THE ORIGIN OF VERTEBRATEI) ANIMALS 189 



brates is thus closely analogous to that of the 

 tadpole with regard to the frog ; a stage through 

 which the frog passes, but at which it does not halt. 

 The inference in the case of the frog is that frogs 

 are descended from tadpole-like — i.e., fish-like — 

 ancestors. Now, if this conclusion is rightly founded, 

 we may conclude that Vertebrates are descended 

 from ancestors like Amphioxus, and that of all living 

 animals Amphioxus most nearly represents the 

 common ancestor of Vertebrates. The only alterna- 

 tive, and one that is urged by Dohrn, Lankester, and 

 others, is that Amphioxus is degenerate. I think 

 this alternative wrong, and that there is no real 

 retrograde development, but that Amphioxus merely 

 stops at what is an early stage in the development 

 of the higher forms. To my mind, Amphioxus is 

 one of the most important of Vertebrates, to be 

 regarded as shifting back the origin of Vertebrates 

 to an extraordinary extent, which is best realised by 

 saying that the differences between Amphioxus and 

 a fish are zoologically of far greater importance than 

 those between a fish and a mammal. 



Ascidians, or Sea-squirts. Some of these animals 

 are solitary and fixed, some in colonies, and others 

 free-swimming. They have a covering or " test " 

 (of cellulose with inhalent and exhalent apertures. 

 (The pharynx has numerous gill-slits. Water enter- 

 ing the inhalent aperture passes into the pharynx, 

 where the food-matter it contains is filtered from 

 it, and passing through the slits in the pharynx 

 -escapes into the atrial cavity, and so out at the 

 •exhalent aperture. In Ascidians the nervous system 



