REMARKS ON THK J^EVELOFMENT OE AMl'HlOXUS. 353 



atrial cavity becomes floored in by ingrowths whicli meet one 

 another, arising from the inner sides of the free ridges. 

 They were not able to trace the lymph canals into continuity 

 with the coelom, but state that they appear to be formed by 

 the hollowing out of the originally solid ridges. The atrial 

 cavity is at first a narrow tube, whicli later expands so "as 

 nearly to surround the alimentary caual.^' 



When in my first paper I stated that Lankester and Willey 

 had confirmed most of Kowalevsky's results, I believed then 

 and believe still that the difference between the flooring in 

 of the space existing between two ridges, and the growing 

 together of the same ridges so as to meet, were practically 

 only a difference in the form of expression. Kowalevsky 

 undoubtedly figured the atrial ridges too far from the mid- 

 ventral line, and does not seem to have observed the first 

 origin of the cavity, but only a later stage, and the account of 

 Lankester and Willey is a needed correction of his observa- 

 tions ; but it seems to me rather unreasonable to require a 

 detailed reference to these minor differences in a paper 

 mainly devoted to the earlier development of Amphioxus ; 

 especially when in a later part of the same paper I further 

 described the views of Lankester and Willey as to the mode 

 of growth of the atrial cavity, and expressed doubts as to 

 their accuracy. 



Both, however, in the resume of their observations given 

 by Lankester and Willey, and in the criticism of my paper 

 by Professor Lankester, there is a view put forward of the 

 origin of the atrial cavity which seemed to me most improb- 

 able, and one main object of the present paper was to test its 

 accuracy by fresh observations. 



In the paper by Lankester and Willey they sum up the 

 development of the atrium in these words : '' a narrow groove 

 which closes and sinks (as it were) into the body of the 

 Amphioxus ;" although tliey are careful to add that the 

 mode of formation suggested by Kowalevsky and that de- 

 scribed by them " ultimately come to the same thing so far 

 as the obvious morphological relations are concerned." In 



