Anatomy of an Error

Marston’s theory for loads on buried pipe dates from the early 1900s and is commonly used for design. The derivation nevertheless contains a rudimentary error whose effect was neutralized by empirical adjustments based on full-scale experiments. The discrepancy went undiscovered until the 1940s when Terzaghi noted an inconsistency relative to failures of cellular cofferdams. The error then was addressed by Krynine, but no attempt was made to make design adjustments until development of a soil arching theory for retaining walls in the 1980s. The error is simply stated: The Rankine active pressure coefficient $K_a$ represents a ratio of principal stresses, and by definition a principal stress cannot contribute to friction. The history, significance, and lessons to be learned are discussed, and suggestions are made for modifications of the Marston presentation.