A new approach to the prevention of sickling in vitro by use of the bifunctional crosslinking reagent, dimethyl adipimidate, is described. Prior treatment of sickle erythrocytes with dimethyl adipimidate will inhibit sickling in completely deoxygenated erythrocytes. Treated erythrocytes do not demonstrate the potassium loss and viscosity increase that usually accompany sickling. The oxygen affinity of hemoglobin in these cells is increased independently from changes in the concentration of 2,3-diphosplioglycerate. The (v/v) formaldehyde in saline for fixation. Five hundred cells were counted by described criteria (6). The viscosity of dimethyl adipimidate-treated and control cells was measured in a Wells Brookfield cone plate microviscomneter model LVT, adapted to maintain controlled atmospheric gas tension (7). After adjustment of the hematocrit to 40 i 1%, the blood was equilibrated for 1 hr with humidified 95% N2 and 5% CO2 in an IL tonometer model no. 237, then transferred in a closed system to the viscometer. The viscosity was measured at shear rates ranging from 4.5 to 90 sec' during continued exposure to nitrogen.Studies in vitro of erythrocyte metabolism and of net K + loss were conducted by incubation at 370 of thrice-washed erythrocytes in Krebs Henseleit buffer with added glucose (10 mM) in the presence of 0.1 mM ouabain. The samples were continuously equilibrated with moistened gas mixtures consisting of nitrogen plus 20%, 3.3%, or 0% oxygen, and pH was maintained at 7.45 4i 0.05 by varying the CO2 concentration with a gasometric pH stat (8). Supernatant K+ concentration was determined by flame photometry. 2,3-Diphosphoglycerate (9) concentrations in erythrocytes were determined in perchloric acid extracts.The oxygen affinity of whole blood, expressed as p50 (that oxygen tension at which hemoglobin is half-saturated), was determined by measurement of hemoglobin oxygen saturation and P02 after equilibration of the samples with gas mixtures containing 2.2%, 2.8%, 3.2%, and 3.9% oxygen plus 5% CO2 and the remainder nitrogen. The p50 was calculated by a best fit analysis of these data and corrected to a pH of 7.4 (10).Hemolysates and globin were prepared by standard methods. Gel filtration in Sephadex G-100 was performed by the method of Andrews (11). The column was calibrated with aldolase (molecular weight 158,000), ovalbumin (molecular weight 45,000), and chymotrypsinogen A (molecular weight 25,000). Hemoglobin electrophoresis was performed on cellulose acetate strips at pH 8.65 in a Beckman microphor system. Barbital buffer, 25 mMI, pH 8.0, was used for electrophoretic separation of globin chains and the cellulose acetate strips were soaked in the same buffer with 8 M urea and 5 mM 2-mercaptoethanol. Spectra were recorded with a Cary 14 recording spectrophotometer.