Femina in line 28 has nagged me subconsciously for years. I have now belatedly realized that it sabotages the poet's prudent disclaimer: it is not women in general who are in question, but only those not ruled out of bounds by stola and uittae. The repetition of the word in the following verse, where it means, as the opposition to uiri indicates, ‘the female sex’, only serves to underline its inappropriateness here. Cristante's defence of the anaphora, that it ‘ribadisce la necessità dell'insegnamento, introducendone la giustificazione’ (Ovidio, L'arte di amare, ed. E. Pianezzola [1991], p. 352), sets up an unwanted connection: lines 25–8 are strictly parenthetic to the main argument, as indeed is signalled by the truly functional anaphora of femina in line 29, whose effect is blurred in the text as transmitted. The form of the couplet, typical of Ovid, dictates that what is wanted is a variation on nil nisi lasciui, e.g. nec or, better, non proba. The source of the intrusive femina is not far to seek, though how precisely it ousted the original reading I do not pretend to guess.