Visually guided eye movements are driven by a mix of current signals (e.g. visual motion) and prior experience (predictive strategies). Previously, large ablations of the frontal eye field (FEF) impaired visually guided smooth pursuit. This study examined if the pursuit decrement could be laid to a selective loss in predictive signals. Normal monkeys demonstrated some of the same predictive eye movements documented in humans. They pursued periodic visual targets with near-zero phase lag. When such targets suddenly disappeared, the monkeys continued smooth pursuit without visual guidance for several reaction times. These epochs of 'blind pursuit' achieved a peak velocity proportional to the prior target frequency just experienced by the monkey. With predictable step-ramp targets smooth eye movements sometimes preceded target motion (anticipatory pursuit). Reaction time to begin pursuit was influenced by the target velocity of prior trials. Small unilateral ablations of 'low threshold' FEF showed smooth pursuit if the fundus of the arcuate sulcus was thoroughly removed. On the step-ramp targets the slowing was evident in both the initial 100 ms and subsequent portions of pursuit. During sine pursuit blind epochs were more slowed by surgery than were visually guided epochs of pursuit. Ipsilateral anticipatory pursuit was abolished in one subject, but not in the other subject with pursuit deficits. Otherwise, pursuit after surgery continued to display the influence of predictive strategies. The average phase lag of periodic pursuit remained much less than the pursuit system's reaction time. Blind epochs persisted after the periodic target disappeared, albeit with a lower peak velocity. Reaction time to begin pursuit of step-ramp targets remained a function of the monkey's experience on prior trials. It is argued that the FEF pursuit deficits do not reflect the loss of visual motion signals or the loss of some 'cognitive' signal such as a prediction about the target's motion, but rather can be explained as a motor deficit.