This paper reviews the findings from a controlled experimental study of the effectiveness of the Reading Recovery Program undertaken in ten primary schools in Sydney, during the second year of its implementation. Children, selected by their teachers as low progress readers after one year at school, were randomly allocated to either Reading Recovery or to a control condition in which they received only the resource support typically provided to at-risk children. Low progress children from five schools of similar socio-economic level, but without a Reading Recovery Program, also took part in the study. All children were pre-tested on a number of norm and criterion referenced reading related literacy measures, post-tested on the same measures fifteen weeks after the Program commenced, and once again after a further fifteen weeks to assess maintenance effects. Results indicated that Reading Recovery students clearly outperformed students in the control group condition on most measures but, at this early stage of Reading Recovery implementation, there appeared to be no spill-over effects to other low progress students and their teachers in Reading Recovery schools. Furthermore, the practice of matching children to appropriate book level in the regular classroom, which is necessary to ensure maintenance of the Reading Recovery intervention, was not being adopted in most schools. Finally, because of the extremely low book level scores found among all Year 1 children in the ten Reading Recovery schools at the commencement of the study, serious implications arise about the selection of students for Reading Recovery in metropolitan New South Wales, their most appropriate point of entry into the Program, and the maintenance of their skills at discontinuation when they return to regular classes where average book levels continue to be low.