The experimental literature on the pyrolysis of C,H2 is reviewed and summarized. These observations divide naturally into three temperature regimes: (i) T < 1100 K, where the homogeneous reaction is a molecular polymerization; (ii) 1100 < T < 1800 K, where the process is still dominated by a molecular polymerization, but a fragment radical chain is clearly involved; and (iii) T > 1800 K, where a fragment chain carried by C2H and H drives a polymerization to polyacetylenes. A combined molecular/fragment chain mechanism is developed and applied to the modeling of three shock-tube studies which span the two higher-temperature regimes. The core of the mechanism is a set of 5 mainly molecular reactions whose rates are derived from an application of unimolecular rate theory and detailed balance to recent measurements on the decomposition of vinylacetylene. The combination of this core with consequent fragment radical reactions provides a satisfactory description of most features of the chosen experiments including several-isotope exchange, hydrogenation to CzH4, and CsH, formation-which are almost entirely the result of fragment radical chain reaction. Possible detailed paths for the molecular dimerization are suggested and some possible molecular paths for the "Berthelot synthesis" of benzene are also proposed.The last 30 years has seen a tremendous increase in our understanding of the pyrolysis and combustion reactions of hydrocarbon fuels. Detailed, very complete, and quite successful mechanisms have now been devised for simple fuels [l-31. Despite all this progress there remain many unresolved questions of mechanism, some of the most serious being those which involve the reactions of acetylene, which is of course a major product, or at least an important intermediate, in all high temperature hydrocarbon pyrolysis and combustion.The C,H2 pyrolysis mechanism has been particularly unyielding. Here both molecular (possibly diradical) and fragment-radical (monoradical) chain mechanisms have been proposed, but neither seems consistent with all the data, and it remains unclear how either process might be initiated.This article presents an attempt to arrive at a fully consistent model of the pyrolysis mechanism, albeit one which only treats the early stages of the process. The basis for this model lies in the recent resurrection of the molecular reaction concept by Duran, Amorbieta, and Colussi (DAC) [41, and Kiefer, Mitchell, Kern, and Yong (KMKY) [51. The core of the paper is a detailed, quantitative modeling of some recent shock-tube data [6-91, using a mechanism which contains both molecular and fragment-radical chain reactions. This is followed by a consideration of the full range of