Ab initio calculations using a (valence) double-zeta pseudopotential (DZP) basis set, with (MP2, QCISD) and without (SCF) the inclusion of electron correlation, predict that the reactions of silyl, germyl, and stannyl radicals with silane, germane, stannane, trimethylsilane, trimethylgermane, and trimethylstannane proceed via transition states of C-3 upsilon or D-3d symmetry in which the attacking and leaving radical centers adopt a collinear arrangement. For reactions involving (SiH3)-Si-., (GeH3)-Ge-. and (SnH3)-Sn-., energy barriers of between 23.4 ((SiH3)-Si-. + SnH4) and 86.0 ((SnH3)-Sn-. + SiH4) kJ.mol(-1) are predicted at the QCISD/DZP//MP2/DZP (+ ZPVE) level of theory. Specifically, the identity exchange reaction involving silane and the silyl radical is predicted to involve an energy barrier of some 53.6 kJ.mol(-1) at the highest level of theory; in good agreement with the available experimental data. The similar reactions involving germyl ((GeH3)-Ge-. + GeH4) and stannyl radicals ((SnH3)-Sn-. + SnH4) are predicted to have energy barriers of 47.0 and 38.9 kJ.mol(-1), respectively, at the same level of theory. Inclusion of alkyl substitution on one of the heteroatoms in each reaction serves to alter the position of the hydrogen atom undergoing translocation in the transition state when compared with the unsubstituted series; the reactions of H3Y. with Me3XH become "later" when compared with the analogous parent reaction (H3Y. with XH4). Energy barriers of between 23.8 ((SiH3)-Si-. + Me3SnH) and 98.3 ((SnH3)-Sn-. + Me3SiH) kJ.mol(-1) are predicted at the MP2/DZP (+ ZPVE) level of theory. The mechanistic implications of these computational data are discussed.