We examined whether presence of vegetation and seasonal changes in flow affected N chemistry and denitrification rates within sandbars of a 7^th-order sandy alluvial river (Wisconsin River, USA). We addressed these questions with a broad-scale approach of measuring parafluvial water chemistry and denitrification rates in multiple sandbars distributed along a 15-km river reach during summer 2004 and 2005. After recession of spring flooding, parafluvial chemistry in unvegetated bars was characterized by moderate dissolved O₂ (DO) and elevated NO₃⁻-N concentrations (>3.5 mg N/L), whereas vegetated bars tended to be hypoxic (<2 mg/L DO) and depleted in NO₃⁻-N (0.2 mg N/L) relative to unvegetated bars and surface water (0.47 mg N/L). As flow declined over the summer, NO₃⁻-N also declined in both bar types, whereas SO₄²⁻ was relatively constant in unvegetated bars but decreased in vegetated bars. Amendment experiments demonstrated that denitrification was limited primarily by NO₃⁻-N and secondarily by organic C in both bar types, but the strength of this limitation varied over time and was greater in vegetated bars, a result suggesting loss of denitrification capacity. Thus, spatial and temporal patterns of water chemistry and denitrification activity among multiple sandbars indicated that unvegetated bars shifted from N transformers/NO₃⁻-N sources early in the summer to N retainers/NO₃⁻-N sinks as discharge declined, whereas vegetated bars always supported anaerobic processes and probably shifted from NO₃⁻-N to SO₄²⁻ sinks. We hypothesize that the contribution of vegetated islands to overall riverine N retention is small because establishment of vegetation reduces hydrologic linkages between bars and surface water. Modern changes in the flow regime of the Wisconsin River have increased establishment of riparian vegetation on exposed bars, a pattern suggesting that parafluvial N retention is being reduced while riverine N loading is increasing.