No abstract

A volumetric method in which the stomach contents are squashed to a uniform depth and the volume is derived from the area of the squash has been evaluated. The apparatus is simple in constiuction and the method applicable to a wide range of fish sizes. The overall error of the method may be as low as 3-5 "/, for well macerated material, though for hard bodied food animals thc coefficient of variation may be up to 8%. These volumetric determinations are not directly proportional to gravimetric measurements but logarithmic conversion curves are readily derived. I. INTRODUCTIONNumerical and occurrence methods of expressing the composition of the diet of fishes are the most rapid in use but have the major disadvantage of over-emphasizing the importance of numerous small items and cannot deal effectively with plant material or indeterminate numbers of fragmented, yet identifiable, animals. Volumetric or gravimetric methods, however, have the opposite bias and over-estimate the importance of larger, yet less frequently eaten, food items. Windell (1968) suggested a combination of volumetric and numerical methods to indicate the importance of food items. In the absence of data on the relative rates of digestion of different food items of various sizes this combination is probably a reasonable compromise.

SUMMARY. 1. The exceptionally dry summer of 1976 presented an opportunity to assess the impact of drought on the fauna of an unregulated stream in Wales. 2. Total number of macroinvertebrates varied from 1808 m−2 in September 1976 to 7412m−2 in August 1977. Two main effects of drought were observed: an initial reduction in abundance during the drought and a change in community structure in the following year, probably through impaired reproduction. 3. Trout density ranged from 0.075±0.028 to 0.402±0.093 m−2 and salmon parr density varied between 0.014±0.004 and 0.414±0.119 m−2. The only detrimental effect of the drought on the fish fauna was the elimination of the 1976 year class of young salmon. probably due to prolonged high water temperature.

Checks were formed on the scales in June and July. A dominant 1959 year class accounted for half the population present in 1964. Females grew faster than males and attained greater ultimate size. Growth in the River Lugg was good in comparison with rates in other British waters. The sex ratio was found to change with age; a predominance of males in the youngest age classes changed to a preponderance of females in the older classes. The youngest fishes caught (3+) were already sexually mature. The gonads were found to be a constant proportion of the body weight for fishes of all sizes at a given season. Spawning was completed by June; males were ripe in May. The ovaries recovered more rapidly than the testes after spawning. Seasonal variation in feeding activity, with higher activity in summer, was observed and could be correlated with temperature. Seasonal changes in the diet could be related to availability. The diet was predominantly vegetable; the most important animal component, aquatic insect larvae, accounted for 11 % of the total diet. Aerial insects were rarely eaten. Diet varied with age: molluscs being particularly important to older roach ; younger roach consumed large amounts of substrate material. There is some evidence that young roach indiscriminately ingest substrate material but older roach are more selective and reject inert material.