Chord length measurements are used to estimate the volumetric distribution of air
voids within hardened concrete samples. The accuracy of the Linear-Traverse method is
questionable because chord lengths do not directly represent the actual volumes of the air voids.
Using computer simulation, randomly generated concrete samples were analyzed using the
Linear-Traverse method to compute the distribution of air voids, the chord length gradation
curve, the chord length-to-traverse ratio, and the air void volume gradation curve. The current
chord-length approach significantly underestimated the air void content. A method that
estimates that air void gradation curve from the chord length gradation curve is presented. The
computer simulation results are supported by a conceptual analysis. To obtain the most accurate
estimate of the air void content, computer analyses showed that at least 15 traverse lines should
be regularly spaced on the concrete sample.