MAKING BLAKE

PART V

Chapter 9: Posthumous Songs Reconsidered

This is the second of three related, sequential essays, preceded by Chapter 8 and followed by Chapter 10.

William Blake’s America, a Prophecy and Europe, a Prophecy were posthumously printed and paired three times in two separate printing sessions, presumably between 1829 and 1832. Two pairs appear to have been printed by Mrs. Blake and the other pair by Tatham (see Chapter 9). The copies of America have 18 plates; the copies of Europe have 17 plates and are missing plate 3. Mrs. Blake’s pairs were printed in black ink and Tatham’s pair was printed in dark reddish brown ink, as were Tatham’s three copies of Jerusalem. America has only four extant loose posthumous impressions; Europe has ten such impressions, and Jerusalem has seven. In short, these copies were deliberately and systematically produced. The ten posthumous copies of the Songs of Innocence and of Experience, all printed by Tatham, appear, in contrast, to have been arbitrarily printed, indifferently compiled, and randomly distributed. Indeed, as recorded in the charts in Blake Books (370-72) and Blake Books Supplement (112), five copies look like they are in various stages of disintegration. Songs copies d, e, g, i, and p have only 40 to 43 of Songs’s 54 plates, while half a dozen or more “copies” are really just clusters of 10 or more plates from Experience, or scatterings of a great many loose impressions named after owners or institutions. Of the five currently complete copies, one has 57 impressions in three ink colors, with three duplicate plates, plate b but no plate 15, and another has 44 black impressions and 10 orange brown impressions. Posthumous impressions of the Songs vary widely in paper size, image quality, and ink color. As recorded in Blake Books, they were printed in “grey,” “black,” “brown,” “red,” “reddish brown,” “yellowish brown,” and “orangish brown.”[1] Keynes and Wolf describe the ink colors as “grey,” “dark grey,” “light brown,” “red-brown,” “sepia,” dark brown,” and “orange-brown” (66-69).

Color, however, is to a great extent subjective, and its saturation is affected by various factors: the thinness of the ink layer, the type and condition of the paper when printed (damp or dry), the amount of pressure used to transfer the ink, and the kind and amount of oil in the ink. Moreover, with a copy the hue can vary depending on when and how the ink was replenished and applied during the session. What Bentley and Keynes refer to as “grey,” I see as thin black with more white paper showing through than thicker opaque black ink. That this greyish ink splotches black where the ink builds up, e.g., along the plate’s sides, reveals that it was a black ink applied thinly (illus. 00). Brown runs along a continuum of either dark reddish brown (sepia) to a lighter, slightly more saturated reddish brown to a lighter orangish brown—and some copies, such as e and h, have all three shades of “brown.” When examined in terms of production, the diversity among posthumous copies of Songs, in leaf size, number and order of plates, and ink colors, appears less frightening and, well, less disorderly. Indeed, when examined closely, together, and materially, the over 550 posthumously printed impressions of Songs reveal order and intent. They were printed per copy and not per plate. In other words, despite appearances, the copies they formed were, with the exception of Songs copy h, never compiled from piles of loose and diversely produced impressions.[2]

Complete copies of Songs have 54 plates. Songs of Innocence, produced first in 1789, originally had 31 plates. The first printing produced seventeen (or possibly eighteen) copies: U and W were printed in black ink on 31 leaves, possibly with copy V; I, J, X, and "Innocence" of Songs of Innocence and of Experience copy F were printed in green ink on both sides of the leaves; A-H, K-M, Z were printed in the same style in yellow ochre or raw sienna ink.[3] In addition, the "Innocence" section of what would later become Songs of Innocence and of Experience copies B-E were printed in this first session. The first copies, printed in black ink on one side of the leaves, were uncolored, appearing more like a book of prints than a book of poems; all the subsequent copies were colored and, with images on both sides of the leaves, had facing pages characteristic of books, though the light imprint, wiped plate borders, and simple washes made these copies appear like “printed manuscripts.”

In October of 1793, when Blake printed his Prospectus describing and advertising illuminated books and some other original graphic works for sale, he listed Innocence as having “25 designs” and Experience, which was forthcoming, as having “25 designs” (E 692). He counted illustrations and vignettes instead of plates and advertised the two sections as separate works. Over time, Blake would move six plates comprising four poems from Innocence to Experience, but he was never to have the same number of poems, plates, or “designs” in the two parts. He executed plate 1, the general title plate, in 1794 to announce “Songs of Innocence and of Experience // Shewing the Two Contrary States of the Human Soul,” that is, to enable him to combine separately printed books into a unified whole. For the first copies of Songs, copies B, C, D, and E, Blake combined copies of Innocence printed in 1789 with copies of Experience printed in 1794.[4] For these copies of Songs, Blake moved two leaves with plates 26/34 and 35/36 from Innocence to Experience. Plates 34-36 are “The Little Girl Lost” and “The Little Girl Found,” and one can, of course, make a case for reading them as Experience poems, but plate 26, “A Dream,” simply went along for the ride, because it was printed on the recto of plate 34. Blake kept plates 34-36 in Experience and eventually added plate 52, “To Tirzah,” to this section. He was to move plate 53, “The School Boy,” and plate 54, “The Voice of the Ancient Bard,” from Innocence to Experience in the last eight copies. Hence, Innocence evolved from its earliest autonomous copies with 31 plates to its last iteration as a 26 plate-section in the Songs.

Copies of Innocence and Songs had varied plate orders until about 1818 (see Viscomi BIB 00). Songs copies T, U, W, X, Y, Z, and AA, the last seven copies Blake produced, between 1818 and 1827, have the same plate order, designated by editors in the early 20^th century as the “standard order.” This plate order, recorded as plates 1-54, has provided the numbers used today to refer to the copper plates and the poems/pages/designs in any of the copies of Innocence, Experience, and Songs. Blake also produced a plate a, a design of five winged cherubs carrying a naked man, which he included only in Songs copies B, C, and D as a tailpiece, and a plate b, “A Divine Image,” that he included only in Songs copy BB.

Tatham, the posthumous printer of the Songs, did not print plate a, but he did occasionally print plate b, and he sequenced his plates according to the standard order. An examination of Songs copy a, which now appears complete with 54 plates, will reveal that it was originally 53 plates, reduced to 40, and then reassembled in one or two stages to its present condition. Tracing its evolution will reveal that Tatham systematically abridged six initially complete copies of Songs by 1840 in the same manner and compiled at least one composite copy from the impressions of two or three printings.

I. The Abridgement of Posthumous Songs copy a

Songs copy a originally consisted of 53 plates, including plate b but not plates 15 and 45 (“Laughing Song” and “The Little Vagabond”), which Blake had etched on the recto and verso of the same copper plate. Plates 15 and 45 are also missing from Songs copies d, g, i, and p, also printed in light black ink, and plate 15 is missing from Songs copy h, most of whose Innocence plates were printed in black ink, and it is also missing from Innocence copy T, a posthumous copy printed in black ink. This consistent absence of the poems in copies of Songs printed in black ink suggests that their shared copper plate was not present when the copies were printed. Plates 15 and 45 are present, however, in copies b, c, f/j, and e, which were printed in reddish and orangish brown inks. The pattern here suggests that impressions in reddish and orangish brown inks preceded black, because it seems more likely that the copperplate bearing plates 15/45 went missing between printings sessions rather than it went missing for the first printing but was then found for the second.

The original 53 impressions of Songs copy a were of uneven quality printed on heavy J Whatman 1831 paper, 24.5 x 19.3 cm, also used in Songs copies g and i (see Chapter 9).[5] Plates 30-32, 37, 44, 47, 48, 50-54, and b were extracted, reducing Songs copy a to 40 pages, which were—and remained—stitched as a unit. However, in 1864, when B. M. Pickering sold Songs copy a to the British Museum, Songs copy a had 54 plates. Fourteen plates had been added, including plates 45 and b, but not 15. Plates 52, 53, 48, and 54 were printed in black ink, followed by plates 30, 31, 47, 37, 32, b, 51, 50, 44, and 45, all printed in orangish brown except plate b, which is closer to reddish brown. Both sets of plates were stitched to the initial forty plates.[6] If they were added at the same time, then they came from the same source; if they were added at different times, then they came from two different sources.

The 14 plates were added randomly, hiding the fact that Songs copy a’s original 53 plates were sequenced in the standard order. Its initial order is revealed when the extracted plates are returned to the forty bound impressions:

30, 31, 32, 37, 44, 47, 48, 50, 51, 52, 53, 54, b

1-14, 16-27 28, 29, 33, 34, 35, 36, 38, 39, 40, 41, 42, 43, 46, 49[7]

The original plate order retained in the abridged 40 plate copy reveals that Songs copy a originally had a full complement of 53 plates deliberately ordered and was not—as it now appears—compiled from diverse piles of black and orangish brown impressions.

The person who initially ordered Songs copy a was not likely the person who added 14 plates randomly, since he presumably would have reinserted the impressions into their original places rather than add them to the back. Tatham appears to have been responsible for Songs copy a’s initial plate order, because the reddish-brown copies (b, c, f/j) that he printed were similarly ordered. Songs copy b was acquired by Hannah Boddington, possibly as early as 1832 or 1833, and remains in this order. Copy c was acquired by her brother Samuel and is now in an order without authority and appears to have been rebound by T. J. Wise (BB 428). An early owner dated copy f “1836” and it was bound by 1869, but then it disappeared, presumably reappearing as copy j, finished in excellent imitation of Songs copy U (Viscomi, BIB 00), rebound in the old binding (BB 428) in the standard order, presumably its initial order. For the plates of these copies to have been sequenced according to Blake’s standard order—long before editors recognized that there was an order—indicates that their original compiler knew one of the last seven copies Blake produced (T, U, W, X, Y, Z, AA), presumably using it as a model or having recorded its plate sequence. Tatham probably knew copy W, which Mrs. Blake had until she sold it in March of 1830 for £10.10.0 to John Jebb, Bishop of Limerick (BB 423). Tatham almost certainly knew copy Y, which belonged to Edward Calvert, his friend and fellow “Ancient.”

The idea that the posthumous printer or an owner replaced Songs copy a’s missing impressions with better quality dark black or orange brown impressions is not credible. The replacements are not, in fact, better quality impressions. They appear as unevenly and poorly printed as most of the initial 40 impressions. Mostly, though, the idea is not credible because the same cluster of plates removed from Songs copy a were also removed from Songs copies d, e, g, i, and p (see Chart 1). Such repetition among six copies indicates intention and location. The plates were necessarily extracted where the six copies were produced, and thus, presumably, by Tatham and apparently at the same time.

Chart I

Cluster of Experience plates missing and not printed in Songs copies a, d, g, i, p (black), e (orangish brown):

a 15, 30, 31, 32, 37, 44, 45, 47, 48, 50, 51, 52, 53, 54 b

d 15, 30, 31, 32, 37, 44, 45, 47, 48 50, 51, 54 b

g 11, 15, 30, 31, 32, 37, 40, [44] 45, 47, 48 50, 51, b[8]

i 15, 30, 31, 32, 37, 44, 45, 47, 48 50, 51, 52, 53 b[9]

p 15, 30, 31, 32 37 44, 45, 47, 48 50, 51, b

e 30, 31, 32, 33, 37, 41, 44, 45, 46, 47, 50, 51, 52, 53

The same cluster of Experience plates missing from Songs copies a, d, g, i, p, and e are present in small clusters of impressions now widely distributed:

grey/light black in copies:

l 29, 33, 38, 39, 41, 42, 43, 46, 49 51, 52

m 54

n 30, 31, 37, 44, 47, 50, b

lc: 30, 31, 32, 37, 44, 47, 50

bentley: 30, 48

Keynes b

Juel-jensen/Danson 44

Harvard 29, 30 37

Tate 48 48

Bieneke 50 b

Dartsmouth 36

Welsyan 46 49

Browns

k (r-b) 30, 43, 45, 48 52

Bentley 28 40, 44, 45, 46 48

JJ/Danson 33

Victoria Library 38 53 (?orange)

Untraced 42, 51

o untraced: 31 [yellowish brown], 38, 39 orangish brown [Bentley] + 14

Posthumously printed loose impressions from Songs of Innocence:

grey/light black

Victoria Library 18, 24

Tate 22

Morgan, from copy n 2, 13

Fitzwilliam, from copy m 3, 10, 11, 19, 22

Bentley 22

Copy n 2

browns

T. D. Danson 7, 10

Brown University 13, 20, 21

Copy n 13

Because the same cluster of Experience plates were extracted from copies a, d, e, g, i, and p, more loose impressions of Experience than Innocence plates are extant. The overlap between extant loose Experience impressions and the clusters of Experience plates extracted is not one-to-one, nor could it be since 9 or 10 of the extracted Experience plates from one of the abridged copies of Songs probably provided the plates used to complete the abridged Songs copy a. But the extracted Experience impressions, mostly in black ink, are too well represented in the extant loose impressions to be coincidental. Also, loose impressions extracted from copies and sold separately—or even as a small set or group that appears already to have been broken up—have a greater chance of going unrecorded, perhaps to be found in extra-illustrated copies of Life of Blake or Nollekens in his Times.[10] Even without a one-to-one mapping, a clear pattern of extraction is discernable from extant impressions to reveal that the plates missing from copies a, d, e, g, i, and p were indeed printed and once part of these copies. The missing impressions were deliberately extracted from completed copies to form subsets of Experience plates. Songs copy a can serve as the model for all the copies of Songs abridged by Tatham. The abridgement was coherent, comprised of 25 Innocence and 15 to 17 Experience plates and one general title plate, two section title pages, and two full-page frontispieces.

II. The Source of Songs copy a replacement plates

When, from whom, and in what condition Pickering acquired Songs copy a are not known. Only 29 when he sold it in 1864, he may have inherited it from his father, William Pickering, the publisher, whose own interest in Blake is evinced in his publishing the first edition of Blake’s Songs, edited by Wilkinson, in 1839. Pickering senior died in 1858, and, conceivably, he could have bought the abridged Songs copy a directly from Tatham. If, however, Songs copy a was acquired rather than inherited, perhaps as stock for his bookshop, then it appears to have been B. M. Pickering’s first work by Blake. His first recorded Blake acquisition appears to have been a manuscript of ballads from Sotheby’s in December 1865. The Pickering Manuscript, as this manuscript was to be called, was quickly followed by acquisitions in May of 1866, from Robert Arthington’s sale at Sotheby’s, Songs copies E and H, Visions copy C, Europe copy F, and a volume of Blakeana that included Thel copy a, eight plates from posthumous Songs copy n, and many life-time proofs and discarded illuminated impressions. Pickering appears to have acquired these as a collector and not as a book seller, holding on to them for years, even stamping his name in the first flyleaf of Songs copy E (BB 414).[11]

His Songs copy a may have been acquired in its abridged form or already completed. It was completed in one or two stages, with plates 52, 53, 48, and 54 (in this order) added first. Plate 53 was printed in the same light black ink on the same heavy-weight J Whatman 1831 paper cut to the same size as the first 40 leaves. Plates 52, 48, and 54 were printed in a darker black ink on longer leaves of thinner J Whatman 1832 paper, approximately 28 x 19.4 cm, a paper that Tatham had used for other copies of the Songs. Following plates 52, 53, 48, and 54 are plates 30, 31, 47, 37, 32, b, 51, 50, 44, and 45, all printed on the thinner Whatman paper in orangish brown ink except plate b, which is closer to reddish brown.[12] These orange-brown impressions are thought to have come from Songs copy d or copy e—since they are missing the same plates as copy a—and/or from copy o (BB 426n1), possibly a miscellany of plates. From this list we can rule out Songs copy d. Though recorded as being in “sepia” (Census 67), “brown” ink (BB 426 n1), “brownish black” (BBS 112), and “dark sepia” (Yale University Library Gazette, vol. 49, #4, April 1975, 330), all forty-two plates of copy d are in black ink (illus. 00). Copy d could not have provided the ten orange-brown plates, nor the black ink impressions of plates 52 and 53, since they are still in copy d. Possibly, it provided plates 48 and 54.[13]

Keynes and Wolf identify the color of the last ten impressions of Songs copy a as “light brown” (66) and Bentley as “brown” (BB 370). The color, however, is more accurately described as an orangish brown, and closely resembles the ink used in Songs copies e and h, which Keynes and Wolf described as “orange brown” (67) and Bentley as “orangish-brown” (BB 373). Essick describes about half the plates in Songs copy h, which he owns, as “orangish brown,” and he has checked their hues against those in copies e and a and believes the inks are the same. As noted, hues of printing ink may differ among impressions in the same copy and from the same printing session. Plate 49 from Songs copy h (illus. 00) and copy e (illus. 00) match in hue, while plate 47 in copy h (illus. 00) and a (illus. 00) match, though the darker h impression is redder and than the lighter and more orangish a impression, much like a darker black looks greyish when thin. The latter impression appears like a second pull, as is evinced by the same monks and blemishes in the right side of the netting.

The orangish brown impressions in Songs copies a, e, and h were from the same printing sessions, which appear to have included or overlapped with the session producing the darker reddish brown impressions. The ten orange-brown Experience impressions in Songs copy a appear to have come from the cluster of 14 Experience plates extracted from Songs copy e. That cluster, however, included plates 52 and 53. In copy a, however, plates 52 and 53 are in black ink, as are two other added plates, plates 48 and 54, which means these four plates did not come from copy e. Songs copy e could have contributed only 9 or, if plate b was in the cluster, 10 of the 14 impressions added to copy a. Impressions extracted from copy e not going to copy a were plates 33, 41, 46, 52, and 53. Impressions of at least three of these appear to be extant but widely dispersed (see Chart I).

For the added plates in Songs copy a to be both in black and orange-brown suggests that they may have come from two different sources, black preceding orange-brown, and that Songs copy a was completed in two distinct stages, with four black impressions stitched to the initial 40 impressions followed by ten orange brown impressions. However, plates in two ink colors also suggests that the impressions came from a shared source, from a miscellany of black ink and orange-brown ink impressions. Over time and with the break up of so many copies, such miscellanies seem inevitably to have been formed. Tatham began selling copies of Songs soon after he printed them in 1832 and to have abridged them before the end of the decade, when Edwell dated his Songs copy d “1840.” By the time Pickering sold the reconstructed Songs copy a to the British Museum in 1864, Tatham had been selling complete copies of Songs, abridged versions, clusters presumably as subsets, and loose impressions in volumes of Blakeana (see Viscomi, Printed Paintings Chapter 12) for at least 30 years. In that time, specific clusters may have been further broken up or enlarged into miscellanies of impressions—larger than a cluster but smaller than a copy or an abridged copy. Songs copy o appears to belong to that last category—and it belonged to Pickering till 1869, when he sold it to Charles Eliot Norton (BB 00).

From whom, when, and in what condition Pickering acquired copy o are not known. Was it ever complete? Was it an abridged copy like copies a, d, e, g, i, or p? Was it a cluster of extracted Experience plates? Descriptions of it long after Pickering owned it suggests that it was a miscellany comprised of an Experience cluster and other loose impressions from various printings. In 1931, it appears to have been reduced to 18 plates. Duttons’ Sale Catalogue of the Private Library of Paul Hyde Bonner (Bentley, Sales Cat. 1900-1999) described it as “eighteen proofs” with "eight of the plates . . . printed in black and ten in sepia.” Keynes and Wolf record copy o as having 18 impressions printed in “black and sepia” (68). By 1938, it was broken up, with four of its plates listed in the December 1938 catalogue for the Weyhe Gallery.[14] Plates 20, 21, and 39 are recorded as being in “yellow-brown,” and plate 38 in “red.” In Blake Books, Bentley records plate 39—which he owns—as printed in “orangish brown” (371); in Blake Books Supplement, he records recently rediscovered plates 20 - 21 and plates 38 and 53 as “brown” and “orange” respectively.

The recently discovered posthumous prints of plates 13, 18, 20, 21, 24, 36, 46, and 49 assigned to Songs copy o are described as brown and grey or black (BBS 112). The 18 plates of copy o may have been the following 10 brown and 8 black plates: 13, 18, 20, 21, 24, 28, 30, 31, 36, 38, 39, 40, 44, 45, 46, 48, 49, 53. If copy o contributed the 14 plates to complete Songs copy a, then it was initially a miscellany of 32 or more impressions in two or three inks with duplicates. Pickering conceivably transferred orange-brown impressions of plates 30, 31, 32, 37, 44, 45, 47, 50, 51 and b, initially from copy e, and black impressions of 48, 52, 53, and 54, all from copy o to copy a. If the recently discovered impressions of plates 30, 31, 44, 45, 46, and 53 were actually part of copy o, then they were duplicates, befitting of a miscellany, or perhaps not all were part of copy o.

Given the multiple inks, Songs copy o was not one of the clusters of Experience plates extracted from a posthumous copy, since clusters would be in the same one ink as the source copy.[15] Copy o appears to have been a miscellany of loose impressions, possibly augmenting a cluster of Experience plates, created over time by dealers and collectors. Or it may have been an incomplete or broken composite copy initially created by Tatham. As we will see, copy h, with impressions mostly in black and orangish brown ink and one, of plate b, in reddish brown, is a true composite copy, with three duplicates, compiled from discarded impressions from two or three printing sessions, rather than a copy deliberately printed as such. If used by Pickering to complete Songs copy a, then Songs copy o was most likely a miscellany of impressions; but if he did not, then copy o appears like an editorial construct, a catch-all category for loose posthumous impressions.

Tatham used sepia or a reddish brown to print Songs copies b, c, f/j. He used orange-brown, which is sepia thinned, lighter, and more saturated with red, for copies e and Experience of copy h. And he used black, dark and light, in Innocence copy T and Songs copies a, d, g, i, and p, and Innocence of copy h. The printing order of the posthumous Songs appears to be: b, c, f; /e, Experience of h, and possibly impressions in copy o; / Songs d, p, Innocence T; /Songs a, g, i. Five copies in black ink and one in orangish brown ink were abridged and their extracted plates provided the majority of extant loose impressions. The main difference among the six abridged copies is that Songs copy a was completed, possibly in two stages from two different sources, with plates extracted from Songs copy e and another copy, or, more likely, at the same time from one source with impressions in both inks.

III. Posthumous Songs copies i and e

Songs copy i follows the same pattern as Songs copies a and d. Initially it was probably 53 plates, minus plates 15/45 but possibly with plate b, all heavily printed in light black ink on the same approximately 24 x 19 cm size thick leaves of J Whatman 1831 paper used for copies a and g. With the extraction of the same cluster of 11 Experience plates, it was reduced to 43 impressions. Bentley records the Innocence plate order for Songs copy i as: 1-14, 16-25, 48, 26-27; he records the Experience plate order as: 28, 33, [29], 34-36, 38-43, 46, 49, 52-4, which he notes corresponds to no known order (BB 378, 380).[16] If, however, we reinsert the extracted Experience plates and the missing plates 15 and 45, then the number and order of the plates comprising the abridged copy i can be recovered:

15 30, 31, 32, 37, 44, 45, 47, [48], 50, 51, b

1… 28, 33, 29, 34, 35, 36, 38, 39, 40, 41, 42, 43, 46, 49, 52, 53, 54

Plate 48 is on thinner, shorter paper and was inserted later (BB 374n47), and plates 33 (“Holy Thursday”) and 29 (Experience title page) were transposed—presumably by accident. Initially complete with 53 plates sequenced in the standard plate order, Songs copy i was abridged to 43 plates, its size when sold at Sotheby’s on 29 April 1862, lot 195, as "Songs of Innocence and Experience 43” for _4.6.0 to James Toovey.[17]

Songs copy e was also sold at the 29 April 1862 Sotheby auction, as lot 196: "Another set, wanting three plates 40," also to Toovey, for _1.6.0.[18] Copy e has a very intriguing history. It was printed in both reddish brown and orangish brown ink—which is to say, the same basic ink diluted and replenished over the course of the printing session—with probably 54 plates—including plates 15 and 45 but probably excluding plate b. Of course, the idea that it was missing “three plates” when sold was relative to the 43-plate copy i in lot 195. It was actually missing 14 impressions, all of which, as recorded in Chart I, were from Experience and match the cluster of Experience plates missing from copies a, d, g, i, and p.

Toovey was a major bookseller turned collector later in his life. According to Roberts, writing in 1895, Toovey,

having acquired a considerable fortune in business, . . . was able to indulge in the luxury, rare amongst booksellers, of collecting a private library for his own entertainment. He retired from active business several years ago, and passed his remaining days in the ever-delightful society of his bibliographical treasures. He died in September, 1893, in his eightieth year, and his stock of books came under the hammer at Sotheby's in March, 1894, when 3,200 lots realized just over £7,090. His very choice private library is still in the possession of his son, and among its chief cornerstones is the finest First Folio Shakespeare known.

(255)

This “very choice private library” was inherited by Toovey’s son, Charles, and sold by him in 1899 to J. P. Morgan, who had it catalogues in 1901. The catalogue’s full title acknowledges that much of Toovey’s library came from the Earl of Gosford: Catalogue of a Collection of Books formed by James Toovey principally from the library of the Earl of Gosford the property of J. Pierpont Morgan (1901). According to the catalogue’s Preface, the collection

was formed by the late James Toovey, the London bookseller, almost entirely during the last twenty years of his life (1873 – 1893), and after his retirement from business. A large proportion of the books formed part of the Library of the late Earl of Gosford, which was bought privately by Mr. Toovey in the year 1878. . . . The Library as a whole was purchased by its present owner, in the year 1899, of Mr. Charles J. Toovey, the son of Mr. James Toovey. . . .1901.

The Catalogue included Songs copy e, which Toovey presumably sold c.1862 to Gosford and bought it back in 1878. Songs copy e, however, is described in the Catalogue as:

Octavo: engraved and coloured by the author, yellow morocco extra, gilt edges. This copy contains a general title for both works, and a separate title for each, in addition to 27 plates of the Songs of Innocence, and 23 plates of the Songs of Experience.

(87)

Thirty-seven years after James Toovey’s acquired posthumous Songs copy e, it was acquired by Morgan as a complete copy, “engraved and coloured by the author.” Songs copy e had metamorphosed by 1899 from 40 posthumously printed and uncolored impressions in 1862 to 53 impressions (minus plate 53) or 54 impressions (if plate b were included) printed and “coloured by the author.” This sequence of events, of course, raises numerous questions: who completed the copy, how, and when? Was Songs copy e, in this completely altered state, one of the works that Toovey acquired from Gosford’s library? If so, Toovey would be excused for not recognizing the entirely transformed copy as the one he sold Gosford years earlier.

The 13 Experience plates (30, 31, 32, 33, 37, 41, 44, 45, 46, 47, 50, 51, 52) that completed copy e came from uncolored Songs copy K, which, with Innocence of Songs copy O, had at one time formed a complete monochrome copy of the Songs (BB 418-19). Songs copies O and K were separated by 1817 (BB 419). While their earliest provenances are unknown, copy K was in the Toovey/Morgan 1901 Catalogue, described as: “Another copy of Songs, Folio; eleven plates, printed in brown and uncolored, half morocco” (87).[19] It had already been cannibalized of 13 impressions to complete Songs copy e. But who did this and when?

An examination of the completed Songs copy e reveals clearly that it was completed in two distinct stages. The initial 40 impressions of Songs copy e were colored before the 13 Songs copy K impressions were added. These 40 impressions, now unbound, were treated as an autonomous and presumably complete artifact. They were finished in translucent watercolors in imitation of Songs copy Y, as a comparison of “The Lamb” from copies e and Y ascertain (illus. 00). The leaves of copy e were gilt top, outside, and bottom and presumably bound. The Songs copy K impressions were added later at the back of the bound volume as a group, with leaves gilt top and outside, but not bottom. The 13 Songs copy K impressions were finished in dry opaquish water colors by yet another, less professional, hand, excessively using gold paint along the margins and interlinear flourishes between stanzas (illus. 00). This artist, unlike the colorist of the 40 impressions of copy e, improvised without a model. Indeed, the colorist was unaware a model had been used for the copy e impressions, or, presumably, he or she would have used it for the sake of consistency.

Songs copy Y, the model for the copy e impressions, belonged to Edward Calvert, who died in 1883. It appears to have remained in his family till around 1893, when it was offered by Ellis and Elvey in their Catalogue #75 (?1893) (BB 424): "This . . . copy has been treasured in the Calvert family since Calvert received it from the hands of William Blake himself"; it was listed at £150. It appears to have entered F. S. Ellis’ library, from which it was sold after his death at Sotheby’s, 4 November 1901, to A. Jackson for £700.[20] F. S. Ellis, according to Roberts, was “doubtless one of the most successful of modern bibliopoles who lived in the vicinity of the Strand.” He

was an apprentice of James Toovey, and who in a comparatively few years built up a business second only to that of Quaritch. . . . Mr. Ellis's shop was at 33, King Street, Covent Garden, and afterwards at 29, New Bond Street, and the prestige of his name is worthily maintained by his nephew, Mr. G. I. Ellis (with whom is Mr. Elvey), at the latter address.

(245-6)

Songs copy e’s first transformation, its coloring in imitation of Songs copy Y, could have occurred between c. 1862 and 1883, while the model was in the possession of Calvert. In addition to raising the possibility of Calvert as colorist of copy e (though the quality of coloring suggests not), these dates mean that Songs copy e could have been colored while in Gosford’s library, by 1878, or while in Toovey’s library, between 1878 and 1883—or, assuming a skilled facsimilist was at work, between 1893 and 1899, while Songs copy Y was in the possession of Toovey senior’s former apprentice, F. S. Ellis. Was Songs copy e colored by order of the 3^rd Earl of Gosford, the collector, or the or 4^th, the son who sold the collection? Or was it completed by James or Charles Toovey?

The Tooveys’ connection and presumed accessibility to Songs copy Y, however, appears to be a coincidence, because copy Y was not used as the model for the 13 added plates. As noted, the Songs copy K impressions were colored without reference to a model, which means that the owner commissioning the colorist was unaware of the model—and, perhaps, unaware that the 40-plate copy he was completing had been posthumously printed and colored.[21] On the other hand, with Songs e/K having had four hands involved in its production—Blake, Tatham, and two posthumous colorists—we need to ask what the second colorist was told to do and why? He or she appears not to have had the other colored impressions to consult, because the type of colors and style of coloring do not match.

Songs copy e appears likely to have undergone its first transformation while owned by Gosford, commissioned by either the 3^rd or 4^th Earl. If so, Toovey is excused for not recognizing the completely altered copy e as the copy of Songs that he had sold Gosford sixteen years earlier. Toovey may have recognized by 1878, though, in light of Gilchrist’s Life and the other copies of Songs then coming to the market (e.g., Songs copy E, at Sotheby’s 17 May 1866) that a complete copy of Songs had 54 plates, not 43.[22] Whether Toovey acquired Songs copy K as part of the Gosford Library or separately for his “choice private library,” he or his son Charles appear more likely than Gosford or his son to have taken 13 impressions from Songs copy K to complete Songs copy e. Gosford appears less likely because, presumably, he knew the model for coloring Songs copy e and, presumably, would have had the second colorist use it for the added plates.

IV. Posthumous Songs copies g and h

Although now loose with an initial plate order that cannot be determined, Songs copy e follows the pattern of the abridged copies and thus, presumably, its 54 plates were in the standard order. Songs copy g (now divided into Innocence and Experience, or copies g1 and g2) also has an interesting, though less colorful history than copy e. It exhibits the same pattern of extractions as Songs copies a, d, e, i, and p, which suggests that it too was initially complete, probably with 53 plates, all printed in black ink on the thicker leaves used for Songs copies a and i. Songs copy g had the following plates extracted:

11, 15, 30, [31], 32, 37, 40, [44] 45, 47, 48, 50, 51, b

Songs copy g has plate 31, but it is on thinner paper, and it has plate 44, but it is in orange-brown ink; both impressions appear to have been added, which suggests that the original black impressions of plates 31 and 44 were among the Experience plates extracted. Also currently missing are plates 11 and 15 from Innocence, but plate 15, along with plate 45, was not printed for this copy (or other copies of posthumous Songs printed in black ink). Plate b, though, was presumably printed and included. How long Songs copy g—or any of the abridged copies—remained in its initial 53-plate state is not known. But, as with the other once completed copies, at some point a cluster of Experience plates—in this case 11—were extracted, reducing Songs copy g to 42 plates. As demonstrated, reinserting the extracted plates in Songs copies a and i revealed their initial number and order of plates. While Songs copy g follows their production pattern, reinserting the extracted plates into the forty-two copy g impressions

30, 31 40, 32, 45, 44, 50, 48, 51, 37, 47, b

28, 29, [31], 38, 42, 34, 35, 36, 33, 49, 41, 39, 52, 54, 43, [44], 53, 46

yields, surprisingly, the plate order that Blake recorded in the “Order of the Songs” manuscript.

Tatham inherited Blake’s “Order of the Songs” manuscript and bundled it as part of a volume of “Blakeana” comprised mostly of illuminated prints and proofs printed by Blake and himself. The volume was acquired by George A. Smith, who had it bound by 1853. Only Songs copy V, produced c. 1818 and acquired by James Vine, followed the manuscript’s plate order. Vine met Blake through Linnell around 1822 and also bought Milton copy D, Thel copy O, and the illustrations to The Book of Job. He bought posthumous Jerusalem copy J from Tatham, which was sold at auction with Songs copy V and Vine’s other illuminated books in 1838.[23] While it is technically possible for Tatham to have ordered Songs copy g in this unique plate sequence, influenced or encouraged by Vine, it seems unlikely that he would have abandoned the standard order that he had been using. In fact, given that its pattern of production mirrors that of other abridged copies of Songs, copy g seems to have been re-ordered by an owner or dealer. Its first known owner was H. Buxton Forman, the bibliophile, editor of Shelley and Keats, and literary forger. He had the impressions mounted on linen stubs and bound in two volumes, Innocence and Experience, now known as Songs copies g1 and g2 (BB 427). He either received copy g in that two-part format or created it. He also owned Songs copy h, which has 57 plates, and had it similarly mounted on linen stubs and bound. The plates in copy h are also, minus one variant, in the “Order of the Songs,” which is not likely to be a coincidence, but neither is it necessarily the work of Forman.

In Songs copy h, plates 2 and 3 are transposed (i. e., “1, 1, 3, 2, 4, 6, 8 . . .”). The two frontispieces (plates 2 and 28) were inserted—presumably at the discretion of the binder—to face the title pages: plate 28 faces plate 29, but plate 2 faces plate 4 instead of 3.[24] Songs copy h is numbered 1-57 in pencil just below the lower left corner of the plates. Robert Essick, who owns the copy, notes that “this numbering was done when the copy was assembled as such” (WBA Copy Information for Songs copy h). But the copy g impressions are not numbered. This omission seems odd if both copies were ordered by the same hand. Could Forman have acquired copy h, the numbered copy, and used it to re-sequence the plates of Songs copy g?

Songs copy h stands out among the posthumous copies. Superficially, it resembles Songs copy a. Its impressions were printed in two inks, light black and the orange brown used in copy e. It is missing plate 15 but has plate b. Its leaves are the same size (28 x 19.3 cm) and type of J Whatman 1831 paper as the black and orange brown impressions that were added to copy a, and it has two leaves of the short (24 x 19 cm) thicker J Whatman 1832 paper used in copy a’s original forty impressions. Copy h’s impressions, overall, are equally uneven in quality. But the resemblance is superficial. Songs copy a has two sets of impressions, black and orangish brown, because an owner and not its printer reassembled it using impressions most likely from a miscellany of posthumous impressions. Copy a was not produced that way, whereas copy h was—though “produced” may be misleading. Songs copy h is a genuine composite copy, put together by Tatham from impressions initially discarded from two or three printing sessions.

Songs copy h, with 57 plates and three duplicate plates, is more than “complete.” Copy h has 31 plates in light and dark black inks, 25 impressions in light and dark orangish brown inks, and its last plate is in reddish brown, the color used in Songs copies b, c, and f/j. It has duplicates of plates 1, 52, and 53 in both inks. Innocence has 26 plates and Experience has 27 plates, and the general title plate was placed first. All but four of the Innocence impressions (18, 19, 24, 26) are in black ink; 20 Experience impressions are in orange brown. Sixteen impressions are on leaves with fragments of J Whatman 1831 paper (9 black impressions and 7 orangish brown), and two impressions are on J Whatman 1832 paper. According to Essick, “The 1832 watermarks” are in plates 43 and b, printed in orangish brown and reddish brown. He states that this watermark indicates “the earliest year in which this posthumous copy could have been printed. It was probably printed in that year or shortly thereafter. It is also possible that some plates were printed in 1831.” Essick is right on both counts; Songs copy h is comprised of prints from two or three different printing sessions, possibly some in 1831 and others in 1832. As we will see, they were compiled by Tatham in the 1830s and not by Forman or the binder, but it may have been resequenced by Forman or another owner.

The printer used two different inking styles. The orangish-brown plates were printed with their borders; most of the black prints were wiped of their borders, except where the borders were part of the design, as in plates 1 and 29. The black ink varies in density and tone (hence, it could be described as either “gray” or “black”) in the manner of the copy a impressions. The dark black blotches of ink in otherwise greyish ink indicate that the ink was black, again as in the copy a impressions, though on the thinner 1832 leaves. Blemishes from the shallows combined with light and uneven printed relief lines are common among both the black and orange-brown sets of impressions, but perhaps more overt among the latter (illus. 00). These visual effects suggest second pulls, that is, impressions printed from plates that were not reinked between pulls. Blake used this production technique to great effect in many of his color prints, large and small. These kinds of prints were fainter and consequently needed more finishing and touching up, which Blake took full creative advantage of, but Tatham did not and could not. For Tatham, pulling second impressions was an occasional gamble. The inking and printing in both sets of impressions are uneven, but in general the orangish-brown impressions have sharper texts than the black impressions, because the orangish-brown ink was slightly thicker and less oily than the black (illus. 00).

Songs copy h is a genuine composite copy mostly made up of discarded prints, i. e., proofs, second pulls, and just poor impressions from two or more printing sessions of Innocence (copy T) and Songs copies a, d, g, i, and p for the black impressions, and Songs copies b, c, e, and f for the reddish to orangish-brown impressions. None of the plates were extracted from copies a, g, and i, all of which were printed in black ink, because these plates in copy h were printed in orangish-brown ink. Nor are they from the Songs copy e, which, as noted, provided nine or ten of the added orange brown impressions to copy a.

Essick, who acquired Songs copy h in 1981, has suggested that Forman may have compiled copy h from loose impressions. He notes that Tatham “very probably printed this copy,” but thinks Forman “acquired at least 98 loose posthumous impressions of Songs of Innocence and of Experience at an unknown time, but probably in the final quarter of the nineteenth century, and bound them into three copies, g1 (23 impressions, now Princeton University Library), g2 (18 impressions, now Library of Congress), and this copy h” (WBA Copy Information for Songs copy h). At first glance, it does indeed look like somebody acquired a very large cache of loose impressions out of which he compiled Songs copy h and, from duplicate impressions, also compiled Songs copy g. This seems to explain the diverse make up of Songs copy h and perhaps the incompleteness of Songs copy g. As we have seen with Songs copy e, “completion” of one copy is often at the expense of another copy.

The idea that an owner, however, acquired a pile of 98—or more—loose impressions of the Songs, 71 of which were in black ink and 35 in orangish-brown ink and 1 in reddish brown and from it compiled a copy with impressions in two inks and three duplicates and compiled another copy in black ink abridged exactly like five other copies of which he could not have known, seems very unlikely. As demonstrated, Songs copy g was never a loose pile of impressions; it was initially 53 plates deliberately abridged to 42 plates by Tatham and apparently sold in that form, as, presumably, were Songs copies a, d, e, i, and p. That Songs copy h has impressions of varying quality from various printing sessions points to the printer as the compiler. Indeed, Songs copy h was deliberately constructed by Tatham out of spare parts, without the direct use of the press but with diverse remnants—some quite fine, most not—from other printings. Songs copy h reveals more of the contents of Tatham’s studio than all the other Songs copies—more of the diverse range and quality of Tatham’s printing. No one but Tatham could have constructed Songs copy h, for no one had on hand so many impressions from different printings.[25]

Essick is right, of course, that Tatham “very probably printed this copy,” but it is more exacting to say that he printed the impressions that formed this copy, because he did not print Songs copy h as a copy per se, in the way he printed the other copies of Songs. The other copies were the products of dedicated printing sessions. Tatham did not print Songs copy h to have two different inks and 57 plates. Moreover, for Tatham to have dumped 98 loose impressions on the market would have been completely out of character (see Viscomi, Printing Paintings Chapter 12). It might seem otherwise because 83 Songs impressions—in the form of copies e and i—and “50 very small plates” of No Natural Religion sold in the 29 April 1862 Sotheby that is thought to have been Tatham’s.[26] Most, if not all, of the Blake items in this auction once belonged to Tatham, but the vendor appears to have been an unknown collector who bought most of the material in “the portfolio of Blake drawings” that Tatham had sold to the printseller Joseph Hogarth and which was acquired but then returned by John Ruskin c. 1843 (see Printed Paintings Chapter 12). Assuming, however, that Tatham was the vendor has created the illusion that he still had substantial inventory in 1862 and had, indeed, dumped large numbers of unsold lifetime and posthumous impressions as a matter of course, making the idea of 98 posthumous impressions showing up on the book or print market perfectly plausible.

Indeed, the idea of a collector finding a cache of Blake impressions has precedent in Blake studies. Keynes described six copies of There is No Natural Religion as compiled from “A pile of leaves of There is No Natural Religion. . . . bought many years ago in a shop in Edinburgh” by Stopford Brooke and R. A. Potts (BB 444). These leaves, though, many watermarked 1811, proved to be facsimiles mistaken for originals (see Viscomi, BIB, chap. 21-22).[27] Blake himself can be thought of as having compiled copies from “piles of leaves” in that he printed a dozen or more impressions per plate when in 1789 he first printed Innocence and compiled copies upon sale, a mode of compilation that accounts for the different plate orders in each early copy (Viscomi, BIB, ch 00). At the end of his life, though, Blake was printing one, two, or three copies of a title, using the same inks and paper cut to different leaf sizes. In c. 1818 he printed in orange ink single copies of Marriage (G), Milton (D) and Urizen (G), as well as two copies of Songs (T and U) and Thel (N, O) and three copies of Visions (N, O, P). In 1821 he printed single copies of America and Europe for Linnell. In the last two years of his life he printed Songs copies W and Y at the same time but using different stacks of paper, one octavo and the other quarto; he may have done the same with copies X and Z; copy AA may have been printed by itself [CHECK]. If Tatham had a production model, it was late Blake, printing mostly per copy and at most three copies in a printing session. Tatham appears to have printed in an equally orderly way and compiled copies systematically.

Forman did not buy Songs copies g and h as an undifferentiated “pile of leaves.” That is mere conjecture inferred from his owning two separate posthumous copies of Songs, one of which seemed without order or intent. No evidence proves Forman bought them from the same source or at the same time. Forman had the leaves of both copies mounted on stubs and bound, retaining their original and diverse character—but probably not their original plate order. Retaining the authentic look and feel of the prints—instead of transforming the original artifact aesthetically by trimming images and laying them into uniform size leaves, as many collectors did—is not surprising for a literary forger who created “first” editions for Shelley, Keats, Browning, and others. Songs copies g and h were compiled by Tatham, but if Forman had a hand in reordering their plates, he did so after 1885. George Smith’s portfolio of Blakeana with the “Order of the Songs” manuscript was sold at Christie’s in 1880 to Quaritch, who “evidently sold it to William Muir,” who reproduced Songs plate b and the “Order of the Songs” as an “Appendix to his facsimile of the Marriage (a copy of which . . . he sent to the Editor of the Anthenaeum with a note on 28 Dec. 1885”) (BB 339).[28] The plate orders for Songs copies g and h were either the work of Tatham, who owned the “Order of the Songs” manuscript and knew the owner of the one copy of Songs in that order, or the work of Forman (or an owner before him) done after 1885 by consulting Muir’s reproduction. Because Songs copies g and h are the only posthumous copies of Songs with this plate order and both copies were owned by Forman, suspecting Forman of rearranging the plate order to realize Blake’s intentions seems reasonable.

If Forman was responsible for the plate order of Songs copies g and h, then these copies were probably once in the standard order that Tatham had used for his other copies of Songs. Forman rearranging the plate order, however, is not the same as his compiling Songs copy h from a pile of 98 loose impressions. I suspect that Songs copy h was compiled by Tatham from discarded impressions from different printings of the Songs before he abridged six copies and assembled the extracted plates as subsets of Experience. Had Tatham merely sold or dumped a loose cache of impressions, the cache would probably have been more random and had more duplicates and possibly not have been diverse enough for a completed copy. Indeed, as far we can tell, Tatham sold posthumous Songs impressions according to some rationale for grouping, from the completed copy, to an abridged form of that copy, to clusters of Experience plates, to loose impressions included in larger volumes or “scrap books” of miscellaneous Blake works (see Chapter 11) that Tatham assembled for sale. Songs copy g now appears much compromised, but it was not compiled randomly or by a dealer. Rather, its original state progressively degraded, reduced from 53 to 42 plates, sold, and passed through various hands till reordered and divided into two parts.

Conclusion

Unlike the other posthumously printed illuminated books, Songs copies a, d, e, g, i, and p appear incomplete and to have been haphazardly compiled or reassembled. But seen in the context of their production, patterns emerge and with them intentionality. They were all initially complete copies, with 53 to 54 plates, but abridged to 40 to 43 plates by the extraction of the same cluster of Experience plates. Songs copy a is unique among these copies because it was reassembled. It was reduced from 53 impressions in black ink to 40 and was given 10 orange brown impressions, presumably from copy e, as well as four black impressions from an unknown source. Chart II summarizes the production of the posthumous copies of Songs and the plate order of the copies:

Chart 2. Plate orders of posthumous copies of Songs

Copy b = Standard Order red-brown

Copy c = ? uniquely re-ordered red-brown

Copy f/j = Standard Order red brown?

Copy d = loose, black Abridged

Copy e = loose reddish and orange brown Abridged extractions possibly in Songs copy o and then to Songs copy a

Copy a = Standard Order black Abridged extractions

Copy i = Standard Order black Abridged extractions

Copy p = Standard order black Abridged extractions[29]

Copy g = ?Order of the Songs black Abridged extractions

Copy h = ?Order of the Songs black, org-br, rd-br. Composite

missing in e in o-b 30, 31, 32, 33, 37, 41, 44, 45, 46, 47, 50, 51, 52, 53

present in h in o-b 28, 30, 31, 32, 36, 37, 40, 41, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54,

missing in h in black: 30, 31, 32, 36, 37, 40, 41, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53

present in h in black 29, 33, 34, 35, 38, 39, 42, 52(b), 53(b)

missing g1-2 11, 15, 30, 31, 32, 37, 40, 44, 45, 47, 48, 50, 51, 52, b

Tatham, with access to both copies a and e and to their extractions, appears logically to have been responsible for reassembling copy a. But that seems very unlikely, because had Tatham reassembled copy a, he most likely would have reinserted the added impressions in their original places to retain the copy’s standard plate order, instead of tacking them randomly at the end of the abridged copy. As unlikely as it seems, Pickering, appears to have had lightening strike twice. He acquired abridged Songs copy a, with its forty impressions uniformly printed on thick 1831 paper, as an autonomous artifact. He also acquired copy o, which apparently was a miscellany of loose impressions in black, reddish brown, and orange brown inks. Pickering’s own Songs copy o appears to have been the source of the four black impressions and ten orange brown impressions used to complete copy a. At first, the idea that an owner could have completed copy a without leaving a depleted source copy—like “copy” o—seems impossible. But the odds of it happening weren’t so bad after all, because copy o was probably never complete to begin with, but an expansion of the Experience cluster of Songs copy e.

The Experience impressions extracted from six copies were apparently extracted for autonomous sale.[30] As unexciting or anticlimactic as this hypothesis sounds, the absence of technical and aesthetic reasons for abridging complete copies of Songs to 40-43 plates supports it. Moreover, just as a print evinces a press, the absence of new prints in the forms of abridged and composite copies evince the absence of a press. They signify the inability to produce more impressions. Indeed, systematically extracting plates from complete copies was a drastic step, one Tatham would most likely not have taken had he been able to print more impressions. As I will argue in Chapter 11, Tatham took that step because he lost the literal means of production, the rolling press, by the end of 1832 and was unable to print more impressions.

WORKS CITED

Bentley, G. E. Jr., Sale Catalogues of Blake’s Works: 1791–2015. http://library.vicu.utoronto.ca/collections/special_collections/bentley_blake_collection/sale_catalogue/