134 BULLETIN OF THE 
gar-pike; this process gives a satisfactory explanation for the feebly 
developed condition of the scales upon the under side of the lower jaw. 
He observed that in the act of deglutition the floor of the mouth became 
very much distended, so much so as to lead him to compare it to the 
pouch of a feeding pelican. The frequent distention to which this part 
of the skin is subjected by the greedy habits of the young fish furnishes 
a sufficient reason why continuous plates of hard bony material are not 
formed in this part of the dermis. The small size of the scales described 
by Hertwig (79, p. 9) as lying at the bases of the fins is doubtless to be 
explained in a similar manner. Instead of being atavistic conditions, as 
maintained by him, they would appear to be due entirely to the action 
of purely mechanical influences. 
I have already spoken of the canals which pierce the central part of 
the scale, and which have been called Haversian canals by Hertwig and 
Klaatsch. Their usual course is from below directly through the scale 
to the upper surface, where the vessels which traverse them spread out 
in all directions through the sub-epidermal layer. There are sometimes 
one or two vessels of considerable size which at the upper surface break 
up into small branches, but quite as often the canal is filled by a large 
number of minute vessels apparently distinct from one another. The 
course of these canals is not, however, by any means constant. I have 
found that sometimes, instead of running directly through, they extend 
for some distance inside the scale parallel with the surface. Indeed, in 
the scales of one fish (287 mm. long) the rule seemed to be for quite a 
number of canals to run horizontally through the scales for considerable 
distances. This condition seems to me to give an additional reason for 
regarding these canals as Haversian, or perhaps better — since the 
osteoblasts are not arranged in concentric lamelle about them — as 
Volkmann’s canals. Figures 28 and 29 (Plate IV.) are from sections of 
such scales. In the former figure the canal extends the greater part of 
the length of the scale and opens to the surface at five points, near 
those marked jfor., as can be seen in other sections not figured. 
The surface contour shows some quite marked peculiarities in several 
instances. Figure 30 (Plate IV.) represents a portion of a scale from 
the same fish as the one last mentioned (287 mm. long). The upper 
surface bears numerous rounded elevations or hillocks, upon which for 
the most part the spines are borne. The variations in shape and height 
are indicated in the figure. Frequently the canals pierce the scale just 
beneath these elevations and open to the upper surface at one side, as 
indicated by dotted lines. These slightly abnormal conditions seem to 
