Behavioural and physiological mechanisms postulated for the control of downstream migration of Atlantic salmon smolts are reviewed briefly, and some new evidence is presented for their refusal to undergo sustained swimming. Although these mechanisms imply passive displacement as the primary means of emigration, it is likely that active components must also exist as the rates of travel of smolts through loch systems are only slightly slower than those recorded for river systems. The timing of these movements within 24 h periods is reviewed and it is shown that the predominantly nocturnal emigration pattern is evident on occasions in alevin, fry and parr stages also. Thus at migration the die1 periodicity probably represents a seasonal locomotor rhythm which, under changed behavioural and physiological circumstances, results in downstream displacement.

No abstract

Nocturnal downstrean migration of juvenile Atlantic salmon is usually interpreted as increased locomotor activity. The frequency of downstream passages ofO-l+ salmon in an endless stream channel was greater by night than by day in both smolting and non-smolting fish in autumn and spring. Movement increased at dusk, and decreased after dawn. Mature male 1 + fish moved slightly less than immatures in October, but significantly more in November. Total movement frequency was lower at full moon than at other moon phases, and movement was reduced when the moon was up. Under turbid conditions by day, the threshold water velocity inducing nett downstreammovement was 8.2 cm s-', and the relativevelocity offish swimmingdownstream was never more than one third that of fish holding station at the normal maximal flow o f 2 M O cm s-'.At the end of their first growing season in October, fish which had been offered food continuously through 24 h did not differ in size from those fed by day only, but the latter were significantly larger than those offered food only at night. We conclude that: (1) the fish fed actively by day, and not by night; (2) station-holding represented activity, and downstream nocturnal movement represented relative inactivity (displacement) which occurred on loss of visual orientation, hence migration resulted from reduced activity; (3) lack of displacement in early autumn has adaptive value for maturing fish, but not for non-spawners.

Sexual maturity of parr reduces the probability of a future seaward migration. In release experiments in two separate years with Atlantic salmon in the Imsa River, Norway, immatures migrated sooner and in significantly higher proportions (P < 0.001) than did previously mature males. Furthermore, higher proportions of 2-yr olds than of 1-yr olds migrated, and 86–92% of the descent occurred at night. Large 2-yr olds migrated before smaller ones. Among those which did not migrate, some (3.2% of those released in 1986) were recaptured in the autumn, of which 91.9% were mature males. At Lussa, Scotland, 5.6 and 5.9% of smolting fish released in two separate years remained resident at the release site throughout the summer, and 91.8 and 93.4% of these matured in the autumn of the release year.

The attachment of dummy, ultrasonic, acoustic tags, of the Stirling Mk. VI design, to pre-smolt Atlantic salmon parr significantly affected fish growth rates. Fish less than 160 mm in length lost weight and showed no change in length. Fish over 160 mm in length put on weight, though at a rate significantly lower than that of untagged controls. For fish between I 6 0 and 180 mm in length, the increase in length was significantly smaller in tagged fish than in untagged controls.Over 180 mm in length, no difference could be detected in the length gains of tagged and untagged fish.Smaller tagged fish gave up maintaining station and showed signs of distress at an earlier stage in the experiment than larger tagged fish. Tag-related mortality was inversely related to fish size. All but two of the control fish survived the experiment and continued to maintain station throughout.The results suggest that the minimum length of Atlantic salmon parr on which this design of acoustic tag could be employed is 160 mm, and that caution is needed in the interpretation oftrack lengths longer than LO days of parr in the size range 160-180 mm.