3G6 NOTES ON SOME BRITISH NUDIRKANCIIS. 



6. Cr. co7icinna (A. & H.).* 



7. Cr. pilata (Gould). Nortli-West Atlantic. 



8. Cr. hylgia, Bergli. riiilijjpines. 



9. Cr. longibursa, Bergli. Pelew Islands. 



10. Cr. stipata (A. & H.). 



11. Cr. cavancce, Bergh. Chile. 



12. Cr. jt)?m//rt, Bergh. Juan Fernandez. 



13. Cr. cucullata, Bergh. Malay Archip. 



Cr. viridis (Forbes) and Cr. f/lottensis are probaldy Aniphorinas, and it 

 seems difficult to separate Cratena gymnota (Gould) from Amphorina 

 aurantlaca. Cr. arenicohi (Forbes), Cr. veronica (Verrill), Cr. couchii 

 (Cocks), and Cr. jrurpiirascens (Fleming) are imperfec;tly described. 

 Cr. cingulata is placed by Alder in Jeffreys Conchology, vol. v. p. .53, 

 under Galvina, which implies that it has a triseriate radula. Cr. luguhris, 

 Bergh, is referred to this genus with a query by the author himself. 



In Jeffrey's Conchology, vol. v. p. 45, etc., Alder distributes the 

 British forms between CiUhoiia and Cavolina. The former, which is 

 defined as having the branchia.* close set and a uuiseriate radula with 

 a central spine and lateral denticulations, includes C. peachii, C. nana, 

 C, stipata, C. angulata, C. inornata, C. concinna, C. olwacea, C a^irantiaca, 

 and C. pustulata. Cavolina, which is defined as having the branchiae 

 set in rather distant rows and a uniseriate denticulated radula with 

 the central spine a little prominent, includes C. couchii, C. avimna, 

 C. nortJmmhrica, C. arcnicola, C. glottensis, C. coerulea, C. viridis, and 

 C. purpurascens. Though something may be said for this division, it 

 ignores the existence of Ampliorina (Quatrefages, 1844), and the point 

 on which stress is laid — whether the branchiae are close set or in 

 distant rows — is often difficult to determine in preserved specimens, and 

 might be doubtful in li\ing ones. 



In the above remarks 1 have merely considered how the described 

 species can be most naturally distributed among the described genera. 

 Whether those genera are valid and necessary is another question. 

 I doubt myself if Bergh's whole subfamily of. Cratenidre, including 

 Amphorina (but perhaps excluding Fhestil/a.) otters more than one good 

 generic type ; and if nature provides a great number of similar forms it 

 seems unscientific to separate them into many genera on the strength of 

 small points of difference. It is convenient, no doubt, to divide a large 

 genus into sections or subgenera, but to split it into several genera if 

 the differences between them are of less than ordinary generic value 

 obscures tlie real uniformity of tlie animals classified. 



* Cr. concinna. Since writing the above I Iuivl' liad lui (ipportunity of exaniiuiug a 

 speeiineii of tliis s])ecies from tlie Menai Straits, given nie l»y Prof. Herdnian. The 

 character.s agree with A. k II. '.s ilcscription. No style was found on the penis. Tlic 

 teeth of the radula are remarkably long, narrow, and pointed. 



