As a result of a large number of experiments on the Hurtley reaction the following tentative conclusions are drawn: (a) the best and most convenient solvent is ethanol; (b) the reaction succeeds best with aromatic o-bromocarboxylic acids although o-iodo acids can be expected to give low yields; (c) replacement of the carboxyl group by any other functional group prevents the reaction; (d) a copper species, probably copper(l), is an essential catalyst. The reactivity of 8-bromo-1-naphthoic acid, which is almost identical with that of o-bromobenzoic acid, is evidence against the intermediacy of a planar resonance hybrid structure. The geometry of the bromo acid-copper intermediate appears to be of paramount importance.