To estimate the biomass of six species, the stratified design used at present for demersal coastal fish stocks is compared against other design alternatives. Density data (t mn-2 between 34ºS and 42ºS, from the coast line to 50m depth, corresponding to whitemouth croaker (Micropogonias furnieri), patagonian smoothhound (Mustelus schmitti), stripped) weakfish (Cynoscion guatucupa), angular angel shark (Squatina guggenheim), brazilian flathead (Percophis brasiliensis) and skate (Sympterygia bonapartii) collected in surveys carried out in spring and winter in the Argentine Sea (SW Atlantic), were analyzed. The spatial distribution of the species was strongly contagious and influenced by bottom salinity. Different marine zones that showed special salinity features for a long time were defined to test bounded strata accordingly. The simple random design caused, in general, greater uncertainty than any of the stratified designs when using the 17 fixed strata surveyed during the last cruises or when two ‘new designs’ that included 7 strata, different for spring and winter, were adopted. Although no stratified scheme showed superiority to assess all and every species, their comparative advantages are discussed. Assessments made with the design used at present seem to be less efficient in winter thus suggesting that, in the future, surveys should be carried out in spring. No great differences among designs were found when the efficiency gain-sample size relation was studied and that resulted very low above 150 trawls if compared to the cost of using the research vessel. Published evaluación de stock, relevamientos pesqueros, pesquería costera, estratificación, análisis espacial, peces demersales, modelado, Micropogonias furnieri, Mustelus schmitti, Cynoscion guatucupa, Squatina guggenheim, Percophis brasiliensis, Sympterygia bonapartii, corvina, gatuzo, pescadilla de red, pez ángel, pez palo, raya marmorada