590 A. WILLEY. 



the branchial epithelium consists of very high filiform 

 columnar cells in which the nuclei occur at different levels ; 

 it is probable that the appearance of stratification is still 

 further increased by the compression of the epithelium as a 

 consequence of contraction. 



What may be described as a sensational conclusion is that 

 which proceeds from the author's comparison of Am phi- 

 oxides with Branchiostoma^ i.e. that the so-called 

 secondary gill-slits which suffer a retardation of develop- 

 ment in the larva of Branchiostoma are heterogeneous 

 formations, not homologous with the primary series. Of 

 course this conclusion is not rendered in an arbitrary manner, 

 but is led up to by a number of arguments based upon the 

 grand assumption that Amphioxides alone is primitive, 

 and that the larval development of Branchiostoma points 

 no farther back, but is a mere recapitulation of the charac- 

 ters of the original pelagic Acraniate as represented by the 

 former genus. 



Dr. Goldschmidfc adopts the method of assuming that his 

 form is primitive, and then explaining the facts on that 

 assumption; and he claims to explain all the facts, whereas 

 other theories only explain some of them. But Amphi- 

 oxides may be as highly adapted to a pelagic life as 

 Branchiostoma to a beuthonic life. I am aware that this 

 is an easy objection, but it is none the less true. In most or 

 many other sharply defined orders, those forms in which an 

 entire organic system is functionally deranged or obsolete, 

 are not usually the most primitive. A simple example is the 

 eyeless condition of many cave animals, deep-sea animals, 

 and otlie cryptozoic forms ; another is the limbless condition 

 of some Teleostean Fishes, Batracliia, and Reptiles. They 

 may be primitive in other respects, but not in respect of 

 their lack of parts. Perhaps a closer analogy is afforded by 

 the pelagic Tunicata of the class Copelata which, like the 

 Acraniata, comprises two families, Kowalevskidae and Appen- 

 dicularidse, the former characterised by the absence of an 

 endostyle, but it is not suggested that the Kowalevskida) 



