Two very different mechanisms of folding have been proposed from experimental studies of disulfide formation in reduced ribonuclease A. (1) A pathway in which the rate-limiting step separates fully folded protein from all other disulfide intermediates and occurs solely in three-disulfide intermediates. (2) A multiple pathway mechanism with different rate-limiting steps for each pathway. The various rate-limiting steps involve disulfide breakage, formation, and rearrangement in intermediates with one, two, three, and four protein disulfides. To distinguish between these two mechanisms, we have carried out further studies of both unfolding and refolding. Refolding of reduced ribonuclease A requires three-disulfide intermediates to accumulate; negligible refolding occurs when only the nearly random one- and two-disulfide intermediate species are populated. Therefore, no rate-limiting steps of the type postulated in mechanism (2) occur in intermediates with one and two protein disulfides. Unfolding and disulfide reduction is an all-or-none process; no disulfide intermediates accumulate to detectable levels or precede the rate-limiting step. Mechanism (2) requires that such intermediates precede the rate-limiting step and accumulate to substantial levels. The different proposals were shown not to result from the use of different solution conditions or disulfide reagents; the two sets of data are not inconsistent. Instead, the inappropriate mechanism (2) resulted from an incorrect kinetic analysis and misinterpretation of the kinetics of disulfide formation and breakage.