1. Analysis of the stable isotope signatures of carbon (C) and nitrogen (N) of foods and consumers has led to some preliminary understanding of the relative importance of autochthonous and allochthonous resources in tropical streams. However, robust generalizations about the dynamics of food webs in these habitats, and their response to shading gradients or season, are still lacking. In addition, the feasibility of employing a baseline δ15N value for estimating trophic positions (TPs) of consumers in small tropical streams has yet to be explored.
2. We analysed data on stable isotope signatures of food sources and aquatic consumers obtained from 14 studies carried out in small streams in monsoonal Hong Kong (22°30′N, 114°10′E) between 1996 and 2006. Emphasis was placed on determining the relative importance of leaf litter and autochthonous foods in supporting consumer biomass, and the extent to which trophic base and TP vary among streams and seasons.
3. Although allochthonous leaf litter was generally 13C- and 15N-depleted relative to autochthonous foods, there were marked isotopic shifts of food sources and consumers in response to season (dry versus wet) and stream shading. Consumer taxa were generally more 13C- and 15N-enriched in the unshaded streams, but seasonal effects were more variable. Despite these changes, there was consistent evidence that stream food webs were based on periphytic algae and/or cyanobacteria with leaf litter serving as a minor food.
4. Heptageniidae (Ephemeroptera), Tipulidae (Diptera), Elmidae (Coleoptera) and shrimps (Atyidae) were used as a baseline for calculating the TPs of other consumer taxa. The maximum TPs in shaded streams remained fairly constant between seasons (dry = 3.93; wet = 3.97), while those in unshaded streams were higher and showed seasonal fluctuations (dry = 5.13; wet = 4.39).
5. Although variations in consumer isotope signatures in response to season and shading gradients did not confound our interpretation of the stream food base, changes in consumer δ15N did affect the calculation of consumer TPs. Misleading estimates of consumer TPs are likely if samples are collected from a narrow range of streams and/or during one season. Overestimation of the TPs of specialist herbivores (e.g. fish grazers) is also possible when autochthonous resources are substantially more 15N-enriched than allochthonous foods.