Iris is not actually killed.]

[Footnote 11: See Chapter XVIII.]

[Footnote 12: See Chapter XX.]

[Footnote 13: Most of the dramatists whom I have consulted are opposed
to the principle of "roughing out" the big scenes first, and then
imbedding them, as it were, in their context. Sir Arthur Pinero goes the
length of saying: "I can never go on to page 2 until I am sure that page
1 is as right as I can make it. Indeed, when an act is finished, I send
it at once to the printers, confident that I shall not have to go back
upon it." Mr. Alfred Sutro says: "I write a play straight ahead from
beginning to end, taking practically as long over the first act as over
the last three." And Mr. Granville Barker: "I always write the beginning
of a play first and the end last: but as to writing 'straight ahead'--it
sounds like what one may be able to do in Heaven." But almost all
dramatists, I take it, jot down brief passages of dialogue which they
may or may not eventually work into the texture of their play.]

[Footnote 14: One is not surprised to learn that Sardou "did his
stage-management as he went along," and always knew exactly the position
of his characters from moment to moment.]

[Footnote 15: And aurally, it may be added. Sarcey comments on the
impossibility of a scene in Zola's _Pot Bouille_ in which the so-called
"lovers," Octave Mouret and Blanche, throw open the window of the garret
in which they are quarrelling, and hear the servants in the courtyard
outside discussing their intrigue. In order that the comments of the
servants might reach the ears of the audience, they had to be shouted in