the catalytic decomposition of 2-propanol has been investigated by a micropulse reactor technique in the temperature range from 461 to 547 k on nickel oxide and alumina doped nickel oxide. nickel oxide is active for the dehydrogenation of 2-propanol, yielding acetone, but the introduction of a1[superscript 3+] ions leads to the development of some dehydration activity; nickel alumlnate is purely a dehydrating catalyst. x-ray diffraction data indicate the formation of the nickel aluminate at the higher dopant concentrations. the variations in catalyst activity and selectivity as a function of catalyst compositions have been roughly divided into three regions: region i, in which there is a significant change in activity and selectivity with very small concentrations of al[superscript 3+] ions; region ii, in which the activity and selectivity remains virtually constant with changes in composition; region iii, in which there is a marked increase in activity and dehydration selectivity with increasing concentrations of al[superscript 3+] ions the behavior in region i is believed to be associated with electronic changes in nickel oxide brought about by the substitutional incorporation of al[superscript 3+] ion, while the behaviour in region iii is ascribed to increasing amounts of nickel aluminate in the surface zones of the samples. self-poisoning in the dehydrogenation reaction has been observed and an equation accounting for the self-poisoning kinetics has been proposed. it is suggested that in the dehydrogenation reaction each alcohol molecule uses one active site on the catalyst, while self-poisoning causes elimination of three sites per deactivating event. strong chemisorption of acetone on the surface is considered to be responsible for the self-poisonlng of the dehydrogenation reaction. the conclusions are supported by conductivity studies of the catalysts under reaction conditions. a characteristic feature of the dehydration reaction is the deviation from linearity in the arrhenius plots at higher reaction temperatures. isotope effect measurements show that this is not due to a diffusion controlled reaction. the deviation is believed to be associated with the decomposition of hydroxyl groups on the catalyst surface and investigations of the thermal dehydroxylation of mixed nickel-aluminium hydroxides support this proposal.