This collective case study examines how four third-grade teachers’ beliefs and knowledge influenced their ways of supporting and limiting student thinking in their first year using a reform-based mathematics curriculum at an urban school. Of focus is the role teachers’ beliefs and knowledge play in supporting and limiting student thinking when student difficulties arise during instruction on multiplication and division. Situated in the growing body of research associated with current reforms in mathematics education, this study is also informed by general education research on urban schools, teacher beliefs, teacher knowledge, and teacher change. Data sources for case studies on individual teachers include classroom observations, pre-/post-observation interviews, beginning/end-of-year measures of teacher beliefs and knowledge, records of an on-going mathematics professional development project, and student achievement data. Each case study describes teacher’s beliefs and knowledge at the beginning and end of the year, presents a case story illuminating the teacher’s patterns of response to student difficulties and their relationship to the teacher’s beliefs and knowledge, and summarizes data from global measures of teaching. In addition to development of multiple case studies, a simultaneous cross-case analysis was undertaken to illuminate patterns across cases and increase the potential for generalizing beyond the particular cases. Findings from this study suggest that some aspects of reform-oriented mathematics instruction are more readily adopted than others. While beliefs and knowledge both appear to influence teacher response to student difficulties, certain aspects of instruction seem more greatly influenced by teacher beliefs while others appear more greatly influenced by teacher knowledge. In addition, evidence suggests that teachers’ differential classroom experiences during initial use of reform-based mathematics curriculum were related to the degree to which teachers’ evolving beliefs and knowledge moved closer to alignment with reform-based mathematics practices. Finally, the urban context of this study was found to influence teachers’ transitions to reform-based mathematics teaching practices in a variety of ways. Study findings have several implications for efforts to support teachers’ transitions to reform-based mathematics programs and practices within and outside of urban school settings. These are discussed along with directions for future research.