250 THE STRUCTURE AND AFFINITIES OF HERDMANIA CLAVIFORMIS. 



are produced on the stem of the T, in connection with which the ova ready for dis- 

 charge are found. 



In the smallest ovaries that I have observed the evagination of the disk has 

 not taken place, the disk, being here in the form of a plain pad, sections in all por- 

 tions of which are similar to that shown in Figure 19. There is, in other words, no 

 intimation of bilaterality in the ovary in its immature state in adult zooids. I have not 

 followed its development in the embryo. 



The ovary, then, resembles that of Fragaroides (Maurice, '88) much more closely 

 than it does that of Clavelina. Van Beneden et Julin ('86) have shown that in all 

 its stages in the blastozooids the ovarian epithelium of the latter genus is more or 

 less distinctly bipartite, and that the two parts are apparently related to the bilater- 

 ality of the zooid. 



The oviduct leads off from the anterior side of the ovary and holds a direct course 

 up to the thorax. It passes to the right of the intestinal loop. Its position through- 

 out most of its length is toward the ventral side of the abdomen a little to the 

 right of the oesophagus. During the breeding period (which appears to be limited 

 to the summer months) the anterior two-thirds or three-fourths of the duct become 

 greatly enlarged in diameter and serve as a uterus in which the embryos are developed 

 to a late stage in larval life (PL XVIII, Fig. 2, ut.). This uterus is capable of con- 

 taining eighteen or twenty embryos at a time, which are arranged in a single series 

 one behind the other, the oldest farthest anterior, or toward the atrial end, the 

 youngest nearest the ovary. The conversion of the oviduct into a uterus or brood- 

 pouch in Herdmania is wholly unique among the Ascidiacea, as indeed it is among 

 all the Tunicata unless the embryonic chamber of Salpa be regarded as a part of the 

 oviduct. 



One may justly raise the question as to whether the condition here is essentially 

 different, morphologically, from that presented by various compound ascidians, 

 particularly of the family Distomidse, where the embryos develop in the atrium or 

 an incubatory pouch produced from the atrial wall. It may appear at first sight 

 that a gradual transition from the oviducal wall to that of the atrial wall, and then 

 as a consequence, of a similar transition of oviducal cavity to that of the atrial cavity 

 might take place, so that in reality the difference between the brood-chamber of Herd- 

 mania and that of Colella or Distaplia, for example, would not be important. As 

 a matter of fact, however, no such transition from oviduct to atrium occurs in Herd- 

 mania. On the contrary I have grave doubts as to whether the embryos ever reach 

 the atrium. Certain it is that the oviduct in all the many gravid zooids examined 

 by me narrows down rapidly, as it approaches the atrium, to a most insignificant 



