blood vessels, and the worms themselves are slender and delicate, being indeed often aquatic in habit. The spermiducal glands, which are as in the former sub-families independent of the sperm ducts though sometimes opening in common with them into a short pocket-like ingrowth of the skin, are reduced in their minute structure and much simpler than in the other types.
The genus Kerria is the least reduced perhaps. It has the male pores on segment XVIII with a pair of spermiducal glands on the segments preceding and following this in the typical Acanthodriline fashion. There are two pairs of spermathecae in VIII and IX, but the spermaries are reduced to one pair in X. The gizzard is present. Ocnerodrilus is a little further reduced from this last. The gizzard has gone; there is but one pair of spermiducal glands (as a rule) opening in common with the extremity of the male duct on to segment XVII; the spermathecae also are reduced to one pair, but there are two pairs of spermaries. The African Nannodrilus is more robustly built. There are two gizzards, the male duct opens into a muscular pouch, into which also open one of the two or three pairs of spermiducal glands. There are two or three other genera and sub-genera not showing any great divergencies from the range of structure indicated in briefly defining those enumerated above.
Finally, we have the largest of all the sub-families of this family, viz. that of the Megascolecinae. These worms are mainly tropical in range and also mainly found in the Indo-Australian portion of the world. In them the sperm ducts open in common with the usually single pair of spermiducal glands and prevalently upon the XVIIIth segment. The glands moreover have not always, though they often have, the tubular form shown in all the sub-families hitherto considered. In many forms they are branched and lobate glands, and if there are two pairs one may be of one type and the other of the second and derived type, as for instance in Megascolex ceylonicus. Furthermore, it is much commoner among the genera of this sub-family for the setae to become numerous and to spread right round the segment; this condition is seen in the genera Pheretima, Megascolex, Diporochaeta, Perionyx, Plionogaster. The spermathecae also are commonly more than the typical two pairs of the forms already considered, and in certain species (for example Pheretima hexatheca) there are as many as six pairs of those organs which are moreover – and in this they resemble the majority of species of the last sub-families – nearly always furnished with a diverticulum or diverticula. The nephridia are either paired or numerous and these various characters allow of the sub-family being split up into sixteen genera or thereabouts.
As an example of another type of organisation and as contrasting with Notiodrilus we may now briefly describe the structure of the genus Pontoscolex (better known as Urochaeta), a member of the American and African family Geoscolecidae.
The worm is some four inches long and composed of as many as 212 segments. Each of these except the first has eight setae which for the first few segments of their occurrence are disposed in four pairs in the usual fashion. Behind this point however the setae cease to present this symmetrical arrangement and are irregularly disposed so that a given seta is not in the same line with the corresponding seta of the segments in front or behind. There is thus no region of the body which has not a seta implanted upon it; and the effect is therefore comparable to the condition obtaining in those worms, such as Pheretima, where circles of numerous setae are met with. There are however only eight in a given segment. The clitellum extends from segment XV to XXII or XXIII and is developed only dorsally and laterally. It has setae like the rest of the body; but those upon the clitellum are longer and more markedly ornamented than are those of the body generally. The latter are not only sculptured with fine ridges but are bifid at their free extremity. The prostomium is often apparently completely absent. It is however really present but is retractile. As to the pores which are visible externally the dorsal pores are completely absent. The pores of the nephridia lie in front of the dorsal pair of setae or in a line corresponding to the position of those setae where the arrangement has become irregular. The spermathecal pores are three pairs and are in the very front of segments VII, VIII, IX. The male pores, very inconspicuous, lie on the ventral side of segment XXI just within the clitellum. The oviducal pores are on segment XIV.
As to internal anatomy the general plan of the segmentation shows no great differences. Certain septa only show a difference, i. e. those separating segments VI-XI which are specially thickened. In the alimentary canal a gizzard in segment VI is to be noted and three pairs of calciferous glands in segments VII-IX. The nephridia are paired structures and commence early. The first two or three segments are occupied by a pair of large glands opening into the buccal cavity which are apparently a slightly modified pair of nephridia and are generally termed 'peptonephridia' since they are associated, as it would appear, with the function of alimentation and are not purely excretory organs. There is but a single pair of spermaries in segment XI, and of ovaries in segment XIII. The sperm ducts open on to the exterior in the position already mentioned and they are not associated at their pore with any glands comparable to spermiducal glands. A pair of sperm sacs depend from segment XI and traverse a considerable number of segments, being thus long and tongue-shaped instead of short and limited to one segment. The spermathecae are three pairs of elongated sacs in segments VII-IX, without any diverticula at all.
It will appear therefore that many and considerable differences divide Pontoscolex from Notiodrilus and indeed from all of the Megascolecidae whose structure has been touched upon in the foregoing pages. The most important of these are the ornamented setae and their arrangement and the modification of the setae upon the clitellum: the absence of diverticula to the spermathecae: the absence of terminal glands associated with the male ducts. Although taken in their entirety these characters are distinctive of the American Geoscolecidae (sub-family Geoscolecinae), there is no one of them which is not to be found in some Megascolecid. Thus the subgenus Ilyogenia (of Ocnerodrilus) has sometimes no spermiducal glands: the genus Perionyx has spermathecae without diverticula in some species, and other genera of Megascolecinae are in a like condition. The setae of Dichogaster are sometimes ornamented, while in Pheretima houlleti the clitellar setae are different from those upon the other segments.
We can in fact only define the family Geoscolecidae by an assemblage of characters which are mainly these: dorsal pores absent, only a few in the neck region being occasionally present; setae generally ornamented, those of the clitellum being larger and more marked than the others; spermathecae without diverticula; often instead of a pair of those pouches in the segment a large number of very small sacs, as in Microchaeta, Kynotus. Sperm ducts without terminal glandular or muscular sac, except in a few cases; setae always eight in a segment except in the genus Periscolex which has acquired the 'perichaetous' condition. The range of variation shown in the anatomy of the Geoscolecidae will be best taken in connection with the several sub-families into which it has been subdivided. In the first of these, the Geoscolecinae, no great differences divide the genera from that selected as the type, viz. Pontoscolex. The long sperm sacs attain to an extraordinary length in Trichochaeta (or Hesperoscolex) where the single pair extends through no less than 109 segments. Though as a general rule the sperm ducts open directly on to the exterior they do so through the intermediary of a large pouch in Glossoscolex (= Titanus). In Onychochaeta the setae on the last segments of the body are very much enlarged and thus form a more effective means of holding on to the soil than is possessed by other species.
The sub-family Hormogastrinae which contains but a single genus Hormogaster is remarkable for being limited in range to the Mediterranean coasts. The genus is mainly distinguished by possessing three gizzards; otherwise it is not very different from the sub-family just described. The African and Madagascar forms are associated (together with a few Asiatic forms) into a third sub-family Microchaetinae. These worms frequently possess a considerable number of very small spermathecae in segments XII, XIII or thereabouts instead of the usual paired arrangement. They have too very often glands connected with the enlarged setae already mentioned which are however (in the genus Kynotus at any rate) usually in front of the clitellum. The latter organ, contrary to what we find among the Geoscolecinae, is often behind the point of orifice of the male pores. This is so with Kynotus.
The
