MOLLUSCAN GENUS COCnLITOMA. 265 



species. Thus, in two examples of Burtoa nilotica the commence- 

 ment of the sculpture varied from whorl 2| to whorl 3|. The 

 result of this is that it is possible to find examples of species 

 placed by Pilsbry in his first group (" embryonic whorls sculp- 

 tured ") with sculpture starting in exactly the same spot as it 

 does in species placed in the second group. For example, in the 

 two examples of C. zebra var. fulgurata in the British Museum 

 the sculpture starts on whorl 2|, the same position as in an other- 

 wise typical Burtoa nilotica. A more serious consequence is that 

 Pilsbry 's arrangement cuts right across several other potential 

 arrangements of equal and possibly greater importance. The 

 character of the penis retractor which is either inserted on the 

 diaphragm or on the right ocular bend, that of the penis ring- 

 muscle, whether a complete sheath or annular, and the character 

 of the admedian teeth, whether having three, two, or one cusps, 

 are three out of many characters which might be used as a basis 

 and, if used, would give a different arrangement of the genera 

 and species. 



If a classification based upon the shell gives (as it certainly 

 does) a different result from that based upon the radular 

 characters, and if the latter yields a different scheme from that 

 based upon the insertion of the penis retractor, it is plain that 

 any classification based upon a single character must be arbitrary*. 

 What is required is a complete knowledge of the variation and 

 mode of inheritance of all the important structures in each genus 

 and a classification based upon the greatest measure of agreement 

 among the various forms. 



The anatomy of all the genera and species of this subfamily is 

 not known. It may be, therefore, argued that a provisional 

 arrangement based upon the only organ known in the species, 

 viz. the shell, is justified. 



If, however, the object of classification is a correct arrangement 

 expressed in terms of relationship, it would seem best, in cases 

 such as the present, to select from all the available data a com- 

 bination of the more important characters as a basis and to 

 arrange the species upon it as far as possible, even if some are 

 left vaguely designated as "Achatina." The selection of a 

 character (in this case the shell) that will enable every species, 

 whether known completely or not, to be given a definite jjlace, 

 involves the production of a scheme that sooner or later has to 

 be revised, if not abandoned altogether. 



With regard to the conchological genus Cochlitoma itself, 

 only three species are known anatomically, and of these only the 

 two varieties of zebra are known in any detail. But it is 

 very doubtful whether these four forms are referable to the 

 same genus. C. crawfordi differs from the others in the 

 remarkable insertion of its penis retractor. C. granulata 



* Note. — It is conceivable that intensive study may show that, within a given 

 genus, one organ or structure reflects genetic modifications better than others. " But 

 this has yet to he proved, 



