There are stringent conditions on any chemical process that might be responsible for the population inversion and the observed properties of the OH maser. So stringent are they that no such process is likely to be found. On general grounds, the high densities and narrow line widths would imply unsaturated amplification. This is most sensitive to small differences in protostar conditions. Yet the intensities are comparable for the different emitters, despite different line widths and polarizations, which attests to a variety of conditions. The emissions from the excited rotational states become progressively weaker at higher positions in the rotational ladder, not, as expected, for excitation temperatures greater than 700°K and for the unsaturated amplification that is required by a total density greater than 1010 cm-3. The time variations, polarization, and preference for certain hyperfine components are inexplicable at high densities. Furthermore, Zeeman splittings would be much larger than observed. However, in the outer parts of the cloud, with optical pumping from a shocked OH layer and heated grains, the conditions are favorable for maser emission with the observed properties.