ABSTRACT A field-portable water-sampling system was designed and fabricated for collecting adequate volumes of seawater to meet the quantitation requirements to support Natural Resource Damage Assessment (NRDA) toxicity determinations and modeling efforts following an oil spill. This system is a significant improvement to conventional water sampling equipment and includes the ability to filter water samples at the time of collection, thereby providing critical differentiation between truly dissolved constituents and dispersed oil droplets. The system can be quickly and easily deployed from shoreline structures (piers and breakwaters) and/or vessels of opportunity to provide essential data during the early stages of a spill. Likewise, data collected with the system can be used to document dispersant effectiveness and provide information relating to seafood exposure, tainting, and toxicity issues. In many oil-spill NRDA efforts, water-column effects from dissolved components and dispersed oil droplets have not been adequately quantified or documented because: (1) samples are not obtained early enough after the spill event; (2) insufficient volumes are collected; and (3) the wrong constituents are analyzed. Generally, EPA hazardous-materials sampling approaches are followed, leading to inadequate sample sizes (e.g., 40 mL for volatile component analyses and 1-L samples for dissolved/dispersed constituents). Analytically, EPA semivolatile gas chromatography/mass spectrometry (GC/MS) SW-846 Method 8270 is often specified for polynuclear aromatic hydrocarbons (PAH). These sample sizes are not large enough to meet the detection limits required for most marine hydrocarbon analyses (de Lappe et al., 1980; Payne, 1997 and references therein), and the EPA PAH target analyte list does not include the majority of alkyl-substituted one-, two-, and three-ring aromatics that are the primary dissolved constituents actually present in the water column following an oil spill (Sauer and Boehm, 1991). As a result, water column effects are often written off as being short-lived or insignificant. Alternatively, impacts are often assessed by computer modeling efforts with limited field validation. In either event, there is inadequate profiling of the extent and duration of petroleum hydrocarbon exposure to marine organisms. Furthermore, when adequate volumes of water have been collected and the proper target analytes have been specified, provisions have not been taken to differentiate between truly dissolved components and dispersed oil droplets. Consequently, later data analyses are unreliable in their ability to reflect conditions as they actually existed during the early stages of the spill. For example, PAH analyses of unfiltered water samples are confounded by the facts that: (1) a significant, but unknown fraction of discrete oil droplets in the water column will rise to the surface with time; (2) high levels of dispersed oil droplets will raise detection limits of dissolved PAH; and (3) it is impossible to determine how much of the PAH is in the truly dissolved state where it will persist as a toxic fraction to exposed organisms and how much is simply associated with slightly less toxic oil droplets that are subject to relatively rapid removal by resurfacing. The equipment and field implementation approach described in this paper can provide samples that are not subject to the aforementioned problems.