AQUATIC MAMMALS 



and ulnare may be identified, but at times (see Platanista, fig. 36) the 

 conditions are more uncertain. Thie second row of bones cannot be 

 named with any feeling of confidence. Some of them are surely com- 

 pound but their constituent elements can not be determined. In Glohio- 

 cephala (fig. 36) the fifth metacarpal may be fused with the ulnare, 

 thus in effect connecting this digit with the ulna, and in Platanista (fig. 

 36) the same end is attained by the direct conjunction of the metacarpal 

 with the ulna. Variations of this character might be enumerated at 

 great length but without much profit. 



In the Mysticeti the carpus may have as many as nine elements (in 

 Eubalaena, according to Holder, 1883) including a presumable pisiform 

 extension, five in Sibbaldus, or even as few as three in Balaena, accord- 

 ing to Flower (1873), but the latter is not clear as to whether this 

 includes the pisiform, which is present. 



I do not see that the undoubted trend of the cetacean carpus can be 

 determined. Is it in the direction of a bony mosaic, like it has been 

 in marine reptiles, or toward a completely cartilaginous carpus, as the 

 condition in Mysticeti might indicate .»* The decreased rate of ossifica- 

 tion even in odontocetes might bear but the latter possibility, but for 

 all that we know the two groups may ultimately attain opposite goals 

 in this respect. 



In the Cetacea the metacarpals of those digits having considerable 

 length are absolutely indistinguishable from the proximal phalanges. 

 They all have precisely the same function, or rather practical lack of 

 function, and hence the two elements may be considered to comprise a 

 homogeneous complex. 



There are invariably five digits in the Cetecea except in the case of 

 the rorquals, which have four. As might be expected the development 

 of the border digits is largely dependent upon the conformation of the 

 flipper as a whole. In those sorts having a pollex this is always short, 

 but the fifth digit may be either practically as long as any of the others 

 {Platanista) or extremely rudimentary in a flipper of such extreme 

 falcate shape as that of Globiocephala. In those sorts with narrow 

 flippers the digits are situated close together, while if these mem- 

 bers are broad the digits are disposed somewhat fanwise. Invariably, 

 I believe, the second one of the digits present is either the longest, or 

 practically as long as any other. {Platanista) . Unlike the case of the 

 Otariidae the digit upon the anterior border of the flipper is (as a rule) 

 not markedly more robust than any other, although Flower and Lydekker 

 (1891) have stated that in Sotalia the two outer digits are heavier than 



[256] 



