BY H. LEIGHTON KESTBVEN. 331 



seems that if we look below the surface we shall find that it is 

 one which will not stand the light of investigation. Although 

 he does not explicitly state so, it is to be concluded that Grabau 

 regarded the character as inherited; for I take it that it could 

 have phylogenetic significance only as an inherited character. 

 Reviewing the two characters, rotundity and umbilication. 

 Firstly, rotundity : at the age when the mollusc deposits the 

 initial whorl it is little more than a viscous particle of protoplasm, 

 differentiated into cells certainl}^ but very little firmer. Now it 

 is inconceivable that this viscous particle could form an angulated 

 shell; it is nob endowed with pseudopodia or power to change its 

 rounded form for an angulated one as are the Rhizopoda. This 

 consideration must, it would seem, deprive the rounded form of 

 the initial whorl of any phylogenetic significance. Secondly, 

 having regard to the umbilication, a moment's thought will show 

 that all torteconchs are more or less umbilicate, and must of a 

 necessity be so. In some instances the umbilicus is filled up with 

 callus, and when the former is small we designate the latter 

 columella. This is not a juggling with words, but a statement 

 of a fact. It is impossible to wind a tube spirally without 

 having an umbilicus at first, when the said tube is round in cross- 

 section. When the whorls are wound in the same plane and are 

 increasing in size there will be a concavity on both sides, one of 

 which is merely a wide, much flattened umbilical cavity. Were 

 it possible to so wind a shell as to envelop the preceding whorls, 

 then would the umbilicus be non-existent, but the initial portion 

 is thimble-shaped and symmetrical, so that such a thing is a 

 practical impossibility. Thus we see that even such genera as 

 Cyprcea, Bulla, Volvula, etc. — types from which both columella 

 and umbilicus are absent — must have had an umbilicus at an 

 early stage of development, for the envelopment of the whorls 

 did not begin till one or two whorls were formed. From this 

 it will be seen that the " naticoid initial whorl" is not an 

 inherited character so much as the result of the twisting; i.e., the 

 twisting is the inherited character and the umbilicus a necessary 

 result. The endeavour to determine a Gastropod radicle seems 



