Mixtures of organochlorine compounds have the potential for additive or interactive toxicity to organisms exposed in the stream. This study uses a variety of methods to identify mixtures and a modified concentration-addition approach to estimate their potential toxicity at 845 stream sites across the United States sampled between 1992 and 2001 for organochlorine pesticides and polychlorinated biphenyls (PCBs) in bed sediment. Principal-component (PC) analysis identified five PCs that account for 77% of the total variance in 14 organochlorine compounds in the original dataset. The five PCs represent: (1) chlordane-related compounds and dieldrin; (2) p,p′-DDT and its degradates; (3) o,p′-DDT and its degradates; (4) the pesticide degradates oxychlordane and heptachlor epoxide; and (5) PCBs. The PC analysis grouped compounds that have similar chemical structure (such as parent compound and degradate), common origin (in the same technical pesticide mixture), and(or) similar relation of concentrations to land use. For example, the highest concentrations of chlordane compounds and dieldrin occurred at urban sites, reflecting past use of parent pesticides for termite control. Two approaches to characterizing mixtures—PC-based mixtures and unique mixtures—were applied to all 299 samples with a detection of two or more organochlorine compounds. PC-based mixtures are defined by the presence (in the sample) of one or more compounds associated with that PC. Unique mixtures are defined as a specific combination of two or more compounds detected in a sample, regardless of how many other compounds were also detected in that sample. The simplest PC-based mixtures (containing compounds from 1 or 2 PCs) commonly occurred in a variety of land use settings. Complex mixtures (containing compounds from 3 or more PCs) were most common in samples from urban and mixed/urban sites, especially in the Northeast, reflecting high concentrations of multiple chlordane, dieldrin, DDT-related compounds, and(or) PCBs. The most commonly occurring unique mixture (p,p′-DDE, p,p′-DDD) occurred in both simple and complex PC-based mixtures, and at both urban and agricultural sites. Mean Probable Effect Concentration Quotients (PEC-Q) values, which estimate the potential toxicity of organochlorine contaminant mixtures, were highest for complex mixtures. Mean PEC-Q values were highest for urban sites in the Northeast, followed by mixed/urban sites in the Northeast and agricultural sites in cotton growing areas. These results demonstrate that the PEC-Q approach can be used in combination with PC-based and unique mixture analyses to relate potential aquatic toxicity of contaminant mixtures to mixture complexity, land use, and other surrogates for contaminant sources.