The relationship between first vertebra width and body length of Atlantic salmon differs between parr and smolts

The relationship between the width of the first vertebra (WV) and the fork length (LF )o f Atlantic salmon diVers between parr and smolts. The previous use of the first vertebra to reconstruct the diets of predators from the bone remains of ingested prey will have underestimated the size of smolts and probably also their frequency of occurrence in the diet. Using the ratio WV:LF provides potential for diVerentiating between intact parr, smolts and pre-smolts in the diet. To model the impacts of predators on populations of fish, it is usually necessary to know what sizes offish are consumed. The diet of predators is commonly assessed by analysing the contents of their guts. Because the rate of digestion may be rapid compared with the rates of feeding and of processing sampled material, it is often the case that few intact prey items are extracted from predators that are killed. To overcome this problem, diet can be reconstructed using the bone remains of prey fishes (Fortunatova, 1951; Popova, 1978). In analysing the diet of sawbill ducks, common bird predators of Atlantic salmon, Salmo salar L., Feltham (1990, 1995a,b) used the first vertebra (the atlas bone) to diVerentiate between salmon and trout,Salmo truttaL., and to estimate the lengths of the fish that had been consumed. He used a relationship between the first vertebra width (WV) and fork length (LF) derived from measurements of 200 juvenile trout and salmon of wild and hatchery origin (Feltham & Marquiss, 1989). Based on results obtained using this method, it has been concluded that sawbill ducks tend to select the smaller fish, and so to select resident parr rather than migrant smolts even during the spring smolt run, and that most of the smolts they eat are small in relation to the length range available (e.g. Feltham, 1990; Carss & Marquiss, 1991; Feltham, 1995a,b). Feltham & Marquiss (1989) derived their relationship between WV and LF from salmon of up to 15 cm but did not diVerentiate between parr and smolts. Salmon change shape, becoming more elongate, as they metamorphose from parr to smolts prior to migrating to sea (Hoar, 1976) and it is not known whether the relationship between WV and LF remains constant as this process occurs. In the present study, the relationships between atlas width and fish length in four groups of fish were analysed: wild and hatchery parr and smolts. The main objective was to assess the validity of extrapolating from regressions derived using parr to infer the size and occurrence of smolts in the diet of predators.