Abstract
Multiple linear regression was used to relate the numbers of the amphipod G. pulex (L.) to the organic and physical characteristics of the stream substratum. Different size classes were treated separately. Substratum particle size and the amount of detritus were recorded both by gravimetric methods and by analysis of photographs of the substratum surface. These methods gave quite different results, particularly if a layer of silt was present on the substratum surface. Information derived from substratum photography was the more effective in explaining microdistribution patterns. G. pulex larger than 6 mm (body length) appeared to congregate in places where there were accumulations (packets) of autumn-shed leaves, but only if these were visible on the substratum surface. Smaller individuals showed no clear relationship with detrital distribution. Leaf packets containing species that decayed rapidly were selected for by large G. pulex early in the autumn. Later, only resistant species were availiable and leaf species did not appear to influence microdistribution. There was a relationship between substratum particle size and the size of G. pulex. Larger individuals were found in progressively narrower ranges of larger particle sizes. The size of interstices between particles controls the sizes of animals that may colonize an area. In experiments in an artificial stream, the rates of movement of all sizes of G. pulex were inversely related to substratum particle size, but the difference between rates of movement over fine and coarase substrata was greatest for large individuals. This result corroborates the relationship between particle size and body size that was obsreved in the field.