The Ecology and Digestive System of the Struthiolariidae (Gastropoda)

The family Struthiolariidae Fischer comprises a small group of proso-branch molluscs, with an extensive time range and geographical distribution in Tertiary seas, but including only four recent genera, all in the southern hemisphere, of which three have but one surviving species. Perisso- donta Martens 1878 is represented by one species at Kerguelen Land and another at South Georgia, and Tylospira Harris 1897 by a single species in New South Wales. The present paper deals with the representatives of the two New Zealand genera—Struthiolaria Lamarck 1818, as now restricted, and Pelicaria Finlay 1928. The two species Struthiolaria papulosa (Martyn 1784) and Pelicaria vermis (Martyn 1784) have long been known to conchology, being accurately figured in Thomas Martyn’s The New Conchologist from shells collected on Cook’s first voyage, probably by Sir Joseph Banks in Queen Charlotte Sound. Our anatomical knowledge of the Struthiolariidae is, however, extremely fragmentary. Quoy and Gaimard (1833) give fairly accurate figures of the external characters of the New Zealand species. Hutton (1882) contributed a note on the anatomy of Struthiolaria papulosa ‘, his description and drawings are rough and inaccurate, though the dentition and operculum were for the first time correctly figured. As regards the biology of the New Zealand forms, Powell (1937) has briefly concluded that Struthio-laria and Pelicaria are deposit feeders. A detailed examination of the structure and feeding mechanism of Struthiolaria papulosa has brought to light features of special biological interest. Pelicaria vermis has been compared with Struthiolaria throughout and found to agree in all essential characters.

Specimens of both Struthiolaria papulosa and Pelicaria vermis used in the present work were collected in New Zealand at Cheltenham and Takapuna beaches on the shore of Rangitoto Channel near Auckland. The mode of life was studied in the field, while material was dissected, with the help of a Zeiss binocular, both alive for the study of ciliary and digestive action, and after fixation with Bouin’s Fluid. This fixative was found excellent for general histological work, while Carnoy’s without chloroform was used for the quick fixation of ciliated tissues. Paraffin sections were cut at 8 microns and double-stained with Delafield’s haematoxylin and van Gieson’s picrofuchsin.

Struthiolaria papulosa (Text-fig. 2) is found widely throughout New Zealand and extends to the Kermadec Islands. It occurs characteristically on the lowermost littoral fringe of clean sand-mud-flats, of which Cheltenham Beach is a typical example, with a wide expanse of shore some three-quarters of a mile long and 300 yards between tides. Wave action is relatively subdued and an area of fine shell sand has become covered with a thin mantle of organic sediment, plant detritus, and benthic diatoms. Water movement is nevertheless sufficient to ensure a well-aerated substratum, the more stable conditions of mud-flat being undeveloped and Zostera present only in isolated tufts. There is a comparatively rich fauna, including selective deposit feeders such as Struthiolaria, and at least six pelecypods; detritus eaters (Echiuris, Amphiura, the synaptid Trochodota, and a hitherto unrecorded enteropneust, Ptychoderd) as well as carnivorous gastropods (Alcithoe and Ancilla) and polychaetes (Glycera, Nephthys, and a maldanid).

Pelicana vermis (Text-fig. 1) is confined to the North Island—accompanying in general the rather more common Struthiolaria papulosa. It is also more tolerant of muddy conditions, being well represented at Waikowhai in the Manukau Harbour, where Struthiolaria does not extend. In addition, both species probably occur fairly widely in the soft benthic sub-littoral, being recorded by Powell (1937) in the Maoricolpus+Dosinula and Tauera-)- Venericardia formations.

Struthiolaria has the general appearance of a typical mesogastropod prosobranch ; the most conspicuous external feature is the highly extensible foot and head region, for the Struthiolariidae— though sedentary feeders—have, like most sand-flat inhabitants, retained active powers of locomotion. Struthiolaria papulosa measures two or three times the length of Pelicaria vermis and is further distinguished by its somewhat handsome appearance—translucent milk-white in colour, with the exposed parts of the head, trunk, and foot marked with fine close-set lines of orange-red. Pelicaria vermis, on the other hand, is yellowish or clay-coloured, with rather less conspicuous lineations of rust-red. The cephalic tentacles in Struthiolaria are tapering and sharply pointed, in Pelicaria less slender and more bluntly tipped.

The foot is very labile and possesses a broad oval creeping surface attached to the trunk by a short, cylindrical ‘waist’ region capable of great elongation. The operculum is reduced to a small chitinous plate with the distal end produced into a strong, sharp claw which is at times employed as an accessory locomotor organ (Text-fig. 3). The Struthiolariidae exhibit two types of movement, the first by using the plantar surface of the foot, the other by the levering action of the operculum. The normal creeping movement on the surface of the sand is performed by the widely expanded sole, the operculum remaining out of use on the dorsal surface of the foot. There is also a second mode of progression by the use of the sole which is observed to best advantage upon a hard surface as at the bottom of a glass dish. The sole is placed firmly upon the substratum and the body-whorl of the shell raised clear of the ground by the elongation of the muscular ‘waist’ region ; the heavy shell is then allowed to fall some distance forward in the direction of movement, and the sole is again moved forward in advance of the shell for the repetition of this rather clumsy lunging movement (cf. Yonge (1937), on Aporrhais).

When the operculum is employed in movement, the extent of the sole is greatly reduced by withdrawal of blood from the pedal sinus, and the foot and trunk elongate to form a narrow muscular column. The sole—now small and oval—is displaced dorsally, and the muscular lobe carrying the operculum is thrust forward so that the sharp claw projects from the tip of the foot. The opercular claw is used mainly in the characteristic movement of ‘righting’ the shell when the animal is overturned; it is thrust round forcibly beneath the left side of the shell and dug deeply into the substratum, when the considerable leverage of the foot serves to heave the shell over to its normal position. The claw of the operculum may also be used as a defensive weapon when a specimen is taken up on the hand, the sharp-tipped foot being extended for 2 to 3 in. and moved about vigorously seeking a point of purchase. The opercular movements of Struthiolaria are chiefly of interest in showing how the normal mode of locomotion in the related highly specialized Strombidae may have arisen. In this group the sole is vestigial and the animal progresses by a series of convulsive leaps by the flexion of the foot as the opercular blade is thrust into the substratum.

During feeding Struthiolaria lies buried immediately beneath the surface of the sand and constructs a pair of mucus-lined siphonal tubes. It is not in the strictest sense adapted for burrowing; despite Oliver’s statement (1923) that it ploughs along below the surface, it is quite clear that progression does not take place while the animal is buried, although the feeding position is abandoned at frequent intervals and the animal crawls about freely upon the surface. The nodulose turreted shell is characteristically that of a surface-dwelling gastropod, in contrast with burrowing forms at Cheltenham such as Ancilla and Pervicacia.

The animal submerges itself by a continuation of the normal crawling movement, and anterior canal being depressed slightly and pushed below the sand, while the labile anterior margin of the foot is thrust downward, obtaining purchase in the sand and drawing the shell behind it. Gradually, the pallial current and the entrenching movements of the foot clear a space into which the whole shell subsides. The buried animal may usually be located by a small raised mound of sand as well as by the two openings of the siphonal tubes (Text-fig. 5), approximately in. apart and about one-sixth of an inch in diameter. These tubes, described originally in the related Aporrhais (Yonge, 1937), are unique among molluscs in being paired and widely separated—an inhalant tube directly above the left side of the animal and an exhalant tube on the right. The side-walls are very regular, being compacted with a thin secretion of mucus as in burrowing polychaetes and Enteropneusta.

The siphonal tubes of Struthiolaria are constructed by the action of the proboscis (Text-fig. 4), which exhibits a high degree of adaptation for its role. The retracted organ forms a short dorso-ventrally flattened tube, some three-quarters of an inch long, with its integument thrown into very regular, close-set annular rugae. By a copious inflow of blood from the body haemocoele into the rhynchocoele, the proboscis may be extended to form a cylindrical siphon-like organ, almost 3 in. in length, terminated by a circular oral disk bearing the vertical slit-like mouth at its centre. The oral disk is radially streaked with yellowish and grey, and may be flat, depressed to form a shallow funnel, or when fully expanded, somewhat convex. The integument of the proboscis is beset with numerous mucous cells, which are especially dense round the marginal rim of the oral disk. There are also large numbers of fusiform sensory receptor cells, of the same type as occur in the integument of the cephalic tentacles ; it appears that the proboscis serves also as a sensory organ for maintaining contact with the surface while the animal is buried.

In the formation of the siphonal tubes the proboscis is at short intervals pushed up through the sand to establish a pair of holes. It is first narrowly compressed by the contraction of the circular muscles passing round the rugae, and at the same time greatly elongated and pointed at the tip by partial inflow’ of blood and the relaxation of its longitudinal muscles (Text-fig. 6). Except for the oral disk, the wall is not everted as in the pleurecbolic introvert type of proboscis. When the oral disk reaches the surface, the circular muscles are relaxed, and the rhynchocoele tensely engorged by the increased blood-supply, the dilation of the rostral artery being assisted by the contraction of extrinsic muscle slips inserted on its Wall. The oral disk is now widely expanded with its marginal rim extending slightly beyond the edge of the siphonal tube (Text-fig. 4). The organ is then quickly withdrawn, and firmly moulds the wall of the tube in its downward passage, expressing a coat of mucus, especially from the periphery of the oral disk (Text-fig. 7). From time to time during feeding the proboscis is extended through one or other tube like a pelecypod siphon, removing obstructions, and maintaining sensory contact with the surface. A specimen removed from sand in a laboratory dish well exhibits the stereotyped tube-forming movements of this organ, with periodical erection, engorgement, and rapid withdrawal.

The pallial cavity in Struthiolaria (Text-fig. 8) is extremely spacious, and a continuous current of water enters from the inhalant tube and leaves by the exhalant, serving respiratory, cleansing, and food-collecting functions. On taking up a specimen in the hand, a strong jet of water is expelled from the exhalant side by the sudden retraction of the head and foot, which serves to close the cavity in front. The aperture is otherwise unprotected and the free margin of the mantle forms a continuous skirt covering the wide callused peristome of the shell. It is liberally supplied with blood, and as the gill is primarily specialized as a current-producing organ, apparently serves an accessory respiratory function. Just behind the right mantle margin and a short distance in front of the anus, lies a single pallial tentacle (Text-fig. 8) resembling in appearance one of the cephalic tentacles. It is extended in life through the exhalant siphonal tube, while the left cephalic tentacle passes through the inhalant tube. The pallial tentacle is densely coated with cilia, which maintain a’ strong outward current through the siphonal tube serving to carry away waste products from the pallial cavity. Aporrhais and the burrowing Strombidae have a similar pallia! tentacle which presumably serves the same function, and Yonge (1947) has described a similar mechanism in the unrelated genus Valvata.

Struthiolaria obtains its food by ciliary means, the inhalant current carrying into the pallial cavity a stream of roughly selected detritus and micro-organisms including diatoms and Foraminifera, as well as a good deal of non-nutritive material such as shell fragments, spicules, and sand grains. Of prime importance in the collecting of food particles for ingestion is the ctenidium, which in Struthiolaria reaches a length of 2 to 3 in. It is immediately apparent on opening the pallial cavity (Text-fig. 8), extending back from the margin of the mantle (across which the anterior filaments may protrude) to the narrow posterior end of the cavity, thus completely encircling the body-whorl. The ctenidial axis (Text-fig. 13) lies along the left side of the mantle, and the monopectinate lamina arches to the right across the whole-width of the mantle cavity. It is composed of some 300 – 400 tubular filaments, each a narrow, laterally compressed rod, attached proximally to the mantle wall on the left side, along the first half of its dorsal edge, while its distal half is entirely free. The great development of the ctenidium divides the pallial cavity (Text-fig. 9) obliquely into two longitudinal compartments each roughly triangular in section. The left compartment or inhalant chamber is the more ventrally placed, being floored by the dorsal surface of the trunk and bounded above by the curved ctenidial septum, extending from the roof of the pallial cavity on the left to near the floor on the right. The right, or exhalant, chamber lies somewhat dorsally to the inhalant, its floor being the gill septum while it is roofed by the hypobranchial, or pallial, mucous gland (Text-fig. 9). The pallial epithelium is here extremely thickened and thrown into wide, yellowish-brown transverse rugae, secreting much colourless viscid mucus.

The individual gill filaments are highly adapted to the mode of life. Like the triangular leaflet of the generalized prosobranch gill (Yonge, 1938), each is formed of a membranous fold of integument enclosing a narrow respiratory blood space, and basally attached to the mantle. In Struthiolaria the filament apex is carried across to the right side of the pallial cavity, to form a long, free distal portion, strengthened by the well-developed, paired, chitinous skeletal rods. In histological structure (Text-figs. 10, 11) the filament is note-worthy for the relatively increased area occupied by the very long lateral cilia (Text-fig. 10). They maintain a strong flexual beat across the filament from ventral to dorsal sides, creating a continuous water current from inhalant to exhalant chambers. In addition to the lateral field there are two principal tracts of shorter cilia; the ventral edge of the filament is clad with short, fast-beating frontal cilia, which direct a current towards the apex, while along the dorsal edge runs a tract of abfrontal cilia, smaller and with a weaker beat than the frontals. The original function of the frontal and abfrontal cilia was no doubt to collect particles for transport to the free edge of the gill and rejection from the pallial cavity (Yonge, 1947). In Struthiolaria, as in other ciliary feeding Prosobranchia (Yonge, 1938), it was only a short step to establish a feeding mechanism by which nutritive particles were collected by the frontal and abfrontal currents.

The inhalant current bearing food and detritus is drawn into the pallial cavity on the left across the small triangular siphonal lappet (Text-fig. 8), passing inwards along the osphradium (Text-fig. 13) which may be regarded as an organ for the detection of entering sediment (Hulbert and Yonge, 1937). In Struthiolaria, the osphradium forms a simple linear sensory tract which runs across the siphonal lappet, and passes backwards parallel to the anterior third of the gill axis. It is edged on either side by a narrow dark-pigmented ridge bearing a tall ciliary fringe whose beat serves to divide the entering stream of particles. Some of these seem to be directed across the trunk to the right, but by far the greater number are drawn by the strong ctenidial current to the left, where—before reaching the gill—they pass across a special mucus-secreting region or ‘endostyle’. This is a narrow tract of tall (40 μ) ciliated and glandular epithelium of very uniform structure (Text-figs. 13, 14) overlying the efferent branchial vessel along the entire gill axis. Narrow wedge-shaped ciliated cells, with ciliary coat 6-7 /x in height, alternate regularly with the long cigar-shaped gland cells, filled with dark-staining secretion. The fusiform nuclei of the ciliated cells are displaced upward to the free surface; those of the gland cells are rounded and basal. Following Orton (1912) and subsequently Yonge (1938) and Graham (1938) this region is referred to as an endostyle ; it is not, however, homologous with the similarly named structure in ascidians and Cephalochordata, which is located within the gut along the floor of the pharynx.

The secretion of the endostyle entangles food and particles of detritus while at the same time the ciliary field maintains a rapid transverse beat, carrying a continuous sheet of mucus with entrapped particles across the ctenidial axis to the frontal surface of the gill. The frontal cilia at once carry the particles across the ventral aspect to the right margin of the gill. The lateral cilia beating inwards between the filaments serve as a sieve mechanism for straining off solid particles from the respiratory current. Such smaller particles as may pass through the sieve between the filaments receive a small accretion of mucus from the scattered gland cells in the respiratory epithelium of the filament, and are finally carried to the apex by the abfrontal ciliary tract.

The termination of each filament (Text-fig. 12) is bluntly rounded and slightly expanded; there are no special cilia creating a forward current along the gill as in Crepidula and Vermetus (Yonge, 1938), but a uniform coat of short terminal cilia beating towards the tip, the columnar ciliated cells being interspersed with a few mucus-secreting cells. Just behind the tip of the filament on the ventral (or frontal) side is a small depression where the frontal and abfrontal currents converge, the abfrontal current with its particles having passed around the apex. Within the depression the particles of the two streams are intermingled by a rapid ciliary rotation, and—assisted by a small amount of mucus—are rounded off into a tiny spherical bolus. The series of depressions on successive filaments together constitute a shallow longitudinal food-collecting groove, and the mucous boluses become continuous to form a thin thread of food material clearly visible to the naked eye along the ventral edge of the gill.

In the living animal the down-curved gill projects into a well-marked excavation of the dorsal surface of the trunk, occupying the right-hand portion of the pallial cavity floor, and referred to as the food groove (Text-fig. 8). It continues forward in front of the pallial cavity along the right side of the trunk as far as the base of the right tentacle, becoming considerably narrower and bounded on either side by tall integumentary folds which may be temporarily approximated to form a closed tube. The bounding fold along the right side encloses beneath its outer edge the ciliated genital furrow in both sexes. The marginal folds of the groove are extremely labile, being liberally supplied with blood and capable of considerable muscular movement. The ciliated epithelium is richly beset with unicellular mucous glands, whose contents render the integument a characteristic greenish-grey in colour, and stain black in haematoxylin.

From time to time the narrow thread of mucus along the edge of the gill is rolled off by ciliary rotation into the posterior part of the food groove. Other particles enter the groove from the surface of the trunk, passing between small muscular crenulations of the left wall. The wide floor of the groove, while less glandular than the margins, maintains a rapid ciliary current which sweeps particles forward towards the region of the head. At the same time a liberal secretion of mucus is received, while the ciliated coat of the side-walls begins to rotate the contents passing forward through the temporary tube, so as to form a long spiral mucous cord, which finally issues from the spout-like opening of the groove near the base of the proboscis. The food string is usually greyish-brown in colour, with a large content of recognizable detritus from the substratum. In animals kept in clear water it becomes opaque white consisting of almost pure mucus and resembling a strand of wool. The contents of the string are surrounded by a delicate pellicle of mucus which becomes condensed on contact with the external medium.

At regular intervals the proboscis is turned backwards to the opening of the food groove and the paired jaws pluck at the tuft of issuing food material, which is pulled away in strands and either rapidly ingested, or allowed to accumulate in a small heap at the side of the animal. The radula appears to be used principally to rake food material through the pharynx, after a bolus has been picked up or detached by the jaws. It would seem likely that the method of ciliary feeding provides the whole means of subsistence in Struthio-laria. The proboscis is evidently modified wholly for constructing the siphonal tubes; though frequently seen to explore the ground like a sensory organ, it was never observed to pick up particles and does not appear adapted for collecting food.

The alimentary canal in Struthiolaria is highly adapted for the slow regular intake of fine detrital particles. As in numerous other microphytophagous mesogastropods a crystalline style is present, and the most distinctive features of the digestive system are :

I. The loss of triturating function by the pharynx.

II. The reduction of the oesophagus to a mucus-secreting region com-eying a food string to the stomach.

III. The reduction of muscular tissue in the gut in general and the increased reliance on ciliary manipulation of food and faeces.

IV. The specialization of the stomach for the sorting of particles.

V.. The absence of extracellular enzymes, apart from that of the style, and the ingestion of particles by the digestive diverticula.

VI. The adaptation of the intestine for producing firm faecal pellets to avoid fouling of the pallial cavity.

VII. The frequent occurrence of wandering phagocytic cells performing an accessory digestive function.

The course of the alimentary canal (Text-fig. 8) is relatively simple. The mouth opens into a small pharyngeal bulb, leading into a long narrow oeso-phagus, which passes directly through the trunk cavity to open into the stomach on the left side. The stomach is a large rounded chamber, occupying with the crystalline style caecum the whole left aspect of the first visceral whorl. It gives exit to paired digestive diverticula which ramify to form the two massive, asymmetrical lobes of the digestive gland ; the right or anterior lobe is the smaller and embraces the deep aspect of the stomach, whilst the left or posterior lobe is spirally coiled, comprising, with the gonad, the greater part of the visceral hump. The style caecum and the proximal division of the intestine (Text-fig. 17) open forward together from the stomach, the caecum on the right side overlying and partly concealing the intestine. Just behind the pericardium the intestine turns sharply below the apex of the caecum, and emerges on the right side of the visceral spire as the narrow middle intestine which loops back around the renal organ and then passes forward along the right pallial wall into the wider rectum. The anus opens anteriorly upon a small spout-like papilla, immediately behind the tentacle on the right pallial margin.

The pharynx presents no special features in Struthiolaria, being greatly reduced in consequence of its loss of function. A pair of cuticular jaws is retained in the form of small triangular plates lining the pharynx wall just within the slit-like mouth aperture. This region is readily reversible, and the sharp chitinized margins of the jaws serve to strengthen the edges of the mouth for grasping mucous boluses from the food groove. The radula, although equipped with sharp, curved marginal teeth as in other ciliary feeding gastropods, was not observed to come into play in seizing food. It is exceptionally small in relation to the animal, and its caecum is very short, scarcely emerging through the floor of the pharynx. There is the typical taenioglossan formula (Text-fig. 15) of seven teeth in each row: the laterals are rectangular in shape with a finely denticulate cusp at the mesial edge. The quadrangular or five-sided central tooth carries a broad, finely serrate triangular cusp. The salivary glands are vestigial, reduced to a pair of tiny white lobules closely flattened against the roof of the pharynx at the base of the oesophagus. They are histologically simple, the epithelial cells each containing a large mucous spherule with no apparent enzymatic contents.

The ciliated, glandular dorsal food channel of the pharynx continues back along the anterior division of the oesophagus, where it is bounded by a pair of rather prominent dorso-lateral folds. There are-other less permanent epithelial folds, and the whole region forms a thin-walled narrow tube, lined with columnar cells bearing a tall ciliary coat (18 /x) interspersed with very numerous fusiform mucous cells of the same type as those in the margins of the food groove. The glandular tracts of the oesophagus are thus a conspicuous grey-green in colour. The middle region of the oesophagus, commencing immediately behind the nerve ring, shows a condition somewhat less advanced than in the ciliary feeders Crepidula and Turritella (Graham, 1939): its topographically dorsal portion forms a structural remnant of the spacious oesophageal crop of less specialized prosobranchs, the food channel passing, as a result of torsion, around the left side to the floor of the oesophagus where it proceeds backward as a wide greenish epithelial tract. The more extensive dorsal portion is transparent and non-ciliated with the lining thrown into small, papillose longitudinal folds. Its epithelium is small-celled and cubical, devoid of glandular elements.

The most posterior part of the oesophagus (Text-fig. 16) possesses—unlike the rest of the alimentary canal—a fairly thick coat of circular muscle (sometimes reaching 100/y). The lining epithelium is uniformly ciliated, thrown into 12-20 regular longitudinal folds. The mucous glands—though less abundant than anteriorly—are still frequent, and it is in this region that the food string receives its final shape, being carried back by the cilia along the summits of the folds to the stomach.

The stomach, style caecum, and proximal intestine form a single functional unit of the alimentary canal (Text-fig. 17). The stomach is roughly flask-shaped, consisting of a wide posterior chamber receiving the oesophagus, and opening forward into a smaller anterior chamber, which leads in front by a common aperture to the style caecum and proximal intestine. The caecum remains in wide communication with the intestine for about three-quarters of its length, the two cavities being incompletely separated by a pair of typhlosoles—the dorsal one very large, and the ventral a low glandular ridge.

The crystalline style projects in life into the anterior chamber, its head bearing like a pestle against the gastric shield, a triangular outgrowth of the ventral stomach wall—transparent and brittle in texture and superficially resembling cartilage in appearance. The surface of the shield is slightly concave and the free edge curves backward towards the posterior chamber. Close to the gastric shield open the paired digestive diverticula. The anterior aperture lies just behind the opening of the style caecum adjacent to the head of the style; the posterior diverticulum opens by a spout-like lip, just below the left side of the gastric shield.

The dorsal wall of the stomach is occupied by an extensive ciliary sorting area, a series of close-set ridges commencing posteriorly and passing obliquely forward around the left side of the stomach to the opening of the proximal intestine. The sorting area is a structure highly typical of ciliary and detritus feeding molluscs, being especially well developed in Struthiolaria. Its limits are easily seen externally through the transparent wall of the stomach, being marked on the right by a broad S-shaped ridge which continues backwards around the fundus of the posterior chamber. Graham (1939) points out that this ridge—with the sorting area within its crescent—probably represents a vestige of the spiral stomach caecum found in archaeogastropoda.

The style caecum in Struthiolaria is a short stout sac of 6 mm. diameter, recognizable externally by its deeply pigmented wall. Its epithelial lining is very regular, beset with small, close-set transverse rugae. The dorsal typhlosole which serves to delimit the caecum from the intestine and to grasp the style in the living animal, forms a wide double fold, L-shaped in section along most of its length, and produced along the right side into a broad style flange (Text-fig. .19, 20) which depends into the caecum and enwraps the style from below. A narrow strip of tall epithelium runs the whole length of the typhlosole, just behind the free edge of the style flange. This is the region of style secretion (Text-fig. 18), the epithelium staining very darkly and forming secretory droplets of similar appearance to the style substance. Posteriorly towards the stomach (Text-figs. 17, 18), the style flange with its secretory ridge becomes very wide and is reflected back across the flat summit of the typhlosole, forming a wide sleeve which completely invests the style from above. Anteriorly the apex of the caecum separates completely from the intestine (Text-fig. 20) by the coalescence of the left typhlosole margin with the ventral wall of the common caecum-intestinal chamber. The now narrow style flange remains in contact with the style within the caecum. The line of fusion lies to the left of the small ventral typhlosole which is thus also enclosed within the caecum, together with a narrow remnant closed off from the intestinal lining.

The crystalline style forms an extremely delicate taper.ed rod, 2 cm. in length, hyaline golden brown in colour, and very flexible. It is rapidly dissolved after cessation of feeding, and was best examined immediately upon removing the animal from water, being completely resorbed within an hour or two of collecting.

The lining epithelium of the stomach, in contrast to that of the oesophagus and of the intestine, is devoid of gland cells. There are two main histological regions, the ciliated sorting area, and the cuticulated area surrounding the base of the gastric shield. The sorting ridges are due entirely to differences in the height of the cells, varying from 80 μ along the folds to 30 μ in the grooves. The nuclei are elongate-ovoid, binucleolate, forming a subcentral row, and the ciliary fringe is especially well developed (12 – 15 μ), supported by dense clusters of fibrillae. The epithelium is conspicuously invaded by wandering phagocytes from the underlying zone of connective tissue and a basal supporting reticulum traverses a series of intra-epithelial canals. The cuticulate epithelial cells are extremely tall and narrow, reaching 200 μ in height, while the cells secreting the gastric shield proper become as tall as 750 μ. The nuclei are compressed and rod-like, forming a crowded subcentral row, while the secreted cuticle (10 p thick) is hyaline and structureless, attached to the epithelium by fine perpendicular strands resembling cilia.

The epithelium of the style caecum (Text-fig. 21) possesses extremely robust cilia forming a dense coat 25 μ in height. The ciliary basal granules form a prominent refractile line, and the cell fibrillae form dense bundles passing from the sides of the nuclei to the bases of the cells. The ovoid to spherical nuclei are 12 μ in length, binucleolate and with sparse chromatin. The cytoplasm is uniformly granular and without secretory contents. At the bases of the cells is developed a system of intracellular supporting fibres, intra-epithelial canals appearing in section as small, non-staining vacuoles.

The digestive gland in Struthiolaria shows typical alveolar structure, each diverticulum ramifying into a system of smaller ductules, which ultimately give rise to clusters of small, dark-coloured digestive lobules, each about 200 μ across. The ductules are lined throughout with columnar ciliated epithelium, the cells reaching 55 μ in height, with a ciliary coat 5 – 6 μ wide and fibrillar bundles well developed. The basal nuclei are ovoid, binucleolate, and 7 μ in length.

The glandular cells of the terminal lobules (Text-fig. 22) are of two distinct types, digestive and excretory. The lumen is generally triangular, bounded on three sides by tall columnar digestive cells which by increase in height may reduce the space to a triradiate cleft. At the bottom of each crypt a smaller pyramidal group of darkly pigmented excretory cells is intercalated between the digestive cells. The digestive cells possess small, darkly staining basal nuclei. The free surfaces—though always intact—are commonly rather convex, sometimes bulging and somewhat pseudopodial. There is some evidence that the border is normally ciliated in the resting condition, though cilia are not apparent after fixation. A row of tiny refractile granules underlies the free surface, with a short fan of fibrillae radiating through the superficial cytoplasm. The distal cytoplasm of the digestive cell is very coarsely granular and evidently contains ingested material, while the basal half of the cell is packed with small greenish spherules, 2 – 3 μ, across, somewhat oily in appearance and containing irregular clumps of particulate matter. The digestive epithelium is frequently invaded by phagocytes from the vascular interstitial connective tissue.

The pigmented excretory cells in the crypts appear to be of several types. There is usually a centrally placed, broad-based, flask-shaped cell, with a constricted neck, containing a single large dark pigment spherule, sometimes 25 – 30 across, round and dark-brown or black with a greenish oily iridescence. The nucleus is small and pressed flat against the cell base. Flanking the central cell on each side are several columnar excretory cells, of the second type, with small round nuclei as in the digestive cells. The free surface is rounded and pseudopodial, and the distal third packed with rather lightly staining granular contents ; the basal portion is uniformly dark-brown pigmented, occasionally invaded by phagocytes. That there are probably two types of columnar cell is indicated by the presence in some cases of much larger, ovoid or flattened nuclei, 10 p, across and with a single prominent nucleolus. From the nature of the granular inclusions in the distal cytoplasm it would seem that some at least of these pigmented cells approach in character the lime cells typically found in the gastropod digestive gland.

In Text-figs. 17 and 18 are illustrated the system of ciliary currents in the stomach and style caecum which together comprise the most specialized region of the gut. In this region the wall of the gut is firmly attached by muscle strands to the external body-wall, and is incapable of peristalsis ; the ciliary fields are thus all-important in securing movement of contents. The long robust cilia of the style caecum serve by their transverse beat to rotate the style on the bearings formed by the epithelial folds. The direction of rotation is clockwise, and in the transparent veliger larva the speed was observed to be about 40 turns a minute. At the same-time the dense but shorter cilia covering the style flange of the typhlosole maintain a strong beat towards the stomach, by which the rotating style is gradually thrust downward so that its head bears against the gastric shield. Yonge (1932) has emphasized the importance of the style as a mechanism for the continuous liberation of amylase in those molluscs, lamellibranchs, and micro-herbivorous prosobranchs, which have a slow continuous feeding process. Strong amylolytic action was detected by digestive experiments on the style of Struthiolaria, while no action was observed on either dissaccharides or cellulose. It would seem, however, that a possibly equally important function of the style is to afford mechanical assistance to the movement of the food contents in the stomach. The soft, rotating style head becomes securely attached to the end of the oesophagal food string, which is slowly but continuously drawn into the stomach, around the fundus of the posterior chamber and forward on the right side to the surface of the gastric shield. The cord is thrown into a close spiral by the style rotation, while the attached end is broken down and the contents freed, partly by attrition by the style head, but also, no doubt, by the lowered viscosity of the mucus in the stomach, brought about by the reduction of hydrogen ion concentration (Yonge, 1935).

The rotation of the style also effects a constant stirring of the dispersed stomach contents by which particles are repeatedly drawn across the ridges of the ciliary sorting area. In Struthiolaria a large amount of non-nutritive material must enter the stomach. The sorting area consists of flat-topped primary ridges alternating with smaller triangular secondary ridges running along the intervening grooves. The primary ridges bear tall ciliary fringes keeping up a steady transverse beat across the stomach wall at right-angles to the direction of the ridges. Coarser, heavier particles such as sand grains, thrown against the sorting area by style rotation, sink into the grooves, and are carried directly forward by the cilia of the secondary ridges to the proximal intestine. The lighter particles, on the other hand, are flicked across the sorting area by the ciliary tufts of the primary ridges, and are conveyed by special ciliary currents into the digestive diverticula.

The diverticula do not secrete; apart from the action of the style amylase, digestion is intracellular. The digestive gland forms the main absorptive region of the gut, and a series of simple spotting tests on extracts showed a normal complement of intracellular enzymes. Of proteins, fibrin and also powdered peptone were readily digested, best in slightly acid media; 5 per cent, methyl acetate was easily hydrolysed showing the presence of an esterase (probably a lipase), while a strong amylase was demonstrated by complete hydrolysis of starch solution. There was no action on the cellulose of cotton fibres and it is probable that, as in other detritus feeders, cellulose, if assimilated at all, demands previous bacterial or autolytic breakdown.

Fine particles appear to be ingested by the epithelium of the digestive lobules in its pseudopodial phase. The basal greenish spherules of the digestive cells invariably contain clumps of solid particles, which apparently represent the non-assimilable residue after intracellular digestion, probably surrounded by enzyme to form a fluid vacuole. At regular intervals these basal cell contents are returned to the stomach by the fragmentation of the digestive cells, and can be detected in sections in the form of tiny boluses of egesta, approximately 50 μ across, passing forward along the sorting area to the intestine. Each bolus contains closely compacted cell fragments containing the greenish vacuoles, as well as nuclei, phagocytic cells, and brown excretory spherules. Mansour (1946) claims that in the lamellibranchs enzymes are liberated into the stomach by the fragmentation of holocrine digestive cells. In Struthiolaria, however, it is quite clear that the fragmented particles are always in the nature of egesta, and pass directly to the intestine. Small traces of enzyme are doubtless liberated in this way into the stomach; in general, however, Struthiolaria conforms entirely to Yonge’s rule as to the intracellular nature of digestion in style-bearers.

Following Macmunn (1900) it may be concluded that the large dark spherules discharged from the excretory cells consist of a chlorophyllous pigment derived from food substances and extracted from the blood by special cells in the digestive lobules. Intracellular digestion of detrital particles must be accompanied by large absorption of plant pigments, and in style-bearing gastropods in particular the excretory mechanism of the digestive gland is well developed.

The abundance of wandering phagocytic cells in the subepithelial layer of most parts of the gut has already been mentioned. These cells readily invade the epithelium and ingest solid particles, though it is questionable whether they are primarily nutritive or are concerned with the removal of waste particles. It is certain that particles of no food value may be ingested, as was demonstrated by the accumulation of neutral red particles within the phagocytic cells of the sorting area of the veliger larva. The most probable hypothesis, in light of the work of Yonge (1926), is that the phagocytes emerge into the lumen where they ingest both free food particles, which are then intracellularly digested, as well as particles of rejected material. The presence of a very thick zone of phagocytes in the wall of the proximal intestine of Struthiolaria would point also to rejectory function.

The intestine, adapted solely for compaction of faeces, consists of three regions, the wide proximal intestine communicating with the style caecum, a harrow middle intestine, and a terminal rectum. The lining is ciliated and glandular throughout (Text-figs. 23, 24, 25), and in the proximal intestine the epithelium is disposed in tall longitudinal tracts, separated by narrower intervening grooves lined by much shorter cells. The ventral wall has two broad tracts, separated by a deeply incised groove along which are carried indigestible particles from the grooves of the gastric sorting area. The dorsal wall consists also of two broad tracts with a narrower suture, while along the left side runs a wide shallow channel of low-celled epithelium. The wall of the proximal intestine is not, as in the stomach, secured to the body-wall, and the edges of the broad tracts work freely upon the contents of the grooves, by means of a thin, continuous zone of muscle in the hind-gut wall. The faecal material is all the while liberally admixed with mucus, and particles are carried by ciliary currents across the tracts into the grooves, where they are conveyed forward to the middle intestine.

The faeces are formed into pellets during passage through the middle intestine, which has a characteristic trihedral structure, the lumen being bounded by three broad ciliated tracts sutured by three grooves of low epithelial cells. The ciliated cells of the broad tracts are extremely tall and narrow (120 μ in height as compared with 80 μ in the proximal intestine). They possess ovoid-elongate central nuclei, and are regularly interspersed with a series of greatly attenuated mucous cells, with a distal rank of cigar-shaped secretion droplets, and a basal row of rounded uninucleolate nuclei (Text-fig.24). In passing along the narrow grooves the faeces are firmly kneaded into a coherent mucous string, from which individual pellets are from time to time nipped off by slight peristaltic movements of the bounding tracts. The pellets are given their final compact form by ciliary rotation against the rectum wall, producing firm grey-green ovoid masses, approximately 0 · 25 mm. long.

The rectum is a long straight tube with its walls smooth or thrown when empty into small impermanent folds. The columnar ciliated cells (Text-fig.25)are much shorter (40-50/x) than in preceding regions, with ovoid basal nuclei. The gland cells are simple and ovoid, with light-staining contents which form the transparent coat finally surrounding the large masses of faecal pellets. The rectal wall near the anus is somewhat muscular, and the faeces are expelled by slight contractions in coherent strings which are immediately carried away in the exhalant pallial current.

Ciliary feeding in prosobranchiate gastropods has now been reported as the result of independent evolutionary change in five style-bearing families which otherwise have little in common, namely, the sessile Calyptraeidae (Orton, 1912), Vermetidae (Yonge, 1932), Capulidae (Yonge, 1938), and in the free-moving Turitellidae (Graham, 1938). To these the also active Struthiola-riidae must now be added. Similar pallial adaptations have been acquired in each family, including strongly ciliated, linear gill filaments, and mucus-producing endostyle and food groove. The Calyptraeidae (Calyptraea and Crepidula) are the most highly specialized group, being suspension feeders, largely on diatoms, and having the gill filaments free and rod-like along their whole length, with the blood space reduced, and the membranous respiratory area entirely lost. Yonge correlates the extreme specialization of the gill filament with a reduction in respiratory needs of the sessile Crepidula’, in the actively moving Struthiolaria the filament has, in the proximal region at least, retained its respiratory function.

The Struthiolariidae most closely resemble in their biology the family Turritellidae (Graham, 1938; Yonge, 1946), as typified in New Zealand by Maoricolpus rosea, examined during the present work. These two groups are the only recorded ciliary feeding prosobranchs free-moving in a soft substratum. In neither case has the gill filament become free along its whole length, while in both groups there are convergent adaptations for dealing with the large amounts of non-nutritive bottom material that must enter the pallial cavity during feeding. The unicellular mucous glands of the gill filament, and the large hypobranchial gland are retained, while the food groove —a mere shallow depression in Crepidula—forms in Struthiolaria and Turri- tella a muscular, temporarily closed tube, capable of compacting large amounts of imperfectly sorted surface detritus. In both forms the gastric ciliary sorting area, and the faeces-compacting region of the proximal intestine, only slightly. represented in Crepidula, are prominently developed. Turritella, with its long tapering shell, is pre-eminently adapted for trailing through mud, and the pallial cavity is screened from sediment by a curtain of pinnate pallial tentacles. The absence of such a mechanism in Struthiolaria may probably be ascribed to the cleaner, sandy habitat, and to the efficient mucus-lined siphonal tubes.

The Struthiolariidae are classified by Thiele (1931) in his Stirps Strombacea, along with the families Aporrhaidae, Strombidae, and Xenophoridae. Unpublished observations of the present writer indicate that the Xenophoridae are somewhat removed from the remaining three families, which together form a compact natural assemblage—with an especially close affinity between the Aporrhaidae and Struthiolariidae. In general characters Aporrhais pes- pelicani, as described by Yonge (1937), is very similar to Struthiolaria. The habitat is alike in both groups, and in mode of locomotion Aporrhais agrees with Struthiolaria in showing both creeping and lunging movements with the widely expanded sole. Yonge states that the reduced operculum in Aporrhais is never used in locomotion, and that the righting movement of the upturned animal is effected not with the operculum after the strombid fashion, but by leverage obtained by placing the sole flat on the substratum. However, examination of the operculum in preserved Aporrhais pes-pelicani reveals that the distal edge is produced into a strong, spade-like plate, not figured by Yonge, which has all the appearance of a locomotor organ like the claw in Struthiolaria.

Aporrhais, like Struthiolaria, constructs paired, mucus-lined siphonal tubes by the action of the proboscis. This adaptation, which is recorded nowhere else among gastropods, would appear to have arisen as a single evolutionary development in a common ancestral form of the Aporrhaidae and Struthiolariidae. Aporrhais has not acquired a ciliary feeding mechanism ; the pallial organs, while forming an efficient cleansing and rejection system, are not modified for food collecting, and the gill filaments remain triangular. The inhalant current brings detrital particles to the vicinity of the rostrum and according to Yonge feeding takes place as readily upon the surface as when buried. ‘The animal presumably feeds normally by collecting by means of the extensible proboscis all particles of plant matter which occur in the mud in the region below and around the expanded lip.’ Yonge stresses the creation of powerful water currents by the gill in Aporrhais and the elaboration of ciliary currents for the rejection of sediment. He emphasizes that these developments ‘have an added interest in that they indicate the way in which ciliary feeding in gastropods may have been evolved’. The exhalant currents might further be modified to form the food channel, while modifications of the tips of the gill filaments would convert the gill into the food-collecting organ found in Crepidula. Such a process is exactly what is now found to have occurred in the Struthiolariidae, which apparently represent an almost direct continuation of the aporrhaid trend of evolution. It is of considerable interest, since Struthiolaria has not hitherto been investigated in life, to find Yonge’s hypothetical conclusion actually realized in the closest living relatives of the Aporrhaidae.

In the digestive system Struthiolaria shows in comparison with Aporrhais a number of structural advances associated with ciliary feeding. The alimentary canal of Aporrhais (Yonge, 1937) examined by the present writer from fixed material has an essentially similar plan to that of Struthiolaria’, the foregut, however, is a good deal more generalized, and the condition of the digestive and other systems indicates that the Aporrhaidae may have given rise fairly directly both to the Struthiolariidae and to the somewhat less closely related Strombidae. The buccal bulb is much less reduced than in Struthiolaria, and the radula—while already small—remains large enough for the long marginal teeth to be employed with the jaws in seizing particles. The anterior oesophagus is much wider than in Struthiolaria, and the mid-oesophagus retains as in the Strombidae a long cylindrical crop traversed by the ciliated glandular food channel, ventral in position following torsion. The salivary glands remain as large lobulated masses within the trunk cavity, sending forward through the nerve ring a pair of long ducts which run along the oesophagal roof to the pharynx. The stomach and style sac complex of Aporrhais is closely similar to that of Struthiolaria, the figure of Yonge (1937) requiring correction by the transposition of the captions of the oesophagus and anterior digestive diverticulum.

Struthiolaria possesses further specializations upon the aporrhaid plan, especially in the nervous system, and the reproductive organs and life-history. The Aporrhaidae are a somewhat ancient family with a world-wide Jurassic range, and are represented by several genera in the New Zealand Cretaceous (Struthioptera, Drepanochilus, Hemichenopus, Dicroloma, Perissoptera). The first undoubted struthiolariids belong to the earliest South American Tertiary^: (Marwick, 1924).

The present writer has examined the shell and animal of Perissodonta georgiana, of South Georgia, one of the two species of the most archaic living genus. On both shell and dentition features this group appears to stand close to the line of descent from an aporrhaid stock. In all essential features, however, the animal is a true struthiolariid, and the pallial organs are already fully adapted for ciliary feeding.

The writer wishes to thank Prof. W. R. McGregor, Head of the Zoology Department, Auckland University College, for assistance during the preparation of the thesis of which these results formed part, Mr. A. W. B. Powell, Assistant Director, Auckland Museum, for much helpful discussion and frequent loan of specimens, and the Director, Marine Biological Association of the United Kingdom, Plymouth, for generously supplying preserved material of Aporrhais. Finally, the writer has had the special advantage of Professor C. M. Yonge’s kindly criticism of the manuscript.