Low-income, inner-city children were involved in a two-year intervention delivered in the regular classroom by regular classroom teachers to develop phonological awareness and word recognition skills. For the treatment children, an 11-week phoneme awareness program in kindergarten was followed by a first grade reading program (extended to grade 2 for some children) that emphasized explicit, systematic instruction in the alphabetic code. Control children participated in the school district''s regular basal reading program. Both groups participated in a phonetically-based spelling program mandated by the district. At the end of grade 1, treatment children (n = 66) significantly outperformed control children (n = 62) on measures of phonological awareness, letter name and letter sound knowledge, and three measures of word recognition, and reached marginal significance (0.056) on a fourth. They also significantly outperformed the control children on two measures of spelling. One year later, at the end of grade 2, the treatment children (n = 58) significantly outperformed the control children (n = 48) on all four measures of word recognition. For the groups as a whole, there were no differences on the one measure of spelling readministered at the end of grade 2. However, there were significant differences in spelling between the treatment (n = 16) and control children (n = 13) who remained in the bottom quartile of spellers at the end of grade 2 when partial credit was given for phonetically correct spelling, and significant differences in reading favoring these treatment children on all four measures of word recognition.