A very useful technique in toxicological investigation is probit analysis. This method, which was developed by Bliss and other workers (notably Gaddum and Finney, but see reference 4, pp. 42–46) consists of plotting the log concentration of a toxicant against the percentage response on a probability scale, and fitting a weighted regression line to the data. Although the technique has found wide application in pharmacology, entomology, and plant pathology, it has been almost ignored by workers in herbicide research. It is the purpose of this paper to demonstrate that the concepts of probit analysis are applicable to herbicide research, and that the mere plotting of data on logarithmic probability paper may permit a very fruitful interpretation of experimental results. The technique is recommended at the screening level, for its results are very useful in the construction of field tests. It is in no sense a substitute for the carefully designed and executed field test, or for the analysis of test results by other statistical techniques.