 Solution

Secondary amines normally favor non-geminal disubstitution, putting the second amino substituent on a different phosphorus. The first installed $\mathrm{N}{\mathrm{R}}_2$ group donates electron density to the P–N framework and deactivates its own phosphorus toward further attack. It is also bulkier than an initially installed primary-amino group. Crucially, $\mathrm{P}-\mathrm{N}{\mathrm{R}}_2$ has no N–H proton, so the proton-abstraction/chloride-elimination pathway that can direct a primary amine geminally is unavailable. Direct attack at a less-deactivated $\mathrm{PC}{\mathrm{l}}_2$ site is favored. The resulting non-geminal products can have cis or trans relative orientations around the ring; non-geminal substitution alone does not select one of these stereoisomers. These are useful trends, with known substituent-dependent exceptions in experimentally characterized mixed derivatives.^[ref]