Longitudinal structuring of stream-fish assemblages: is niche partitioning observed in two-species systems applicable to three-species systems?

Longitudinal distribution and niche partitioning phenomena have been extensively reported for streams containing two salmonid species but remain poorly understood for streams containing three or more salmonid species. Herein, we examined the abiotic environmental variables and population densities of fishes, including three salmonids, in 30 reaches of the Setose River, Hokkaido, Japan, across altitudes (120–450 m) and maximum water temperature (15–26 °C) gradients to determine the distribution of major fish species in this system. The densities of white-spotted charr Salvelinus leucomaenis , southern Asian Dolly Varden Salvelinus curilus , masu salmon Oncorhynchus masou , fluvial sculpin Cottus nozawae , and stone loach Barbatula oreas varied along environmental gradients despite their wide distribution sympatrically. We further investigated the interspecific interactions between three potentially competitive salmonids, white-spotted charr, Dolly Varden, and masu salmon. We observed that in the three-species system, including upstream Dolly Varden, salmon occupied upstream areas colder than those inhabited by charr, which is contrary to the pattern reported for charr and salmon two-species systems, wherein charr are found in the colder upstream areas. In underwater observations, salmon preferred faster water velocities than charr and Dolly Varden. Dolly Varden were found exclusively on the riverbed, followed by charr, whereas salmon occurred further away from the riverbed. Analysis of the stomach contents revealed that all three salmonids depended on terrestrial insects, but only Dolly Varden consumed benthic aquatic insects. In our three-species system that included Dolly Varden (preadapted to benthos foraging), no difference was observed in the degree of terrestrial insect consumption between charr and salmon. These results are in striking contrast to the reports of charr preying on relatively more aquatic insects than those consumed by salmon in charr and salmon two-species systems. Therefore, we speculate that Dolly Varden and salmon influence interspecific interactions between charr and salmon and between charr and Dolly Varden, respectively, explaining the distinct longitudinal distribution and feeding habits observed in two- vs. three-species systems. These variant interspecific interactions among the three competing species may have driven the expansion of their sympatric zones.