There are more than 70,000 man-made chemicals in widespread usage in our world today, with additions coming along at a clip of more than 5,000 per year. Besides these, hundreds of thousands of new structures are synthesized each year in many academic and industrial laboratories. Toxicological problems which may be associated with the latter must be ignored, at least until some serious accident forces them onto us or until they move into the marketplace. It is impossible to test all of these chemicals in all of the tests which have been devised and are continuing to appear. To make matters more confusing, there is no generally accepted way to collect and organize the streams of data from all kinds of tests on all kinds of compounds that flow from many laboratories. With governments and industrial laboratories planning to spend billions in the coming decades, it is urgent that more systematic means for the organization and discussion of these results be developed. It is viewed that a long-term objective of toxicology must be to devise a computerized data base of numerically defined, statistically validated, structure-activity relationships.