With 5 figures in the text) Salmon (Salmo salar) formed the largest proportion by mass of fish found in the stomachs of redbreasted mergansers (Mergus serrator) shot during the smolt run in north-east Scotland. Salmon parr represented approximately two-thirds by mass of juvenile salmon eaten, whilst smolts were present in a smaller proportion than previously predicted. The median lengths of parr and smolts eaten were 70 and 115 mm, respectively; the latter being significantly smaller than the smolt population sampled during annual production estimates on the river.

With 3 plates and 3 figures in the text)A method isdescribed in which a single bone, the first vertebra, is used to distinguish bone remains ofjuvenile salmon (Salmo salur) and trout (Salmo rrufia) with 90% certainty. A single regression of salmon and trout first vertebra-width versus fish-length of fish predicted the latter with an accuracy of f 10 mm (95% confidence limits) for salmonids of 45-150 mm fish-length. First vertebrae were assigned to salmon or trout based on three visual characters with 89%,-90% certainty. Salmon are more variable than trout for the three characters. Of first vertebrae, 6.5% were not readily identifiable as either salmon or trout, and a further 3.0%) were misclassified using our criteria.

The field metabolic rates (FMRs) of nine captive goosdnders, Mergus merganser, released on two Scottish rivers were estimated using the doubly-labelled water (DLW) technique. Mean ( * s.E.) FMR was 2.322 & 0.239 nil CO, g ~ ' h ~ ' and daily energy expenditure (DEE) 1939 f 184 kJ per day. This was significantly greater ( x 1.5) than previous estimates which assumed DEE to be three times the basal metabolic rate (BMR) based on regression equations predicting BMR from body mass. FMR of captives and dietary data from previous studies were used to estimate daily consumption of salmon, Salmo salar, smolts and parr by natural populations of these ducks on the river North Esk, north-east Scotland. Goosanders are likely to consume 480-522 g fish per day of which two-thirds are juvenile salmon; equivalent to a daily intake of 10-11 smolts and 48-52 parr. Annual predation of smolts by goosanders was estimated to be between 8000 and 15 000 or 3 and 16%) of annual production.

Current models estimating the impact of red‐breasted mergansers, Mergus serrator L., on salmon, Salmo salar L., fisheries in Scotland fail to take account of any annual variation in the proportion of the diet that is smolts. During the 1987–1990 smolt runs, the annual variation in the diet of mergansers was estimated from the stomach contents of birds shot on two Scottish rivers. The proportion of salmon in the diet was greatest early in the smolt run (76–91% by weight), and contained proportionately more smolts than later in the run, when coarse fish were more prominent. There was little annual variation in the proportion of the diet that was juvenile salmon. However, the proportion of these fish that were smolts, was twice as great in some years than in others and this appeared to be independent of estimated annual smolt production in the rivers.

No abstract