Modern accounts of the seventeenth-century French stage have repeatedly asserted that plays were divided into short acts of some twenty to thirty minutes in performance because the candles that lit the theatres had to be snuffed at frequent intervals. This article claims that there is no evidence for this assertion and aims to evoke the technological constraints of candle usage at the time so as to suggest that candles could be managed in such a way that they did not actually dictate dramaturgical practice. The article considers seventeenth-century theoretical discussion of the division of plays into acts: such discussion never alludes to candles, but refers to historical precedent and spectator attention spans as perceived explanations for the phenomenon of act division. It aims to adduce compelling evidence against the traditional view and concludes that the snuffing of candles took advantage of the opportunity offered by act division, but was never its cause. There is a tenacious orthodoxy in accounts of seventeenth-century French theatre concerning the use of candles and the length of the acts of plays. Acts, it is often alleged, had to be relatively short, and followed by intervals, so that the candles could be frequently snuffed. Wilma Deierkauf-Holsboer says that Henri Lyonnet was the first to express this view, and, though she gives no reasons for doing so, she categorically states that it is wrong: 'cette opinion est erronée'. 1 In the most up-to-date survey of lighting on the seventeenth-century French stage, Jan Clarke has repeated Deierkauf-Holsboer's opposition to any causal link