Toxicology is both a science and an art. The science of toxicology cha
racterizes the adverse effects of a chemical by integrating all availa
ble scientific information. Toxicologists must ask whether current pro
tocols truly characterize the toxicity of a chemical in terms of ident
ifying all the potential adverse effects. The art of toxicology involv
es the use of a database to predict toxic risk. A key question for the
future of toxicology is ''can alternative models contribute to progre
ss or are they useless?'' This paper concludes in favour of progress i
f such alternative models help to improve the science of toxicology. A
s a practical approach to improve the art of toxicology it proposes re
inforcement of flexibility of current regulations, encouragement of th
e production and submission of data of alternative tests, and evaluati
on of these data on a strictly scientific basis. ECVAM, the European C
entre for the Validation of Alternative Methods, will play a major rol
e in implementing this new and more scientific approach to toxicology.