Twenty-one students in their second quarter of Romanian language study at The Ohio State University completed a language background survey, took a perception test and made a tape recording of their pronunciations of a list of Romanian words. The tape used in the perception test consisted of a list of fifteen minimal pairs with initial voiceless or voiced plosives representing four places of articulation: labial, dental, palatal, and velar. The results show that the students confuse identification of the labial plosives; identify the voiced dental plosive correctly, but confuse the voiceless dental plosive; and identify both the voiced and voiceless palatal and velar plosives correctly. Acoustic analyses of the spoken stimuli show that the average duration of the release of the voiceless plosives varies systematically depending on the place of articulation: labials--25 ms., dentals--15 ms., palatals--90 ms., and velars--60 ms. The students apparently used the longer duration of the palatal and velar voiceless plosives as a cue in identifying them correctly. The shorter release of the labials and dentals resulted in perceptual confusion. No students consistently produced initial, unaspirated plosives in their tape recordings of the Romanian words. (Author)