Studies of the micro-distribution of herbivorous lotic macroinvertebrates have tended to rely on short-term experiments. While these experiments have provided information on the factors that determine the distributions of herbivorous invertebrates, they ignore the complexity of the epilithon as a food source, any interaction between physical and biological factors and the possibility that the factors that determine an invertebrate species' distribution may vary through time. Five colonization experiments were performed in an Australian upland stream to determine the role of epilithon abundance in the distribution of an abundant invertebrate (Agapetus monticolus: Trichoptera). Artificial cobbles with different types and amounts of epilithon were placed in the Taggerty River at two sites and the response of A. monticolus recorded. The experiments revealed that, although A. monticolus expressed a preference for patches of epilithon with abundant periphyton, the response varied among experiments performed at different times and was always secondary to the flow preferences of A. monticolus. Studies that fail to acknowledge the presence of temporal variation in the response of animals to their environment run the risk that they will over- or underestimate the importance of particular factors because of the timing of their experiment.