The current study examined the relation between the amount of literacy instruction and student reading achievement (word reading and reading comprehension), using data from 927 students and their 189 teachers from Grades 1 to 3 in the US. Literacy instruction was observed once in the Fall and once in the Spring each year from two projects. We categorized literacy instruction into two types by content: (1) code-based instruction that focuses on phonics and decoding skills; (2) meaning-based instruction that focuses on vocabulary, oral language, and comprehension skills. We used multilevel multivariate models to separate classroom-level effects from student-level effects and accounted for cumulative effects across grades. Approximately 16-32% of the total variation was due to variation across classrooms. Overall, we detected a positive longitudinal relation of meaning-based instruction to students' reading comprehension across grades and a negative relation of code-based instruction in Grade 2 to students' word reading achievement. The inconsistent relations between the amount of instruction and students' reading achievement indicate the complexity of instructional effectiveness.