1. Survival, growth and downstream dispersal of trout (especially 0 group) and the relationships of these variables to initial stocking density were studied in north Pennine streams. 2. Two methods were used. First, electrofishing censuses were made in a marked reach of each of four streams over a period of about 20 years. Second, downstream moving trout were trapped in two streams over a 10-year period. Each stream upstream of the trap was experimentally stocked with 'swim-up' trout fry, using a different population density each year. 3. Before 1970 the four census reaches showed very large year-upon-year variations in August trout parr densities, with local failures of recruitment in some years. Population densities after completion of Cow Green Reservoir (1970) were generally higher but still showed wide fluctuations. 4. Survival (including the effects of losses by dispersal) from swim-up to early August, for starting population densities of 0-10 fry m(-2), was about 10% regardless of initial density. Estimates of survival from August to early October were 30-50% for the census reaches and 55-65% for the areas upstream of the traps. However, for August 0 group densities of 0-0.9 m(-2), estimated instantaneous loss rate from August of the first year of life up to age 40-65 months showed a positive curvilinear relationship to population density in the first year of life. Loss rate was, therefore, density-dependent during this period. 5. Estimated instantaneous growth rate day(-1) of 0 group fish from swim-up to August and from swim-up to October was inversely related to the natural logarithm of August population density and this was most apparent for August densities of <0.15 fish m(-2). 6. Although survival from swim-up to August was proportional (about 10%, at starting densities of 10 m(-2) or less), the percentages of the total loss attributable to mortality and to downstream dispersal varied considerably with starting density. At starting densities around 4-5 fish m(-2) dispersal was negligible. As initial density rose above 4-5 fish m(-2) and towards 10 fish m(-2) the percentage of loss attributable to dispersal rose towards 30%. As initial densities decreased from 4 to 1.4 fish m(-2), the percentage rose to around 20%. Below a starting density of 1.4 fish m(-2) the percentage decreased.
