University of Virginia Library

I have now examined all known copies and separate proofs of Jerusalem, as listed below and with the exceptions noted, and have made photographic collation of the difficult plates. My debt to all the collectors, curators, photographers, and Blake scholars who have assisted in this collation is infinite and my gratitude is infinite. As also to the American Council of Learned Societies for a grant in aid. For permission to publish photographs of their copies of Jerusalem I wish to thank the Trustees of the Pierpont Morgan Library, Mr. Charles J. Rosenbloom, and Mr. Lessing J. Rosenwald and the Library of Congress. It seems advisable in the present report to refer to copies of Jerusalem by their present locations rather than by the letters employed, with some ambiguity, in Geoffrey Keynes and Edwin Wolf 2nd, William Blake's Illuminated Books: A Census (1953). Though valuable in much of its detail, this revision of earlier lists by Keynes in his Bibliography of 1921 and Blake Studies of 1949 does not always follow, in its notes on individual plates, the new enumeration tabulated on pp. 113-114. Moreover it gives a mistaken foliation for one of the copies of Jerusalem. (For a corrected List of Copies, see below.)

Professor Karl Kiralis long ago called my attention to the presence of variant numbers on some of the plates and to discrepancies in descriptions of the order of plates in different copies. Throughout the preparation of this paper, he has been unstinting of his critical assistance.

The Complete Writings of William Blake with all the variant readings, ed. Geoffrey Keynes (Nonesuch Press, 1957), p. 621. (Herein cited simply as Keynes.)

A strong magnifying glass helps to keep the work in focus. Enlargement by three diameters sometimes facilitates progress; further magnification proves of no service. Photographs are not appreciably better than photostats for the kind of deletions here examined, but for the recovery of erased or palimpsest mss. photography is more helpful, infrared photography most helpful. High contrast paper is less satisfactory than the regular glossy photostatic paper in which shadings and shadows of letters are retained. Only such paper captured the ghost letters on Plate 95.

A. C. Swinburne, Essay on Blake (1866), p. 284; S. Foster Damon, William Blake: His Philosophy and Symbols (1924), p. 434. The British Museum copy has some crumbs of letters throughout Plate 3 but of a most discouraging faintness.

These methods are also applicable to handwriting, with some adjustment. The uniformities of shape and spacing are less precise but still present, even in such cramped inscriptions as Blake's marginalia. Photographic enlargement plus magnifying glass bring many an apparently lost passage up to legibility. An appendix of recovered deletions, in the prose as well as the poetry, will be given in the Blake Concordance which is under preparation for the concordance series of Cornell University Press.

The exact particulars of Blake's methods of printing are unknown, but the hypotheses of Ruthven Todd, tested experimentally in the workshop of Stanley William Hayter, supply a convincing reconstruction. (See Ruthven Todd, "The Techniques of William Blake's Illuminated P[r]inting," Print, VI (1948), No. 1, pp. 53-65; the title is misprinted "Painting".) One fragment of copper alone survives from all Blake's Illuminated Printing, a piece cut from a discarded plate for page 5 of America, now in the Rosenwald Collection. The plate and a proof of it are shown in Todd's article, p. 57, but captions referring to the "worn condition of the plate" are quite mistaken. No plates of Blake's were so much used as to become worn; moreover this plate was discarded before use: the blurring of letters in a print made from it is due to the imperfect original etching, the varnish or ground having oozed between the strokes of the letters in such a way that the acid did not cut a clear outline.

Professor John Grant, in correspondence, points out that "28 chapters is 'the same' as 4 or 100 in symbolic arithmetic: i.e. it equals apocalypse."

No known copy of Milton was printed before 1808, no known copy of Jerusalem before 1818-1820 (according to watermark evidence). The title-page dates mean that Blake had at least planned these two epics by 1804. Just possibly he etched the title pages at that time; yet none of the pages of text (including the Jerusalem frontispiece) can have been inscribed or etched before the middle of 1805 (see note below on Dating of Blake's Script). He may have written Milton by this time; in letters of April and July, 1803, he declares that he has in the previous "three years composed an immense number of verses" descriptive of his "Spiritual Acts" of those three years — an account which fits Milton better than Vala or Jerusalem — and that these verses are already "perfectly completed into a Grand Poem" waiting only to "be progressively Printed & Ornamented with Prints & given to the Public." Possibly the title pages of "1804" mark a division of the grand poem into two. Biographical and historical allusions in Jerusalem, however, range in date from 1804 to 1814 or 1815 (the end of the war) and suggest a much later date of composition for most of that poem. By the end of 1805 Blake was caught up in Cromek's schemes, to the neglect of his epics, and by the time he and Cromek separated Blake's attention must have been absorbed by preparations for the extremely ambitious Exhibition of 1809. In that exhibition's Descriptive Catalogue (of which the printing had begun by 19 Dec. 1808; see below) he announces that he has "written" and will, if encouraged (the language is close to that of his 1803 letters), "publish" a "voluminous" work that sounds more like Jerusalem than Milton. He had evidently not yet etched or "Printed & Ornamented" many of the plates, if any, or his announcement would be of copies for sale. At this point there is a very important document in which Blake asserts that he has not done any Illuminated Printing for a long time and is not free to do any in the foreseeable future. George Cumberland had written to Blake on 18 Dec. 1808 (Keynes, Letters, p. 171) with news that he had a customer for "a compleat set of all you have published in the way of Books" of colored "etchings" (i.e. in Illuminated Printing). The significance of Blake's reply, 19 Dec. (Keynes p. 865) has been overlooked because of a mistranscription of "printing" as "painting" in the first sentence: "I am very much obliged by your kind ardour in my cause, & should immediately Engage in reviving my former pursuits of printing if I had not now so long been turned out of the old channel into a new one, that it is impossible for me to return to it without destroying my present course." (B.M. Add. MSS 36501, f. 314) He goes on to indicate that preparations for his coming Exhibition "preclude all possibility of promising any thing." His time is taken up with the (conventionally) printed Descriptive Catalogue, which he calls "an account of my various Inventions in Art". (Some mystery has been made about this reference, from taking "Inventions" in a wrong sense, but Blake is using the very language of his Catalogue title: ". . . Poetical and Historical Inventions, painted by WILLIAM BLAKE. . . ".) And his time "in future must alone be devoted to Designing & Painting"; there were 16 ambitious painted Inventions which must be ready by the opening of the Exhibit in May, 1809, and he was also painting for Thomas Butts. The printing of the Catalogue had already begun, he said. It is difficult to know whether Blake's "so long" should be measured in months or years. He cannot have meant that he had done no Illuminated Printing since the Prophecies dated "1795", however, for he had more recently told Cumberland something about printing "60 plates of a new prophecy"—according to a memorandum in a Cumberland notebook of 1807, as reported to me by Professor G. E. Bentley, Jr. Bentley points out that the news may not have been very recent, since Cumberland at the same time refers to Blake's designs for Gray as new, though they existed perhaps as early as September, 1800, and certainly before September, 1805. (See Bentley's "Blake's Engravings and his Friendship with Flaxman," Studies in Bibliography, XII, 1959, 184.) Evidently the "voluminous" work announced in the Catalogue (see above) had been carried as far as 60 plates, in plan at least, before Blake's complete absorption in the Cromekian schemes. Can 60 plates have been actually etched and printed from by 1806 or 1807? The only possible relics, if so, would be the three Jerusalem proofs on 1802 paper; yet Blake's own later allusions will hardly bear this construction. At any event, when, after the failure of the Exhibition, he did return to his old channel, he must first have turned his attention to the etching and printing of Milton, for it is to this work that he alludes in his "Public Address" (paragraphs written in his notebook not earlier than 1810) as "a Poem . . . which I will soon Publish" (Keynes, p. 592). This implies a more advanced stage of production than the Catalogue allusion to Jerusalem. Milton is a more modest undertaking — the plates being fewer than half the number and only half the size of those for Jerusalem — and several things including the date of paper indicate that he finished Milton first; even the latest copies use paper with no later date than 1815. By that time Blake may have begun etching his plates for Jerusalem; he began printing in 1818; by "30 Decembr 1819" he had delivered to Linnell (and received payment for) "Jerusalem Chap 2" of what is now the Rinder copy; presumably he had earlier delivered Chapter 1: both are on 1818 paper; the next two chapters use some 1820 paper; this dated receipt forces us to assume that the paper was used close to the time of its dates (Letters, p. 183). The British Museum copy was evidently printed more or less simultaneously, for it too was begun on 1818 paper and completed on 1819 and 1820 paper. The first two chapters use only 1818 paper, but the third and fourth are on a combination of papers. The distribution, kindly checked for me by Professor E. E. Bostetter, is: Chap. 1—ff. 2, 3, 5, 9, 21, 22, 25 (1818); Chap. 2—ff. 27, 32, 38, 41, 44, 47, 48 (1818); Chap. 3—ff. 53, 62, 66, 72 (1818), ff. 58, 69, 71, 74 (1820); Chap. 4—ff. 78, 88 (1818), f. 76 (1819), ff. 82, 89 (1820). The Cunliffe copy (Chap. 1 only) must have been printed before the later chapters of any copies (whichever state Plate 20 may be), since it is on 1818 paper. It may or may not be an indication of the work's nearing completion that Thomas Waineright was able to describe Jerusalem in the September 1820 London Magazine as "an eighty-eight pounder" now "casting" and about ready to "fire off." Yet the fact that the next two copies are also on 1820 paper suggests that they were begun in or soon after that year. This leaves 9 or 10 years after 1809 for revising and rearranging both before and during what must have been a protracted period of etching. Things pointing to a long period of etching include the deletions and changes dealt with in this article, the trial arrangements of etched plates noted below in the discussion of the Numbering of the Plates, the great variety in such details as thickness of varnish, steadiness of line, handling of paragraph indention, quality of etching and engraving, and other matters hardly to be attributed to symbolic or aesthetic intention. The only other Illuminated work at all approaching this amount of technical variation is Songs of Experience, a work which was a minimum of three years in the etching (to judge from the etched but rejected "A Divine Image", a plate inscribed in 1790-91 according to the evidence of the script).

Printed on blank side of a proof of the title page of Europe; first described in Blake Studies, pp. 110-113. But Sir Geoffrey may be mistaken in deducing that line 8 is "erased after the printing of the plate, leaving a white narrow space," for the same erasure appears in posthumous copies, where one can see that the space is white because the erasure, made on the plate before printing, was gouged too deep to take the ink. Yet Blake may have made his deletion doubly sure by erasing the paper as well. Deletion in the plate plus erasure of the paper is found in some copies of the Book of Urizen.

[A Commentary on] William Blake's Jerusalem (William Blake Trust, 1953), p. 103. (Cited without page when the reference can be found by plate number.)

A conjectural reading of Milton, Plate 10, line 6, severely deleted, would be: "Thus Space ?becomes ?Serpent-Formd & the Womb ?Englobes". The queried words are quite uncertain, but context requires the line to deal with "the nature of a Female Space".

For the monumental plans Blake was interested in, see his letters of March and April. Of the four known copies of Milton, it is the two earlier that read "in 2 Books" and the two later "in 12", the reverse of what we might have expected. Nor can we suppose that Blake merely neglected in his later copies to cover the first digit. In copy C (in The New York Public Library) his brush and spatter work on the title page brings the "1" of "12" into prominent relief in a way that cannot be accidental. Even as he puts finishing touches on a copy in 2 books, he reaffirms his original goal of 12. Yet "Finis" on the last page of Chapter 2 is part of the original etching.

Wicksteed (p. 113) notes that the words "Sheep" and "Goats" are "engraved directly on the copper . . . a device very rarely resorted to . . . ." Failing to observe that it was technically necessary for Blake to resort to engraving when making an addition after etching, Wicksteed takes the rarity as celestial and finds that these words "suggest a Last Judgment scene with the Saviour seated above the Title but unseen."

I hesitate, however, to press the theory that Blake came to feel about Jerusalem the way Coleridge told Mrs. Barbauld he felt about the Ancient Mariner.

Keynes (p. 635) reads it "end of chap. I" and so do D. J. Sloss and J. P. R. Wallis, eds., The Prophetic Writings of William Blake (1926) I, 469.

The facial expressions vary somewhat with inking; in the Morgan copy the woman's face is whitened slightly with a thin grey wash. In the Mellon copy ringlets of brown hair (of about the same degree of curl as that on the woman's head) have been added between the bodies; a different, more golden tint has been given to the tassel-like inflorescence on either side — which Damon identifies as a symbolic "golden net."

When two plate numbers are given, the first is that of the "standard" order (Harvard and Mellon copies), the second, in brackets, is that of the order found in the other three copies.

The phrase is not "Albion coverd", the last words on Plate 30[44], as might have been the case were that plate an insertion. But some variant it may be: "Albion hid" looks feasible — i.e. "hid his heaven with clouds", a meaning that would fit the top of Plate 31[45].

Sloss and Wallis (1, 509) observe: "Page 35 interrupts the sequence of pp. 34 and 36. It is either interpolated or misplaced: no better place can be suggested." Late insertion in a given position does not, however, need to mean late composition.

Meaning engraved instead of etched? If we call Blake's etched plate a stereotype (as he himself did in the colophon to The Ghost of Abel), we ought to apply the same term to his etched and engraved plate, still a single unit of metal, one of his "types".

In the first part of the line he was trying to simulate his usual relief etching (e.g. the rest of the plate). He may have transferred his line of text to the copper in somewhat the usual way (though not in varnish), but he then had to cut it out unusually.

The ascenders added to the copper to make "pa" into "bl" have not been erased, but the letters "pal" have been restored by pen and ink. The restoration does not, however, obscure the evidence of the earlier mending of the plate from "pale" to "blue". Wicksteed (citing Max Ployman's note in the Everyman edition, p. 198) tells only half the story-and permits the mistaken supposition of a variant plate-when he describes this copy as simply "altered in ink" from "blue" to "pale".

Karl Kiralis notes that in the Mellon copy the text of this page is washed in blue (I would call it slate grey), suggests that blue may represent the creative spirit, and wonders whether Albion's soul may be leaving his body, according to a 17th-century supposition, through his feet.

This page in the Morgan copy is labeled "from another copy" and is on a narrower and shorter leaf (6 mm. narrower, and sloping from 5 to 10 mm. shorter, than the adjacent leaves). It is foliated, by Blake, "36". Does this mean that there was still another copy of Jerusalem, in the non-standard order and of the same late vintage as the Morgan copy? (The watermark of the page is 1826.) Or did this leaf only return to the copy from which it had accidentally been separated?

See Blake: Prophet Against Empire (1954), pp. 440-444. Wicksteed (pp. 178-179, 184) on grounds I cannot fathom supposes an allusion in this plate to Blake's own Job illustrations of 1825 — forgetting for the moment that the plate is in all copies, let alone the 1802 proof. Yet at another moment (p. 178) he speaks of the latest plates as made by 1818, recalling the probable date Jerusalem began to be printed.

Professor Harold Bloom suggests that the inversions in Los's speech are meant as satiric mockery of the speech of the Daughters of Albion — i.e. that the difference is functional not chronological.

Keynes, pp. 778, 785. This may be a good place to put forward a revised reading of Blake's note on his 7th design in his Illustrations to Dante (Fogg Museum; collotype in Albert S. Roe, Blake's Illustrations to the Divine Comedy, 1953). A misreading of "Memory" as "Mistress" and failure to read the word "Imagination" distorts the emphasis in present readings. Blake first wrote: "Every thing in Dantes Comedia shews That for Tyrannical / Purposes he has made This World the Foundation/ of All & the Goddess Nature & not the Holy Ghost / as Poor Churchill / said Nature thou art / my Goddess". In Dante, that is, the Goddess of all is Nature rather than the Holy Ghost. Blake then put carets after "Nature" and "not" and wrote above the line the words "Memory" and "Imagination", to replace or explain "Nature" and "the Holy Ghost" respectively. Dante, he means, made the Goddess of all Memory rather than Imagination. But then he shifted the syntax, writing below the line the word "Nature" at the first caret and "is his Inspirer" at the second caret — but then crossing out "Nature" because the word already stood in his original line. If we follow carets and guide-lines for the final reading, we must stop after "of All" and make a new clause thus: "& the Goddess Nature, Memory, is his Inspirer, and not Imagination, the Holy Ghost." (The Churchill clause, written alongside, may not have been added until this revision was completed.)

The Census ambiguously reports that the names "in incised lettering can be seen in copies A, B, and E" and that "fruits are hanging on the branches" in D and E. But in the Census code, copy B is the Cunliffe copy, which stops at the 25th plate. The code must be that of the 1921 Bibliography, where A is the Rinder, B the British Museum, D the Stirling (Mellon), and E the Morgan. Yet the statements are inexact for either code. In the 1921 code it should be said that both names appear in copy B, "Albion" only in A and E, the fruits in B, D, and E. In the Census code that both names appear in A, "Albion" only in C and F, the fruits in A, E, F.

Keynes, p. 726, gives correct readings of lines 48 and 67; he now has verified my readings of the other lines, from the Fitzwilliam copy.

A gap in the center of "weeping" in line 48 in the Rinder facsimile suggests that some damage was done there too, though the word prints fairly well in other copies.

Says Wicksteed (p. 228): "The critical significance of the story is revealed in two broken lines (47-8) and (66-7). In each case something has been deleted by Blake from his original text, but the underlying meaning is only emphasized by this breach of continuity." Which is perhaps only to say that we can follow the continuity despite the lacunae.

Wicksteed (p. 232) in a glancing phrase calls the deletion "a purposeful break" but gives no indication of what he means.

One curious detail is that the horizontal wavy lines in the hair of young Albion were drawn along the lines of text, perhaps to make certain that the words were concealed.

This commercial broadside must have been produced in 1797 or early 1798. Keynes hazards a date of 1790, but in London directories the firm name that fits the advertisement, "Moore & Co. Carpet-warehouse, 45 Chiswell St", occurs only in 1797 and 1798. There are no later listings. Earlier variants, none of which would quite fit the engraved wording, are: "Moore, Foskett and Foskett, Hosier and Carpet-maker, Chiswell St" (1792-96), "Moore & Fosketts, Hosier and Carpet-maker Chiswell St" (1789-91), and "Thos. Moore Hosier and Carpet-maker, 63 Chiswell St" or simply "Chiswell St" (1766-1788). We may guess that Blake's acquaintance with Moore & Co, suppliers of hosiery and carpets to "Shopkeepers", came through his family's hosiery shop. From 1754 to 1765 the directory listing is simply "Thos. Moore Hosier, 63 Chiswell St." without the carpet-making. I wish to thank Mr. Arthur Hudd, of the British Museum staff, for this directory-searching.

Keynes gives the dimensions as "Engraved surface 26.5 x 24 cm. Plate-mark 35.5 x 27.3 cm.", but the sides of this print are trimmed edges not plate-marks. From the evidence of J 96 we can deduce that the original plate was at least 32 cm. wide, the cut being made considerably short of the center of the engraved portion. One more plate of Jerusalem size could have been cut from the remaining top portion, potentially supplying two pages, but no other page shows traces of the carpet advertisement. One of the pages measuring around 15 cm. in width could have been made from the back of such a cut, but without knowing its height we cannot guess which. Blake would, of course, use the backs first. It seems likely that J 64, identical in dimensions to J 96 and bearing a plate-maker's stamp which shows it to be the back of a plate, was etched on the same piece of copper. Cut to match other Jerusalem plates, this we might expect to be neither very different from the others in shape nor precisely similar to plates cut in a batch. The measurements given below support this expectation (see "Sizes of Plates").

Professor John Grant, in correspondence, suggests that "Blake made the twofold mistake because he forgot that he was no longer fighting with the scientists." Sloss and Wallis (I, 635) read "Sexual Twofold" and find it "inexplicable." Wicksteed (p. 247) says that Blake first etched "Threefold" then mended "with infinite trouble" to the reading "Twofold" and then, anticipating such criticism as that of Sloss and Wallis, "contrived to change the word back to 'Threefold'" in the Mellon copy. The only contrivance was to leave the print untouched!

Only very slight beveling is found on the uncut edge of the America fragment (see note 6 above) and on the edges of the plates for the Dante illustrations (Rosenwald Collection). Moreover the plate outlines that show especially in the heavily impressed posthumous copies indicate very little rounding of the edges of the Jerusalem plates. From examination of these materials I judge that the difference between the front and the back of a Jerusalem plate would seldom amount to as much as 1 mm.