NO. 3 TELEOSTEAN FISHES — GOSLINE 23 



scribed in the eels by Allis (1903) ; this will not be discussed here. 

 Very different canal structures occur in the snouts of the albulids and 

 halosaurids. These will be described below. 



At least some of the difficulties with the canal bones of the head in 

 Albula and Pterothrissus undoubtedly arise in relation to the cavern- 

 ous trough system of lateral line canals in these fishes. Many deep- 

 water fishes, but also some shallow-water and even some fresh-water 

 forms, have such a system of troughs replacing the usual sensory 

 canals of the head. Such troughs differ from the usual tubes in a 

 number of ways. In the first place the external surface of these 

 troughs is covered by a layer of thin, taut skin. In the second, such 

 a system may have a large number of minute openings to the exterior, 

 but it lacks the usual type of large (primary) sensory canal pores. 

 Whatever the significance of these differences as far as function is 

 concerned, it seems certain that a trough system shows considerable 

 variation as compared to the relatively stable normal canals. Con- 

 nections may be established between some parts of a cavernous sys- 

 tem and membranous closures may be developed between others. 

 Sometimes, e.g., in Halosauropsis, there seems to be a separate ossicle 

 developed below each lateral Hne sense organ (neuromast). With 

 this background, which is based on an examination of such diverse 

 fishes as engraulids, notopterids, macrourids, Acerina, and brotulids 

 (see also Garman, 1899; Pf filler, 1914; Omarkhan, 1949), the 

 troughs in the snout region of Albula, Pterothrissus, and Halosaurop- 

 sis will be dealt with. 



In Albula the supraorbital canal of each side passes forward over a 

 nasal and a prenasal ossicle to end blindly at the border of the snout. 

 There is also a rather extensive median cavity in the tip of the 

 mesethmoid. Though this cavity extends laterally into the flesh on 

 either side of the mesethmoid, it does not appear to connect with the 

 supraorbital or infraorbital sensory canals. Whether this cavity 

 represents the ethmoidal commissure of Elops, the peculiar median 

 sensory system of eels described by Allis (1903), or whether it is 

 even a part of the sensory canal system is not clear to the present 

 author. In the related Pterothrissus (fig. 7), the supraorbital canal 

 extends forward to the snout rim, down which it extends for a short 

 distance before ending blindly. Near its anterior end it is in open 

 connection with its fellow on the other side via a hole through the 

 mesethmoid ; on the floor of this transverse opening there is a well- 

 developed neuromast. Other than this opening, there is no median 

 cavity in the mesethmoid. 



