In the case of sea-fishes it is becoming increasingly recognized that the millions of cod fry which are annually turned out of the American, Newfoundland and Norwegian hatcheries are but an insignificant fraction of the billions of fry naturally produced. A single female cod liberates, according to its size, from one to five million eggs in a single season. Yet the annual output of fry from each of these hatcheries rarely exceeds 200,000,000, i.e., the natural product of a few hundred cod at most. In Britain, marine hatcheries have been established by the Fishery Board for Scotland in the Bay of Nigg, near Aberdeen, by the Lancashire Sea Fisheries Committee at Peel, and by the Government of the Isle of Man at Port Erin. These establishments have been prin cipally devoted to the hatching of the eggs of plaice. But again the maximum output of fry from any one of these establishments has not exceeded 5o,000,000 in any single year. As a single female plaice produces about 200,000 eggs per annum, this output does not exceed the natural produce of a few hundred fish. Under these circumstances the probable utility of the operations could be admitted only if the fry were sedentary and could be planted in suitable localities where young fish were naturally scarce. But the fry drift with the currents as helplessly as the eggs, and the a priori objections to the utility of the operations have in no case been met by satisfactory evidence of tangible results.
The Scottish Fishery Board carried out a large scale experiment from 1896 to 1908 which was very illuminating (Fulton, 26th Report). They attempted to measure the density of young plaice on the beaches of Loch Fyne during six years when plaice fry were liberated in the loch and a similar period when none were added. From 1896 to 19o1 the quantities of fry added were 4, 21, 19, 16, 3o and 51 millions annually, the number of eggs naturally present in the loch being estimated at nearly soo,000,000 in 1898 (Wil liamson, 17th Report). Although the figures of young plaice on the beaches yielded an average density which was twice as great in the first period as in the second (88:4o per hour), there was little or no correspondence from year to year between the density of young fish and the number of fry planted. The highest record corresponded with the year of maximum output of fry (190I), but the next highest occurred in 1905, when no fry were planted, and the third occurred in the first year (1896), when the number planted was negligible (4,000,000). Assuming (which is not cer tain) that the coincidence in 1901 was significant, it nevertheless appears that the natural fluctuations in the number of young plaice were unaffected by the hatching operations except in the one year when the number of fry liberated exceeded 5o,00o,000 and thus formed an appreciable proportion (1o%) of the number estimated to be naturally present. Even these numbers do not provide Loch Fyne with a regular plaice fishery.
In the United States the utility of the cod-hatching operations has, in the past, been often asserted by representatives of the Bureau of Fisheries, but practically the only evidence adduced is the occasional appearance of unusual numbers of cod in the neighbourhood. It has not been established that the fluctuations
in the local cod fisheries bear any relation to the extent of the hatching operations, while the earlier reports of the commissioner of fisheries contain evidence that similar fluctuations occurred before the hatching of "Fish Commission Cod" began.
The situation is much the same with regard to the operations of the Newfoundland hatcheries at Dildo. (See Fryer, Reports x.—xii. of H.M. inspectors of sea fisheries, E. and W.) The most exact investigations bearing upon this problem are those undertaken in Norway in connection with the cod-hatching operations at Arendal under Captain Dannevig. Four fjords were selected on the south coast of Norway in proximity to the hatch ery, and the usual number of fry (io to 3o millions) was planted in the spring in alternate fjords, leaving the intermediate fjords unsupplied. The relative number of young cod in the various fjords was then carefully investigated throughout the succeeding summer and autumn months. It was found that there was no relation between the abundance of young fish and the presence or absence of "artificial" fry. In 1904, 33,000,00o fry were planted in Sondelfjord and young fish were exceptionally abun dant in the following autumn (three times as abundant as in 1903, when no fry were planted). But their abundance was equally striking in other fjords in which no fry had been planted, while in 1905 all the fjords were deficient in young cod whether they had been planted with fry from the hatchery or not. (For a summary of these investigations see Papers on "Artificial Fish hatching in Norway" by Captain Dannevig; and Dahl, Report of the Lancashire Sea Fisheries laboratory for 1906, Liverpool, 1907.) It would seem clear that the attempts hitherto made to increase the supply of sea-fish by artificial hatching alone have been unsuccessful. The experience gained has doubtless not been wasted, but the direction to be taken by future work is plain. The energy and money devoted to hatching operations should be diverted to discover a means of rearing on a large scale the just-hatched fry of the more sedentary species to a sturdy adolescence. When that has been done it will be possible to deposit the young fish in suitable localities, on a large scale, with a reasonable prospect of influencing the local abundance of the species of fish in question. The increase of the supply of the migratory marine species, such as the cod, depends on a more complete knowledge of habits, life histories and migrations. With this knowledge in hand these fisheries may be maintained or in creased either through regulation or by artificial propagation or by both means; until it is in hand the effect of artificial propaga tion cannot be estimated.