EPR was used to study the influence of formate on the electron acceptor side of photosystem II (PSII) from Thermosynechococcus elongatus. Two new EPR signals were found and characterized. The first is assigned to the semiquinone form of Q(B) interacting magnetically with a high spin, non-heme-iron (Fe²(+), S=2) when the native bicarbonate/carbonate ligand is replaced by formate. This assignment is based on several experimental observations, the most important of which were: (i) its presence in the dark in a significant fraction of centers, and (ii) the period-of-two variations in the concentration expected for Q(B)(•-) when PSII underwent a series of single-electron turnovers. This signal is similar but not identical to the well-know formate-modified EPR signal observed for the Q(A)(•-)Fe²(+) complex (W.F.J. Vermaas and A.W. Rutherford, FEBS Lett. 175 (1984) 243-248). The formate-modified signals from Q(A)(•-)Fe²(+) and Q(B)(•-)Fe²(+) are also similar to native semiquinone-iron signals (Q(A)(•-)Fe²(+)/Q(B)(•-)Fe²(+)) seen in purple bacterial reaction centers where a glutamate provides the carboxylate ligand to the iron. The second new signal was formed when Q(A)(•-) was generated in formate-inhibited PSII when the secondary acceptor was reduced by two electrons. While the signal is reminiscent of the formate-modified semiquinone-iron signals, it is broader and its main turning point has a major sub-peak at higher field. This new signal is attributed to the Q(A)(•-)Fe²(+) with formate bound but which is perturbed when Q(B) is fully reduced, most likely as Q(B)H₂ (or possibly Q(B)H(•-) or Q(B)(²•-)). Flash experiments on formate-inhibited PSII monitoring these new EPR signals indicate that the outcome of charge separation on the first two flashes is not greatly modified by formate. However on the third flash and subsequent flashes, the modified Q(A)(•-)Fe²(+)Q(B)H₂ signal is trapped in the EPR experiment and there is a marked decrease in the quantum yield of formation of stable charge pairs. The main effect of formate then appears to be on Q(B)H₂ exchange and this agrees with earlier studies using different methods.