Neutron angular distributions from the ($p, n$) reactions in ${\mathrm{C}}^{13}$, ${\mathrm{B}}^{11}$, and ${\mathrm{Be}}^{9}$ have been measured using a long-counter detection technique in conjunction with the Livermore 90-inch variable-energy cyclotron. Proton energies ranged from threshold (2.0 Mev to 3.2 Mev) up to 5.7 Mev. The aim here was to find qualitative experimental evidence bearing on the direct reaction mechanism proposed by Bloom, Glendenning, and Moszkowski wherein the ($p, n$) reaction connecting the ground states of mirror nuclei should go via a direct mode which is derived principally from the residual two-body interaction between the incoming proton and the bound neutron (or neutrons). It is found that the experimental evidence supports this hypothesis in that the angular distribution changes slowly in the direction of increasing complexity with increasing energy, largely ignoring the occurrence of resonances except in their immediate vicinity. Also a tentative grouping by pairs of the ($p, n$) angular distributions for (${\mathrm{C}}^{13}$, ${\mathrm{N}}^{15}$) and (${\mathrm{Be}}^{9}$, ${\mathrm{B}}^{11}$) shows marked similarities between the members of each pair in conformity with the twin-reaction picture stemming from the same theory. Finally, preliminary results are presented of an IBM-704 computation program using a distorted-wave Born approximation theory formulated originally by Glendenning. The comparison between theory and experiment, although based on early returns, is in general encouraging. It is found that a triplet-singlet interaction strength ratio is required here which is about $\frac{2}{3}$ of that derived from the Gammel-Thaler phenomenological potential. However, in view of the basic differences between the free and the bound two-body problem it is felt that more knowledge will be required in order to properly compare the present results with the free-scattering analyses.