The two forms of the g approximate to 4.1 signal recently identified in photosystem II (Smith, P J.; Pace, R. J. Biochim. Biophys. Acta 1996, 1275, 213) have been simulated at several frequencies as near-axial spin 3/2 centers. In both cases, an explicit spin coupling model is assumed, involving two magnetically isolated Mn pairs, one for each signal type. For that pair assumed to give rise to the spin 1/2 multiline signal as the ground state, the modeling of the first-excited-state 4.1 signal gives estimates of the fine structure parameters for the individual Mn centers and the exchange coupling constant for the pair. The fine structure terms suggest that one Mn ion is a conventional Mn-III ion in a highly axially distorted environment. The other Mn center, which is formally spin 3/2, is unlikely to be a conventional Mn-IV ion, but rather a Mn-III-radical ligand pair, strongly antiferromagnetically coupled to give a net spin 3/2 state. The coupling between this Mn-radical center and the other Mn-III is weak (J = -2.3 cm(-1)) in the absence of alcohol in the buffer medium, as determined earlier (Smith and Pace). The model is shown to be quantitatively consistent with the behavior of other signals proposed to arise from this coupled dimer. Comparison of our own data with those of others (Haddy, A.; ct al. Biochim. Biophys. Acta 1992, 1099, 25-34) on one-dimensionally ordered photosystem IT samples shows a generally consistent orientation of the molecular axis system for the dimer in the membrane plane. The second 4.1 signal, which exhibits ground-state behavior, may be simulated at X- and Q-band frequencies as an isolated system with D = +1.1 cm(-1) and E/D = 0.037. The spin center is suggested to arise from a radical-bridged Mn homodimer, and the modeling parameters have been interpreted within this framework. The resulting proposal, involving two isolated dimers for the Mn organization within the oxygen evolving center, is critically examined in the light of recent work from other groups.