Two kinds of problems associated with developing standardized procedures for the safety evaluation of new compounds are identified. The first of these is the question of the desirability of using such standard methods. It is concluded that a basic set of procedures is to be recommended, but this should be supplemented with special tests as may be indicated. The second problem is connected with the technical difficulties of any given type of assay and is normally dealt with in terms of the state of the art at the time. Assays by the inhalation route tend to be custom designed and do not follow standard protocols. One of the causes of this situation is the propensity of individual investigators to design de novo the equipment used to effect exposure of animals to airborne substances. Second, some confusion exists because investigators do not always appreciate that the concentration-time product is not the same as the true dose received by the exposed subjects and this may lead to anomalies when dose-response relationships are being characterized. It is suggested that interlaboratory studies be undertaken to ascertain the variability that might be expected in independent assays of inhalation toxicity.