PHYLOGENY OF THE PELECYPODA. 317 



of the spat period if the shell does not project freely from the olijeet of fixation, then the 

 relative shape of the valves characteristic of the spat is retained, the attached valve still 

 spreading over the support, is flat, and the free valve is convex. In such forms the true 

 spat jjeriod is marked as a smooth rounded convex area at the umbos. 



In yonng Ostrea virginiana spat, as noted by Ryder (GO), bands of purple color I'adi- 

 ate from the hinge line, widening toward the free edges of the valves; but these colora- 

 tion bands are in the prismatic layer and are therefore limited to the upper right valve. 

 In none of the oysters growing on glass, have I seen coloi- bands in the lower valve, 

 which is nearly or quite white; in later growth, however, faint traces of color bands may 

 be occasionally seen in the left valve. I found spat in the living chamber of a Nautilus 

 Poinijilius, when unpacked on arrival from Singapore, and these are similar to the spat 

 of 0. virginiana, showing similar prodissoconchs and nepionic stages, but no coloration 

 bands, the upper valve being yellowish-white. The eai'ly sjiat growth of Ostrea edulis, 

 fig. 27, p. 314, is very similar in foi-m to that of 0. virginiana, PI. xxiv, fig. 20, but I 

 have not had specimens for a comijarison of later stages. 



YII. Gryph^a, Exogyra and Adult Ostrka, with a Consideration of the 

 Mechanical Origin of the Ostrean Form of Shell. 



The genus Gryph^ea is described as free or attached in the young, yet I have not 

 been able to find a species in which it is free at a very young stage. All the species 

 studied have a flat area in the lower valve and a curved area in the upper valve at the 

 mnbos, corresponding to the condition found in the attached spat stage of the oyster. 

 G. arcuata of the Lias is commonly considered as free, yet I have found well preserved 

 specimens, PI. xxiv, fig. 22, in which the lower valve is flat at the apex, where it was 

 attached when young, and correspondingly the upper valve has a small convex area at the 

 uml)o, as noted by Sowerby. Flat areas of attachment have also been found at the umbos 

 of two other highly arcuate species, namely, G. calceola, var. 7iehrascensis, M. & H., PI. 

 XXIV, figs. 23-24, from the Jurassic of this country, and G. 'pitclieri, Morton, from the 

 Cretaceous of Texas. 



The area of attachment in many species of Grypha'a is uniform in size, indicating that 

 the shell dehisced at a very definite period from the object of attachment. In fact, highly 

 arcuate forms must have separated early, as the involute form could not have been de- 

 veloped unless the shell was free. Other species were attached for a very variable period, 

 as is seen by the differing size of the flat area of the lower valve and corresponding 

 curved area of the upper. This is well shown in any large collection of the Enro])ean 

 form of Grgphrea vesicularis from the Cretaceous, where the arcuateness of the adult 

 varies invei'sely as the length of period of attachment. 



Exogyra, also, is said to have been free or attached; but here the case is somewhat 

 more comi^lex. Take, foi- example, the typical E. costata. Say, of the Cretaceous. Spec- 

 imens are found in which a flat area at the apex of lower valve distinctly shows them to 

 have been attached when young, but rarely are adults found with the object of eai-ly fixa- 

 tion still clinging to them, showing that as in Gi'yi)h;ea and many oysters, tlu'y sepai'atcd 

 naturally. Other specimens are found which, upon most carefid examination, sIkjw no 



MEMUIUS BOSTON SOC. NAT. HIST., VOL. IV. Jj 



