The dependence of the rate of hydrogenolysis of C2H6, C3H8, and n-C4H10 on 0.3% Pt/Al2O3 (AKZO CK303, EUROPT-3) has been determined at several temperatures, using short reaction pulses to minimise deactivation by carbon deposition. Rate maxima, which with C2H6 occur at very low H-2 pressures, become broader and move to higher H-2 pressures as the chain-length or temperature increases. Criteria for selecting an appropriate rate expression with which to model the results is discussed in detail : one is selected that is based on the competitive chemisorption of the reactants, that of the alkane requiring the loss of more than one H atom to activate it. For each data set, optimum values of the constants of the rate equation are obtained by computation. Changes in the form of the kinetic curves are described by (i) values of the rate constant k(1) and the equilibrium constant for alkane chemisorption K-A that increase with alkane chain length and with temperature, (ii) values of K-H, the adsorption coefficient for H-2, that decrease with alkane chain-length but are not very temperature-dependent, and (iii) a decreasing degree of the needful alkane dehydrogenation as the chain-length increases. For C3H8 and n-C4H10, true activation energies are respectively 82 and 76 kJ mol(-1), and enthalpy changes for alkane chemisorption are 88 and 79 kJ mol(-1). Apparent activation energies increase with Hz pressure in consequence of the Temkin equation and show compensation effects. Rate dependences on H-2 pressure were measured at a single temperature using 0.3% Re-0.3% Pt/Al2O3 (AKZO CK433, EUROPT-4) : rates were faster than for Pt/Al2O3, and values of k(1), K-A, and K-H were larger. Measurements were also made with catalysts partially deactivated by carbon deposition; slower rates were associated with lower values of k(1) and K-A. Changes in product selectivities with operating variables are recorded, but are only significant for PtRe/Al2O3, and for n-C4H10 on Pt/Al2O3, where at 547 K the extent of internal C-C bond fission decreases as Hz pressure increases. The difficulty of devising a simple rate expression to embrace all the experimental observations is discussed.