FISH — MIGRATIONS. 
249 
which its anal fin is provided ; it then pushes itself up- 
wards by straightening the tail, while it closes the gill- 
covers not to prevent progress, and so on. Sir E. Tennent, 
however, without disputing the evidence that these fish 
do climb trees, says, — 
The probability is, as suggested by Buchanan, that the 
ascent which was witnessed by Daldorff was accidental, and 
ought not to be regarded as the habit of the animal. 1 
A great number of species of fish perform migrations. 
In relation to intelligence, the most interesting of these 
is the migration of salmon, which annually leave the sea 
to spawn in rivers, though there is some doubt whether 
the same individuals spawn every year. There is no doubt, 
however, that the same individuals frequently, though not 
invariably, revisit the same rivers for their successive 
spawnings. This fact may be due either to the remem- 
brance of locality, similar to that wdiich is unquestionably 
manifested by birds, or to the salmon not swimming far 
along the coast during other seasons of the year, and there- 
fore in the spawning season when seeking a river happen- 
ing to hit upon the same one. The latter hypothesis is 
one which Mr. Herbert Spencer tells me he is inclined to 
adopt, and, being a salmon-fisher, he has paid attention to 
the subject. He informs me of an observation by a friend 
of his own, who saw a salmon, when about to spawn, swim- 
ming along the coast-line, and all round a boathouse, 
apparently seeking any stream, that it might first en- 
counter. 
The distances up rivers to which salmon will swim in 
the spawning season is no less surprising than the energy 
with which they perform the feat, and the determination 
with which they overcome all obstacles. They reach 
Bohemia by the Elbe and Switzerland by the Rhine. On 
encountering a waterfall they display astonishing agility 
and perseverance in surmounting the obstacle. This fact, 
of course, is well known to all salmon-fishers ; but the 
actual vertical height to which a vrell-grown salmon is 
able to leap has only recently (1886) been made the 
subject of exact measurement. By means of upright 
1 Natural History of Ceylon , p. 351. 
