370 FOSSIL TURTLES OF NORTH AMERICA. 



series of fossil turtles of the various geological periods would doubtless show a gradual reduction 

 of the size of the rib-heads in this family. 



The distal ends of the costal ribs of the Testudinidae have become reduced in an interesting 

 way. In the great majority of turtles the distal ends of these ribs project more or less beyond 

 the margins of the costal plates, and the projecting ends are usually received by pits in the 

 adjacent peripherals. In most of the Testudinidae the ends of the ribs have disappeared. 

 In Gopherus we find that, instead of the rib entering the peripheral, a broad process of the 

 peripheral rises and enters a notch on the inside of the costal plate, an interesting reversal of 

 conditions. The relations of the ribs and peripherals in 7 '. radiata are very similar. In T. 

 tabulata no conspicuous process is found on either the costal plate or on the peripheral. In 

 Hadrianus, as in Gopherus and most species of Testudo, a process of the peripheral enters an 

 excavation in the inner surface of the costal. On the other hand, in Stylemys the extremity 

 of the rib fits into a pit in a peripheral, just as in the Emydidae. 



In typical Testudinidae of to-day the peripherals and contiguous costals are articulated, 

 not only by the processes just described, but also by digitating sutures, and these are usually 

 better developt than in the Emydidae. In T. radiata long slivers of bone arising from the 

 peripherals interdigitate with shorter processes from the costals. On the other hand, in 

 Stylemys and Hadrianus there appear to have been no such digitations; but the rounded 

 border of each peripheral joins the similar border of the neighboring costal. 



In the majority of the Testudinidae the epiplastral bones are thickened for some distance 

 on each side of their symphysis. The thickening takes place on the upper side of the bones. 

 At the hinder border of the thickening, on each side of the midline, the elevation drops off 

 suddenly to the level of the entoplastron. In many species the ledge thus formed and looking 

 backward is excavated more or less deeply. Usually the thickened epiplastra project more or 

 less beyond the general curve of the rest of the anterior lobe, and there is thus formed a con- 

 spicuous epiplastral lip. This lip presents various forms in the different species and, tho 

 subject to variations, it furnishes valuable specific characters. The lip is conspicuously 

 developt in Hadrianus, as in most species of Testudo. In Stylemys the front of the epiplastra 

 is thickened, but it does not project beyond the curvature of the rest of the lobe. 



The hinder end of the plastron is usually, but not always, notcht. 



At the anterior ends of the bridges the hyoplastron of each side sends upward a process, 

 the axillary buttress, which articulates with the first costal plate. Usually this buttress does not 

 pass above the lower border of the costal. At the hinder end of the bridge the hypoplastron 

 sends up an inguinal buttress, which also rarely rises above the lower border of the costals 

 involved. In Gopherus and the species of Testudo already mentioned this buttress is articu- 

 lated with the middle of the lower border of the sixth costal; but it appears that in some 

 other living genera of the family, possibly in some other species of Testudo, the buttress is 

 attacht somewhat further forward. In Hadrianus corsoni the buttress ascends a short distance 

 on the inner side of the sixth costal. In Stylemys nebrascensis it is articulated between the fifth 

 and sixth costals, rising a short distance above their lower borders. It is evident that there has 

 been a tendency among the Testudinidae to push the buttress backward, and thus increase 

 the width of the bridge. The purpose has apparently been to diminish as much as possible 

 the hinder opening of the shell. 



The shoulder-girdle of the Testudinidae appears to be distinguish! from that of the 

 Emydidae in two respects: The coracoid is greatly expanded at its inner, or mesial border, and 

 the procoracoid process makes an obtuse angle with the body of the scapula. Probably in all 

 Emydidae the coracoid is only slightly expanded at the mesial border and the procoracoid 

 process makes an acute angle with the body of the bone. 



The humerus, in probably all members of the family, is characterized by having the radial 

 and ulnar processes both twisted toward the ventral side of the bone, thus making the angle 

 between them small. Usually the humerus is considerably bent, but often no more so than in 

 some species of Emydidae. 



In most species of the family there are 5 digits belonging to each forelimb; but occasionally 

 there are only 4. No digit possesses more than 2 phalanges. How this reduction has been 

 brought about is an interesting question. A similar reduction of the phalanges to two in each 

 digit occurs in the forefoot of Terrafene as observed by Dr. George Baur (Zool. Anzeiger, xv, 

 1892, p. 159), 



