This thesis takes its departure from the on-going debate about teachers´(collective) ‘continuing professional development’ (CPD). Teachers’ CPD through an imposed nine-step model of peer group mentoring (PGM) is focused on. The study draws on data from a two and a half yearlong interactive project that took place in a teacher team in a Swedish school. The general aim of the thesis is to study a practice of professional development in a teacher team involving peer group mentoring and to find out how and what kind of teachers’ expertise that is constructed. Furthermore, the aim is to examine how the PGM-practice was constrained and enabled and what kind of CPD was made possible. The theoretical and methodological framework is mainly based on practice theory. Practices and practitioners are seen as mutually interrelated. Practice architectures (Kemmis & Grootenboer, 2008) are used to uncover the relations between the PGM-practice and its historical, material-economic, social-political and cultural-discursive conditions. Furthermore, Foucault’s notion of power was adopted as an analytical tool to examine how power came into play during the mentoring sessions and how the teachers’ discursively constructed a ‘good teacher’ and teachers’ expertise. The methodological approach is action research. A main finding of the thesis is that professional and personnel development may be imposed through peer group mentoring. Furthermore, democratic processes increased during the PGMmeetings and seemed to have an impact on classroom practice and the practice of parent-teacher meetings. The results show how the PGM–practice and its outcomes are deeply interconnected to global and local historical, material-economic, social-political and cultural-discursive arrangements which constrained and enabled it. When economic cut downs (i.e. materialeconomic arrangements) began to take effect in the local school, along with a neo-liberal discourse (i.e. cultural-discursive arrangements), democratic processes were challenged and threatened. The focus in the PGM discussions shifted from the teachers’ perceived need for pedagogical knowledge development to talk about students as costs. The constrained nine-step model disciplined some individuals more than others. The teachers disciplined each other through e.g. confessions, corrections and differentiations. Inconsistent discourses about good teaching and teachers’ know-how were constructed and the teachers positioned themselves and each other as ‘good’ or ‘bad’ teachers. The interactive research approach partly enabled the PGM-practice but at the same time effected the teachers’ positioning of each other. The interactive research approach disciplined both the teachers and the researcher. Anyhow, power relations became fluent and mutual among the participants. A collegial approach and the ability to carry out reflexive cooperation were both fostered by the model and articulated in the PGM-practice as important teacher skills.