The activities of artemisinin (QHS) and a number of its semi-synthetic analogues, as well as Fenozan B07 (B07), a synthetic 1,2,4-trioxane, and arteflene (ATF), a synthetic surrogate of yingzhaosu, were compared in mice infected with drug-sensitive Plasmodium berghei or chloroquine-resistant P. yoelii ssp. NS. The studies were stimulated by the observation that B07, in certain aqueous preparations, appears to be equipotent by the subcutaneous (sc) or oral (po) routes in the rodent model but not in a simian model. In the rodent model, B07 was found to undergo rapid alteration (with a half-life of <24h) in an aqueous stock solution prepared using dimethyl sulphoxide (DMSO) to pre-dissolve the drug. Therefore, for all later experiments with aqueous preparations, the test material was newly formulated each day. In a carboxymethylcellulose formulation used as a 'standard suspending vehicle' (SSV), B07 and dihydroartemisinin (DIHYD) were found to be, respectively, one sixth and one 10th as active po as when the drugs were pre-dissolved in DMSO and then diluted with water. ATF in DMSO given po was less than one 20th as active as when used sc in the rodent model, and this drug in SSV was almost inactive po. The relatively low oral activity of these three compounds (especially DIHYD and ATF) may be attributable to extensive first-pass metabolism in the mouse. Oral beta-artemether (AM) and beta-arteether (AE) were highly active when used in SSV. ATF has been found to have low activity in simian models and clinical trials because of its poor absolute bio-availability. In in-vivo studies of the blood schizontocidal action of anti-malarials, in rodent malaria models, the data collected on the structure-activity relationships (SAR) of the drugs must be viewed critically when selecting specific compounds from a chemical series for further development. A study of the influence of drug formulation on the activity of other, novel antimalarials is crucial to the evaluation of the drugs, and merits high priority.