Shelters are a key component of animal population ecology, as they provide protection from predators and promote visual isolation among competitors. From a behavioral perspective, how shelter availability affects the way individuals allocate their activity in space has been extensively documented. However, how shelters affect the distribution of activity in time (i.e., diel activity) has been less studied in natural conditions. Here, we report results from a field study that used stream enclosures with either high or low shelter availability and stream-dwelling juvenile Arctic charr as a model species. We collected repeated measurements of individual activity 8 times a day (every 3h) for six 24-h periods during a span of 9 days. In shelter-limited enclosures, fish were more active, became diurnal, and were active over a wider span of time each day, compared with fish with access to abundant shelters. In addition, fish were more aggregated and attacked prey over shorter distances and at lower rate when shelters were limited. Body mass did not affect individual activity rates, but smaller individuals extended their activity over longer periods, possibly as a result of interference competition, and were more isolated. Growth rates were similar across treatments and were positively correlated with individual activity rates and the average distance to the nearest competitor. However, additional nonmeasured effects on fitness may occur, such as increased predation risk associated with daytime feeding and extended activity in shelter-limited conditions.