The influence of refugia on predation impacts in a Hong Kong stream

Abstract

A field experiment compared macroinvertebrate colonization of three types of introduced substrate, which differed only with respect to the provision of prey refuges, in the presence of natural densities of predatory, stream fish (Ctenogobius duospilus: Gobiidae) in the Lam Tsuen River, Hong Kong. Replicated Hester-Dendy multiple samplers (MPS's) were used to manipulate prey refuges by adjusting the spacing between plates while keeping surface area and complexity (heterogeneity) constant. Where plates were closely spaced, predators could not gain access to the surface of the MPS's to capture prey; wide spacing permitted access by all predators, while small predators only could gain access to plate surfaces where the spacing was intermediate. Densities of macroinvertebrate colonizers on MPS's were monitored after 7, 14, 21 and 29 days in the stream. Density (all species combined) and morphoscpecies richness of macroinvertebrates colonizing MPS's over a 29-day period decreased in the order close > intermediate = large spacing providing evidence of a beneficial effect of refuges on prey assemblages. There was also a trend towards greater abundance of colonizers on MPS's with increasing time. Densities of the majority of taxa were highest on MPS's with close spacing, but not all taxa exhibited increased abundance with time. The thiarid snail, Brotia hainanensis, which was not eaten by Ctenogobious duospilus, was the only macroinvertebrate taxon to exhibit significantly lower densities on MPS's with close spacing. Abundance was greater on MPS's with intermediate and wide spacing. By contrast, densities of those macroinvertebrate taxa which were vulnerable to fish predation were enhanced by the refuges on closely-spaced MPS's. In order to confirm the importance of refuges in reducing predator impacts, replicated MPS's with close and wide spacing were enclosed individually inside coarse-meshed (3.5 mm) net bags and placed in the stream. The bags prevented predator access but not allowed macroinvertebrate colonization. Faunal densities were monitored after 14 and 28 days colonization. Macroinvertebrate densities (all species combined as well as individual taxa) and morphospecies richness on MPS's enclosed in nets were rather similar after 14 and 28 days regardless of spacing treatment. Where inter-treatment differences were noted, there was no consistent trend: i.e., some taxa were more numerous in the closely-spaced treatment, whereas the abundance of others (and densities of all species combined) was greater in the wide-spacing treatment. In essence, the beneficial effects of refuges on prey density and morphospecies richness disappeared when predators were excluded.