Moment-to-moment variations in the engagement of young children's cognitive capacity by televised material were examined using a secondary task paradigm. Thirty-five 5-year-olds watched a 35-min Sesame Street program containing three types of segments: normal segments, segments with scenes reordered, and segments with incomprehensible language audio tracks. While watching the program, children were to respond quickly to auditory probes distributed across all types of segments and positions within segments. Probe response times and visual attention were recorded. Major findings were: As indicated by longer probe response times, capacity was more effectively engaged if language was comprehensible, provided children were looking at television when probes were presented. If not looking, response times were equally fast across segment types. For normal segments only, there were increases in the engagement of cognitive capacity if a look at the television or the program content had been continuous for some time. The findings provide evidence for, but important refinements of, the hypothesis that young children's ongoing comprehension is a major determinant of their attention to television.