Still and Morris (2008) discovered that nonword anagram primes interfere with word target processing when the letters in the prime appear in reverse order of the letters in the target (e.g., yruf - FURY). This finding was unexpected as facilitation is typically found when a word target is preceded by an orthographically similar nonword prime (e.g., Grainger & Jacobs, 1999). The present study was designed to replicate and extend Still and Morris' finding of anagram interference. Results across three experiments indicate that anagram interference is modulated by target word frequency, stimulus length, prime exposure duration, and whether or not the anagram prime and target share letters in the same relative positions (e.g., enorht vs. oetnrh vs. htoren for the target THRONE). In addition to replicating the finding of anagram interference, these results show that the anagram interference is robust and is not limited to a specific set of stimuli. Current models of word recognition are unable to account for the finding of anagram interference.