Means of reducing the quantity of water required for the washing unit operation in surimi processing were evaluated. The yield of flesh mechanically dewatered with a screw press and the degree of reduction in trimethylamine oxide (TMAO) content were used as indices of washing efficiency. Variations in water:minced flesh ratio, multiple exchanges and washing conditions near the mean isoelectric point of flesh proteins were techniques evaluated to optimize the yield and quality of washed minced flesh. Single exchanges in potable water produced solids and protein recoveries superior to those observed for multiple exchanges utilizing the same total amount of water. A single exchange of a water:minced flesh ratio of 3.0 produced the highest recoveries (73.32 and 74.07 % of flesh solids and protein, respectively) with the composition characteristics desired. Multiple potable water exchanges (water:minced flesh ratio = 1.0/exchange) at equal total water volume produced desired composition characteristics, but pressed flesh yield was reduced (65.9 and 68.5 % of flesh solids and protein, respectively). Washing (water:minced flesh = 1.0) under acidic conditions near the isoelectric point (pH 4.9-5.3) of the minced flesh protein followed by a potable water wash (water:minced flesh = 1.0) produced recoveries (73.02 and 76.20 % of flesh solids and protein, respectively) superior or equal to all of the wash regimes of equal total water volume. These recoveries were observed even with considerable loss during screening between exchanges and at a water usage 66 % of the single exchange (water:minced flesh ratio = 3.0) of potable water. The moisture content of pressed flesh was increased by both the total potable water:minced flesh ratio and by the number of exchanges to which minced flesh was exposed. Moisture contents mediated by the number of exchanges were a function of water volume and exposure time. Washing conducted in an acidic water:minced flesh mixture at pH 4.9-5.3 (near the mean isoelectric point of flesh proteins) produced pressed flesh with a moisture content 10.09-15.09 percentage points lower than minced flesh wash in an identical (total water:minced flesh ratio; means of separation between exchanges) potable water regime. Washing under acidic conditions produced the most efficient reductions in TMA0 and lipid content (91.4 and 31.1 % of flesh TMA0 and lipid, respectively). Superior reductions were accomplished through pH conditions favoring the extraction of amines and by the greater pressures produced during screw pressing by the altered physical characteristics of the flesh. Pressing between multiple potable water exchanges also showed better reductions than single exchanges involving only one pressing operation. "Folding test" evaluation of heat-set gels (kamaboko) produced from a standard surimi formulation (containing 4 % each of the cryoprotectants sucrose and sorbitol and 0.5 % condensed phosphate) with the addition of 5 % each of potato starch and dried egg white revealed a difference between potable water and acid-washed flesh. Potable water-washed flesh (one exchange; water:minced flesh ratio = 3.0) graded SA (the highest possible grade for the "folding test used by Japanese processors to test the gel strength of surimi". An A-B grade (second and third grade on a five point scale) was assigned gels produced from acid washed flesh. Texture profile analysis of gels revealed acid washed flesh to yield significantly lower gel elasticity (P [greater than or equal to] .007) and cohesiveness (P [greater than or equal to] .0223) than those prepared from potable water washed flesh. The gels were equal in hardness (P [greater than or equal to] .05). The two washing regimes produced gels with comparable (P [greater than or equal to] .05) expressible water contents. Low potable water (a single exchange; waterrminced flesh ratio = 3.0) and acidic (two exchanges; water:minced flesh ratio = 1.0/exchange; exchanges = first pH 5.0-5.3, second potable water) washing regimes improved surimi yield by 26 % and 34 %, respectively, over yields reported for the conventional shore-based processing operations. This was accomplished at water use levels conservatively estimated to be 20 % of those reported for these processes. / Graduation date: 1987