m or 6 m. Qualitatively, the same pattern Ij.. 2 was seen at site 2. The interaction at site 1 resulted from S 1 2 S 5 1 5 8 10 11 12 1@ most individuals occurring on the channel walls at 2 m Ex.cution Rock and 6 m, but more occurring on channel bottoms and ridges at 4 m (Fig. 2). The distributions are similar be 1 G â€”¿41.62 tween sites at 6 m, but differed substantially at 2 m and P < 0.001 4 4 m. At site 2, very few individuals occurred on channel walls at 2 m. At 4 m, approximately equal numbers oc

curred on the walls and channel bottoms (Fig. 2). 0 1 2 3 4 S 6 7 9 9 10 11 12 13 14 15

Frequency distributions of the abundance data from Number of Individuals the small-scale photographic grids were compared with poisson distributions to determine the nature and sig Figure 3. Observed (shaded bars) and poisson (open bars) fre quency distributions of Molgula pacj/ica in small-scale photographic nificance of small-scale clumping (Fig. 3). With the ex grids. In computing the goodness offit (G) values for each site, classes ception of Blowhole grid 1, which had a much lower with expected frequencies less than 2.5 were combined on the tails to overall density than the other three, all of the observed make the significance tests more conservative. DISTRIBUTION AND ANURAL DEVELOPMENT IN MOLGULA 43

Blowhole Grid 1 Blowhole Grid 3

C/) 0 -D > C Blowhole Grid 2

â€˜¿4- 0 Execution Rock a) .0 E z

Quadrat Coordinates

FIgure 4. Small-scale abundance patterns of Molgula pacifica at four sites.

with respect to established individuals (Young, 1982; material showed a few test cells present (R. Qoney, pers. Young and Braithwaite, 1980). None ofthe correlation comm.). Each follicle cell contains a single large vacuole coefficients were significant (P. haustor: r = 0.005; C. which occupies most of the cell volume. Shortly after productum: r = â€”¿0.002;C.Jinmarkiensis: r = â€”¿0.244;P spawning, these vacuoles undergo a holocrine secretion > 0.05 for all three). of their contents to form the adhesive coat (Fig. 7, 8). This coat, which stains intensely with toluidine blue, Reproduction and development swells to occupy a space many times the volume of the original follicle cell (Fig. 9). Those portions ofthe follicle Unlike many ascidians that spawn during the day fol cell membranes not disrupted by the secretory process lowing dark adaptation, Molgula pacifica released ga remain closely applied to the chorion after secretion (Fig. metes during periods of darkness. All eight individuals 8), but all traces ofthe follicle cells disappear by 24 h after tested spawned, and all released gametes during, not aS fertilization. ter, the dark adaptation period. Approximately 100-200 All but a few ofthe naturally spawned eggs were firmly eggs were released by each individual. Adults were main attached to the substratum at the time they were discov tamed in separate dishes duringthese experiments, yet all ered. Many of these were clumped together and joined embryos were undergoing development when they were by a common adhesive coat (Fig. 9). discovered in the morning. In dissecting numerous speci Eggs removed from the ovaries by dissection did not mens we never found embryos being brooded within the produce adhesive coats spontaneously. Nevertheless, atrial chamber. Thus, M. pac@/icais an oviparous her many were sticky and their follicular vacuoles could be maphrodite capable of self fertilization. induced to release their contents by exposure to a hypo The eggs are 180 @min diameter, including the follicle tonic medium. Following secretion, such eggs appeared cells. At spawning, the egg has a single outerlayer of folli identical to naturally spawned eggs with adhesive coats. dc cells surrounding the chorion (Fig. 5). The chorion is Complete embryonic development occurs within the closely applied to the egg plasma membrane; thus, the adhesive coat; no tailed tapole larva is ever produced. perivitelline space is much smaller than in most typical The early details of anural development in this species oviparous ascidians (Fig. 6). Sections ofthe eggs revealed are obscured by the opacity ofthe adhesive coat and em no test cells (Fig. 5, 6), but careful observation of living bryo. At 10â€”12Â°C,the first juveniles hatched 36 h after @..

Figure 5. One-micron section ofnaturally spawned 2-cellembryojust priorto secretion ofthe adhesive coat. Note the very thin perivitelline space and apparent absence oftest cells. Stained with toluidine blue. F: follicle cell. C.P.: cleavage plane. C: chorion. Figure 6. Electron micrograph ofa follicle cell prior to secretion ofadhesive coat. V: vacuole. C: cho rion. Y: yolk granule. Figure 7. Thick (l-@m) section ofan egg following secretion ofthe adhesive coat. AC: adhesive coat. FM: follicle cell membranes. Figure 8. Electron micrograph ofan artificially removed eggjust after secretion ofadhesive has been induced by exposure to hypotonic seawater. C: chorion. P: plasmalemma. AC: Adhesive coat. Y: yolk granule. Figure 9. Aggregation ofnaturally spawned eggs(E) bound together by common adhesive coat (AC). Figure 10. Juvenile shortly after hatching. A: ampulla. T: tunic. DISTRIBUTION AND ANURAL DEVELOPMENT IN MOLGULA 45 fertilization. Just before hatching, a variable number (1â€” drooping blades. The spores flow down the grooves and 5) of epidermal ampullae were extended onto the sub fall on or near the haptera ofthe parents, where they ad stratum (Fig. 10). As in other molgulids, embryos hatch here (Dayton, 1973). by rupture ofthe chorion. One ampulla was often larger Some ascidians with active tadpole larvae settle gregar than the others, and juveniles generally extended this iously, whereas others settle randomly (Young, 1982; ampulla through the initial fissure in the chorion. No Young and Braithwaite, 1980). Both kinds ofspecies oc embryos were observed to hatch before any ampullae curred in the same habitats as Molgula pacifica. Chelyo had formed, but some hatched with only a single am soma productum, which aggregates behaviorally in vitro pulla. The broken adhesive coats remained attached to (Young and Braithwaite, 1980) and in the field (Young, the substratum for several days after hatching, then dete 1982) formed clumps at about the same scale as those of riorated. M. pacifica at our study sites. However, the densities of the two species were not correlated, suggesting that the Discussion aggregations are formed and maintained independently. M. paczjica densities were also not correlated with those Molgulapaa/ica is one ofonly three molgulids known of Pyura haustor, another gregarious species, nor with to have anural development while occupying hard sub Cnemidocarpafinmarkiensis, a species that does not ag stratum, and it is the only known anural species in the gregate. Ifsome sites were consistently better for juvenile Pacific Ocean. It is particularly surprising that the species survival than others, or if current patterns concentrated lives in wave-swept habitats. We propose that the adhe propagules in certain regions, we would expect positive sive eggs may be an adaptation that permits this unex correlations among ascidian species since neither of pected habitat distribution, as the eggs probably stick to these processes should discriminate among tadpoles of surfaces near the parents and develop on the bottom. different species. The low correlation coefficients do not Tadpoles of other species living in this habitat presum support the idea that ascidians accumulate passively. ably use their swimming ability to assist in reaching the Thus, clumps of Chelyosoma productum and Pyura substratum. haustor are probably established by tadpole behavior, The highly clumped small-scale distribution may be whereas similar clumps of M. pacifica are probably explained by two mechanisms associated with egg adhe formed by philopatric dispersal mediated by sticky eggs. sion. First, eggs frequently attach to each other after In the egg of a typical urodele ascidian, test cells are spawning; indeed, they often share a common adhesive located within superficial concavities ofthe oocyte (Kes coat. Such egg masses falling to the bottom would place sel and Kemp, 1962). A single layer ofinner follicle cells numerous individuals in close proximity. Although am is separated from the oocyte and test cells by the chorion. pullar locomotion probably disperses individuals away Prior to ovulation, a layer ofouter follicle cells surrounds from the initial attachment site at hatching, such loco the inner follicle cell layer (Kessel, 1983). At ovulation, motion would not be expected to obliterate the clumped the outer follicle cell layer remains in the ovary with the pattern. Aggregations produced in this manner by sib germinal epithelium, the chorion lifts away from the lings should be dominated by single size classes. Second, plasma membrane of the oocyte, and the test cells are local aggregations could be established and maintained extruded from their superficial concavities and come to by limited (philopatric) dispersal. It seems reasonable to lie within the perivitelline space. The eggs ofsome anural assume that eggs should be more likely to contact the ascidians (e.g., Molgula bleizi) contain test cells and a substratum near their parents than further away. These perivitelline space, whereas the eggs of others (e.g., M. juveniles in turn would produce short-distance dispers retortiformis) have neither(Berrill, 1931).The eggs of M. ing offspring oftheir own, resulting in aggregations with pacjfica have very few test cells lying within a perivitel polymodal size distributions. We were unable to make line space much smaller than that seen in typical urodele precise size measurements of individuals in our photo species. graphs because of adjacent epifauna. Nevertheless, it is The functions ofthe follicle cell layer vary among eggs clear from the photos that clumps tend to contain ani from different ascidian species. The follicle cells of Co mals of all sizes, suggesting that the second method of rella infiata are filled with ammonia and cause the eggs to aggregation could be important. float (Lambert and Lambert, 1978). Highly vacuolated Phiopatric dispersal is known in many other sessile follicle cells of other species probably reduce the egg's organisms (reviewed by Jackson, 1986), including some specific gravity to slow the sinking rate of the egg (Har that occupy high-energy habitats. For example, the inter vey, 1927; Berrill, 193 1). Kessel and Kemp (1962) de tidal alga Postelsia palmeformis, which lives only on ex scribed a secretory product in the follicle cells of Ciona posed headlands, often forms aggregations containing in intestinalis and Molgula manhattensis. However, these dividuals of all sizes (Dayton, 1973). During low tides, secretory masses break down during oocyte maturation spores are released from sori located in grooves of the and are not present at ovulation; at spawning, the eggs 46 C. M. YOUNG ET AL. contain a granular material ofunknown function. M. pa Berrill, N. J. 1950. The Tunicata, With an Account ofthe British Spe@ cifica is the only speciesreported to have secretoryfolli des. Ray Society,London. 354pp. cle cells after ovulation and the only one whose follicle Damns, D. 1902. Recherces sur le developpement des Molgules. Arch. Biol. 18:599â€”664. cells function in attachment. The cells themselves are es Dayton, P. K. 1973. Dispersion, dispersal, and persistence ofthe annual sentially destroyed when the adhesive material is se intertidal alga, Postelsia palmaeformis Ruprecht. Ecology 54: 433â€” creted. 438. Lucas (1927) reported that the eggs ofMolgula robusta Harvey, L. A. 1927. The history ofcytoplasmic inclusions ofthe egg of are often held together by strings of adhesive mucus Ciona intestinalis during oogenesisand fertilization. Proc. R. Soc. which facilitate dispersal by reducing sinking rates. Lond.B1O1:137â€”161. These strings could be wafted from the bottom of the Humason, G. L. 1967. Animal Tissue Techniques, 2nd Edition. W. H. culture dishes by stirring, a feat that would be impossible Freeman and Co., San Francisco. 569 pp. Huntsman, A. G. 1912. Holosomatous ascidians from the coast of west with the securely attached eggs ofM. pacifica. era Canada. Pp. 103â€”185in Contributions to Canadian Biology. Kupffer (1875; cited in Berrill, 1931) argued that an C. H. Parmelee, Ottawa. ural development was the forerunner ofurodele develop Huntsman, A. G. 1922. The ascidian family Caesiridae. Proc. R. Soc. ment, and Lacaze-Duthiers (1877) classified anural and Canada16:211â€”234. urodele species in separate genera. However, virtually all Jackson, J. B. C. 1986. Modes of dispersal of clonal benthic inverte later authors consider direct development to be the de brates: consequences for species' distributions and genetic structure rived state (Berrill, 193 1). Anural development is proba oflocal populations. Bull. Mar. Sci. 39: 588â€”606. bly polyphyletic (Berrill, 1931). It has arisen indepen Kessel, R. G., and N. E. Kemp. 1962. An electron microscope study of the oocyte, test cells, and follicularenvelope ofthe tunicateMolgula dently in at least two different families (Styelidae and manhattensis.J. Ultrastruct.Res.6: 57â€”76. Molgulidae), and within the Molgulidae it has probably Kessel, R. G. 1983. Urochordataâ€”Ascidiacea. Pp. 655-734 in Repro arisen in at least four different clades (Berrill, 1931; ductive Biology oflnvertebrates, Vol. I, K. G. Adiyodi and R. G. Huntsman@ 1922). Berrill (193 1) argues convincingly Adiyodi, eds. John Wiley and Sons, New York. that in each case anural development arose as an adapta Lacaze-Duthiers, F. J. H. de. 1874. Histoire des ascidies simples des tion to sedimentary environments, where neither swim cotes de France I. Arch. Zool. Exp. Gen. 3: 119â€”656. ming nor habitat selection had a strong selective advan Lacaze-Duthiers, F. J. H. de. 1877. Histoire des ascidies simples des tage. It follows then that the few anural species occurring cotesde France II. Etudes desespeces. Arch. Zool. Exp. Gen. 6: 457- 676. on hard bottom represent reinvasions ofthe original hab Lambert, C. C., and G. Lambert. 1978. Tunicate eggs utilize ammo itat by species that had evolved the soft-bottom develop nium ions for flotation. Science 200: 64â€”65. mental mode. In Molgula pacifica, adhesive eggs may be Lucas, A. M. 1927. The validity ofMolgula robusta (Van Name) as a one character that facilitated such a reinvasion and that species distinct from Molgula manhattensis (DeKay). Occas. Pa allowed the species to recruit and maintain populations persBostonSoc.Nat. Hist. 5: 243â€”246. on hard bottoms in surgy habitats just as well as species Luft, J. H. 1961. Improvements in epoxy resin embedding methods. J. with typical tadpole larvae. Biophys.Biochem.CytoL9: 409â€”414. Miller, R. H. 1954. The breeding and development ofthe ascidian Pel. Acknowledgments onaia corrugateForbesand Goodsir. J. Mar. BioL Assoc.U.K. 33: We thank R. E. Foreman, director of Bamfield Marine 681â€”687. Station, and R. L. Shimek, associate director, for provid Millar, R. H. 1962. The breeding and development of the ascidian ing lab facilities and logistical support for the Ascidian Bi Polycarpatinctor. Q.J. Microsc. Sci. 103:399â€”403. ology course during which this project was initiated. A. Torrence, S. A., and R. A. aoney. 1981. Rhythmic contractions of the ampullar epidermis during metamorphosis ofthe ascidian Molgula Bergey, R. Shimek, and S. Smith assisted with field work. occidentalis.Cell TissueRes.216:293â€”312. J. L. Cameron, B. L. Bingham, R. R. Olson, and two Van Name, W. G. 1945. The North and South American ascidians. anonymous reviewers improved the manuscript with BUlLAm. Mus. Nat. Hist. 84: 1-462. their comments. J. L. Cameron and B. L. Bingham also Whittaker, J. R. 1979. Development ofvestigial tail muscle acetylcho assisted with data analysis and computer plotting. We are linesterase in embryos ofan anural ascidian species. BioL BulL 156: especially grateful to R. A. Cloney for his timely assistance 393â€”407. during the course and his critical review of an earlier Young, C. M. 1982. Larval behavior, predation, and early post-settling manuscript. Harbor Branch contribution number 613. mortality as determinants ofspatial distribution in subtidal solitary ascidians of the San Juan Islands, Washington. PhD Dissertation, Literature Cited University ofAlberta. 260 pp. Berrill, N. J. 1931. Studies in tunicate development, II. Abbreviation Young, C. M., and L. F. Braithwaite. 1980. Larval behavior and early ofdevelopment in the Molgulidae. Phil. Trans. R. Soc. B. 219: 281â€” post-settling morphology in the ascidian Chelyosoma productum 346. Stimpson. J. Exp. Mar. Biol. EcoL 42: 157â€”169.