41(1 H B. POLLAKD, 



From such a scheme it may be learnt that all parts are inde- 

 pendently variable. Root piece and tentacle in Amphioxu» are essen- 

 tially similar in structure. In Myxinoids the root piece is differentiated 

 from the tentacle by the great development of the intercellular matrix, 

 the hardening of the same, and the degeneration of nuclei and proto- 

 plasm. In Siluroids much more comi)licated differentiations have 

 arisen. The tentacles are not of similar histological nature in different 

 families. The root piece differs from the tentacle in the same indi- 

 vidual, the root pieces differ among themselves, some being of pro- 

 cartilage, some of the Myxinoid tissue, and some of true hyaline 

 cartilage. Usually in one animal all the tentacles are of similar 

 histological nature, but nevertheless this is not always the case e. g. 

 in Motella tricirrata. 



To a certain extent the grade of histological differentiation is a 

 measure of the constancy of the piece. For example Meckel's cartilage 

 and the hyoid cartilages occur in all the Craniata. The prepalatine ' 

 piece, the hyaline premaxillary piece of Teleostei, the mental piece of 

 Holocephali and of the Dipnoi are less omnipresent, but still run 

 through whole orders, while less differentiated tentacles are present 

 or absent in different genera. Of course there is no universal rule. 



In view of the extraordinary amount of variation in histology 

 and structure, it is very remarkable that, when tentacles do occur 

 sporadically, they can be refered to certain of definite 6 or 7 pairs. 

 Following the conception of Weismann, it would seem that the archi- 

 tecture of the germ plasm is more constant than the quality of the 

 determinants. Exceptions to this rule may be discovered. 



The embryonic development of the tentacles in Ictalurus albidus 

 has been investigated by Ryder. Those present in the adult develope 

 early in situ and there is no parallelism with the phylogeny. 



Reversion and Larval Forms. 



From comparison of long lists illustrating the occurence of the 

 individual tentacles, and from consideration of the fact that they 

 api)ear sporadically, I have come to the conclusion that it would be 

 extremely rash to maintain that the tentacles have come down in 

 unbroken ancestral line from an early progenitor. In other words, 

 their presence must often be due to reversion ' ). They are not always 



1) Or, in many cases, inasmuch as rudiments of tentacles are almost 

 always present, by "redevelopment from rudiments" (Darwin). No sharp 

 'listinction can be drawn between the phenomena of reversion and 

 redevelopment from rudiments. 



