This article concerns the use of dance in the Act 3 finale of Mozart's Le nozze di Figaro, starting from the problem articulated by Alan Tyson in his Mozart: Studies of the Autograph Scores. Tyson points out that the absence of the fandango from the Viennese musical sources is at odds with Da Ponte's statement that the dance scene was restored at the emperor's command. New evidence shows that the fandango was performed for the three performances that constituted a première at this time in Vienna and was then removed from the score. However, before its removal, the score with the fandango intact was copied for at least one other theatre, hence accounting for the two versions that circulated through Europe. The article goes on to consider the dramatic function of the fandango by exploring the nature of the dance itself and examining the stage directions in the autograph in combination with those in Beaumarchais's play, several early librettos and editions, and the original first-desk first-violin part.