On an approximate method of determining the median effective dose and its error, in the case of a quantal response

Käber's method is, for many purposes, a sufficiently accurate way of determining the median effective dose, when we have a series of doses increasing in a constant ratio and a small number of animals on each. The standard error of the determination may also be rapidly found by a simple approximate method. Illustrative examples are given. The standard error calculated by the approximate method is based on the assumption that the animals are homogeneous. It must be used with caution when any heterogeneity is suspected. With animals of similar heterogeneity to those used here, the differences between two different toxic-fractions should only be regarded as significant if they are from 3 to 3.5 times their standard errors so calculated.