To test the hypothesis that during unsupported arm exercise (UAE) some of the inspiratory muscles of the rib cage partake in upper torso and arm positioning and thereby decrease their contribution to ventilation, we studied 11 subjects to measure pleural (Ppl) and gastric (Pga) pressures, heart rate, respiratory frequency, O2 uptake (VO2), and tidal volume (VT) during symptom-limited UAE. We used leg ergometry (LE) as a reference. Exercise duration was shorter for UAE vs. LE (207 +/- 67 vs. 514 +/- 224 s, P less than 0.05) even though the end-exercise VO2 was lower for UAE (9.3 +/- 1.1 vs. 30.8 +/- 3.2 ml.kg-1.min-1, P less than 0.05). Eight subjects had positive Ppl-Pga slopes and less negative end-inspiratory Ppl during UAE vs. LE (-11.8 +/- 6 vs. -19 +/- 7 cmH2O, P less than 0.05). This was not due to the lower VT's achieved during UAE, since at a similar VT, UAE resulted in a rightward and downward displacement of the Ppl-Pga slopes. Three of the subjects had irregular breathing rhythm and negative Ppl-Pga slopes as early as 1 min after initiation of UAE. They had shorter UAE duration and more dyspnea than the eight with positive Ppl-Pga slopes. In most subjects UAE decreases the ventilatory contribution of some of the inspiratory muscles of the rib cage as they have to partake in nonventilatory functions. This results in a shift of the dynamic work to the diaphragm and abdominal muscles of exhalation. In a few subjects UAE results in an irregular breathing pattern and very short exercise tolerance.(ABSTRACT TRUNCATED AT 250 WORDS)