A radioimmunoassay (RIA) for estrone (E1) in (human) plasma was developed using an ether extraction, a partition between NaOH and light petroleum, and a TLC as sample purification and an antiserum cross-reacting with E2 for the final assay (Method A1). The reliability criteria are given. In order to simplify this method a highly specific antiserum was developed by using E1-3-hemisuccinate-bovine serum albumin as an antigen. Using this antiserum for the final assay but omitting the TLC (Method B) the E1 concentration in male plasma was 76% overestimated (Method B compared with Method A1). In pregnancy plasma E1 could specifically be determined after a simple ether extraction (Method C). The use of a highly specific antiserum (as determined by cross-reaction studies) for the final assay does not necessarily imply that a chromatographic sample purification can be omitted without loss in assay specificity. This appears to be true mainly in cases where the steroid concentration of the sample is low. Normal values of E1 from 80 healthy adult males were ascertained by Method A1. Age group I (22-61 yr, n = 48) ranged from 1.22-5.60 ng/100 ml, median 2.81; age group II (67-90 yr, n = 32) from 1.55-6.40, median 3.41. The small increase of E1 with age was highly significant.