The O I 989-¿ and 1173-¿ intensities in the mid-latitude dayglow are modeled for the conditions of a January 17, 1985, rocket flight. It is found that the branching ratio required to fit the 1173-¿ data is in agreement with the laboratory measurement of 1.3--1.5¿10-4. The results also suggest that the electron impact excitation cross section for the 3s' 3D¿ state in current use may be too large by a factor of ~2--3, in agreement with the most recent laboratory results. If this is indeed the case, then the 1173-¿ triplet is by far the most important branch for the 989-¿ multiplet. ¿ American Geophysical Union 1987