a few comedies and in the histories and historical tragedies, where Shakespeare's sources leave him less discretion, he shifts his scenes with a readiness outdone by no other playwright. The third act of Richard II requires no less than four localities, three of which have a castle, perhaps the same castle from the stage-manager's point of view, in the background. The second act of 1 Henry IV has as many. King John and Henry V pass lightly between England and France, All's Well that Ends Well between France and Italy, The Winter's Tale between Sicily and Bohemia, Cymbeline between Britain, Italy, and Wales. Quite a late play, Antony and Cleopatra, might almost be regarded as a challenge to classicists. Rome, Misenum, Athens, Actium, Syria, Egypt are the localities, with much further subdivision in the Egyptian scenes. The second act has four changes of locality, the third no less than eight, and it is noteworthy that these changes are often for quite short bits of dialogue, which no modern manager would regard as justifying a resetting of the stage. Shakespeare must surely have been in some danger, in this case, of outrunning the apprehension of his auditory, and I doubt if even Professor Thorndike's play of curtains would have saved him.
It is to be observed also that, in Shakespeare's plays as in those of others, no excessive pains are taken to let the changes of locality coincide with the divisions between the acts. If the second and third acts of All's Well that Ends Well are at Paris, the fourth at Florence, and the fifth at Marseilles, yet the shift from Roussillon to Paris is in the middle and not at the end of the first act. The shift from Sicily to Bohemia is in the middle of the third act of The Winter's Tale; the Agincourt scenes begin in the middle of the third act of Henry V. Indeed, although the poets regarded the acts as units of literary structure, the act-divisions do not appear to have been greatly stressed, at any rate on the stages of the public houses, in the actual presentation of plays.[1] I do not think that they were wholly disregarded, although the fact that they are so often unnoted in the prints of plays based on stage copies might point to that conclusion.[2] The act-interval did not necessarily denote any substantial time-interval in the action of the play, and perhaps the actors did not invariably leave the stage. Thus the lovers in A Midsummer-Night's Dream sleep through the interval between the third and fourth acts.[3] But some sort of break in the*
