Pork insulin was labeled by the chloramine T technique (phosphate buffer 0.25 M; pH 7.5; EDTA 0.001 M; chloramine T: 0.2625 mg/ml; sodium metabisulfite 2.4 mg/ml) in a reaction volume of 50 microliters, adding chloramine T every 30 seconds twice (2.1 micrograms in 1 minute) and halting the reaction with 5 microliters metabisulfite. Three fractions were separated in preparative starch gel: F1 (mono-125I-insulin contaminated with cold insulin), F2 (mono-125I-insulin free from cold insulin), and F3 (di-125I-insulin). Insulins with low and high specific activity (iodine/insulin ratios 0.1/1 and 1/1 respectively) were prepared for study purposes, and quality was assessed by means of dose-response curves with antibodies and with liver cells. Specific activity of F2 as obtained from dose-response curves utilizing Scatchard's plot was 323 and 382 mCi/mg. Specific activity of F1 varied according to the extent of contamination with cold insulin. A reduction in the F2 B/F ratio was observed upon iodination with iodine/insulin ratios of 1/1 or in the neighborhood. The mass and immunoreactivity of F3, as well as its B/F ratios were constant, regardless of specific activity. The behavior with antibodies was ratified upon observations on uptake by liver cells and glucose consumption by isolated fat cells. In conclusion, F2 labeled with 0.1/1 iodine/insulin ratios was separated from cold insulin in preparative starch gel, thus increasing its specific activity (360 mCi/mg approximately) without alteration of its immunologic or biologic properties.