An experimental silicon steel has been used in a detailed kinetic and structural study of the bainite transformation in an attempt to resolve some of the controversies concerning the reaction mechanism. Distinct reaction ‘C’ curves and transformation mechanisms were observed for the upper and lower bainite reactions. The observed set of three minima in transformation kinetics were found to be incompatible with the solute drag explanation of the kinetic Bs temperature. Transmission electron microscopy indicated the growth of both upper and lower bainite by the propagation of displacive subunits, with adjacent nucleation in the latter case. Definite evidence for carbon supersaturation was obtained for the lower bainitic ferrite. The results are best explained in terms of a shear mechanism for the ferritic component of bainite rather than a ledge mechanism (as is observed in Widmanstatten ferrite growth). Carbide precipitation events were also characterized and the evidence suggested that precipitation resulted from the aging of a supersaturated matrix in lower bainite. The evidence also suggests that carbide precipitation events are of secondary importance to the essence of bainite formation. It was further proven that the concept of a metastable equilibrium1 controlling the transition from upper to lower bainite was not applicable to the present steel and indeed, if any metastable equilibrium does exist in any other steel, it does not constitute a general phenomenon and hence is not essential to the bainite transformation mechanism.