We compare the results of our JCMT spectral line survey of molecular g
as towards ultracompact HII regions with the predictions of models of
sulphur chemistry in hot cores. We investigate the range of evolutiona
ry models that are consistent with the observed physical conditions an
d chemical abundances, and see to what extent it is possible to constr
ain core ages by comparing abundances with the predictions of chemical
models. The observed abundance ratios vary little from source to sour
ce, suggesting that all the sources are at a similar evolutionary stag
e. The models are capable of predicting the observed abundances of H2S
, SO, SO2, and CS. The models fail to predict the amount of OCS observ
ed, suggesting that an alternative formation route is required. An ini
tial H2S abundance from grain mantle evaporation of similar to 10(-7)
is preferred.