530 MAYNARD M. METCALF 



all cases the secretion is foriued by the degeneration of cells pro- 

 liferated from the endothelium of the gland. 



Herdman in his report upon the Tunicata of the Challenger 

 Expedition ^) argues in favor of a diphyletic origin of the compound 

 Ascidians, The results of my study upon the intersiphonal organs 

 would agree with this hypothesis. The neural gland is dorsal to the 

 ganglion in the BotrylUdae as in the Cynthiidae. In the other 

 compound Ascidians studied (i. e. in the Distomidae, the PolycUnidae, 

 the Didemnidae, and the Diplosomidae) the gland is ventral, as in 

 the Ascidiidae and Clavelinidae. In the gland of some species of 

 the latter family (e, g. Ecteinascidia turhinata and Perophora viridis) 

 we find the same sort of vacuolated cells that are so characteristic 

 of the four families of compound Ascidians last mentioned. 



The neural gland of the BotrylUdae differs from that of any of 

 the other compound or simple Ascidians in having two somewhat 

 distinct portions, an anterior chamber and a posterior more nearly 

 solid portion. 



It is somewhat puzzhng to find in the posterior portion of the 

 gland of Polycyclus (one of the BotrylUdae) gland cells similar to 

 those in those families of the compound Ascidians which Herdman 

 beheves have had an origin distinct from that of the BotrylUdae. 

 Herdman derives both groups of compound Ascidians from the Cla- 

 velinidae as their remote ancestors. He might therefore say that one 

 (or more?) species of the BotrylUdae and all species of the other 

 great group of compound Ascidians have retained the vesicular gland 

 cells characteristic of their remote ancestors, the Clavelinidae. 



In all the Ascidians I have studied, both simple and compound, 

 a more or less gangliated nerve cord is present in the raphe in close 

 contact with the rapheal muscle. This has been described for mauy 

 Ascidians, but has not been regarded as universally present. The 

 ganglionic swellings upon this nerve in some of the Molgulidae do 

 not, I believe, indicate metamerism. 



Very great diversity is shown among the Ascidians as to the 

 condition of the rapheal duct. It is present and well developed in 

 some of the Clavelinidae (Clavelina, Bliojxdaea, Diazona) while ab- 

 sent in other species {Ecteinascidia and Perophora). Among the 

 Ascidiidae it is present and fully developed in the Cioninae, and a 

 rudiment is present in the Ascidiinae. Turning to the Cynthiidae 



1) Herdman, 1886, p. 387—399. 



