148 THE VOYAGE OF H.M.S. CHALLENGER. 



apertures which was seen obscurely in Ecteinascidia is very well marked : they have 

 eight lobes around the branchial aperture and six around the atrial. In Chelyosoma, 

 however, the branchial aperture is six-lobed, and in Abyssascidia the number has 

 become greatly increased. In A hyssascidia loyvlllii there are twelve or fourteen branchial 

 and nine atrial. The Botryllidae, like most other Compound Ascidians, have six-lobed 

 branchial apertures. The ancestral Simple Ascidians in the neighbourhood of the point 

 M. (table, p. 150) pro1)ably acquired four-lobed or cross slit apertures, since it is the 

 general rule both in the Styeliuje and the Cynthinaj that both branchial and atrial 

 apertures should have four lobes. The Polystyelidse, like the Styelinse from which 

 they are derived, have both apertures four-lobed or quadrangular. 



In both the side branches which arose from the ancestral Cynthinse modification 

 has taken place. In the Molgulidse the branchial aperture has became six-lobed 

 while the atrial has remained unchanged {i.e. has four lobes). In the Bolteninse, 

 finally, Boltenia has both apertures four-lobed, while Culeolus and Fungulus have the 

 branchial aperture triangular and the atrial bilabiate. 



From the distinctness of the lobes and the constancy of their number in most 

 groups, there can be no doubt that the number of lobes around the apertures is of 

 some importance to the Ascidian, and yet it is not easy to say what difference it 

 makes whether there are six loljes or eight around the branchial aperture ; and 

 although it may possibly Ije an instance of correlation, — the lobes varying in accord- 

 ance with the condition of some other part of the body, — still there is no known organ 

 in the Ascidian which will account for all the variations. I have sometimes thought 

 that the structure of the branchial sac might be the cause of the number of lobes in 

 some cases ; for example, in the Styelinse, where the branchial and atrial apertures are 

 always four-lobed, the branchial sac has always four folds upon each side ; then, again, 

 in the Botryllidae, where there are six lobes around the branchial aperture, there are 

 always three well-marked internal longitudinal bars upon each side of the branchial sac. 

 But, on the other hand, in many cases there appears to l^e no connection between the 

 structure of the branchial sac and the number of folds. Perhaps the most "unaccountable 

 case is that of the genera of Polyclinidoe derived from the ancestral form occupying the 

 point H. (table, p. 150). In this little group, most closely related genera such as 

 Sidnyum, Fragarium, MorchclUum, and Morchellioides are found, some with six-lobed 

 liranchial apertures and others with eight lobes, while their branchial sacs and most of 

 the other organs of the body are almost indistinguishable from one another. 



Uljanin in his sketch of the phylogeny of the Tunicata^ shows the Appendiculariidae 

 giving rise to the Simple Ascidians, from which three lines then diverge, one to Salpa, 

 the second to Doliolum through Anchinia, and the third to the Compound Ascidians, 

 1 Fauna imd Flora des Golfes von Neapel., Monogr. x. Doliolum, p. 123, 1884. 



