146 ARCI1WE. 



colouring. The first ground is evidently untenable, 

 because size is notoriously dependent on food, shelter, 

 and the proportionate quantity of carbonate of lime con- 

 tained in sea-water, according to its proximity to the 

 shore, or distance from it, and to the action of fresh- 

 water and marine currents. The second ground, or the 

 degree of those angles which affect the contour of 

 Nucula, is influenced by the growth of the shell in any 

 particular direction. All the Nucula represent an iso- 

 sceles triangle, the beak or point of one angle always re- 

 maining the same, while the sides or points of the other 

 two angles extend pari passu by the addition of new 

 layers in either direction. If one of these sides is more 

 produced than the other while the angle of each is 

 greater or less than that of the beak, a scalene triangle 

 is the result. This is the case with N. radiata ; and if 

 it were a permanent or invariable character, I should 

 regard it as having considerable weight in deciding the 

 question. But were the most acute mathematician to 

 measure the angles in certain specimens of the two so- 

 called species, he would assuredly fail to detect any 

 sensible difference. As to convexity, it is true that spe- 

 cimens of N. radiata are usually more compressed than 

 those of N. nucleus. Other specimens, however, of both 

 species are equally convex. The coloured rays are 

 clearly a varietal, and not a specific character. They 

 are even more conspicuous in another species (N. nitida), 

 which has commonly a plain and sober hue. Instances 

 of a similar diversity in this respect occur in Astarte 

 triangularis and many other marine shells, as well as in 

 species of the freshwater genus Unio. The nature and 

 cause of colour in shells has not yet received that atten- 

 tion from philosophical chemists which the interest of 

 this curious subject demands. Probably the most in- 



