The concept of flow was examined in relation to the play behavior of preschool children. Thirty-five preschoolers were encouraged to play a beanbag toss game on three separate occasions called Phases I-III. In Phase I, children were allowed to match their skills to the challenges of the task in a choice condition. In Phase II, one group of 12 children again played the game under the choice condition. Two other groups of 11 and 12 children were required to play the game under conditions in which the challenges of the task were assumed to be either greater than or less than their levels of skill. In Phase III, all children again played the game under the choice condition. Five measures of the characteristics of flow, including number of attempts, percentage of successes, ratio of evaluation-seeking behaviors to number of attempts, ratio of off-task behaviors to number of attempts, and ratio of time spent off task to time spent on task were obtained for all phases. A repeated measures MANOVA with three groups and three phases was used to examine the data. A significantly greater number of evaluation-seeking behaviors per attempt were found at Phase I than at Phases II and III, regardless of condition. A significant phase x group interaction for the variable percentage of successes was also found. Post hoc analyses indicated that children in the assigned easy group had a significantly higher success rate at Phase II than at Phases I and III, while the assigned difficult group had a significantly lower rate. The percentages of successes between the three groups were significantly different at Phase II only. The single exposure to an assigned level of challenge may have been insufficient to produce the predicted effects on the characteristics of the flow experience. Subjective comments of the children obtained in Phase II indicated there were some differences in the levels of enjoyment of the task between the three groups. These subjective perceptions offer a possible explanation for the lack of differences on the measures of flow. The subjective comments indicate that throwing the beanbag at the target was enjoyable to the children regardless of their ability to hit the target; therefore the level of challenge may not have been critical. Several suggestions for future testing of the flow model are offered. / Ph. D.