1. Despite its critics, the self-thinning rule with a -3/2 power has been applied to a wide range of plant species. A similar relationship has been proposed for animals, energetics allometries suggesting a -4/3 power. This relationship has been rarely tested for mobile animals; the present investigation examines its suitability for juvenile brown trout, Salmo trutta L. Data came from a 25-year study of a trout population in the English Lake District. 2. The relationship between mean live weight and mean density was well described by a negative power function, using a functional regression (GMb); the power for the pooled data for all year-classes being 1.33, exactly as predicted. Functional regressions for separate year-classes showed that both b and the intercept log a varied between year-classes with overall weighted mean values of 1.35 and 3.01 respectively; exactly the same values were obtained for median estimators from non-parametric regressions (complete Theil estimators). 3. Apart from two outlying drought years, there was no significant variation in b: parallel lines with a common slope of 1.35 provided an adequate fit to the data. Apart from one high value for the 1981 year-class, intercepts varied between year-classes in the range 2.54-3.74. Intercepts of parallel lines ranged from 1.77 to 2.70. Unfortunately, no obvious factors could be found to explain this variation, but the intercepts of parallel lines were positively related to geometric mean density and negatively related to geometric mean weight in each year-class. 4. Differences between year-classes may reflect differences in the carrying capacity of the habitat. The 'self-thinning' relationship could therefore provide a useful method for comparing temporal and spatial variations in carrying capacity both within and between trout streams.
