An expression for the sound transmission loss of an obliquely incident plane wave striking a double partition consisting of two identical septa separated by an air space has been derived. It is assumed in the theory that each septum is infinite in extent and behaves like a rigid mass. Previous investigators have shown that in the case of a normally incident wave, there exists a characteristic frequency fc for the air-mass sandwich for which there is total transmission of sound, given by fc = 12π(2ρc2md) 12, where ρ is the density of air, c the velocity of sound in air, m the mass per unit area of the septum, and d the air space thickness. In the present investigation it is shown that Eq. (1) is a special case of a more general expression which holds for any angle of incidence θ, i.e., fθ = 12π cos θ(2ρc2md)12 (2) or, a wave incident at angle θ will be totally transmitted at frequency fθ, given by (2). Since in a reverberant sound field energy strikes the partition at random angles of incidence, it follows that total transmission will not be observed at any frequency. While there may be a particular component of the sound field which has zero transmission loss, there will always exist other waves which do suffer a loss. The expression for the transmission loss corresponding to a wave at angle θ may be integrated over all angles of incidence in accordance with the usual reverberant sound field statistics. The resulting average transmission loss expression may be written as a function of two non-dimensional parameters: (1) the mass ratio, μ = ρd/m, (3) i.e., the ratio of mass per unit area of air in the air space to the mass per unit area of the septum and (2) the frequency ratio parameter, r = f2/fc2. (4) A family of transmission-loss versus frequency-ratio curves with μ as a parameter, having general applicability to any double partition, is plotted. Comparison of the results of this theory with actual double walls used in building construction shows that in some cases the latter are poorer than the theoretical predictions, possibly due to: (1) solid sound conducting paths existing between septa due to studs, and the like, (2) the existence of transverse standing waves in the air space, (3) the occurrence of bending waves in the septa, (4) and the appearance of resonance frequencies of the septa. The results of measurements on some experimental panels designed to estimate the extent to which these worsening influences are operative will be discussed.