942 THE AMERICAN NATURALIST: (N0L: XXXII. 
been advantageous to have a more resistant armor than that 
afforded by a layer of small articulated dermal bones. Fewer 
and larger bones, resting on and perhaps breaking joints with 
the as yet perhaps rather indifferently developed fascia bones, 
would have rendered the shield less vulnerable. It was only 
natural that the osteodermal plates of the already present keels, 
and indeed only a few of these plates, should grow at the 
expense of the smaller surrounding plates. As these few 
plates extended themselves at their base, they rose above the 
surface in the form of tubercles or spines. Possibly it was 
their function as spines that determined their growth. The 
result was finally, as I view the matter now, that one of these 
plates, with its correspondingly extended epidermal scute, occu- 
pied most of the space now covered by each of the scutes of 
our living turtles. At length the deeper elements of the cara- 
pace and plastron attained such a stage of development that 
the dermal bones were of small service, and they began to 
undergo reduction ; but this reduction did not necessarily inter- 
fere with the subsequent growth of the epidermal 'scute. In 
some cases the extirpation of outlying isolated patches of horny 
epidermis is not yet complete, as may be seen on the plastron 
of Chelydra. As already suggested, not only have the keels 
disappeared from many turtles, but in many cases even the epi- 
dermal scutes, which became associated with the ossicles of 
those keels. The supramarginals have disappeared from all 
except Macroclemys. The Cheloniidz possess inframarginals; 
so, too, does Dermatemys. Staurvotypus triporcatus' has a row 
of only two inframarginal scutes lying across its shortened 
bridge. In most genera the pectoral and abdominal scutes 
have come into contact with the marginals. There are, then, 
often found at each side of the bridge a scute, the axillary and 
the inguinal. These are doubtless vestiges of the inframarginal 
keels. 
That the epidermal scutes have originally taken their start 
from the individual tubercles of the various keels, may be seen 
on examination of the scutes in almost any of our turtles, more 
especially the lower forms, Chelydra, Malaclemys, etc. In the 
1Gray. Catalogue Shield Reptiles, pt. i, Pl. XX B. 
