04 January, 2009

Recasting Caslon Old Face


The heading to the image above claims authentic historical origins for the type that is shown. Something like it was often used in presenting the Caslon ‘Old Face’ types to the customers of the typefounders H. W. Caslon & Co. The example is from a finely-printed large quarto specimen that the foundry produced in about 1896 in order to promote the type more widely. The title page reads, Specimens of the original Caslon Old Face printing types, engraved in the early part of the 18th century by Caslon I.
‘Caslon’ is an example of what became known in the commercial world of the 20th century as a ‘brand’: a family name that was not only widely recognised by customers but which stood as a guarantee of long-standing integrity. George Bernard Shaw had the editions of his plays set in the Caslon Old Face types on the recommendation of Emery Walker, the friend and adviser of William Morris. Printing-offices rooted in the principles of the Arts and Crafts movement, like the Dun Emer Press, later the Cuala Press, of Elizabeth Corbet Yeats, the Cranach Press of Harry Graf Kessler and the St Dominic’s Press in Ditchling, used Caslon Old Face. The printed versions of the Declaration of Independence of the United States having mostly been set in Caslon types (probably by necessity rather than choice, since there were more modern alternatives in use in the colonies), there was a comparable revival of interest in the face there towards the end of the 19th century. John E. Powers (1837–1919), who acquired a reputation as ‘the father of honest advertising’, had ‘a partiality, which became a fetish, for dressing up his advertisements in Caslon Old Style type. Rivals who imitated his make-up are said to have found great initial difficulty in telling a lie in Caslon Old Style’ (E. S. Turner, The Shocking History of Advertising. London: Michael Joseph, 1952. p. 134).
The preface to the new specimen, signed by Thomas W. Smith, the proprietor of the foundry, contains this passage:
‘The modest specimens issued by the first Caslon were quite inadequate to render justice to his work, and, admiration and demand for these remarkable founts being steadily on the increase, we venture to hope that the following quarto pages, showing in ample form the complete series, from Five-line Pica to Nonpareil, and, at the same time, giving an account of the life and labour of their originator, with the history of the Caslon Foundry to the present day, will be acceptable to the Literary as well as the Printing world.’
One must concede that the types, ‘engraved by Caslon I’, look splendid. Perhaps their impression on the highly-calendared paper is a little pallid, but the comfortingly familiar, old-fashioned shapes and the clarity of outline and the quality of their casting do honour both to the punchcutter and to the typefounder. But which punchcutter? That is not a simple question to answer, because this is how the same type had appeared in a specimen just a few years earlier, and then it did not look nearly so smooth, but rougher and less refined. This text will try to resolve the puzzle.

The story of the revival of the Caslon types at the Chiswick Press in the 1840s, recast by the Caslon foundry from original matrices that were still in their hands, is a familiar one. Advised, it seems, by Henry Cole, a figure who would be active in the organizing of the Great Exhibition of 1851, Longman, one of the major London publishing houses, had published the pseudonymous Diary of Lady Willoughby, printed in 1844 by the Chiswick Press in an elaborately imitated 17th-century style, both literary and typographical, using the Great Primer Caslon type, with its long s, that had been ordered for the printing of an Eton leaving present, a quarto edition of the Juvenal Satires. This, in the event, appeared in 1845. Works using other sizes of the Caslon types were printed by the Chiswick Press in 1844. In 1852, Longman published Thackeray’s Henry Esmond, a novel in the form of a memoir that purports to have been written in the early 18th century. It is set wholly in the Caslon Pica, using the long s. The printer was Bradbury & Evans. Anne Manning’s popular Maiden and Married Life of Mary Powell, afterwards Mistress Milton, published by Hall, Virtue and Co. in 1855 and printed by Richard Clay, set in the English size of Caslon type, using long s, was another exercise in pseudonymous historical typographical pastiche. And of course the Chiswick Press continued to use the types for the tasteful editions, notably of the rediscovered works of 17th-century high-church clergy like Herbert and Taylor, that were published by William Pickering. During the 1850s, then, the types achieved a discreet success as a choice for the publishing of nostalgic evocations of historical texts.

Title page of a showing of the Caslon Old Face types that was included as a section on tinted paper in many of the foundry’s specimens during the later 19th century and is sometimes found bound as a separate specimen.
During the 1850s two related events took place that left their mark on typography. Types that appear to be identical with those cast by the Caslon foundry in London appeared under the name of ‘old style’ in the specimens of three type foundries in the United States: John K. Rogers, Boston (The Boston Type Foundry), 1856, Peter Cortelyou, New York (Cortelyou & Giffing), 1857, and in the Typographic Advertiser in 1859, the promotional journal of L. Johnson, Philadelphia, whose foundry later became MacKellar, Smiths and Jordan.
The other event was the making of a type with the name of ‘Old Style’ by Miller & Richard, Edinburgh. This is the first page of the earliest known specimen, dated 1860:

The text, as can be seen, is knocking copy, playing on the unease of some clients with the irregular and unconventional appearance of the original ‘old face’ or ‘old-faced’ types, the term that appears in a specimen from the Caslon foundry that can be dated 1854, the earliest appearance of the type in a specimen that I have yet found:

Text sizes of the Caslon ‘old face’ type appear in the big new specimen book of the foundry that is dated 1857. The punchcutter of Miller & Richard’s type was Alexander Phemister, who emigrated to the United States in 1861. He was said by T. L. De Vinne, in the second edition of his Plain Printing Types (New York, 1914), to have begun his work on the type in 1852. If this was so, and if Phemister’s old style type became at all widely known at this date or shortly after, perhaps its name was adopted by typefounders in the United States for their versions of a type to which they were perhaps reluctant to give the name of the English artist of the 18th century who had cut it and the foundry which had recently recast it. Or perhaps, conversely, Miller & Richard borrowed the name from them for the improved version of the archaic ‘old face’ type that they had made. There is still uncertainty about the exact chronology of these events, but we know that the design of the bland and regular Old Style type produced by Miller & Richard was quickly copied by other typefounders in Britain and the United States, including H. W. Caslon & Co., and that it would become a generic typeface that was widely used by English-speaking publishers for literary texts (and to some extent, under the name of Mediäval, in Germany), ‘modern face’ types being kept for works of technology and information.
The middle years of the century were difficult ones for the Caslon foundry. It was put up for sale in 1846, but then withdrawn, one of its advertised attractions to buyers having been that it included ‘the original works of its founder, William Caslon, which have recently been much in request for reprints’. A strike in 1865, followed by a protracted lockout, sapped confidence in the management of Henry William Caslon, the last lineal descendent of William Caslon I.
Two years before his death in 1874, Caslon invited back a former employee, Thomas White Smith, who had left the firm at its low point in 1865. The effect of Smith’s energy as manager soon became apparent. In 1875, he set up a journal, Caslon’s Circular, to promote its products. A branch of the foundry opened in Paris. By the 1890s, he had become the sole proprietor of the foundry and was modernizing the firm to face competition from other foundries at home and abroad, and from the new Linotype machine. A new and well-equipped typefoundry, shown below, was built at Hackney Wick in 1900.

However the value to sales of the firm’s name and its traditions did not escape T. W. Smith, and his own sons, when they entered the business, were instructed to change their surnames from Smith to Caslon.
In 1878 Caslon’s Circular published an article with the title, ‘Hand-cast v. machine-cast type’. It opens with this text:
‘In one department of our venerable foundry may still be seen the old process of type-casting by hand, such as was in use nearly two-hundred years ago: indeed we may say such as was in use, with but little alteration, in the days of Caxton. Four or five old men, whose heads have grown grey in the service of the Caslons, bend over their melting fires, and with tiny spoons pour the fused metal into the quaint old moulds, jerking and swaying about with grotesque monotonous movement. They look very much behind the time in the midst of revolving wheels and clanging machinery turning out type at incredible speed. During recent years hand-casters have learnt machine-casting, only a few having been kept at the old process, for reasons which we shall hereafter explain. The art is not taught to new hands, and the consequence is that in a few years hand-casters and their art will be unknown. Machine-cast type can be easily distinguished from hand-cast. It is bright as silver; moreover it has a small round mark on its side near the face of the letter. On the other hand, type produced by the old hand-process does not look so bright, is not so sharp in its angles, and gutters or air-holes may be seen on its sides and foot. We venture to say, however, that beyond its appearance, which is certainly inferior to that cast by machinery, there is little or no superiority in machine over hand-cast type.’
The reason for publishing this explanation then becomes clear:
‘Most of the original old-face founts, for which a demand has sprung up within recent years, are still cast by hand, and we have been led to make the foregoing remarks on the hand-casting process through having received letters from purchasers of an old-face fount, drawing attention to what they concluded to be inferior workmanship. The face of some of the letters of these old founts is no doubt rough and inferior to the modern type in finish—but in finish only. Notwithstanding that the matrices from which they are cast are more than a century old, the type produced by them is not only excellent but unique. […] We may state that the demand for these original founts, instead of declining, as some have predicted, is steadily on the increase, and we are taking steps to improve them so far as smoothness of face is concerned, and to produce them by the machine-casting process, without altering their shapes in the least degree.’
There seems little doubt that the Caslon types that appeared in the United States in the 1850s derived directly from those cast in London, and the most plausible explanation is that they were cast from electrotyped matrices made from the newly-cast Caslon types. There is a persistent story that the matrices for the Johnson type were made with the consent of the Caslon foundry, and indeed possibly by it, a suggestion that is supported by Johnson’s reputation as an honourable man of business. How Rogers and Cortelyou got their copies is unexplained, but a suspicion of piracy is inevitable. (Rogers, like Cortelyou and Johnson and the Caslon foundry in London, included long s in the specimen texts, but mistakenly used it in place of f.)
In 1858 the Caslon foundry supplied electrotyped matrices for the roman of the English and Small Pica sizes of Caslon Old Face to Charles Whittingham at the Chiswick Press. He passed them on to William Howard, the punchcutter and typefounder who had made its Basle and Caxton types during the earlier 1850s and who was no longer capable of such demanding work. Howard, who appears to have died in 1864, cast type from them by hand for filling the cases at the Press. The matrices survive among the materials of the Chiswick Press at the St Bride Library. Those for the English lower case were adapted for machine casting. Some examples are shown below. Note the different colour of the electrotyped part, inset into the copper of the matrix.

Electrotyping, that is the growing of a copper shell from an impression of typeset matter, which could be backed up with metal and used to print from as a substitute for cast stereotype plates, was invented in about 1840 and spread rapidly in the printing trade. The use of electrotyping to make matrices from cast type was the subject of US Patent 4130 of 1845, granted to Thomas Starr. By the 1850s, the electrotyping of matrices had entered the normal practice of typefounders. Increasingly, later in the century, punchcutters turned from cutting their designs in steel – especially the more elaborate ones – towards making them in typemetal, from which electrotyped matrices could be grown. The practice and its historical background are well described in detail by Roy Rice. Unlike the original matrices that were designed for use with the hand mould, electrotyped matrices could be shaped to work with any of the new typecasting machines that were developed during the second half of the 19th century, and by preserving sample types, the founder could generate any number of identical replacements for matrices that suffered wear or damage. But the process was the cause of unease among the major founders, since an unscrupulous rival could make an undetectable copy of a type from a small fount that had been bought commercially. The Caslon foundry was a strong and vocal critic of this practice.
The modernizing of Caslon Old Face was studied in detail by Justin Howes, who was able to spend some time at Stephenson, Blake in Sheffield before all the foundry’s punches, matrices and specimens were acquired for the Type Museum in 1996 and moved to London. He published his report on what he found as ‘Caslon Old Face: an inventory’, which appears as an inset in the article that he wrote for the journal Matrix, no. 20 (2002). It is the result of long and painstaking work, and it throws a great deal of light on the reworking of the smaller sizes of the Old Face types. His conclusion was that a process of remaking the Caslon Old Face punches took place from around 1893. This was the date of the first recutting that he found recorded in the Punch Notes, the documentation kept at the Caslon Foundry. The size was the Great Primer, now cast on 18-point, which was the work of Emile Bertaut. George Hammond, another punchcutter, took over where Bertaut left off, and was responsible for most of the recutting by hand of other sizes that took place between October 1894 and 1908. Later punches for revised characters were mostly machine-cut.
In the light of what he had put together about the state of the ‘Caslon Old Face’ that was cast during the 20th century, Justin went on to make his own digital version of the type, Founder’s Caslon, taking it back where he could to original forms, and purging it of some of the anachronistic characters that had been introduced when the types had first been recast in the 19th century. (These characters are also, incidentally, to be seen in both the Cortelyou and Johnson ‘Old Style’ types, making it abundantly clear what their direct source had been.)

Smith – the article of 1878 in Caslon’s Circular must be his – had been quite frank about the reason for reworking the Old Face types. It was simply no longer practicable to continue to cast any substantial part of the output of the foundry by hand. But there is no evidence – and this I find puzzling – that the expedient of making electrotype matrices from existing types was resorted to, at least not on any significant scale. Perhaps the original matrices had deteriorated too far. (Where are they, by the way?)
In fact a substantial move towards achieving the ‘smoothness of face’ that was promised in 1878 had undoubtedly been made by the date of a specimen book of about 1884, when on a page that shows the four biggest sizes, the ‘Two-Line Double Pica’ (which would later be cast on a 42-point body), a type that first appears in a specimen in 1742 and which is in fact the work of William Caslon II, still shows the irregular lining of type hand-cast from original matrices. But the first three sizes are now ‘smooth’.

Moreover the same specimen includes a slip showing swash italic capitals based on a 16th-century model that had been added to the type. The wording is studiedly vague. One could read it as suggesting that the matrices for these sorts had come to light among the many treasures of the foundry. They had in fact been newly and very expertly cut. (An account of them in Caslon’s Circular, intended for printers, is more frank about their origin.)

The first full presentation of the newly made-over and presumably machine-cast type to printers was in 1890, when a four-page showing of all sizes of the Caslon Old Face roman types, in which each page was headed, ‘Original Caslon Founts’ was given in Caslon’s Circular. The public relaunch of the new ‘smooth’ Old Face took place with the issue of the specimen of 1896, directed at ‘the Literary as well as the Printing world’, in which all the large sizes, including the Canon (the roman lower case of which was Joseph Moxon’s type of the late 17th century), which had looked so crude in the earlier specimens of the ‘ancient types’, were now irreproachably regular in their appearance. The inescapable conclusion is that they have been recut. Here is the lower case a of the Five-line Pica, in the old and new castings:

The image in the older impression is distorted to some extent by heavy inking, and the defects of hand-casting are evident, but it is clear that in the new type the opportunity has been taken to improve the form of the letter. In fact we have proof of the extent to which the whole type was altered. A album from H. W. Caslon & Co. has survived from the 1890s which gives synopses of newly-cut types. One of these, shown below (it is unfortunately neither dated nor annotated), has what are clearly the old and new versions of the Five-line Pica or 72-point size of Caslon Old Face, with the new version, in which the tidying-up can clearly be seen, above the old one. Serifs are more even and regular, the deviation of long s from the vertical is corrected, and the weight of strokes generally has been made more uniform. The corrections of anomalies are not overdone: the ascender of d still does not line with that of b, and j is far too short, but the overall impression is that of a type just a little too beautifully remade by a highly-skilled punchcutter of the 19th century. (It seems to me that Matthew Carter’s ‘Big Caslon’, 1994, based on these large sizes, especially on the 4-line Pica which was later cast on a 60-point body, manages to preserve more of the energy of the originals.)

The inventory compiled by Justin Howes was only a start, as he was well aware, and some entries raise questions that only a careful examining of the surviving materials can begin to answer, something that is hardly possible in the present state of the Type Museum. There are, for example, 48 surviving punches for the Five-line Pica, but only two of these seem to be original. Justin Howes writes that the ‘hand-cut punches for the remaining authentic sorts presumably date from the nineteenth century’. There are also machine-cut punches for a further 28 characters. Are the hand-cut punches those that were made for the revised type that is first seen about 1890? It seems likely, since although the 149 surviving matrices are largely ‘punch-struck’, he does not suggest that these date from the 18th century.
As it happens we have a small piece of more accessible evidence that became detached from the Caslon materials, having apparently at some time formed part of a display for exhibition: a set of four original punches, for K O U and m, for the Four-line Pica, later cast on a 60-point body, together with matrices for these letters. They are now in the St Bride Library.
It can hardly be doubted that these are the original 18th-century punches, in poor condition. Here is the face of m. The width, from one extremity of the foot serifs to the other, is 15.5 mm.
Here are impressions of the type, before and after recutting.
The second counter is slightly narrower than the first, and its upper curve is higher. In the earlier impression, on the left, the initial stroke aligns with the first of the upper curves. In the new type, on the right, these features have been kept, and the oddly-angled ends of the central foot-serif have been preserved, but it seems clear that the drawing of all the parts is more accurate. Moreover the first vertical stroke now rises above the line of both subsequent curves. It is the earlier impression that matches the original punch.
If there could be any doubt about the suggestion that the type was recut, the struck matrices that accompany the old punches confirm that something of the kind took place. The old punch and the new matrix do not fit together, but rattle uncomfortably when one is placed in the other. Here is the matrix for the Four-line Pica m, made for machine casting, and stamped with the code for its character number (47), the point size (60) and the name of the type, OF for ‘Old Face’.
The right-hand letter in the pair of m’s shown above is from the specimen printed in London for H. W. Caslon & Co. Ltd in 1924 by George W. Jones, which is one of the most elaborate and carefully-printed presentations of the Old Face type that the foundry ever produced. This claim, which forms the ‘unique selling proposition’ for the product (to use another piece of 20th-century marketing jargon) is made on the title page:

These words, an echo of those that had appeared in many specimens of the Caslon foundry, were clearly designed to sustain the faith of their customers, among whom were so many devoted craft printers, in the genuineness of types that bore one of the most respected names in typography. But the suggestion that the matrices of the type had been ‘produced’ directly from the original punches was now wholly misleading. The type was now cast from matrices made with new punches, and the direct link with the originals had been permanently and deliberately broken. Sources
The list by Justin Howes, ‘Caslon Old Face: an inventory’, is an eight-page insert in his article, ‘Caslon’s punches and matrices’, Matrix no. 20 (2000), pp. 1–7.
Here are some other related sources. The Caslon types as they appeared in the 18th century can be seen in the specimen book of the foundry published in 1766, reproduced in facsimile in Journal of the Printing Historical Society, no. 16 (1981/2).
G. W. Ovink, ‘Nineteenth-century reactions against the didone type model’, Quaerendo, vol. 1 (1971), pp. 18–31, pp. 282–301; vol. 2 (1972), pp. 122–43, is a series of articles, the first of which is the most wide-ranging survey that has been published of the appearance of ‘old face’, ‘old style’ and ‘elzévir’ types, in the 19th century. Similarly, A. F. Johnson’s survey of the English scene, ‘Old-face types in the Victorian age’, which originally appeared in the Monotype Recorder in 1931, and which is incorporated in his Type designs, their history and development, third ed. (London, 1966) and in his Selected essays on books and printing, 1970 (pp. 423–44), though much in need of updating, is the most thorough account yet attempted.
For details of the revival of Caslon Old Face, the account by Janet Ing (now Janet Ing Freeman), based on work with the surviving accounts of the printer as well as the books and other materials, is the most detailed study: ‘Founders’ type and private founts at the Chiswick Press in the 1850s’, Journal of the Printing Historical Society, 19/20 (1985–7), pp. 63–102. I am most grateful to her for guidance to the sources for the history of the Caslon Old Face matrices used by William Howard that are illustrated above. In her article she makes the suggestion that the early appearance of Caslon Old Face capitals in a set of five title pages that were proofed in 1839, of which an example is shown below, some four years before the setting of Lady Willoughby and other related texts from newly-cast type, may be due to the finding by the younger Charles Whittingham of old Caslon types in his uncle’s cases when he took over responsibility for the shop. This seems highly plausible – and if this is what happened, perhaps it was the discovery of the old types and their use in these few books that set off the whole revival.

The image of the new Caslon foundry at Hackney Wick shown further above is from an album made for a member of the Caslon-Smith family. It was bought by the St Bride Library from the book- and print-seller Ben Weinreb, who generously added the Caslon ‘Synopsis book’ from the same source as part of the deal. I published some of the images in 1993: James Mosley, ‘The Caslon foundry in 1902: selections from an album’, Matrix 13 (1993), pp. 34–42.
ATF Caslon
Here is the presentation of the series known as Caslon 471 in the Specimen book and catalogue of the American Type Founders Company issued in 1923:

This curiously opaque statement appears to suggest that the original Caslon matrices were brought to the United States. It fails to mention that in 1859 the type had already been cast in London by the Caslon foundry for over a decade from early matrices, and that it continued for many years to cast the type from the same matrices. The reference to matrices ‘brought to America’ in 1859 seems likely to be to electrotyped matrices imported by Johnson that were possibly made in London and perhaps by the Caslon foundry itself. If these matrices had been derived from the types that were being cast from surviving matrices, it can be argued that by the 20th century the Caslon types cast in the United States had a closer relationship to the 18th-century originals than the recut types that were being produced in England.
Electrotyped matrices deriving from MacKellar, Smiths & Jordan for the Small Pica size of the Caslon type, later 11 point, survived the break up of ATF in 1993. See the account and images given by Theo Rehak at the web site of the Dale Guild Type Foundry. I learn that they have now been bought by Rich Hopkins.
I should like to express my thanks to Steve Saxe and Alastair Johnston for help with the documenting of the versions of the Caslon types that appear under the name of ‘Old Style’ in the USA. Last edited 4 July 2009