Extrapolation of animal data to assess pharmacokinetic parameters in humans is an important tool in drug development. Allometric scaling has many proponents, and many different approaches and techniques have been proposed to optimise the prediction of pharmacokinetic parameters from animals to humans. The allometric approach is based on the power function Y = aWb, where the bodyweight of the species is plotted against the pharmacokinetic parameter of interest on a log-log scale. Clearance, volume of distribution and elimination half-life are the 3 most frequently extrapolated pharmacokinetic parameters. Clearance is not predicted very well (error between predicted and observed clearance > 30%) using the basic allometric equation in most cases. Thus, several other approaches have been proposed. An early approach was the concept of neoteny, where the clearance is predicted on the basis of species bodyweight and maximum life-span potential. A second approach uses a 2-term power equation based on brain and body weight to predict the intrinsic clearance of drugs that are primarily eliminated by phase I oxidative metabolism. Most recently, the use of the product of brain weight and clearance has been proposed. A literature review reveals different degrees of success of improved prediction with the different methods for various drugs. In a comparative study, the determining factor in selecting a method for prediction of clearance was found to be the value of the exponent. Integration of in vitro data into in vivo clearance to improve the predictive performance of clearance has also been suggested. Although there are proponents of using body surface area instead of bodyweight, no advantage has been noted in this approach. It has also been noted that the unbound clearance of a drug cannot be predicted any better than the total body clearance (CL). In general, there is a good correlation between bodyweight and volume of the central compartment (Vc); hence, Vc does not face the same complications as CL. The relationship between elimination half-life (t 1/2 beta) and bodyweight across species results in poor correlation, most probably because of the hybrid nature of this parameter. When a reasonable prediction of CL and Vc is made, t 1/2 beta may be predicted from the equation t 1/2 beta = 0.693 Vc/CL.