Résumé
* As nature-based tourism grows, its impacts on aquatic ecosystems follow, requiring effective management techniques to conserve river integrity. Nevertheless, strong indicators of tourism impacts are scarce and have seldom been studied for many species. * This study evaluated whether microhabitat use and activity period of a variety of fish species are effective for assessing tourism impacts in a headwater stream in which recreation (snorkelling) occurs. Microhabitat use and activity period of the most representative fish were observed in river stretches used for recreation and compared with reference stretches. * Fish neither altered microhabitat use nor became more cryptic owing to tourism, maybe because they have evolved to occupy a certain microhabitat and cannot inhabit others, thus making habitat use a poor indicator of tourism impact. * Fish respond to the presence of tourists by changing the promptness to begin and end activity, making the time of activity a good indicator of tourism impact, which can be easily assessed by tourism managers or government agencies and used to control the environmental impact of recreation involving fish. * It is suggested that aquatic conservation protocols should include local indicators, and that behavioural indicators (activity period) might be more relevant than ecological indicators (microhabitat use) for early recognition of tourism impacts. These findings can be extended to the conservation of other fish communities subject to intensive tourism and with a strong circadian rhythm, such as coral reef fishes.