The reported relative biological effectiveness (RBE) for thermal neutrons has a large range (5-51, for cytogenetic endpoints), which can confound radiation protection decision-making. To determine whether thermal neutron spectra can influence RBE, the RBE of reactor-derived thermal neutrons of average energy 31 meV was evaluated in human peripheral blood lymphocytes using two classical DNA double-strand break endpoints: the dicentric chromosome assay (DCA) and the cytokinesis-block micronucleus assay. Dose-response curves for 41 to 408 mGy revealed a preference for linear regression. Maximum RBE (RBEM) values of 6.7 ± 0.9 and 4.4 ± 0.7 were calculated for the DCA and the micronucleus assay, respectively. These 31 meV RBEM values were significantly lower than our prior results for 64 meV thermal neutrons, which yielded a DCA RBEM of 11.3 ± 1.6 and a micronucleus RBEM of 9.0 ± 1.1. Dose-specific RBE values decreased with increasing dose for both assays. Microdosimetry simulations demonstrated similar quality factor values for both thermal neutron spectra. Dose deposition differences on the cellular scale, the difference in dose rate between irradiation configurations, or a not-yet understood phenomenon may be responsible for the RBE difference between the 31 and 64 meV thermal spectra. These findings indicate that the currently accepted radiation weighting factor wR value of 2.5 for thermal neutrons may underestimate the radiation detriment to small or shallow tissue targets including the lens of the eye.