INTRODUCTION. 11 



This arrangement of parts is not to be found in otber genera of the 

 Delphinulcc, but is characteristic of the tiuviatile dolphins {Platanista, 

 etc.), to which indeed the Delpkinaptenmc show many marks of affinity. 

 Their separate cervical vertebric, prolonged pterygoids, broad pectorals, 

 and rudimentary dorsal fin, taken together, entitle them, I believe, to be 

 regarded as a distinct subfamily. Elsewhere in the group I do not 

 perceive that broad divisions are called for. Professor Flower employs 

 provisionally the characters furnished by the shape of the head as a 

 means of dividing the family into two groups. These characters, as 

 Professor Flower himself admits, though useful and seemingly in accord- 

 ance with natural affinities, within certain limits, are not trenchant. 



The characters of the two divisions as regards the form of the head 

 are as follows:* 



a. With ronuded head, without distinct rostrum or beak. (Among the genera in- 

 cluded here are Ce])haIorhynchus and Lagenorlnjnchus.) 



h. Dolphins with distinctly elongated rostrum, or beak, generally marked off from the 

 antenarial adipose elevation by a V-shaped groove. (Comprises Delplimus, 

 Tarsiops, Prodelphinus, Steno* and Sotalia.) 



Leaving Monodon and DelpMnapterus out of consideration, this dis- 

 tinction is valid for the majority of the genera, but is broken down by 

 Lagenorhynclms and Cephalorhynclms. In the former genus (included 

 in section a) the beak, though shorter than in Tursiops (included in 

 section 6), is quite distinct and well marked off from the forehead, while 

 in some species of GepUalorhynchus the head is certainly not "rounded" 

 in the sense of being globose, but is conical. 



The second character of the sections has to do with the length of the 

 rostrum as compared with the total length of the skull. Here again 

 Lagenorhynchus and Geplialorliynchiis appear intermediate. Cephalo- 

 rhynchus eutropia (section a) has the beak relatively as long as Tursiops 

 tursio (section b); the same is also true for some species of Lageno- 

 rhynclms. 



In spite of these considerations, however, I have employed these char- 

 acters in the artificial keys to the genera, given on jiages 152 and 153, 

 believing them to be as useful, for that purpose at least, as any which 

 can be formulated at present. 



Among the sui)ergeneric distinctions employed by Professor Flower 

 is one which was brought into requisition for the firsttime and seems to 

 be of value; this relates to the position of the two pterygoid bones, 

 lu a number of genera these bones meet in the median line of the palate, 

 while in others they are widely separate. The value of this distinction 

 is, however, diminished by the fact that in some species of Lageno- 

 rhynchus these bones are in contact, while in otliers they are widely 

 divergent; also by the fact that the two positions appear to occur in 

 some species, e.g., Sotalia gadamu, as an individual variation. Within 

 certain limitation, however, the character is apparently of much value. 



Characters and Divisions, pp. 504 and 511. 



