SummaryWe examined the effects of ventilatory muscle endurance trazmng on resting breath ing pattern in 12 C6-C7 traumatic quadriplegics at least 1 year post-injury. All subjects had complete motor loss below the lesion level. Subjects were randomly assigned to a training (N = 6), or a control group (N = 6). Baseline tests included measurement of resting ventilation and breathing pattern using mercury in rubber strain gauges for 20 minutes in a seated position; maximum inspiratory mouth pressure (MIP) at PRC, and sustainable inspiratory mouth pressure for 10 minutes (SIP); lung volumes, and arterial blood gases (ABG's). The training protocol con sisted of breathing through an inspiratory resistor equivalent to 85% SIP for 15 minutes twice daily, 5 days a week for 8 weeks. Both trainers and controls attended the lab every 2 weeks for reassessment of MIP and SIP and the inspiratory resist ance was increased in the training group as SIP increased. At the end of 8 weeks, baseline tests were repeated. All subjects had normal ABG's. There was a significant increase in mean MIP and SIP in both the control group (30% ± 19% and 31% ± 18% respectively), and in the training group (42% ± 24% and 78% ± 49% respectively). Although the absolute values for both MIP and SIP were greater in the training group than in the control group, the differences were not significant. The alterations in resting breathing pattern were also the same in both groups. Mean frequency decreased significantly in the control group (20' 2/min ute to 16'9/minute) and, while insignificant, the change in frequency in the training group was the same, 19'4/minute to 16·4/minute. Mean tidal volume (Vt) in creased 18· 2% of baseline Vt in the control group and 17-0% baseline in the trainers, resulting in no change in minute ventilation. As MIP and SIP increased similarly in both groups, the data from the control and trainers was pooled and timing changes re-evaluated pre-and post-study. A significant decrease in mean Ti/ Ttot was obser ved, while no change in Vt/Ti was found. We concluded that the testing procedure