10 INTRODUCTION. 



ouly the author's bare stateiueut that the external characters of the 

 individual whose skeleton is described were identical with those of a 

 previonsly-described species. 



In the case of species founded upon single skulls, absolute certainty 

 as to theii* distinctness can be reached only when large series of indi- 

 viduals known to be alike in their external and skeletal characters shall 

 havebeen acquired. When such series shall be at command, the limits 

 of specific variation can be determined with accuracy, and it will be pos- 

 sible to judge whether the characters held out as distinguishing the 

 species in question are really of specific value or only represent such 

 variations as are common among individuals of the same species. In 

 the mean time it is only possible in many cases to form opinions which 

 may or may not coincide with the truth. 



In this, as in all other families of animals, an arrangement of the 

 genera in a single linear series does violence to their natural afiinitios, 

 while the attempt to introduce subfamily distinctions, with a view of 

 approximating the arrangement more closely to a natural sequence, is 

 here attended with great difficulties. Dr. Gill * has recognized four sub- 

 families: PoHtoporiincv, Ddplimapterinci', Belphiabuv^ixwA Globioccphali- 

 1103. The genus Pontoporia {=Po7ito2)oriin(e) I do not regard as belonging 

 to the DeJphinirlae, and shall, therefore, omit all further reference to it. 

 ThGGlohinccpltaJincc {=Glohic€2)halus and Gram2)us) are characterized as 

 having'^digits (second and third) segmented into numerous phalanges," 

 and to this are opposed the DelpMnapterincc and DeJpMninw, which have 

 "digits (second and third) not segmented into more than h-\j phalanges 

 each." The facts do not appear to warrant this distinction, since Bel- 

 phinus deJphis commonly has from seven to nine phalanges in the sec- 

 ond digit, and Tur slops tursio and other species seven phalanges, which 

 figures also represent the number of phalanges in the second digit of 

 Grampus. 



The character which Dr. Gill employs for the separation of the Bel- 

 pMninw from the Belpli'inapterinw seems to me to be of much greater im- 

 portance. This relates to the condition of the cervical vertebrae In 

 M^>nodon and Belphinapterus {=Belphinapterince) the cervicals are all 

 distinct, while in the other genera of the family they are more or less 

 consolidated. I should be inclined, therefore, to unite Dr. Gill's Bel- 

 phinime and Glohiocephalince under the former name, and to oppose to 

 them the Belpldnapterince as a second subfamily. I am the more in- 

 clined toward the adoption of this division on account of having dis- 

 covered a character, which, in addition to that of the separate cervicals, 

 is common to Monodon and Belphinapterus, but wanting in the other 

 genera. This is that in the narwhal and white whale the pterygoid 

 bones, instead of merely forming the walls of the posterior nares, extend 

 backward in the form of broad plates across the optic canal and articu- 

 late with the squamosals. 



*Gill. Arrangement of the Families of Mammals, 1872, p. 95. 



