DEVELOPMENT OF CEPHALODISCUS i95 



the testis is continued into a short narrow tube, the vas deferens, which opens externally. 

 I was not able to discover any trace of pigmentation in the vas deferens or in the ecto- 

 derm surrounding its external opening. In C. sibogae the two testes are large and fill the 

 greater part of the trunk cavities. Their structure is not clear from the account of 

 Harmer (1905), probably because the zooids were imperfectly preserved, but it would 

 appear from the description of his figures that the lining epithelium is comparatively 

 thin, and that it surrounds a large central cavity which is filled with ripe spermatozoa. 



In the female zooids the ovaries are conspicuous organs and in a mature specimen of 

 C. fumosus they appear like swellings on either side of the anterior part of the trunk 

 (John, 1 93 1, pi. XXXV, fig. 3). Each ovary is an ovoid or flask-shaped body which is filled 

 with a varying number of immature eggs. In an adult individual one or two eggs, at the 

 end furthest from the external opening, are large and well developed and fill the greater 

 part of the ovary. In C. gilchristi each of these large eggs is lodged in a follicle of small 

 cells, while Schepotieff (1909) has described the large ova of certain species as lying be- 

 tween a central lumen, clothed with epithelium, and the wall of the oviduct, in a position 

 where they are surrounded by a large quantity of blood. The indication of a central 

 lumen in the ovary is also suggested by Harmer (1905), but no trace of any lumen has 

 been observed in any of the species collected by the Discovery Expedition. 



The oviduct is a thick-walled pigmented tubular prolongation of the anterior end of 

 the ovary, and its external opening corresponds in position to the external opening of the 

 male duct. In some species, such as C. levinseni and C. dodecalophiis, the external 

 opening of the oviduct is situated at the dorsal end of an ectodermal recess. 



Gilchrist suggests that the ova originate from the wall of the oviduct, but this view is 

 not supported by any convincing evidence. Very little is known about the origin of the 

 oviduct and its homology. From a comparative study of the developing buds and young 

 zooids it would appear that it is an ectodermal invagination, to the proximal end of 

 which is attached the primitive germ cell, surrounded by a thin mesodermal epithelium. 

 The oviduct is deeply pigmented like the red line on the ventral wall of the proboscis. 

 In addition to this red pigment the ectoderm of the trunk carries a black or brown pig- 

 ment in C. nigrescens and C. fumosus. In the developing buds and larvae, which will be 

 described later, pigment granules are confined only to the ectoderm. If, therefore, the 

 occurrence of the red pigment in the wall of the oviduct is taken as evidence for deciding 

 its homology, there can be no doubt of its ectodermal origin. The proximal end of the 

 pigmented tube is continued into a very short unpigmented part, which in turn opens 

 into the ovary. This latter part is more clearly seen in the figures given by Harmer 

 (1905, pi. xii, figs. 153, 154) than in my own sections. The continuation of this tube with 

 the lining epithelium of the ovary suggests that it is mesodermal in origin. The oviduct 

 is therefore really formed of two parts : a distal pigmented tube, which arises as an ecto- 

 dermal invagination, and a short proximal part which is probably mesodermal in origin. 

 According to this view the ectodermal recess in C. levinseni and C. dodecalophiis is the 

 anterior part of the invagination, which serves as a receptacle into which the egg is ex- 

 truded before it is finally liberated from the body of the parent. 



