A previously unpublished very high resolution solar spectrum has been analyzed to determine the relative contributions of O II and O III at 834 angstrom. The resulting ratio of O II to O III emissions is found to be 0.35, which is lower than derived from lower resolution data or expected from calculations based on standard emission measure theory. In order to reproduce the ratio observed in the high-resolution spectrum, it is proposed to lower the emission measure in the vicinity of 40,000 K by a factor of 2. This would also reconcile differences among apparently discrepant lines of C II and is consistent with the physical interpretation that the emission measure between 10(4) and 10(5) K is comprised of components arising from physically different regimes in the solar atmosphere.