Infant eye tracking is becoming increasingly popular for its presumed precision relative to traditional looking time paradigms and potential to yield new insights into developmental processes. However, there is strong reason to suspect that the temporal and spatial resolution of popular eye tracking systems is not entirely accurate, potentially compromising any data from an infant eye tracking experiment. Moreover, "best practices" for infant eye tracking, such as knowing which software tool enhances experimental flexibility, remain to be determined. The present investigation was designed to evaluate the temporal and spatial accuracy of data from the Tobii T60XL eye tracker through the use of visual latency and spatial accuracy tasks involving adults and infants. Systematic delays and drifts were revealed in oculomotor response times, and the system's spatial accuracy was observed to deviate somewhat in excess of the manufacturer's estimates; the experimental flexibility of the system appears dependent on the chosen software. (Contains 7 figures.)