BACKGROUND: The rapid selection of pyrethroid resistance throughout sub-Saharan Africa is a serious threat to malaria vector control. Chlorfenapyr is a pyrrole insecticide which shows no cross resistance to insecticide classes normally used for vector control and is effective on mosquito nets under experimental hut conditions. Unlike neurotoxic insecticides, chlorfenapyr owes its toxicity to disruption of metabolic pathways in mitochondria that enable cellular respiration. A series of experiments explored whether standard World Health Organization (WHO) guidelines for evaluation of long-lasting insecticidal nets, developed through testing of pyrethroid insecticides, are suitable for evaluation of non-neurotoxic insecticides. METHODS: The efficacy of WHO recommended cone, cylinder and tunnel tests was compared for pyrethroids and chlorfenapyr. To establish bioassay exposure times predictive of insecticide-treated net (ITN) efficacy in experimental hut trials, standard three-minute bioassays of pyrethroid and chlorfenapyr ITNs were compared with longer exposures. Mosquito behaviour and response to chlorfenapyr ITN in bioassays conducted at night were compared to day and across a range of temperatures representative of highland and lowland transmission. RESULTS: Standard three-minute bioassay of chlorfenapyr produced extremely low levels of mortality compared to pyrethroids. Thirty-minute day-time bioassay produced mortality closer to hut efficacy of chlorfenapyr ITN but still fell short of the WHO threshold. Overnight tunnel test with chlorfenapyr produced 100% mortality and exceeded the WHO threshold of 80%. The endogenous circadian activity rhythm of anophelines results in inactivity by day and raised metabolism and flight activity by night. A model which explains improved toxicity of chlorfenapyr ITN when tested at night, and during the day at higher ambient temperature, is that activation of chlorfenapyr and disruption of respiratory pathways is enhanced when the insect is more metabolically and behaviourally active. CONCLUSIONS: Testing according to current WHO guidelines is not suitable for certain types of non-neurotoxic insecticide which, although highly effective in field trials, would be overlooked at the screening stage of evaluation through bioassay. Testing methods must be tailored to the characteristics and mode of action of each insecticide class. The WHO tunnel test on night-active anophelines is the most reliable bioassay for identifying the toxicity of novel insecticides.