BOYCOTT : THE RADULA OF HYALINIA. 



23S 



difficulties of accurate measurement, which have already been noticed, 

 render this conclusion uncertain, though it hardly seems likely that 

 errors of observation can have led to such a large difference. With 

 regard to all the figures it must also be remembered that the shells 

 from which the radul^e were derived were not all the same size but 

 ranged either way from their mean by 6%, and that every item varies 

 with the size of the shell. ^ 



The two teeth which obviously vary in their configuration are the 

 third lateral and first marginal. With regard to the third lateral we 

 may distinguish five types, shewn in the diagrammatic sketch, which 

 occur with the following frequency :— 



Type I. Endocone absent, ectocone absent 



Type II. Endocone small, ectocone absent 



Type III. Endocone small, ectocone a shoulder ... 



Type IV. Endocone present, ectocone absent 



Type V. Endocone present, ectocone a shoulder 



Type VI. Endocone present, ectocone present ... 



0=0% 



10=15% 



6= 9% 



10=15% 



34 = 51% 



6= 9% 



SO ^ 



Type I. did not occur in the series. In the whole series the endo- 

 cone was " small " in 24% of the specimens, the ectocone absent in 

 30%, represented by a shoulder in 60% and well developed in 9%. 

 With regard to the first marginal, the endocone was present in seven 

 cases, though always small or rudimentary, and in two on only one- 

 half of the radula. It should also be noted that in six of these 

 specimens, the third lateral had a definite ectocone; in the other the 

 ectocone was represented by a shoulder only. This simultaneous 

 progress in differentiation is a regular feature of the Hyalinia radula. 

 Thus an ectocone on the third lateral is commonly associated with a 

 well-developed endocone, and if the latter is very small the ectocone 

 is often completely absent. The difference may sometimes be seen 



I The sort of difference which is introduced by this factor may be judged o: by the a\ 

 figures of half millimetre groups : — 



Group Mean diameter Rows Marginals Long Wide 



7'o — 7"4 7'24 4i'2 ii"4 2'30 o^^^ 



7'5 — 7'9 7'64 407 if2 2'50 o'54 



8'o— S'4 815 4i'6 II 9 2'54 o'59 



The correlation between total number of teeth and area of radula is only +o'22 X°'°7- 



Specimens 

 34 

 32 



