Metamorphosis of the marine gastropod Phestilla Sibogae Bergh (nudibranchia: aeolidacea). I. Light and electron microscopic analysis of larval and metamorphic stages

J. exp. WKW.Biol. Ecol.,

METAMORPHOSIS

1974, Vol. 16, pp. 227-255;

OF THE MARINE

Bergh (NUDIBRANCHIA: MICROSCOPIC

ANALYSIS

c

North-Holland

GASTROPOD

AEOLIDACEA). OF LARVAL

Publishing

Company

PHESTZLLA SZBOGAE

I. LIGHT AND ELECTRON AND

METAMORPHIC

STAGES

DALE B. BONAR’ Department

of Zoology,

Uniwrsity

of Hawaii.

Honolulu,

Hawaii

and

MICHAEL G. HADFIELD Kewalo

Marine

Laboratory,

Pacific

Biomedical

Research Hawaii

Center,

Unicersity

of Hawaii,

Honol~du,

Abstract:

The structure and fate of transitory larval organs (velum, shell, operculum, retractor muscles, part of the epidermis) of Phestilla sibogae Bergh were studied before, during, and after metamorphosis with both light and electron microscopy to elucidate the morphology of these organs and the mechanisms by which they are lost. Loss of the velar lobes is the first morphological sign of metamorphosis, and involves selective dissociation and subsequent ingestion of the ciliated velar cells; the remaining aggregate of supportive cells is apparently incorporated into cephalic epidermis. Attachment of the larval body to shell and operculum is primarily at sites of retractor muscle insertions: once the velum is gone, the attachment between shell and larval body is lost and the shell is cast off as the visceral organs exit through the shell aperture. Merger of visceral and cephalopedal elements results in flattening of the postlarval body and reorientation of internal organs. Simultaneously, a rapid spreading of epipodial epidermis over the lateral, dorsal, and posterior sides of the body produces the definitive integument. The squamous cells which comprise the larval perivisceral epidermis are pushed ahead of the definitive epidermis and are seen shortly after the shell is cast as a constricted aggregate of cells on the posterior end of the body. Autolysis of the left and right retractor muscles begins during metamorphosis and no trace of them is left after 24 to 48 h. The metapodial mucous glands which hypertrophy before metamorphosis are also lost within 48 h following exit of the post larva from the shell. Metamorphosis produces a detorsion caused in part by muscular action and in part by continuing growth and development INTRODUCTION

For most invertebrate groups, some behavioral and ecological aspects of metamorphosis are generally known, but anatomical and morphological studies are few. Ultrastructural investigations of metamorphosis have been reported for very few marine invertebrates, those mainly for ascidians (Cloney, 1963, 1966, 1969, 1972; Cloney & Grimm, 1970; Ishikawa and Numakumai, 1972; Ishikawa et al., 1972), and the ectoprocts (Woollacott & Zimmer, 1971). Many studies of gastropod metamorphosis have been reported but few were histologically oriented: the most detailed investigations are those of Crofts (1937, 1955) Fretter (1972) Smith (1935), and Thiriot-Quitvreux (1967, 1969) for the proso’ Present U.S.A.

address:

Department

of Biology,

University 227

of Maryland.

College

Park.

Maryland

DALE B. BONAR AND MICHAEL

228

branchs,

and Horikoshi

and Thompson (1958, extensive, investigations

(1967)

Rao (1961), Tardy

G. HADFIELD

(1970), Thiriot-Quievreux

1962) for opisthobranchs. Complete reviews in these groups are given in Thiriot-Quievreux

( 1970)

of other, less (I 969) Tardy

(1970) and Russell (1971). The investigations of Thompson and Tardy are most pertinent to this study since they also deal with nudibranch gastropods. Phestilla

sibogae

Bergh,

an aeolidacean

nudibranch,

was chosen

for an ultra-

structural and cytological study of gastropod metamorphosis for several reasons. The animals are easily cultured in laboratory aquaria, spawn throughout the year, have a short non-feeding larval stage, and can readily be induced to metamorphose. The life history, behavior, and ecology of P. sibogae in Hawaiian waters have been described by Harris (1970). Adult P. sibogae feed on the scleractinian coral Porites compressa Dana and lay their egg masses on the undersides of the coral heads. The veliger stage is reached within the egg mass three days after spawning, and natural hatching occurs about the ninth day (Hadfield, 1972). Larvae become competent to metamorphose about nine days after oviposition, so that newly hatched larvae are able to metamorphose shortly after hatching. Once competent to metamorphose, they will do so only in the presence of living P. compressa. The active factor inducing metamorphosis is probably part of the coral mucus (Hadfield & Karlson, 1969) and presence of the coral mucus alone in the culture milieu will stimulate metamorphosis. Metamorphosis of Phestilla sibogae follows a pattern in which selective destruction of tissues specialized for larval life allows the continued development of the adult form. Our work was undertaken to characterize the transitory larval tissues and to analyze their destruction or transformation during metamorphosis, Since there are no descriptions of larval gastropod ultrastructure in the literature, much of the research was of necessity an anatomical analysis. Subsequent papers (in preparation) will describe, in greater cytological details, specific phases of the metamorphic process and experimental procedures utilized to clarify basic mechanisms underlying the substantial histolysis and morphogenesis which occur during this period of rapid developmental change. MATERIALS AND METHODS oBTAINING THE LARVAE

Adult P. sibogue were kept continuously cultured in aquaria supplied with fresh, flowing sea water. Periodic replacement of Porites compressa with living coral heads ensured a constant food supply for the nudibranchs. Newly laid Phestillu sibogae egg masses were collected from the undersides of coral heads and placed in mesh cages which were then submerged in the aquaria. After 5 days the egg masses were rinsed and removed to small dishes with sea water filtered through 0.22 or 0.45 pm Millipore filters. The egg masses were teased open with forceps, allowing the larvae to escape from their capsules. Newly hatched larvae are positively phototaxic: a bright light placed on one side of the dish ‘attracts’ the larvae

METAMORPHOSIS

which may then be pipetted

OF AN AEOLlD

off into a culture

NUDIBRANCH

beaker of Millipore

229

filtered sea water

stirrer, (MPF-SW). Th e culture beakers were placed under a low speed, reciprocating and the larvae collected from the cultures for examination and experimentation as needed. Cetyl alcohol (Hexadeconal) was often sprinkled on the water in the culture vessels to lower surface tension, since the larvae were otherwise easily trapped at the air water interface

(Hurst,

1967). A stock mixture

of 0.5 mg penicillin

G (Sigma)

with 0.6 mg streptomycin sulfate (Sigma) in 100 ml of distilled water added to larval cultures in proportions of 1 ml antibiotic stock per lOOm1 culture effectively controlled bacterial

infestation.

LIGHT MICROSCOPY

In preparingtissues for histological examination, it was necessary to ensure adequate penetration of fixative by anesthetizing larvae prior to fixation so they would not retract into the shell and seal the aperture with the operculum. A mixture of 3 parts MPF-SW to 1 part Chloretone-saturated distilled water effectively relaxed animals within 5 to 10 min. For routine histology, anesthetized larvae were fixed at successive stages of metamorphosis in Bouin (made up in MPF-SW) or in 4 “/, formalin in I y. CaCI,, dehydrated in ethyl alcohols, cleared in toluene, and embedded in paraffin or resin. For light histochemistry, a hydroxyethyl methacrylate-polyethylene glycol embedding medium (Ruddell, 1967) was most satisfactory. The resin was found to polymerize best when only 1/2 or 2/3 of the recommended amount of accelerator was used. A thin layer of Tacky-Wax (Central Scientific Co.) applied to the upper and lower surfaces of the trimmed block facilitated sectioning of the 1 to 2 pm sections cut with either steel or, preferably, glass knives (Ruddell, 1967). Staining and histochemical methods used included hematoxylin-eosin, periodic acid-Schiff’s (P.A.S.), alcian blue series, mercury-bromphenol blue, and acroleinSchiff’s (Pearse, 1968). ELECTRON MICROSCOPY

The most effective fixation for ultrastructural preservation was obtained with 2.5 “,:, glutaraldehyde in 0.2 M Millonig’s phosphate buffer brought to 960 milliosmoles with 0.14 M NaCl, followed by postfixation in 1 or 2 y0 0~0, in 1.25 % sodium bicarbonate buffer (pH 7.4), all at room temperature (Cloney, 1972). Dehydration via alcohols preceded embedding in Epon (Luft, 1961). Embedding was done in flat molds so that larvae could be oriented in the resin prior to polymerization. Larval stages with shells required decalcification when using non-acid fixatives. After an initial fixation period of 30 min in glutaraldehyde, 10 “/: disodium EDTA was added in a I : 1 ratio with fixative to make a final concentration of 5 7: EDTA. As soon as the shells were decalcified (usually after 20 to 30 min) the larvae were washed twice in fixative before proceeding to postfixation and dehydration. Thin sections for electron microscopy were cut with a glass or diamond knife on a Reichert OM-U2 ultramicrotome and collected on loo-mesh formvar-coated grids or

230

on 300-mesh

DALE

uncoated

B. BONAR

AND MICHAEL

grids. Sections

were stained

G. HADFIELD

for 5 to 10 min in 50 9,; ethanol

containing 2 % many1 acetate followed by 2 to 4 min in lead citrate (Reynolds, and examined on a Philips 201 electron microscope. Thick (0.75 to 1.0 pm) sections

taken alternately

with thin sections

1963),

were mounted

on glass slides and stained with 1 yd azure II and 1 “//, methylene blue in 1 p;, sodium borax solution (Richardson, Jarett & Finke, 1960). These sections were used for orientation

as well as general

histological

analysis.

RESULTS STAGES OF METAMORPHOSIS

Metamorphosis of P. sibogae is an ordered sequence of events, both constructive and destructive, which convert a shelled, swimming larva into an elongate, shell-less, benthic juvenile. The events of metamorphosis are interde~ndent and must proceed in their natural order for normal development. Superficially, however, these events are visibly distinguishable and allow construction of an artificial staging of metamorphosis. This staging is based on anatomical changes easily visible with a dissecting microscope, and was intended for practical convenience. Arbitrarily defining set points in the metamorphic sequence produced a system by which larval and metamorphic stages could be fixed and compared with some degree of constancy (Bonar, 1972). As ilIustrated in Fig. 1, there are seven easily recognizable stages, labeled here A through G. Stage A larvae are newly hatched individuals which are not yet competent to metamorphose. As a larva develops metamorphic competence, it becomes a stage B larva, physicaIiy recognizable by the inflated propodium, slightly reduced visceral mass, and increased retraction of the mantle fold from the shell aperture. Larvae remain in this stage until they metamorphose or perish; the length of this stage is dependent on environmental factors. At this stage, larvae which were positively phototaxic at hatching become negatively phototaxic and undergo the swimmingcrawling behavior reported for other settling larvae (Thompson, settled larva is now exposed to Porifes compressa, metamorphosis field & Karlson, 1969). The first visible feature of metamorphosis

1958, 1962). If a is induced (Hadis the loss of the

velum: such larvae are termed stage C. Transition from stage B to stage C is abrupt, loss of the velum occurring in 10 to 20 min. Within 20 to 60 min after velar loss, the attachment of the Larval body to the shell is lost, giving a stage D individual. At this time the animal crawls about the substratum on its foot and after a variable period of time (usually less than 30 min) casts off the shell and operculum, entering stage E. During the next 30 to 60 min, the visceral hump becomes incorporated more smoothly into the body, forming the stage F postlarva. Continued development involves flattening and elongation of the body until within two days stage G, the pseudovermis or limapontiform stage (Tardy, 1970), is attained. Development of tentacles, rhinophores, and cerata will produce the typical adult form.

METAMORPHOSIS

Ill

z -

f=!

u

OF AN AEOLID NUDIBRANCH

DALE

232 STRUCTURE

OF THE

LARVA

B. BONAR PRJOR

AND

MICHAEL

G. HADFIELD

TO METAMORPHOSJS

Shell, oelum, and head

The inflated, Type-2 larval shell (Thompson, 1961) measures 240 pm x 160 pm, shows no evidence of sculpturing, and does not increase in size during larval life (Fig. 2). The veliger larva is able to withdraw its body entirely into the shell, sealing the aperture with the simple, unadorned operculum. The most distinctive feature of the gastropod veliger is the velum, a cephalopedal locomotory and feeding organ.

Fig. 2. Larval

shell of Phestillu sibogae: A, right lateral

view; B, ventral

view:

x 160.

A detailed description of velar structure and function in Phestilla sibogae, as well as its fate at metamorphosis, will be given in another paper (Bonar & Hadfield, in prep.). This transitory larval organ, an outgrowth of the cephalic region, arises from the larval body dorsolaterally to the mouth. From the mouth, the esophagus passes between and slightly beneath the partially fused cerebral ganglia and opens into the anterodorsal surface of the stomach. Just behind the cerebral ganglia an evagination of the floor of the esophagus forms the radular pouch which contains 3 or 4 radulat teeth in competent, stage B larvae. Dorsally, beneath the epidermis and just behind the origin of the velar lobes, the pair of darkly pigmented eyes is visible. Tentacles and rhinophores are never present, even rudimentarily, before metamorphosis. Foot and glandular

components

Below the mouth the anterior portion of the foot of a premetamorphic, stage B larva has expanded into a propodium, due primarily to an increase in size of the propodial and metapodial glands within the foot. The ventral surface of the foot, which has a distinct groove down its midline, is heavily covered with short (5 to 6 pm), simple cilia and microvilli. Most of the lateral surfaces of the foot are unciliated (Fig. 3) although microvilli are present. The foot epidermis is seated on a dense basal lamina and is composed chiefly of columnar cells interspersed with at least two morphologically distinct types of nonciliated secretory cells (Fig. 4). Histo-

METAMORPHOSIS

Fig.

3. Frontal

section

through

OF AN AEOLID

NUDIBRANCH

the larval foot of Phestilla sibogae: ciliated lateral face of foot: x 300.

OP.,

233

operculum;

U.L.F.,

un-

Fig. 4. Epidermal morphology of the larval foot of Phesrilla sibogae: a-c, electron micrographs of epidermal cells and epidermal mucous glands: B.L., basal lamina; I.V., vacuoles with dumbbellshaped inclusions; M.G.C., mucous gland cells: note the numerous cilia and microvilli present on the surface of the foot: a x 1670; b x 5830; c x 4516.

DALE B. BONAR AND MICHAEL G. HADFIELD

234

chemical

tests for carbohydrates

using periodic

acid-Schiff’s

and alcian

blue tech-

niques were only weakly positive, suggesting that the secretions, presumably mucoidal in nature, contain very little neutral mucopolysaccharide or nonsulfated acid mucopolysaccharide. which contain

The ciliated columnar cells have basal nuclei and large apical vacuoles faintly visible, dumbbell-shaped structures (Fig. 4). These vacuoles

have only been reported from the epidermis of aeolid nudibranchs (Hyman, 1967) and have been studied at the ultrastructural level by Schmekel & Wechsler (1967), who suggest they have an absorptive, filtrative, and mechanical function. In P. .vibogae larvae, the vacuoles are usually elongate in an apical-basal direction, with their inclusions roughly perpendicular to the cell surface. The inclusions do not stain noticeably with P.A.S., alcian blue, mercury-bromphenol blue, or acrolein-Schiff’s, indicating they are neither protein nor neutral mucopolysaccharide. It should be noted that these diffuse inclusions are barely visible with the light microscope, so that they may have reacted to these tests but not stained intensely enough to be seen. The operculum is tightly attached to a layer of squamous epidermal cells on the metapodium. The structure of the cells and the nature of their attachment to the operculum will be described in a future paper (Bonar, in prep.). Beneath the epidermis, the interior of the foot is largely occupied by well-developed propodial and metapodial mucous glands. In other species (e.g., Addaria proxima and Tritonia hombergi; Thompson, 1958, 1962) these glands are distinguished by their respective positions in the propodial and metapodial (actually mesopodial) portions of the foot, but in Pllestillu sibogae they are in close proximity, especially in the propodium (Fig. 5a). For the sake of consistency, the names propodial and metapodial have been retained. The metapodial glands fill most of the mass of the foot and consist of enlarged, flask-shaped cells with narrow necks which penetrate the epidermis and expel the cell contents on the ventral foot surface, primarily in the midventral groove (Fig. 5b). These cells have basal nuclei and extensive rough endoplasmic reticulum, the latter apparently involved in manufacturing the abundant, darklystaining, membrane-bound secretion granules stored apically in each cell. Electron micrographs show that metapodial-gland secretion granules are released intact; these probably break down on contact with the external environment. Secretions of the metapodial glands are morphologically similar to those described in mucous the nemertine, Lineus ruber (Starch glands of other organisms, for instance, & Welsch, 1972). These authors speculate that the presence of extensive rough endoplasmic reticulum in mucous gland cells indicates a high protein content of the mucus produced by those cells. The propodial glands, much smaller and occupying a more anterodorsal position in the foot, are morphologically distinct from the metapodial glands. Cells of the propodial glands contain more Golgi apparatus, and the membrane-bound secretions are more electron lucent (Fig. 5b) than in metapodial gland cells. The propodial glands usually section poorly, as if fixation were inadequate. Peterson & Leblond (1964) suggest that Golgi origin of mucus reflects a high proportion of carbohydrate

METAMORPHOSIS

in the mucus:

this is also suggested

polysaccharides (Pearse,

1968:

are not always Hayat,

1970).

OF AN AEOLlD

NUDIBRANCH

here by the evidently

poor fixation,

235

since muco-

fixed well by gluteraldehyde and osmium The propodial glands stain more intensely

tetroxide with the

P.A.S. test for carbohydrate than do the metapodial glands; however, since the propodial glands overstain with any stain, the results of the P.A.S. test cannot be accepted as qualitative

evidence for higher carbohydrate

levels in these glands.

5a 6

Fig. 5. Subepidermal pedal glands in Phestilla sibogae larvae: a, diagram of the relative propodial and metapodial glands; b, electron micrograph of propodial and metapodial M.M.G., metapodial mucous gland; P.M.G., propodial mucous gland: bx 1370. Fig.

6. Light

micrograph

of a sagittal

section

through

mantle fold; P-V EP, perivisceral

positions of gland cells:

a stage B larva of PhestiNa sibogm: epidermis: x 160.

MF,

DALE

236

Also present

B. BONAR

AND

MICHAEL

in the foot are the pedal ganglia

muscle fibers from both the left larval retractor

G. HADFIELD

with associated

nerve tracts,

muscle and the cephalopedal

and

muscle

complex. Mantle cavity andperivisceral

Within

epidermis

the shell, the viscera are covered

with a very thin layer of squamous

epi-

thelial cells, averaging 0.5 pm in thickness and termed the perivisceral membrane (Thompson, 1958). Except for the floor of the mantle cavity, this epidermis is part of the mantle epithelium, or shell gland, a flap of which continues above the mantle cavity as the mantle fold (Fig. 6). The mantle fold, no longer active in shell secretion at the time of hatching, is not in intimate contact with the larval shell, but extends only part way toward the shell aperture. Like the rest of the mantle epithelium, this fold is composed of very thin squamous cells, although its leading edge is slightly thickened. With increasing age the mantle fold retracts further and further from the shell aperture. The mantle cavity is situated above the dorsal surface of the larval body and extends laterally and ventrally on the right side where the mantle fold is continuous with the perivisceral epidermis low on that side. The larval body is closely applied to the shell on the left side (Fig. 7) but the freedom of body from shell on the right side becomes increasingly evident as the larvae ages and the volume of the visceral mass decreases.

R.R.M

Fig. 7. Diagrammatic representation L.R.M., left larval retractor muscle:

of a stage B larva of Phestillasibogae from a dorsal view: N., nephrocyst; R.R.M., right larval retractor muscle; S., shell: \’ 200.

A slight mid-dorsal ridge of cilia is present on the floor of the mantle cavity. This ciliated ridge, which extends only part way toward the rear of the mantle cavity, functions in circulating water through the mantle cavity. On the ventral floor of the mantle cavity on the right side of the body is the anal opening. A narrow band of cilia, originating just behind the anus, arches over the anus, emerges from the mantle

METAMORPHOSIS

cavity,

OF AN AEOLID

NUDIBRANCH

237

and runs down the right side of the foot. The beat of these cilia effectively

sweeps fecal material out of the mantle cavity and away from the body. The floor of the palhal cavity is composed of cuboidal cells which are morphologically distinct from those of the foot (Fig. 8) in that they do not contain vacuoles with dumbbell-shaped inclusions described above,

the peculiar

Fig. 8. Electron micrograph of epidermal cells on the floor of the mantle cavity in a stage B larva 01 P. ~j~~~~e: EP.C., epidermaf cell; M.F., mantle fold: y 5670.

Visceral mass The internal anatomy of Phestilla sibogae veligers is similar to that of other lecithotropic nudibranch veligers, as described by Thompson (1958, 1962), Rao (1961), and Tardy (1970) and will only be treated here superficially except where significant differences occur. The visceral mass of P. sibogae contains abundant yolk reserves so that larvae can be maintained for as much as two weeks without food while retaining the ability to metamorphose. In newly hatched larvae, the individual visceral organs are difficult to distinguish and must be traced by sectioning, but in older larvae many of the visceral elements may be readily observed. Connected with the stomach are the right and left digestive diverticula which contain most of the larval yolk reserves. The left

DALE

238

diverticulum

B. BONAR

AND

MICHAEL

G. HADFIELD

is much larger than the right, so that the stomach

of a newly hatched

larva is displaced to the right. As the larvae age and yolk is utilized, diverticula decrease in size and the stomach comes to occupy a medial

the digestive position in a

stage B larva. The intestine exits from the posterior, dorsal wall of the stomach, passes down the right side of that organ, loops beneath it to the left, then back to the right,

where it opens

into the mantle

cavity just

below and behind

the larval

kidney. Excretory

organs

Three, apparently independent, organ systems contribute to the larval excretory apparatus: 1) the larval kidney; 2) the paired nephrocysts; and 3) epidermal deposition granules. The unpaired larval, or secondary, kidney almost entirely surrounds the anus and posterior intestine on the right side of the body. There are at least three morphologically distinct cell types in this organ, with more than 15 cells involved in the total structure (Figs 9a, b). Two of the cell types contain large vacuoles, and the third cell type, represented by only a single cell, appears to be a mucous gland. This glandular-appearing cell is the most dorsal of the three types, takes stain very heavily. and sections poorly, as if insufficiently fixed. Mitochondria, Golgi apparatus, and endoplasmic reticulum are present basally, further suggesting its glandular nature. It is possible that it produces mucus which aids cilia in eliminating feces from the mantle cavity. If this hypothesis is correct, this cell should not be considered part of the larval kidney. The two vacuolated cell types of the larval kidney are distinguishable primarily by the nature of their vacuolar inclusions. Those cells immediately adjacent to the posterior intestine and anus contain several vacuoles in each cell; these vacuoles are seen in sections to be empty, to contain a fine homogeneous precipitate, or to contain crystalline structures (type I, Fig. 9b). Free ribosomes are present in these cells. but abundant endoplasmic reticulum and Golgi apparatus are lacking. No discrete ducts leading from these cells to the exterior were seen: however, most of the cells border on the mantle cavity and vacuoles are often separated from the outside only by the cell membrane. Occasionally, sections were seen in which these vacuoles appeared to be bursting to the exterior, although this may be a fixation artifact. There are at least four type-2 cells in the larval kidney, lying in the most ventral portion of that organ just beneath the posterior intestine (type 2, Fig. 9b). The basal cytoplasm is similar to that seen in type-l cells, but the very large apical vacuoles are Under high magnification these filled with highly convoluted myelin structures. myelin figures do not show the trilaminar structure of a unit membrane, but appear as stacks of diffuse lamellae. Fig. 9b shows what appears to be a ‘maturation’ of these vacuoles, from the medial to lateral cells, involving progressive filling of the vacuoles with the myelin material. In the medial cells, this material is seen in smaller vacuoles which appear in the more lateral cells to be merging with the larger vacuoles. Type-2

METAMORPHOSIS cells

do not border

on the mantle

OF AN AEOLID

NUDIBRANCH

cavity and no ducts are seen that connect

239

them

with the exterior. How these cells release their storage products is not known. In life, the larval kidney is colorless and could not be observed to release the contents of any of its component

cells.

Fig. 9. Sections through the larval kidney of a stage B Phestillu sibogae larva: a, light micrograph of a cross section through a larva just anterior to the anus; b, composite electron micrograph of the larval kidney immediately adjacent to the anus; c, enlargement of area outlined in b: A., anus; U.M.G., unicellular mucous gland; 1, type-l vacuolated cells; 2, type-2 vacuotated cells. a x250: b r 1525, c y 10,800.

DALE

240

The nephrocysts

B. BONAR

AND

are a pair of unicellular

MICHAEL

G. HADFIELD

structures

z 20 pm in diameter,

located

beneath the epidermis, just behind the velum and above the statocysts on either side of the esophagus (Fig. 7). The left member of the pair is located somewhat more

Fig. 10. Electron micrograph of a nephrocyst in a stage B larva of Phestifla sibogue: a, whole nephrocyst; b, higher magnification of a section through the tubular area: EP.C., epidermal cell; L.R.M., left larval retractor muscle; LUM., lumen; R.R.M., right larval retractor muscle; T.S., tubular system; VAC., vacuole: arrows show the cilia seen within the nephrocyst: a Y 3140: b Y 6585.

METAMORPHOSIS

OF AN AEOLID

NUDIBRANCH

241

posteriorly than the right. Each nephrocyst has many irregular vacuoles with sparse particulate contents (Fig. 10). The nucleus is located apically, and vacuoles, yolk platelets, numerous mitochondria, and a highly convoluted tubular (condensing?) system are basal. An apical lumen encloses a few cilia, but no connections between the lumen and the mantle cavity or the hemocoel could be detected. One vacuole in each nephrocyst, considerably larger than the rest, is always seen bordering on the mantle cavity, separated from it only by the membrane of the nephrocyst. As with the larval kidney, sections sometimes show this vacuole bursting to the outside: again, this may be a fixation artifact. Nephrocysts are closely apposed to the larval retractor muscles, and electron micrographs suggest an intimate association between muscles and nephrocysts (Fig. 10). During early larval life, abundant yolk is stored in the muscle cells, especially adjacent to the nephrocysts; as larvae age this yolk diminishes.

Fig.

I I. Electron

micrograph of crystalline deposits Phestillu sibogae: E.D.G., epidermal

in the pedal epidermis of a stage deposition granules: r 6250.

With increasing age, white pigmentation becomes evident head and propodium of the larva. Electron micrographs show form of intracellular crystalline deposits (Fig. 1 I). Although primarily in the cephalic region of the larva, patches of white in the epidermis of much of the adult body. It is quite likely

B larva

of

in the epidermis of the the pigment to be in the this deposition is seen pigment are later visible these patches are nitro-

DALE B. BONAR AND MICHAEL

242

genous

waste products

of amino-acid

metabolism,

G. HADFIELD

possibly

pterines

or purines

(see

Vuillaume, 1969, for the biochemical derivation of these products), which are reported to produce the white pigmentation seen in other nudibranchs (Fox & Vevers, 1960; Burgin,

1965).

Larval musculature

The larval musculature is concentrated primarily in two bands: I) a left larval retractor with its origin on the extreme left posterior side of the larval body inserts on the operculum and pedal tissues; and 2) a larger right retractor muscle, which originates posteriorly, just to the left of rhe dorsal midline, and inserts in the tissues of the head and velum (Figs 12A, B, C). Contraction of these muscles retracts the head, velum, and foot into the shell and pulls the operculum tightly into the shell aperture. Within the foot there is also an opercular retractor which may be only a subdivision of the left retractor. The opercular muscle inserts on the right side of the operculum and onto the ventral lip of the shell aperture (Fig. 12C); its action is to aid in closure of the operculum. Longitudinal sections of these muscles reveal a striated appearance with a I .25 btrn periodicity. These striations are not the Z-bands seen in vertebrate

L.R.M.

Fig. 12. Diagrammatic representation of the larval musculature of PhestiNa siboyue: A, dorsal view; B, right lateral view; C, oblique frontal view; L.R.M., left larval retractor muscle; 0P.R.M.. opercular retractor muscle; R.R.M., right larval retractor muscle.

METAMORPHOSIS

skeletal muscle,

but are overlapping

OF AN AEOLID NUDIBRANCH

alternations

of filaments,

presumably

243

actin and

myosin, as shown in bivalve ‘quick muscle’ by Hanson & Lowy (1961). Many subepidermal muscle fibers are present throughout the anterior parts of the larval body. and correspond in part to the cephaIopeda1 muscle complex described by Thompson (1958). The position and function of these scattered fully in a future paper (Bonar, in prep.).

fibers will be discussed

more

Attachnlent between shell, oper~ulum, and larval retractor muscles is mediated by specialized cells of the perivisceral epidermis. The retractor muscles insert just below the junction of mantle fold and perivisceral epidermis and, for at least part of their route to the head and foot, pass just beneath the floor of the mantle cavity. The left retractor passes immediately lateral to, and slightly below, the Ieft statocyst and arches smoothly into the foot. The right larval retractor muscle passes directly forward beneath the floor of the mantle cavity and inserts on the basal lamina of head and velar epidermis. This muscle is responsible for withdrawing the head and velum into the shell cavity ahead of the foot and operculum, but is not responsible for closure of the shell aperture by the operculum; the latter function is assumed by left larval retractor and opercular muscles. No apparent morphological differences were seen between cells of the different muscles, such as is seen in the quick and catch muscles of bivalves (Hanson & Lowy, 1960). Nrrvous system The nervous

system was not analyzed

in detail, but is very much like that shown for

Aeolidiella alderi by Tardy (1970). Comparison of larval stages A and B shows the ganglia of the latter to be larger and more condensed. Peripheral nerve fibers are small and di@icult to find with the light microscope, but electron microscopic examination shows nerve fibers to be more abundant in older larvae. Attempts at selective staining of the nervous system of fixed specimens with methylene blue or gold chloride (Conn, Darrow & Emmel, 1960) were unsuccessful. Heart No larval

heart

was visible

at any stage, nor was there indication

of pulsatory

activity on the floor of the mantle cavity. At metamorphosis, there are no signs of definitive heart primordia, and as Tardy (1970) notes for A. alderi, it is only after metamorphosis and during elongation of the pseudovermis stage that the adult heart becomes noticeable, and even then it is at first poorly defined. METAMORPHOSIS

The first indications of the onset of metamorphosis in P/?estilia sibogae are behavioral. Larvae which are competent to metamorphose will settle and crawl actively on the substratum, apparently testing for the suitable environmental conditions which promote metamorphosis. Often larvae will swim off, settle elsewhere, and repeat the crawling behavior. When Iarvae settle in conditions conducive to metamorphosis,

DALE B. BONAR AND MICHAEL

244

which in nature

will presumably

morphological process sense the environmental

G. HADFIELD

be on or near Porites compressa

of metamorphosis inducing factor

coral heads,

the

begins. The mechanism by which larvae is unknown, and so no accurate estimate

can be made of the time between ‘sensing’ the inducer and the beginning of any visible morphological changes. The shortest time observed between introduction of P. compressa

to the larval culture

and onset of visible metamorphic

change

was 2 h.

It was, however, common to see metamorphosis take place only after larvae had been in the presence of P. compressa for I2 h or more, and sometimes there was a delay of 24 to 48 h before metamorphosis began. As the larvae crawl about on the substratum, their velar lobes are partially extended and the preoral cilia are constantly beating. The areas around the mouth are in fairly close apposition to the substratum, giving the appearance of ‘testing’ the ground over which the larva travels. If interaction of larvae with metamorphic inducer is by way of contact chemoreception (Carthy, 1958) or a ‘chemical-tactile’ perception as suggested for larvae of barnacles (Crisp, 1965) and annelids (Wilson, 1968) then it may be that larvae can perceive the inducer only when that inducer is adsorbed onto the culture vessel. Whatever the nature of metamorphic induction, the first indication of irreversible metamorphosis is loss of the velar lobes. The settled larvae cease crawling and remain in one place with the velum partly extended from the shell aperture. Cilia of the velar lobes begin to beat with increasing asynchrony and ciliated cells lose their contact with adjacent cells, round up, and are ingested by the larva; two crumpled piles of cells mark the sites of the lost lobes. This process is brief, occupying IO to 20 min, and is followed by increased activity of the animal. Within another hour, the body will have lost intimate contact with the shell. Insertion points of the left and right larval retractor muscles are the primary sites of adhesion between larval body and shell. Most of the perivisceral epidermis lies closely against the shell, but no obvious structural adhesion exists except at muscle insertion sites. After loss of the velum the body pulls increasingly away from the shell until the muscle connections are the last visible sites of attachment. When these connections are lost the visceral mass rounds up in the anterior part of the shell and there is no longer any physical connection between body and shell (Fig. 13). In this stage, larvae may continue creeping about on the foot, swinging the shell from side to side as if attempting to shake it off but will usually cease crawling to complete the exit from the shell. Since the diameter of the visceral mass is larger than that of the shell aperture, the larva must squeeze out of the shell. Numerous observations support the hypothesis that muscular action, against a foot firmly attached to a substratum, is utilized in levering the body from the shell. Partly metamorphosed larvae which were held in such a way that the foot had no firm anchorage could not shed the shell. The visceral mass is not firmly held by the overlying epidermis, and can be moved about within the hemocoel. During the exit from the shell, the visceral organs are pulled toward the ventral surface of the foot and merge with the cephalopedal mass.

METAMORPHOSIS

ANT.

OF AN AEOLID

NUDIBRANCH

245

14

13

ANT.

POST, Fig.

13. Light micrograph

of a stage D larva of Phestilla sibogae: note larval body and shell: S., shell: x 197.

Fig. 14. Light micrograph of a stage F postlarva remnants of larval epidermis: ANT., anterior;

lack of attachment

between

of Phesjilla sibogae from a dorsal view showing POST., posterior; REM., remnants: ): 171.

Movement of the visceral mass during shell loss is closely related to a rapid spreading of epipodial epidermis to cover the dorsal and posterior aspects of the body. This complex process, reviewed here briefly, and the role muscles play in its execution, will be fully discussed in a forthcoming paper (Bonar, in prep.). Shortly after the shell has been cast, the postlarva has a hump-backed appearance (Fig. IE). The body is now almost

completely

covered

with definitive

epidermis,

although this epidermal layer is quite thin on the dorsal and posterolateral surfaces. The remnants of the mantle fold and perivisceral epidermis are visible posteriorly as a crumpled pile of cells on the left posterior aspect of the postlarval body (Fig. 14). Within the next I8 to 24 h these cells are lost, though whether by rejection or resorption is not known. Immediately after epidermal migration, the layer of epidermal cells on the dorsal and lateral sides of the body is quite thin, only 4 to 5 pm, compared with the 9 to I4 pm thickness of the ventral foot epidermis. This thin layer of definitive epidermis will increase in thickness over the following I8 to 24 h until it has approximately the same thickness as the foot epidermis. Cell division may be partly responsible for this increase, but the large epidermal vacuoles which were quite flattened at completion of migration have become spherical to oval, as if having

246

filled with liquid.

DALE

B. BONAR

AND

MICHAEL

Swelling of these vacuoles

G. HADFIELD

due to absorption

of water may account

for the noted increase in cell volume. Scattered patches of cilia are visible on the epidermal surface, but only the foot is uniformly ciliated. As would be expected, the ventral surface of the foot is densely and evenly covered with cilia. Re-arrangement of internal anatomy brought about at metamorphosis

is similar

to that described by Tardy (1970) for AeoEidielh alderi. Much of the apparent change is due to a shift in the anteroposterior axis, coupled with a dorsoventral flattening of the body (Fig. 15). This axial shift is made possible shell, or more specifically, by loss of the constricting

in nudibranchs by loss of the influence of the shell aperture

Fig. 15. Diagrammatic representation of axial shifts in PhesriNu sibogae flattening of the dorsoventral axis and rotation of the anterior-posterior stage G postlarva.

at metamorphosis showing axis: A, stage B larva; B,

on the viscera1 mass. At metamorphosis, the original line of contact between shell and operculum is freed to move to a posterior position in the ‘hump-backed’ stage, and without the constricting shell, the visceral organs move to assume a more ventral position in the postlarva. The anus, originally at the anterior, ventral limit of the mantle cavity near the point of shell-opercular articulation, moves to a posterior position on the right lateral flank of the metamorphosed animal, carrying along with it some of the large vacuolated cells of the larval kidney and the rudiment of the adult kidney. As the viscera1 mass moves to a more ventral position, the left digestive diverticulum becomes more dorsally situated over the stomach; this is accompanied by a slight rotation of the stomach to the right. Continuing development and growth of the left digestive diverticulum, as seen later in the pseudovermis stage, will bring it to occupy a position posterior to, and filling most of the body cavity behind, the stomach. The much smaller right digestive diverticulum is located posterolaterally to the stomach on the right side. The openings of intestine and digestive diverticula into the stomach are all in very close proximity at this stage. The diverticular openings are side by side on the ventral portion of the posterior stomach wall. The intestine opens at the junction of the stomach and right digestive diverticulum, and appears to be more closely associated with the diverticulum than the stomach. As it passes posteriorly, the intestine loops to the left, then to the right, all entirely beneath the

METAMORPHOSIS

left digestive

diverticulum,

OF AN AEOLID

before opening

NUDIBRANCH

on the posterior

247

right side of the animal

(Fig. 16). In the stage G postlarva, the adult kidney rudiment is visible just anterior and dorsal to the anus, and the large type-l vacuolated cells of the larval kidney are still evident lying alongside the posterior intestine and adjoining the adult kidnev rudiment (Fig. 16). Type-2 vacuolated cells are not seen after metamorphosis, and are probably lost during, or soon after, casting of the shell. No traces of the nephrocysts are seen after metamorphosis; it is believed that they are lysed along with the larval retractor muscles with which they are in close association. Dissolution of muscular tissue is evident quite early in metamorphosis. Within minutes of velar loss, lytic bodies containing muscle fibers are evident in the tissues of the head (Fig. 17); these are remnants of fibers of the right retractor muscle which

A.

L.D.D.

0. R.D.D. Fig. 16. Diagrammatic representation of visceral reorganization in PhesfiNa sibogae at metamorphosis: A, stage B larva; B, stage G postlarva: A, anus; B.A., buccal apparatus; E., esophagus; K.RUD., adult kidney rudiment; LA.K., larval kidney; L.D.D., left digestive diverticulum; O.R.D.D., opening to the right digestive diverticulum (removed in drawing B); R.D.D., right digestive diverticulum: STOM., stomach.

248

DALE

B. BONAR

AND

MICHAEL

G. HADFIELD

Fig. 17. Electron micrograph showing autolysis of muscle fibers in the cephalic area larva of Phestilla sibogae within minutes of velar loss: B.A., buccal apparatus; 1Y.B.. ” 5560.

Fig. 18. Electron micrograph of a section through a larval retractor of Phestillu sibogae. Note the loss of fibrillar organization, suggesting retractor muscle: * 7260.

of a stage C lytic bodies:

muscle in a stage F postlarva autolysis: L.R.M., left larval

METAMORPHOSIS

inserted

OF AN AEOLTD NUDIBRANCH

on the velar tissues. As metamorphosis

larval retractor right retractors

muscles,

shell, and operculum

lose their internal

progresses

249

and attachments

between

are lost, the main bulk of the left and

organization

as autolysis

begins

(Fig. 18). The

many small circular, longitudinal, and oblique subepidermal fibers are not lysed at metamorphosis. These muscles, most noticeable in the cephalopedal region, will contribute to the adult musculature. Pedal glands are visible in the stage C postlarva. These are, however, unicellular glands which do not resemble the original metapodial glands; the latter have disappeared within two days following metamorphosis. The small propodial glands are still visible anteriorly. DISCUSSION

The function of metamorphosis in all nudibranchs is the conversion of a shelled. pelagic larva into a shell-less, benthic juvenile. It is thus not surprising that the order of metamorphic events described for other nudibranchs is similar to that of Pl~estillu sibogae, so much so that the staging method used in this study is widely applicable. Where differences occur, they may generally be attributed to variations in developmental patterns in different species. The developmental patterns of nudibranchs, and discussions of their differences, have been given by Thompson (I 967a) and modified by Tardy (1970). Tardy has essentially reorganized the three developmental types proposed by Thompson into two types with six major subgroupings; this reorganization divides developmental types less artificially than does Thompson’s classification, and emphasizes the possible phylogenetic progression in the Nudibranchia. Tardy’s two major divisions are based primarily on larval shell type, which he feels is representative of basic ontogenetic differences during veliger development and metamorphosis. Species with Type-One development all possess Type-One protoconchs (Thompson, I96 I ) and derive the definitive adult epidermis of the dorsal surface from thickening and spreading of the mantle. Thompson has described the thickening and proliferation of mantle fold cells during larval life in Ad&aria proxinza (Thompson, 1958) and Tritonia hombergi

(Thompson,

1962). This fold reflects

back over the body just

prior to, and during. metamorphosis, and forms the dorsal integument. Tardy (1970). describing Aeolidiella alderi, has shown the definitive epidermis to come from the “bourrelet pallCal”, a thickening of cells of the mantle which later everts and spreads over the dorsal and lateral surface of the body at metamorphosis. It is not clear whether he considers the “bourrelet palltal” to be completely homologous to the thickened mantle fold described by Thompson, but they appear to be identical. Virtually all reported descriptions of metamorphosis, including the studies just mentioned, deal with Tardy’s Type-One development. Little is known of the metamorphosis of species corresponding to Tardy’s developmental Type-Two, since most of these species have long planktonic lives. Tardy (1970) briefly describes metamorphosis of five species in this category, and notes

DALE B. BONAR AND MICHAEL G. HADFIELD

250

that the “bourrelet

palleal”

forms

on the floor of the mantle

cavity

and spreads

directly over the sides of the animal, rather than everting as in Type-One development. His figures (No. 15a, b, page 340) show what he considers to be the developing layer of definitive epidermis on the floor on the mantle cavity of Tenelliu uentilIuhrunz, a species with Type-Two thinner

than

development;

that seen in the thickened

this presumptive mantle

epidermis

is much

of species with Type-One

adult

develop-

ment. Phestilla sibogae is representative of Type-Two development, but varies somewhat from that briefly described by Tardy. The most significant difference is in the origin of the dorsal epidermis, which in P. sibogae originates entirely from the lateral surfaces of the larval foot. Within an hour of exit from the shell, definitive epidermis, characterized by the presence of vacuoles containing dumbbell-shaped inclusions, is present over almost the entire surface of the body. When metamorphosis begins, vacuolated epidermis is found only on the larval foot. The epidermis on the floor of the mantle cavity does not contain these unique vacuoles and is too thin (2.5 to 3.0 itm) to spread and form the epidermis which is evident around the body shortly after shell loss. It is the fortunate occurrence of these unique vacuoles in aeolid nudibranch epidermis and the opportunity to view them clearly with the electron microscope which has revealed the true origin of the adult epidermis. Whether this origin is consistent for all species evidencing Type-Two development, or whether some species derive the epidermis from the mantle cavity floor, as Tardy suggests, must await further examination of other aeolid species. The use of the terms left and right for the two larval retractor muscles refers to their position in P. sibogue larvae. Since the embryology of this animal was not followed, it is not known whether the relationship between left and right muscle bundles in this species is the same as that reported for prosobranch gastropods (see Fretter & Graham, 1962, for review). To compound this uncertainty, the larval musculature as described for various opisthobranch species is not consistent in position, insertion or in function. There is only a single larval retractor muscle reported for Aduluriu proximu, Tritonia hombergi, (Thompson, 1958. 1962) and Aeolidiellu ulderi (Tardy, 1970), while Saunders & Pool2 (19 IO) figure two retractors for AplJlsia punctuta, and Casteel (1904) shows three for Fiona marina. In all of these species except Fionu the main larval retractor, located dorsally and slightly to the left of the larval midline, branches into the velum, head, and foot. The equivalent muscle in fhrstilla sibogue, which we have called the right retractor, inserts only in the velar and head area, whereas only the left retractor inserts on the foot. Casteel (1904) figures a large central retractor in Fionu marina which inserts only on the cephalic tissues, while two much smaller lateral bundles serve the foot. It is evident that opisthobranch larval musculature is quite variable, and the presence or absence of muscle bundles may explain, at least in part, why torsion is present as a mechanical process in some species (Tardy, 1970) and not in others (Thompson, 1958, 1962, 1967a, b). It has been assumed since the late 1800’s that the structure termed a larval kidney

METAMORPHOSIS

in opisthobranch

larvae

is involved

OF AN AEOLID NUDIBRANCH

in excretion,

although

there

251

is only

diffuse

evidence to suggest this is so (see Casteel, 1904, and Saunders & Poole, 1910, for reviews). Saunders & Poole discounted earlier suggestions that this structure, darkly pigmented

in many species, is an ‘anal eye’, and stated that, “There can be no doubt

it is an excretory

organ,

since it is easy in living larvae of some opisthobranchs

to

observe the process of excretion” (p. 513). Although secretory activity could not be detected in living Phestilfu sibogue larvae, the morphology of cells of the larval kidney suggests that this organ is involved in accumulation of metabolic wastes, if not in excretion from the larval body. The nature of the vacuolar inclusions of the type-2 cells is suggestive of myelin figures produced in the formation of artificial phospholipid membranes (Miyamoto & Stoeckenius, 1971) and concentric lamellar lipid accumulations seen in cells of persons with Tay-Sachs disease (Brady, 1973). Possibly these lamellar structures are remnants of lipid-rich yolk reserves which have been utilized during development. The other type of vacuolar cells may correspond to a urinary bladder as suggested by Tardy (1970) for similar structures in Aeolidiellu alderi. The crystalline deposits sometimes seen in these vacuoles are of unknown origin, though Fawcett (1966) notes that intracellular crystalline deposits are usually assumed to be proteinaceous. Nephrocysts, by their very name, have been ascribed an excretory function, yet very little is known about these cells. Previous studies of other opisthobranchs with light microscopy showed nephrocysts to be paired, unicellular, filled with vacuoles (sometimes refractive), and not associated with the larval kidney (see Saunders & Poole, 1910, for review). Although Tardy (1970) reports seeing a duct connecting one of these cells of A. alderi to the outside, suggesting a protonephridial nature, he was unable to repeat this observation on other larvae. No ducts were seen from the nephrocysts of Phestifla sibogae, although some sections show a small lumen enclosing a few cilia. Kume & Dan (1968) note that paired nephrocysts are seen in the larvae of many marine gastropods which do not have protonephridia. According to Portmann (1930) who studied these structures in larvae of the prosobranch gastropod Buccinum undatum, nephrocysts have no excretory tubules and function by receiving waste materials from wandering cells (“Wanderzellen”) in the body cavity. Because at least one large vacuole in each nephrocyst of Phestilla sibogae is separated from the mantle cavity only by the cell membrane, it is conceivable that vacuoles may rupture to release their contents. The close association between nephrocysts and larval retractor muscles suggests an interrelationship of some sort. Reduction in muscle yolk reserves during larval life is especially evident in this area; perhaps the nephrocysts enhance the larval capacity for metabolic waste elimination from muscle cells. Maturation of the pedal glands is apparently a necessary requisite for metamorphosis and muscular action for successful exit from the larval shell; in order for muscles to function there must be a structural anchorage on which they insert. Once the larval body loses its attachment to the shell, the only anchorage available for the

252

DALE B. BONAR AND MICHAEL G. HADFIELD

muscles is on the foot, and then only if the foot is firmIy anchored

to the substratum.

The propodial and metapodial mucous glands are not developed in newly hatched larvae, and stage A larvae are rarely observed to crawl actively about on the bottom of the culture vessels. Disappearance phosis of P. sibogae further indicates

of the metapodial glands following mctamortheir transitory importance. Thompson (1962)

notes that the pedal glands of Tritonia hombergi also disappear after metamorphosis, but of even more interest is his observation (Thompson, 1967a) that metapodial glands do not develop in nudibranchs with direct development. Certainly, the development of pedal mucous glands is not the only prerequisite to metamorphosis in planktonic veligers, but the presence of these glands is apparently necessary. According to Tardy (1970) detorsion does not occur in nudibranchs. Based on his own studies and others on several species of nudibranchs, he claims that ail nudibranchs undergo a 180” torsion during development, and that the slight reversal at metamorphosis of the originaI effects of torsion is not true detorsion, but a superficial reversal due to elongation of the postlarval body. His studies show correctly that metamorphosis of nudibranchs results in the merger of visceral and cephalopedal elements, and he suggests that a 180” detorsion would involve not only the viscera, but would cause a helical twist of the cephalopedal area as well. Extrapolating from adult aeolid anatomy, Tardy suggests that the location of the anus on the right lateral side of the body and the exit of the intestine from the dorsal side of the stomach shows that there was no true detorsion. This view deviates SubstantiaIly from Thompson’s (1958, 1962) thesis that torsion in dorid nudibranchs is a result of early, asymmetrical development, that this torsion does not attain the full 180” twist seen in diotocardian prosobranchs, and that the reversal of torsion, detorsion, occurs at metamorphosis. Our work on Phestilla sibogae indicates that detorsion does occur in this species. Sections through elongating postlarvae (stage G) show that the intestine opens from the ventral, posterior wall of the stomach at the junction of the stomach and right digestive diverticulum, and that the anus is located on the right posteriolateral surface of the body, well behind the visceral organs. It is possible that detorsion does take place at metamorphosis in the aeolids which Tardy (1970) studied, but later growth and development to the adult form results in a secondary shift of the position of intestinal opening from the stomach wall. The right lateral position of the aduIt anus can also be explained as a result of posteriorIy-directed growth of the body and viscera behind the anus. Detorsion of P. sibogae is a two part process, the first as a result of casting the shell at metamorphosis, and the second due to growth and elongation of the postlarval body. Migration of the anus from the anterior limit of the larval mantle cavity to a posterolateral position in an early postlarva clearly indicates detorsion. Unfortunately, the expression of torsion is not seen in the nudibranch nervous system, due to condensation of ganglia during early development, so that this system cannot be used as an indicator of torsion or detorsion as it is in many other gastropods.

METAMORPHOSIS

There is a remarkable gastropods

similarity

and the reversal

brought about ential growth

OF AN AEOLID NUDIBRANCH

between

of torsion

the process

of torsion

253

in diotocardian

in P. sibogae. In diotocardians,

torsion

is

by an initial rapid muscular contraction, followed by a slower differprocess (see Fretter & Graham, 1962, for review). In P. sibogac,

detorsion is produced in the same way; first, by contraction which results in casting of the shell and initial reorientation

of the retractor muscles, of visceral elements; and

secondly, by a slower growth process. A further similarity is the subsequent loss of muscles that produced these effects, both in P. sibogae and in Calliostoma (Crofts, 1955) and presumably in other prosobranchs that have a single shell muscle as an adult (Fretter & Graham, 1962). Thompson (1958) also noted that the process of detorsion in Adalariu proxima was remarkably similar to torsion in diotocardians; however, he implicated eversion of the mantle fold and progressive growth of that tissue as the motive force for detorsion. It would be interesting to see whether the larval musculature of A. proxima aids in reflection of the mantle fold. ACKNOWLEDGMENTS

This research

was supported

in part

by NSF

Grant

GB 36702 to Dr Hadfield.

REFERENCES BERGH, R., 1905. Die Opisthobranchiatu der Siboga-Expedition. Monographi 50, 248 S., E. J. Biell, Leiden. BONAR, D., 1972. Fate oflarval organs at metamorphosis in a gastropod. Am. Zool., Vol. 12, p, 722 only. BRADY, R. O., 1973. Hereditary fat metabolism diseases. Scienf. Am., Vol. 229, pp. 88-97. BVRGIN, U. F., 1965. The color pattern of Hermissenda crassicornis (Erscholtz, 1931). Veliger, Vol. 7, pp. 205-215. Carthy, J. D., 1958. An introduction to the behaviour of invertebrates. George Allen and Unwin, Ltd., London, 380 pp. CASTEEL, D. B., 1904. Cell lineage and early larval development of Fiona marina, a nudibranch mollusk. Proc. Acad. nat. Sci. Phillad., Vol. 56, pp. 325-405. CLONEY, R. A., 1963. The significance of the caudal epidermis in ascidian metamorphosis. Biol. Bull. mar. biol. Lab., Woods Hole, Vol. 124, pp. 241-253. CLONEY, R. A., 1966. Cytoplasmic filaments and cell movements: epidermal cells during ascidian metamorphosis. J. Ultrastruct. Res., Vol. 14, pp. 300-328. CLONEY, R. A., 1969. Cytoplasmic filaments and morphogenesis: the role of the notochord in ascidian metamorphosis. Z. Zellforsch. mikrosk. Anat., Bd 100, S. 31-53. CLONEY, R. A., 1972. Cytoplasmic filaments and morphogenesis: effects of cytochalasin B on contractile epidermal cells. Z. Zellforsch. mikrosk. Anat., Bd 132, S. 167-216. CLONEY, R. A. & L. GRIMM, 1970. Transcellular emigration of blood cells during ascidian metamorphosis. Z. Zellforsch. mikrosk. Anat., Bd 107, S. 157-173. CONN, H. J., M. A. DARROW & V. M. EMMEL, 1960. Staining procedures used by the Biological Stain Commission. William and Wilkins Co., Baltimore, 2nd edition, 289 pp. CRISP, D. J., 1965. Surface chemistry, a factor in the settlement of marine invertebrate larvae. Proc. Fifth Biol. Symp., Botanica gothoburgensia, Goteborg, Vol. 3, pp. 52-65. CROFTS, D. R., 1937. Development of Haliotis. Phil. Trans. R. Sot. Land.. B, Vol. 208, pp. 219-268. CROFTS, D. R., 1955. Muscle morphogenesis in primative gastropods and its relation to torsion. Proc. zool. Sot. Lond., Vol. 125, pp. 711-750. FAWCETT, D. W., 1966. The cell, an atlas offine structure. W. B. Saunders Co., Philadelphia, 448 pp.

254

DALE

B. BONAR

AND

MICHAEL

G. HADFIELD

Fox, H. M. & G. VEVERS, 1960. The nature of animal colours. Sedgwick and Jackson, Ltd., London, 246 pp. FRETTER, V., 1972. Metamorphic changes in the velar musculature, head, and shell of some prosobranch veligers. J. mar. biol. Ass. U.K., Vol. 52, pp. 161-178. FRETTER, V. & A. GRAHAM, 1962. British prosobranch molluscs. Ray Society, London, 755 pp. HADFIELD. M. G., 1972. Flexibility in larval life patterns. Am. Zool., Vol. 12, p. 721 only. HADFIELD, M. G. & R. H. KARLSON, 1969. Externally induced metamorphosis in a marine gastropod. Am. Zool., Vol. 9, p. 317 only. HANSON, J. & J. LOWY, 1960. Structure and function of contractile apparatus in the muscle of invertebrate animals. In, The structure and function of muscle. Vol. I, edited by G. H. Bourne, Academic Press, New York, pp. 265-335. HANSON, J. &J. LOWY, 1961. The structure of the muscle fibres in the translucent part of the adductor of the oyster Crassostrea angulata. Proc. R. Sot. Lond., B., Vol. 154, pp. 173-193. HARRIS, L. G., 1970. Studies on the biology of the aeolid nudibranch I’hestilfa melanobranchia Bergh 1874. Doctoral Dissertation, University of California, Berkeley, 318 pp. HAYAT, M. A., 1970. Principles and techniques of electron microscopy. Vol. I. Biological applications. Van Nostrand Reinhold Co., New York, 412 pp. HORIKOSHI, M., 1967. Reproduction, larval features and life history Philine denticulata. Ophelia, Vol. 4, pp. 42-84. HURST, A., 1967. The egg masses and veligers of thirty northeast Pacific Opisthobranchs. Veliger, Vol. 9, pp. 255-288. HYMAN, L. H., 1967. The inaertebrates. Vol. VI. Mollusca I. McGraw-Hill Co., New York, 792 pp. ISHIKAWA, M. & T. NUMAKUMAI, 1972. Changes in the activity of acid phosphatase and the appearance of metamorphosing potencies of the larvae of the ascidian Halocynthia roretzi. Develop. Growth Different., Vol. 14, pp. 75-83. ISHIKAWA, M., T. NUMAKUMAI, M. DE VINCENTIS & M. LANCIERI, 1972. Enzymatic activities relating to respiration and the onset of metamorphosis of the ascidian tadpole. Deuelop. Growth Dilyerent.. Vol. 14, pp. 229-236. KUME, M. & K. DAN, 1968. Inuertebrate embryology. NOLIT Pub. Co., Belgrade, 605 pp. LUFT, J. H., 1961. Improvements in epoxy resin embedding methods. J. biophys. biochem. Cytol., Vol. 9, pp. 409-414. MIYAMOTO, V. & W. STOECKENIUS, 1971. Preparation and characteristics of lipid vesicles. J. Membr. Biol., Vol. 4, pp. 252-269. PEARSE, A. Ci. E., 1968. Histochemistry, theoretical and applied. Vol. I. Williams and Wilkins Co., Baltimore, 759 pp. PETERSON, M. & C. P. LEBLOND, 1964. Synthesis of complex carbohydrates in the Golgi regions as shown by autoradiography after injection of labeled glucose. J. Cell Biol., Vol. 21, pp. 143-l 54. PORTMANN, A., 1930. Die Larvennieren von Buccinum undatum L. Z. Zellforsch. mikrosk. Anat., Bd 10, S. 401-410. RAO, K. V., 1961. Development and life history of a nudibranch gastropod Cuthona adyarensis Rao. J. mar. biol. Ass. India, Vol. 3, pp. 186-197. REYNOLDS, E. S., 1963. The use of lead citrate at high pH as an electron-opaque stain in electron microscopy. J. Cell Biol., Vol. 17, pp. 208-213. RITHARDSON, K. C., L. JARETT & F. H. FINKE, 1960. Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technol., Vol. 35, pp. 313-323. RUDDELL, C. L., 1967. Hydroxyethyl methacrylate combined with polyethylene glycol400 and water; an embedding medium for routine l-2 micron sectioning. Stain Technof., Vol. 42, pp. 191-123. RUSSELL, H. D., 1971. Index Nudibranchia 1554-1965. Delaware Mus. nat. Hist., Greenville, Delaware, 141 pp. SAUNDERS, M. & M. POOLE, 1910. The development of Aplysiapunctata. Q. Jl microsc. Sri.. Vol. 55, pp. 4977540. SCHMEKEL, L. & W. WECHSLER, 1967. Electronmikroskopische Untersuchungen tiber Struktur und Entwicklung der Epidermis von Trinchesia granosa (Gastropoda, Optisthobranchia). Z. Zellforsch. mikrosk. Anat., Bd 77, S. 95-114. SMITH, F. G. W., 1935. The development of Patella rulgata. Phil. Trans. R. SOL-. Lond., B, Vol. 225, pp. 95-125.

METAMORPHOSIS

OF AN AEOLID

NUDIBRANCH

255

STORCH, V. & U. WELSCH, 1972. The ultrastructure of epidermal mucous cells in marine invertebrates (Nemertini, Polychaeta, Prosobranchia, Opisthobranchia). Mar. Biol., Vol. 13, pp. 167-175. TARDY, J., 1970. Contribution a I’dtude des metamorphoses chez les nudibranches. Annls Sri. not. (Zoo/.), T. 12, pp. 299-370. THIRIOT-QUIBVREUX, C., 1967. Observations sur le developpement larvaire et postlarvaire de Simnitr spelta Linne. (Gasteropode Cypraeidae). Vie Milieu, T. 18, pp. 143-I 5 I. THIRIOT-QUI~VREUX, C.. 1969. Organogentse larvaire du genre At/unto (Mollusque, HUropode). Vie Milieu, A, T. 20, pp. 347-396. THIRIOT-QUI~VREUX, C., 1970. Transformations histologiques lors de la metamorphose chez C?smbulia peroni de Bainville (Mollusca, Opisthobranchia). Z. Morph. okol. Tiere, Bd 67, S. 106-l 17. THOMPSON, T. E., 1958. The natural history, embryology, larval biology, and postlarval development of Adofuriu proximu (Alder and Hancock) (Gastropoda, Opisthobranchia). Phil. Trans. R. Sot. Land., B, Vol. 242, pp. l-58. THOMPSON, T. E., 1961. The importance of the larval shell in the classification of the Sacoglossa and the Acoela (Gastropoda, Opisthobranchia). Proc. mulac. Sot. Land., Vol. 34, pp. 233-238. THOMPSON, T. E., 1962. Studies on the ontogeny of Tritonia hombergi Cuvier (Opisthobranchia). Phil. Trans. R. Sot. Land., B, Vol. 242, pp. 171-218. THOMPSON, T. E., 1967a. Direct development in a nudibranch Cadlinu lueuis, with a discussion on developmental processes in Opisthobranchia. .I. mar. biol. Ass. U.K., Vol. 47, pp. l-22. THOMPSON,T. E., 1967b. Adaptive significance ofgastropod torsion. Malacologiu, Vol. 5, pp. 423-430. VUILLAUME, M., 1969. Les pigments des invertebrks. Monogr. No. IO. In, Les grandsproblt?mes de lo biologic, ed. par P-P. Gras&, Masson et Cie, Paris. WILSON, D. P., 1968. The settlement behaviour of larvae of Sabelluria ulwolota. J. mu. biol. Ass. U.K., Vol. 48, pp. 387-435. WOOLLACOTT, R. M. & R. L. ZIMMER, 1971. Attachment and metamorphosis of the Cheilo-ctenostome bryozoan Bug& neritina (Linne). J. Morph., Vol. 134, pp. 351-382.