The effects of midazolam on animal behavior often are evaluated under a chronically food-limited regimen, which is used to implement food-reinforced performance, but the corresponding pharmacokinetics are lacking. The present study investigated the pharmacokinetics of midazolam after i.v., s.c., i.p., and p.o. administration in food-limited rates. A two-compartment model best described the concentration-time profiles for the four routes of administration. The rate of midazolam absorption was rapid, and peak concentrations were attained in less than 7 min for the three extravascular routes. The mean volume of distribution of the central compartment and clearance were 0.77 l/kg and 2.03 l/h per keg, respectively. Midazolam elimination half-lives for the four routes of administration ranged from 23.1 to 49.5 min, and metabolites could not be detected. The mean absolute bioavailability was route-dependent: 39.3% (s.c.) 19.2% (i.p.) and 4.6% (p.o.). The markedly low oral bioavailability found in food-limited rats contrasted to the value reported for free-feeding rates (45%). Although the i.p. route yielded the highest maximum concentration on occasion, serum midazolam concentration-time profiles were variable, but did correspond to respective sedative responses. DRL 45-s performance after s.c., i.p., and p.o. administration further supported the advisability of using the s.c. route of administration, as opposed to the i.p. route, for studying midazolam dose-response relations. The bioavailability values assessed from DRL performance also agree with the measured pharmacokinetic values.