A modification of Posner's letter-matching paradigm proposed by Beller (1971) was employed to study the development of abstract visual and name codes for letters. Second-, fourth-, and sixth-grade students (ages 8, 10, and 12 years, respectively) were presented in advance one of a pair of to-be-matched letters. Case agreement (upper- and lowercase) between the prime letter and the letter pair was manipulated to differentiate facilitation based on concrete features of the stimulus letter from that based on the letter's abstract visual code. For physical matches second graders showed greater facilitation from same-case primes than from different-case primes, whereas sixth graders benefited equally from both types of prime. For same-name and different conditions, no qualitative developmental differences were revealed as all subjects benefited from letter primes irrespective of case agreement. The acquisition and utilization of abstract visual codes for letters are a later acquisition than name codes, which were clearly operative at the lowest grade level in the present experiment. © 1978.