But this may be accounted for by the topography of the bay. Tomales Bay is a long and narrow channel, about twenty miles long and from one to five in width, isolated from other rivers and with but one tributary stream. Probably the salmon had not wandered far from it; some may not have left it at all. In any event, a large number certainly came back to the same place. That the salmon rarely go far away is fairly attested. Schools of king-salmon play in Monterey Bay, and chase the herring about in the channels of southeastern Alaska. A few years since Captain J. F. Moser, in charge of the _Albatross_, set gill-nets for salmon at various places in the sea off the Oregon and Washington coast, catching none except in the bays. Mr. Davis gives an account of the liberation of salmon in Chinook River, which flows into the Columbia at Baker's Bay: "It is a small, sluggish stream and has never been frequented by Chinook salmon, although considerable numbers of silver and dog salmon enter it late in the fall. A few years ago the State established a hatchery on this stream, and since 1898 between 1,000,000 and 2,000,000 Chinook fry have been turned out here annually. The fish are taken from the pound-nets in Baker's Bay, towed into the river in crates and then liberated above the dike, which prevents their return to the Columbia. When ripe the salmon ascend to the hatchery, some two or three miles farther up the river, where they are spawned. "The superintendent of the hatchery, Mr. Hansen, informs me that in 1902, during November and December, quite a number of Chinook salmon ascended the Chinook River. About 150 salmon of both sexes were taken in a trap located in the river about four miles from its mouth. At first thought it would appear that these were probably fish which, when fry, had been liberated in the river, but unfortunately there is no proof that this was the case. According to Mr. Hansen, the season of 1902 was remarkable in that the salmon ran inshore in large schools, a thing which they had not done before for years. It is possible that the fish, being forced in close to the shore, came in contact with the current from the Chinook River, which, since the stream is small and sluggish, would not be felt far from shore. Once brought under the influence of the current from the river, the salmon would naturally ascend that stream, whether they had been hatched there or not." The general conclusion, apparently warranted by the facts at hand, is that salmon, for the most part, do not go to a great distance from the stream in which they are hatched, that most of them return to the streams of the same region, a majority to the parent stream, but that there is no evidence that they choose the parental spawning grounds in preference to any other, and none that they will prefer an undesirable stream to a favorable one for the reason that they happen to have been hatched in the former.