Fractionation of testicular, epididymal, and serum extracts containing rat androgen-binding protein (rABP) on a Concanavalin A-Sepharose (Con A) column resolved two peaks of immunoreactive protein. The first peak was present in the void volume, and the other was bound by the column and specifically eluted by alpha-methyl-D-glucoside. These two peaks of immunoreactive rABP have been designated form I and form II for the portions of rABP that do not and do bind, respectively, to Concanavalin A. In the course of studying this heterogeneity, we observed that the distribution of the two forms of rABP was the same in the blood and cytosols prepared from testis and epididymis of young rats before the formation of the blood-testis barrier; that is, the ratio of form I to form II ranged from 1:1 to 1:2. Similar heterogeneity was observed in extracts of the reproductive tract from mature animals. However, the blood of adult rats contained reduced amounts of form I relative to form II, so that their ratio was about 1:5. Subsequent studies of infertile rats heterozygous for the Hre gene (Hre/ +), in which total rABP secretion was decreased, and of their normal littermates, indicated that the reduced amount of form I ABP in the sera of mature rats is typical of adult animals regardless of strain or genetic abnormality. The reduced amount of form I relative to form II observed in the blood of adult rats could result from either reduced secretion or increased metabolic clearance of form I in the blood compartment. To distinguish between these possibilities, the blood clearance of the two forms was estimated after orchiectomy. The disappearance rate of form I was not significantly different from that of either form II or unfractionated serum. These results are consistent with reduced release into blood of form I relative to form II rABP rather than increased clearance of form I in adult animals.