328 ARTHUR WILLEY. 



mata which lie interspersed between the normal sized stigmata 

 have not arisen by fresh perforation^ but by unequal division of 

 other stigmata (PI. XXXI, figs. 23 and 24). It is impossible, 

 to say absolutely that no fresh perforation occurs in these later 

 stages, but at least no evidence of any such has been brought 

 forward, while contrary evidence is now being adduced. 



A portion of two adjacent transverse rows from the bran- 

 chial sac of an individual with eleven rows is represented in 

 PI. XXXI, fig. 22. On the left are two stigmata occupying 

 their normal places in their respective rows. In the middle is 

 seen one long slit stretching from the anterior limit of one 

 row to the posterior limit of the other. The blood-vessel 

 which is seen to cross the slit has nothing to do with it, as it 

 lies on a different level, and forms part of the well-known 

 internal vascular scaffolding of the branchial sac. Finally, on 

 the right is seen a similar slit, which, however, is in process of 

 dividing into two in the usual way. 



The complicated origin of the transverse rows of stigmata of 

 Ciona having now been described, it remains to point out that 

 in Clavelina, as shown by Seeliger (31) and Garstang (14), 

 and as I have also satisfied myself, these transverse rows of 

 stigmata are developed in the first instance; that is to say, 

 that ontogenetically they are primary and not secondary in 

 origin, and each aperture arises by independent perforation. 

 Here, then, is a very good demonstration of abbreviation of 

 development in the case of the branchial stigmata of Clave- 

 lina. I have already drawn attention to this remarkable 

 evidence of the modified character of the development of 

 Clavelina (34). 



2. Prseoral Lobe and Endostyle of Ciona. 

 By the name prseoral lobe I refer to a structure which 

 has long been known, but which has been assumed hitherto 

 to have no special significance, to which, however, from a 

 morphological point of view, I propose to attach a capital im- 

 portance. I have already given a preliminary account of this 

 structure under the name of proboscis cavity (34). The 



