A theoretical model describing acoustic attenuation in a flexibile walled duct (having neither internal acoustic treatment nor external lagging) which passes through a reverberant space is described. This model is kept as simple as possible, consistent with retaining the salient physical features of the problem. Both "breakout" and "breakin" sound power, as well as acoustic dissipation by damping in the duct walls, are taken into account. Predictions of attenuation, based on the model, are compared with published experimental data and are observed to embody the principal features of the measurements; quantitative agreement is reasonable, considering the sensitivity of the attenuation to system parameters. The composite nature of the attenuation mechanism is made clear by the results.