THE SENSORY CANAL SYSTEM OF FISHES. Silk? 
a. The branch a@ is the most dorsal one (PI. 40, fig. 11, a.), 
and gives off innumerable fine branches from either side, which 
pass to the combined supra- and sub-orbital branch of the main 
canal. In the space of 26 mm. fifty-three branches were 
counted and traced to the canal. In Lemargus, Ewart (20) 
states that there are over 1500 twigs given off from the facial 
nerve to the sensory and ampullary canals. In Polyodon 
there must be 2000 to 3000. In the anterior portion the 
branch further divides into smaller branches, which terminate 
either in primitive pores or the anterior portion of the canal. 
6. This is only a small branch passing along the rostrum for 
a little more than half its length. It supplies the primitive 
pores on the dorsal region of the rostrum. 
ce and d. These two branches supply the primitive pores on 
the lateral and ventral portion of the rostrum; most of the 
branches from d pass to the two lateral branches given off from 
the combined supra- and sub-orbital branch of the sensory 
canal. 
e. There is a small branch passing off from the ramus buc- 
calis posterior to the orbit, labelled e (PI. 40, fig. 12). It 
innervates the circum-orbital series of primitive pores. 
Van Wijhe figures the ramus buccalis as a single branch 
passing directly forwards beneath the olfactory capsule (Pl. 40, 
fig. 9, r.m.8.). 
3. The ramus mandibularis—Van Wijhe’s description and 
figure do not at all correspond to the condition I have found in 
Polyodon. He figures a ramus mandibularis passing across 
the centre of the hyomandibular bone, a condition common to 
the Teleostei, but not at all correct for this fish. In his figure 
(Pl. 40, fig. 9) it is at once evident that he has the hyomandi- 
bular at a wrong angle. Supposing his figure were otherwise 
correct, which it is not, it would appear in the position I have 
shown it in in fig. 10,4. m. Matters are further complicated by 
his figuring the ramus hyoideus as a branch of the ramus mandi- 
bularis (fig. 9,7..). This latter is shown as passing beneath 
the suspensorio-quadratum ligament, and is described as divid- 
_ ing into an internal and external branch. He speaks of the 
