264 PROCEEDINGS OF THE CANADIAN INSTITUTE, 
sequently the direction of the canals can only be approximately 
determined by the study of the surface. 
Within recent years the study of the course of the mucous canals 
has received an impetus from the discovery of their relation to the 
morphology of the skull, and accordingly it will be found detailed 
in Prof. McMurrich’s paper on the osteology of this species. 
The canals in the head vary considerably in their dimensions ; 
their diameter is on the whole greater, sometimes twice as great as 
that of the lateral canal, and their walls are different in so far as 
the protective canal is formed of true osseous substance throughout. 
Except in respect to the greater size of the nerve-hillocks, the lining 
epithelium appears to be very similar. A transverse section through 
a nerve-hillock from a young specimen is represented in Fig. 7. The 
upper half of the tube is occupied by the ordinary epithelium, which 
becomes thicker as it approaches the neuro-epithelium, projecting 
inwards so as to lessen the cavity at this place. Two kinds of cells 
are to be distinguished in the neuro-epithelium : sensory cells, short 
and oblong, occupying the inner half of the height of the epithelium, 
and indifferent cells (Stuetzzellen) occupying the whole height with a 
basal nucleus. The latter are more frequent at the point of passage 
into the ordinary epithelium. Fig. 8 represents a section of a macula 
acustica from a fish of the same age, drawn under similar conditions ; 
the resemblance of the two kinds of neuro-epithelia is particularly 
striking. In Fig. 7 the whole height of the neuro-epithelium is 37 p, 
of the sensory cells 15:5 »; the nuclei of these are 6:5 yp, of the 
indifferent cells 4:5 ». The latter stain very densely in carmine, con- 
trasting with those of the sensory cells in this respect. Here and 
there between the indifferent cells are structures which are possibly 
nerve fibres in section. 
To return to the course of the canals in the head. It will be 
observed from Fig. 6 that the lateral line rises as it passes forwards 
towards the posterior upper angle of the gill-cover. Before reaching 
that a short tube is given off which opens in the skin over the 
ascending process of the supraclavicle. Directly over the posterior 
upper angle of the gill-cover is another pore (Figs. 4 and 6) and in 
front of that another. At the plane of the latter the canals of the 
two sides communicate by the ‘occipital commissure,’ which again 
has two apertures near the middle line. The canal proceeds forwards 
from this plane, and again opens by a short tube over the articula- 
