It has previously been demonstrated that both [(C5Me5)Ir(PMe3)(CH=CH2)H] and [(C5Me5)Ir(PMe3)(H2C=CH2)] are formed when [(C5Me5)Ir(PMe3)] is thermolytically generated in the presence of ethylene. At higher temperatures, the vinyl hydride is converted to the eta2-ethylene adduct. Density functional theory has now been used to investigate this reaction, using the B3LYP functional, two types of basis sets (LanL2DZ and TZV*), and two models of the [(C5R5)Ir(PR3)] species (R=H and CH3). The study consists of full optimizations of local minima, first-order saddle points, and minimum energy crossing points (MECP). The experimental results are best accounted for by considering both singlet and triplet spin surfaces. The relative energies of singlet [(C5R5)Ir(PR3)(CH3)H], [(C5R5)Ir(PR3)(CH=CH2)H], and [(C5R5)Ir(PR3)(H2C=CH2)] are in good agreement with experiment, as is the calculated barrier for the conversion from the vinyl hydride to the eta2-alkene complex. However, the singlet surface alone fails to explain the experimentally observed product ratio, or the intermediate inferred from experimental isotope effect studies. Locating the MECP between singlet and triplet surfaces indicates that the thermolysis of the singlet alkyl hydride precursor directly forms triplet [(C5R5)Ir(PR3)]. The weak vanderWaals adduct of triplet [(C5R5)Ir(PR3)] and ethylene is proposed to be the key intermediate in the overall reaction. The interchanging of the available ethylene C-H bonds in this triplet sigma complex accounts for the observed kinetic isotope effects, and partitioning between alkene pi-complexation and C-H bond activation may also occur from this common intermediate. The possible role of steric factors and molecular dynamics are also discussed.