RESULTS OF SALMON MARKING 87 
can scarcely be caught in the small streams of the 
West Highlands, so quickly do they get away to the 
sea. Ina, large river, on the other hand, or in any 
river which broadens out into a large lake or into 
deep pools, the descent of female kelts at least is 
very slow indeed. A large number of recaptures of 
such descending kelts have been made in the Spey 
and Tay, and it would appear that so long a time as 
three months may be occupied in descending forty 
miles of river. Grilse kelts, as already stated, descend 
to the sea with some promptitude, but as salmon 
grow older their descent after spawning becomes 
slower, although this appears to be less true of males 
than of females. Large female kelts linger longest. 
The discovery of this habit makes a considerable 
difference in our interpretation of the increments 
put on in the sea. One fish may relatively be so 
much longer in reaching the sea than another even 
if both be, say, short period fish. In judging of in- 
crease from kelt weight this uncertainty of knowing 
when the sea was entered is always present. We 
may have in long period fish considerable uniformity 
of increase, since an actual delay at starting may in 
the longer period be made up. For instance, in 
Helmsdale long period fish of 480, 498, 504, and 506 
days we find increments respectively of 9, 9, 9, and 
10 lb. Yet clearly there may also be great variation. 
In short period fish we have similar results. Two 
Brora fish were taken, the one in 162 and the other 
in 163 days: the one had added 4 lb., and the other 
11 1b. Many Tay fish recaptured are described as 
‘‘bull-trout.” In the Tay a bull-trout is in reality a 
