Multireference wave functions were used to study the ethylene and 2-butene surface reactions on Si(100) in their lowest energy singlet states. In addition to the diradical pathway, a pi-complex pathway on the ethylene surface was found. The net barrier for the latter process is 4.5 kcal/mol higher than that for the former, making the pi-complex pathway kinetically less accessible. Therefore, although there is a competition between the two initial channels, the diradical path is slightly favored, and rotational isomerization is possible. However, since the initial potential energy surfaces of the two channels are different, depending on experimental conditions, the branching ratio between the two channels may change. Consequently, the combined effects that would favor one channel over the other may not derive directly from the initial reaction barrier. This provides an explanation of the experimental controversy. As a result, the final distributions of surface products may depend on the experimental kinetic environment, especially when the population change due to the rotational isomerization is expected to be very small. A significantly different reaction channel is found in the 2-butene surface reaction on Si(100), in which a methyl hydrogen easily transfers to the surface yielding a new type of surface product other than the expected [2 + 2] cycloaddition product, with a comparatively small activation barrier. Consequently, the overall surface reactions of ethylene and 2-butene may be quite different. Therefore, direct comparisons between ethylene and 2-butene experimental results would be very useful.