Image analysis of aggregates does not measure the same size as sieve analysis. The size of aggregates, determined by sieve analysis, is presented with respect to the percent cumulative mass, whereas image analysis does not measure mass. Results are often presented in percent particles or percent area. Several researchers have claimed that more accurate volume and mass determinations are necessary for accurate construction of grain-size curves. In the present work, several methods for reconstructing volume and thus mass of aggregates from image analysis (IA) have been tested to see how they influence the grain-size distribution curves. The actual mass of the individual particles was found to have little or no influence on the grain-size distribution curve, which is normalized and thus insensitive to mass. Accurate conversion of image-analysis size to sieve size is dependent upon how particles pass through sieves. Most existing methods base their conversion of image-analysis size to sieve size on the intermediate axis, multiplied by some factor. The present work shows that there is no direct correlation between the intermediate axes and sieve size. A universal conversion of image-analysis size to sieve size has been developed, using the minimum-bounding square around the minimum projected area. This measure yields very good correlation with sieve-analysis results.