THE SALMON I9t 
of salmon have been the subject of gross exaggeration, owing 
to one writer blindly repeating the statements of another. 
Thus it is certain that Professor Seeley, usually a conscientious 
witness, was not speaking from his own observation when 
he described the limit of their perpendicular spring as “ about 
twelve or fourteen feet.” I have watched hundreds—thousands 
—of salmon and grilse leaping at falls, and I agree with 
Dr. Day that “a clear jump of six feet’ is the maximum 
of a salmon’s performance. A sheer fall of seven feet is 
an effectual bar to the ascent of salmon. 
The question whether salmon feed in fresh water is 
the nucleus of a controversy which there seems no prospect 
= of bringing to an early conclusion. On the one 
o Salmon Z i 
feedin hand is the belief of many anglers that they do 
fresh Bee oe supported by the fact that salmon take not 
only artificial flies, cunningly but arbitrarily fashioned to 
meet their hypothetical taste, but worms, prawns, and 
minnows. On the other hand there is the all but universal 
failure to discover food in the stomach, or traces of food 
in the bowel, of salmon taken in fresh water. Those rare 
instances in which the remains of food have been so found, 
have occurred only in the examination of kelts—that is, 
salmon returning to the sea after spawning. Yet if the 
voracity with which the hungry aspect of kelt salmon has 
caused them to be credited were well founded, such remains 
of food ought to be present in almost every kelt examined. 
The well-recognised phenomenon of the absence of all trace 
of food in the stomach and intestines of salmon taken in 
fresh water has led persons who decline to believe in a 
physiological fast to account for it by a purely hypothetical 
power inherent in the salmon of ejecting the contents of 
both stomach and intestines immediately upon being hooked 
or netted. Needless to say that such an explanation, in 
the complete absence of any evidence to support it, is worth 
no more consideration than ought to be shown to any a priori 
