DOES THE SPATIAL SCALE OF EXPERIMENTATION MATTER - A TEST WITH TADPOLES AND DRAGONFLIES

We report the results of a study designed to examine whether the spatial scale of experimentation in artificial ponds (cattle watering tanks) influences the results of manipulative studies of competitive and predatory interactions. The experiment consisted of two initial densities of Rana utricularia (Ranidae) tadpoles (0.035 and 0.14 animals/L) and the absence or presence (0.005 animals/L) of predatory Tramea lacerata (Odonata, Libellulidae) naiads in each of two water depths in cattle tanks of equal dimensions. Depth of tank did not significantly affect survival rate of tadpoles either as a main effect or as part of any interactive effect. At low tadpole densities the addition of predators decreased tadpole survival rates significantly, but at high tadpole densities they increased them significantly. Length of larval period increased significantly with increases in the density of surviving tadpoles at similar rates at both water depths, but deeper tanks produced significantly longer larval periods for the same tadpole density. Size at metamorphosis decreased significantly with increases in the density of surviving tadpoles, but it decreased at a significantly faster rate in the shallower tanks. The results indicate that experiments at smaller spatial scales do not necessarily distort the numerical dynamics in this system but that they can overestimate the importance of density variation in producing variation in some phenotypic characters.