Insulin decreased markedly the adenylyl cyclase activity associated with fat cell membranes purified by centrifugation in sucrose gradients. The hormone effect was not readily evident in crude membrane preparations. The kinetics of this effect indicate that some time was required for the onset of the insulin-induced inactivation. This lag period decreased when the insulin concentration was increased. The hormone dose dependence for adenylyl cyclase inactivation measured at a fixed time (3 min) showed a 10 to 15% decrease in activity at 1 to 30 muU per ml insulin; 30 to 40% at 100 to 1000 muU per ml; and 75% at 0.1 U per ml. The insulin effect was completely abolished by 0.1 mM GMP-P(NH)P, 10mM fluoride, or 50 ng per ml glucagon, or by increasing the Mn++ concentration to 4 mM. In addition, it was partially reversed by the addition of a fraction from the sucrose gradient, which contained soluble factors. The kinetics of the adenyl cyclase-catalyzed reaction were studied using ATP or AMP-P(NH)P as adenylyl donor, and Mn++ or Mg++ as divalent cation, in the absence or presence of insulin. With ATP and Mg++ there was a striking reduction of the transient reaction rates after 1.5 min of incubation. Under these conditions the insulin effect was not evident. On the contrary, with ATP and Mn++ this spontaneous reduction of activity was less evident; however, in the presence of insulin there was a clear and marked reduction of the transient reaction rate measured after 1.5 min of incubation. With AMP-P(NH)P the kinetic data were qualitatively similar to those observed with ATP. It is concluded that under certain assay conditions adenylyl cyclase may be converted to an inactive enzyme form, and that such a conversion is more evident in the presence of Mg++ than with Mn++. In the latter case, insulin appears to enhance the rate of this conversion.