The "Building Blocks PreK Math Curriculum" (Clements & Sarama, 2007) was designed to facilitate children's engagement in math and talk about math. Much research investigates the effect of curriculum on classrooms or teacher practices. This study used a mediational model to look at a curriculum's effect on children's achievement gain, operating through specific child behaviors in the classroom. Specifically, this study looked at how a math curriculum affected children's focus in math alone or in all learning activities (math, literacy, science, social-studies, and other), talking during math-related activities or in all learning activities, and engagement during math or during all learning activities. Additionally, this study examined how those child behaviors predicted children's math achievement gain. It is hypothesized in the existing literature that much of the variability in student achievement across prekindergarten programs can be explained by the amount of time children are engaged in learning through talking, listening, or sustained focus on academic content. Behaviors like a child's focus on instruction (Barr & Dreeben, 1983; NCES, 2002), verbal behaviors (Winsler & Naglieri, 2003; Dickinson & Tabors, 2001; Snow, Burns, & Griffin, 1998), and educational engagement (Brophy & Good, 1986; Howse, Lange, Farran, & Boyles, 2002) are all considered critical elements of learning. This study was based on the hypothesis that a curriculum which encourages teachers to focus on such critical elements in the classroom can lead to changes in child achievement if changes in children's behaviors are also affected. This study was conducted as part of the SUNY Buffalo/Vanderbilt scale-up of the "Building Blocks Prekindergarten Math Curriculum" in Nashville, Tennessee. Fifty-seven classrooms from twenty sites, 16 Public Schools and 4 Head Start centers, participated in one of two study conditions. Thirty-one classrooms participated in the new math curriculum while twenty-six classrooms conducted business as usual. Across both conditions, children were observed in their classrooms and during mealtimes on three typical days--once in fall, once in winter, and once in spring,--but not when they were outdoors. The "Building Blocks" curriculum, designed to facilitate children's engagement in math and talk about math, was predictive of children's participation in math-focused activities, which predicted higher gains on standardized math measures. Without the resulting change in children's behaviors in the classroom, however, changes in achievement would not have been possible through curriculum implementation alone. Although getting children to talk more about math was a goal of the curriculum, children's talk with a math focus did not predict their gain. It is possible that teachers in the treatment condition actually spent more time in learning related activities, but spent that time instructing rather than probing children's thinking or asking higher-order inferential questions. The results of this study demonstrate the difficult task of helping teachers promote an environment where children can investigate mathematics through talk and play, rather than through direct instruction. The results suggest that classrooms organized to keep children focused in learning activities will better equip children with math-readiness skills, and thereby improve early math education in large-scale public programs for children from low-income households. (Contains 4 tables.)