Tolerance of bioresmethrin, permethrin, deltamethrin and the natural pyrethrins with and without synergists (piperonyl butoxide and propyl prop-2-ynyl phenylphosphonate), and the frequency of kdr homozygotes (flies surviving a discriminating dose of DDT+FDMC (1,1-bis(4-chlorophenyl)-2,2,2-trifluoroethanol)) was investigated in 42 strains of Musca domestica L. collected on British pig farms and in one Canadian strain. Although all field strains were at least 4–10 times more tolerant of synthetic pyrethroids at the LD50 than the standard susceptible Cooper strain, and 27 strains (26 British + 1 Canadian) had detectable levels of kdr, pyrethroids failed to control flies on only nine farms where 10% of flies or more were homozygous for kdr and/or the LD50s (in μg/female) in bioassays equalled or exceeded 0·1 for bioresmethrin, 0·12 for permethrin, 0·012 for deltamethrin or 1·1 for pyrethrins. These figures correspond to resistance factors (Cooper strain = 1) of 20, 15, 30 and 4, respectively. Regression analyses showed a close relationship between the LD50s to pyrethroids and the frequency of kdr homozygotes. Cross-resistance to trichlorfon, which in part results in the 4–10-fold increased tolerance to pyrethroids observed in strains lacking kdr, was insufficient to cause control failure. The quantification of the relationship between the frequency of an insecticide-resistant genotype, levels of resistance as measured by bioassays and control failure in the field provides the means of forewarning when control failure is likely to occur and of confirming resistance as the cause when it has taken place.