396 
THE VOYAGE OF H.M.S. CHALLENGER. 
of the Ascidiozooid which has been already described in the case of Leptoclinum moseleyi 
(see p. 273), where it has resulted in the abdomen having come to lie at right angles to 
the thorax in place of being in the same straight line with it. In other cases the 
Ascidiozooids have remained unflexed, but have become scattered irregularly in the test 
inclined at various angles to the surface of the colony. 
In the ancestral Diplosomidse the reproductive organs have remained in a more 
primitive condition than in the Didemnidse, and the vas deferens has become straightened. 
The testes are usually two in number, and are therefore in an intermediate state between 
the numerous spermatic vesicles of the ancestral Distomidse and the single large testis of 
the Didemnidse. The property of producing calcareous spicules in the test has become 
gradually lost in the Diplosomidse, Spicules are still found in the upper layer of the 
colony in Diplosomoides (see p. 309), but have completely disappeared in the genus 
Diplosoma. As a result the test has become softer and more transparent, and the 
system of canals and cavities in connection with the common cloacal apertures has 
become at the same time so increased as to greatly reduce the relative amount of test 
present in the colony (see p. 308). Diplosomoides is less modifled than Diplosoma 
(see flg. 13, p. 393), and is more nearly related to the Didemnidse. 
In order to trace the evolution of the remaining Compound Ascidians, it is necessary 
to return to the ancestral Ascidian allied to Ecteinascidia (B. in flg. 11, p. 388). This 
form was the common ancestor of the various groups of Simple Ascidians, and it is 
described in the first part of this Eeport (vol. vi., 1882, p. 285). It gave rise to an 
ancestral series which, after losing the power of reproducing by gemmation and under- 
going a certain amount of modification, led to dona and the other genera of the 
Ascidiidae.^ From the ancestral Ascidiidae an important branch leads to a great series 
of forms in which the body has become shortened antero-posteriorly by the alimentary 
canal being placed alongside the branchial sac instead of extending behind it, while the 
branchial sac has become more highly developed and has had its surface greatly increased 
by being thrown into a series of longitudinal folds. Before this last change took place, 
however, the branch probably divided (at J. in flg. 11, p. 388) into two ancestral series, 
one leading to the family BotryUidse and the other to the primitive Cynthiidse (see 
flg. 11, J.). 
In the long line of descent leading from the point J. to the fa mil y Botryllidse the lost 
property of reproducing by gemmation was regained, and as a result colonies were once 
more produced. The Ascidiozooids in these primitive Botryllidse were completely 
imbedded in a common test, and they became arranged in systems as in the case of the 
Polyclinidse, and finally in each system all the atrial apertures came to open into a 
* For the phylogeny of the Simple Ascidians see Part I. of this Eeport in vol. vi., 1882, Summary, p. 285. The 
subject will be discussed more in detail in Part III. of this Eeport, where the relationships between the various groups 
of the Tunicata will be fully considered. 
