360 DEEP SEA FISHES. 



groups. Collinge and others also have made attempts in similar directions 

 placing the stress mainly upon the innervation. Though the distribution of 

 the nerves can be used to advantage in connection with the higher divisions, 

 in an approach to the species and varieties it becomes less practical than the 

 more obvious features, the arrangement and special characters of the system 

 itself. It is in the numbers of the disks, their distribution, and the compar- 

 ative sizes and degrees of development that the most effective aids to classi- 

 fication are available ; it is by means of these that most light is shed upon 

 tlie closer and more recent affinities among deep sea fishes. 



In the following notes attention is directed to a few of the more obvious 

 special features of each of twenty-six species of Telcosts from which outlines 

 of the cephalic portions of the system are j^resented. 



Ectreposcbastes imiis, Cottnncidus Thomsonii, and Iloplostctlws paa'ficus of 

 Plate LXXI. illustrate three types of the canals on the cheek behind the 

 eye : in the first the orbital and the spiracular are reduced to a single canal, 

 in the third they are distinct though tolerably close together, and in the 

 second they appear to be partially reduced. The aural branches of the first 

 and the second are transverse tliough they may not unite across the occiput, 

 but on the third they evidently nnite with the frontal branches of the 

 cranials and form a loop, as in Lamprogrammus. The disks are small and 

 nearly uniform in size in each case ; they are more developed than those of 

 the shoalwater allies ; on the head E. imiis has 52, C. Thomsonii has 50, and 

 H. pacificus has 70. 



Caulolcpls suhuUdcns and Melamphacs nigrofidvus, Plate LXXIL, though 

 differing in details, show considerable evidence of relationship. The post- 

 orbital and spiracular branches of the canals are distinct; the frontal 

 branches and the aurals are similar, but the latter bear two disks on Melam- 

 phacs and only one on Caulolepis. The probability is in favor of an aural 

 commissure on the occiput on both forms, and it may be the frontal 

 branches are connected with the aurals ; these connections have not yet 

 been made out. The disks of Caulolepis appear to be rather more complex 

 than those of Melamphacs ; C. suhuUdens has 72 disks on the head and 

 M. nifjrofulmts has 68. 



Chaunax coloraim and Lcpophldhmi emmeJas, Plate LXXIII., possess very 

 different developments of the lateral system. The first represents the 

 pediculates ; it shows postorbital and spiracular as a single series, the spira- 

 cular from its position, and there appears to be an angular and a jugular 



