The overall purpose of this research is to explore how bilingual children in India learn to read. Therefore, this study aims to identify factors associated with attainment and progress in reading in the early years of school. Second, it investigates reading and writing experiences in the classroom, from the perspective of high, average and low attaining readers. Five-year-old children's reading was assessed at the end of nursery, using tasks selected on the basis of curriculum validity. The 161 children were selected from seven schools in Calcutta in which English is the medium of instruction. A questionnaire, surveying home background and literacy experiences, was completed by the parents at this stage. Eight months later, when the children were in the kindergarten, the teachers' ratings of the children's achievement, adjustment and overall development were obtained. Seventeen months after the initial assessment, the children's reading was reassessed, using graded passages, isolated words and written spellings. Multiple regression was used to analyse the data. Word reading, rate, accuracy, comprehension and written spellings were included separately, as dependent variables. The independent variables included the scores in the initial assessment; the age and sex of the child, the survey data, and teacher ratings. Word recognition consistently emerged as a variable strongly associated with reading when all factors were taken into account. Teacher ratings and a few home background factors were also associated with progress in reading. The data indicate that there were also sizeable differences between the impact of the schools on reading acquisition. In order to investigate low-achieving children's responses to classroom based literacy tasks in comparison to others, a small sample of the children was divided into categories of low, average and high achievers in reading on the basis of their scores in the initial assessment. These children were observed in the classroom, using both systematic and interpretive observations, in the period between the initial and follow-up assessment. Non-parametric statistical analysis revealed significant differences in the systematic observation data in the case of two categories of behaviour - 'child-child interaction', and looking 'lost'. Qualitative methods of analysing interpretive observation data however, show that there were substantial differences between the responses of the low-achieving children and the others to reading and writing tasks. Both phases of observation indicated that teachers' focus on individuals in the classroom was limited and that individual needs were not met. No differences were noted in teachers' behaviour or responses to low, average and high achieving readers. The findings of this research will be discussed with reference to Indian classrooms, so that teaching and learning contexts can be upgraded with a greater understanding of factors associated with progress.