Legacy phosphorus concentrations resulting from historic additions of phosphorus (P) to the landscape may impede rapid remediation of P pollution and achievement of water quality management goals. Herein, we hypothesized that the capacity of stream biofilms to assimilate new polyphosphate (polyP) will vary as a function of stream legacy phosphorus. To test this hypothesis, we deployed a series of in situ enrichment experiments at five sites of varying land cover in central Pennsylvania, United States. Incremental P-loading was delivered using vials fitted with porous lids, that contained agar enriched with six levels of P (as Dissolved inorganic phosphorus, dissolved inorganic P) loading with rates ranging from 0 to 1,540 µg PO 4 −3 /day; these loading rates mimicked natural stream P loadings. Substrata were incubated at stream sites for a relatively short incubation period (12 days), to measure uptake rates; after which, biofilms growing on the lids were removed and their tissue content was analyzed for biomass (as chlorophyll) and various forms of particulate phosphorus. Polyphosphate (polyP) accumulated by stream biofilms at all sites closely tracked the release of dissolved inorganic P from experimental enrichment assays. Comparatively, biofilms accumulated relatively small amounts of Particulate inorganic phosphorus and other forms of organic P that we assume constitute a third group of P-rich biochemicals (e.g., DNA, RNA, lipids, proteins). Viewed at the watershed scale, land use appeared to affect P accumulation, where sites dominated by forest cover had a higher capacity for P storage, while sites dominated by agriculture did not; this underscores the importance of polyP storage as an indicator of legacy P pollution.