1908.] 



NATURAL SCIENCES OF PHILADELPHIA. 



105 



by insensible degrees from simply arched with apical umbo to angular 

 with the umbo remote from the apex. 



In the group of Scalpellum scalpellum and the group of S. cHceratiim 

 there is similar variation in the carina, yet one cannot doubt that these 

 are natural groups. These and other like instances show that, to 

 rank the shape of the carina, whether simply bowed or angular, as a 

 character of the first importance, is to oppose it to the evidence of 

 all the rest of the organism. In ScalpeUum stearnsi I find that the 

 umbo of the carina is very close to the apex in quite young individuals, 

 becoming more remote with age; hence the angular shape of the carina, 



Fig. 1. — Types of the Scalpelliform genera: a, b, Calantica villosa; c, d, Smilium 

 peronii; e, f, EvsealjieUum rostratum; g, h, ScalpeUum scalpellum; i, j, Scalpel- 

 lum (Arcoscalpellum) relutinum. Upper line males (very much enlarged); 

 lower line hermaphrodites or females. The figures are somewhat diagram- 

 matic; carinal side of all towards the right, s.c, subcarina; c.l., carina! latus. 



being a feature acquired late in the ontogeny of the individual, has 

 probably been assumed only recently in the evolution of the groups. 



These considerations teach, I think, that an angular carina has been 

 independently acquired by unrelated species of many phyla. It can- 

 not be considered a criterion of relationship. 



Another character which has not received due weight in taxonomy 

 is the morphology of the complementary males. Former classifica- 

 tions have been based solely upon the hermaphrodite or female form, 

 which has been far less diversely modified than the male. When we 



