Jeremy Gunawardena Ultrasensitivity, or an input-output response which is "steeper" than hyperbolic, was first noticed in the response of allosteric enzymes to metabolic regulators. It enables large changes in flux rates to be controlled by relatively small changes in metabolite concentrations. Its significance in signalling is quite different. I'll explain why an ultrasensitive mechanism seems to be a necessary component of any decision making system. What do these mechanisms look like? There is evidence that some MAP kinase cascades can implement an ultrasensitive response and that this underlies the decisions made by the corresponding signalling pathway. Simulations show that this ultrasensitivity emerges from basic properties of protein phosphorylation. I'll discuss a simplified ("crude") model of this, which can be studied analytically, to see if we can get some intuition for how it works and how it might be used biologically.