The delay in the onset of fatigue of the limb skeletal muscles following ingestion of caffeine is reported to be mediated via a direct effect of caffeine on skeletal muscle (Mohr et al. 1998). The diaphragm is a skeletal muscle and weakness and fatigue of the diaphragm is a cause of respiratory failure. The objective of the present study was to determine if caffeine directly affects the contractile properties and delays the onset of fatigue in the diaphragm. Data are presented as mean ± 1 SD. Analysis of variance and Student’s t test were used to test for significance (P<0.05). Adult male Wistar rats were anaesthetised (pentobarbitone sodium 60 mg kg-1 i.p.) and the diaphragm was rapidly removed. Rats were then humanely killed. The diaphragm was divided along its central tendon and a small costal strip was dissected from each hemi-diaphragm. Each strip was suspended in a separate organ bath containing oxygenated (95% O2-5%CO2) Krebs solution containing 10 μM d-tubocurarine and maintained at 30oC and pH 7.4. Contractile properties were determined at optimal length using field stimulation (supramaximal voltage,1ms duration) with platinum plate electrodes. Following an equilibration period of 30 min, a single twitch was elicited in each muscle strip. Caffeine was added to one bath to give a final concentration of 1mM. After 2 min a single twitch was again elicited in each strip. The percentage changes in twitch tension (Tw), contraction time (CT) and half-relaxation time (1/2 RT) in control strips (n=7) after 2 min were +2.1±3.9, -5.0±12.6 and +5.9±8.4 compared with the corresponding changes in caffeine-exposed strips (n=7) of +18.5±5.7 (p< 0.001), +2.6±10.2 and +22.8±10.7 (p< 0.01) respectively. The force frequency relationship was determined using stimulation trains of 300 ms at frequencies of 10-100 Hz. Force, expressed as a percentage of maximum force, was significantly greater in the caffeine-exposed strips compared to control strips at frequencies from 10 Hz to 60 Hz. Fatigue was induced using a fatigue protocol of 30 Hz trains every 2 sec for 5 min. Tension after 5 min., expressed as a percentage of initial tension, was 48.0±7.6 in the control strips and 46.0±6.7 in the caffeine-exposed strips. In conclusion, caffeine increased Tw and 1/2 RT in the rested isolated diaphragm and increased the relative force generated at frequencies from 10-60 Hz. Caffeine did not prolong the time to fatigue in the diaphragm.