THE SECRET ROSE J. B. Yeats Portrait of W. B. Yeats. November. 1896. first used as frontispiece to The Tab/es ofthe Law. The Adoration of the Magi. (1897) The Secret Rose, Stories by W. B. Yeats: A Variorum Edition

EDITED BY

WARWICK GOULD PHILLIP L. MARCUS, and MICHAEL J. SIDNELL

Second Edition Revised and Enlarged

M The first edition of this book was published with the aid of a grant from the Run Memorial Publication Fund of Cornen University

Copyright © 1981 by Cornell University Press Previously unpublished material by W. B. Yeats© Michael B. Yeats and Anne Yeats 1981 New editorial matter © Warwick Gould, Phillip L. Marcus and Michael]. Sidnell, 1992

Softcover reprint of the hardcover 2nd edition 1992 978-0-333-49257-4

All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 33-4 Alfred Place, London WCIE 7DP Any person who does any unauthorised act in relation to this publication may be liable to criminal prosecution and civil claims for damages.

First edition (Cornell University Press) 1981 Second edition (Macmillan) 1992

Published by MACMILLAN ACADEMIC AND PROFESSIONAL L TD Houndmills, Basingstoke, Hampshire RG2 I 2XS and London Companies and representatives throughout the world

British Library Cataloguing in Publication Data Yeats, W. B. (William Butler), [865-[939 The secret rose: stories. - 2nd ed. I. Tide 11. Gould, Warwick 111. Marcus, Phillip Le Duc, [94[- IV. Sidnell, Michael J. (Michael lohn), [935- 823.8 [F] ISBN 978-1-349-10879-4 ISBN 978-1-349-10877-0 (eBook) DOI 10.1007/978-1-349-10877-0 Preface to the First Edition

Yeats's stature as a writer of fiction is justly less than as a poet or dramatist; but his work in fiction is considerably better than has generally been recognized, and his stories deserve a wider audience. Familiarity with his fiction is essential for an accurate appreciation of his career. The development of Yeats was one of his major themes. He often directed attention to his earlier life and work in the later work, and it becomes ever clearer that our understanding of Yeats is impeded if we do not know the early work in its early forms and the transforma• tions it underwent through revision. Thus, for example, 'The Wisdom of the King' is one of his earliest explorations of the wisdom-power antinomy that appears in 'Leda and the Swan', 'Blood and the Moon', and other poems; and the modifications of the passage about the King's 'wisdom' would have to be considered by anyone tracing Yeats's attitudes concerning this concept. On the other hand, though 'Proud Costello' was first published in I896, it was not until I925 that Yeats symbolized Costello's gyring life by making his hero mount a winding stair. This Variorum Edition of a major unit of the fiction will enable the reader to follow in all its complexity the evolution ofthe stories included. In addition, the information it contains about the history of the collected edition of Yeats's works undertaken in the I930S has important implica• tions for the editing of virtually all of Yeats's published texts. Nearly a decade ago, we began working independently on editing projects involving the stories, then decided to combine our efforts; for the final product we share the responsibility equally. In work of this kind errors are almost inevitable, and this volume probably contains them; but we have done everything possible to make the edition complete, easy to use, and worthy of the author whose name it bears. The preparation of this edition required the help of many individuals, and it was generously given. Our foremost debt is to Senator Michael B. Yeats and Miss Anne Yeats, both for aiding our research and for granting

v Preface to the First Edition

permission to print the published and unpublished materials by W. B. Yeats included in this edition. Michael Horniman of A. P. Watt, Ltd., T. M. Farmiloe of the Macmillan Press, Ltd., and Lydia Zelaya of Macmillan Publishing Co. were particularly helpful in our dealings with their firms. Among the many librarians who patiently provided informa• ,tion we are particularly grateful to Nora Niland of the Sligo County Library and Museum; Lola L. Szladits ofthe Berg Collection, New York Public Library; P. Baker of the Sterling Library, University of London; Ann Hyde ofthe Kenneth Spencer Research Library at the University of Kansas; N. Hyde of the Library, Royal Holloway College; and the stafTs of the British Library, the Bodleian Library, the National Library of Ireland, the Library of Trinity College, Dublin, and the Manuscript Division ofthe New York Public Library. The stafTofCornell University Press ofTered invaluable assistance in the production of a very complex volume. Other individuals who have helped us in various ways include Barry B. Adams, Fredson Bowers, David R. Clark, Michael J. Colacur• cio, R. Cooper, Douglas Cowling, Alan Denson, Rache Lovat Dickson, Donald D. Eddy, Paul Epril, David Farmer, Richard J. Finneran, lan Fletcher, Joan Grundy, George M. Harper, the late T. R. Henn, SarahJ. Jaenike, ColtonJohnson, Sydney Josephs, John V. Kelleher, John Kelly, Richard Londraville, W. W. Lyman, Alexander Macmillan, the Rt. Hon. Harold Macmillan, Mrs. Thomas Mark, William McCarthy, Daniel J. Murphy, William M. Murphy, MaryO'Connor, Roger Parisious, Susanne Schaup, Felicity, Michael G., and Anne Sidnell, Jane Solomon, Robert Stacey, Jon Stallworthy, Henry Summerfield, Stephen Tifft, Peter Tim• merman, and Deirdre Toomey. Financial support for research was provided by the University of Queensland Foundation; the University of London Central Research Fund; Trinity College, the University ofCambridge; Trinity College, the University of Toronto; the Social Sciences and Humanities Research Council of Canada; and Cornell University. The texts from the proofs of Mythologies and The frish Dramatic Movement (1932) are published with the permission of Senator Michael B. Yeats, Miss Anne Yeats, and the Macmillan Press, Ltd. Other previously unpublished materials by W. B. Yeats are published with the permission of Senator Michael B. Yeats, Miss Anne Yeats, and the British Library. Unpublished correspondence from the Macmillan Archive is quoted with the perm iss ion ofthe Macmillan Press, Ltd., and the British Library. 'A Very Pretty Little Story' is published with the perm iss ion of Mr. Colin Smythe and the Henry W. and Albert A. Berg Collection, the New York Public Library, Astor, Lenox and Tilden Foundations.

VI Preface to the First Edition

Published writings ofW. B. Yeats are quoted by permission ofSenator Michael B. Yeats, Miss Anne Yeats, the Macmillan Press, Ltd., and the Macmillan Publishing Co., Inc.

P.L.M., Ithaca, New York; W.G., London; MJ.S., Toronto, 1980

Vll Preface to the Second Edition

For this new edition we have revised the Introduction to take fuller accoun t ofmaterials for the Dublin Edition ofthe Works ofW. B. Yeats wh ich Charles Scribner's Sons planned to issue in the late 1930S and later. The archives ofthese ma terials provide val uable insigh ts in to Y eats' s pu blishing methods in the 1930s, and much new information on the posthumous handling of his texts. They do not, however, alter our choice of basic texts, and so the main body of this book is reprinted without change. We have also taken the opportunity to correct one ofYeats's notes (that to The Tables of the Law and The Adoration of the Magi [1904 and 1905 reissue]), and to insert a note, omitted from the first edition, which Yeats included in volume 8 of the 1908 Colleeted Works in Verse and Prose. We are grateful to Professor Richard J. Finneran for pointing out this omission to uso We have selected from inscriptions to copies of Yeats's books those which shed light on the shared authorship of the revisions to Stories of Red Hanrahan (1905), and on Yeats's changing attitudes to his stories. We have also included so me of his memoranda for revision of certain stories, and added further information about the co pies he used when preparing copy for new editions, as weIl as augmenting the lists of surviving manuscripts and related materials. This revised edition has also provided the chance to include plates of the illustrations and cover designs of both The Seeret Rose (1897) and Stories of Red Hanrahan and The Seeret Rose (1927). Yeats was heavily involved in the decoration ofthese 'talismanic' books, and his involvement is the subject of a new appendix, which also includes a short passage found only in the page proofs of The Seeret Rose (1897) which provided the iconographical basis for Althea Gyles's cover design for that volume. The surviving archive of copy submitted by Yeats for the Dublin Edition is in the Harry Ransom Humanities Research Center, University of Texas at Austin, and we are especially grateful to the Director, Professor Thomas Staley, and his staff including Dr Cathy Henderson,

IX Preface to the Second Edition

Mr lohn Kirkpatrick and others for their expert assistance offered to Warwick Gould during a visit to the Center in the autumn of 1988. Miss Jean Preston and others at the Department ofRare Books, The Firestone Library, Princeton University, provided hirn with access to the Robert H. Taylor collection of rare books and kindly made available to hirn the Charles Scribne's Sons Authors' Files relating to the Dublin Edition. Unpublished materials from this archive, from the National Library of Ireland, the British Library, a private collection in Dublin and from the Charles Scribner's Sons materials in Texas are published with the permission ofMr Michael B. Yeats, Miss Anne Yeats, Charles Scribner's Sons, A. P. Watt Ltd., and the libraries. Warwick Gould is very grateful to the British Academy for the award of a travel grant to undertake the research necessary for this new edition, and to Professor Ronald Schuchard, Dr Keith Schuchard, Professor A. Walton Litz, Dr Omar Pound and Mrs Elizabeth Pound, Professor Stephen M. Parrish and Mrs Johnnie Parrish, the la te Dr Lola L. Szladits and others whose assistance and hospitality made his trip fruitful and memorable. He is also grateful to the Earl of Stockton for facilitating his research es at Birch Grove House among the archives of Harold Macmillan, first Earl of Stockton, and for arranging an interview with hirn in August 1980. The la te Norah McGuinness provided hirn in 1975 with copies ofYeats's letters to her about the 1927 illustrated edition. Anne Yeats provided the xeroxing and facilitated the contact with Norah McGuinness. The County Librarian of Sligo also provided xerox copies of Yeats's letters to Norah McGuin• ness. Dr Mark Dimunation, Rare Books Librarian ofStanford University Library, Patience-An ne W. Lenk, Special Collections Librarian of the Miller Library, Colby College, and Vincent Giroud, Curator of Modern Books and Manuscripts at the Beinecke Library, Yale all obliged with copies of annotated or inscribed materials. Omar Pound indefatigably provided further research in Princeton and Roger Nyle Parisious patiently deciphered astrological symbols for a neophyte. We are also grateful to the late Professor Kevin Danaher of University College, Dublin; and to Mr Colin Smythe of Colin Smythe Ltd., for checking various references to inscribed copies ofYeats's books. We are grateful to Sarah Roberts• West, Margaret Cannon and jill Lake of the Macmillan Press for their enthusiasm for this project, and to A. P. Watt Ltd., Dr lohn Kelly and the Oxford University Press for their permission to publish extracts from Yeats's letters. We are grateful yet again to Deirdre Toomey, Research Editor of Yeats Annual, for her shrewd critique of these revisions to the first edition.

W.G., London; P.L.M., Ithaca, New York; M.j.S., Toronto, 1989

x Contents

Frontispiece Preface to the First Edition V Preface to the Second Edition IX Abbreviations and Symbols Xlll Introduction XV

THE SECRET ROSE: TEXTS AND VARIANTS

Texts [rom Mythologies and The frish Dramatic Movement (I932 ) 3

THE SECRET ROSE 5 To the Secret Rose 5 The Crucifixion of the Outcast (1894-1932) 6 Out of the Rose (1893-1932) 16 Plates 1-8: Cover Design and Illustrations from The Secret Rose (1897) The Wisdom ofthe King (1895-1932) 25 The Heart of the Spring (1893-1932) 34 The Curse ofthe Fires and ofthe Shadows (1893-1932) 40 Where there is Nothing, there is God (1896-1932) 48 The Old Men ofthe Twilight (1895-1932) 54 Proud Costello, MacDermot's Daughter, and the Bitter Tongue (1896-1932) 61 STORIES OF RED HANRAHAN 83 Red Hanrahan (1903-1932) 83 The Twisting ofthe Rope (1905-1932) 96 Plates 9-37: Cover Design and Illustrations from Stories 01 Red Hanrahan and the Secret Rose (1927) Hanrahan and Cathleen, the Daughter ofHoulihan (1905-1932) 102

Xl Contents

Red Hanrahan's Curse (1905-1932) 105 Hanrahan's Vision (1905-1932) III The Death ofHanrahan (1905-1932) 117 ROSA ALCHEMICA, THE TABLES OF THE LAW, and THE ADORATION OF THE MAG! Rosa Alchemica (1896- I 932) The Tables of the Law (1896- I 932) The Adoration of the Magi (1897- I 932) NOTES

Other Texts I 75

The Binding of the Hair (1896- I 897) 177 The Book of the Great Dhoul and Hanrahan the Red (1897) / The Devil's Book (1892) onJacingpages 183 The Twisting of the Rope and Hanrahan the Red (1892- I 897) 198 Kathleen the Daughter of Hoolihan and Hanrahan the Red (1894-1897) 205 The Curse ofHanrahan the Red (1894-1897) 209 The Vision ofHanrahan the Red (1896-1897) 21 5 The Death ofHanrahan the Red (1896-1897) 221 The Rose ofShadow (1894-1897) 227

Dedications and Notes 233 Appendixes 237 r. V arian ts from Mythologies (r 959) 239 2. Line-End Word Division in the Copy Texts 25 1 3. Related Documents 255 Memoranda from a Notebook of 1893-95 255 'A Very Pretty Litde Story' 257 4. Census ofOther Manuscript and ProofMaterials 26r 5. The Illustrations and Cover Designs 27 r This appendix also contains a list oJ plates included in this volurne 6. Bibliography

xu Abbreviations and Symbols

CB The Chap Book IR The Independent Review MM McClure's Magazine NO The National Ohserver NR The New Review P The Pageant S represents variously The Sketch, The Savoy, and The Speaker, depending on the story in question; full details of periodical printings are given in the first note for each story. WS The Weekry Sun Literary Supplement 1897 The Seeret Rose (London: Lawrence & Bullen, 1897; New York: Dodd, Mead & Co., 1897; Dublin: Maunsel, reissued 1905) The Tahles of the Law. / The Adoration of the Magi. (London: privately printed, 1897) The Tahles of the Law and The Adoration of the Magi (London: Elkin Mathews, 1904) 190 5T The Tahles of the Law and the Adoration of the Magi (London: Elkin Mathews, 1905) [reissue of 1904T with new tide page] Stories of Red Hanrahan (Dundrum: Dun Emer Press, 1904) [actually published 1905] 1908 The Collected Works in Verse and Prose, vols. 5, 7 and 8 (Stratford-on• Avon: The Shakespeare Head Press, 1908) Stories of Red Hanrahan: The Secret Rose: Rosa Alchemica (London and Stratford-upon-Avon: A. H. Bullen, 1913) Stories of Red Hanrahan / The Secret Rose / Rosa Alehemica (New York: The Macmillan Company, 1914) The Tahles of the Law; & The Adoration of the Magi (Stratford-upon- Avon: The Shakespeare Head Press, 1914) Earry Poems and Stories (London: Macmillan & Co., 1925) Earry Poems and Stories (New York: The Macmillan Co., 1925) Both the above editions. [Where the two editions vary identically from the basic text the variant reading is identified as '1925.'] Stories of Red Hanrahan and The Seeret Rose (London: Macmillan and Co., 1927 See also Appendix 6 (pp. 29 1-97).

Xl1I Introduction

The stories by W. B. Yeats presented in this edition form a distinct and coherent unit within the body of his work. As the collective title for the Variorum Edition of these stories we have chosen The Seere! Rose. This entire unit has never before been published in a single volume, under this or any other title. Yeats used The Seeret Rose as the title of his I897 collection containing most of the stories, and three other stories that belong to this unit were published separately. Yeats later arranged the stories in three groups, using 'The Secret Rose' as the tide for only one of the groups. The three groups had no comprehensive tide, though together they correspond to the collection Yeats had originally called The Seeret Rose. We have used the title in a way the author himselfhad ceased to use it because it is the only Yeatsian title that has any claim to comprehensiveness and offers a traditional and convenient way to refer to alt the related stories. Despite his conviction that the revision of prose never culminated in the 'dick of the box' that signalled the final form of verse, Yeats revised these stories as assiduously as he did his poems. The stories were frequently republished and Yeats made full use of the opportunities for modifying his texts. The revision of the prose for stylistic, symbolic, and thematic reasons was extensive and protracted; he also grouped and regrouped the stories in various ways and even dropped some stories and added one new one. Further complications in the evolution of the unit arose from conftict with a publisher's judgment and experimental collaboration with Lady Gregory in rewriting the Red Hanrahan stories. In The Wild Swans at Coole (I9I9 edition), Yeats reintroduced into his work so me of his fictional characters of the nineties and referred, in the poem 'The Phases ofthe Moon', to 'that extravagant style ... learnt from Pater' of the stories in which these characters had earlier appeared. The reference to the Paterian style loses much of its point if the stories are remembered only in their later versions, for Yeats had been at some pains

xv Introduction to subdue that Paterian extravagance through reVISIOn. The stylistic revision of the stories was one aspect of a continuing and comprehensive attempt to bring aH his work to a symbolic and thematic unity. Revision of early productions was not, however, the only means by which Yeats attempted to unifY his work. By 1919, he was also beginning to derive a new synthesis from the contrast between the earlier and later work and selves. The new synthesis depended on the perception of distinctions between early and la te work which revision tended to blur. To appreciate these contrary aspects of Yeats's struggle for unity we need to know the early work in its early forms and the process of the revision as weH as the 'final' versions (where they survive) of his texts. That is to say we must come to terms with the reality of Yeats's 'multiple texts'. This edition is designed to conform to that reality and to stimulate apprecia• tion of the struggle for unity embodied in them. Given the peculiar character of the texts of the stories, we have concerned ourselves only in part with Yeats's 'final intentions'. His texts have been transmitted (with few exceptions) in a way quite different from those of most modern and nearly aH earlier authors. There were many printings of the stories and few over wh ich Yeats did not exercise elose supervision. I Each new printing was an opportunity for revision and rewriting as weH as correction. New versions were built on old ones• often, quite literaHy, on printed copy. Since authorial revision is so extensive from printing to printing, the nonauthorial elements become scarcely discernible. Textual errors may disappear not because they are corrected hut hecause the passages in which they occur are extensively revised for literary reasons, or even expurgated. On the other hand, an editor's punctuation or even an error may be adopted after the fact, in the context of a careful revision, and become Yeats's new intention. In this situation, the texts that were actually printed at each stage have the highest importance as the substantial embodiments ofYeats's intentions

IThe evidence of Yeats's supervision of the production of most of the editions concerned is found in published and unpublished correspondence and in extant proof materials. In three instances it is difficult to be certain that Yeats exercised supervision of the texts, though even with these editions there is some indication of Yeats's involvement in publication. One of them is Stories of Red Hanrahan: The Seeret Rose: Rosa Alehemiea, published in 1913; a letter of the same year (Letters of W. B. Yeats, ed. Allan Wade [London: Rupert Hart-Davis, 1954], 575-77; hereafter L) shows Yeats concerned that Bullen should keep the various parts of the Colleeted Works available in satisfactory versions. Another of the three editions contains twelve slight variants. This is The Tables of the Law; & The Adoration of the Magi, which appeared in the same year, 1914, as the heavily revised Stories of Red Hanrahan: The Seeret Rose: Rosa Alehemiea. Stories of Red Hanrahan anti The Seerel Rose (1927) was planned by Yeats (in collaboration with Norah McGuinness) as a talismanic book, though in this case there is no evidence of any renewed concern for the text, which he had recently revised for Early Poems and Stories (1925).

XVI Introduction of the moment and (even where they are in small particulars deficient in that respect) as the bases of subsequent versions. The chief interest of the printed texts lies in their relationship with one another, not in their bearing on Yeats's 'final' intentions; consequently this edition is designed to display not a 'best' text and the means (elimination of nonauthorial elements, adjudication among authorial readings, and so on) by which a 'best' text is produced-not this, but the full relations hip among the texts. In this edition, all the texts printed in Yeats's lifetime can be reconstructed in their entirety from the collations, and the evolution of each story traced through successive printings.

Canon and Order 01 the Stories Yeats published twenty-three stories,2 seventeen of which were included in The Seeret Rose (1897). 'Dhoya' appeared with the novella! John Sherman, as aseparate volume. 3 Two others ('Michael Clancy, the Great Dhoul, and Death' and 'The Cradles of Gold') were published in periodicals in the nineties. Unlike 'Dhoya', 'Michael Clancy, the Great Dhoul, and Death' was early on entertained by Yeats as part of the Secret Rose unit. 4 Yeats also might have entertained the idea of including 'The Cradles ofGold' in The Seeret Rose, since as late as November 21, 1896, United Ireland could report that he intended to include 'twenty stories' in that volume and that story appeared during November in The Senate. However, as the story does not seem to have been included in the page proofs (see below P.263), and as there is no evidence that Yeats was forced to omit that story, it would appear that, had he indeed entertained the idea of including it, he made a clear artistie decision to exclude it. In the 1897 collection the seventeen stories were in the following order: The Binding of the Hair The Wisdom of the King Where there is Nothing, there is God The Crucifixion of the Outcast Out of the Rose The Curse of the Fires and of the Shadows

2This tally does not incIude John Sherman, the stories in The Celtic Twilight, Stories 01 Michael Robartes and His Frienris, or the posthumously published The Speckled Bird. 3John Sherman and Dhoya by 'Ganconagh' was originally published, in 1891, in Unwin's Pseudonym Library. The volume has been edited by RichardJ. Finneran (Detroit: Wayne State University Press, 1969). 4See P.256, and Uncollected Prose by W. B. Yeats, Vol. I, collected and edited by John P. Frayne (London: Macmillan, 1970) for both 'Michael Clancy, the Great Dhoul, and Death' (pp. 310- 17) and 'The Cradles ofGold' (pp. 413-18).

XVll Introduction

The Heart of the Spring Of Costello the Proud, of Oona the Daughter of Dermott and of the Bitter Tongue The Book of the Great Dhoul and Hanrahan the Red The Twisting of the Rope and Hanrahan the Red Kathleen the Daughter of Houlihan and Hanrahan the Red The Curse of Hanrahan the Red The Vision of Hanrahan the Red The Death of Hanrahan the Red The Rose of Shadow The Old Men of the Twilight Rose Alchemiea InJune 1897, a small volume eontaining two other stories, 'The Tables of the Law' and 'The Adoration of the Magi', was published 'privately', though with the fleuron of Lawrenee and Bullen (the publishers of The Secret Rose) on the title page. Yeats himself was responsible for an unsigned note in this book informing readers that 'these stories were originally intended to follow "Rosa Alchemiea" in "The Seeret Rose.",5 Yeats elaborated the note in Early Poems and Stories (1925), in whieh he declared that the two stories 'were intended to be part of "The Seeret Rose", but the publisher, A. H. Bullen, took a distaste to them and asked me to leave them out, and then after the book was published liked them and put them into a little volume by themselves'. 6 EIsewhere Yeats was even more emphatie, saying Bullen 'made me leave them out'. 7 In the Collected Works of 1908 the two stories were reunited with 'Rosa Alchemiea' but in the 1913 and 19 I 4 editions of The Secret Rose they were omitted onee again. In 1914, the two stories were published as aseparate volume again. Early Poems and Stories (1925) reassociated 'The Tables of the Law' and 'The Adoration of the Magi' with 'Rosa Alchemiea' as the triptyeh Yeats had first intended, and the Mythologies and The lrish Dramatic Movement proofs of 1931-32 preserve this arrangement. In la te 1897, then, nineteen stories probably eonstituted the unit in Yeats's final intention for the first inearnation of The Secret Rose. In 1905, the separate publication of the 'Hanrahan' stories, redone with Lady Gregory's help to bring them 'closer to the life ofthe people',8 rearranged

5The note appears on page 4. See below p. 234. 6Early Poems and Stories (1925), p.528. Early in ]anuary Yeats had written from Paris to 'Fiona Macleod' (William Sharp) to say that the book had been delayed from December until February, indicating that at that stage, 'The Adoration of the Magi', a story 'half prophecy of a very veiled kind' (L, 280), was still included in its scheme. 7See pp. 26g-70 below for the fuH statement from which this phrase is taken. S Early Poems and Stories (1925), p.528 and below p. 173. See also L, 36 I.

XVlll Introduction

the overall pattern of the unit. In the I897 volume, the Hanrahan stories had not been presented as a distinct group, though they did form a connected set within the collection; from I905, the stories of Hanrahan were treated as a group distinct from the Secret Rose stories. Moreover, 'The Book of the Great Dhoul and Hanrahan the Red', the first story in the set as published in The Secret Rose, was replaced by an entirely new story, 'Red Hanrahan', in the separately published set. In I908, Yeats abandoned 'The Binding of the Hair' and 'The Rose of Shadow'. 'Where there is Nothing, there is God' survived into the I908 volume but thereafter Yeats vacillated. He revised the story extensively for the I9I4 volume, omitted it from the editions of I925 and I927, then restored it to the collection in I 93 I when he began to prepare the Mythologies and The frish Dramatic MODement text. This text of I93I-32 includes seventeen stories arranged in three groups: 'The Secret Rose' (eight stories); 'Stories of Red Hanrahan' (six stories); 'Rosa Alchemica', 'The Tables ofthe Law', and 'The Adoration of the Magi'. The present edition comprises the seventeen stories that appear in Mythologies and The Irish Dramatic MODement, and the three stories (including 'The Book of the Great Dhoul .. .') that were aban• doned after their appearance in The Secret Rose-in all their printed forms. The scope ofthis edition, that is to say, is determined by the most inclusive possible conception of the unit. The stories of The Secret Rose and The Tables of the Law & The Adoration of the Magi were arranged as a sequence of spiritual events through history, from the Heroic Age to the present. 9 In subsequent editions this order was alte red by omission and rearrangement and its attractive thematic implications lost. The arrangement in this edition, respecting the conscious evolution of Yeats's intentions, follows that of the 1931-32 text. Abandoned texts are placed in aseparate section.

The Basic Texts The basic text for most of the stories is that of the page proofs of Mythologies and The Irish Dramatic MODement, set in I 93 I for Macmillan of London, corrected by Yeats in I932, and never published. This text differs in numerous details from the posthumous Mythologies published by Macmillan in I959. For 'The Binding of the Hair' and 'The Rose of Shadow', abandoned after their appearance in The Secret Rose (I 897),

9See Phillip L. Marcus, Yeats and the Beginning of the frish Renaissance (Ithaca, N.Y.: Cornell University Press, 1970; and second edition, Syracuse, N.Y.: Syracuse University Press, 1987) pp. 49-50.

XIX Introduction the 1897 volume is the source of the basic text. Between the periodical and 1897 versions of the Red Hanrahan stories and the versions published in 1905 and after the differences are so great that they cannot be presented through variant readings: thus in this edition the 1897 texts of the stories are printed separately as a basic text, with variants from the periodical vers ions that preceded them; while the 1905 and later versions are coIlated against the basic text of the Mythologies and The frish Dramatic Movement proofs.

The Origins of the Mythologies and The Irish Dramatic Movement Text of /93/-32 with an Account of Yeats's two Edition de Luxe Projects

/. The Macmillan Edition de Luxe The page proofs of Mythologies and The frish Dramatic Movement now in the National Library of Ireland 10 were prepared as Volume 11 of a projected Edition de Luxe of the works of Yeats. Correspondence concern• ing the edition and proof materials of most other volumes can also be found in the Macmillan Archive at the British Library, II in the firm's own archive at Basingstoke,12 as weIl as in the National Library of Ireland. The surviving fragments of the edition and the correspondence provide crucial information about Yeats's last version of the Secret Rose stories and a rationale for the choice of a basic text. They are also crucial documents for the re-editing of aB of Yeats's 'canonical' works. 13

IOMS. 30030, formerly in the collection of Michael B. Yeats. Hereafter MSS. in the National Library are ci ted in the text using the symbol NLI followed by the MS. number. 11 Dept. ofWestern MSS., The British Library. Materials from this Archive are hereafter cited in text using the symbol B.L. Add. MS. and a MS. and folio number. Uncatalogued pagers sold by Macmillan to the British Library in 1990 are cited as 'B. L. Uncat.'. 2Which includes papers formerly in the private collection ofthe late Earl ofStockton• who as Mr Harold Macmillan had been Yeats's editor during the thirties-at Birch Grove House, Chelwood Gate, W. Sussex. These are cited following the symbol Basingstoke. Warwick Gould interviewed Mr Harold Macmillan at Birch Grove on August 12, 1980. 130n the Macmillan Edition de Luxe and its derivatives, the Coole Edition (also abandoned), Poems (1949), and the Uniform Edition of the Prose Works in the 1950S and 1960s, as weil as the abandoned Dublin Edition which Charles Scribner's Sons proposed from 1935 and did not abandon until 1953, see Warwick Gould, 'The Definitive Edition: A History of the Final Arrangements of Yeats's Work' which is Appendix 6 of A. Norman Jeffares (ed.), Yeats's Poems (London: Macmillan, 1989), pp. 706-49. It is hereafter ci ted as 'Gould, The Definitive Edition'. See also Gould's forthcoming Yeats's Permanent Self: An Essay in Textual Biography. On the Macmillan Archive see W. E. Fredeman, 'The Significance of a Publisher's Archive: The Macmillan Papers', Studies in Bibliography 23 (1970), 183-91; and Philip V. Blake-Hill, 'The Macmillan Archive', British Museum Quarterly, 36 (Autumn 1972), 74-80. A partial account of both the Macmillan Edition de Luxe and the Dublin Edition projects will be found in Richard J. Finneran, Editing Yeats's Poems: A Reconsideration (London: Macmillan Press, 1989, hereafter EYPR).

xx Introduction

A collected edition ofYeats's works had been a condition ofhis 1916 agreement with Macmillan, and by December 23, 1920 Yeats wrote to the firm: 'I don't think it likely with book producing at its present expensiveness that you are anxious to publish that six volume edition of my collected work, which you will find in our original agreement. I wonder if you might not do something short of it with profit' (B.L. Add. MS. 55°03, f. 64). There follow Yeats's plans for what in fact became The Collected Works wh ich Macmillan issued from 1922 onwards. But the question of an Edition de Luxe, as the project was termed until after Yeats's death (when it was named the Coole Edition by George Yeats), remained open. By early February, 1930, however, Macmillan were thinking of clearing away the obstacles in the form of the rights by this time held by Ernest Benn Ltd. in Poems (1899, reprinted many times) in preparation for publishing the Edition de Luxe (B.L. Add. MS. 55704, f. 195). The rights were due to expire in May, 1933 anyway, but the matter was not finally settled until March 1, 1933 (B.L. Add. MS. 55733, ff. 424, 589; 54902, ff. 4, 16; 55737, f. 291), and in the intervening period many details of the Edition de Luxe were gradually settled (see EYPR 7- 8). Yeats was able to write to Olivia Shakespear on December 27, 1930:

Macmillan are going to bring out an Edition de Luxe of all my work published and unpublished. The unpublished to include AVision (rewritten of course), my Wheels and Butterjiies (a book of plays and essays almost finished) and By;:;antium (the new book of verse [i.e. Tke Winding Stair and Otker Poems]). The new stories of Hanrahan and those Cuala diaries and the Sophocles versions. I am to be ready next autumn at latest. Months of re-writing. What happiness! (L 780)

Yeats accepted the financial arrangements and approved a specimen page by January 15, 1931 (B.L. Add. MS. 54901, f. 114), and the Memorandum of Agreement (dated April 17, 1931) was returned by A. P. Watt & Son from Yeats to Macmillan on May 4 (B.L. Add. MS. 54901, f. 160). Yeats began to think about the deployment ofhis work in seven volumes by May 7, 1931 (B.L. Add. MS. 55003, f. 121). Macmillan proposed then to publish by September 30, 1932, but included in their contract the usual clause which effectively gave them the power to judge if the market conditions were right for the release of a seven volume, signed edition limited to 375 copies, containing old and new work (see EYPR 8). By June I Yeats had met with Harold Macmillan to settle the contents of the individual volumes of the set (B.L. Add. MS. 54901, f. 171). In an undated letter of late 1930 to Sir Frederick Macmillan, Yeats added in a postscript, 'You have the best reader for the press I have ever

XXI Introduction come across' (B.L. Add. MS. 55003, f. 119). The tribute was to be relayed to Thomas Mark (1890--1963), and it is clear that throughout the 1930s a special relationship developed between Yeats and his publisher's reader. Charles Morgan has given some idea of the responsibility which Yeats entrusted to Mark:

When a complete edition was projected, his whole works were elaborately studied on his behalf, phrase by phrase and comma by comma, before being submitted to his personal care. The edition has been delayed by the war, but all ofit was seen and revised by Yeats. He attended to every point that was raised, explaining his meaning where he thought it might have been missed, and writing: 'For the first time there will be a satisfactory text of my work, thanks to your watchfulness and patience.' 14

Morgan accurately states Yeats's interest in the work, although he overstates his carefulness and the extent of the edition which he had in fact seen and approved, for much newer work (and some older work, such as Essays [1924]) was not in fact seen by Yeats into Edition de Luxe proofs, and so the final text of the entire edition was not seen and approved by Yeats. Also, Yeats did not in fact comment upon every point raised by Mark on proofs, but his silence is usually explained by some general delegation of responsibility to Mark. However, his widow and executrix had in Mark a publisher's reader whom she knew Yeats had trusted and the comment does give a clear sense of Mark's role. Yeats, for example, had been happy to delegate to Mark full responsibility for checking the final revises of Tke Winding Stair and Otker Poems (1933), and they were not seen by the poet hirnself. The extent of Mark's responsibilities and Yeats's trust in hirn in• creased as work on the edition advanced. Yeats wrote to Harold Macmillan on September 8, 1932: 'I would be very much obliged ifyou would give the enclosed letter to the admirable scholar who is assisting in the correction of the proofs of my new collected edition. It is partly a letter of thanks and partly an explanation of certain metrical tricks of mine which have puzzled hirn' (B.L. Add. MS. 550°3, f. 136). The 'enclosed letter' is not in the Archive, and is likely to have been among certain mementos of Thomas Mark's which were sold to an untraced American after Mark's death, 15 but it seems likely that this was the letter which contains the tribute quoted above in Charles Morgan's account. Thomas Mark showed the letter to Jon Stallworthy, who quoted from it:

14Charles Morgan, The House 01 Macmillan (1843-1943) (London: Macmillan, 1943), p.223· 15Mrs Thomas Mark kindly discussed her husband's work with Warwick Gould before the first edition of this book.

XXlI Introduction

'Again, as late as 1932 he wrote to his publisher: "I have never been able to punctuate properly. 1 do not think 1 have ever differed from a correction of yours in punctuation. 1 suggest that in the remaining volumes you do not query your own corrections.'" 16 That this was the letter in question seems likely in the light of a letter from Mark to Yeats dated September 16, 1932, in which Mark refers to the 'metrical tricks' and thanks Yeats for having offered hirn 'a free hand with the punctua• tion'.17 This was also the letter recalled by George Yeats in a letter of April I 7, 1939, when she wrote to Mark: 'WBY wrote to you in September (or October) 1932 about punctuation and generally asking your help, without which he knew he could never get his work into the final form he wished [emphasis added]. There were a few metrical "tricks" as he called them, and tricks of repetition of words and phrases, deliberately used, which we should, 1 think, carefully preserve'. 18 The first volume of the seven was to have contained the poems, and the second was entitled Mythologies and The frish Dramatic Movement. Its contents included The Celtic Twilight; the 'Secret Rose' and 'Red Hanrahan' groups ofstories, 'Rosa Alchemica', 'The Tables ofthe Law', 'The Adoration of the Magi'; as weIl as The frish Dramatic Movement. It be comes clear, then, that the tide Mythologies embraced all ofthese works except the last named. The Mythologies proofs bear printer's date stamps of various dates between September 30, 1931, and October 26, 1931. Anxious to plan his summer's work, Yeats wrote to Harold Macmillan on May 15, 1932, asking for proofs, wh ich were promised in 'about a fortnight's time' (B.L. Add. MSS. 55003, f. 124; 55728, f. 397). So me delay was caused by Thomas Mark's thoughtful reading of the proofs, but it was already clear that Macmillan were in no hurry to publish until Benn's rights in Poems (1899) had expired, by which time the market, they hoped, might be more promising for an Edition de Luxe. Besides, there were other delays. Yeats had to advise the publishers to notify hirn separately when sending proofs: the de Valera government had, said Yeats, 'reduced our Custom House to chaos' (B.L. Add. MS. 55003, f. 126, June 3, [1932]). By June 17, however, Yeats had expressed so me

16Jon Stallworthy, Between the Lines: Yeats's Poetry in the Making (Oxford: The Clarendon Press, 1963), p. 12. 17 Letters to W. B. Yeats edited by Richard J. Finneran, George Mills Harper and William M. Murphy, 2 vols., (London: Macmillan, 1977) 11, pp. 543-44. Mark in that letter prided hirnself on being 'specially responsive to the subtleties of the beautiful rhythms' in Yeats's work, and indicated that he had only queried lines 'in which it often seemed possible that a word had been accidentally dropped or retained in an earlier printing ofthe text' (ibid.). 18NLI 30248. George Yeats's wording strongly suggests that Yeats kept a copy of his tribute to Mark, or perhaps a draft of it. If so, it has yet to come to light among the Yeats papers. For Mark's reply, see below, p. xxxviii.

XXIIl Introduction impatience (B.L. Add MS. 55003, f. 128), and Harold Macmillan replied, chiding Yeats for failing to supply all the revised copy:

Your letter reached us this morning, and I write to say that we are sending, by registered post, a complete marked set ofproofs ofVolume I (Poems) of the Edition de Luxe of your works, together with the printed pages and other material with which you supplied us for use as 'copy'. We expect to be able to forward Volume 11 on Monday. Perhaps you could let us have an idea of when you are likely to go to the United States, as we could then do our best to let you have the bulk of the proofs before you leave. No doubt you remember that we have not yet received any material for Volume VII (Discoveries), and that you are going to supply us with additional essays for Volume V. You were also thinking of substituting a new version of the opening essays in Volume IV (the second volume of Plays) for the one now in type, and if you could let me have this 'copy' we could set it up and need not trouble you with the proofs except in this revised form. (B.L. Add. MS. 55729, ff. 477--8). The proofs for Volume II (Mythologies and The frish Dramatie Movement) were forwarded as promised on June 22, 1932, together with the printed copy from which they had been set up (B.L. Add. MS. 55729, f. 605). Yeats went over the proofs hirnself and returned them to Harold Macmillan on July 5, 1932: 'I return the proof sheets of two volumes of the new collected edition. The volume ealled "Mythologies" f need not see again. Your reader ean eomplete the revision better than f eould [emphasis added]. 19 I want however a "revise" ofthe whole ofthe volume ofpoems. It is I think the only troublesome volume owing to the fact that I have written a great deal since I sent you the typed copy.' (B.L. Add. MS. 55003, f. 129). The prose volume had been easier to prepare than the volume of poetry, not because the correction of prose was intrinsically less arduous, but because the volume of poetry contained so much recent work. Yeats certainly felt that the prose was being given the same careful treatment as the poetry. Correction where there had not been substan• tive revision he could confidently trust to the consistent and precise mi nd of Thomas Mark.20 The next problem with the Edition de Luxe project was precipitated by Yeats hirnself. On August 9, 1932 when he was correcting proofs of Volume III (Plays, Vol. I), he wrote to Macmillan to say that he was publishing 'a sheaf of lyrics' (Words for Musie Perhaps and Other Poems)

19See also below, pp. 3-4· The short tide Mytlwlogies was frequently used: see B.L. Add. MS. 55727, ff. 271-3, 15 April 1932. 21The errors which in fact do remain in the text are discussed in 'The Page Proofs of 1931-32' (pp. xl-xliii) and also at pp. 3-4 below.

XXIV Introduction with the Cuala Press (B.L. Add. MSS. 55°°3, ff. 131-2; 55731, f. 85,July 25, 1932), and thereby violating the letter of the agreement he had entered into for the Edition de Luxe, which was to contain any new work. He was also running the risk of losing his copyright in America if he did not publish the new poems there within six months. He preferred to secure the copyright in America after publishing the new lyrics in the Edition de Luxe, and thus, in effect, he was asking Macmillan whether there was any chance of its appearing within six months. Harold Macmillan's reply cast the shadow of the Depression over the immediate future of the whole edition:

I do not think we could have any objection to this, [Cuala volume] in spite of our agreement, because I am afraid we have taken a long time to produce the Edition de Luxe. The fact is that the work has taken us rather longer than we expected, and also, to be quite frank with you, conditions are so bad at the present time that I am not particularly anxious to hurry on the production, as I think it would hardly be to our interest to do so .... I am very glad to hear that you can let us have the rest of the contents of the seven volumes before you leave for America in October. I will look very carefully into the whole situation again and write in a week or ten days' time. My own feeling is that we ought to wait now and see whether this Autumn may bring rather beUer times to the world in general and the book trade in particular, with a view to announcing and subscribing the Edition de Luxe early next spring. This is only my own present idea, and circumstances might arise which might alter it. I am very much obliged to you for writing to me so franklyon the whole matter. (B.L. Add. MS. 55731, ff. 405-7, August 11, 1932) Yeats was 'relieved' to get this reply, but his relief was due to the fact that the firm had not remonstrated with hirn concerning 'the spirit of the agreement'; delay held no advantages for hirn. Indeed, in the summer of 1931 when anxious about his income, he had written to Watt 'counting on some substantial gain from the edition de luxe'. The alternative then (as ever) appeared to be to publish in America, against the spirit but not the letter of his agreement with Macmillan, wh ich only forbade publica• tion in America of signed limited editions. But Yeats 'hate[d] correcting American proofs', and resented the loss of control over the finish of his work which he felt the established relations hip with Thomas Mark assured hirn. American publishers were just too far away (NLI31018, draft undated letter to Watt). On August 13, 1932, Yeats pressed for the revise of Volume I, and Macmillan's note accompanying the proofs read: 'this revise, which is accompanied by your own marked proofs, has been checked, but will be

xxv Introduction read through carefully once more before it is passed for press.'21 It appears that the firm hoped, even though the project was for the moment in active abeyance, to finish preparatory work on Volume I quickly. At the same time, proofs of Volume VI (Autobiographies) , together with the copy from which it had been set up, were sent to Yeats (B.L. Add. MS. 55731, f. 569). Proofs of Volume IV (Plays, II) began to be sent on 29 August, 1932, but the printers could go no farther than page 288 until Yeats supplied the Introduction to Wheels and Butterflies, which was sent shortly afterwards (B.L. Add. MS. 55732, ff. 42, 122). Yeats had determined to leave for the United States, via England, on October 6, 1932 (B.L. Add. MS. 550°3, f. 137). At that time he was still working on the revise of V 01 urne I; had considered hirns elf done wi th the marked proof of Mythologies and The frish Dramatic Movement on July 5; had still in his pos session proofs ofVolume III and the first 288 pages of Volume IV (both volumes of plays); had promised extra material for Volume V (Essays); and still had the proofs of Volume VI (Autobiog• raphies) (B.L. Add. MSS. 55731, f. 85; 55732, ff. 42, 366; 55003, ff. 129, 137). He had not submitted copy for Volume VII (Discoveries) (which would have contained what in fact became AVision (1937)). The project was well in hand by the autumn of 1932, but hopelessly behind the schedule Yeats had happily planned as 'months of re-writing' in the letter to Olivia Shakespear of December 27, 1930, which envisaged that his own work on the edition would be completed by 'next autumn at latest' (L 780). From this point on the Edition de Luxe remained in active abeyance until Yeats's death, delayed but by no means abandoned. Yeats's newer work was appearing all the time, and being published in the usual way, and throughout the intervening years, Yeats, Harold Macmillan and Thomas Mark corresponded from time to time about absorbing such work into the Edition de Luxe. In October, 1936 new lists of contents for that set were prepared by Macmillan and annotated by Yeats during November (NLf 30202). OnJune 22,1937 Yeats and Harold Macmillan met to discuss various of his projects (B.L. Add. MSS. 55795, f. 298; 55796, ff. 128-g; 54904, ff. 14-18, June 23, 1937). While the Edition de Luxe was delayed, Macmillan nevertheless wanted to get newer work into proof, and obtain 'clean revises' (B.L. Add. MS. 55800, f. 447, October 13, 1937). At that time, Yeats was also eager to consult Thomas Mark about rearrangements ofhis plays and essays which he had effected for the other de luxe collected edition of his works, the Dublin Edition,

21B.L. Add. MSS. 55003, f. 135,55731, f. 569, August 23,1932. Pages 1-384 only were sent on this date.

XXVI Introduction which also was never to be published, but for which he had submitted copy (B.L. Add. MS. 55003, f. 174, October 15, [1937]).

2. The Charles Seribner's Sons Dublin Edition Late in 1935, Charles Scribner proposed to George P. Brett of the Macmillan Company, New York (Yeats's American publishers), that his firm be allowed to produce a subscription set ofYeats's works to be 'sold solely by mailorder and house to house canvas [sie]'. 22 Brett anticipated that Yeats stood to earn some $3500 to $4600 from the issue. Yeats's agent accepted the offer by December 3, 1935 (Callan 91), but not before negotiation with Harold Macmillan. He wrote to Watt after Yeats had written on November 18: 23

... you and he [YeatsJ will realise that the publication of our edition, a great deal ofwhich is now in type, had better be postponed until wejudge that the general situation will justify the publication of another collected edition. Although I do not suppose that many of the American sets will reach this country, some ofthem will; and it is therefore important that we should be allowed to judge as to the moment when it may be possible to produce our edition. Since, however, we have, quite frankly, felt difficulty in pursuing our edition with any vigour, in the present state of the market, I am very glad that this American proposal has come along, which will give Mr. Yeats a substantial profit and will not impede the publication of an English edition. By the way, I shall be glad ifa condition can be made that we shall have a set of the American edition. It will be useful to us for many purposes. (B.L. Add. MS. 55774, ff. 153-4, November 27, 1935)

22See Edward Callan, Yeats on Yeals: The Last Introductions and the 'Dublin' Edition (Dublin: The Dolmen Press, New Yeats Papers Series XX, 'g8" hereafter cited as Callan) , p. go, November 8, '935. The correspondence relating to the Dublin Edition is largely held in the Department ofRare Books, Firestone Library, Princeton University, in the Charles Scribner's Sons' Authors' Files Series I, Box '74 and Series II, Box 37. It is hereafter ci ted using the symbol Princeton and a date. The copy for the Duhlin Edition is now at the Harry Ransom Humanities Research Center, University of Texas at Austin, and is cited by Box and Folder number after the symbol HRHRC. For an account ofsome of these papers see Richard J. Finneran, 'A Note on the Scribner Archive at the Humanities Research Center' in Finneran (ed.), Yeats: An Annual 01 Critical and Textual Studies (Ithaca: Cornell University Press, 'g84) Vol. 2, pp. 227-32. The papers have, however, been rearranged somewhat since that artide was published. 23'1 dont want to do anything unfair' (B.L. Add. MS. 55003, f. '93; NLI 30248). He left it to Watt and Macmillan to negotiate, instructing Watt on the same day that while he was 'not legally bound' to consider his English publisher in the matter, he knew that Macmillan had their Edition de Luxe in type. The two editions 'would not dash in any way' and '[s]hould Macmillan ask me to refuse the American offer I would expect hirn to say at what date he can bring out the English edition ... [which] has been put off from year to year' (NLI 30248).

XXVII Introduction

One of those purposes, we may be sure, would have been that of checking their own text in standing type against the revisions Yeats would inevitably make in proof for the Scribner's Sons edition. Appre• ciating the 'one-off nature of the venture, out of the normalline of book production for his American market, Yeats evidently thought it a simple matter to consign copy from in-print texts and to correct proof. Through Watt he insisted that Scribner's Sons 'must print from the latest London text ofmy work'. He knew that the New York editions ofhis work were frequently badly misprinted, and so stipulated that the Collected Poems (London: Macmillan, 1933) and Collected Plays (London: Macmillan, 1934) be used. 24 This choice of volumes created problems which he anticipated: both had been designed for a popular market and lacked notes, 25 which were of course desirable in a collected edition, and were to be taken, said Yeats, from Plays in Prose and Verse (London: Macmillan, 1922), which would also if necessary provide the prose version of The Hour-Glass. 1t was left up to Scribner's Sons to decide if they wanted Fighting the Waves. Yeats concluded, 'I must see final proofs' (Callan 91, December 3, 1935)· Ten months later a contract was made between Macmillan (New York) and Charles Scribner's Sons. The delay was caused by Scribner's Sons' uncertainty about the viability of the venture, unless Yeats would contribute prefaces and possibly unpublished materials to make it more attractive. At the time he was unwilling to do so, and put them off throughout the summer of 1936, eventually agreeing to write three prefaces but not to supply any more unpublished writings. 26

24princeton, list accompanying letter from A. P. Watt tojohn Hall Wheelock ofCharles Scribner's Sons, New York,january 28,1937, with Yeats's annotation informing hirn that the American edition of Collected Poems (1933) had 'many misprints'. This point has however been misunderstood by Finneran, (see EYPR 169). 25See Gould, 'The Definitive Edition', pp. 710-14 for an account of the plural arrange• ments ofYeats's poems and plays in the 1930s. 26 As early as December 17, 1935, internaIoffice memos mention details of the edition, which C. B. Merritt of Scribner's Sons wanted to be 'definitive' and to contain new material. Scribner's own doubts led hirn to write to George P. Brett to say that 'the least [Yeats] could do' to help sales would be to write 'three short forewords for the three natural divisions ofthe set-namely, Poems, Plays, and Essays-also some ofhis admirers and friends ... feel certain that he would be able to dig up a little unpublished material that would tend to make the set more definitive as weil as more ofa collector's item. Please do not think for a moment that I am trying to encroach on your preserves as his publisher, and anything in our set you would naturally be perfectly free to use, but it is important for the success ofthe project that we get his cooperation in making it as attractive as possible' (Princeton, December 27, 1935). Yeats agreed to write abrief introduction to the set, but thought further new writing 'unnecessary' (Princeton, Watt to Wheelock, june 9, 1936). Wheelock, claiming he had met Yeats once and putting pressure on hirn via Padraic Colum, pressed Scribner's Sons' case for the prefaces (Princeton, February 24, june 17, 1936), while Yeats put hirn off due to pressure of work (Princeton, Watt to Wheelock,

XXVIlI Introduction

On September 30, Charles Scribner informed George P. Brett that Yeats had agreed to their terms. The contract, which had initially stipulated that the prefaces were not to exceed three pages each, was changed to allow 'whatever prefatory material' Yeats might provide (Callan 93). Scribner wrote to Brett, 'I do not think that any of them will exceed three pages but if the boy should be inspired to write five or six pages in any instance, I do not think you would complain. After all, the material will be yours except for use in this limited edition' (Princeton, October 7, 1936). The contract was signed on October 9 (Princeton, October 15, 1936). In the earliest plans, 750 signed sets were to be for sale, with a further 30 for presentation. The number was later increased, and Yeats signed 850 leaves (Princeton, March 14, 1954). Yeats, 'overwhelmed with work' on The Oxford Book 01 Modern Verse, was not ready to turn his mi nd to the prefaces, submission of copy, or even a title for the set (Princeton, December 3 I, 1936). He was also being forced to think again about the Macmillan Edition de Luxe. As he grew more enthusiastic about the Scribner's Sons' project he obviously wondered if they could not use the Edition de Luxe proofs, instead of his current in-prints and volumes from the I920S Collected Works. Through Watt, he asked Macmillan for up-to-date lists of contents of the Edition de Luxe (B.L. Add. MS. 54903, f. 133; October 23, 1936) which were supplied in duplicate through Watt by October 29 (B.L. Add. MS. 54903, f. 136). Yeats kept one annotated copy of these lists (now NLI 30202), and returned the other to Macmillan via Watt by November 12 (B.L. Add. MS. 55787, f. 362). With the lists, he sent the following question:

Are Messrs, [sie] Macmillan going ahead at once with the de luxe edition? The list of contents I propose to send to Scribner is exactly the same as that which I have sent to Messrs. Macmillan, though slightly different in form [emphasis added]. I shall send it to you in a couple ofdays, this applies to the portraits also. If Messrs. Macmillan were going on with their edition at on ce would it not be possible to make some arrangement both about contents and portraits. Could I not send the sheets or corrected proofs ofthe Macmillan edition to Scribner? If Messrs. Macmillan are not going on at

August 25, 1936). By January 28, 1937, when Yeats did forward lists of contents, he again protested that he did 'not promise apreface for every volume' adding that further prefatory material to AVision might 'look ridiculous'. Wheelock's reply of February 9 requested a General Introduction and three prefaces to Poems, Plays, and Essays, making four in all (Princeton). Yeats however, did not write aseparate preface for his poems and, although Wheelock remained confused about the number ofprefaces he had in fact received, he was finally forced to concede to George P. Brett of the Macmillan Company, New York that Yeats 'never did the fourth' (Princeton, January 16, 1953).

XXIX Introduction

once the process might be reversed. Of course if it is desirable to have the two editions completely different, I could make some rearrangement, though not much. On consideration I will delay sending you my list for Scribner until I get an answer on this point. (BL Add. MS. 54903, f. 148, November 10, 1936).

That slight difference in form 27 and the opened possibility of 'some rearrangement' indicate that Yeats had to the forefront of his mind the plural arrangements of his text. He was well aware that the Edition de Luxe volume of poems was in a chronological arrangement, while he had quickly proposed to send Scribner's Sons his latest in-prints, ineluding Collected Poems. He had rearranged his plays into a chronological series of plays and volume sections on the Macmillan lists, too, and naturally saw the desirability of superseding his provisional instruction to Scrib• ner's Sons to print from Collected Plays.28 So it is elear that had Yeats had his way, the two de luxe editions would have been identical, and would ideally have been edited from London and set by Macmillan's printers (R. & R. Clark, Edinburgh). Thomas Mark would have overseen the proofing and both sets would have used the same portraits. Doubtless, too, an agreement between the two firms to sell in separate markets (similar to that which they eventually coneluded after Yeats's death)29 would have resulted in the use ofthe prefaces in the Macmillan Edition de Luxe also. Scribner's Sons, for the moment, would have taken over the role usually taken by Macmillan (New York). Macmillan (New York) would, of course, have been losers by this arrangement. They had, after all, issued the various volumes of the Collected Edition in the 1920S, ineluding a signed limited edition for the American collectors' market. Nothing had been agreed with Macmillan in London over American publication of the Edition de Luxe, but had it ever come elose to publication, the question would have arisen. Macmil• lan (New York) therefore, also had an investment in delay to allow the Scribner's Sons venture its market: they too, it would seem, were gambling that Yeats would live long enough to permit a later collected edition of their own. In any event, Watt, normally a very discreet man, was to agree with Harold Macmillan that it was simply not in Yeats's

27The remark might also envisage the projected prefaces for the Scribner's Sons edition. 280n these lists and Yeats's instructions to Mark inscribed upon them, see Gould, 'The Definitive Edition', pp. 720-22. 29see below, p. xxxv.

xxx Introduction

interest for both the Macmillan and Scribner's Sons editions to be identical and to appear simultaneously.30 So, wh at is notable about the episode is Yeats's entire openness and practicality, his desire to get on with the job, and his willingness to be advised by his agent and publisher. He was happy if necessary to entertain 'some rearrangement' if it were to be judged on his behalf that to do so would be in his interests. He was prepared to accept that if Macmillan could not 'go on' with their plans they then might receive 'sheets or proofs' from Scribner's Sons, which would of course have embodied a new stage of corrections and revisions. Macmillan's reply was clear and yet rather canny. Again, it makes plain that so far as the firm and Yeats were concerned, there were two arrangements of his poems in play, with the later, that of Collected Poems (1933), still seen as a 'departure' from the former:

On the whole I think it would be judicious for us to hold over our edition for the time being, as we want to make it a very fine piece of work and to see that it has the special features that makes [sie] so much difference to the success ofsuch an edition. In the circumstances, ifMr. Yeats does not mind, we should prefer not to let Scribners have the sheets or corrected proofs of our own edition. For one thing, we want to put the revises of two or three of the volumes in hand, now that we have the further instructions Mr. Yeats has given on our lists, and for another, I think that the more divergence there could be between our edition and Scribners [sie], the better it would be in the interests ofthe former. My own suggestion would be that Scribners should be told to follow the text of our two volumes of Collected Plays and Collected Poems for those works - it is, after all, the latest text in both cases - and that of the Uniform (IO/6d.) Edition for the prose works; but we should not wish to oppose Mr. Yeats' own wishes in this matter. Perhaps you will kindly put these observations before hirn ... (B.L. Add. MS. 55787, ff. 444-45)

Yeats's reply does not appear to have survived, but we know that he concurred with the plan. 31 He could hardly do otherwise, given that he must have been urged by his agent to follow Macmillan's suggestion. As

30Watt told Harold Macmillan 'As you know, I entirely share your opinion that the more divergence between this edition and that of Messrs. Scribner the better it will be' (BL Add. MS. 54903, f. 151, November 16, 1936). 31Because when he did submit contents lists to Scribner's Sons on January 28, 1937 (Princeton) , the directions suggested by Macmillan were largely followed, except with regard to the volumes of plays, for which he took notes, music etc., from the Collected Edition of the 1920S. His compliance has illogically been read as preference: see EYPR 168--69. It is likely he did not reply at all, given that Watt sent no remarks on to Macmillan.

XXXI Introduction he was to write to Watt on a subsequent occasion: 'Macmillan's interest and mine should not dash' (B.L. Add. MSS. 54904, f. 45, July 28; 55798, f. 127, August 4, 1937)· Yeats's lists of contents sent to Watt for despatch to Scribner's Sons on January 28 were essentially pre1iminary, and more detailed thinking was to follow. 32 That for Volume II had been typed up on George Yeats's typewriter from the Macmillan contents list oflate 1936, and Yeats had altered its tide, MYTHOLOGIES AND THE IRISH DRAMATIC MOVEMENT, in his own hand to read MYTHICAL STORIES AND THE IRISH DRAMATIC MOVEMENT (Princeton,January 28,1937). On this list he has also stipulated in his own hand that the volume was 'To be printed from "Early Poems & Stories" and from "Plays and controversies" in edition published by Macmillan & Co, London': as ever, he was vexed by the mi sprints in the American editions ofhis work. Subsequendy, George Yeats despatched the copy for the Dublin Edition from Dublin while Yeats was in London in June. The copy for Volume II was packed off to Watt on June 14, and she did indeed despatch a copy of Early Poems and Stories (London, 1925) for use as copy. With it she sent a new list showing the arrangement ofthe volume. That list now confirmed the earlier (Macmillan) tide, for she typed MYTHOLOGIES & THE IRISH DRAMATIC MOVEMENT. Just why the weaker Mythical Stories had been temporarily entertained by Yeats for the American market is not known. The question of a name for the edition was next settled, with Yeats's two suggestions-'Rathfarnham Edition' and despite its 'thin sound', 'Coole Edition' -overturned by the publishers in favour of their less pertinendy local suggestion of 'Dublin Edition' (Princeton, March 2, 10, 1937). Scribner's Sons were somewhat frustrated by Yeats's de1ays-he was revising the notes and arrangement of the volumes of plays-and J ohn Hall Whee10ck had to restrain himse1f from setting from unrevised copies of a set of Macmillan in-prints he had had his London agent, Charles Kingsley, purchase (Princeton, June 8, 1937). Considerable confusion surrounded the submission of the copy for the Dublin Edition: 33 matters were made no easier by the fact that so many people in so many pi aces were involved: Yeats was away (with Edith Shackleton Heald at The Chantry House, Steyning); George Yeats was typing up lists and tying up parce1s in Rathfarnham; Watt was in London, as was Charles Kingsley, at the Scribner's Sons' London office, where he was confusing, yet exasperatedly trying to reconcile, lists and

32Yeats had always intended to 'arrange the volumes to suit hirnself (Princeton, October 7, 1936), but the wish was overtaken by the advice of Watt and Macmillan. 33See Gould, 'The Definitive Edition', pp. 72 3-5.

XXXIl Introduction parcels; whileJohn Hall Wheelock in New York was awaiting receipt of the material. So it was predictable that a replicated error should have gone unnoticed. It had after all escaped Yeats and Mark on the 1931-32 page proofs and would continue to elude Mark (see below, p. 83). George Yeats in retyping the J anuary 28 list silently passed over one of the dating errors on p. 395 of Early Poems and Stories-that of the revision of the Red Hanrahan stories, (which should be 1904)-in presenting the following as the arrangement of the volume:

VOL 11. MYTHOLOGIES & THE IRISH DRAMATIC MOVEMENT THE CELTIC TWILIGHT 1893 THE SECRET ROSE 1897 STORIES OF RED HANRAHAN 1897 (rewritten 1907) ROSA ALCHEMICA ( THE TABLES OF THE LAW (1897 THE ADORATION OF THE MAGI ( THE IRISH DRAMATIC MOVEMENT 19°1-1919 (Princeton, accompanies letter of June 22, 1937)

The copy of Early Poems and Stories submitted with this list was clean, unrevised copy. 34 It did not even carry over the reinstatement of 'Where there is Nothing, there is God' which Yeats had made when preparing copy for the Macmillan Edition de Luxe. 35 It seems likely that the Dublin Edition Volume 11 is one of the volumes wherein Yeats was adhering to the spirit of Macmillan's wishes. But, as he made no attempt to change any of the Secret Rose or Celtic Twilight materials, it is also probably true that he was not very interested in further rewriting of his prose. He did make three changes to poems in the Collected Poems, and alter a lyric in the plays, as weIl changing some other aspects of them, including their notes. 36 The rearrangement of the volumes of plays apart, this was, for Yeats, a very cursory review of his work prior to submission. But of course, he was also much given to revision in proof, and the Dublin Edition never proceeded to proof. Finally, all copy was in New York by November 8, but by this point Scribner's Sons, ready to proceed to setting, became markedly less enthusiastic about the project and by February 1938 were predicting a

34HRHRC Box 3, Vol. 8, two folders of unbound pages. For a description of the sub• editing done upon them at Scribner's Sons, see beIow, p. xxxix. 35See 'The Page Proofs of 1931-32' beIow, p. xli. 36See Gould, 'The Definitive Edition' pp. 709, 738 n.12 on these changes. See also EYPR passim and Finneran, 'A Note on the Scribner Archive at the Humanities Research Center', loc. eit..

XXXIll Introduction year's delay in publication (Princeton, February 23, 1938). Apart from sending Yeats proofs of the illustrations, no further work was done on the project before his death. By August 29, 1938 publication was not envisaged before the au tu mn of 1939. These delays, of course, meant that the Macmillan Edition de Luxe, planned to appear after the Dublin Edition, was temporarily marooned also, but Macmillan were, as we have seen, also in no particular hurry.

3. The Two Projects after Yeats's Death While Yeats was alive he could overrule suggestions made on his proofs. Without hirn, the editorial team split between London and Dublin had to produce something which was not just the latest in aseries ofincreasingly stable but always provisional 'final' arrangements (which is what Yeats's canon could only ever amount to while he was alive). It had also to be, in words first used by Harold Macmillan in a letter to Mrs Yeats of February 8, 1939, 'complete and definitive' (NLI 30248; B.L. Add. MS. 55819, ff. 18g--g0). The words were also used in a 'Preliminary Notice' headed 'The Collected Edition of the Works of W. B. Yeats' which was released to the trade press before April 4, 1939. 37 Two imperatives shaped what Mrs Yeats, Harold Macmillan and Thomas Mark set out to do. First they were committed to producing a distinctive and handsome set for collectors, to seIl in a market perceived as a very narrow one. Uniformity and finish were important, and Macmillan knew that they had a riyal in the Dublin Edition which, since 1935, had made them even more aware of the importance of 'special features' (including, doubtless, the chronological arrangement of the

37The copy sent to John Hall Wheelock by CharIes Kingsley together with a clipping from The Bookseller (March 30, 1939, 501) is filed at Princeton, April 4, 1939. An early proof of this prospectus, leaving a blank space where the word 'Collected' would appear, was sent to George Yeats on February 28, 1939 (B.L. Add. MS. 55820, ff. 203-5), and returned with her annotations and is now filed at B.L. Add. MS. 55890, f. I. A second state ofthe notice was prepared after Mrs Yeats had settled on the name Coole Edition. It was printed on the ivory paper chosen for the volumes themselves. A fragment of it (the first half of abifolium, printed on both sides) is filed at BL Add. MS. 55890, f. 3. It might actually have been issued. I ts blurb ran: 'THE COOLE EDITIO;,\T / OF / THE WORKS OF / W. B. YEATS / MESSRS. MACMILLAN will have ready for publication this autumn a complete and definitive edition ofthe writings ofWILLIAM BUTLER YEATS, to be known as THE COOLE EDITION. MR. YEATS had corrected and revised the text of his published [emphasis added] works for this purpose, and had autographed the first volume of each set of the edition. The later poems, plays, essays, introductions, and other material which remained unpublished at the time of his lamented death have now been added to the appropriate volumes, and the arrangement is given in detail overleaf' [arrangement page is missing). This version was as honest a statement of the extent of Yeats's involvement as could have been expected in abrief compass. The force of the word 'published' has been quite missed in EYPR 1-2.

XXXIV Introduction volume ofpoems) which could make 'so much difference to the success of such an edition' (B.L. Add. MS. 55787, ff. 444-45). The two firms clearly felt nervous about their respective commercial prospects, and agreed to divide the potential market between them. 38 The second imperative was that the Edition de Luxe had now become a memorial edition, just as Yeats had foreseen. 39 I t was immediately enhanced from seven to eleven volumes (a decision later copied in New York). Charles Scribner's Sons' London officer, Charles Kingsley, at once feared that Macmillan 'had put a fast one over on us'. 40 The Macmillan rearrangement was submitted to George Yeats, but Scribner's Sons did not bother to consult her about their own eleven volume rearrangement. Macmillan were also in quest of'completeness and definitiveness' (as a publisher thought a buyer or collector would und erstand such a term). Yeats had written a lot since the edition had last been in production, and the contract had always called for the inclusion of all his newest work (EYPR 8). But, while both firms and Mrs Yeats made uncoordinated attempts to harmonize the contents of the Coole and Dublin Editions, no• one after Yeats himself sought to make their respective arrangements of contents uniform, and neither set was planned to be the exhaustive re cord of Yeats's writings. Macmillan wanted a canon for a limited edition. Other editions of his works for other markets could follow. Abandoned writings were not to be included. Nor were poems which Yeats had published only in the context ofprose, stories or plays to be dislodged and dragged into the volumes ofpoems, nor were uncollected writings as book editor, except as chosen by Yeats, to stand in his works. The eleventh volume, originally entitled Essays and Reviews by Macmillan, was left vague, with a mention of 'miscellaneous later and unpublished papers', presumably to account for Essays 19JI to I936 (Cuala Press, 1937) and other pieces. The early proof sent to Mrs Yeats was returned by her with the tide changed to Essays and Introductions and she indicated firmly that 'no REVIEWS' would ever be republished in her lifetime (B.L. Add. MS. 54904, f. 171, April 17, 1939). Not that the canon was to be entirely closed by these two limited

38See Princeton, Wheelock/Macmillan letters, March 27, April 8, April 2[, April 28, [939; B.L. Add. MS. 55823, f. 457, May 9, [939. By the terms oftheir agreement, the two firms were to refer buyers for the British market to the Coole Edition and buyers in the American market to the Duhlin Edition. 39In her reply to Harold Macmillan's letter of condolence, George Yeats had remarked that Yeats had 'always said' that Macmillan 'would only bring out [the Edition de Luxe] after his death' (NLI 30248, February [3, [939). For the context ofthis remark see Gould, 'The Definitive Edition', p. 7[0. 4oPrinceton, April 4, [939, a letter accompanying the 'Preliminary Notice' and press cutting. But in fact Macmillan had already sent a courtesy copy ofthe prospectus direct to the :\'ew York office on March 27, [939.

xxxv Introduction

editions. Mrs Yeats and Harold Macmillan thought of adding a twelfth volume-of autobiographical writings-at some future date, and Mrs Yeats certainly thought of extending the idea to Scribner's Sons as weIl (NLI 30248, April I4, I5, 24, I939; BL Add. MS. 54904, f. I 7 I, April I 7, I939). SO neither the Yeats Estate nor the publisher had in mind a Complete Works, and Scribner's Sons were in no position to do anything but proceed with what copy they had in hand while looking nervously at Macmillan's latest plans to see if they themselves had everything they were entitled to. They never did realise, however, what Macmillan had in their volume of Mythologies and The früh Dramatic Movement, and so never caught up with the fact that Yeats had reinstated 'Where there is Nothing, there is God'. Scribner's Sons banked on making their set unique with distinctive plates (apparently unaware that Yeats initially had wanted the plates to be uniform in both sets),41 and with the three prefaces written especially for the Dublin Edition. Later, as we shall see, Scribner's Sons were to be happy to agree to supply Macmillan with the special prefaces once they had been set into proofform, by which time Scribner's Sons would have been commit• ted to publication. Such an arrangement was a natural consequence of the two firms' gentlemen's agreement to seIl in different markets. After Yeats's death, Macmillan had asked Scribner's Sons for proofs (Prince• ton, February 28, I939) when seeking a list of Dublin Edition contents, and were told on March I6 that galleys would be sent when composition had begun. By March 27, having seen a list ofthe Dublin Edition contents and judged their own to be the [uller, the more recent and the better arranged, Macmillan feit that they 'need not trouble' Scribner's Sons for duplicate galleys of volumes they already had (Princeton, March 27, I939). By April 8, Scribner's Sons realised that while they had the prefaces which Macmillan did not have, Macmillan had 'Last Poems' and other vital items of Yeats's late work which they now had to seek from the Yeats Estate. George Yeats was very slow to provide such material, doubtless because she was waiting for Macmillan to get it into prooffor the Coole Edition before supplying Scribner's Sons with corrected copy in that form (Princeton, April2I, 28; May 2, 9,23, I939). By May I2, with the two editions apparently converging, Scribner's Sons offered Macmillan a deal. If Macmillan would take electrotypes of the Dublin Edition, so as to make the two editions uniform and give Scribner's Sons some vital Macmillan assistance on 'various problems' which would appear at proof stage, Macmillan could have the 'General Introduction' and other prefaces. On May I9, Harold Macmillan declined the offer,

flB.L. Add MS. 549°3, f. 148, H. Watt to Harold Macmillan, November 10, 1936.

XXXVl Introduction stating that 'the whole of our Coole Edition is in type and will shortly be passed for press', and nevertheless asked again for proofs of 'the general introduction and the three prefaces' (B.L. Add. MS. 55824, f. 179). Harold Macmillan decided, for the moment, to dispense with the prefaces sight unseen, accepting from Mrs Yeats (and convincing him• self), that they would probably contain some sense of the occasion for which they had been written. This was a hasty decision, as Scribner's Sons had given Macmillan a clear hint of the size and significance of 'A General Introduction for my Work', and Mrs Yeats was prepared to be persuaded to allow Macmillan to use it. 42 (In the end, Coole and Dublin Editions having been abandoned, Macmillan did indeed first publish all three prefaces, in the 1961 Essays and Introductions.) In mid-1939, Mrs Yeats and Thomas Mark began to work intensively on the text and proofs ofwhat was now the Coole Edition, which, with the setting of Yeats's newer work and the vexing questions of what from his unpublished essays was to be included, was a considerable task for the widow, especially as she was moving house and coping with other matters consequent on Yeats's death. One of their concerns was unifor• mity. In reply to her letter of April 17, 193943 Thomas Mark wrote to assure George Yeats that '[ n]o alterations will be made, of course, in words and phrases that have been deliberately repeated, but I will draw your attention to anything that appears purely accidental in the prose. I shall feel much more satisfied if I can refer matters of this kind to you.' (B.L. Add. MS. 55822, ff. 551-Y2, April 21, 1939). In another letter he solicited further assistance: 'I should feel more at ease if you would let me send you my marked proofs of all the volumes before they go to press, as there are always one or two points on which I should like your advice' (B.L. Add. MS. 55825, ff. 2I7-18,june 12,1939). As time went on, Mark naturally became less sure about just what Yeats

42B.L. Add. MS. 55825, Ir. 301-2, JunI' 13, 1939; NLl 30248. Harold Macmillan had heard [rom Scribner's Sons in a letter of May 12 that the 'General Introduction' was of 'about 28 pages' and that there were also 'three [in fact two] prefaees of about seven or eight pages' and that these prefaees were 'of very great value as casting light upon his theory of poetry and his methods of eomposition' (Princeton; B.L. Add MS. 55824, Ir. 586- 8,June 2,1939). He had already enquired ofGeorge Yeats about them on April 21 (B.L. Add. MS. 55822, f. 551), having been alerted to their existenee by Seribner's Sons' letter of8 April (Princeton). OnJune 7, she replied to his letter ofJune 2, saying that the prefaees and 'General Introduetion' 'were written by WBY for the exclusive use of the Seribner edition; I do not know if they would allow Maemillans to use them. Perhaps you or Mr. Watt would ask them? I am not sure that the short prefeaees [sie] are very essential; I think they would hardly "separate" from the partieular purpose for whieh they were written' (Basingstoke). Thomas Mark has written ';-.ro' into the margin whieh suggests that perhaps A. P. Watt & Son wer I' eonsulted and advised against Maemillan's pressing further in the matter. But see below, pp. xxxix-xl. 43Quoted above, p. xxiii.

XXXVII Introduction

had wanted, and his uncertainty was increased by the fact, evidenced by the proofs of Mythologies and The frish Dramatic Movement, that Yeats hirns elf corrected proofs with an eye to a particular correction in a particular context rather than to any principle ofuniformity, even though he wished to be offered the choice of imposing uniformity, if only to decide to reject it. So Mark was to write again to George Yeats:

As regards the Gaelic names, I should be grateful ifyou would check them. I made a kind ofindex ofthem as I went along to see that they were uniform. As the aitch in 'Uladh' and other names had been dropped in the first volume-I think Mr. Yeats took them out-they were alte red in the other volumes, but it is quite simple to correct them. (BL Add. Ms. 55825, ff. 532-3, June 20, 1939)

On June 26, Mark wrote to thank George Yeats far her comments on his list of suggestions and letter of June 22 (NLf 30248). He would make 'Gaelic spellings uniform' when he had 'proofs of the plays' and was 'sending ... to-day the proofs ofVolumes VIII, IX and X, "Mytholo• gies", "Discoveries", and "Essays", with the marked proofs of Volume VIII. The other two volumes were not read by Mr. Yeats.' (B.L. Add. MS. 55826, f. 50). The marked proof of the old Volume 11 of the Edition de Luxe was thus returned at this point and remained in the Yeats family collection until Mr. Michael B. Yeats gave it to the National Library of Ireland in 1986 (MS. 30030). The new proof must have been a marked set ofthe 'revise' ofthe 1931-32 proofs prepared at Mark's request after Yeats had returned them. 44 There is no evidence that Yeats had ever seen the 'revise', and his instruction that 'The volume called "Mythologies" I need not see again' is strong evidence to the contrary. The new proofs of Mythologies and The frish Dramatic Movement were returned by July 1 I, 1939 (BL Add. MS. 55826, f. 436) but George Yeats did not return the marked first proof of Mythologies and The frish Dramatic Movement from 1932. In the late 1950s, when Mark was preparing Mythologies for the press, he did not have Yeats's own proof comments with hirn, and his access to George Yeats and her advice was by then severely restricted. So far as Mythologies and The frish Dramatic Movement was concerned, a major change had been instigated by Thomas Mark at the time of drawing up the 'Preliminary Notice'. Then, while enhancing the Edition de Luxe from seven to eleven volumes, he had moved Mythologies as he now entitled it, from Volume 11 to Volume VIII, substituting Per Amica Silentia Lunae for The frish Dramatic Movement and shifting that work to

44Basingstoke, Printers' and Authors' Records, Vol. 32, f. 91uly 8, 1932. R. & R. Clark's records, formerly in the possession of that firm, show that no further revises were printed. The new proofs have not been located.

XXXVlll Introduction

Volume X, the first of the two volumes provisionally entitled Essays and Reviews. The change was later to be echoed in New York when Scribner decided-at an unknown moment but before March 22, 1940 (Princeton)-to follow Macmillan's 'Preliminary Notice' in enhancing their set to eleven volumes. At that stage, the Charles Scribner's Sons' sub-editor known only from his initials 'B.S.' changed their Mythologies volume from Volume II to Volume VIII and placed The frish Dramatic Movement in Volume X, moving Per Amica Silentia Lunae from Volume V of the old seven volume arrangement to Mythologies. All this was done without reference to Mrs Yeats or A. P. Watt. 'B.S.' worked without any clear idea ofwhat he was doing, but, following the Macmillan 'Prelimin• ary Notice', he was also able, for the most part, to put the two volumes of poems back into the chronological arrangement of the Coole Edition on new lists drawn up at this time (HRHRC, Box 4, Mise. Vol.). The Coole Edition was put 'on ice' by October 19, seven weeks into the war, when Thomas Mark wrote to George Yeats to tell her that it had 'to wait for better times' (B.L. Add. MS. 55830 f. 334), but proof materials were still being prepared for possible despatch to her in January and February 1940 (B.L. Add. MSS. 55830, f. 334; 55833, f. 223; 55834, ff. 522-3). By February 20, 1940, Lovat Dickson of Macmillan had told Charles Kingsley of Scribner's Sons that the Coole Edition had been indefinitely postponed. However, on March 7, 1940, Harold Macmillan was still trying to get proofs ofthe 'Introduction ofthe Collected Edition' from Scribner's Sons while informing them that 'war-time conditions' were preventing his firm from 'proceed[ing] with our original plans for our own Coole edition' (B.L. Uncat., Black Letter File 463, f. 440). A follow-up letter on August 7, 1940 elicited the reply that Macmillan could indeed have the 'General Introduction' in proof but that the Dublin Edition was also postponed and that proofs would not be available 'for many months to come' (Princeton, August 23, 1940). The Cuala Press publication of If I were Four and Twenty in September 1940 prompted Daniel Macmillan to write to Watt seeking the contents for the Coole Edition, but Mrs Yeats did not respond (B.L. Uncat., Black Letter File 469, f. 197, November 22, 1940). By July 29, 1947, Harold Macmillan was prepared to abandon the whole idea of the Coole Edition because of the 'staggeringly high' costs and the absence of a market 'for this kind oflimited edition de-Iuxe' as he wrote to Watt (B.L. Uncat., Letter Book 498, f. 392). He suggested the idea of the two-volume Poems incorporating the signed sheet. Watt agreed with this 'excellent idea' by August 14. By December 1948 both an expanded Collected Poems and the two volume Poems were under way (B.L. Uncat., Letter Book 507, f. 150, December 13, 16, 1948). The set was

XXXIX Introduction

published on November 25, 1949, and sold out that day (B.L. Uncat., 509, f. 567; B 51, f. 114). The Uniform Edition of the prose works got under way more slowly, and it was not until March 12, 1959 that Mythologies appeared in London in an edition of 3000 cop{es, with publication (4500 copies) in the United States to follow on July 28. On sight ofthe Macmillan 'Preliminary Notice' in the spring of 1939, John Hall Wheelock realised that Scribner's Sons did not have 'Last Poems', Purgatory, The Death of Cuchulain and 'On the Boiler' (Prince• ton, Wheelock to Macmillan, April 8, 1939). Mrs Yeats supplied the extra copy for Volume I in time for Watt to forward it to Kingsley on May 23. All the copy for the edition, including 'On the Boiler' was in New York by September 19. By January 22, 1940, Scribner's Sons were beginning to wonder ifthe Coole Edition had appeared after all, but Lovat Dickson's letter of 20 February told them of its indefinite postponement, probably 'until after the war'. For the moment, Macmillan's difficulty seemed to Scribner's Sons to be their own opportunity and they hoped to have their first volume out in November, 1940 (Princeton, Wheelock to Kingsley, March 27, 1940). Scribner's Sons did not formally abandon their project untilJanuary 5, 1953. Thereafter they tried to sell the signed sheets, plates and prefaces to the Macmillan Company of New York for $4000, but without success. Photostats of the prefaces were, however, supplied to Macmillan (New York) by September 15, 1953 and the original typescripts were sold through the Scribner Book Store to Cyril I. Nelson of the E.P. Dutton Co .. The 850 signed leaves were turned over to the Macmillan Company in 1954 for $800. They were eventually used in the limited edition of The Variorum Edition of the Poems of W. B. Yeats edited by Peter Allt and Russell K. Alspach, (New York: The Macmillan Company, 1957) (Princeton, March 12, April 22, 1954). In the end, no definitive edition of the stories was to be published in an Edition de Luxe.

The Page Prooft of 1931-32 The 1931-32 page proofs ofthe Edition de Luxe Volume 11, Mythologies and The frish Dramatic Movement (NLI 30030), which incorporate Yeats's considered and corrected revision ofthe 1925 text, are to be preferred to that text in its unrevised form, even though Yeats submitted the unrevised text of Early Poems and Stories (London, 1925) as copy for the Dublin Edition well after he had carried out the revision. In 1937 he clearly did not have second thoughts about the changes he had made for the 1931-32 proofs. Probably he was trying to ensure that English and American editions would be different, preserving the superiority of the

xl Introduction

Edition de Luxe proofs by following the request of his publisher and the judgment of his agent. He evidently felt no motivation to manufacture divergence by new or further revision ofthe 1925 text for Scribner's Sons. The 193 I -32 page proofs consist of thirty sheets, each of sixteen pages, folded and cut. These are numbered (with the first page of each sheet given after the sheet number) B (Halftitle), C 17, D 33, E 49, F 65, G 81, H 97, I 113, K 129, L 145, M 161, N 177,0 193, P 209, Q 225, R 241, S 257, T 273, U 289, X 305, Y 321, Z 337, 2A 353, 2B 369, 2C 385, 2D 401, 2E 417, 2F 433, 2G 449, 2H 465. These sheets were printed on various dates in late 193 I. The first page of each sheet has the printer's stamp ofR & R Clark, Ltd., usually in the bottom left-hand corner of the page. The first sheet is dated '30 Sept 1931', the last '26 Oct 1931'; all are stamped 'First Proof'. The pages are larger than the pages in the Uniform Edition, and although the type-face is the same, and the number of lines of type on each page is the same, the wh oie has been set with more generous use of space between lines and between words. The texts of the Secret Rose unit are on pages 127-299. It is clear that the Secret Rose and Hanrahan groups, 'Rosa AI• chemica,' 'The Tables ofthe Law', and 'The Adoration ofthe Magi' were set up from a printed copy of the English edition of Early Poems and Stories (1925). Where the English and American editions of 1925 differ, 193 I nearly always has the same reading as the English version. In one of the two exceptions, 'then' for 'them' in 'Red Hanrahan's Curse', line 33, the printer merely corrected a typographical error in the English edition; the other, 'toward' for 'towards' in 'Proud Costello ... " line 288, could have been either a slip or another intentional change by the printer. The volume Stories 01 Red Hanrahan and The Seeret Rose, published in 1927, had virtually the same text as the 1925 English edition, but was an unlikely choice as copy for 193 I as it did not contain 'Rosa Alchemica', 'The Tables of the Law', and 'The Adoration of the Magi'. In three instances ('The Twisting ofthe Rope', lines 195, 196, 202) in which 1927 and the English version of 1925 differ, 193 I shows the same reading as 1927; all ofthese involve only the placement of a comma or period inside rather than outside a final quotation mark, and provide no evidence that 1927 was being 'followed'. The 1931 page proofs included one story, 'Where there is Nothing, there is God', that had not been included in 1927 or in either the English or the American versions of 1925. This story was set up from either the 1908 or the 1913 volume. (See the headnote to the collation of 'Where there is Nothing, there is God'). There is no evidence that Yeats revised the text of the stories before they were set up in page proof in 193 I. The brief section 'Notes' was

xli Introduction

changed substantively because it included a comment on the poetry section ofthe 1925 volume; the 1931 version of'Notes' omitted this and included instead a paragraph on The frish Dramatic Movement (which was of course part of the 1931 proofs bu t was replaced by Per Amica Silentia Lunae in the 1959 volume). However, in the texts of the stories themselves, all differences between the copy text and the uncorrected 1931 proofs are slight, and almost certainly attributable to the printer. Thus in line 71 of'The Tables ofthe Law', the dipthong of'medic:evalism' was separated, and in 'Rosa Alchemica', line 542, aperiod was placed outside a final quotation mark. In at least one instance, the printer corrected a substantive error in his copy, having observed that for reasons ofconsistency 'Aodh' in line 131 of'Where there is Nothing, ... ' should have been 'Olioll'. 45 On several other occasions he introduced new errors, some of them not caught in proofreading. The most serious example ofthis occurred in li ne 160 of'The Tables ofthe Law', in which the change from 'are' to 'and' distorts the syntax. In the same story, line 366, 'faint' was dropped from 'a faint voice' and never restored. When preparing copy for his 1925 collected edition, Yeats had subjected the stories to substantive revision; but by 1931 he was generally satisfied with them, and could leave minor problems involving accidentals and consis• tency to be taken care of in correcting the page proofs. The 1931 page proofs had been read very carefully by Thomas Mark, who made many corrections and queries to Yeats in light blue ink. (Occasionally they are written over pencil, perhaps indicating that he had skimmed the proofs and noted a few points before undertaking full• scale correction.) Yeats's responses and comments are in dark blue or blue-black ink. In addition, there are throughout the proofs a consider• able number of queries and a few unqueried changes in pencil. The half• tide page ofthe proofs bears the note 'Marked by Sutherland', and it may have been this (unidentified) reader who was responsible for the pencilled notes. It is possible that a few of the pencilled notes were also made by Mark; but, while Yeats responded to almost all of Mark's queries in pencil-a fact which may indicate that the latter group, whether by Sutherland, Mark, or someone else, were added after Yeats had returned the proofs with his own responses. 46 It is clear from Yeats's correspondence that he was satisfied with the corrected page proofs and content to leave any further changes in the

4S'Aodh' was a vestigial reading from the earliest version of the story. 46The Old Men ... " line 12, may be an exception, but the cancellation line is not certainly Yeats's. Mrs. Thomas Mark very kindly examined the pencilled notes, but could not say conciusiveiy that any of them had been made by her husband.

xlii Introduction hands of Mark, who could 'complete the revision better than I could'. However, when the Mythologies text was finally published over a quarter ofa century later, in 1959, it contained many changes that Yeats had not specifically authorized, and some actually at variance with principles agreed upon between Yeats and Mark-as weIl as a few typographical errors. The differences between 1932 and 1959 are listed in Appendix I. Some of them resulted from changes suggested after Yeats's death by Mrs. Yeats, others from the fact that when the idea of a collected edition was revived in the 1950s, Mark did not feel bound by the 1932 text and made some decisions different from those he and Yeats had made long before. In other words, what Mark would have sent to press as the 'final' text in the early 1930S was not wh at he did send in the late fifties. Yeats had carefully checked the 1931-2 proofs, which must thus, in their corrected state, be considered the most appropriate choice for the primary basic text for the Variorum Edition. (See also pp. 3-4·)

The Collations Certain purely typographical features have not been treated as variants. In this edition single quotation marks are basic and double quotation marks are used for quotation-within-quotation. All readings from the printed texts are regularized to this form. Capitalization of the entire first word or the first li ne of a story (common in the periodical versions) is disregarded; as is the non-indentation of the first line of the opening paragraph. Apart from the purely typographical features we have recorded all instances in which the earlier printings differ from the basic text, induding apparently trivial differences of spelling and punctuation. Yeats permitted others to correct spelling and punctuation for hirn but he was far from indifferent to the final result. The 1931-32 page proofs for Mythologies show that Yeats paid dose attention to the suggestions made by his editor. The following editorial conventions are used in recording the variants: I. The lines of each story have been numbered; the lineation in this edition does not correspond to previous printings of the basic texts. 2. In addition to line numbers, contextual key-words are used to locate the variant readings in relation to the basic text. Thus, in the entry for lines 25-26 of 'The Crucifixion ... "

had often looked upon a cross ere now. But presently the fit passed, and he hurried on. He . .. NO.

xliii Introduction

the words 'had' and 'He' appear in the basic text as weIl as in the National Observer text and are the contextual key-words. The matter between these words is the variant reading appearing in the National Observer. The line numbers refer to the contextual key-words. Where ambiguity would otherwise be possible, more than one key-word on either side of the variant may be given. 3. Contextual key-words are used also to relate variants in punctua• tion to the basic text. Thus in 'The Crucifixion ... " lines 38-39, the two earliest printings place a comma between 'turf and 'that' and the variant is entered as:

turf, that . .. NO, 18g7.

The two contextual key-words fall in different lines in the basic text; hence the reference to lines 38 and 39. 4. In the case ofvariant spellings, only the variant word is entered in the collation, thus:

blessed . .. NO to Ig08.

Here ('The Crucifixion ... " line I06), the word appears as 'Blessed' in the basic text but with the lower case initial in the printings up to and including that of 1908. Several stories were subject to the consistent revision of spellings in one or more printings, and in these cases the variant spellings, keyed to the printings, may be listed at the beginning of the collations. Variant spellings appearing in these lists are not specifically recorded again in the line-by-line collation, though every entry reproduces the spelling actually found in it. 5. Contextual key-words are printed in italics to distinguish them from variant words, which are printed in roman type. This procedure serves also to distinguish verbal variants from those involving punctuation only (as in the contrasting examples given in 2 and 3 above). No effort has been made to differentiate by typeface between variant and nonvariant punctuation, which may be either in roman or in italic, depending upon the typeface used in the contiguous words. For variants involving only spelling, the type used is italic as for contextual key-words. An asterisk (*) following a variant printed entirely in italics indicates that the variant involves the addition of one or more words in the basic text. 6. The suspension points ( ... ) preceding the abbreviated identifica• tion of the printing(s) in which the variant occurs indicate that the text now becomes identical with the basic text and continues so until the next recorded variant.

xliv Introduction

7a. In general, each entry gives all variants from a particular portion of the basic text: the less frequent form of the variant (or, in cases of equal frequency, the later form) is contained in square brackets within the more common (or earlier) form. Occasionally the resultant entry would be too complex, and separate entries are given. The combined form is illustrated in the following example:

leaves, had soon a leaping blaze that cast its flickering light over the knight [cast a flicker over the face of the knight 1897] and over the piled-up heads of the wood-thieves. Then . .. NO, 1897.

In the National Observer the passage appeared as above, except the matter within square brackets; in 1897 (The Seeret Rose), the 'cast ... knight' section, only, was changed in the manner indicated by the reading within the square brackets. After 1897 the passage was reduced to the form in which it appears in the basic text:

leaves, made a very good blaze. Then ...

This example from 'Out of the Rose' (line ro8) is typical of the many entries in the collations that show a passage being at first slightly, and then more extremely, revised by Yeats. b. Where the variant of a variant is punctuational, the tilde (-) is substituted for the word preceding the variant punctuation mark, and is followed by the punctuation mark or, where the variant is absence of punctuation, by a caret (1\); as in 'Out ofthe Rose', li ne 154:

shining out upon the world to keep it alive, [~I\ NO] with a less clear lustre ... NO to 1913.

In this instance, the National Observer printing omitted the comma after 'alive' but was in other respects identical (in this portion ofthe text) with all printings up to 1913. In 1914 a further slight change occurred and, not to make the single entry too complex, this further variant is added after a semicolon, so that the full entry reads:

shining out upon the world to keep it alive, [~A NO] with a less clear lustre . .. NO to 1913; shining out upon the world to keep the world alive, with a less clear lustre . .. 1914.

In keeping with this example, the variants from a particular portion of the basic text are either compressed or arranged in chronological sequence as darity demands and permits. c. Where variants are related to each other but do not begin or end at

xlv Introduction

precisely the same point (in relation to the basic text) contextual words may be added to one variant in order to highlight the relationship. 8. In the example under 7b above, 'NO to 1913' indicates that the particular variant is found in the two printings so indicated and in all the printings falling chronologically between them. All variants which occur in an unbroken sequence of three or more printings are identified in this way. A list of all printings is given at the head of the collation for each story. 9. Word division is treated as folIows: a. When a compound word in the basic text has appeared in a different form (i.e. as two words, hyphenated, or unhyphenated) in the earlier printings, a variant is recorded; as also are instances in which two words in the basic text have appeared hyphenated or joined in earlier printings. Thus:

birthplace 1925, 1927; birth-place NO to 1914. In this example ('The Crucifixion ... " li ne 7), the variant is keyed to the basic text by repetition of the form used in the basic text; this is followed by a list of all the printings (in chronological order) which use the same form as the basic text; then, following a semicolon, the variant forms are given in chronological order. In the present example, 'birthplace' appears in the basic text and was first printed in 1925. The form 'birth-place' appeared first in the National Observer and was repeated in all printings up to and including that of 1914. b. Where, in one or more printings, word division occurs at the end of a line a slash indicates the feature, thus (from 'The Old Men of the Twilight,' lines 23-24):

shot-gun WS, 1897 to 1913, 1925, 1927; shotgun CB; shot-I gun 1914. c. Where a word is hyphenated at li ne-end in one or more printings and is consistently joined or consistently hyphenated within the line in all other printings and is thus (and in itself) unambiguous, no variant is recorded. (See also Appendix 2.) d. In the following example there are variant forms of word division for which there are no corresponding forms in the basic text:

it, the Rose shone a deep blood colour in the fire-light, and, ever as he spake the lad ... NO; it, the Rose shone a deep blood-colour in the fire- / light [fire-light 1897; firelight 1914], and the lad . .. 1897 to 1914.

These variants correspond to apart of the basic text ('Out of the Rose', line 140) wh ich reads simply 'it, the lad', all reference to rose and fire

xlvi Introduction having been omitted. Thus 'blood-colour' and 'fire-light' are not speci• fically relatable to the basic text. There is, however, a relation between the variants themselves. Unremarkably, 'blood-colour' is hyphenated in 1897 and for as long thereafter as the word was retained at all. In the National Observer it had been printed as two words. The word 'fire-light' has likewise no existence in the basic text but a more interesting one amongst the variant versions. The word appeared within the line and hyphenated in 1897; subsequently (in 1908) it was hyphenated at line• end and continued thus until, in 1914, it appeared as an unhyphenated word. Thereafter the word disappeared along with the passage ofwhich it was apart. As in this example, so throughout the collations variations in word division are presented (where there is no corresponding element in the basic text) in the regular forms for the sequential or combined entries of variants. e. Because the basic texts have been newly set for this edition, there are variations in line-end hyphenation between the 1932 and 1897 texts and the text of the Variorum. The original readings can be reconstructed as follows: if a word hyphenated at line-end in the original of the basic text does not fall at line-end in this edition, (I) it will be hyphenated and will be listed in Appendix 2 (words hyphenated at line-end in the originals of the basic texts and also hyphenated in earlier printings when they appeared within the line); or (2) it will be joined, its proper form as determined by consistency in earlier printings; or (3) it will be hyphenated in the text but printed with a slash in the collations (e.g., 'Glen- / Gar') and its full his tory included as a variant. If a word not hyphenated at line-end in the original of the basic text is hyphenated at line-end in this edition, its normal form is joined. In a few instances, words hyphenated within the line in the originals of the basic texts fall at line-end in this edition; such instances are indicated by a hyphen symbol (=).

xlvii