Studies have repeatedly shown (Harlow, 1959) that the performance of infrahuman primates given single-object training trials with a nonrewarded object, the Negative (Neg) Stimulus, will, in subsequent two-choice discrimination trials, exceed that of animals given comparable training trials with a rewarded object, the Positive (Pos) Stimulus. This phenomenon has become known as the Moss-Harlow effect (Moss & Harlow, 1947), and the phenomenon lends support to a uniprocess, inhibition theory of discrimination learning. Employing the Moss-Harlow procedure (Moss & Harlow, 1947), observed the MossHarlow effect in the discrimination performance of younger, but not older children; i.e., the younger children's performance was better following training with the Neg stimulus while the older children's performance was better following training with the Pos stimulus.The Moss-Harlow effect has been observed in squirrel monkeys even when the single-object training trials were interposed between the acquisition and reversal phases of discrimination reversal (Cross & Brown, 1965;Cross, Fickling, Carpenter, & Brown, 1964). However, when Vaughter & Cross (1965) employed the prereversal discrimination procedure with five-year-old children, the Moss-Harlow effect was not observed. These children showed the greatest reversal learning following either training with both the Pos and Neg stimuli or with the Pos stimulus alone. Training with the Neg stimulus alone was no more beneficial than a control situation involving no single-object, prereversal training trials.The present study was designed to test for the occurrence of the Moss-Harlow effect in the performance of younger children when the prereversal discrimination procedure was used to test the effect. Both older and younger children were tested in order that an immediate comparison could be made; however, it was hypothesized that the older children's performance would replicate the findings of Vaughter & Cross (1965). In addition, the study was designed to investigate the effects of the amount of single-object prereversal training (PRT) and the amount of overlearning in the discrimination training upon discrimination reversal performance. The nature of the relationships between these variables might be of significance to any theory proposing that discrimination reversal does not involve ex tinction of one habit and acquisition to a new one (Lovejoy. 1966).Method. Thirty normal children, IS below five years of age with a mean age of three years and four months, and IS above five years of age with a mean age of five years and four months from the Lubbock Day School nurseries were employed as Ss.A modified WGTA for children was utilized; the apparatus has been described in detail elsewhere (Olson, Cross, & Vaughter, 1966). The essential features of the WGT A are a divided tabletop with a sliding three-well test tray and a one-way mirror assuring E a clear view of S. The stimuli consisted of 48 multidimensional or "junk" objects mounted upon uniform bases and randomly paired to co...