South Africa participated in the Progress in International Reading Study (PIRLS) for the first time in the PIRLS 2006. PIRLS was the most complex research conducted on reading literacy in South Africa to date. Internationally most countries participating in PIRLS assessed pupils in one grade and in one language. In South Africa, pupils were assessed in two grades and the children were assessed in all of the official 11 languages. This took place in more than 400 schools with children in both grade 4 and grade 5 being assessed in the medium of instruction they had received in grades 1-3. However, this does not imply that all children had had the opportunity to learn in their home language. Nonetheless, after four years of schooling, children internationally are expected to be able to read and now start the process of reading to learn. Initially, South African children start their learning at school in their home language and this practice continues until grade 3. However, in the majority of schools, the language of instruction changes, and in grade 4 more than 80% of South African pupils learn in a second language. The first aim of the research was to investigate the extent to which South African children's reading performance in the medium of instruction was affected by the fact that so many children learn in a second, third or "foreign" language. It is widely believed that the mixing of the medium of instruction is partly to blame for the continued underperformance in both primary and secondary school level of South African children, particularly in international studies (Howie, 2002). Secondly, the authors want to ascertain to what extent are South African pupils proficient readers in their home language (or alternatively the language which they have received reading instruction in for four years). The South African situation is complex due to the fact that most white, Indian and coloured children will continue to receive their schooling in the same language of instruction from grade 1-12, whilst most children speaking African languages at home will switch at grade 4 to either English or Afrikaans despite the current government language policy permitting learning to take place in their home language throughout school. Finally, the research sought to ascertain the level at which the children were proficient in English (as the majority of children would be studying in English) and how this affected their performance in their home language.The findings suggest that children from English and Afrikaans backgrounds and who speak those languages at home achieved higher scores on the PIRLS tests. Furthermore, African children attending suburban schools requiring them to speak English from grade 1 achieved higher scores than those African children attending township schools and learning through the medium of their home language. These and other findings are presented in this paper using the PIRLS 2006 national data comprising about 15 000 pupils per grade from more than 400 schools.