A comparison of tolerance distributions in bioassay
Bioassay deals with the problem of determining the potency of a stimulus by analyzing the responses it produces in biological organisms, such as experimental animals, plants, bacteria, etc. The same methods used to determine the potency of a stimulus in biological organisms can be used to determine the sensitivity of non-biological organisms to various stresses and will be discussed in more detail in Chapter III. Nevertheless, responses vary in type and intensity from one member of the test population to the next. Each test subject, biological or non-biological, has a threshold or critical stress level, which must be exceeded in order to produce the response of interest.
For example, in testing the potency of an insecticide, there is a certain dosage above which an insect will die and below which it will not. In testing the sensitivity of explosives, there is a certain height above which a dropped explosive will detonate, and below which it will not. This gives rise to the concept of a continuous distribution function F, which models the response of interest. The distributions used to model the responses are referred to in the literature as tolerance distributions and are described in detail in Chapter IV. This paper will begin to compare how well five specific distributions; the normal, logistic, WeibuU, Gumbel, and Gompertz, fit quantal-response data in sensitivity experiments, which readily apply the methods of bioassay to non-biological investigations, as those dealing with explosives.