THE SALMON 189 
that salmon leaving the sea are not always acting under the 
reproductive impulse. 
This is no matter for dogma, but the explanation of 
the periodical movement of salmon which most nearly 
corresponds with my own observation is as follows. 
Like many of the lower vertebrate animals, especially 
fish, salmon manifest intermittent periods of appetite during 
which growth proceeds, and of abstinence, during which 
growth is suspended.* The salmon feeds and grows without 
intermission through the stages of alevin, parr, and smolt, 
and continues to do so in the sea in its transition to the 
grilse stage. Under the stimulus of unlimited nutrition, the 
fish rapidly increases in bulk, until the time comes when its 
tissues are so fully stored with fats and proteids, that no more 
nourishment can be assimilated, appetite ceases, and the fish 
returns to its native river. There is no evidence to connect 
the return of grilse in May, five or six months before the 
spawning season, with the sexual impulse. It appears probable 
that, the purpose with which the fish visited the sea having 
been accomplished, it simply returns home, i.¢., to its native 
fresh water, which is its most congenial environment. After 
an uncertain period, appetite may return, and the fish obeys 
the impulse to seek again the salt water, teeming with various 
forms of life offering an abundance which cannot be found 
in the river. 
* Asan extreme instance of periodical fasting and arrest of growth, 
the so-called lepidosiren (Protopterus annectens) may be cited. This fish, 
which belongs to the Ganoid Order, abounds in many rivers of tropical 
Africa. These rivers usually disappear in the dry season, when the fish 
bury themselves in the mud, lining their resting-place with a protecting 
case of mucus. The mud-balls thus formed may be dug out and have 
been brought to Europe, where the fish inside can be released by 
immersing their domicile in tepid water. Here it is obvious that there 
must be absolute abstinence from food for long periods, and the rigid 
envelope renders the corresponding cessation of growth a necessary 
consequence of the situation. 
