Documenting sexual dimorphism for structures that exhibit indeterminate growth can be more difficult than for structures exhibiting determinate growth. Most proboscidean tusks are ever-growing structures that change size and shape throughout life. Sexual dimorphism is pronounced in tusks of mature individuals, but the external form of tusks offers no clear evidence of maturation, and it is difficult to distinguish a young male's tusk from that of an older female. Thus, with previous approaches, knowledge of age was often required to assess sex from tusk measurements. This study examines sexual dimorphism of American mastodon (Mammut americanum) tusks through principal components analysis to determine which aspects of tusk form contribute most strongly to the variance among measurements and to explore the relationship between tusk form and individual age and sex. Twenty-one mastodon tusks from the Great Lakes region were evaluated in two analyses, the first focusing on geometrically distinct aspects of tusk form and the second adding measurements that reflect ontogenetic changes in a single aspect of morphology (circumference). Both analyses separated mastodons by sex (PC-I) and sorted them by age (PC-II). The distribution of tusks on the PC-II versus PC-I plane provides better discrimination of sex than univariate or bivariate methods because tusks of similar size and opposite sex appear near opposite ends of an age spectrum. The second analysis enhances sorting by age, thereby clarifying assessment of sex. This work contributes to studies of mastodon paleobiology by presenting a reliable method for assessing the sex of an individual from tusk measurements without requiring independent knowledge of age.