Dungeness crab studies conducted in 2002 for the Portland District of the U.S. Army Corps of Engineers (Corps) constituted a major step forward in quantifying crab entrainment through statistical projections of adult equivalent loss (AEL) and loss to the fishery (LF) from proposed construction and maintenance dredging in the Columbia River navigation channel (Pearson et al. 2002, 2003). These studies also examined the influence of bottom salinity on crab abundance and entrainment rates. Additional sampling was conducted in 2004 to improve loss projections, further develop the crab salinity model, and apply the model to assess correlations of entrainment rates and projected losses with seasonal salinity changes. Measurements of crab-entrainment rates were made aboard the Corps Dredge, Essayons, during August and September 2004 at Flavel Bar, Desdemona Shoals, and the mouth of the Columbia River (MCR). In both 2002 and 2004, rates decreased from the MCR to upriver locations. An appreciable increase, however, was observed in entrainment of age 0+ crab, populations of which are known to vary among years and episodically within a year. In 2004, entrainment rates at Flavel Bar were similar to those at Upper Sands, which also has similar salinity regimes. Desdemona Shoals, on the other hand, had bottom salinities intermediate between those of the MCR and the upriver locations and also had the most variable entrainment rates. The MCR, dominated by oceanic water with bottom salinities above 28 psu for 98% of the summer observations, had consistent entrainment rates (0.05 to 0.10 crab/cubic yard). A statistical procedure for ratio estimation was used to hindcast the 2002 entrainment rates for Flavel Bar from the 2002 and 2004 data, and a modified Dredge Impact Model (DIM) was used to project crab AEL and LF for the dredged volumes proposed for the Channel Improvement Project. The results improved the previous estimates of loss in two ways: first, the estimates were based on site-specific data, and second, the additional 2004 data reduced the variance and narrowed confidence limits. No crab population estimates are available against which to judge the AEL at ages 2+ and 3+, although the commercial crab landings from the regions adjacent to the Columbia River enable a comparison based on the LF values. Regression analysis with the 2002 and 2004 data was used to develop the salinity model for crab occurring in the MCR and navigation channel. The model showed that the natural logarithms of entrainment rates for ages 2+ and 3+ crab are significantly related to salinity. As the salinity decreases from oceanic values, the entrainment rates for this age group decrease exponentially. In areas where salinity observations are lower than 16 psu at least 50% of the time (e.g., near the Astoria Bridge in summer), the model predicts that entrainment rates will be a fraction (less than 6%) of those where salinity remains below 16 psu only 2% of the time (e.g., the MCR in summer). According to results, the entrainment of younger crab are g...