into copulatory organs, the anterior pair of the seventh segment, are in the Colobognatha entirely unmodified. Hence we must suppose that either the legs or the function have migrated, in case we assume a common origin and attempt to homologize the copulatory legs in the two orders. The theory of migration, however, has no facts to support it, and would be equally fatal to the idea that affinity or the want of it can in the Diplopoda be inferred from the position of the copulatory legs.

The fact that the Colobognatha have eight precopulatory legs is not new, but up to this time the whole eight have been supposed to belong to the first six segments. Both Latzel and Pocock give the distribution of these legs as in the second column. In reality the arrangement is that of the third column.

My attention was first attracted to these facts while engaged in examining specimens of Siphonotus collected in Sierra Leone in December, 1893. The creatures were abundant in decaying banana stumps in Freetow, and I secured a large quantity. Instead of curling up as nearly all the representatives the present order are accustomed to do when placed in alcohol, my specimens remained conveniently straight and pliable so that they could be mounted in alcohol or balsam and studied to advantage. Of the arrangement of the legs as here stated there can be no doubt. The drawings are mostly camera tracings made from preparations in balsam. In order to make sure of the condition in Polyzonium, the genus studied by Latzel, I cut animals in two horizontally, brushed away the internal structures and mounted in balsam.

5 Mr. Pocock seems to have come to doubt this disposition, for he uses a “?” in front of his last statement on the subject.—Max Weber's Reise, p. 335.

Without such a preparation satisfactory observation is very difficult if not impossible in Polyzonium, for all the parts, especially the bases of the legs, are crowded together. I have examined in addition Andro, gnathus, Platydesmus, Pseudodesmus, Siphonorhinus and Siphonophorawithout finding any indications that the condition described is not present in all, though a final demonstration would in most cases not be easy without dissection.

Probably correlated with the comparatively slight degree of specialization which appears in the copulatory legs of the Colobognatha is the fact that in young males of Siphonotus the copulatory legs are severaljointed before maturity. Such a condition seems to be unknown in the other helminthomorphous groups.

In the previous discussion there has been no intention to imply that the orders Oniscomorpha and Limacomorpha are not valid; the contention is merely that the position of the modified legs does not of itself justify holding them as divisions of greater weight than other natural groups of Diplopoda, some of which have been designated by ordinal names. It is to be expected that future study may result in a natural arrangement of the groups now designated as orders, but until their affinities are demonstrated nothing is to be gained by attempting to retain under one ordinal name and description animals which may prove to be widely divergent in their development history. Thus it is by no means impossible that the Colobognatha are really a group farther removed from the other Helminthomorpha than are some of these latter from the Oniscomorpha. Many of the peculiar characters of the Oniscomorpha are evidently the result of their power to roll themselves into a sphere, and are not to be assigned great weight in estimating affinities.

Systematic Note.-The genus Siphonotus has not until very recently been reported since its establishment by Brandt in 1836. Within the last year or two Mr. Pocock has described species from St. Vincent (West Indies), Java and Celebes. To me it seems doubtful whether any of these species are congeneric with Brandt's type S. brasiliensis. Provisionally, however, the species of which drawings are presented may be described under Siphonotus, no doubt being possible that its affinities are here rather than with any other genus yet established.

Siphonotus africanus sp. n.

Body slender, the sides parallel to near the ends, or very slightly converging cephalad.

Head smooth and shining, sparsely hirsute distad.

Eyes of a single ocellus on each side, large and strongly pigmented. Antennæ sparsely hirsute, strongly crassate, scarcely clavate; joints increasing in length from the first to the sixth.

Hypostoma distinct, medianly deeply excavate.

Mentum distinct, on each side a large cardo (?). Other parts of gnathochilarium indistinct; the lines of the figures may represent internal structures only.

First segment longer than the others, which increase in length to the middle of the body and are scarcely shorter caudad; surface of all the segments smooth and shining.

Repugnatorial pores in a continuous series from the fifth segment, removed considerably from the margin of the scuta, except those of the fifth segment, which are also distinctly larger than the others.

Pleuræ large and entirely free, smooth and shining.

Pedigerous lamina distinct, free; sometimes the edges of the pediger. ous lamina and the pleura lie one upon the other.

Penultimate segment without legs, in contact or overlapping, but not closed below.

Anal valves narrow; preanal scale wanting.

Legs sparsely hirsute, all six-jointed, with a rudimentary second joint which would make seven.

Males with a conic process from the coxæ of the second legs. All the other legs except the first have a coxal aperture from which projects a transparent membrane or secretion.

Male genitalia, see figures.

Color of living animals, pinkish, pale to dull reddish-pink.

Length of mature individuals up to 15 mm., width .75 mm.; segments of adults 39-55. The males are distinctly smaller than the females. The young male specimen from which fig. 15 was drawn has 24 segments, all but two of which bear legs, as in the mature animals.-O. F. Cook.

EXPLANATION OF PLATE.

Siphonotus africanus.

Fig. 1. Head and first nine segments, ventral view, to show arrangement of legs.

Fig. 2. Head and first three segments, ventral view, the legs removed to show pedigerous laminae, spiracles, etc.

Fig. 3. Head and five segments, dorsal view.

Fig. 4. Segments 4-6, lateral view, to show the peculiarly enlarged pore of the fifth segment.