The ideal flow predicted for an inviscid adiabatic shock tube is not realized in experiments. A correction of the ideal theory for the viscous stresses and heat losses arising from the interaction of the gas flow with the tube walls is calculated. The boundary layer is regarded as a distribution of sinks of energy and momentum. The actual distribution is approximated by that behind a constant strength shock front and is calculated for laminar flows. The change in shock strength caused by a single sink located at a fixed distance behind the shock front is determined and this result is integrated over the distribution previously calculated. Interactions between the expanding compression chamber gas and the wall are neglected. It is shown by comparison with experiment that the interactions treated account for the major part of the attenuation of the shock front. It is pointed out that more data are required to evaluate precisely the assumptions used in the calculation.