Perhaps the first reaction to the title of this paper may be “why discuss probability? If that approach in hydrology has not already been discarded, it ought to be.” Certainly the critical attitude toward studies of probability expressed in much recent engineering literature would suggest this conclusion.The following quotation from an article “Possible and probable future floods” by William P. Creager in the November, 1939, issue of Civil Engineering is representative:“About 1914 the theory of probabilities was applied to flood studies; that is, curves were derived indicating by past records on a stream the frequency with which, during a long period, a given flood should be expected. Notwithstanding the fact that periods of record sometimes did not exceed 20 years and very seldom exceeded 30 or 40 years, these probability‐curves were extrapolated to estimate the flood which would be expected during long periods—once in 1,000, 5,000, 10,000 years, etc. Then, according to the judgment of the engineer, the 1,000‐, 5,000‐, or 10,000‐year flood was selected for the design‐cap