The current study investigated the spectral distribution of vowel cues to Mandarin and English sentence intelligibility. Sentences were segmentally interrupted to preserve various amounts of vowel information. Interruption parameters ensured that similar durations of speech were presented across the two languages. The remaining vowel cues were then either high-pass or low-pass filtered. Results demonstrated significant contributions of information present during vowels to sentence intelligibility. Performance was much higher in most of the conditions for Mandarin speech. In addition, Mandarin vowels contained a higher distribution of cues in the low-frequencies, below 1000 Hz. Results suggest that Mandarin vowels carry additional cues in the low-frequencies, most likely related to lexical tone. This is further supported by the highly similar performance between Mandarin and English sentence intelligibility when spectral cues are confined to frequencies above 2000 Hz.