In order to evaluate the potential risks of human visitation on macroinvertebrate communities in streams, we investigated the effect of trampling using two short-term experiments conducted in a Brazilian ecotourism karst region. We asked three questions: (a) Does trampling increase the drift rate of aquatic macroinvertebrates and organic matter? (b) Does trampling change the macroinvertebrate community organization? (c) If trampling alters the community structure, is a short time (5 days, a between weekends interval - peaks of tourism activities) sufficient for community restructuring? Analysis of variance of richness, total abundance, abundance of the most abundant genus (e.g., Simothraulopsis and Callibaetis), and community composition showed that trampling immediately affects macroinvertebrate community and that the intervals between the peaks of visitation (5 days) are not sufficient to complete community restructuring. Considering that bathing areas receive thousands of visitors every year and that intervals of time without visitation are nearly nonexistent, we suspect that the negative effects on the macroinvertebrate community occur in a cumulative way. Finally, we discuss some simple procedures that could potentially be used for reducing trampling impacts in lotic environments.