large number of forms the poriferous segments are wholly or partly red or yellow, while the remainder of the body is nearly black, giving a most striking appearance. Prepodesmus includes several such forms, all with the anterior corner of the second segment greatly produced and embracing the first segment. Tylodesmus has the corner rounded and not produced. Cheirodesmus is similar to the last in general shape, but is more slender and with the male genitalia resembling in shape a gloved hand. Anisodesmus is peculiar in that the fourth segment is distinctly, though slightly, narrower than the third or fifth. The species are uniform dark red in color and the type is closely allied to Polydesmus erythropus Lucas. Isodesmus is evidently related, but with the fourth segment not narrowed, and remarkable in that the pores are not borne on a distinct callus as in the other genera of the group. The copulatory legs are also very peculiar being deeply divided into several lacinia. In all these genera the dorsal surface is finely and evenly granular, though differing somewhat in other respects. The family is probably distributed along the entire West Coast, and I have seen two forms from South Africa, one of which, Lipodesmus, is in the Berlin Museum.

Family OXYDESMIDÆ.

The Liberian forms which belong to this family in the more limited sense are all referable to the genus Oxydesmus, and belong to three species, O. grayii Newport, O. medius and O. liber, both new. The first is a very striking form, black in color with a narrow median stripe of bright vermillion. The other species are also black, O. liber with bright yellow submarginal ridges.

Family POLYDESMIDÆ.

Of Liberian species referable to this family in its stricter sense there seem to be but two; small pinkish-red forms, similar in general appearance to some species of Brachy desmus. The dorsal elevated areas are each supplied with a clavate hair. The antennæ are strongly clavate, though rather slender, and the second pair of legs is crassate in the

The African forms having the apex of the last segment broad, the femora spined, and the carinæ with a submarginal ridge, constitute the family Oxydesmidæ. There are five genera now known, two confined to the east side of the continent, three to the west. Of the east coast forms, Orodesmus includes those with strongly tuberculate segments, Mimodesmus those with the body slender and the dorsum nearly smooth. Of the west coast genera Oxydesmus has three rows of dorsal tubercles and surrounding areas; Scytodesmus has five or six rows, while Plagiodesmus resembles Oxydesmus, but has the submarginal ridges very broad and oblique, and the copulatory legs large and exposed.

male. For this genus the name Bactrodesmus is proposed; it will probably be found to be next related to the form described from Ceylon by Humbert as Polydesmus cognatus, but which is generically different from the European P. complanatus, and may be denominated Nasodesmus.

Family PTERODESMIDE, New.

This family is proposed for Polydesmus gabonicus Lucas and its African relatives, by more recent writers referred to Cryptodesmus. I have examined the type of Cryptodesmus olfersii in the Berlin Museum. The diversities seem to be of family importance. The African forms are very curious, the development of the lateral carinæ being carried to its greatest extent. They are very much flattened, elliptical in outline, and only four or five times as long as broad. They never coil into a spiral, even when placed in alcohol. At least five genera are

found in Liberia.

All the African forms yet known to me have repugnatorial pores, and we may expect to find these in the others, notwithstanding the statements of several writers to the contrary. The location of the pores is, however, very unusual. They are far remote from the lateral margin, in the anterior part of the carinæ, in some cases so far ahead as to be concealed by the posterior margin of the preceeding segment. An even more remarkable condition obtains in Pterodesmus brownellii, the type of the genus and family. The fifth segment has no pore! The Liberian forms are further peculiar in that all are more or less pruinose. Pterodesmus is the largest of the Liberian genera. It is pure white when young, but mature individuals are usually dusted with earth which adheres to the pruinosity and gives them the advantage of protective coloration. Gypsodesmus, on the other hand, is pure white, even when mature. Lampodesmus is partly pruinose and appears to be black and white when alive, though it is brown in alcohol. It is structurally peculiar in that the sternum of the sixth segment bears two hollow processes fringed along their apical edges with long hairs. These may be of use as a protection to the copulatory legs. Compsodesmus is the broadest of the Liberian forms. When alive it is one of the most varied and brilliant of Diplopoda. A large median area of the dorsal surface of each segment is dark brown, while the space between it and the posterior margin on each side is nearly white or bright yellow. Carinæ tinted with bright orange or pink, or both. Below, except near the edges of the carinæ, the body is covered with a pure white bloom or chalky powder. Last segment nearly white. Motions very sluggish.

From the German colony of Togo comes a genus evidently allied to the last, but distinct by reason of the more slender body and narrower carinæ, which are also scarcely produced at the posterior corners. From Lampodesmus it is distinct in the absence of the process from the sternum of the sixth segment, and in the form of the copulatory legs. A small horn-brown or yellowish creature with remarkably agile movements it is proposed to name Choridesmus citus. The first segment is pure white, pruinose, and abruptly different in color from the remainder of the body. The pores are large, and are located in the middle of the carinæ, remote from the margins. The quick, jerky movements remind one strongly of Polyxenus.

Family STRONGYLOSOMATIDÆ.

Of this group there are two genera in Liberia, both new, though probably not confined to the West Coast. Scolodesmus grallator represents the usual Strongylosoma type, with long legs and antennæ. It is dark wine-color, nearly black. Habrodesmus latus is a rare species apparently confined to the darkest forests. It is exceedingly quick and agile, very graceful in form and brilliantly colored. The legs are orange and pink, and the segments have the posterior margin yellow, shading through orange and brown to black on the remainder of the segment.

In gardens at Monrovia Orthomorpha vicaria (Karsch) is not uncommon; it is probably not indigenous.

Family STYLODESMIDE.

The type of this family is a bizzare creature named Stylodesmus horridus. The generic name alludes to the fact that the pores are borne on long stalks placed near the lateral margins of the broad, decurved carinæ. The pore-formula is the usual one, 5, 7, 9, 10, 12, 13, 15-19. The whole dorsal surface of the animal is setose and coal-black. There is almost always an incrustation of dirt which furnishes a completely protective coloration. The head is completely concealed under the flabelliform, anteriorly lobed, first segment, and the last segment is reduced, included in, and concealed by the penultimate. The most striking feature is that each of the segments except the last bears dorsally a pair of long slender processes. Those of the anterior and posterior segments are close together and show a tendency to unite at the base. These processes are also rough and setose, and almost always so incrusted with dirt as to appear several times their actual size. If segments of Stylodesmus had been found in fossil condition they would probably have been looked upon as allied to some of the

Archipolypoda, so much greater is the general resemblance to the fossils than to previously known extant genera. Yet there are in Liberia at least three other genera which have evident affinities with Stylodesmus. In all the pores are located on special processes or tubercles, and the first segment is enlarged and scalloped in front. Udodesmus telluster differs from Stylodesmus in being much more slender and without the long dorsal processes. These are replaced by two longitudinal crests of two or three large tubercles. The body is very rough, setose, and incrusted with earth. Hercodesmus aureus is a beautiful little species, more slender than Udodesmus, and usually without a covering of earth. In Stiodesmus the dorsal ridges of tubercles are not much more prominent than the numerous large, rounded tubercles with which the whole surface is beset. The result is a creature which on first view might be supposed to have affinities with Scytonotus.

Besides these, the present family will contain four East Indian genera, Pyrgodesmus and Lophodesmus, described by Pocock, and two new ones from Java. In the Canary Islands is a beautiful and evidently allied form inhabiting the nests of ants, and called Cynedesmus, on account of the form of the first segment. The Stylodesmidae do not coil up into a close spiral; they usually remain nearly straight, even when in alcohol. Though there is no close resemblance in form or structure between the Stylodesmidae and Campodesmidæ, yet both are strikingly different from other Diplopoda. That two groups of such remarkable creatures should inhabit the same locality seemed at first very strange, but as the various new and equally interesting forms continued to be found it was soon apparent that we were really in the presence of a new fauna.

That the new families are not all confined to Africa is shown by recent papers, notably those of Mr. Pocock. As yet, however, the Ammodesmida and Campodesmida are known only from African representatives. Of the larger, long known forms the Oxydesmidae and Prepodesmidæ, appear to be confined to Africa. In East Africa is another family of several genera, none of which has yet been reported from the West Coast. Indeed, speaking with regard only to families and genera, there are four very distinct diplopod fauna in the African continent, the northern, southern, eastern and western parts having little in common. The species are, of course, even more local. I have examined the collections of the Berlin Museum and the British Museum, as well as the literature of the subject, and with the exception of Oxydesmus grayii and Orthomorpha vicaria, collected at Sierra Leone, know of no Liberian diplopod from any other part of the West Coast.

We are thus assured of an African fauna of surpassing richness, not a tithe of which has yet been revealed.-O. F. Cook.

Entomological News.-Prof. Clarence M. Weed of the New Hampshire College spent several weeks in December and January, studying the Bermuda Islands. Many species not before recorded from there were collected.

EMBRYOLOGY.'

The Sense Plates, the Germ of the Foot, and the Shell or Mantle Region in the Stylommatophora.'-To our knowledge of these subjects, Dr. Ferdinand Schmidt contributes the results of his numerous observations upon the embryos of Succinea, Limax and Clausilia. Concerning the sensory plates he shows that immediately behind the budlike rudiments of the future egg-bearing and the simple tactile tentacles, in Limax where the development is most easily followed, there arises a third pair of buds like the first two pairs in all respects except in size. From these buds arises the so called oral lobes, subtentacular lobes, or labial tentacles. They have no relation to the velum whatever, since they arise in a pre-velar region. This is completely at varience with the observations of Jayeux-Laffuie on Onchidium and those of Ray Lankester on Limnæus in which the subtentacular lobes are asserted to arise from the velum or a rudiment of the same. Should further studies upon these forms substantiate the assertion, we would then have two groups of oral lobes, one in which they arise from the velum and to be homologized with the oral lobes of the lammellibranchiata, where they undoubtedly have such an origin, and the other in which they arise from the sensory plates and are homologous with the tentacles.

In his account of the development of the foot in Succinea he supports the conclusion long since put forth by Lankester, namely, that the typical form of the blastopore is an elongated cleft on the ventral side of the embryo, from which arises in some cases the mouth, in others the anus, according as the cleft persists anteriorly or posteriorly. This form of a blastopore is certainly important, considering his con1 Edited by E. A. Andrews, Baltimore, Md., to whom abstracts reviews and preliminary notes may be sent.

2 Beitrage zur Kenntniss der Entwicklungsgeschichte der Stylomatophoren, with 9 text figs. Zool. Jahrbücher, VIII, 318.