PEROPHORA ANNECTENS. 49 



aiich bei den socialen und zusammengesetzten Ascidien 

 constatiren." Salensky's paper having appeared about 

 the time, or only shortly before Kowalevsk3^s observa- 

 tions were being made, the latter seems not to have been 

 aware of what the former had said on this point. 



As the testicular mass in these fully compounded col- 

 onies of Perophora is presumably growing constantly, it 

 seemed to me that they offered a good opportunity for 

 testing the assumption of Kowalevsky, quoted above, that 

 in the social and compound ascidians, also, the test cells 

 have a mesodermal origin. I have accordingly studied 

 the point with considerable care, with what results the 

 sequel will show. 



The cells of the test of the growing colony must be 

 derived from one or more of three sources : First, they 

 may be the direct descendants of the original cells of the 

 test of the larva from which the colony has been produced. 

 Second, they may arise by division from the growing por- 

 tions of the epithelial linings of the stolonic vessels, or, 

 what is the same thing, the external epithelium of the 

 body of the zooid. This epithelium is ectodermal, as we 

 know from the method of development of the stolon, as 

 first shown by Kowalevsky ('74). Third, they may arise 

 from the mesenchymatous cells contained in the blood. 



Their origin by the first method I can neither affirm nor 

 deny, as I have studied neither the development of the 

 larva nor its first-formed stolons. Even if some of them 

 do have this origin, certain it is that not all of them do, 

 as the following account will show. 



Their origin by the second method, /. c, from the 

 ectoderm, I fully believe does not take place. I have 

 searched through many sections prepared by many meth- 

 ods for evidences of it, but have failed utterly to find it. 



There remains, then, only the third method by which 



2d Sek., Vol. IV. ( 4 ) October 26, 1893. 



