Digital Type Foundries Respond to Letterpress

This is the article, “An Affinity by Design: Digital Type Foundries Respond to Letterpress” (sans images), which was published in Parenthesis: The Journal of the Fine Press Book Association, Number 7, November, 2002. Reprinted here with minor post-publication editorial corrections.

Copyright © 2002, 2005 by Gerald Lange.

The appearance of the work itself is of more importance than any quibble over the method of its translation into the vehicle of thought, since its legibility or beauty is determined by the eye and not by the means employed to produce the type. —Frederic Goudy

For well over the past decade fine press printers have increasingly turned to the photopolymer plate process as the alternative source for metal type. In recent years the productive use of metal type in the overall studio-letterpress community has declined, but to a great extent this is no longer due just to dwindling resources. While not all would agree, the advantages and capabilities of the computer in rendering traditional typography, along with the high printability of photopolymer plates, have proved the process beyond dispute. Residual resistance to the digital/photopolymer process is now less and less based on its technical capabilities, but more on the technique itself. It violates the romanticized view of historical letterpress. This is its cardinal sin, for which there is little forgiveness.

What plagues the photopolymer process is not these considerations, but the nature of digital type itself. It is not that digital type is inferior to metal type, either technically or aesthetically, but that it is simply not designed for the relief process. (Many metal typefaces it should be noted, do not perform well when printed letterpress either. This is self-evident to any seasoned printer, but rarely acknowledged.)

The printer of photopolymer plates, however, does stand at a unique disadvantage to the printer of metal type. The physical formation of the photopolymer plate letterforms—though facilitated by the technical processes that are available to the printer—has not been undertaken at a previous stage in the technology. Metal type is cast at the foundry or on the printer’s casting machine and is im-mutable. The printer works with the physical form of the typeface he or she has been given. The processing of photopolymer plates is much more fraught with the possibilities of disruption at any intermediary step. Even if these steps are taken care of exactingly, the printer is still dependent upon the very first step in the process: the proper configuration of a digital typeface. Digital typefaces face two age-old problems for which they are not equipped: ink spread and lack of optical size compensation.

These problems also affect metal typefaces. With the invention of the pantograph engraving machine in the late nineteenth century, for instance, twentieth-century typefaces themselves had significantly reduced optical size ranges. Typeface designs were scaled to fit a range of group sizes. Ink spread, a thickening effect of the letterform that is unique to relief printing, is caused by impression and accumulating ink gain. This has always been a persistent demon to be controlled by the concerned printer, but the term itself came into currency more as a consequence of twentieth-century revivals of historic typefaces. Typeface designers have traditionally viewed it as an obscuring effect than a problem to be dealt with technically. The punchcutter’s view was perhaps the same. In his Manuel typographique (1764), the great typefounder Pierre Simon Fournier declared, “It is not right to blame the letter for the fault of the ink.”

The digital environment is quite complex and a technically sound typeface must perform well under a variety of diverse conditions: viewing devices from the small hand held to the large screen, printing devices from low resolution dot matrix to high resolution imagesetter. It has to work with varied forms of print delivery systems; inkjet, laser, film-based. It has to function on screen, as print, in electronic transmission. It must set well in the lowliest of text-based programs as well as perform the typographic requirements demanded by sophisticated page-layout programs. Not all digital typefaces function in all these aspects, but to some extent professional design practices ensure such capability in the more sophisticated releases. The evolutionary process of digital type has not been as rapid as one might assume. A good part of the latter third of the twentieth century went into its development, and at every step of the way it had to satisfy parallel developments in electronic technology. At no point in this evolution was any serious consideration given to the unique restraints of the letterpress process, nor should it have been expected.

Digital typefaces also inherited the dubious typographic conventions of the preceding printing technology, that of photomechanical typesetting. With a notable exception or two, photofilm did away with optical size ranging entirely. In following this practice, digital faces are traditionally offered in only one point size, usually 9- or 10-point. All other representations are then extrapolations of this core size. Thus 72-point could actually be 10-point scaled to 720 percent. For a typographer this is problematic in that there is little harmony in weight between the sizes and as a face increases in size it just gets thicker and appears bulky and there is no built-in compensation for this. For legibility, smaller sizes need a large x-height, increased weight and width, opened counters, and wider settings. Larger sized faces need a reduction in weight and width, with serifs refined to a visually pleasing thinness.

Hand punchcutters from the mid-fifteenth to the late nineteenth century are thought to have redesigned a typeface as a matter of course, on a per size basis, building in a natural optical compensation as they graved and filed the letterforms on the ends of the steel punches. Though the pantograph engraving machine allowed this practice to whither away, there was still an attempt throughout the first half of the twentieth century to build in certain optical size ranges in metal type designs. With notable exception, firms selling machine composition matrices commonly restricted their ranges. Monotype often provided only a small text range (6- to 10-point), a text range (12- to 14-point), an intermediary range, and a large size range. Some firms, such as Linotype, had more diverse ranges, but not by much. With the introduction of photofilm there seemed no point in continuing the practice as typesetters could just scale sizes photographically, and, were not willing to pay for ranging. Why buy four versions of a typeface when you could buy four typefaces? Without client resistance there was no need to modify the practice.

The development of digital type followed the same path but, oddly enough, took more seriously the gift the past had handed over to it. The first digital typefoundry, Bitstream Inc., was co-founded by the well-seasoned type designer, Matthew Carter, who as a young man cut punches at the Enschedé foundry. When Adobe Systems Inc. began to manufacture digital typefaces they brought on board individuals who had serious interest in letterforms; professional calligraphers, several with ties to the fine press community. Adobe began to offer original designs that had a historical connection to metal type and they patterned the best offerings of our typographic history, including character forms that were rarely seen in the photofilm years: small capitals, oldstyle figures, alternate characters, and ornaments. Most exciting of all, they revived the type specimen book. Adobe had a strong influence on the industry, as it had also developed one of the more sophisticated font formats, PostScript Type 1.

A significant development at Adobe was its eventual line of multiple master typefaces, the first of which was released in 1991. These were unique in offering optical size ranging. A multiple master face came equipped with one or more axes. These axes were primary redrawings of variance, which Adobe called a dynamic range. A weight axis would carry a lighter version of a design as well as a heavier version. Adobe also provided instances of redrawings at selected intervals between the primaries. Thus a multiple master weight axis might carry a number of weight variations built into the face. Multiple master fonts included a software program that allowed users the possibility of interpolating between instances to create their own instance. Adobe eventually offered faces with a variety of design axes: weight, width, style (used primarily in applications equipped for font substitution) and, significantly, optical size.

In 1992, when the influential book arts review Bookways made the switch from machine composition to photopolymer plates, its printer, Bradley Hutchinson, configured Adobe’s multiple master font Minion MM for letterpress printing by reducing the stroke of the face on its weight axis, increasing the optical size axis to open the counters, and slightly expanding the letterform width. This then replaced the metal face, Monotype Bembo, that had previously used in the journal. Subscribers were apparently unaware enough of the transition until informed by the publisher W. Thomas Taylor in a subsequent issue.

Interestingly, in 1994, when Adobe released its version of the historic Jenson as a multiple master (Adobe Jenson MM), it included in its remarkable specimen book a tipped-in letterpress printed poem with a caption reading:

“The use of polymer plates to print digital type by letterpress has become popular in recent years. This method combines the convenience and flexibility of typesetting on a computer with the traditional look of letterpress printing. With multiple master fonts that include axes for optical size and weight, a custom instance can be generated to better suit this method of printing. In letterpress, ink spread occurs naturally, resulting in a denser printed image. To compensate for this in the printed poem, a finer and lighter multiple master font was used.”

Unfortunately, with its co-development of the recent OpenType font format, Adobe has abandoned further offerings in its PostScript Type 1 line. Though several of the multiple master fonts have been reformatted they are now only equipped with “opticals,” the existent multiple master primaries and preconfigured instances.

Generally, digital typefaces need to be altered in a font-editing program such as Fontographer or FontLab to make them more suitable for letterpress printing. This is a fairly simple task but must be approached quite carefully as altered faces may not function properly if not configured correctly. But this is only an alternate solution as, unless one is quite familiar with such programs, only a slight weight reduction is possible without changing the characteristics of the letterforms or damaging the font metrics.

Not all digital typefaces need to be configured for letterpress. A number of them work quite well without any alteration whatsoever. These tend to be from foundries with some historical connection such as Monotype Typography Ltd. or Linotype AG that had inherited large typeface libraries from the years when their predecessors sold matrices for machine composition. One of the last faces released by Monotype Typography, before it merged with Agfa, was Monotype Pastonchi, a quite near replication of the machine composition face issued by The Monotype Corporation Ltd. Unaltered digital revivals of historical typefaces, however, are the exception rather than the rule. Some faces such as HTF Didot Light, issued by The Hoefler Type Foundry Inc., function quite well for letterpress in that they are almost too spindly and anaemic for general digital work. Other than Adobe, however, the only digital foundry previously “friendly” to letterpress was Lanston Type Co. Ltd., which issued a number of digital faces reproduced from the American Lanston Monotype Machine Company’s historic type library. Lanston Type acquired the remnants of Lanston Monotype, along with the original brass master patterns in 1989.

Lanston Type’s goal was to digitize the faces with all the unique spatial and visual characteristics of the original typeface so that the eventual type, when printed, would look as if produced on a Monotype casting machine. All the variant characters associated with the original font scheme would be made, including small caps, ligatures, ranging and non-ranging figures, tied, swash, and accented characters, alternate short or long descenders, etc. At the time, this was quite an ambitious undertaking; no other digital foundry was offering such amenities. (These were not, by the way, often available to the printer in the metal type era, as alternate character matrices were optional purchases, and very few composition houses carried them.) The large masters were pulled as reverse proofs on a Vandercook proof press, then scanned and digitized, with the result that at text sizes the digital recreations were exacting replications of the metal type designs. These digital reenactments perform remarkably well when printed letterpress. Lanston Type digitized a good portion of the Lanston Monotype library but the project came to a sudden halt in the early 90s. There is, however, recent intention to revive the endeavor.

Digital faces such as these do to some extent address the problem of letterform weight exacerbated by ink spread, but unlike Adobe’s multiple masters, there is still no compensation for optical size range. Optical size range can be built into a digital typeface by creating several alternate weight instances with a font-editing program (as mentioned above). But this is not as desirable as if optical size range was offered directly from the typeface designer. Several digital foundries, such as Adobe, Agfa Monotype, Carter & Cone Type Inc., and Hoefler Type Foundry have released certain typefaces drawn at larger size with the recommendation that they not be used at text sizes. If used properly, these can provide satisfactory results when printed letterpress.

An interesting development occurred as an offspring of multiple master technology and possibly of these other considerations. In 1994, International Typeface Corporation completed work on a historically accurate typeface based on the types of Giambattista Bodoni, which was unique in that it was a digital typeface issued in the optical size ranging pattern common to twentieth-century machine composition practice. Originally intended as a multiple master typeface, ITC Bodoni was brought out instead in three separate fonts: 6-, 12-, and 72-point, each representing a redesign of the character set. (This is now the pattern for Adobe OpenType fonts released with opticals.) This was followed in 1998, by ITC Founder’s Caslon (based on specimens and printed text sheets of the types of William Caslon), which was released in 12-, 30-, and 42-point, as well as an 8-line based on wood type. These were less redesigns than exacting replications of originals by type historian and designer Justin Howes.

Revivals have been a part of digital type releases almost right from the beginning, but the idea of resurrecting historic typefaces first appears about mid-nineteenth century with the Caslon revival. It does not begin across the board until the late nineteenth century—under the popular sway of William Morris and his Kelmscott Press, and also as a consequence of increased consumer access to the photographic process. This was a two-phase revival that reached its typographic fruition in the 1920s (primarily conducted by the Monotype companies, British and American). One of the more interesting faces produced by Monotype at that time was Poliphilus (1922), which was based on Aldine types but refreshingly, rather than being completely redesigned, was left somewhat with its original “printed” look.

Only relatively recently have digital type designers thought to go beyond this and preserve historic faces in situ. Besides Justin Howes’ work, Hoefler Type Foundry’s Historical Allsorts (a remarkable specimen collection comprised of the Fell Types—Roman, Small Caps, Italics; St. Augustin Civilité—with alternates; English Textura—with alternates; and Great Primer Uncials), which was released in 1997, is an attempt at rendering historic faces without any designer interpretation or intrusion. Characters from original texts were traced algorithmically by software and left be. Hoefler’s comment on the faces, that they “have a pleasantly arrhythmic pace... and overall, an agreeably (if not ironically) un-digital warmth,” hits the mark. Unlike the ITC revivals, however, the faces that comprise this unique experiment are not size optimized.

Astonishingly, in 2001, Justin Howes reissued Founder’s Caslon from his newly “claimed” foundry, H. W. Caslon and Company Limited, in 8-, 10-, 12-, 14-, 18-, 22-, 24-, 30-, 36-, 42-, 48-, 60-, and 72-point, as well as in 8-line. Each size offered is based on a specimen showing at that size. In doing so, the twenty-first century was given a gift from the past that the twentieth may never have had—a fully size-ranged type design with true optical compensation.

The reissue consists of some 68 fonts in the fourteen sizes: Caslon Text consisting of 8- to 18-point and Caslon Display, consisting of 22- to 72-point (also included with the Display is the 8-line Poster and a suite of Caslon Ornaments). Separately available is a beta version of the fonts in the OpenType format as well as Founder’s Caslon 1776, a singular font based on the text type (at 14-point) used by the Philadelphia printer John Dunlap for the printing of the Declaration of Independence.

Founder’s Caslon is an “exact and scholarly resuscitation.” Howes’ digital reconstruction attempts to bring back “a simple, really basic typographic tool which earlier designers had been able to take for granted from about 1720 up to about 1980.” By reproducing from high-resolution scans of new proofs of existent founder’s type (or where not available, from printed text and specimen sheets), he was able to “sidestep the conceptual and problems associated with re-interpreting punch-cut letters in terms of pencil outlines and computer curve points.” He suggests his Caslon is “richer in texture than designs based on a single master design, and still with something of the vagaries and beauties inherent in punchcutting and letterpress printing.” Indeed, Howes felt no need to attend to the problem of ink spread in his digitization:

“The irregularities of letterpress printing are also faithfully captured by the process. I see no reason to remove them, since they become troublesome only under high magnification and, at normal size, make an important but virtually invisible contribution to the text of the page... Ligature Caslon [Founder’s Caslon], if it cannot produce diversity, makes an attempt to reproduce it.”

Howes’ work on Founder’s Caslon is quite serious. His scholarly analyses of “Caslon’s Punches and Matrices” (which includes a detailed inventory of the existing Caslon materials) was published in Matrix, in 2000. He had access to the type, punch, and matrix collections of St Bride Printing Library, as well as the Type Museum of Stephenson Blake and Company Limited (Stephenson Blake acquired the remnants of the Caslon firm when it went into liquidation in 1936).

Measured by contemporary standards, the original Caslon old face form is quite quirky in its spacing and lining attributes—the letterspacing is unusually wide and the alignment of the individual characters is not “regularized.” The letterforms themselves suffer from odd contrasts in stroke and color. The face seems designed less for the harmony and details of its letterforms than for the intriguing pattern of its text setting, which provides to it a remarkably deceptive, simple and fluent legibility.

William Caslon cut, in 1722, the first roman for the typeface that is now the ancestor to most faces by that name. It was derived from Dutch type, which then dominated the English market. The popularity of the typeface in years to follow, however, gave England an international reputation as the leader in the type industry. The Caslon old face fell out of use by the turn of the century (the Caslon specimen book of 1805 shows no such fonts) but was revived in 1844 when the Chiswick Press began using it in its books.

In America, the old face style had degraded over the years by successive recuttings. Following the English resurrection, the first “authentic” American version of Caslon dates from 1859, when the Laurence J. Johnson foundry began issuing an old face derived from “borrowed Caslon foundry punches.” By 1892, the American Caslon revival was in full swing. Interestingly, most twentieth-century American Caslon revivals were based on the Johnson types. In the digital era there have been several important revivals, notably Lanston Type’s replication of Lanston Monotype Caslon Oldstyle 337 (considered the closest rendition of the Johnson Caslon), Adobe Caslon (based on Caslon foundry specimen sheets from 1738 and 1786), and Matthew Carter’s Big Caslon, based on titling sizes that appeared in Caslon’s famous broadsheet specimen of 1734.

Most Caslon revivals, however, even these digital versions, are not historically truthful to the originals as they are regularized versions; old face style letterforms were, and are, considered too eccentric for contemporary taste. Founder’s Caslon, on the other hand, has an engaging authenticity about it that hasn’t been seen since the decline of commercial punchcutting. To ensure this, Howes deliberately eliminated inauthentic characters added to the various Caslons since the revival began. To preserve the integrity of Founder’s Caslon he has similarly removed from his definitive version the oldstyle italic figures that are proffered with ITC Founder’s Caslon, and only included those swash and alternate forms that were created by William Caslon himself. The reissued Founder’s Caslon also carries with it character forms that are not in the ITC version, such as a full long-s ligature set. Also included are original Caslon tied characters, small caps, oldstyle nut-fractions, and so forth. Even the 8-line poster font is based on the only authentic wood letter version of Caslon old face (issued in the 1890s).

Working with Founder’s Caslon is a refreshing experience and not too dissimilar from that of standing at the composing bank. The compositor needs to select the font of that size rather than sizing. This might be a bit awkward at first for those who have never handset metal type. And even the most tempered of typographers might be reluctant to adjust the kerning of oddly spaced characters. (Howes, in fact, did not initially intend to apply kerning to the fonts).

Since the face literally changes in its design per size it can be a bit perplexing and one might be tempted to violate the intention of the typeface by using one designed size at other sizes. One anomaly, an unusual jump in size from the 10-point (Long-Primer) to the 12-point (Pica) is not atypical in a metal type design, but could force the compositor to either shrink the Pica or enlarge the Long-Primer. Howes has acknowledged the usefulness of the further addition of a 11-point (Small Pica). This would serve better in the digital Caslon than it might have in metal. In his Mechanick Exercises on the Whole Art of Printing (1683-4), Joseph Moxon revealed why the Small Pica could prove problematic in the printshop: “I account it no great discretion in a Master-Printer to provide it; because it differs so little from the Pica, that unless the Workmen be carefuller than they sometimes are, it may be mingled with the Pica, and so the Beauty of both Founts may be spoil’d.”

The work on Founder’s Caslon [Ligature Caslon], began in 1995/6. At that time Howes was intending to include eight sizes of Caslon’s Old Black ranging from 8- to 36-point, a reconstruction of Caslon’s Long Primer No. 1, and William Caslon II’s “Proscription” letters, as well showings of Caslon’s Greek, Hebrew, musics, and script type. Since these are not part of Founder’s Caslon as presently configured one can only hope that they will materialize in the new future. Interestingly, in his article, “The Compleat Caslon,” which was published in Matrix, in 1997, there are digitized specimen showings of the Old Black and that of William Caslon II.

Howes has even entertained the possibility of further optimizing Founder’s Caslon specifically for letterpress. I asked him why he would offer this considering the complexity of Founder’s Caslon and the amount of effort that this would take—with little expectation of return. He replied, “There’s the sheer pleasure of getting it right... We all know that Caslon looks best printed on hand-made paper... it should be possible for Caslon to look as good as it’s ever done.”

Equally amazing is the empathetic yet entirely differing approach in the creation of Rialto, which was released by the Austrian foundry dfTYPE in 1999. Rialto is an original digital typeface that was designed with letterpress printing in mind. Begun in 1995, Rialto is the result of a collaboration between Lui Karner, the founder (in 1990) of the fine press, Die Fischbachpresse, and the calligrapher Giovanni de Faccio. dfTYPE is the type design offspring of the press.

Representative of the new breed of fine press printer, Karner is holder of a remarkable collection of handset foundry type (which includes such rarities as Cancelleresca Bastarda, Delphin, Elisabeth Antiqua, Lectura, Peter Jessen Schrift, Romulus, Shakespeare Antiqua, Trump Mediaeval, and Walbaum Antiqua) and prints his work (both metal and “photopolymere Druckplatten”) on the highly regarded European-made FAG “Cylinder-Handpresse.”

A typographic tour de force, Rialto has been released in an optically sized range of four fonts: Rialto, Rialto Piccolo, Rialto Bold (all in roman, italic, and small caps), and Rialto Titling (caps only). A recent addition is Rialto Pressa, a special version of Rialto Piccolo specifically optimized for letterpress printing with photopolymer plates. The nicely produced Rialto specimen book (which is available as a promotion) reveals the concerns that went into the design of the typeface.

Beginning with the idea that the shapes of all roman and italic lowercase derive from roman capitals (which may or may not be the case) the team of Karner and De Faccio sought to create a set of capitals that shared the characteristics of both roman and italic, attempting to develop a face that would allow for a harmonious combination of the various type forms while retaining a certain character and contrast to each. Thus the capitals are shared by both the roman and italic lowercase. Rialto is heralded by its makers as “a bridge between calligraphy and typography.” And this was certainly my initial reaction to the face. It does not quite have the calligraphic look of typefaces that are typographic renditions of hand written letterforms; there is, instead, an enhanced typographic feel to the face because of the calligraphic approach to its design. Rialto is uncommonly beautiful. While not an Aldine, it has a familiar Aldine typographic sensibility, but with a tempered contemporary flair.

Though based on the stone-carved letterforms of Roman inscriptions (Capitalis Monumentalis) with an inspiration from the model of well proportioned Dutch type forms, such as those popularized by the seventeenth-century typefounder Christoffel van Dijck, Rialto’s capitals are “distinct from all historical types” in that they have calligraphic serifs derived from the broad pen. The movement of the pen is also reflected in the “flow of the stems into the serifs” which imparts to them “a delicate lightness and dynamism.”

Karner and De Faccio provide an interesting synopsis of the historical evolution of the height and slope of capitals in defending their relatively small cap height and degree of slope. The capitals are noticeably lower than usually presented and the rationale was that this would allow them to harmonize better with the italic. This seems to be the case and certainly sets Rialto apart. Based on historical romans, a one degree slope to the right was considered by the team to be “an important requirement for optimum legibility.” The calligraphic features imparted to the capitals were also given over to the lowercase roman font, which are likewise sloped. The mark of the broad pen is revealed in the sharp angles where the stroke changes direction and in the precision of the serifs (especially the baseline serifs which have a remarkably engaging dip, extended draw, and sudden lift).

Concern for legibility is also shown in the distinctive movements of the individual letterforms relative to the x-height and in the traditional small counters provided the a and e characters. The italic was allowed upright shapes without regard to slope but to ensure harmony with the capitals the slope was held to three degrees. There is also a considered empathy with the roman in the spacing of the italics, which are set a bit wider than normal and therefore show well in text mass.

Other typographic amenities provide typographic breadth to the font. Small caps are provided for the three Rialto upper- and lowercase fonts and are unique is that they are self-spacing. The oldstyle Dutch inspired figures (e.g., Jan van Krimpen’s Romanée) follow the example of the capitals and serve both the roman and italic, being set at the x-height of the small caps and supplied for both text and columnar uses. A healthy palette of ligatures is provided for both the italic and roman, including both tied-characters and a full set of long-s ligatures. Rialto also contains special characters such as early Italian italic es is us ligatures, a double-storied italic g, and variants of capitals.

Most importantly, Rialto is optically ranged. dfTYPE recommends Rialto Bold for sizes smaller than 6-point, Rialto Piccolo for sizes up to 14-point, and Rialto [regular] for sizes 16-point and larger. The Titling font itself reveals clearly the weight reduction, narrow shapes, and elongated serifs necessary for display.

The special letterpress configured Rialto Pressa is absolutely remarkable. At first I could not detect any significance difference between its Roman and that of Piccolo’s because I was looking for stroke or stem weight reduction. In fact, Pressa has a slight heavier appearance. In examining the italic, I noticed outline point disturbances. When the fonts were matched at 148-point several alterations were apparent. The face appears to have been modified for ink trapping, as well as slightly extended serifs and changes in junctions.

Rialto is a typographer’s dream. Its typographic beauty is not just skin-deep, and it is pleasing to work with. The settings require little fussing and the typographic amenities are configured to each specific font. I have found the roman lowercase combinations of ch and ck a bit too tight for my pleasure, as is the unit space setting of the parentheses, and the characters comprising the tied italics are too widely set apart, but these are quite minor quibbles compared to the deficiencies of most typefaces. All this is made up for by an unbelievably disarming italic f with its wonderfully exaggerated descender—a defining typeface identification mark if there ever was one.

In the Rialto specimen book there is a tiny gloss next to the description of the optical ranging capability of the typeface: “Voilà—there are no more excuses!” Lewis Allen’s statement on the selection of metal type, “choose only after thorough study, for inferior tools corrode the spirit,” is as appropriate today as it was in 1969, when he published Printing With the Handpress. There are no more excuses.


[Jonathan Hoefler and various]. Catalogue of Typefaces: Fourth Edition. The Hoefler Type Foundry, NY. 2000.
Justin Howes. “The Compleat Caslon.” Matrix: A Review for Printers & Bibliophiles, Herefordshire, UK. 17, 1997.
Lui Karner and Waltraud Stefan. Rialto df: a bridge between [calligraphy and typography]. dfTYPE, Texing, Österreich. 2000.
Gerald Lange. “Intelligent Letterform Scaling: Adobe System’s Forthcoming Multiple Master Font Adobe Jenson.” Bookways: A Quarterly For the Book Arts, Austin. Number 10, January, 1994.
Gerald Lange. “Monotype Type Revivals (Part III): Lanston Type’s Caslon Oldstyle 337.” The Typographer (Typographers International Association), Washington, DC. Volume 19, Number 5, July/August, 1993.
Gerald Lange. Printing digital type on the hand-operated flatbed cylinder press. Bieler Press Monographs, Marina del Rey. Second Edition, 2001.
Robert Slimbach [and various]. Adobe Jenson: A Contemporary Revival: A new multiple master typeface family based on the original types of Nicolas Jenson and Ludovico degli Arrighi. Adobe Systems Incorporated, Palo Alto. 1994.