Assimilation-based techniques such as stoichiometric analysis, bulk tissue stable isotope analysis (SIA), fatty acid (FA) profiling and compound-specific SIA of FAs can be used to resolve ambiguities in consumer-resource relationships, but comparisons of their effectiveness are lacking. Feeding trials and concurrent application of these techniques were used to investigate the trophic base of two shrimps from Hong Kong streams: the atyid Caridina cantonensis, which is a primary consumer, and predatory Macrobrachium hainanense (Palaemonidae). Leaf litter and periphyton were fed to C. cantonensis (CC-LF and CC-PF, respectively) reared in the laboratory for 2 months, when C. cantonensis, that had fed on a mixed diet (CC-WC), were collected from the field. Atyids from each group (CC-LF, CC-PF and CC-WC) were fed to M. hainanense (MH-L, MH-P and MH-W, respectively) during a further 2-month trial, at the end of which M. hainanense were also collected from the field (MH-R). FA biomarkers present in CC-WC and CC-PF indicated that C. cantonensis depended primarily on autochthonous foods, and FA profiles of CC-WC were distinct from those of CC-LF. Differences in C/N ratios and isotope signatures of leaf litter and periphyton were not reflected in tissues of atyids or palaemonid predators. FA profiles of M. hainanense groups were similar, but FA stable C isotope ratio (delta(13)C) signatures of MH-R and MH-W were distinct from those of MH-L and leaf litter. FA delta(13)C signatures of MH-R were similar to those of MH-P and MH-W, indicating that autochthonous resources constituted the trophic base of production for this predator. This is the first study using compound-specific SIA to study stream food chains. Compound-specific SIA and FA profiling allowed accurate elucidation of consumer-resource relationships that were not revealed by stoichiometry or bulk tissue SIA. Compound-specific SIA was particularly sensitive for detecting rapid shifts in the predator trophic base. This tool will have wide applicability for investigating food webs in a range of ecosystems.