MOSS AND BOYCOTT : OBSERVATIONS ON THE RADUL/E OF HYALINMi. 1 59 



the mesocone to the tip of the ectocone. This has been done 

 systematically for the (8) centrals and (9) first laterals only; it is, 

 however, clear that useful information would have been obtained by 

 the extension of this method of examination to the second and third 

 laterals and to the first two marginals. The rest of the table requires 

 no further explanation. 



The four radulse concerned are constructed on the same plan. 

 Apart from differences in size, the main distinctions I'e in the 

 characters of the third and fourth teeth in each row. Thus in drapar- 

 naldi, the third lateral has a definite ectocone in fifty four per cent, of 

 the specimens, and a well-developed endocone in one hundred per 

 cent., while in forty-five per cent, the first marginal also has an endo- 

 cone: the radula, in short, is tending to have four laterals. In glabra, 

 on the other hand, it is tending to have but two laterals : for in no 

 case has the first marginal any endocone, and the endocone of the 

 third lateral is noted as definitely small in sixty-eight per cent., and 

 absent in eight per cent., while the same tooth shewed an ectocone in 

 only one instance. Between these two lie cellaria, in a fairly inter- 

 mediate position, and alliaria close to glabra. The chief difference 

 between the two latter seems to be that alliaria has a well-developed 

 endocone to the third lateral in ninety per cent, as against twenty-four 

 per cent, of the specimens, and glabra shews the curious truncation 

 of the mesocone of the central tooth in forty per cent, of radulae. 



There is thus a tendency for the first marginal tooth to move 

 outwards along the series glabra, alliaria, cellaria, draparnaldi. This 

 is associated with a similar progression of the largest tooth, which is 

 generally the first tooth with truly marginal characters, and in 

 draparnaldi is frequently as far out as the fifth, while in glabra it may 

 be the third. The differences are obviously not absolutely diagnostic; 

 the following formulae give, perhaps, fair representations: 

 draparnaldi ^-f4- + — x 29 

 cellaria \ + ; = ^ .^ + JLl x -:4 



3 3 2 or 3 1 'J'^ 



alliaria -V -f- : " ^ + 11 x -ic 



glabra -1- -f -|-L_^ + 2_i x 39 



We do not propose to enter into much detail as to the individual 



specimens examined; with regard, however, to the ''glabra'' specimens 



something more must be added. These fall into two distinct groups: 



(a) Specimens from Finchley and Bettws with 37 to 45 



(average 41) rows; average size of radula 2-3 x 0-53; length 



occupied by 10 rows =0-56; average number of marginals in. 



(/?) Specimens from Anglesea and Northampton with 31 



to 38 (average 34) rows; average size of radula 17 x 0-45; length 



occupied by 10 rows =0-50; average number of marginals io-6. 



