A quantitative study of saccadic eye movements in the cat was undertaken to attempt to account for the high degree of variability in the trajectory of feline saccades compared with the more stereotyped monkey saccades. Cats were trained to make oblique saccades so that a large variety of saccadic component amplitudes, maximum velocities, and durations could be obtained. The horizontal and vertical components of oblique saccades were either stretched or compressed, relative to equal amplitude movements without orthogonal components, so that the two components were nearly equal in duration. On average, the components began approximately synchronously but were more asynchronous in their termination times. Consistent with the stretching or compression of a saccadic component, there was the complementary decrease or increase respectively of its maximum velocity. The product of maximum velocity and duration was constant for saccades of a given size. Furthermore, the constant relating maximum velocity and duration was a linear function of saccade size. Therefore, any saccade size was uniquely determined by knowing both duration and maximum velocity while knowing either alone resulted in considerable ambiguity in specifying saccade size. In addition to uniquely specifying feline saccades with different degrees of obliquity, this two parameter description allowed us to fit not only the monkey data but data obtained from rabbit and human as well. Therefore, specification of both duration and maximum velocity might be a basic organizing principle of the neural mechanisms generating saccades in many species.