Curriculum materials explicitly designed to foster teacher learning represent a prominent route toprofessional development (PD) for teachers. However, it is unclear whether PD can be deliveredsuccessfully in the form of teacher self-study of curriculum materials, or whether it has to be scaffoldedadditionally by an expert. This study investigated effects of expert scaffolding in science-related PD forelementary school teachers with regard to proximal teacher outcomes (i.e., teacher beliefs and motiva-tions), instructional quality, and student achievement. Moreover, mediation of PD effects throughproximal teacher outcomes and teachers’ instruction was examined. Seventy-three teachers and 1,039students participated in the study. Expert scaffolding was implemented in a 3-tiered way: A first group(18 teachers) received PD with extensive scaffolding, a second group (18 teachers) was provided withreduced expert scaffolding, and a third group (18 teachers) received no expert scaffolding and wasprovided with the curriculum materials only (self-study group). A baseline group (19 teachers) did notparticipate in science-related PD and completed questionnaires on teacher outcomes only. Scaffolded PDwas significantly superior to PD through self-study in terms of teacher beliefs and motivation, instruc-tional quality, and student achievement. Contrary to our hypothesis, PD effects on student learning weremediated only to a small extent by teacher beliefs. However, teachers’ instruction emerged as asubstantial mediator of PD effects on student achievement. The results highlight the advantages ofadditional expert scaffolding in PD based on curriculum materials for the preparation of elementaryschool teachers for teaching science.