Purpose: The purpose of the current study was to determine the effects of vowel identity and temporal onset asynchrony on identification of vowels overlapped in time. Method: Fourteen listeners with normal hearing, with a mean age of 24 years, participated. The listeners were asked to identify both of a pair of 200-ms vowels (referred to as "double vowels") presented either simultaneously or with a temporal asynchrony ranging from 25 ms to 150 ms in 25-ms steps. The stimuli were synthetic steady-state vowels /i ae [script] u [reversed epsilon]/ arranged in seven combinations: /u i/, /ae [script]/, /[reversed epsilon] [script]/, /[reversed epsilon] ae/, /ae i/, /[reversed epsilon] i/, and /[reversed epsilon] u/. Results: Listeners' responses revealed that one vowel of a pair was identified correctly more often than the other vowel (known as "vowel dominance"). Vowel dominance effects were seen for 6 of the 7 vowel pairs, and there was improvement of vowel identification with increasing temporal separation between vowels for 5 of the 7 pairs. Vowel pairs with the vowel /[reversed epsilon]/ consistently yielded improved identification with increases in temporal asynchrony. Discussion: Peripheral masking cannot explain the patterns of results of this study. A more parsimonious explanation may be perceptual anchoring.