The present research determined whether low-verbal Ss, as compared with high-verbal Ss have (a) a greater production deficiency, and (b) a greater mediational deficiency. The design included 3 grade levels: K, II and V. At each grade level approximately 40 Ss were identified as low-verbal and 40 Ss as high-verbal, with these groups differing on WISC vocabulary, PPVT IQ and in mean number of communication units. Two paired associate learning tasks were presented for testing both the production deficiency and the mediational deficiency hypothesis. The production deficiency hypothesis was analyzed by means of a 4-way ANOV with independent variables: grade, verbal groups, sex and PA tasks. For both silent and aloud rehearsals, the F ratios for grade and verbal group were significant. Between kindergarden and grade II there is a marked increase in the mean number of silent rehearsals and an equally marked decrease in the number of aloud rehearsals. The low-verbal group have significantly fewer silent rehearsals but significantly more aloud rehearsals than do the highverbal group. Evidence from the Ss own report of rehearsal supports conclusions drawn from the analyses of silent and aloud rehearsals. Data relevant to the mediation deficiency hypothesis were analyzed by means of a 4-way ANOV, with independent variables: grade, verbal group, sex and trials. F ratios for grade and verbal group were significant across two PA learning tasks. The most marked increase in mediational behavior occurs between kindergarden and grade II, with the lowverbal group performing at a significantly lower level than the highverbal group. Downloaded by [New York University] at 11:16 19 February 2015 182 The overall objective of the reported research was to analyze differences in the mediational performance of groups of children who have been exposed to contrasting linguistic environments. It examined ontogenetic changes in both the ability to produce verbalization and the ability to use such verbalization to mediate performance.Despite the assumption that increasing linguistic development is associated with the increase in mediating responses accompanying increasing development level, very few studies have directly compared language skills of mediators and non-mediators. The recent explorative work of Kendler and Kendler (1970) showed little, if any, relation between Peabody Picture Vocabulary score and probability of making a mediated response. Results from a study by Knowles and Boersma (1971) also suggest that no relationship exists. It is suggested that more precise hypotheses than those investigated by either of these two explorative studies must be formulated before the relationship between linguistic development and mediating behavior can begin to be understood. Jensen (1968) considers it reasonable to hypothesize that certain children do poorly on paired-associate and other tasks requiring mediating responses because they have a high threshold for the elicitation of spontaneous verbalization.In a recent formulation of his socioli...