is a little projecting point at the middle of the apical margin, and the sides are crenulate as if in reminiscence of earlier segmentation. But these characters, which Hansen does not notice, only became conspicuous after the pleon had been detached. The stilet-like mandible shows a microscopic serration at the distal end. The almost linear maxilliped has a terminal seta.

The male has a faint greenish tinge, as contrasted with the pallid colour of the female. The relation of these tints to the colours of the living animals is uncertain.

Size. The female measures 11 mm. in length, and about the same in breadth. The male attains the unusual length of 8 mm., with a breadth of 2.5 mm.

Locality. Cape Point N.E. by E. E. 40 miles; depth, 800-900 fathoms; bottom, green mud. In Glyphocrangon sculptus (S. I. Smith).

AMPHIPODA.

GAMMARIDEA.

FAMILY LYSIANASSIDE.

For the families in this tribe a general reference may be given to Das Tierreich, Lieferung 21, 1906. But continually additions are being made to the number of genera and species.

GEN. TRISCHIZOSTOMA, Boeck.

1853. Guerinia (preocc.), Hope MS., Costa, Fauna Reg. Napoli, Apr., 1853, p. 1.

1861. Trischizostoma, Boeck, Forh. Skand. Naturf., Möde 8, p. 637. 1893. Guerina, Della Valle, Fauna und Flora des Golfes von Neapel,

vol. xx., p. 775.

1905. Guerinella, Chevreux, Bull. Mus. Monaco, No. 35, p. 7. 1906. Trischizostoma, Stebbing, Das Tierreich, Amphipoda, p. 12.

The rather rare crustaceans for which Hope and Costa, in 1853, established the genus Guerinia present some remarkable and some still perplexing features. The original generic name being preoccupied must be relinquished. In my opinion its place should be taken by Trischizostoma, which was independently defined by Boeck in 1861. That author thought it advisable to assign it to a distinct tribe, Prostomata, which, however, he subsequently reduced to a family, Prostomatidæ. In the meantime Lilljeborg had named a

sub-family Trischizostomatina, for which Sars, in 1882, substituted the family name Trischizostomidæ, improved by Bovallius, in 1886, into Trischizostomatidæ. In the preface to the report on the Challenger Amphipoda (1888), I remarked that "the Prostomatidæ are in close relationship with the Lysianassidæ, and might, in my view, well be included in the older family." This opinion has been fully confirmed by Professor Sars in his "Crustacea of Norway," vol. i., p. 30 (1890). It is unfortunate that Boeck, Bovallius, and Sars, when discussing and figuring the species called T. raschii, by Boeck, appear to have been in ignorance of Costa's Guerinia nicæensis, so that they have expressed no opinion as to the identity or distinctness of the Norwegian and Mediterranean forms, which until recently I have regarded as specifically one. Della Valle, on the other hand, has separated them not only specifically but generically, substituting the name Guerina for the preoccupied

He admits the great superficial similarity of the two forms, but in his contrasted generic definitions makes out a strong case for keeping them apart. Thus, Guerina has the mandibles with moderately broad trunk, the first maxilla without inner plate and with the palp reduced to a little simple tubercle, the second maxilla with the inner plate short, carrying a single setule, the maxillipeds with the first joint of the palp longer than the second, the first and second peræopods with the second joint much dilated. Trischizostoma has the mandibles with very narrow trunk, the first maxillæ with inner plate and with minute, slender, two-jointed palp, the second maxilla with the plates subequal, the inner carrying various little setules, the maxillipeds with the first joint of the palp considerably shorter than the second, the first and second peræopods with the second joint not dilated.

Before attributing overwhelming weight to these differences, one must remember that the Scandinavian authors had, among adult examples of their genus, only females to examine, while the three specimens assigned by Della Valle to his genus were, he says, probably all males. The mouth-organs are very delicate in structure, and, as often happens in tubiform arrangements, not very easy to separate. When successfully drawn apart their appearance, after flattening out, is in some respects extremely different from that which they bore in their natural position. There is also scarcely any doubt that some of the details differ with the age of the specimen.

Some additional light may be thrown on the question by specimens from South Africa, which certainly require a new specific name, but

may rest contentedly in Trischizostoma, unless that genus be separated from Guerina. They have the monstrous eyes and huge peculiar first gnathopods which are so remarkable alike in T. raschii and G. nicæensis. They have the mandibles of the latter, the first maxillæ of the former. As to the first joint of the palp in the maxillipeds, they agree better with the latter than the former. But they differ from both in the shape of the side-plates to the second gnathopods and the first and second peræopods, in the second joint of the fifth peræopods and still more in the flattened leaf-like appearance of the sixth joint, and lastly in having the telson divided for two-fifths of its length. The specimens are certainly of the male sex, and perhaps not fully adult, as the torsion of the first gnathopods had not been effected or not completed. To have three genera for three (or possibly only two) species, which share characters so very notable, while the real or supposed differences refer chiefly to degraded mouth-organs, seems to me to be for the present unadvisable.

In the new species the first maxilla has a well-pronounced inner plate of slight texture, an outer plate with five spine-teeth on the apex, four of them much curved, and a small one- or two-jointed palp tipped with a long seta. For T. raschii, Boeck describes the first maxilla as consisting apparently only of the outer plate; Bovallius endows it further with a one-jointed palp, Sars alone awarding to it a slender unarmed inner plate together with "a very minute, but distinctly biarticulate" palp. The supposed absence of the inner plate cannot, therefore, be relied on for distinguishing Costa's species from Boeck's. It is clear that this small pellucid structure sometimes escapes observation under the difficulties of dissection. That the small palp is at first one-jointed, and later on biarticulate, is far from impossible. Bovallius gives the telson of the young male, "probably just out from the incubatory pouch of the mother," as having the apical margin excavate. Boeck also speaks of the telson as apically divided, but probably by a misprint, as his Latin text says that it is apically rounded. That the new species with an incised telson should stand in the same genus with one (or two) species in which the telson is whole-rimmed, agrees with what is found in another Lysianassid genus, Onisimus, Boeck.

TRISCHIZOSTOMA REMIPES, n. sp.
Plate XXXIV.

The head is almost entirely covered by the large dark eyes, broadest above, meeting in the medio-dorsal line, and leaving

unoccupied a small down-bent frontal space. The last two segments of the peræon are the shortest. The small first side-plates are all but completely concealed by the large second pair, which are horizontally ovoid with almost vertically truncate hind margin; the third are much shorter, but nearly as deep, subquadrate; the fourth similar to the third, with more rounded outline, and faintly emarginate behind. These are much at variance with the corresponding parts as figured by the authors named in the generic discussion. The fifth and sixth side-plates are bilobed. The posterolateral angles of the third pleon segment are minutely dentate, with sinuous outline above the denticle. The fourth segment has a deep dorsal depression; the fifth is much shorter than the fourth or the sixth.

The first antennæ have a short stout peduncle, not longer than the long and broad first joint of the flagellum, which is thickly fringed with filaments and followed by ten slender joints, together scarcely as long as the accessory flagellum. This is scarcely shorter than the first joint of the primary, and is composed of one long laminar joint followed by a short linear one and a minute apical joint; in comparison with the primary flagellum it is much larger than that figured by Della Valle for G. nicæensis.

The second antennæ have the first three joints of the peduncle short, the fourth rather longer than any of them, but less than half as long as the slender fifth; the twenty-five jointed flagellum is scarcely as long as the peduncle.

The upper lip and the forward lobes of the lower lip are apically narrow, but below the apices they show considerable breadth when flattened out.

The mandibles in the rear part of the trunk afford a substantial base for the very large palp, but the front part of the trunk is feebly laminar, with truncate cutting edge and a microscopically tridentate accessory plate. Whether in the female the cutting edge becomes narrower remains to be seen. The second joint of the palp carries twenty-one long spines, and the rather shorter third joint twelve that are shorter, serrate, and more widely spaced. But these numbers are not constant, since in the smaller specimen examined they were respectively thirteen and seven.

The first maxillæ have been already described. It may be added that the outer plate is slightly contracted a little below the apex, and what I have suggested in the generic discussion as a possibility seems to be here clearly the case in regard to the palp. The smaller specimen shows a one-jointed palp surmounted by a single seta; the

larger specimen has a two-jointed palp carrying two setæ, one much longer than the other.

The second maxillæ appear to be of great tenuity, the plates apically rounded, the inner slightly the shorter, tipped with one curved setule, the outer carrying two such setules.

The maxillipeds have both the inner and outer plates of considerable breadth, though in position for forming the oral tube they seem very narrow. The first joint of the palp is relatively much longer than that which authors show in T. raschii. In our smaller specimen it is at least as long as the second joint, but in the larger specimen it is a little shorter than that joint, which is there the longest of the four, not the shortest, as in Della Valle's G. nicæensis. The third joint has a few spines and some fur-like armature on one margin; the fourth joint, subequal to it in length, is narrow and straight, having the apical part furnished with very fine outstanding hairs.

The first gnathopods scarcely, if at all, differ from those in the northern and Mediterranean forms. The enormous hand, as mounted for the microscope and figured in the plate, does not show the true extent at right angles to the spinose palm, because it refused to be fully flattened out. Possibly it was preparing for the curious torsion by which, in adults of this genus hitherto known, the great curved finger of the first gnathopods looks as if it were attached to the wrong end of the palm. The second gnathopods have the third joint much shorter than the fifth, in this respect strikingly differing from G. nicæensis, in which these proportions are reversed; the small finger is almost concealed among the spines of the sixth joint.

The second peræopods are somewhat shorter than the first, and have neither the second nor the fourth joint notably expanded. In the third and fourth pairs the second joint is broadly oval, larger in the fourth than the third pair. In the fifth pair the second joint is narrower, with the upper part forming a kind of neck; the fourth, fifth, and sixth joints are longer than in the preceding pairs, the fifth longer than the fourth and wider, with its lower front corner squared like the corresponding angle in the other peræopods; the sixth joint is longer than the fifth and slightly wider, forming a narrowly oval blade-like lamina, quite different from the sixth joint in the other peræopods and from that of the last pair in forms previously described. The finger is small. To this pair, as to the other peræopods and to the second gnathopods, there are attached large much-pleated branchial vesicles.