Experimental results are presented for (e, 2e) ionization of the Ar(2p) orbital in a highly asymmetric geometry. Despite the magnitude of the momentum transfer being only 0.5 au, it is found that the triple differential cross section is not symmetric about the direction of momentum transfer. We have performed distorted-wave Born approximation calculations which not only do not show the observed symmetry breaking but are also little different from the first Born results. This suggests that distortion of the fast electron by the atom is not significant. We conclude that a new mechanism must be invoked to explain the data which depends on a two- or higher-step process.