We reconsider the rotational analysis of the absorption electronic A-X
and D-X band systems of OH and show that interstellar observations of
OH in front of bright stars have been interpreted with oscillator str
engths that are too small by a factor of 2. However, the column densit
ies reported from the various observations are approximately right, as
a result of the neglect, in their interpretation, of the two differen
t initial levels, which compensates the previous error.