In vivo pharmacokinetic studies were conducted in healthy adult human volunteers and the appropriate test species to determine the adequacy of the default uncertainty factors for inter-species differences and human variability in toxicokinetics. Toxicokinetic parameters from studies in humans were used to estimate the proportion of the population that would be adequately protected by the default factor of 3.16 fold. These data were compared with the results of similar studies in animals (rats or mice) to determine the adequacy of the toxocokinetic default factor of 4.0-fold for inter-species differences. Four food additives were chosen as test chemicals, butylated hydroxytoluene (BHT), curcumin, propyl gallate (PG) and thiabendazole (TBZ). Separate studies were conducted in which human volunteers received the food additive at the acceptable daily intake (ADI) or 10 x ADI, and studies in which the animal was dosed at the no-observed adverse effect level (NOAEL) or 0.1 x NOAEL. Plasma samples from these subjects were analysed for the additive and used to construct concentration-time curves; where appropriate a mathematical model was used to extrapolate parameters to infinity. The results of the studies show that the default factor of 3.16-fold for human variability is adequate for greater than 92% of the population for all test chemicals based on all PK parameters derived from the ADI studies. The inter-species default of 4.0-fold was adequate for all 4 test chemicals (comparing ADI and NOAEL data after correction for dose) based on oral clearance, and for PG and TBZ only based on Cmax. Interestingly, there was evidence of saturation kinetics in both humans and animals with TBZ. Oral clearance was ~9-fold and ~13-fold lower at the higher dose for humans and animals respectively. In conclusion these results support the current regulatory framework and the usual default uncertainty factor. However the data suggest that the dose used in the generation of pharmacokinetic data for deriving chemical-specific adjustment factors, for substitution into the framework, will need to be carefi-illy considered. The dose-response will need to be carefully characterised for the species of interest, and the doses adininistered should reflect the likely human exposure, so that the correct adjustment factor and therefore a more accurate ADI can be derived.