1. In neurotoxicity, functional indices may be the only available measures of effect, as many potent neurotoxic agents produce no morphological change. Examples of these are strychnine, dieldrin and pyrethroids, which produce excitation but no pathology, and barbiturates, xylene and lithium, which produce depression but no pathology. 2. In other cases where both functional and morphological effects are seen, functional measures often produce the most convenient, if not always the most specific, indices of toxicity. Appropriate functional measures can be highly sensitive, both in humans and in experimental animals, and can also give vital mechanistic information. However, it is essential that functional measures are reproducible and interpretable (some behavioural measures are not) and also provide a reasonably exacting test of function (passive observation of resting behaviour can miss many effects). 3. In addition to their use as an index of toxicity, changes in function, even within the normal range, can themselves influence susceptibility to toxins. Tissue perfusion can determine delivered dose and is influenced by function, while metabolic transformation is modified by nutritional state. Nutritional state can also influence absorption, with anaemia enhancing manganese toxicity and calcium deficiency enhancing lead toxicity. Functional activity can influence target susceptibility directly: thus, noise exposure enhances the ototoxicity of carbon monoxide, toluene or aminoglycoside antibiotics; noise, motor activity or anaesthesia all influence the central neurotoxicity of dinitrobenzene or metronidazole; motor activity enhances the peripheral nerve toxicity of lead or thallium; and nerve regeneration enhances the toxicity of hexane. These functional factors can be very important in determining individual susceptibility.