Two experiments explored rates for introducing grapheme-phoneme correspondences (GPCs) and the types of correspondences taught for optimal alphabet and early literacy skills learning. In both studies, children entered with minimal alphabet knowledge and were randomly assigned within classrooms to one of two treatments delivered individually over 5 weeks. In Study 1, children grades K-1 were assigned to instruction in a set of either 10 (Slow rate, n = 33) or 15 (Fast rate, n = 32) single- and two-letter GPCs. Study 1 findings indicated that children who learned five added GPCs did not reduce learning of the common set of 10 learned GPCs for any measure (including letter names, sounds, letter sound writing, word reading, and spelling), and learning favored Fast items over Slow for letter sounds, letter sound writing, and word reading (median d = 0.30). In Study 2, kindergarteners were assigned to instruction in either single letters only (Single, n = 30) or mixed-size GPCs (Mixed, n = 31). Instruction included application of GPCs to decoding and spelling. Results showed that kindergarteners in the Mixed condition made significantly greater gains learning the four two-letter GPCs across measures (median d = 0.86), and no significant differences between groups on measures of the 11 one-letter GPCs common to both conditions. Findings add precision to understanding how rate and order of introducing GPCs influence children’s initial alphabet learning. Further study of empirically validated methods of alphabet instruction may benefit in particular those children most at risk for acquiring this foundational knowledge.