The results of a 1974 survey of albumin measurements as performed by more than 1,300 laboratories are presented. The most widely used methods are the dye-binding technics: bromcresol green (BCG) and 2-(4'-hydroxyazobenzene) benzoic acid (HABA). These are followed by electrophoresis and salt fractionation. All methods yielded comparable albumin concentrations except electrophoresis, which manifested a consistent low bias. This close agreement is attributed, in part, to the normal-range concentration of albumin in the test specimen. Type of standardization, i.e., commercial serum, bovine serum albumin, human serum albumin, or pooled serum, did not appear to be a factor in the estimation of albumin in the normal serum submitted for analysis. Surprisingly, interlaboratory variation, from method means, was the lowest for salt fractionation and electrophoretic technics.