ASCIDIACEA DEVELOPMENT OF THE FREE-SWIMMING LARVA. 367 



to VAN BENEDEN and JULIN, are for the most part of entodermal 

 origin, a large although not sharply marked portion of the peri- 

 branchial cavity must be lined with entoderm. These observers 

 therefore maintain that these spaces cannot be homologised with the 

 atrial cavity of Ampkioxus, and the future perforations in their inner 

 walls can in no way be homologised with the gill-slits of Amphioxus 

 and the Vertebrates, and are therefore distinguished from true gill- 

 slits by these authors as branchial stigmata. We have some hesita- 

 tion in accepting the view of VAN BENEDEN and JULIN. The origin 

 of the peribranchial cavities in the Ascidians does not appear to us 

 sufficiently understood to justify its utilisation in the formation of 

 such important conclusions. It should be pointed out that, in the 

 embryo of Pyrosoma, the purely ectodermal origin of the peribranchial 

 cavities can hardly be doubted (see below, p. 394), and this seems- 

 also to apply to Doliolum (see the statements as to the Anchinia buds,, 

 p. 481). Again, we are probably not justified in concluding without 

 further evidence that the condition of Append I cularia is primitive. 

 Since an Appendicularian was found by Moss (No. 5) possessing 

 many gill-clefts like those of Doliolum, we may have to regard the 

 apparently simple respiratory organs of other Larvacea as degenerate. 

 We shall have to speak of the further transformations in the 

 branchial region when describing the metamorphosis of the attached 

 larva, although, in some cases, the multiplication of the gill-slits 

 commences in the free-swimming larva (Botryllus, Distaplia, Fig. 

 230, p. 457). 



LAHILLE (No. 38) and WILLEY (No. 54a) have both recently expressed their 

 belief in the purely ectodermal origin of the peribranchial sacs. According to 

 LAHILLE, they arise, in the Didemnidae, through the enlargement of the 

 cloacal vesicles. Their outer aperture then disappears and an unpaired 

 ectodermal invagination appears on the dorsal side ; this is the cloacal in- 

 vaginatioii proper, which only secondarily becomes connected with the 

 peribranchial sacs. WILLEY (No. 54a) observed, in Clavelina, as the pre- 

 cursor of the cloacal vesicle, a longitudinal furrow, from the posterior end of 

 which the cloacal vesicle develops. WILLEY is therefore inclined to assume 

 that the peribranchial cavity of the Ascidians is homologous with that of 

 Amphioxua. 



The outer apertures of the two peribranchial spaces (Fig. 231, 

 p. 460) shift continually towards each other and towards the dorsal 

 middle line till they meet and fuse. In this way arises the cloacal, 

 atrial or inhalent orifice (Fig. 230, e). VAN BENEDEN and JULIN 

 (No. 9) have pointed out that, during this fusion, the part of the 

 ectoderm which lies between the two apertures becomes depressed 



