The study investigated the effects of the direct and exploration methods of teaching on the overhand throwing performance of kindergarten children. Variable throwing practice was provided in the exploration method by allowing children to throw 5 different types of balls at a variety of targets. Children given the direct method threw only one type of ball and received specific instruction and demonstrations. Throwing for distance and throwing accuracy were measured on pre- and posttests. A novel throwing task was administered on the posttest to measure skill transfer. The experimental groups received instruction three times per week for 4 wk. in throwing while a control group received no instruction. Significant sex differences in throws were found for distance and accuracy. Analysis of data from the novel task gave no significant effects. The two methods of teaching did not produce different levels of throwing skill.