The R2 subunit of E. coli ribonucleotide reductase contains a tyrosyl radical-diiron(III) cofactor which assembles spontaneously when apo R2 is mixed with Fe2+ and O-2. The kinetic/spectroscopic characteristics of this assembly reaction were previously shown to depend on the Fe2+/apo R2 ratio (Bollinger, J. M., Jr. et al. Science 1991, 253, 292-298). In the case of the reaction carried out with limiting Fe2+ (Fe2+/apo R2 2.0-2.4), two intermediates were proposed to accumulate. On the basis of its broad absorption feature at 560 nm, the first intermediate was suggested to contain a mu-peroxodiferric cluster. The one-electron oxidation of tyrosine 122 by this cluster was postulated to generate the second intermediate, which was proposed to contain the tyrosyl radical (*Y122) and the diferric radical species (Ravi, N. et al. J. Am. Chem. Soc., first of three papers in this issue; Boilinger, J. M., Jr. et al. J. Am, Chem. Sec. 1991, 113, 6289-6291). In this study, stopped-flow absorption, rapid freeze-quench electron paramagnetic resonance, and rapid freeze-quench Mossbauer spectroscopies have been used to characterize in detail the kinetics of the assembly reaction carried out with limiting Fe2+, The time course of development and decay of the 560 nm absorption transient, and the quantities of diferric radical species, *Y122, and product diferric cluster as functions of time, are consistent with the proposal that the 560 nm absorbing species generates *Y122. The data indicate, in contrast to our previous hypothesis, that the 560 nm absorbing species is not iron based. It is proposed, instead, that the species is a tryptophan radical cation produced by the one-electron oxidation of W48. The results of this study are consistent with our previous proposal that Fe(IV) intermediates are not involved in generation of *Y122.