16 February, 2007

Scotch Roman

William Miller, A specimen of printing types. Edinburgh, 1822
In his Typographia, 1825, p. 361, T. C. Hansard wrote that the English punchcutter Richard Austin ‘executed’ most of the founts of the two Scotch foundries of Alexander Wilson in Glasgow and William Miller in Edinburgh. In his Practical hints on decorative printing (1822) William Savage had made the comment that Miller’s types ‘so much resemble [Wilson’s] as to require a minute and accurate inspection to be distinguished.’ William Miller, formerly foreman to Alexander Wilson, appears to have set up his foundry in 1808 and issued his earliest surviving type specimen in 1813. Wilson’s ‘modern cut’ types appeared in a specimen dated 1812.
Although Hansard was a careful writer, it is difficult to know how much reliance to place on his unsupported statement. Richard Austin, a professional punchcutter and engraver in London, cut types during the 1780s for John Bell’s British Letter Foundry, and in 1806 he cut the ‘Porson’ Greek type for the University Press at Cambridge from the writing of the scholar Richard Porson. He set up his own foundry in London, the Imperial Letter-foundery, from which he issued the specimen dated 1819 with its critical preface that is quoted in a previous post. Even if he may have had some involvement with one or both of the Scotch typefoundries, it seems unlikely that he would have had the time to cut many of their punches if he was also making types at the same time for his own foundry. Moreover, although there is indeed some similarity between the Miller and Wilson types, none of Wilson’s has anything resembling the distinctive capitals of the Miller Pica No. 2 that is shown above, with a detail below, which became the model for the Pica Scotch Roman.

Detail of William Miller’s Pica No. 2, as shown in 1813, from the specimen of 1822.
The term Scotch Roman originates in the United States towards the end of the 19th century. It probably has something to do with Scotch-face, the name given to some types of the typefounder S. N. Dickinson in Boston that are said to have been made to his own design by Alexander Wilson & Son in Glasgow, and which were first cast by him in 1839 with matrices imported from Scotland. Other founders made similar types with this name. They do not in fact look much like the Miller and Wilson types of the second decade of the 19th century. De Vinne tells the story in his Plain printing types (1900), defining the Scotch-face as ‘a small, neat, round letter, with long ascenders, and not noticeably condensed or compressed.’ He adds that James Conner of New York appears to have shown the first complete series of the face, and gives this example:

The types that became known as ‘Scotch Roman’ are derived from a quite different model from this one. In 1882 two sizes of the early types of William Miller, a Pica and a Long Primer, were recast by the Miller foundry in Edinburgh, by now known as Miller & Richard, for a reprint of Sir Walter Scott’s edition of the works of John Dryden, originally printed by James Ballantyne and published in 1808. The new edition was printed by T. & A. Constable of Edinburgh, who according to a little history of the firm published in 1937 appear for a time to have had exclusive rights to the recast type.

Founts for five sizes of these types from Pica to Brevier (12 to 8 point) were bought from Miller & Richard in 1901 by the University Press at Oxford, where the type, as shown above, was known as Dryden. The same sizes were in due course marketed by Miller & Richard under the name of ‘Old Roman’. Several recut sorts had been included in this recasting, perhaps because the old matrices were missing or perhaps because the older forms did not satisfy contemporary taste. A new, wider capital S was substituted in the Pica, and in all sizes t was given a flat top in the French, rather than the English, manner. The Pica size, the Pica No. 2 of the early Miller specimens with its bold capitals, has sometimes been picked out as characteristic of the whole range of sizes, although the capitals of the smaller bodies are more delicate. The contrast in the Pica, often noted critically, between the weight of the capitals and the lower case characters was made more extreme in the revival by mistakenly choosing one of the heavier sets of capitals for revival among the matrices for the Pica italic and casting a light lower case for it.
The Miller & Richard types with their telltale added modern sorts were cast and sold under the name of Scotch Roman by the A. D. Farmer foundry of New York from about the beginning of the century, though the date is uncertain, and whether by arrangement with Miller & Richard or by opportunistically making their own electrotype matrices from imported types is not known.

The Farmer foundry added some larger sizes of its own that appear to have been crudely redrawn from the existing design. T. L. De Vinne showed them in his firm’s specimen book of 1907, and his words are worth quoting: ‘This Scotch Roman, as it is now called, was a contribution made to novelty made for and first used by the Ballantyne Printing House in Edinburgh, in the first decade of the nineteenth century. Its most striking peculiarity to the inexpert is the greater breadth and openness of the letters without appearance of undue obesity. It has no eccentricity save the almost unnoticeable flat top to the lower-case t.’
A type that appears to be Farmer’s was chosen in 1903 for setting the text of the new journal Printing Art. The design was followed in the same year for the Scotch Roman of the Mergenthaler Linotype Company, and also by the Monotype Company in Philadelphia for its Scotch Roman, Series 36. At about the same date ATF introduced the design from 6 to 12 point under the name of Wayside, a name that suggests the cooperation of Will Bradley, who was responsible for their series of ‘Wayside Ornaments’. In the specimen of 1912 a full series of Scotch Roman was introduced, perhaps derived directly from materials from the Farmer foundry that had belatedly been joined with ATF. Unlike Wayside, which continued to be produced under that name, it has descenders that, like those of the US Monotype version, are brutally truncated to make them fit the foundry’s ‘standard line’. Perhaps one should add that about the same date the H. C. Hansen foundry, Boston, made a version of the Miller & Richard type, which it marketed as National Roman.
The flat-topped t of the revived ‘Scotch Roman’, to which De Vinne refers, shows that the model for all the new American ‘Scotch Roman’ types that appear around 1900 was the revived version cast in Edinburgh by Miller & Richard, since this form of t is not seen in the early Miller types, and very rarely appears in British modern face types. The acceptance of this form in the US seems in its turn to have induced the designers of some other types to add flat tops to the t of some types related, however distantly, to ‘Scotch Roman’. Dwiggins’s Caledonia for Linotype is an example. Even more oddly, ATF’s Bodoni (1911), the model for countless other types with this name, with its flat-topped t may possibly be another example of the influence of the Miller & Richard ‘wrong font’ character, since the top of Bodoni’s own t was generally pointed.
Since he made distinguished use of the type, it is perhaps worth quoting D. B. Updike’s remarks on ‘the modern face known in America as Scotch’ (Printing Types, ii. 230). However, it is an oddly confused piece of writing:
‘Perhaps the most beautiful version of it ever brought out was that cut by William Martin; and a very close copy if not actually the same face was produced in Scotland in the last century—notably in the ‘Series of Old Founts’ by Messrs. Miller & Richard of Edinburgh. The Wayside series of the American Type Founders Company, if in its original form, with long descenders—is a fairly satisfactory equivalent.’
The types cut by William Martin for William Bulmer are so very different from those of Miller that the use of his name here is difficult to understand, unless Updike was confusing it with that of Richard Austin. And the Miller & Richard types were called ‘Old Roman’, not ‘Old Founts’. Nonetheless the endorsement of the Wayside types suggests that these may have been the founts that Updike himself used with distinction.

In 1906 the English Monotype Corporation made a robust but rather clumsy adaptation (above), based on a single size of the Miller & Richard types, the Pica, but with the capitals made more uniform in width. This type, Series 46, was known simply as Old Style. It became so widely used in Britain that it was often taken to be the normal form of the type. It was used for the edition of Jane Austen’s novels, edited by R. W. Chapman, that was published by the Oxford University Press in 1923. This was a pity, since Series 137, a more faithful recutting of all the different sizes of the Miller & Richard ‘Old Roman’, including the delicate Small Pica and Long Primer on 11 and 10 point bodies, was made by Monotype at the request of the Edinburgh printers R. & R. Clark in 1921:

Like Series 46 it was called simply ‘Old Style’, but eventually (and confusingly) both types were given the US name Scotch Roman. The sales of Series 137 were never seriously promoted (these were the years of Garamond, Baskerville, Fournier and Bembo), and very few other British printers bought matrices. (The larger sizes of Series 137, on the analogy of those made in the USA, were enlarged from one of the smaller ones (the 14-point in this case) and were not based on original Miller types for these bodies.) Monotype’s digital version was based on its mediocre Series 46, and the opportunity to make a better Scotch Roman was missed until the appearance of Matthew Carter’s Miller (1997), based on the early Miller types.
My own digital New Scotch (below) is derived from the Long Primer (about 10 point) size of the Miller types. I am well aware that the spacing, alignment and kerning are all very far from perfect and the outline looks pretty rough, but I have been agreeably surprised to see that it is robust and readable in high-resolution typesetting. The @ character (the earliest that I have found in type) is from the Miller specimen dated 1822. Monotype cut the Miller italic w that is seen here for its Series 137, and it is seen in some printing with this typeface, but their nerve failed them and they later substituted a more conventional character. The fist is adapted from one that I once cast in a Figgins matrix.

Footnote 1
The text above is derived from a note that I originally wrote in 1972 for Brooke Crutchley, University Printer at Cambridge, when a public edition was being prepared of the text by Stanley Morison that had been privately printed in 1953, A tally of types cut for machine composition and introduced at the University Press, Cambridge 1922–1932. I had suggested to him that, since the new edition was not limited to types acquired by the Press between 1922 and 1932, a note on ‘Scotch Roman’ should be included. Modern-face types had been great favourites of the former University Printer, Walter Lewis. He welcomed the idea.
The new edition of the Tally was published in 1973. For some time afterwards I thought that the text I had sent to Crutchley had not been used. I took this philosophically until one day, looking rather more carefully at Morison’s preface, as ‘revised and amplified by P. M. Handover’ (his research assistant at The Times), I found every word of my note, without attribution, on pages 27 to 30 of the new edition. I would not labour the point if it were not for the possibility that to anyone who reads this passage in the preface I might appear here to be plagiarizing it.
A recent history of printing types suggested that the account of ‘Scotch Roman’ in this preface of 1973 (there had been no reference at all in Morison’s original preface) has ‘what remains the best description of Miller’s finest Modern’. Whether this is the case or not, the words are neither Morison’s nor Handover’s. Morison had, however, written a brief account of what he called ‘Scotch Roman’ in 1935, and this is published in Footnote 2 below. I notice with mixed feelings that my narrative in the Tally was evidently one of the sources of Alexander Lawson’s worthy but rather garbled account of ‘Scotch Roman’in his Anatomy of a typeface (1990), which reproduces my words at some length, something that it seems all the more worth noting now that parts of his text have turned up on Google Books, which makes the story more than ever in need of the additions and corrections that are made in the present post. The 1973 edition of the Tally was reprinted by David Godine, Boston, with the addition of an excellent introduction by Mike Parker, in 1999.
A revised version of my original essay was published by Alastair Johnston in Ampersand: quarterly journal of the Pacific Center for the Book Arts, San Francisco, autumn/winter 1998, pp. 2–11.
There is a useful note on ‘Historic Scotch Roman: a design originated by an American Printer?’ in issue 24 of Richard L. Hopkins’s ATF Newsletter, with a ‘Follow-up to Newsletter Article on Scotch Roman’, including a facsimile of a sheet of Specimens of Book and Job Faces issued by the Dickinson Type Foundry, in issue 26 of the Newsletter. In this instance ATF, with its echoes of the initials of the American Type Founders Company, stands for the American Typecasting Fellowship, set up by Hopkins and other enthusiasts in 1978.

Footnote 2
This is Morison’s own account of Scotch Roman, published in 1935, which seems to have been written with some knowledge of the version of the story that was believed at the Edinburgh printing firm of T. & A. Constable. The paragraph that follows it is an extract from Constable’s little house history, issued privately in 1937. Morison’s reference to the date of 1808 is clearly derived from the story believed by Constable, and refers to Scott’s edition of Dryden. But this, as noted above, is not set in any of the types that appear in the Miller specimens, and the claim that they were cut in 1808 was baseless.
‘In 1808 Messrs. Miller & Richard cut a new fount whose virtues seem to have been completely overlooked at the time. This design, now known to the trade variously as Scotch Old Face or Scotch Roman, would have had a fair claim to rank as the Scottish National Face if circumstances had been otherwise. But in the fifties the Miller & Richard foundry brought out ... a sort of diluted version of Caslon’s Old Face known to the trade as ‘Revived Old Style’. The Scotch Roman, cut originally in 1808, was laid aside immediately, and never given the consideration it deserved until Messrs T. & A. Constable secured its exclusive use and printed a number of very handsome books in it, for some years after 1882. The Scotch Roman, though a modern face within the Bell-Austin tradition, possesses an individuality distinguishing it from the other members of the group. The fount, released by Messrs. Constable, is much used at the present day, being suitable for working upon smooth and coated papers. As these words, set in Scotch Roman, prove, the fount possessed boldness of character and orderliness in the disposition of colour.’
Stanley Morison, Introduction to W. T. Berry and A. F. Johnson, Catalogue of specimens of printing types by English and Scottish printers and founders 1665–1830 (London, 1935), p. xliii. The last sentence of Morison’s text was set in Monotype’s Series 46. As can be seen he makes no reference at all to the more faithful version, Monotype Series 137, cut at the request of R. & R. Clark in 1921. His reference to its ‘release’ by T. & A. Constable suggests that the latter firm still had a notion that the type was in some sense their property, something that seems odd, given its wide adoption in the USA.
‘During the next twenty years there was a great development in the typography of the firm. The partners were dissatisfied with the types then available, and on exploring the possibilities of improvement they discovered a fount of type which had been cut in 1808 by Miller & Richard, the Edinburgh typefounders. For some reason the possibilities of this type had not been realised, and the sales had been slight. The firm of T. & A. Constable were granted for a period of years the exclusive rights to this type, and the first books to be printed in it were the Works of John Dryden, with notes, and a life of the author by Sir Walter Scott, edited by George Saintsbury. As a result of this the type was known as Dryden, and other books printed in this fount were the Edinburgh Edition of Stevenson, Carmina Gadelica, and the Edinburgh Folio Shakespeare.’
Brief notes on the origins of T. & A. Constable Ltd (Edinburgh, 1937), p. 9.

Footnote 3

In 1888 T. & A. Constable, who had printed Scott’s edition of Dryden in 1882, wrote the note shown above to the London typefounder Sir Charles Reed, asking for help in locating the matrices of the ‘New Long Primer No. 5’, a type from the Wilson foundry that had been used – so they claimed – ‘for the original edition of the Waverley Novels, printed by the Ballantyne Press’. The materials of the Wilson foundry had been sold at auction in 1845, and it seems likely that this request was related in some way to the revival of the Miller types that is discussed above. The first of the ‘Waverley Novels’ of Sir Walter Scott was Waverley, or ’tis sixty years since, published in 1816, and they were indeed printed in Edinburgh by Scott’s friend James Ballantyne, originally of Kelso. The first edition of Waverley was not set in the Wilson type named above, which does seem to have been cut later, but it is possible that it was later adopted. Nor were the early Miller or Wilson types used for Scott’s Dryden of 1808. But the document (which was kept by Talbot Baines Reed in his interleaved and annotated copy of his History of the Old English Letter Foundries, 1887) does suggest that the revival of the early Miller types may have encouraged more attempts to recreate the appearance of Scott’s works.

This text has been much revised and expanded since it was first posted, but the story it tells is so involved and the published versions are so confused and contradictory that it seems worth doing this as thoroughly as possible. Moreover, thanks to many helpers and contributors, much new material has come to hand. The last revision was on 14 February 2009