The direct dissociative chemisorption of ethane, propane, n-butane, isobutane, and neopentane on Pt(111) was investigated as a function of the initial translational energy, E-T, polar angle of incidence, theta(i), initial vibrational temperature, and surface temperature using supersonic molecular beam techniques. For each alkane, the initial probability for direct dissociative chemisorption scales with the initial normal energy of the alkanes, E-n=E-T cos(2) theta(i), and is independent of both the surface temperature and initial vibrational energy of the alkanes under the experimental conditions employed. Above initial normal energies of approximately 125 kJ/mol, at constant E-n, the dissociation probability decreases with increasing chain length of the C-2-C-4 linear alkanes; however, the dissociation probability of neopentane is greater than that of isobutane, and both isobutane and neopentane are more reactive than n-butane. By assuming that cleavage of primary C-H bonds is the dominant reaction pathway for all of the alkanes investigated here, the trends in reactivity are best explained by considering the differences in the steric factors for primary C-H bond cleavage for these alkanes. Secondary C-H bond cleavage does appear to contribute to the reactivity of propane and n-butane but only at the highest energies examined. Additionally, the reaction probabilities of each of these alkanes were estimated using a statistical model recently proposed by Ukrainstev and Harrison [J. Chem. Phys. 101, 1564 (1994)]. Assuming cleavage of only primary C-H bonds, the trends in reactivity for ethane, propane, n-butane, and isobutane were qualitatively reproduced by the statistical model; however, except for ethane, which was used to obtain the necessary parameters for the theory, there was poor quantitative agreement, and the predictions for neopentane were significantly lower than the measured values. The model also predicts that the dissociation probability is enhanced by increasing the energy in all vibrational modes, which is inconsistent with the experimental results. Thus, we believe that direct alkane dissociation would be better described using a dynamical rather than statistical approach.