Volume 16 · Issue 4

Two Forged Plates in America Copy B

BY THOMAS V. LANGE

For nearly a century, copy B of William Blake’s America, now in the Pierpont Morgan Library, has been considered by dealers, auction houses, and scholars as a complete and uncolored copy with the full complement of eighteen plates on eighteen leaves. Detailed examination of this copy has revealed that two plates are, in fact, facsimiles, and were inserted in copy B between 1874 and 1878, most probably with intent to deceive.

In his Blake Books,¹↤ ¹ G.E. Bentley, Jr., Blake Books (Oxford: Clarendon Press, 1977), pp. 100 et seq. G.E. Bentley, Jr. discusses the provenance of copy B, which was presented to the Morgan by Mrs. Landon K. Thorne in 1973.²↤ ² Morgan Library accession number 63938. The earliest evidence of ownership of this copy is a penned inscription on the recto of the frontispiece (mistakenly located by Bentley on the verso of the title-page), reading, “From the author / to C H Tatham Oct.^r 7 / 1799.” While this inscription was once thought to be in Blake’s hand, later authorities have dismissed this claim. Bentley suggests that copy B might be the one listed by Lowndes as having been sold by “Sotheby, 1855, 2£.7s,” described only as a folio with 18 designs.³↤ ³ W.T. Lowndes, The Bibliographer’s Manual. . ., (London, 1856), I, 215. There appears to be no justification for this suggestion. This otherwise unidentified sale at Sotheby’s has not been traced. Evidently copy B was again sold for £18 in 1874 (according to the note in the 1878 sale catalogue reproduced below), although no trace of this transaction at auction or in dealer catalogues has been found.

The first printed reference to copy B appears in the sale catalogue “of a very choice library and a small but rich collection of ancient engravings & modern drawings” belonging to Albert George Dew-Smith, sold at auction by Sotheby’s on 29-30 January 1878, as lot 247: ↤ ⁴ Reprinted with the kind permission of Sotheby, Ltd.

247 Blake (W.) America, a Prophecy. Engraved throughout by this extraordinary artist. Excessively Rare, presentation copy with author’s autograph inscription, splendidly bound in citron morocco, ornamented with variegated leathers and gold tooling, g.e. by F. Bedford; two leaves said to be wanting, but Blake’s original prospectus says—“America, a Prophecy in illuminated printing,” folio, with 18 designs, Lambeth, W. Blake, 1793
^*_*^* This copy, unbound, sold for £18 in 1874.⁴

According to the notes in the Sotheby master catalogue deposited at the British Library, lot 247 was sold to the London dealer John Pearson for £16.5.0. The next traced appearance of copy B occured twelve years later, on 23 April 1890, when it was again offered by Sotheby’s, as lot 189 in the sale of the library of Thomas Gaisford,⁵↤ ⁵ Thomas Gaisford, 1799-1855, Dean of Christ Church, Oxford. See De Ricci, English Collectors of Books and Manuscripts (Cambridge: Cambridge University Press, 1930), p. 165. who had presumably acquired the book from Pearson. At the Gaisford sale it fetched £61 to the London firm of Bernard Quaritch, which added its customary pencilled note to the rear flyleaf, “Collated and perfect/ Ap.¹ 24 90/ J.T.”⁶↤ ⁶ In a letter of 10 January 1978, Mr. Nicholas Poole-Wilson of Bernard Quaritch, Ltd. kindly informed me that there were two employees of the firm at the time with the initials J.T.; it has proven impossible to differentiate between them: J. Tuckett and J. Thorowgood. Quaritch included copy B in catalogue 104, May 1890 (no. 460), from which it was purchased for £68 by Bernard Buchanan Macgeorge. Interestingly enough, Macgeorge had attempted to purchase copy B at the Gaisford sale and had given his bid of £30 to Quaritch; it was not until the appearance of Quaritch’s catalogue 104 the following month that the collector was persuaded to pay more than twice his earlier bid.

Let us return to the Dew-Smith sale catalogue, which contains the first printed notice of copy B. There is no doubt that the copy under consideration is the one described in the catalogue: the unusual inscription and the readily-identifiable binding are both unique to this copy. The Dew-Smith sale entry enables us to supply an approximate date for the binding of copy B: it must have been done between 1874 when sold “unbound,” and 1878 when the elaborate binding of “citron morocco, ornamented with variegated leathers and gilt tooling” by Francis Bedford was so carefully described in the Dew-Smith catalogue.

There is a glaring ambiguity in that catalogue entry: it is not at all clear how many plates were in copy B. As printed, the catalogue entry reads, “two plates said to be wanting, but Blake’s original prospectus says—‘America, a Prophecy in illuminated printing,’ folio, with 18 designs, Lambeth, W. Blake, 1793.” This phrase makes little sense. The conjunction “but” should be used to connect two countering statements; in this case, the second part, “America, a Prophecy in illuminated printing” in no way contradicts or counters the statement that two leaves are said to be wanting. This awkward and illogical sentence can be explained by a simple printer’s error which removed the words “folio, with 18 designs” from the quotation of Blake’s prospectus text and made the words appear a part of Sotheby’s sale entry. The printer, evidently accustomed to setting type in the auction house’s style, assumed that the number of leaves was Sotheby’s description of the copy at hand rather than a part of the quotation. The meaning of the sentence is obvious begin page 213 | ↑ back to top when one repairs Blake’s prospectus description. It thus becomes clear that the Sotheby cataloguer could not determine why two plates had at some earlier point been described as lacking, when the copy before him agreed with the prospectus in having the full 18 plates. While the meaning of this description is clear, it has not proven possible to locate any statement of imperfection: a note (since lost or erased) in the volume, or a catalogue entry for the untraced 1874 sale are equally plausible sources.

The Sotheby cataloguer’s insecurity about the number of plates in copy B can be explained by the fact that two plates are not a part of the original volume: careful examination has revealed that plates 4 and 9 not only are not original to this copy, but are facsimiles. Neither plate is sewn into the binding as an integral leaf; instead, both leaves (in their correct positions) are tipped in, pasted to adjacent leaves deep in the gutter of the volume. This is most evident when turning pages: when plates 4 and 9 are turned, plates 5 and 10 obediently follow since they are joined by adhesive at the inner margins. The remainder of the leaves in the volume, as one would expect of single bound leaves, turn quite independently of their neighbors. Further, the two leaves in question do not display the gilt edges of the rest of the volume, but form minute gaps at the top, fore, and bottom edges; both plates are shorter than the remaining leaves. As it is virtually impossible to apply matching gilt to a single leaf, inserted leaves could never display the flawless gilding of the edges of the bound leaves. It might be argued that the paper of plates 4 and 9 was merely shorter when bound than the remainder of the volume, and thus escaped gilding; two suggestions can be made to show that this is unlikely. Before gilding the edges of this book, the binder was scrupulously careful in cutting and smoothing the edges to provide a perfectly flat surface for the gilt. A binder of Francis Bedford’s reputation and ability would certainly not have permitted these two leaves to remain shorter, but would have trimmed down the remainder of the volume to even all edges. While we would gasp at giving a binder such freedom today, brutally trimming all uncut edges and elaborately gilding the edges was a perfectly normal—in fact, a desirable—way of achieving a finer copy for the nineteenth-century collector.

Another argument against plates 4 and 9 being an original part of copy B concerns the sewing of the plates before binding. Once again, Bedford, a binder of the highest quality, would have sewn the two leaves with the bulk of the volume (on stubs, if necessary); it is unlikely that he would have simply glued them to adjacent leaves. There is nothing to indicate that the original plates 4 and 9 were ever present: there are no stubs, and the binding is extremely tight. If Blake’s original plates 4 and 9 were at some time extracted from copy B, they are nowhere recorded as existing today. Since the binding was executed between 1874 and 1878, the false plates 4 and 9 must also have been inserted between those years.

Further evidence to demonstrate the alien nature of plates 4 and 9 is found in their hard and smooth wove paper, in general appearance similar to the wove Whatman hand-made paper used in the remainder of the volume.⁷↤ ⁷ Watermarks appear on only five of the sixteen genuine plates, but measurement of paper thickness has demonstrated that Whatman paper was used throughout. Were it not for matching stabholes I would be tempted to suggest that the titlepage was substituted from another copy: it alone has an obvious central horizontal crease, and it alone is touched—unconvincingly and with disregard for the design—with grey wash. But the surface of plates 4 and 9 is far smoother and does not show the fibrous surface of genuine Whatman paper; the paper of these two plates is without a doubt machine made and is considerably stiffer than Whatman paper. When viewed in a strong light, the paper is also considerably browner in tone than the surrounding Whatman paper. Bentley’s description of copy B in the Thorne catalogue⁸↤ ⁸ G.E. Bentley, Jr., The Blake Collection of Mrs. Landon K. Thorne (New York: Pierpont Morgan Library, 1971), pp. 26-27. notes that (only) “plate 9 seems to be on stiffer paper than the rest”; Joseph Viscomi noticed the same difference in thickness of plates 4 and 9 when examining copy B in 1978, suggesting in conversation with me that these two plates were from another copy of America. After more detailed examination it became clear to me that the two plates were not authentic. It is interesting to note that Bentley deleted his question about the thickness of the paper when incorporating the Thorne catalogue description into Blake Books and evidently felt no doubts as to authenticity. This is not at all surprising, since both facsimile leaves are of very high quality.

Copy B has been variously foliated in pencil—not by Blake—at the upper right corner of each sheet (1-18) and at the lower left corner of each design. Both foliations have been erased and rewritten at various times, all before acquisition by the Morgan. Written below the upper right foliation on plates 4 and 9 appear what could be described as a penciled European “7” with a crossbar, but which may more plausibly be described as a capital “F,” perhaps indicating the word “Facsimile.” It is telling that no other plates in copy B display these markings. Further evidence is provided by the foliation at the lower left, which omits plates 4 and 9 entirely, resulting in a count of 16 plates.⁹↤ ⁹ Further examination of the foliation of the sixteen genuine plates may reveal a different arrangement of the plates before binding, as some erased numbers quite out of sequence are still discernible at the lower left.

The printing ink of the two plates is a dull, dark brownish-black, similar in color to the ink used in the remainder of copy B. When compared with original Blake pulls, however, the printing ink of the spurious leaves appears flat and without surface. Both the flatness of ink and the quality of impression make it evident that these two plates were not printed by Blake’s method of relief etching. Authentic pulls show the ink impressed into the surface of the sheet to such an extent that a clear blind impression of both design and text is visible on the verso of each plate. The versos of plates 4 and 9 are smooth and unblemished.

Perhaps the most conclusive evidence against the authenticity of plates 4 and 9 is an obvious platemark, measuring 22.8 × 17 cm. (plate 4) and 24.7 × 17.5 cm. (plate 9). There would be no justification for such a platemark on an original pull by Blake, since the size of the artist’s copper plates generally determined the size of his image: for him, margins were undesirable, whether on financial or aesthetic grounds. Bentley and others have discussed the likelihood of Blake using both sides of some of his copper plates, as evidenced by close similarities in size and by the presence of maker’s stamps (usually found on the backs of copper plates).¹⁰↤ ¹⁰ Blake Books, pp. 381-82. The only instances of substantial margins in Blake’s colorprinting occur in The Book of Los and begin page 214 | ↑ back to top The Book of Ahania, both conventionally etched (intaglio) works.

The impossibility of the platemark dimensions on plates 4 and 9 can readily be shown on plate 4, where the lower portion of the design was masked by Blake in the printing of most copies, so the area beneath the crouching figure is without the four-line sentiment (so masked in copies C-F, H-M, R, a, and the single Morgan pull). In genuine impressions of this plate, an examination of the surface of the paper will demonstrate that this area was masked by placing a slip of paper over the inked plate; under raking light one can, in fact, see the blind impression of its deckle edge in the single Morgan pull. There is abundant evidence that the original copper plate was not simply cut off to delete this part of the plate. On genuine impressions the copper can be shown to extend 40 mm. below the tip of the man’s foot; in copy B the platemark is clearly visible 24 mm. below the foot, locating what would be the bottom of the copper plate in the middle of the text lines.

Facsimiles—particularly Blake facsimiles—deserve more study. In general, the earliest facsimiles of printed book pages were pen-and-ink facsimiles, often sophisticated redrawings of type. In the case of Blake, such work (as in Works by William Blake) is readily identifiable by comparison with known originals. In the field of early printed books, the most accomplished and annoying facsimilist was John Harris, who worked in England in the mid-nineteenth century, primarily on sixteenth-century religious works. His earliest method was to prepare a tracing of the original leaf to be copied, to transfer the design with a stylus or carbon-like paper to a sheet of appropriately old paper, and to fill in the letters by hand with pen-and-ink. His later technique involved the use of a very lightly printed photolithograph of the page to be reproduced, which was darkened by hand with ink, to form a remarkable facsimile. It is said that he was at times unable to tell his facsimile leaves from originals, and he therefore later insisted on signing his facsimiles with his initials in a lower corner.

Facsimiles relying on photographic means can be divided into two types: those printed planographically (as lithographs, printed from one surface) and those printed relief (as from zinc blocks). There is no indication that intaglio technique was used for the facsimiles in copy B. Photolithographic prints could be printed from litho stones, or from zinc plates (known as photozincographs). Photorelief blocks were, confusingly enough, also known as zincographs, but the difference in printing is crucial.

Whichever process was used for plates 4 and 9, there is no doubt that it was rooted in photography. Photolithography is, I feel, the more likely process used for these leaves: that process was executed by photographing an original design, transferring the photographic image to a photosensitive litho stone or zinc plate, and printing that plate or stone through the lithographic (planographic) process. Photolithography was in common use by the 1860s when Henry James instituted the use of “photozincography” for the reproduction of detailed ordnance maps for the British government.¹¹↤ ¹¹ See the biography of James in the Dictionary of National Biography, XXIX, 210-13. The extreme detail obtained was “such that the greatest error in a photozincographic reduction did not amount to 1/400 part of an inch, a quantity quite inappreciable, and much less than the error due to the contraction of the paper on which the maps were printed.”¹²↤ ¹² For exhausting details of the anastatic and photozincographic processes see Geoffrey Wakeman, Aspects of Victorian Lithography (Wymondham: Brewhouse Press, 1970). Because of James’ successful use of the process and at his urging, a photozincographic facsimile of Domesday Book appeared, demonstrating that the process was already considered suitable for the making of exact facsimiles. By the 1880s the technique was totally accepted: William Griggs employed it for his esteemed series of Shakespeare quarto facsimiles, the principal feature of which was exact duplication, barring the variations inherent in paper shrinkage and stretching, and the vagaries of the photographic process. The exact dimensions of the images on plates 4 and 9 of copy B vary somewhat from those of originals, but such variation amounts to less than 2% and can be explained by any of the factors mentioned above. The smoothness of impression argues for a photolithographic process; I have not encountered a craftsman or artist able to differentiate between impressions from stone and those from zinc plates.

It is also possible that the two plates were printed from relief zinc blocks (also known as linecuts or lineblocks), which are made by transferring a photographic image to a photosensitive sheet of metal. The unexposed areas are etched away, leaving a raised metal surface from which to print with a normal printing press. The plate is mounted on a type-high block of wood for printing. This process is suggested by the overinking on plate 9 around the periphery of the design; on the other hand, there is no trace of impression on the versos of plates 4 and 9 to indicate a relief process.

Neither process would explain the presence of plate-marks on these two impressions. If printed lithographically from stone there would be only a slight smoothing of the surface of the paper from the stone; if printed from zinc plates there might be a slightly more defined platemark, but barely visible. There is no justification for any platemark if the plates were printed from relief blocks. It may be that in a mistaken understanding of Blake’s technique, a false platemark was applied by running the dampened sheets through a rolling (etching) press next to a sheet of metal, the edges of which would impress a noticeable indentation into the paper.¹³↤ ¹³ A curious example of false lithographic platemarks is described by Michael Twyman, “A Note on some Lithographic Stones relating to Alken’s Ideas and Notions,” Journal of the Printing Historical Society, 14 (1979/80), 82-88.

In any case, it is virtually certain that a photographic process was used in the preparation of these facsimile leaves, for minute flaws and printing spots quite incidental to Blake’s designs are reproduced with astonishing fidelity. It would, in fact, serve little purpose to illustrate the two facsimile leaves in their entirety here, since they would appear virtually indistinguishable from originals. Some small details, however, are lost in the reproductions, most noticeably the two small birds hovering in the sky on plate 9, between text and tree boughs in the center of the plate.

begin page 215 | ↑ back to top

It has not proven possible to identify positively the copy (or copies) of America from which these two facsimile leaves were made. Plate 4 in copy B shows certain affinities with that plate in copy F (British Museum Print Room), which is uncolored, and which has been available since 1859. In particular, the broken flourish of the top of letter “S” beginning the text appears to be an identifying feature, as are other spots and printing flaws. Such a close relationship does not appear between the facsimile of plate 9 and that plate in copy F, where there are distinct differences. The birds beneath the tree boughs are absent in the facsimile, and other details of inking also differ.

When the dubious nature of plates 4 and 9 in copy B was pointed out to Robert N. Essick, he informed me that in his Blake collection are impressions of the same two plates, purchased by him as facsimiles. The ink, impression, and images are identical to plates 4 and 9 in copy B. Essick acquired these leaves from the London firm of antiquarian booksellers, Walter T. Spencer & Co., in whose stock they were found without further identification, and where it is believed they had lain for many years.

I have not been able to identify the maker of the two facsimile leaves in copy B. The facsimiles were certainly made in England, where (if, indeed, it was the model) copy F was available to scholars, collectors, and dealers during the period of 1874-1878. I have no doubt that photographic services were available at that time, although I have not examined the records of the British Library to verify this. No known facsimilist of Blake’s works can be connected to copy B.

William Muir, the leading facsimilist of Blake, was active beginning only in the 1880s; in fact, no facsimiles of any of Blake’s works are known to date from the 1870s. It appears that those involved with Blake’s illuminated printing were less concerned with completing individual copies than they were in making Blake’s works available to a wider audience (this in spite of the fact that the reviewer of a Muir facsimile wrote in the Times of 29 July 1886, “All the world does not admire Blake, but then [Muir’s] edition of fifty copies is not for all the world”).¹⁴↤ ¹⁴ Unnumbered advertising brochure of Blake materials issued by Bernard Quaritch, dated Nov. 1886.

It would be tempting to associate John Pearson, the first recorded dealer to handle copy B, with the making of these facsimiles, as he is known to have been responsible for various other facsimiles. Pearson was active in the London Book trade from about 1870, and is known to have handled at least three copies of America between 1878 and 1896: he owned copy B in 1878, copy D between 1886 and 1888 (he had difficulty in disposing of this copy, offering it in catalogues for £52, later for £45, and finally auctioning it off for £23 in 1888), and copy E in 1896. But evidence shows it is unlikely he was involved with the facsimiles since they were both present in copy B when he purchased it at auction in the Dew-Smith sale in 1878. If the shadowy sales of 1855 and 1874 could be identified, further light might be shed by the names of sellers and purchasers. It would also be tempting to associate the firm of Walter T. Spencer with the preparation of these facsimiles, since Essick purchased his two leaves from their stock, and, as Bentley has shown, the firm was responsible for the fraudulent coloring of Europe (L) and America (Q).¹⁵↤ ¹⁵ Blake Books, pp. 105-06. In collecting circles it is generally known that the firm was less than careful in informing customers about the presence of facsimiles, and it handled a fair number of works by Blake. However, no evidence connects Spencer with the facsimiles in copy B; it is not known whether company records exist.

It is also somewhat obscure why these facsimile leaves were made, other than the obvious desire to complete an imperfect copy. It seems unlikely to have been for financial reasons, as the sums involved were never great. Copy B (uncolored) sold for £16.5.0 in 1878, copy L (uncolored) sold for £21 in 1879, copy R (colored) sold for £31.10.0 in 1880, and poor copy D finally fetched £23 in 1888. One wonders if the relatively small sums merited the production of what must have been expensive facsimiles. It is only possible to surmise that an English dealer or collector was responsible for the facsimiles; the fact that the two plates were nowhere described as facsimiles can only lead one to the inescapable conclusion that they were intended to deceive.

One further—and puzzling—inscription in copy B should be noted. On the free front endpaper, an unidentified hand has written in now-smudged pencil, “Lowndes gives 18 designs. There are two more in some copies but I believe these to be supplementary numbers and that the book as published had only 18.” No copy of America is known to have had more than eighteen plates, nor are the three known proof plates likely to be what is meant in the inscription.¹⁶↤ ¹⁶ These proofs are reproduced in David V. Erdman, The Illuminated Blake (New York: Doubleday, 1974), pp. 392-95. They are unique proofs, one heavily corrected by Blake, and must represent false starts on Blake’s part: the images are much improved in the published versions. There is, in fact, no copy of any of Blake’s work in illuminated printing with twenty plates. While it might be argued that the Sotheby cataloguer for the Dew-Smith sale incorporated this curious note in his description for the catalogue, the above evidence demonstrating the insertion of plates 4 and 9 make this unlikely.

It thus appears that these two facsimile leaves, reproduced by a fundamentally photographic process, were inserted in copy B of America between 1874 and 1878 and have escaped detection by all owners and scholars before identification by Joseph Viscomi and myself in 1978. The quality of reproduction is so deceptive that the facsimiles went undetected in exhibitions at the National Gallery (1913), Manchester (1914), Nottingham (1914), the National Gallery of Scotland (1914), the Fogg exhibition (1930), the Philadelphia exhibition (1939), and at various exhibitions at the Morgan Library, including that of the Thorne Collection (1971), in the catalogue of which both facsimile leaves were reproduced as genuine. One can only wonder what other facsimiles have for so long remained undetected.¹⁷↤ ¹⁷ I would like to thank Herbert Cahoon, Barbara Prince, and Lisa Vercollone for their assistance.