﻿44 6 



all these, which forms this connection in the Gastrosteidcv , Aiilorynchidœ and 

 Pegasidœ, thus exactly as in the Scleroparei). In none of the genera does the 

 pieorbital contain any canal for the lateral line; on the other hand the lateral 

 line passes through the nasal in Amphisile and Centriscus. 



Structure of the Genera Amphisile and Centriscus. 



The grouping together of Amphisile and Centriscus is very old. Linné as is 

 known placed in the same genus Centriscus the species C. scutatus {=■- A. scutata) 

 and C. scolopax (Syst. nat. Ed. XII, p. 415); later ichthyologists have however mostly 

 endeavoured to emphasize the differences and to remove the two forms as far as 

 possible from one another. They have rightly maintained two genera, each with 

 its species; in recent limes these have again been raised to Ihe level of families, 

 2 in number, each with 2 genera. Too much concentration on the fairly obvious 

 differences, which for a great pari express themselves quite externally, seems however 

 to have led observers to forget or to overlook the essential resemblances which 

 really exist. These seem to me so considerable that I consider it right to maintain 

 one family, Centriscidce, with only two genera Amphisile and Centriscus*. Of the 

 former, I know the species A. scutata L. (Gthr.), strigata Gthr. and punctulata Bian- 

 coni; of the latter, C. scolopax L., gracilis Lowe and humerosus Richards. I shall 

 not discuss here the validity of the other supposed species of both genera, but I 

 am most inclined to believe that they are not maintainable; nor does the division 

 of each of the genera into 2 separate genera seem to me valid, but this question 

 also will not be discussed. 



In first dealing with Amphisile and thereafter Centriscus in the following pages, 

 I have no intention naturally of giving the impression that 1 consider Amphisile 

 the more primitive and simpler form — the opposite is indeed the case — but my 

 investigations began with Amphisile, which seemed to me from Günther's descrip- 

 tion somewhat enigmatical and to have been on the whole curiously dealt with; 

 it was only later that I took up Centriscus and saw clearlj' how much was insuf- 

 ficiently known in this form also, and how much in it threw light on the condi- 

 tions in Amphisile. 



Amphisile. 



In appearance Amphisile is quite remarkable. The body is extremely com- 

 pressed; the diameter at the broadest part of the trunk, as Günther remarks, is 

 scarcely greater than the diameter of the orbit. The ventral edge is as sharp as a 

 razor. The head is produced into a long tube, with a small terminal, toothless 

 mouth, as in a pipe-fish. The trunk runs out posteriorly into a long spine, under 

 which is seen two dorsal fins and a quite short caudal fin directed obliquely 



* I disregard here the objections, which might be raised witli a certain amount of right, against 

 using this generic name for the species scolopax, gracilis and humerosus. In using the names the 

 main thing for me, here as elsewliere, is tliat there can be no doubt what forms are being discussed. 



