The necessity to protect humans from the adverse effects of chemicals on structure and function of the developing and/or mature nervous system is increasingly recognized among regulatory bodies throughout the world. Whereas structural changes dominated much of neurotoxicological research in the past, functional markers of neurotoxicity are gaining acceptance as early signs of insult. Among the more ambitious protective efforts are legal requirements of premarket testing for "behaviour-disrupting properties" of chemicals. Assumed or proven validity of cross species extrapolation underlies the use of animals in primary or secondary screening schemes. Two steps must be distinguished here. The first step is endpoint-base or qualitative, whereas the second one is dose-based or quantitative. Species comparisons in terms of endpoints is typically done within a framework of broad functional categories, such as sensory, motivational, cognitive, motor or social functions. Dose-based extrapolation requires knowledge about species differences in terms of toxicokinetics or metabolism in order to arrive at valid translations of dose-response contingencies. Principles of cross-species extrapolation in neurotoxicology will be exemplified by means of representative neurobehavioural and neurophysiological findings for neurotoxic chemicals of environmental concern, for which an adequate data base is available for comparative purposes, such as inorganic lead and polychlorinated biphenyls (PCB).