The treatment efficiencies of two prefabricated stormwater-treatment devices were tested at a freeway site in a high-density urban part of Milwaukee, Wisconsin. One treatment device is categorized as a hydrodynamic-settling device (HSD), which removes pollutants by sedimentation and flotation. The other treatment device is categorized as a stormwater-filtration device (SFD), which removes pollutants by filtration and sedimentation. During runoff events, flow measurements were recorded and water-quality samples were collected at the inlet and outlet of each device. Efficiency-ratio and summation-of-load (SOL) calculations were used to estimate the treatment efficiency of each device. Event-mean concentrations and loads that were decreased by passing through the HSD include total suspended solids (TSS), suspended sediment (SS), total phosphorus (TP), total copper (TCu), and total zinc (TZn). The efficiency ratios for these constituents were 42, 57, 17, 33, and 23 percent, respectively. The SOL removal rates for these constituents were 25, 49, 10, 27, and 16 percent, respectively. Event-mean concentrations and loads that increased by passing through the HSD include chloride (Cl), total dissolved solids (TDS), and dissolved zinc (DZn). The efficiency ratios for these constituents were -347, -177, and 20 percent, respectively. Four constituents—dissolved phosphorus (DP), chemical oxygen demand (COD), total polycyclic aromatic hydrocarbon (PAH), and dissolved copper (DCu)—are not included in the list of computed efficiency ratio and SOL because the variability between sampled inlet and outlet pairs were not significantly different. Event-mean concentrations and loads that decreased by passing through the SFD include TSS, SS, TP, DCu, TCu, DZn, TZn, and COD. The efficiency ratios for these constituents were 59, 90, 40, 21, 66, 23, 66, and 18, respectively. The SOLs for these constituents were 50, 89, 37, 19, 60, 20, 65, and 21, respectively. Two constituents—DP and PAH—are not included in the lists of computed efficiency ratio and SOL because the variability between sampled inlet and outlet pairs were not significantly different. Similar to the HSD, the average efficiency ratios and SOLs for TDS and Cl were negative. Flow rates, high concentrations of SS, and particle-size distributions (PSD) can affect the treatment efficacies of the two devices. Flow rates equal to or greater than the design flow rate of the HSD had minimal or negative removal efficiencies for TSS and SS loads. Similar TSS removal efficiencies were observed at the SFD, but SS was consistently removed throughout the flow regime. Removal efficiencies were high for both devices when concentrations of SS and TSS approached 200 mg/L. A small number of runoff events were analyzed for PSD; the average sand content at the HSD was 33 percent and at the SFD was 71 percent. The 71-percent sand content may reflect the 90-percent removal efficiency of SS at the SFD. Particles retained at the bottom of both devices were largely sand-size or greater.