Two studies were conducted to evaluate short-latency vocal tract air pressure responses to sudden pressure bleeds during production of voiceless bilabial stop consonants. It was hypothesized that the occurrence of respiratory reflexes would be indicated by distinct patterns of responses as a function of bleed magnitude. In Study 1, 19 adults produced syllable trains of /pa/using a mouthpiece coupled to a computer-controlled perturbator. The device randomly created bleed apertures that ranged from 0 to 40 [mm.sup.2] during production of the 2nd or 4th syllable of an utterance. Although peak oral air pressure dropped in a linear manner across bleed apertures, it averaged 2 to 3 cm [H.sub.2]0 at the largest bleed. While slope of oral pressure a(so decreased in a linear trend, duration of the oral pressure pulse remained relatively constant. The patterns suggest that respiratory reflexes, if present, have little effect on oral air pressure levels. In Study 2, both oral and subglottal air pressure responses were monitored in 2 adults while bleed apertures of 20 and 40 [mm.sup.2] were randomly created. For 1 participant, peak oral air pressure dropped across bleed apertures, as in Study 1. Subglottal air pressure and slope, however, remained relatively stable. These patterns provide some support for the occurrence of respiratory reflexes to regulate subglottal air pressure. Overall, the studies indicate that the inherent physiologic processes of the respiratory system, which may involve reflexes, and passive aeromechanical resistance of the upper airway are capable of developing oral air pressure in the face of substantial pressure bleeds. Implications for understanding speech production and the characteristics of individuals with velopharyngeal dysfunction are discussed.KEY WORDS: stop consonants, oral air pressure, subglottal air pressure, respiratory reflexes, velopharyngeal dysfunction