MUSEUM OF COMPARATIVE Z05LOGY. 341 



The Cirrhipeds, for instance, pass through a" nauplius " stage, which 

 has been retained because it is necessary, but if a Cirrhiped should 

 now become adapted to a locomotive life, the nauplius stage would 

 lose its importance, and might, in time, disappear by the process 

 of acceleration of development. This seems to have taken place 

 in Salpa. The " tadpole larva " appears to represent a form similar 

 to Appendicularia, from which the Tunicata are descended ; and this 

 tailed stage has been retained by most of the fixed Ascidians, but 

 in Pyrosoma, where the embryo does not need special locomotive 

 power, the tail is not formed, although the embryo in its development 

 passes through this stage, which is represented by the so-called Cyatho- 

 zooid. 



In Salpa, not only the tail but all traces of the larval stage have dis- 

 appeared, and the egg shapes itself at once and directly into the per- 

 fect animal. As Herbert Spencer would state it, the indirect method 

 of development has given place to the direct,* as is so often the case 

 in the higher forms of a group ; the Cephalopods, among the Mollusca, 

 for instance. 



The relations of the branchial sac and atrium may be explained in 

 a similar way. Leuckart has pointed out (loc. cit.) that the " gill " 

 of Salpa is simply a sinus between two large branchial slits, and we 

 have seen that the atrium of the embryo has the two lateral chambers 

 upon the sides of the branchial sac, which in ordinary tunicates com- 

 municate with the branchial chamber by the formation of the bran- 

 chial slits. In Salpa it has been shown that no branchial slits are 

 ever formed upon the sides of the sac, and that the lateral atria be- 

 come converted into the respiratory muscular girdles. If we attempt 

 to follow out in imagination, the manner in which a tunicate might 

 become adapted to a free locomotive life, we can see that if an ordinary 

 ascidian were to be loosened from its attachment the branchial current 

 would give it more or less motion in the water, and if this free 

 life were advantageous natural selection might in time lead to the 

 separation of the branchial and atrial openings until they reached 

 opposite poles of the body, in which position the current of water 

 would be most efficient as a motive force. (Sars calls attention to the 

 fact that the two openings are nearer each other in the embryonic 



* According to Krohn (Wiegem. Arch. 1852, XVIII., I., 53, Taf. 2), Doliolum passes 

 through a larval stage, and at this time is provided with a tail. 



