The pharmacokinetics and bioavailability of dihydroartemisinin (DQHS), artemether (AM), arteether (AE), artesunic acid (AS) and artelinic acid (AL) have been investigated in rats after single intravenous, intramuscular and intragastric doses of 10 mg kg(-1). Plasma was separated from blood samples collected at different times after dosing and analysed for parent drug. Plasma samples from rats dosed with AM, AE, AS and AL were also analysed for DQHS which is known to be an active metabolite of these compounds. Plasma levels of all parent compounds decreased biexponentially and were a reasonable fit to a two-compartment open model. The resulting pharmacokinetic parameter estimates were substantially different not only between drugs but also between routes of administration for the same drug. After intravenous injection the highest plasma level was obtained with AL, followed by DQHS, AM, AE and AS. This resulted in the lowest steady-state volume of distribution (0.39 L) for AL, increasing thereafter for DQHS (0.50 L), AM (0.67 L), AE (0.72 L) and AS (0.87 L). Clearance of AL (21-41 mL min(-1) kg(-1)) was slower than that of the other drugs for all three routes of administration (DQHS, 55-64 mL min(-1) kg(-1); AM, 91-92 mL min(-1) kg(-1); AS, 191-240 mL min(-1) kg(-1); AE, 200-323 mL min(-1) kg(-1)). In addition the terminal half-life after intravenous dosing was longest for AL (1.35 h), followed by DQHS (0.95 h), AM (0.53 h), AE (0.45 h) and AS (0.35 h). Bioavailability after intramuscular injection was highest for AS (105%), followed by AL (95%) and DQHS (85%). The low bioavailability of AM (54%) and AE (34%) is probably the result of slow, prolonged absorption of the sesame-oil formulation from the injection site. After oral administration, low bioavailability (19-35%) was observed for all five drugs. In-vivo AM, AE, AS and AL were converted to DQHS to different extents; the ranking order of percentage of total dose converted to DQHS was AS (25.3-72.7), then AE (3.4-15.9), AM (3.7-12.4) and AL (1.0-4.3). The same ranking order was obtained for all formulations and routes of administration. The drug with the highest percentage conversion to DQHS was artesunic acid. Because DQHS has significant antimalarial activity, relatively low DQHS production could still contribute significantly to the antimalarial efficacy of these drugs. This is the first time the pharmacokinetics, bioavailability and conversion to DQHS of these drugs have been directly compared after different routes of administration. The results show that of all the artemisinin drugs studied the plasma level was highest for artelinic acid; this reflects its lowest extent of conversion to DQHS and its slowest rate of elimination.