The recent contribution on the olivine-spinel geothermometer by Roeder, Campbell, and Jamieson (1979) contains several noteworthy inconsistencies that result in what we consider misleading conclusions. The paper fails to present an up-to-date “reevaluation” of the geothermometer. We note the following points: (1) The criteria of geological reasonableness used by Roeder et al. to evaluate previously proposed versions of the geothermometer are not applied to their own revised model. (2) The experimental results presented are (a) of questionable quality as equilibrium data, and (b) if anything, more supportive of other published calibrations than of the revised thermometer put forward by the authors. (3) Despite the repeated acknowledgment by Roeder et al. of the problems inherent in formulating a thermodynamic model of the geothermometer based on a set of (independently gathered) free-energy data for the spinel end-members, the authors do just that. We conclude that the thermometric Eq. (3) derived by Roeder et al. does not give meaningful temperatures. Reconciliation of their isotherms with those inferred from suites of natural samples would suggest kinetic problems in interpreting the latter, for which there is no evidence. We do not dispute the likelihood that olivine and spinel undergo exchange re-equilibration at subsolidus temperatures in slowly-cooled intrusions. However, we believe that the suggested closure temperatures (in the range 500° 800° C) are inaccurate, since their proposed geothermometer yields temperature-composition relations that are entirely at odds with those indicated by metamorphic assemblages in that temperature range.