We examined models of individual change and correlates of change in the growth of reading skills in a sample of 40 children from kindergarten through third grade. A broad range of correlates was examined and included family literacy, oral language, emergent reading, intelligence, spelling, and demographic variables. Individual growth curve analysis was used to model change in Letter Word Identification (LWID), Word Attack (WA), and Passage Comprehension (PC) subtests of the Woodcock-Johnson Psychoeducational Battery-Revised. Third grade LWID was predicted uniquely by family literacy, phonological awareness, and emergent reading skills. Growth in LWID was predicted uniquely by emergent reading skills. Phonological awareness, spelling, and emergent reading were unique predictors of third grade WA, whereas family literacy and emergent reading skills uniquely predicted third grade PC. The general oral language factor defined by semantic and syntactic variables did not contribute significant unique variance in any of the models. Thus, the pattern of results extends the model of emergent-to-conventional literacy proposed by Whitehurst and Lonigan (1998) to third grade and suggests that early contextual understandings necessary for competent reading (family literacy and emergent reading) become more influential as reading skills develop.