It is argued that chemical substances can be meaningfully ranked or classified according to their persistence (P), bioaccumulation (B), toxicity (T), and potential for long-range transport (LRT) only if these attributes can be shown to be intensive, as distinct from extensive, properties of the substance, i.e., they are independent of quantity of substance. It is shown that P, B, and LRT can be considered intensive or quasi-intensive properties, but toxicity is more problematic. To obtain an intensive metric of toxicity requires selection of one of several possible extensive quantities that define exposure or dose. Ranking of a group of chemicals by toxicity is shown to be very dependent on which quantity is selected. It is suggested that toxicity metrics, such as lethal concentration to 50% of the population (LC50), lethal dose to 50% of the population (LD50), and threshold limit value (TLV) suffer the severe disadvantage of being dependent on the efficiency of delivery of the substance to the site(s) of toxic action in the organism. The use of measured or calculated internal dose is a preferable measure of toxicity since it reduces ambiguities inherent in the other metrics. Also, the primary concern is not the quasi-intensive property of toxicity; rather, it is the risk of toxic effects, an extensive quantity. To adequately assess the risk of toxic effects, both the toxic hazard and the degree of exposure must be characterized. Since exposure cannot be estimated without knowledge of the emission rate of chemicals to the environment, a compelling case can be made that screening to identify priority P, B, T, and LRT substances should be expanded to include quantity released to the environment as an additional factor.