etc.; colour, white; and I have referred them previously to the g. Platybdella. A sixth specimen was found in our preparation room after a mixed collection of fish, including sharks and hagfish, had been injected and preserved. This specimen, contracted but in good condition, showed much of the appearance of a Platybdella; but dissection revealed the absence of posteriorly directed caeca.

In March, 1951, two specimens were taken from a collection of hagfish sent to the Department from Paraparaumu. These were alive and survived in seawater for two hours after reaching me. Both were small, between 3.0 and 3.5 cm. when fully extended. The general colour was uniformly white, under magnification the superficial tissues showing no chromatophores, but the deeper tissues were flecked with well-spaced, short, rod-like bodies tinged with yellow. The venter was highly transparent, so that the major organs of the reproductive system, the nerve cord, etc., were generally visible; but the body wall over the dorsal and lateral aspect was much denser and concealed the internal morphology.

These living specimens were invaluable since each showed pulsatile hemispherical vesicles on the abdominal region. No trace of such structures was visible in the previous material, but after relaxation with coal gas the vesicles remained when these specimens were preserved. Each bore eleven pairs of vesicles, each pair situated at the mid-level of the somite from the first abdominal somite posteriorly on the following ten somites. In dilation, the vesicles are rounded and prominent, light reddish in colour. When contracted, there is no indication of their presence. There seemed to be no particular rhythm in pulsation, either transversely or longitudinally although the contraction synchronised with the contraction of the lateral longitudinal vessel. When at rest the pulsatile vesicles are inflated on the both sides at the one time.

Further live material taken by Mr. Garrick from Eptatretus trawled off Cape Campbell by the “Maimai” early in March of this year were held alive at the laboratory in an aquarium with circulating water, but lived only a few days. These were unable to swim but were otherwise relatively active, more so at least than Branchellion. Attempts to allow these leeches to attach to Myxine biniplicata which we were keeping alive at the same time, were not successful. It can be noted here that although Eptatretus and Myxine are trawled in the same area, this leech is not yet known from Myxine.

The presence of pulsatile vesicles, the absence of posteriorly directed caeca, the presence of six pairs of testes, etc., shows that this species cannot be located in the g. Platybdella. There appear to be twenty-seven currently recognised genera in the F. Piscicolidae (syn. Ichthyobdellidae). The relationship of these is by no means clear; but for the present purpose it is adequate to divide these into two groups on the presence or absence of simple external pulsatile vesicles. Such vesicles are recorded as present in Calliobdella, Cystobranchus, Ganymedebdella, Oxytonostoma, Piscicola, Trachelobdella, and Trulliobdella. A subdivision of these can be made in terms of the last gastric caeca which are present in one form or another and posteriorly directed in these genera excepting Trulliobdella, and possibly Trachelobdella for which I lack information on this point. Trachelobdella has four preclitellar vesicles which are lacking from the present specimens. Using the above criteria, it appears that the leeches from the hagfish are referable or immediately comparable only with Brinkmann's genus Trulliobdella, also from southern waters, but they differ in the greater degree of annulation, the absence of eyes, the presence of six pairs of testes, the more uniform body

width, etc. Brinkmann (1948) has associated his genus with Branchellion, both having a fold organ located in Trulliobdella in XXIV, five pairs of testes, and eyes but Branchellion has posteriorly directed caeca. In contrast to a fold-organ this species shows a peculiar distortion and asymmetry of the intestine in XXII/XXIII (Fig. 1) which is not unlike that described by Badham (1916) in his Austrobdella translucens and having much the same somital arrangement; but of course A. translucens has well-developed fused posterior caeca. Other genera lacking the posteriorly directed last gastric caeca are Phyllobdella, Pterobdella and Pterobdellina. These lack the fold organ, have five pairs of testes, excepting Phyllobdella, which possibly has six pairs; but these genera have more or less extended lateral fin-like flanges and lack external pulsatile vesicles. Accordingly a new genus is established for the species from the hagfish as follows:—

Bdellamaris n.g.

Neck cylindrical, body depressed; body wall thin, transparent ventrally, thick and opaque dorsally; muscular envelope weakly developed; integument lacking papillae or tubercles; eyes lacking; the fully developed somite divided into twelve annuli; eleven pairs of pulsatile respiratory vesicles on the abdomen; no gills; testes, 6 pairs; seminal vesicle, small; posterior gastric caeca broadly lobed, not digitate; no posteriorly directed gastric caeca. Marine. Genotype, Bdellamaris eptatreti n.sp. as follows:—

Bdellamaris eptatreti n.sp (Figs. 1-8.)

The anterior sucker is essentially circular, its diameters equivalent; but the aperture somewhat less than the width of the sucker and directed ventrally or anteroventrally. The margins are thick and entire in life, but appear incised in the preserved specimen (Fig. 2). The mouth is posterior, about three-quarters back from the anterior margin and concealed by the ventral margin. The dorsal aspect is marked by two faint annuli which cannot be assigned to segments, the first annuli ascribable to segments are the three complete equivalent annuli of segment V which form the origin of the neck. There is no indication of eyes, submarginal papillae, or other such structures.

The neck increases slightly in diameter posteriorly to merge with the abdomen. The division of the two is distinguished at best by slight shoulders. The neck is 1/4 to 1/5 of the length of the abdomen. The segments are much subdivided and the annulation obscure since the intersegmental furrows are not evident and many tertiary furrows are incomplete. The ventral surface is more clearly marked than the dorsum (Figs. 2, 3.) Segment V, as described, is included in the neck, and equals segment VI, which is also trimeric. Segments VII to XII are fundamentally bimeric, the annuli of each being more or less divided by incomplete furrows so that some of these segments appear quadrimeric in ventral view, but bimeric dorsally. In detail, the three annuli of V are equivalent, as also of VI. The first annulus of VII is divided, the second is intact so that VII consists of two short anterior annuli each the half in length of the posterior annulus. Segments VIII and IX are bimeric, each subdivided into four equal but incomplete annuli. The first annulus of × is incompletely furrowed even ventrally, the second is divided into three incomplete subequal annuli. Segment XI is bimeric and not further furrowed. Segment XII is bimeric, the anterior annulus divided into two equal parts, the posterior into three. Segment XIII is the first segment