Reduction of (+)-[3,3−²H₂]camphor ([3,3−²H₂]1) with lithium, sodium or potassium in ammonia and a co-solvent gave; 1) the enolate of [3,3−²H₂]1 and the alcoholates of (−)-[2,3,3−²H₃]isoborneol ([2,3,3−²H₃] and (+)-[2,3,3−²H₃]borneol ([2,3,3−²H₃]3); 2) the alcoholates of [3,3−²H₂]2 and [3,3−²H₂]3; 3) the dialcoholates of the pinacols [3,3,3′,3′−²H₄]4 and [3,3,3′,3′−²H₄]5. It is proposed that these are formed from the ketyls [3,3−²H₂]1- M+, by: 1) disproportionation; 2) H-atom abstraction from the medium; 3) dimerization. Protonation upon work-up afforded [endo−32H]1, [2,3,3−²H₃]2, [2,3,3−²H₃]3,[3,3−²H₂]2, [3,3−²H₂]3, [3,3,3′,3′−²H₄]4 and [3,3,3′,3′−²H₄]5. Pinacol [3,3,3′,3′−²H₄]5 was the main and pinacol [3,3,3′,3′−²H₄]4 a minor product in the reductions with lithium and both were minor products in the reductions with sodium; pinacols were not formed in the reductions with potassium. Parallel reductions of 1, unlabeled, analogously led to 2, 3, 4 and 5, and the ratios 2/3 differed from the ratios ([2,3,3−²H₃]2+[3,3−²H₂] 2/([2,3,3−²H₃]+[3,3−²H₂]3) under certain conditions. Different values for these ratios were found in the reductions with each metal, all of which corresponded to low overall diastereoselectivities. Reactions 1 and 3 persisted when the reductions were carried out in ammonia/water/co-solvent mixtures and the enolate formed via reaction 1 was protonated and the resulting [endo-3-2H]1 recycled. Reaction 2 cannot be monitored under these conditions. Reactions 1 and 3, and by inference also reaction 2, were almost completely suppressed when analogous reductions were carried out in the presence of ammonium chloride, [3,3−²H₂]2 and [3,3−²H₂]3 being obtained almost exclusively, in a 6: 94 ratio. It is proposed that the mechanism outlined in House [1] was dominant when, and only when, ammonium ion was the proton source; it may have competed when water was the proton source.