MUSEUM OF COMPARATIVE ZOOLOGY. 339 



impregnation is impossible, since the egg must pass into the bud 

 before it can reach a position to which the spermatic fluid can gain 

 access ; but the zooid which contains the egg, and the embryo into 

 which the latter is to develop are the children of the same parent, and 

 very close interbreeding would be apt to occur. if the testis became 

 mature before the egg had been impregnated. In Salpa, which is not 

 hermaphrodite, the same provision against incest exists, and seems to 

 have been inherited from a hermaphrodite ancestral form, which in 

 turn may have inherited it from a still more remote ancestor before 

 the peculiar method of throwing off the eggs had been acquired, in 

 which its object was, as in most invertebrates, simply the prevention 

 of self-impregnation. 



As the remainder of this paper will be mainly theoretical and 

 speculative, I must state here that while I fully realize the difference 

 between observations as to the way in which a phenomenon has been, 

 and speculations as to the way in which it may have been brought 

 about, it does not seem best to omit all theoretical discussion, 

 although the views here advanced may be very much modified or en- 

 tirely replaced by subsequent discoveries. 



Exact observations are permanent additions to our stock of knowl- 

 edge, and although they may be supplemented they cannot bo super- 

 seded ; while any theoretical views which are reached in the present 

 imperfect state of our knowledge of zoology are liable to be entirely 

 set aside by the discovery of new facts, the history of our knowl- 

 edge of Salpa shows that a theoretical interpretation may be of the 

 greatest utility, and yet be entirely false. Chamisso's theory of the 

 " alternation of generations " in Salpa has resulted in the discovery of 

 the numerous and instructive instances of true " alternation," which 

 now form so large a chapter of zoological science, although the facts 

 detailed in this paper show that, as applied to Salpa, this theory is 

 absolutely without basis. I am encouraged by this to give my views 

 of the relationship of Salpa, and of the origin of the separation of the 

 sexes, although I have kept this discussion as distinct as possible 

 from my record of observations. 



The free-swimming Tunicata have generally been regarded as the 

 lowest representatives of the group, and Huxley (Salpa and Pyrosoma), 

 gives a scries of diagrammatic sections, to show a gradual transition 

 from Appendicularia through Salpa and Doliolum to Pyrosoma and 



