InDidaskalos 1968 I have argued that the chief task of the commentator on a short poem (an ode of Horace, an elegy of Propertius, or a poem of Catullus) is to reconstruct the hypothesis upon which the poem rests by piecing together the data actually provided by the text, or deducible from the text as a plausible, necessary deduction; to these data, I argued, one need only add what the poet feels he can rely on any reader to know, though two thousand years later it may of course require some research to recover what was common knowledge at the time. I cited Catullus 17 as a case in point—a poem where reconstruction of the hypothesis depended on things outside the text, things which, though part of the cultural background of the poet's contemporaries, were hardly part of ours, yet within our grasp if we cast around a little.