The spatial pattern of antibiotic resistance in culturable sediment bacteria from four freshwater streams was examined. Previous research suggests that the prevalence of antibiotic resistance may increase in populations via indirect or coselection from heavy metal contamination. Sample bacteria from each stream were grown in media containing one of four antibiotics-tetracycline, chloramphenicol, kanamycin, and streptomycin-at concentrations greater than the minimum inhibitory concentration, plus a control. Bacteria showed high susceptibilities to the former two antibiotics. We summarized the latter two more prevalent (aminoglycoside) resistance responses and ten metals concentrations per sediment sample, by Principal Components Analysis. Respectively, 63 and 58% of the variability was explained in the first principal component of each variable set. We used these multivariate summary metrics [i.e., first principal component (PC) scores] as input measures for exploring the spatial correlation between antibiotic resistance and metal concentration for each stream sampled. Results show a significant and negative correlation between metals PC scores versus aminoglycoside resistance scores and suggest that selection for metal tolerance among sediment bacteria may influence selection for antibiotic resistance differently in sediments than in the water column. Our most important finding comes from geostatistical cross-variogram analysis, which shows that increasing metal concentration scores are spatially associated with decreasing aminoglycoside resistance scores--a negative correlation, but holds for contaminated streams only. We suspect our field results are influenced by metal bioavailability in the sediments and by a contaminant promoted interaction or "cocktail effect" from complex combinations of pollution mediated selection agents.