118 THE VOYAGE OF H.M.S. CHALLENGED 



As to the first point, it has been hitherto the current idea tliat in the Nemertea the 

 generative sacs, alternating with the intestinal cseca, are paired and more or less meta- 

 meric. This is no doubt the case in the very large majority, and relieves us from 

 the duty of further describing the position of the generative sacs in Etqjolia, the 

 Schizonemertea and most Hoplouemertea. In other Hoplonemertea, however, we find a 

 multiplicity of generative sacs in one transverse section (PL IX. fig. 4) which cannot 

 possibly be made to answer to the type just alluded to, and this irregularity reaches 

 its extreme expression in Amphijyorus onoseleyi. The specimens of this species are 

 literally full of sacs, which I was able to notice in their first stages, as well as in their 

 later develoj^ment and ripest stages. In each transverse section (PI. IX. fig. 4) a great 

 number of them may be seen; in horizontal sections (PL IX. fig. 7) the same irregular 

 multiplicity is met with. Externally I did not notice the openings, but it must be 

 remarked that only in very ripe specimens are these distinctly present. 



Now this aberrant distribution is not wholly limited to AmphiiMrus moseleyi. There 

 are even reasons for considering it as an arrangement which has been retained in this species, 

 but which was common in the more primitive ancestral species of both the Hoj)lonemertea 

 and Schizonemertea. At least I find it in a similar condition in Carinella anmdata, with 

 this difi'erence, (1) that it is here only in the dorsal half, above the intestine, that the gene- 

 rative products are found ; and (2) that the external openings are generally very distinctly 

 visible as numerous irregular whitish dots on the dorsal surface of the animal, in the dark 

 coloured space between the dorsal median white line and two successive transverse ones. 



Whether in Carinina similar conditions exist could not be verified, because the 

 preserved fragments contain no generative products. At any rate the fact is fully 

 established for one of the most primitive genera of Pala^onemertea, and this may justify 

 our insisting upon its having an archaic significance. 



AVith the exception of Amjihiponts moseleyi, Drepanoplwrus lanhesteri is perhaps 

 the only Hoplonemertean in which still another trace of it is preserved, at least if we 

 may consider the fact that in this species (PL IX. fig. 1) two ventral generative sacs 

 are present on each side, as a reduced phase of the phenomenon, which we have found 

 re2:)resented in Amphiporus moseleyi, instead of looking ujaon it as a secondary dupli- 

 cation of primarily simple generative caeca. For myself, I should feel inclined to take 

 the first view as rather the more probable. 



The other Hoplonemertea have not this peculiarity. The generative sacs are paired 

 and metameric, and if in ripe specimens two of them are cut in one section (PL IX. 

 fig. 3) it is generally an indication that the section was not perpendicular. There is then 

 also found an adequate distance between the external pores. 



This same figure clearly illustrates the fact that the generative products (in this in- 

 stance ova) may attain a considerable development and closely approach the ripe condition 

 before the generative sac itself communicates by a pore with the exterior, as indicated 



