352 DEEP SEA FISHES. 



For coufirmfvtorj instances on bony fishes see Plates LXXI. to LXXXIV. of 

 the present work, which represent forms on which the aurals are not con- 

 necting and on which subcephalic connections between the system on the 

 two sides of the head are generally absent. The facts that the neuration 

 of the system on each side of the body proceeds from its own side of the 

 brain developing from near the brain to the farther portions in the early 

 stages, and that oral, jugular, and, in bony fishes, aural connections arc 

 somewhat rare, indicate rather conclusively that on ancestors of the fish- 

 like vertebrates the lateral system was in two parts, one on each side of the 

 middle of the head and the body. Exceptional instances of transverse 

 connections in the system are to be seen in the Chima^rans, Lat. Canal Syst. 

 Plates III. and IV. figs. 2 and 3, where the canals are highly differen- 

 tiated, also on the greatly specialized Ilalieutoids on which the oral is 

 continuous from one side to the other, Plates XVIII. to XXV. below, and 

 on Chaunax, Plate LXXIII. fig. 1, which apparently has a transverse sub- 

 mental series of nerve papilla3. 



So far as the Plagiostomes are concerned the intention at this writing is 

 merely to compare the system in the several species figured on Plates 

 LXIX. and LXX. The distribution of the canals on Ceiitrosci/lHuni nignim, 

 Plate LXIX. fig. 1, and on Isistius hrasiUcnsis, Plate LXIX. fig. 2, approaches 

 that of the simplest arrangement obtaining among the sharks (Antacea). 

 Excepting in regard to slight differences in directions and curvatures the 

 canals of the mentioned species are similar to one another, with the further 

 exception perhaps of tlie division of the aural (an) on Centroscyllium, a 

 division which may or may not be a peculiarity of the individual corre- 

 sponding to that noticed above as occurring on ILptahranchias vincnlatus. 

 In both Centroscyllium and Isistius the median {m) is short and longitu- 

 dinal, but on Isistius oral (o) and angular {ang) are more elongate than on 

 the other. Neither of them possesses a jugular (/), a gular {g), nor a spira- 

 cular {sp), as seen on C/tIam//doselacJivs aiiguineus, Plate LXX., which see for 

 the lettering. On the last mentioned the system was originally traced as 

 indicated by the outer openings of the tubules leading from the tubes (Lat. 

 Canal Syst., Plate XV.) ; on specimens obtained subsequently the canals them- 

 selves have been followed and sketched, Plate LXX., with a result corre- 

 sponding nearly to the arrangement in the diagram first published. The 

 median canal {m) proved to be transverse, in this particular agreeing with 

 PrioHotldii Milhcrll M. H. and with Alopias viilpes Gmel. The functions of 



