Renewed interest in the measurement of fluid velocities in boundary layers by the hot-wire method has called attention to a difficulty that arises when the wire approaches a conducting body of lower temperature. The additional convection caused by the body may be large and is not easy to assess by experiment alone, while theoretical investigation is somewhat complicated. Some time ago, the present authors obtained an approximate solution of the cognate problem of a hot wire moving adjacently to a cooling flat plate through an inviscid fluid, assumed to be initially stationary, and found a measure of agreement with corresponding experiments which they conducted in air. It is anticipated that the results should provide reliable corrections in the practical problem when the wire is small compared with the thickness of the boundary layer and sufficiently clear of the surface of the body. Section 1 of the present paper gives an account of the analysis, which is in the nature of a development of the work of King and of Piercy and Winny. Section 2 records observations of the cooling of a hot wire whirled through air round a large circular brass cylinder of normal temperature. The results of the experiments are compared with the theory by means of examples worked from the final equations.