Changing classroom practice is a difficult and complicated business. It involves challenging both what teachers know and do in their classrooms. It involves a process of re-configuring both teacher knowledge and practice. Many attempts at teacher development have had less than spectacular results in this regard. This study explored one such attempt of a teacher development programme that was aimed at challenging and reshaping teacher’s science content knowledge (CK) and pedagogical content knowledge (PCK), with the aim of influencing classroom practice. Using Shulman’s argument on the importance of uncovering teachers’ knowledge, this study investigated teacher clusters or networks and the opportunities they provide for science teachers to collaborate and share their knowledge and classroom practices. The collaboration was intended to strengthen science content knowledge (CK) and pedagogical content knowledge (PCK) in order to improve classroom practices. The context of the study was a professional development intervention on science and mathematics in Mpumalanga; the Mpumalanga Secondary Schools Initiative (MSSI) funded by the Japanese International Co-operation Agency (JICA), where the teachers met regularly to share knowledge and experiences in smaller groups commonly referred to as Teacher Clusters. Qualitative research methods in the form of case studies were used to investigate and understand how these clusters helped science teachers in particular to challenge and change their CK and PCK by working collaboratively. The study presents two case studies of science teacher clusters, and examines the interactions and mechanisms by which the clusters constituted resources for teacher learning and improvement in teaching practice. The major findings of this research are that: -- Teacher clusters indeed provided better resources for changing the classroom practice of science teachers by allowing them to focus specifically on their CK and PCK and the interactions between the two forms of teacher knowledge -- Most importantly, teacher clusters also functioned in such a way that they allowed teachers a substantially more and better set of collective resources from which to begin and sustain changes in classroom practice. In making these major findings, this research established the critical role of the structure and leadership of the cluster in helping to create conditions for successful clustering and teacher change. The study concluded by observing that clustering or networking does provide teachers with enhanced opportunities for professional growth and classroom change. Variations in forms of clustering and teacher leadership issues within the clusters still require further investigation and research than this study was able to provide.