A number of superficially similar evolutionary phenomena are often lumped together as "alternative reproductive behaviors (ARBs)." Several authors have previously organized the broad array of reproductive alternatives by their relation to ESS theory. Because such an organizing scheme begs the question of how much ESS models can contribute to our understanding of ARBs, I suggest a different scheme-devoid of ESS connotations-which classifies alternatives according to whether they represent genetic differences between individuals (genotypic versus phenotypic alternatives), whether the alternatives can be expected to manifest equal or unequal fitnesses (isogignous versus allogignous alternatives), and whether individuals may switch back and forth between alternatives (reversible versus irreversible alternatives). I point out that plausible selective regimes other than frequency-dependence can maintain genotypic alternatives, and that explanations for the maintenance of phenotypic alternatives usually can be examined only theoretically. I also urge more rigor in the field-testing of ESS models, review criteria which putative ESSs should meet, and suggest a statistical approach for evaluating evidence with regard to equality of fitnesses of reproductive alternatives. © 1984 by the American Society of Zoologists.