It is a major goal of clinical pharmacology to understand the dose-effect relationship in therapeutics. Much progress towards this goal has been made in the last 2 decades through the development of pharmacokinetics as a discipline. The study of pharmacokinetics seeks to explain the time course of drug concentration in the body. Recognition of the crucial concepts of clearance and volume of distribution has provided an important link to the physiological determinants of drug disposition. Mathematical models of absorption, distribution, metabolism and elimination have been extensively applied, and generally their predictions agree remarkably well with actual observations. However, the time course of drug concentration cannot in itself predict the time course or magnitude of drug effect. When drug concentrations at the effect site have reached equilibrium and the response is constant, the concentration-effect relationship is known as pharmacodynamics. Mathematical models of pharmacodynamics have been used widely by pharmacologists to describe drug effects on isolated tissues. The crucial concepts of pharmacodynamics are potency — reflecting the sensitivity of the organ or tissue to a drug, and efficacy — describing the maximum response. These concepts have been embodied in a simple mathematical expression, the Emaxmodel, which provides a practical tool for predicting drug response analogous to the compartmental model in pharmacokinetics for predicting drug concentration.The application of pharmacodynamics to the study of drug actionin vivorequires the linking of pharmacokinetics and pharmacodynamics to predict firstly the dose-concentration, and then the concentration-effect relationship. This may be done directly by equating the concentration predicted by a pharmacokinetic model to the effect site concentration, but this simplistic approach is often not appropriate for various reasons, including delay in drug equilibrium with the receptor site, use of indirect measures of drug action, the presence of active metabolites, or homeostatic responses, thus often necessitating the use of more complex models.The relative pharmacodynamic bioavailability of different preparations of the same drug may be determined from the time course of a drug effect. Bioavailability determined in this way may differ markedly from bioavailability defined by measurements of drug concentration if active metabolites are formed or if effects are produced in the non-linear region of the concentration-effect relationship.The influence of changes in the extent of plasma protein binding may be important in the interpretation of drug concentration measurements since it is generally held that only the unbound fraction is pharmacologically active. Clear examples of this phenomenon are few, but this reflects the general paucity of adequate observations rather than casting doubt on the usual assumption.The design of rational dosing regimens for clinical therapeutics cannot be performed with a knowledge of pharmacokinetics alone. The time course of drug effect may be essentially independent of concentration when a dose produces near maximal effects throughout the dosing interval. If effects are between 20 and 80% of maximum, the response will decrease linearly even though concentrations are declining exponentially. Finally, at relatively small degrees of effect, the time course of drug effect and concentration will be in parallel. The usual ‘rule of thumb’ of dosing every half-life is a conservative strategy for limiting wide fluctuations in drug effect, but demands more from the patient in terms of dosing frequency than may be necessary to achieve consistent drug action. On the other hand, if therapeutic success is dependent more on cumulative response than moment to moment activity, the use of extended dosing intervals may markedly reduce the effectiveness of the same average dose. Considerations of these factors can be incorporated into a dosing scheme by combined application of the principles of pharmacokinetics and pharmacodynamics.