oesophagus to intestine is slight, being marked by a slightly broader lumen, a thicker gut wall and longer cilia. In fixed material the distinction is clearer than in the living worm (Fig. 16).

The vascular system is illustrated in Fig. 10. The ventral loop unites in III and in all the anterior segments is joined to a single dorsal vessel by a pair of commissures in each segment. This arrangement closely resembles that of Tubifex (Dixon, 1915) and is simpler than that described by Beddard (1891) for Phreodrilus subterraneus which has two dorsal vessels. The nervous system was not specially studied.

The nephridia are tubificid in structure (cf. Dixon, 1915), much coiled and with a small diamond-headed nephrostome. The nephridiopore opens ventrally in line with the ventral setae near the anterior margin of the segment. The nephridia occur either paired or single. In the latter case they occur on alternate sides. Typically they occur paired in × and XIII, although occasionally alternate nephridia occur in these two segments. They are present in a number of subsequent segments either paired or single.

Besides the nephridia and reproductive organs, the coelom, especially of the anterior segments, contains gland cells which are often aggregated into septal glands in V-VII but may also lie freely in the coelom. The setal follicles of the stout ventral setae and their muscles are relatively large and project well into the coelom. Those of the dorsal setae are much smaller.

Chloragogen commences in VII.

Only a few worms have been observed with reproductive organs, and these show considerable variation not only from the typical phreodrilid arrangement but even between individuals. This is doubtless associated with the unusual form of asexual reproduction characteristic of the species. A similar variation is recorded for Lumbriculus variagatus which also reproduces by multiple fragmentation (Stephenson, 1930, p. 789).

Female organs have not been observed. The testes are found in IX, and the sperm masses lie freely in the coelom (Fig. 8). There are no seminal vesicles. The sperm duct and efferent organ, or penis, occur in X. They are paired. The ciliated funnel of the sperm duct opens in IX close to the site of the nephrostome. The sperm gland, or prostate, is an elongate body joining the sperm duct just posterior to the IX/X septum. The penis (Fig. 9) is a simple evertible organ lined with large cells. It opens ventrally, and is strongly muscular. A fine penial bristle, about 80 mu in length, may occur in XI (Fig. 12, 16). A pair of large spermathecae occur in XI. A second pair may occur in XII. The spermathecae open ventrally, in line with the ventral openings of the penis. They are elongate bodies and may be coiled on themselves within their segments or may protrude backwards into the succeeding segments.

Externally there are few signs of sexual maturity. The clitellum is not distinguishable in living worms but in fixed material there is an obvious thickening of the body wall in X, XI and XII, and to a lesser extent in IX and XIII (Fig. 16). The dorsal setae are typically absent from X–XIII and the ventral setae from X–XI, the penial bristles occurring in XI, posterior to the spermathecal openings.

This arrangement of the reproductive organs agrees with the phreodrilid pattern generally, and in particular in the posterior position of the spermathecae. It differs chiefly in the anterior displacement of the principal organs—viz., the testes, efferent organ and spermathecae, which are usually in XI, XII and XIII respectively. Beddard, however, recorded an individual of Phreodrilus albus in which the organs were displaced one segment forwards and in P. zeylanicus the efferent organ opens in XIII and the spermathecal pores in XIV, i.e., they are displaced one segment further back (Stephenson, 1913). The exact location of these organs is then of less significance than the posterior relation of the spermathecae to the other organs. The present species is unusual in having occasionally two pairs of spermathecae.

Despite the presence of reproductive organs in some worms I have observed no evidence of sexual reproduction. Reproduction appears to take place chiefly asexually by a process of fragmentation and subsequent regeneration. Typically the worm secretes a mucous sheath within which the process of fragmentation takes place. It may break up into ten or more fragments, the smallest consisting of only three or four segments (Fig. 11). The most anterior and posterior segments lose their setae and are thickened terminally, the alimentary canal being, for a time, a closed tube. Only the first fragment, with the original head, and the last fragment, with the original tail, are exceptions to this, and in these two cases the activity of the cilia of the alimentary canal is inhibited. Within a week to a fortnight the worm develops a head and a tail and very rapidly regenerate a large number of segments, so that by the end of a month the separate parts are identical with the original worm. Long

before regeneration is complete, however, the worms will move actively through the culture.

Asexual reproduction is commonly correlated with remarkable powers of regeneration. In some experiments with this worm it was found that amputated fragments would regenerate, following the closure of the original wound, in very much the same way as the naturally occurring fragments.

This asexual method of reproduction and the great capacity for regeneration of damaged fragments of this worm closely parallels Lumbriculus variegatus, another aquatic worm in which sexually mature individuals are very rare and which reproduces asexually by a similar process (Stephenson, 1930, p. 538). Like the present species L. variegatus belongs to a family, the Lumbriculidae, in which sexual reproduction is the rule and asexual reproduction by budding does not occur. The only other worms for which fragmentation and regeneration have been recorded all belong to the Naididae, a group in which asexual reproduction, typically by budding (i.e., regeneration prior to separation) is the rule.

Schizodrilus major n.gen., n.sp. Figs. 17–22.

This is a small, white, very soft bodied worm, generally tubificid in appearance, with a delicate transparent body wall which is highly contractile but not strongly annulated. There are up to 120 segments in a large worm which may measure almost 20 mm when fully extended. They are about 250–350 mu broad. The prostomium is rounded and flexible and about 150–200 mu broad.

Ventral setae commence in II and dorsal setae in III. The ventral setae consist of two per bundle; both are sigmoid setae, but one is single pointed, the other bifurcate (Figs. 19, 20). The single pointed seta is about 100–170 mu in length with a well defined nodulus between a-half and one-third the length from the distal extremity. The bifurcate seta is the same length with a nodulus between a quarter and a fifth the length from the distal extremity. The proximal tooth is slightly larger than the distal tooth.

The dorsal setae consist of a single (or rarely two) needle setae per bundle, and sometimes one or two pectinate setae. The needle setae and the pectinate setae resemble those of the previous species but differ in size. The needle setae (Fig. 22) are 200–300 mu in length and the length of the shaft about 60–120 mu and its breadth about 10 mu. The pectinate setae (Fig. 21) are 55–75 mu in length and almost 5 mu in breadth. In the anterior segments and in regenerating segments the setae are shorter.

The alimentary canal consists of a pharynx in II–IV, a winding oesophagus in V–IX which broadens into the intestine in X. The intestine which nearly fills the coelom is partially constricted by the septa. Chloragogen commences in V.

The vascular system presents some unusual features (Fig. 18). The dorsal vessel, closely apposed to the gut, shows a remarkable double loop in III. From the anterior portion of the second loop the first left commissural vessel arises, while the right commissural vessel arises from the posterior part of the loop just anterior to the septum. The ventral vessel is double in the first five segments, the two parts uniting just before the junction of the commissural vessels in V.

Nephridia generally occur paired in IX, and are either paired or single (and then alternate) in subsequent segments. The nephridia consist of closely apposed coils with double lumina. There is a small nephrostome with a slightly swollen base leading into the succeeding segment. Here, after several coils, the tube leads into a swollen body or ampulla and then to the ventral nephridiopore.

Septal glands occur in V–VIII. Coelomocytes up to 20 mu in diameter occur throughout the coelom.

No reproductive organs have been observed in any of the worms even when the same worm has been observed regularly over a period of ten months. No instance of multiple fragmentation has been observed, but a number of small fragments—of about 12 segments—have been observed. These fragments regenerate a head and ultimately increase in size, by the addition of further segments. Further, like the previous worm, there are very great powers of regeneration of mutilated fragments. I believe, therefore, that this worm reproduces by a process of simple fragmentation and regeneration. In this case it appears that only a single posterior fragment becomes detached, and consequently it is very difficult to observe the actual process of fission.

To support this interpretation a number of experiments were carried out in which a worm was cut into two and in one case, into three parts. In every instance all the parts regenerated quite satisfactorily. Briefly the process falls into the following stages: first, the closure of the wound; secondly, the regeneration of a small pros-

Figs. 17—22.—Schizodrilus major n.gen., n.sp. Fig. 17.—Semi-diagrammatic drawing. Fig. 18.—Vascular system. Fig. 19, 20—Ventral setae. Fig. 21.—Pectinate seta. Fig. 22.—Needle seta.

tomium and peristomium; thirdly, the intercalation of new segments immediately posterior to the new head; fourthly, the reorganisation of the body to a reduced size and the addition of further segments posteriorly. The rate of regeneration, as with the previous species, is rapid. The posterior half of a worm cut into two parts shows regeneration of a “stub” prostomium in seven days, and regeneration of the whole head within a fortnight, although the setae take longer to develop. The severed tail of the anterior half regenerates a new anus just as quickly. When the worm was cut into three parts the centre section consisted of only 12 segments. In less than a week a head was developing on both the posterior parts, and in less than three weeks the middle section had developed a head, six new segments immediately behind the peristomium and sixteen new posterior segments. In other words it had trebled the