72 
BULLETIN OF THE UNITED STATES FISH COMMISSION. 
the larger streams, though not necessarily those emptying through estuaries. The 
other two species are of little importance. They spawn in any kind of stream, 
frequently in mere brooks that empty directly into the ocean. 
The Pacific salmon, of whatever species, passes most of its life in the ocean, and 
upon arriving at maturity ascends the rivers to spawn. Sometimes enormous 
numbers pass upstream together; stories of their being “ thick enough to walk 
across on” are often told, but I have never seen them quite so numerous as that. 
The upstream migration occurs sometime during the warmer half of the year, 
the earlier fish going farther upstream. Spawning occurs late in the summer or 
during the fall, and as soon as it is completed the salmon dies. 
Most of the eggs deposited are eaten by other fishes, or are killed by being cov- 
ered with sand and gravel. Those not destroyed hatch in from seven to ten weeks, 
according to the temperature of the water. In the cold waters of Alaska they are 
four or five months in hatching. It requires about six weeks more for the yolk-sac 
to be absorbed, when the fry are able to swim and are ready for their seaward 
migration. Most of the alevins, however, are devoured by other fishes before they 
are able to swim. It is to prevent this great mortality among eggs and alevins that 
artificial propagation lias been employed. 
The young salmon start downstream as soon as they are able to swim, and 
reach brackish water when three to five months old, according to the distance they 
have to go. Those from the vicinity of Battle Creek hatchery reach Benicia in 
about seven weeks. It is not known when they reach the ocean, but probably soon 
after. The variation in the time of spawning and hatching makes the period of 
migration very long. The fry from the summer run begin passing Battle Creek in 
September, and from that time until April following there is a continuous stream of 
young salmon, about I t inches long, passing that point. 
Although a large majority go downstream as soon as they can swim, many, 
especially those hatched in the spring, remain in the pools in the headwaters all 
summer and fall. There were estimated to be from 700 to 1,000 in each of several 
pools in the vicinity of Sims during the summer of 1898, and there were probably as 
many as 10,000 to the mile in that portion of the river. These remained in the 
headwaters until the first of December, when rains caused a rise in the river. 
The food of young salmon in fresh water, at all times, places, and ages, consists 
of insects, either larval or adult. 
SALMON EGGS AND MILT. 
The following notes embody observations and experiments made at Battle Creek 
hatchery in 1897 and 1898. Although dealing largely with artificial propagation, 
they are not intended to give even a general account of the methods of fish-culture 
as applied to the Pacific salmon. For such an account reference is made to the 
Manual of Fish-culture issued by the United States Fish Commission.® 
For facilities placed at my disposal I am under obligations to Mr. G. H. Lambson, 
superintendent of the station. Special acknowledgments are due to Mr. William 
Shebley, superintendent of Sisson hatchery, and to Mr. Robert Radeliff, of Baird, 
who had charge of the spawn-taking operations in 1897 and 1898, respectively, and 
heartily cooperated in the experimental work. 
" Report United States Fish Commission 1897, pages 1-340. 
