This dissertation is an acoustic description of the phonemic inventory of Nez Perce [nez], a Penutian language of the United States. Acoustic work has been conducted on the consonants of Nez Perce, but no acoustic work has been conducted on the vowels or vowel harmony system. This work begins with an overview of the dissertation, language situation, and previous research. Following the introduction are chapters on ejectives and plain obstruents, plain and glottalized sonorants, vowels, vowel harmony, and the conclusion. Nez Perce has both plain and ejective stop series, a plain and ejective affricate series, and a plain fricative series. I examine these segments for acoustic correlates comparing them to previous research, other languages and current theory. The ejectives are described with f0, intensity, jitter, burst amplitude and VOT. I discuss fricatives in terms of spectra, duration, formant transitions, and moments. The timing and realization of glottalization on glottalized and plain sonorants is investigated. Segments are measured for duration and are visually and aurally inspected for variation of glottalization, realized using pitch, laryngealization, glottal stops, or a combination of these features. It is nearly always realized on the sonorant rather than before or after. Vowels are plotted and compared to previous phonological descriptions. The inventory is /i, æ, a, o, u/ rather than the canonical five-vowel system, leading to the description of the inventory as having a “gap” and not maximally contrastive. I suggest that if Nez Perce vowels are considered using a shifted axis then the vowels are maximally contrastive. The non-canonical vowel inventory leads to two seeming unrelated vowel harmony sets: dominant, /i, a, o/, and recessive, /i, æ, u/. The proposed shifted axis view becomes important for reanalyzing the vowel harmony to reconcile these unusual sets. Previous analyses have described Nez Perce vowel harmony as based on advanced tongue root (ATR). I investigated Nez Perce vowels for ATR acoustic correlates; however, the results provide evidence both supporting and not supporting an ATR analysis. I propose an alternate analysis for the vowel harmony based on the principle of maximal contrast, evidenced by the shifted axis model.