region, where it is six or more cells thick. The dorsoventral muscles are meagrely represented. In general there are only isolated individual cells in this plane.

Internal to the muscle layers, throughout the abdominal region there is a thick layer of large clitellar gland cells. This layer is some 0.3 mm. in width and about the width of four or five gland cells. These cells have the usual cytology. Their ducts extend medially to form a thick band of firm tissue which simulates longitudinal muscle tissue in macroscopic appearance, but is more brittle. This mass of ducts (Fig. 1) extends as a definite layer from about XXIII anteriorly on either side of, and above, the intestine, then lateral to the testes into the neck, where it terminates in the clitellar region. It is not obvious anterior to X, where there is on either side a thick mass of clitellar gland cells of smaller size with short ducts passing directly to the exterior. The anterior region of the neck is essentially filled by these gland cells, there being left only the passageway for the pharynx and room for the oesophageal glands which enter the pharynx from the rear. The transparency of the body wall is related to the distribution of the large gland cells. These are essentially lacking from the midventral line where the nerve-cord, testes, etc., are visible in life; but their presence beneath the dorsal and lateral surfaces renders these surfaces opaque.

The ventral nerve cord contains the usual 21 independent ganglia with one immediately dorsal to the male genital pore (Fig. 6) and five anterior to this. Accepting the male pore as situated in XI, as is usual, then the first free ganglion is that of VI and the last will be XXVI, which suggests that the posterior ganglionic mass must represent eight instead of the usual seven somites. Examination of the circumpharyngeal ganglionic mass showed apparently five pairs of emergent nerves, which is in agreement with the interpretation of the segmentation as above. The first nerve is stout, the next three nerves are more delicate, the last nerve appears to divide immediately into a nerve of the diameter of the second, third and fourth, and a more delicate branch. These two have no common root (Fig. 5), and diverge widely at emergence, suggesting that they represent not one segmental nerve, but two, which is confirmed from sections. This means that the anterior mass represents six fused ganglia which would place the male genital pore in XII. The independent ganglia are each formed of the usual six ganglionic capsules; one pair, anterodorsal, one pair, posterodorsal; and an anterior and a posterior midventral pair. The circumpharyngeal ganglionic mass consists by count of 36 ganglionic capsules. There is a median ventral series of four transverse pairs with a posterior longitudinal pair, 10 in all. There are 13 ganglia above these on each side. Accordingly it is clear that there are six segmental ganglia fused into the circumpharyngeal mass and that the last two nerves are those of segments V and VI. The first free ganglion is accordingly VII, the last is XXVII. There are seven ganglia in the posterior ganglionic mass. The independent ganglia are not noticeably misplaced, but are central in their respective somites and at the level of the pulsatile vesicles.

The male reproductive system (Fig. 6) consists of six pairs of testes situated intersegmentally at XIX/XVIII, XVIII/XVII, XVII/XVI, XVI/XV, XV/XIV, XIV/XIII. Each is joined by a short duct to the dorsal longitudinal vas deferens which expands into a seminal vesicle in XIII/XII and continues loosely coiled as the epididymis in XII. This terminates in the wider dorsal limb of the atrial cornu which extends anteriorly to IX (but in one specimen reaches VII on the left), turns sharply ventrally and posteriorly to join the atrium in

XII. The latter is small in all material so far examined and opens in the middle of XII. The pars glandularis is extensive. In one specimen the area around the male pore was extended ventrally as a simple cone, of a length equal to the width of the neck, with the male pore at the distal end as though this were a copulatory device. The female system (Fig. 6) consists of two elongate ovaries in XV, the left commonly larger than the right and partially overlying the ganglion of the segment to which both are closely pressed. The oviducts are thinwalled, extend anteriorly dorsal to the ganglia of XIV and XIII, and enter a very small, thin-walled atrium ventral to the nerve-cord in the anterior portion of XIII. The atrium has a small dorsal median pocket and opens through a small pore at XII/XIII. There are at most three annuli between the genital pores. Neither in dissection nor in longitudinal horizontal sections have I been able to determine any obvious conductive tissue or connections between the male and female systems, not even as Brumpt (1900) describes for Branchellion torpedinis where the conductive tissue is superficial, immediately adjacent to the female genital pore and largely concealed beneath the preputial fold.

The mouth pore is minute. The proboscis is cylindrical, short, and reaches only to IX in retraction (Fig. 1). The oesophagus is very delicate and is situated immediately above the nerve cord. The large oesophageal glands are few, ranged alongside the pharynx and possibly not more than a dozen in all. In life these glands are white and opaque. The adjacent clitellar gland masses are transparent. The oesophagus expands into the thin-walled, six-chambered stomach which extends from XII/XIII to XVIII/XIX. The first three chambers are crowded between the testes, simple dilations and not lobed. The last three chambers are lobed, subdivided laterally, and extend dorsal to the testes. No sphincter can be detected between the oesophagus and the first chamber, nor between the first, second, or third chambers; but there is sharp constriction and possibly sphincters, between the fourth and fifth, and the fifth and sixth chambers. A distinct sphincter separates the last chamber from the intestine.

The intestine extends from XVIII/XIX to XXII/XXIII (Fig. 1). Its wall is glandular and variously corrugated externally and internally. There are no diverticula in XIX, the intestine traverses this segment as a simple tube which is constricted at XIX/XX where the anterior portion of the next section of the intestine expands into paired lateral pouches which occupy the anterior half of XX and taper to a tubular region extending through the latter half of the segment. This is repeated in XXI and XXII; but there is a peculiar distortion in XXII and XXIII. The posterior tubular part of the intestine in XXII narrows to enter an asymmetrically-lobed region which consists of a small left lobe and a larger right lobe. Segmentally, this should be part of the normal complement of the anterior half of XXIII, but the lobes are even more anterior and situated actually in the posterior portion of XXII. The corresponding tubular portion takes its origin from the posterior aspect of the right lobe, extends transversely to the midline where it turns posteriorly to constrict at about the XXIII a₁/a₂. Here it enters a small bilobed chamber. The wall of this chamber is thin, nonglandular, smooth, similar to that of the rectum into which it opens after constriction, but still freely. This is apparently the first section of the rectum. There has been extension and torsion of the latter part of the intestine producing this asymmetrical form of the last region of the intestine. Accordingly the intestine has essentially four pairs of lateral pouches. The rectum commences

at XXIII a₂ in the bilobed section described above. This is followed by a wide tapering region which has four or five pairs of simple lateral pouches, the last are the smallest and in XXV. Behind this the rectum continues and terminates with a small but definite pouch extending behind the anus into XXVII. The anus opens at XXVI/XXVII. There is no suggestion of posteriorly directed gastric caeca.

The smallest specimen is 29.0 mm. in length; its anterior sucker, 2.0 mm; the neck, 5.0 mm. long; the abdomen, 21 mm. long and 3.0 mm. wide; the posterior sucker, 5.0 mm. The largest specimen is 35.0 mm. long.

Host, Eptatretus cirrhatus (B. & S., 1801), Cook Strait, N.Z.

The holotype is held in the collection in the Department of Zoology, Victoria University College.

Branchellion parkeri Richardson 1949.

A single specimen was collected by Professor H. W. Manter from Callorhynchus milii during his examination of New Zealand marine fishes for their parasitic trematodes. This record from a holocephalan seems to be the first such for a leech of the g. Branchellion. The specimen was quite typical, but a specimen collected by Miss P. M. Ralph from a skate trawled off Dunedin and several specimens collected by Mr. J. A. Garrick during trawling operations of the “Maimai” off Cape Campbell show a colour pattern and other features which are novel to me.

The Dunedin specimen is large, and the first I have seen with a dark phase and a marked colour pattern. The total length is 58 mm.; the neck, 9.0 mm. long and 3.0 mm. at the widest part; the abdomen, 41.0 mm. long and 7.0 mm. wide excluding the gills. Preserved, the venter is essentially white, almost immaculate. The dorsum of the anterior sucker is richly coloured with black chromatophores; the neek, sparsely so; but the dorsum of the abdomen (Fig. 9) is very richly covered with melanophores, which are also present, but to a less extent, on the dorsal half of the gills. The ventral half of the gills and the entire pulsatile vesicles are immaculate. The number of melanophores diminishes on the last few gills, so that the gills of XXIII are plain, as is also the dorsum of XXVI, XXVII and of the posterior sucker.

Segments XIII to XXV are richly covered with melanophores excepting for some irregularly outlined marginal areas half or more the length of a segment and extending no more than a-third across the abdomen. The patterning is derived from small cream-coloured circular or subcircular areas having a diameter generally about a third of the length of an annulus. These are situated in the anterior portion of each annulus in front of the first two furrows which incompletely divide the dorsum of each annulus. They form a transverse row of four circles with two approximately in the paramedian position and two in the paramarginal so that in general view the dorsum of the abdomen appears marked with four longitudinal rows, a pattern which is somewhat obscured under magnification by many smaller pigmentless patches.

As mentioned in the original account, the dorsum of the typical abdominal segment shows clear and complete division into annuli of the “a” category, each such annulus being traversed by two shallow furrows which rarely extend to the actual margin, so that the dorsal aspect appears as though divided into nine annuli. Although not everywhere obvious, careful examination shows that

these nine annuli are not equivalent (Fig. 9). The circular pigmentless patches are located essentially in the anterior portion of the “a” category annuli, and if these patches can be given morphological significance it appears that this incomplete annulus is of the “b” category and that the posterior portion of the annulus shows incipient division into the “c” category. This suggestion is the stronger in the fact that the anterior incomplete annulus (b₁, b₃, b₅) is quite equal in length to the remainder of the annulus.

A further feature of the specimen from Dunedin is that it shows the fully everted male organ. (Figs. 10, 11.) This protrudes from the greatly enlarged male genital aperture as a short broad-based truncate cone terminating in a flattened disc of equal or possibly greater diameter than the base of the cone. The thin margin of this disc is divided into a pair of posterior lobes which form a small posterior lip set off cleanly from the greater part of the disc which is in turn divided into six subequal lobes together forming an anterior lip. This is cupped and folded over the central portion of the disc which is perforated by a slit-like, orange-coloured male aperture. This ornate everted male organ seems novel in the genus. In this case, the female genital aperture is slit-like, transverse, and as usual immediately posterior to the male aperture.

The specimens collected by Mr. Garrick were brought alive to the laboratory, and survived for several weeks in a tank with circulating water. Two were of medium size, and colourless as usual, the other four were large and showed the colour pattern of the Dunedin specimen. This leech does not swim. When released in the water there is not even an indication of the usual undulation by which a leech swims, although they are capable of such undulatory movement when attached. They are remarkably inert. None moved from the place of attachment unless grossly disturbed when they would shift by rapid looping, but only a matter of inches. Other than a very rare and simple undulating movement of the respiratory type, there was a slight simple forward sweep of the gills from time to time. At rest, the pulsatile vesicles are equally distended on both sides. Their contraction is brief and followed by a much longer period of relaxation.

These leeches were sensitive to a disturbance in the water, such as caused by a glass rod passed near them. A young live Squalus was placed in the tank, and as this shark moved past them the leeches swung their bodies towards the shark; but in no way was it possible to induce these leeches to attach to the shark, although their condition was then good, for they survived for a week beyond these experiments. It can be appreciated that in their inability to swim, in their sedentary habit, there are behaviouristic limitations which, while favouring attachment to skates and rays, must greatly reduce the opportunities for the attachment of this leech to sharks so that their “preference” for the former is more probably an outcome of their behaviour rather than an expression of some peculiar physiological distaste.

This additional material covering a fair range in size and generally in excellent condition has conformed reasonably to the original description, but none has so far shown the simple blunt undivided cylindrical process previously described as present on the gill-less first pulsatile vesicle. These have been gill-less in all cases; and it is possible that the lobe described from a preserved specimen may have been no more than a prominent development of the lateral end of the

anteriorly reflected folds which are such a striking feature of the venter of the abdomen. In several specimens, the XIII a₂ is deeply furrowed above, and this furrow extends on to the pulsatile vesicle, which then appears incompletely divided.

G. aff. Pontobdella.

The leech here described superficially is known to me from a single specimen 12.0 mm. in total length, which was accidentally destroyed before a full description could be taken. In twelve years no other specimen has come to hand. Although the data are meagre, the leech is sufficiently striking in appearance that the following account should permit an identification of this specimen should the species be found and described from elsewhere. Previously I have referred to this specimen as belonging to the g. Ichthyobdella as used in the later sense, but re-examination of my data indicates a possible affinity with the g. Pontobdella. The obvious features of this leech are: the lack of differentiation between the anterior sucker and the neck, which is cylindrical and expands and is continuous with the abdomen, which is convex above but flattened below; the opaque tuberculated integument; the absence of eyes, of external pulsatile vesicles; the triannulate abdominal somite with more, and smaller, tubercles on a₁ and a₃, than on a₂; the absence of tubercles from the venter; the absence of submarginal papillae from the anterior sucker. The general facies is pontobdellid, but in the absence of tubercles from the venter and of papillae, this leech differs sufficiently from pontobdellids that it may be more properly located in some other genus.

The specimen was taken during October, 1941, from a much perforated and eroded rock brought to the surface at East Bar, off the Wellington coast. It was found free in a crevice and could not be related to any host. The specimen was alive, somewhat gorged so that the annulation could be determined only as shown (Fig. 12). The anterior sucker and the origin of the neck were of uniform diameter and the two not differentiated from one another in dorsal or lateral view. Seen ventrally, the appearance of the sucker is more that of a glossiphonid than a piscicolid, the sucker being a shallow open excavation with thick margins; the mouth, a perforation slightly posterior to the centre. There was no indication of pigmented eyes or papillae and the sucker and adjacent region of the neck were free from obvious tubercles. The cylindrical neck widens gradually and is continuous with the abdomen, there being no indication of shoulders or prepuce, etc. The abdomen is flattened ventrally and boldly convex above, so that the appearance is that of an asymmetrical club. The posterior region of the abdomen tapers abruptly to a short peduncle, which is attached excentrically to the thin, well-formed, shallow, unornamented posterior sucker. This has a lesser diameter than the abdomen but is wider than the neck.

No pulsatile vesicles or other similar suckers are visible. Throughout the anterior sucker and adjacent part of the neck, the segmentation is obscure, but indicated on the greater part of the neck by the cutaneous tubercles so that while on the neck and abdomen together seventeen somites could be determined, only nine of these can be recognised as triannulate, which was obviously the condition of the fully formed abdominal somite. In these, the annuli are equal, the a₁ and a₃ both carry fourteen cutaneous tubercles which are quite low, and each has a minute but distinct apical cream patch. The annulus a₂ has six large