We compared nine linear measures of body size and their appropriateness as estimators for body size in apoid wasps. A correlation and a multiple regression analysis with backward elimination of non-significant variables of the linear parameters with dry weight were carried out for the entire data set and for various subsets (defined by sex, body size classes, and systematic position). All variables are significantly correlated with dry weight. Since all variables are also correlated with each other, multicolinearity effects are likely to occur, which reflect a biological fact rather than an extraneous factor to be eliminated. Head width, mesoscutum width, and forewing length are generally better predictors of body size in apoid wasps than any other variable. This result is in accord with the fact that these variables are already commonly in use in the literature. Our results provide selection criteria to choose among the variables to be best used in body-size related studies. If a single variable is required to estimate body size in apoid wasp, mesoscutum width should be used.