In many methods of use of toxic substances to kill injurious organisms, the toxicant cannot be applied directly to the organism, but only to an easily accessible part of its habitat. This is true of fumigation of empty dwellings, stores of grain and soil. Pre-emergent herbicides may be applied to the surface of soil in which weed seeds lie at various depths. Bactericides may be ingested by hosts of bacterial parasites and systemic insecticides can be applied to accessible plant parts to attack hidden insects. In all these cases the dose received by the target organism depends on transport processes, chemical reactions and other means of loss, as well as on the dose applied. Within single organisms also there are usually localized sites of toxic reactions and the relationship of dose applied at some accessible part of the organism and dose eventually available at the site of reaction is subject to similar dependence on chemical and physical factors. There is, in fact, no distinction in kind, as far as physical factors are concerned, between site and organism on the one hand and organism, host and environment on the other. In this paper the word “dose” means an amount of toxicant applied or received, usually related to volume or area. For gaseous toxicants, the product of concentration in the ambient environment by the duration of exposure is often used as an alternative measure. A dose applied to some accessible part of a system will give rise to a variation of concentration with time in an internal part, rising from zero to a maximum and falling to zero again. The complete integral of concentration with respect to time is an internal equivalent of the external concentration time product. It is convenient to give this quantity a name as it must frequently be used but cannot properly be called a dose as it has different dimensions. The name “chemical impulse” is proposed as the quantity measures the capacity to produce chemical change. An alternative is “availance”. If the physical properties of the systems are assumed time invariant, diffusion and permeability coefficients are assumed constant and reactions are assumed first order, it can be shown that relationships between doses and impulses take on very much simpler form than can apply to rates of transfer and concentration-time curves. Relationships are derived, for simple model systems, between doses applied and doses received at an internal site in the systems, between doses applied and internal impulses generated and between impulses applied and doses received at an internal site. The models examined are of limiting simplicity and the conclusions arrived at are therefore applicable quantitatively only to simple problems of fumigation. Living organisms are more complex as physical systems, as well as in other ways, and therefore the relationships of impulse and dose will be more complex. Since toxic reactions are almost necessarily not first order, some examination is also made of concentration-time curves in order to predict qualitatively the type of relationship to be expected on purely physical considerations between toxic damage and duration of impulse. It is hoped that this analysis can help towards a better definition of the quantities recorded on the “dosage” axis of dosage response curves and towards a better understanding of the physical background against which more subtle physiological factors must be studied.