Nonwords created by transposing two phonemes of auditory words (e.g., /buʒãle/) are more effective primes for the corresponding base word target (/bulãʒe/) than nonword primes created by substituting two phonemes (e.g., /buvãʀe/). In one in-lab experiment and one online experiment using the short-term phonological priming paradigm, here we examine the role of vowels and consonants in driving transposed-phoneme priming effects. Results showed that facilitatory transposed-phoneme priming occurs when the transposed phonemes are consonants (/buʒãle/-/bulãʒe/; /lubãʒe/ /bulãʒe/), but not when they are vowels (/bãluʒe/-/bulãʒe/; /buleʒã/-/bulãʒe/). These results add to existing findings showing differences in the processing of vowels and consonants during spoken and visual word recognition. We suggest that differences in the speed of processing of consonants and vowels combined with differences in the amount of information provided by consonants and vowels relative to the identity of the word being recognized provide a complete account of the present findings.