There is now considerable evidence (e.g., Perea & Lupker, 2003a, 2003b) that transposed-letter nonword primes (e.g., jugde for JUDGE) are more effective primes than replacement-letter nonword primes (e.g., jupte for JUDGE). Recently, Perea and Lupker (2004) demonstrated that, in Spanish, this transposed-letter prime advantage exists only when the transposed letters are consonants (C-C transpositions) and not when they are vowels (V-V transpositions). This vowel-consonant difference causes problems even for models that can successfully explain transposed-letter effects (e.g., SOLAR, Davis, 1999). In Experiment 1 in the present paper, we demonstrated a parallel result in a language with a different syllabic structure (English) in both a masked priming experiment and an unprimed lexical decision task in which the transposed letter strings (e.g., ADACEMY, ACEDAMY) were used as the nonwords. Results in Experiment 2 suggest that at least part of the reason for the vowel-consonant difference is because of the higher letter frequencies of the vowels. Possible alternative interpretations of the vowel-consonant difference are discussed.