Fluorescence in situ hybridization (FISH) is widely used to characterize bacterial community structure. However, a major limitation of the FISH technique is that the effectiveness of target cell detection varies widely over ecosystem types and with differences in methodology. Samples collected at sea often are stored for weeks or months before analysis using FISH, therefore quantifying the effects of storage on the detection of bacterial cells is crucial for comparing studies of bacterial community structure from diverse regions and ecosystems. Presented are the results of a 12-month time-course study during which replicate seawater samples were prepared, stored frozen, and hybridized after 0, 1.5, 3, 6, and 12 months to determine the effects of long-term sample storage on hybridization efficiency and the characterization of community structure. The time-dependent slope of the probe for Bacteria, but not the Cytophaga-Flavobacteria cluster or the α- and γ-Proteobacteria, was significantly different from zero, with a 6.3% change in target cell detection per year. This change in detection was small and within the typical error reported for bacterial counting. We conclude that during this 12-month timecourse study there was a minimal effect of long-term storage on the detection of bacterial cells using FISH.