1914-] S.Kemp: Oiiychophora. 481 



vesicle. The vesicles themselves are full of spermatozoa in various 

 stages of development and among them considerable numbers of 

 sperm-mother-cells are to be found. 



The structure of the vas deferens is, in one respect, peculiar. 

 At the point where it leaves the seminal vesicle it has thick walls, 

 the small and sharply-defined lumen being surrounded by long 

 columnar cells with nuclei at their bases (pi. xxxvi, fig. A). It 

 continues backwards in this character for a short distance and 

 then rapidly changes to a thin- walled tube with a vastly larger 

 lumen surrounded b}^ flattened cells (pi xxxvi, fig. B). Further 

 backwards still it returns to its original form and again becomes 

 a thick-walled duct, while finahy, before reaching the posterior 

 limit of its length, it reverts once more to a thin- walled condition, 

 wholly similar in structure to that found further forwards, and in 

 this condition passes upwards and eventually joins its fellov; 

 from the other vesicle. 



That two types of structure are to be found in the vas defer- 

 ens is well known and sections through the thin- and thick-walled 

 portions of the Abor species bear a close resemblance to figures 

 gi\-en by Gaffron and Evans. But that the duct should revert to 

 its previous condition after once having changed to the thin- 

 walled type has not, X believe, been noticed previously in an}- 

 species of Peripatus, and it is extremely difficult to suggest any 

 reasons for such a modification. 



The thick-walled portions of the duct are distinguished bv 

 oblique hatching in the central figure on plate xxxvi. In prepar- 

 ing this figure I was obliged to have recourse to reconstruction 

 from serial sections, a task which, owing to the great length of 

 the system (2500 sections were made) proved somewhat tedious: 

 the vasa deterentia form such an intricately convoluted mass that 

 it was found impossible to unravel them in dissection. 



When the curious change in the character of the ducts was 

 first noticed, it was natural to conclude that some error had been 

 made in tracing their course; but a repetition of the process led 

 to the same result and precisely similar phenomena were ob- 

 served in tracing the vas deferens on the other side of the animal. 

 It may also be pointed out that, in sections passing through the 

 interspaces between the fifteenth and sixteenth pairs of legs, the 

 v/alls of all coils of both ducts seen in cross-section are of the thin 

 type, whereas the thick type is to be found both in front of and 

 behind this region. This, in itself, is sufficient to prove that the 

 thick-walled part of each duct at the posterior end of the animal 

 must be separated by a thin-walled interval from the portion of 

 the duct of similar structure that opens from the seminal vesicle. 



The common duct, at its point of origin, is composed of an 

 inner layer of cells of considerable depth provided with numerous 

 nuclei and an outer muscular layer of no great thickness (pi. xxxvi, 

 fig. C). The inner layer, which is doubtless glandular in function 

 and is concerned with the formation of the spermatophore, in- 

 creases gradually in thickness until the anterior limit of the loop of 



