Gerard Unger

Experimental № 223, a news­paper typeface, designed by W.A. Dwiggins

This page is an adaptation of an article by Gerard Unger (1942–2018), originally published in Quaerendo, Vol. 11, № 4, (Amsterdam 1981), pp. 302–324. It is republished here in a new responsive web format with the kind permission of Mr. Unger’s heirs.

In 1929 William Addison Dwiggins (1880–1956) started working for the Mergenthaler Linotype Co., thus entering once again, at the age of 49, upon a career as typedesigner. As it turned out, most of the work he was to do as a typographer still lay ahead of him. He had only been working one or two years for the publisher Alfred A. Knopf, and it was not until the following year that he got his first commission from the Limited Editions Club. His puppet-making, too, something that was to become an important hobby for him, did not start until 1930.1 This is not to say that he was entirely without fame: in 1929 the American Institute of Graphic Arts awarded him its Gold Medal and organized an exhibition of his work. Dwiggins was well-known as a calligrapher, author, typographer and illustrator.2 In his typography he made considerable use of ornaments which, like many of his illustrations and sketches for type-designs, he created with the aid of celluloid stencils.3

From 1904 to 1922 Dwiggins had supported himself largely by drawing countless advertisements and other forms of printed work for publicity. He must have written and drawn hundreds of texts in those years, few of which have survived. In 1922 he was discovered to be suffering from diabetes. At that date the medicine for treating it was still under development and thus not yet available to him. His illness shocked him and made him decide that, as far as possible, he would spend the rest of his life only on things which gave him pleasure. He turned his back on the advertising world, though he did gain satisfaction from recording his experiences in that field in an excellent and still very useful book, Layout in Advertising.4

In this latter work he concludes the chapter on ‘Type’ by expressing his dissatisfaction with the current sans serifs available and by issuing a call for better designs. Harry L. Gage, then Assistant Typographic Director with the Mergenthaler Linotype Co., was spurred on by this passage into getting in touch with Dwiggins in February 1929. A few months later they came to the agreement that Dwiggins would become a salaried typographic consultant and that as a typedesigner he would work exclusively for Mergenthaler. Dwiggins’s further contacts with Mergenthaler were conducted via the Director of Typographic Development, Griffith.

C.H. Griffith portrait

C.H. Griffith in 1936

Chauncey H. Griffith (1879–1956) started work in the graphic industry at the age of sixteen, in Lexington, Kentucky. One of his employers there was the Lexington Herald, with whom he learnt to use Linotype machines. Between 1898 and 1906 he was a so-called ‘tourist comp’ up and down the Mississippi Valley for more than 20 different newspapers. He then made an unusual step: he gave up his $80 a week job with the New Orleans Items for a job at $25 a week as salesman for Mergenthaler, manufacturer of Linotype composing-machines. His rise with this firm was rapid, and in 1915 he was put in charge of the division of typographical development. He retired, as vice-president, in 1949, but continued to act as a consultant to the firm right up to his death.5

Griff, as he was usually called, was an extremely conscientious craftsman, who avoided publicity but contributed greatly to the way Mergenthaler’s type programme developed. He was one year older than Dwiggins. Although unable to design himself, he possessed an outstanding capacity for assessing designs. He also had good assistants, including Nils Larson, who worked for Mergenthaler from 1922 to 1959—from 1925 as assistant supervisor and five years later as supervisor of the letter-drawing department. Like Griffith, Larson was a skilled craftsman who had learned his trade the hard way.6 These men were largely responsible for the work leading up to the creation of a typeface. Rudolf Ruzicka, who also worked along with them, wrote as follows: ‘It was with complete confidence that both Dwiggins and I left “fitting”, the last vital step in type design, to Griffith.’7

Dwiggins designed typefaces for a variety of purposes. His first design for Mergenthaler, the sans serif Metro, was a display face. The other faces he created and which reached the market were mainly used for books. From 1929 onwards Dwiggins made repeated attempts to design a newspaper face. But despite all his experiments nothing by him succeeded in being published in this admittedly difficult field. It was precisely this terrain, on the other hand, in which Griffith won his spurs. He was the man behind the ‘legibility group’, some of the most frequently used newspaper faces, which even today have retained their popularity. These types are somewhat dry in character and derive their form from the unyielding technical restrictions imposed by the demands of newspaper production. Ionic and its progeny are undoubtedly useful faces. Dwiggins, however, wanted more than merely types which could match technical requirements. He was aiming at the creation of an attractive newspaper face with a clarity and vivacity that would dominate technique, and which would be more attractive and more readable for the public.

On 8 August 1929 Dwiggins wrote to Griffith that he was interested in the latter’s news faces, and particularly in Ionic. ‘I wondered what would happen if a feller tried to get some of Jones’8 sleight-of-hand into a newsface, observing the requirements that you canvassed in making Ionic.’ His remarks were accompanied by a sketch. Not long after this Griff got on to the phone and ordered WAD ‘to take a good rest and keep away from newspapers.’ This produced an irritated reaction from Dwiggins, who immediately wrote back: ‘Dear Griffith: Yessir. Keep away from news, sir. Very good, sir. Anything else, sir? Me I’m agoin’ to leave these yer parts on Friday an come back after Labor Day. Wishing you the same, WAD.’

Drawings for the letters g, f, n, a, t, p, r, s

Dwiggins’s first sketch for a newspaper face with an extreme x-height, which went with his letter of 1 August 1929 (Orig. drawn as 10 × 6½ point)

This altercation is very much the exception in the twenty-year-long collaboration and extensive correspondence between the two men. Dwiggins wrote much and enthusiastically. His letters are liberally sprinkled with bright ideas, detailed information and witticisms, all expressed in terse sentences. The handwriting is beautiful, and the letters are frequently interlarded with drawings. Griffith’s letters are much more businesslike. They provide a great deal of technical information, and are also often adversely critical, but never without supplying arguments to support the criticism and suggestions and encouragement to soften it.

The reason for Griffith’s uncompromising reaction was his involvement with the development of newspaper faces. These types were only just beginning to make headway in the newspaper world, and there can have been little enthusiasm at Mergenthaler’s just then for a new approach which threatened to undermine their chances of success. Dwiggins sensed this too, for his letter of 8 August ends: ‘… and not let any new stuff out of the cage, along news body-letter lines, until Ionic is totally absorbed.’ Griffith’s phone-call resulted in this project being shelved until 1936, with the exception of a brief reminder in 1930.

Griffith himself was not displeased with his technically inspired solutions. He was ‘nourri dans le serail’ and knew exactly what was required of a newspaper face. In the introduction which he wrote to his correspondence with Dwiggins (now in the library of the University of Kentucky) concerning the latter’s design for a newspaper type (Experimental № 223) he claims:

In support of my “engineering” approach in the design of Ionic, Excelsior, Textype, Opticon and Corona, I submit while heartily in agreement with WAD’s criticism of these types which is sound, fair and wholly warrantable in every particular, that the performance of type in the various processes by which a newspaper is produced is entirely foreign to that which is involved in ordinary printing. The factors of stereotyping, soft pulp paper, thin ink, heavy impression, and high speed presses, all serve to defeat or handicap the true performance of small type on the basis of this performance in book and other classes of printing. In the first place, design detail is lost in making the papier-mâché matrix from which the stereotype printing plate is cast, also in the casting of the plate; ink is less perfectly applied to the plates by the hard form rollers, and this operation with the addition of lint from the fast moving web of pulp paper tends to fill up the small counters in such letters as the “e” and many others—all of which submerge the fine nuances of letter design, and in many cases the identity of the design itself, and result in a relatively blurred text for you to read. […] Fundamentally an engineering problem, requiring the analysis of the structure of each letter and its behaviour characteristics under the conditions imposed on it.

Dwiggins, with his background and experience, had an entirely different approach. But he was also fully aware of the technical problems involved. Things were easier for him than for Griffith, who had had to pull the chestnuts out of the fire. When in 1936 WAD, started working finally on his newsface, Ionic (1926), Textype (1929, almost identical to Ionic but with longer ascenders and descenders) had already come out, followed in 1931 by Excelsior (slightly lighter and with a slightly smaller x-height than Ionic). These types comprised the first part of the legibility group and they enabled Dwiggins to build on Griffith’s experiences.

8 single-line font samples, shown with the text “faces are standard throughout”

The ‘legibility group’ with two forerunners, from the top: (1) a typical Modern from the second half of the 19th century, Linotype DeVinne, (2) Century Expanded, (3) Ionic № 5, (4) Textype, (5) Excelsior, (6) Opticon, (7) Paragon, (8) Corona. All specimens enlarged from 8 point, printed on smooth paper

Dwiggins’s approach is characterised by the unconventionality of his ideas and by great imaginativeness. He attempted to give his newsface much more extreme proportions and a more pronounced character than was possessed by Griffith’s designs. On 30 September 1940, for example, he wrote: ‘I am just cuckoo enough to be “intrigued” by this problem of getting a thundering big character on a 7 pt body; it ties up in a way with that “half-uncial”9 stunt that I talked about once.’ He quoted Arthur Rushmore’s10 opinion of this approach: ‘should print to beat Hell, even on the tail of the Devil’s night shirt.’ The curious thing is that Experimental № 223 refused to grow into a homogenous and convincing design. Most of the Dwiggins designs brought out by Mergenthaler turned out successful. Alongside these Dwiggins created numerous designs which failed to get further than the experimental stage. Into one or two of them, such as № 223, the designer poured idea after idea, so that it became difficult to achieve anything in the way of a unified result. For this reason № 223 is not so worth studying as a type in itself as for the large number of ideas Dwiggins put into it: it provides a beautiful illustration of the way in which Dwiggins approached a design. Griffith’s great merit lies in his giving WAD every opportunity to develop this approach.

Dwiggins did not have any one fixed method for setting about designing a face. Sometimes he set out from his own original ideas and calligraphy—‘Electra’ is clearly the result of such a procedure. At other times he started out from existing types and moulded them to fit his own insights. ‘Caledonia’, for instance, is based on a type cut by William Martin (c. 1790) and adapted by Bulmer, and on the ‘Scotch Modern’ of Alexander Wilson as displayed in his type specimen of 1883.11,12 Despite this, Caledonia is not a pastiche. WAD’s sense of form was so strong that it permeates even the smallest details of the design while preserving the ‘modern’ flavour. Contemporary designs, too, were capable of serving him as starting-points, for example Cheltenham and Times New Roman. Other people’s designs usually operated as catalysts for his own ideas. He sought for a design with a particular atmosphere, took over the proportions and the colour or the technical fundamentals, and then proceeded to go his own way.

Three-line font samples of ”Electra with Italic, 12 pt” and “Caledonia with Italic, 12 pt”

Electra (top) and Caledonia

Dwiggins never worked constantly at a type design. He could not have done so even had he wanted, since after the working-drawings had been completed, for example, Mergenthaler needed time to carry out a whole series of activities in order to make a test-casting and produce a trial fount, and the proofsheets then had to be sent to Dwiggins for his appraisal. All in all this could take a number of months. But quite apart from necessity, WAD often temporarily shelved a design of his own accord. After some time had passed he would come back to it with fresh inspiration and new ideas. Electra was created in a relatively short time. He started working on it in 1930 and it was put on the market in 1935. Falcon, on the other hand, was a bird of a very different feather. It was started as an original design as early as 1929. Ten years or so later Dwiggins pushed it more in the direction of a Garamond. By 1942 the project had been virtually abandoned since neither Dwiggins nor Griffith could see much promise in the results. Later on, however, Dwiggins became enthusiastic again, resuscitated his design and worked on it for a number of years. In the end Falcon was the last of Dwiggins’s designs that Mergenthaler was to bring out. It appeared on the market in 1961, five years after Dwiggins’s death. The newsface project ran from 1929 till 1944.

Dwiggins has a remarkable way of expressing himself. To me it seems even more remarkable that Mergenthaler published his texts, such as the type specimen for Electra (1935) with illustrations by WAD and poems by William Rose Benét, together with the amusing account by Dwiggins of his meeting with Kobodaishi, the Japanese patron saint of lettering; and the Caledonia specimen (1939), in which WAD allows his alter ego Hermann Püterschein to fulminate against him.

Dwiggins also wrote a number of unusual texts by way of accompaniment to Experimental № 223. Between 1937 and 1941 Griffith received, in addition to large numbers of letters, fifty or so documents containing drawings in which WAD explained his ideas about this design. In February ’37 he sent off a manuscript running to about 18 sides and entitled ‘Newsface’ (document 1). In July he wrote a piece called ‘M-formula’ (2), and this was followed in August ’38 by a further contribution in which he explained various aspects of his design (3). In May ’39 he wrote a document entitled ‘Action’ (4), and on 30 September ’40 he sent Griffith an untitled set of proposals for a completely new approach to № 223 (5). These five documents, together with a number of shorter ones, are representative of the rich variety of ideas Dwiggins could come up with and also of his style of writing.


In the spring of 1936 the newsface project was resuscitated. On 27 March Dwiggins wrote a letter in which he now based his plans on Excelsior (1931). He had a lot of criticism to make of this type—he found it too stiff, some of the letters were cramped in their design. Griffith in his reply expressed the hope that this criticism was directed at Textype (1929), a type which he claimed not to be happy with either, although he did not back up the claim with arguments. Dwiggins’s enthusiasm was fired again by reading Allen’s book on Newspaper Make Up,13 which had been sent him by Mergenthaler. This book is set in Textype. The type has been adapted for books by the addition of lengthened ascenders and descenders, which results in its making a somewhat hybrid impression. This time Griffith yielded to WAD’s pleas and wrote back: ‘We are interested in any effort that will lead to an improvement in newspaper body types from a legibility angle.’ Mergenthaler had so far been incredibly successful with Ionic, Textype and Excelsior (1931). Now, unlike the situation in 1929, proposals for change were obviously no longer felt as a threat but as a chance to gain a still firmer hold on the market.

In 1937 Dwiggins intensified his criticism of the Mergenthaler approach. In the ‘Newsface’ document (1) he wrote:

Your (Lino’s) attack on the problem of newsface, via Ionic, Excelsior, etc.,—is the first intelligent effort in the line ever made, and results in a solution that is adequate in every detail so far as I can learn, from all engineering angles. […] You have taken the style of letter established by custom as “the” newspaper face, and have modified it step by step until it will print under modern conditions, and can be read when printed. I think very little can be added on the side of the problem that comes under the head: It Will Print. If anything more can be done, it will have to be in the department: It Can Be Read. […] The Lino. work of improvement has been done on one specific style of letter. This style was taken over as the subject for experiment for historic reasons: it was the style of type-letter in fashion when newspapers began their expansion after the Civil War, and when Lino. began its career. There wasn’t any other kind of letter to work on, so far as body matter went. So, naturally, Lino.’s experiments fastened upon this style, and by so doing fixed it as the only kind of typeface proper for newspapers to use. The style was (and is) a much modified bastard descendant of an original “Scotch Modern”. In its much modified Post Bellum form it was not what one could call a highly legible face.

Dwiggins’s criticism is only partly justified. There was certainly room for improvement—as later years were to show—but nevertheless Ionic, with its sturdy, in a sense even awkward, forms and proportions represented a significant advance in readability. Prior to its arrival newspapers were frequently printed with 19th-century ‘moderns’ which, with their thin hairlines and serifs, presented a fragile picture and which rapidly became broken and thus made reading more difficult. Ionic could resist breaking, but proved to be somewhat too solid for easy reading unless it was leaded. The arrival of Excelsior represented an improvement as far as this problem was concerned. Long before 1925 an important step had been taken towards strengthening the spare, fragile image of the moderns (or ‘didones’). Theodore Low de Vinne, together with Linn Boyd Benton, developed for Century Magazine—which was printed by his firm—a face called Century Type (1894), which in an improved version was put on the market by American Type Founders Co. (ATF) in 1902 under the name of Century Expanded. This face, which was also available on the Linotype from 1904 onwards, was still too cramped however, too restricted in the size of its image, to be really suitable for modern newspaper printing. Twenty years later a wider, sturdier and more open version was brought out, the Century Schoolbook, designed by L. B. Benton’s son, Morris Fuller Benton.

Font samples showing basic character sets of Century Schoolbook and Italic

Century Schoolbook (30 point orig.)

As regards the historical background to Ionic and Excelsior, Dwiggins was perfectly right. Although the Century Schoolbook clearly formed a precedent for Griffith, he went back to a 19th-century Egyptian of British origin—specifically to a variant brought out by a number of English and Scottish foundries in the 1860s. This Egyptian was designed to serve not only as a display type but also in small size as a text face, for example for catchwords in dictionaries. It offered a maximum of clarity in a small compass.14 The version of this face brought out by the Miller & Richard typefoundry at Edinburgh in 1863 under the name Ionic № 1, and which was copied in the United States, is the true forerunner of Griffith’s Ionic № 5.15 It is interesting to note, by the way, how for a long time Dwiggins in his sketches stayed close to the newsface image popularized by Griffith. He changed the vertical proportions, it is true, but he went on planning his newsface in the conventional ‘Egyptian-idiom’—little difference between thick and thin, substantial serifs that flow into the verticals, a round, broad image, and short ascenders and descenders. The changes he was to introduce later were concerned with the bracketing and the shape of the serifs, and the shape of the curves, especially on their inner sides.

When in 1936 Dwiggins turned back to his design once again and Griffith gave him the go-ahead, it was not exactly clear what Mergenthaler stood to gain by this experiment. With their Ionic, Excelsior and Opticon (1935, a little heavier than Excelsior for printing on harder paper) and Paragon (also 1935, a bit lighter again to allow it to consort with heavy display types and illustrations) they dominated most of the newspaper market, certainly in the United States. The year 1941 was to see the addition of Corona, designed to cope with problems in stereotypy (high-shrinkage matrices)16 and enabling more text to be fitted into a column. The arrival of this face completed the legibility group as far as Griffith’s lifetime was concerned.17 In 1960 Aurora was added, made under the supervision of Jackson Burke, Griffith’s successor, and designed for the Canadian market and for teletypesetting. The latter process imposes extremely narrow limits on the widths of the letters.

In Boston there is a resumé, drawn up by Mergenthaler in 1946, of all the work that they were carrying out for Dwiggins at that time. Thirteen types are included plus a number of individual items. As regards newsface № 223, there were at that moment 8 roman lower-case letters on a 5½-pt body; in the 7-pt the roman was complete, the italic was still lacking punctuation marks, ligatures and numerals; and two capitals and three lower-case letters had been made for the bold-faced roman. In addition the lower case had been cut for the normal roman on a 10-pt body. The decision to start on a body of this last size was taken because Archbishop Spellman was interested in a type of this sort for setting a Missale Romanum. Spellman later abandoned these plans, thus pulling away the carpet from under the 10-pt body.

After this the whole project was shelved, the correspondence relating to it dried up, and the decision was finally taken to drop the experiment. Documents dating from after 1946 provide what is only a partial clarification of the reasons for this decision. In 1947 № 223 was put on public show at an exhibition in Chicago, together with a number of other experiments and finished designs.18 In 1949 Griffith wrote that he had named the type Hingham, after Dwiggins’s place of residence. What is both striking and confusing is that this name had originally been assigned to Experimental № 267D, WAD’s experiment on the basis of Times, Plantin and Bookman. In Griffith’s letter we read: ‘… and ordered the small caps and a few odds and ends cleared up, so the face can be put into production at short notice.’ He also makes mention of two ways in which the type might be used: ‘for small format bibles for Harper’s, and on 4 and 5 point bodies for “a big Webster Job”.’ Griffith retired in this same year and was succeeded by Jackson Burke. The only Dwiggins types that were brought out under the latter’s management were Eldorado (1951) and Falcon (1961).

Drawings for the letters v, n, i, f, e, g, t, B

The second attempt at a large-bodied newspaper face. Drawing dated 27 March, 1936 (Orig. drawn as 20 × 7 point)

In 1953 Griffith made the following observation: ‘This face was not completed. The design characteristics, however, were incorporated to some extent in a later news body face, “Corona” [1941] now in extensive use.’ This could imply that the design was worked on for another five years or so, even though it had outlived its usefulness. It is most improbable that Griffith and his collaborators would have persevered so long if they had not felt that what they were doing held forth some sort of promise. In his introduction of 1956 Griffith puts forward an alternative reason for the scrapping of № 223: ‘It so happened that about the time this project was completed, in 1943, [?] the trend in newspaper style was toward the use of larger type, with 8 point as the minimum—while today, 1956, 9 point for body type is becoming the standard.’ He preceded this by suggesting that the ‘M-formula’ which Dwiggins applied to the design in the summer of 1937, while satisfactorily enough for the smallest bodies, failed, when it came to 10 point, to provide a pleasant and easily readable image.


In the spring of 1936 Dwiggins had sent Griffith a number of drawings with his letter of 27 March or shortly after it, and some of these were put into execution. The attempts were then consigned to the archives. From 1937 up to 1943 work on № 223 continued steadily, with breaks of a few months from time to time. From August ’38 onwards Dwiggins made a great quantity of ‘thin paper drawings’—accurate working-drawings (64 × 12 pica points)—from which Mergenthaler made brass patterns to guide the punch-cutter, engraved punches, punched matrices, and cast and set type and made pulls. All the while Dwiggins poured out a constant stream of ideas. The document entitled ‘Newsface’ (1) was written on 22 February 1937 for Griffith when the latter was briefly in hospital for examination and rest. ‘Written to be read by CHG while he’s in convent’, is what WAD wrote on the title-page. Further on he launched into another close analysis of the ‘Post Bellum many-times-mishandled modern’. One important point of detailed criticism that he made was that the ‘engineers approach’ of Excelsior had not succeeded in creating a unified whole. To WAD it looked as if letters of different origins were standing side by side: ‘… the “hand” that swung the o never struck off the loop of the d, and the d hand didn’t swing the p’. It remained a bunch of loose letters ‘… unwilling to fuse into words—to shake down into sentences and paragraphs.’

Side-by-side gestural illustrations, one labeled “freehand” with loose flowing lines and the other labeled “engineer” with more rigid geometric lines

Freehand versus engineer

In ‘Appendix I’ he explained what he meant by ‘fuse’: ‘Something happens when everything is OK, what is it?—something that gets all the animals under one “top” and brands them as A Circus. It is more than “blend”.’ And he went on: ‘One of the things that make any design “fuse” is what art slang calls Variety in Unity.’ WAD initially thought that salvation was to be found in ‘the control of one “hand”. This quality of a hand, or similarity of motion running through a face is what I think of as the action of the face.’ […] ‘And I think that the action that makes a type come alive is got only by free-hand curves—not by geometrical curves.’ With a significant illustration he contrasted ‘engineer’ with ‘free-hand’ and made a clear choice for the latter. He was to return later to the concept of ‘action’. He wasn’t happy either with the proposal about fusing quality. He continued to brood on it and would soon come up with a much further-reaching proposal.

Stencil sheet with a variety of modular elements that can be combined to build up letterforms

The celluloid stencil that Dwiggins made for the specimens in the ‘Newsface’-document of February 1937

In this document Dwiggins also wonders whether he really had to persist in using the idiom of the ‘moderns’: whether it was necessary to go on introducing thick-thin differences in strickly vertical and horizontal planes. Nevertheless, in the sketches which accompanied the document he remained close to the Egyptian character of the legibility group. The only real difference was that all the curved lines were drawn more freely and widely. Later on he was to show a very clear preference for an ‘old style’ formation, in which the thickest points of the curves clearly lay above and below the centre of the x-height and the top serifs were unmistakably triangular.

Typeface sample with a very large x-height, labeled “Experiment for a 5 point face for time-table, with its related boldface.”

Page 18 from the ‘Newsface’-document showing the 5 point timetable face

Dwiggins then proceeded to add a delightful twist to his document. He told Griffith—who, don’t forget, was supposed to be recuperating from overtiredness—‘At this point empty your mind of all prepossessions, convictions, prejudices, about newspaper type. Get yourself in a condition to look at things with a mind as fresh as a baby’s. Forget that you ever worked on a newspaper …’ After presenting this text in large, red letters, he fired a proposal at Griff for a 5-point ‘timetable face’—a type with what he later described as ‘a thundering big character’, with an extreme x-height and relatively narrow. The only demonstrable result of this experiment was his attempt later on to adapt this forceful image to larger bodies as well, 7 and 10. Some months later he sent in a new written contribution, having in the meantime been busy with working-drawings for 12 lower-case letters and 3 capitals referred to as ‘Babylon Newsface’. Once again trial shots were made of a number of the drawings, after which they disappeared into the archives, as had happened with the earlier attempts. Obviously these tests were not considered successful either.

Two illustrations of marionette heads and a photograph of another

Illustrations from the ‘M-formula’-document showing the different ways of making a marionette head. Right: an experimental marionette, made from cast parts, called Marina, showing the angular features for the M-formula

The document of 3 July 1937 (2) comprised 4 memoranda. In the first Dwiggins reasserts his belief that there was a genuine need for the sort of face he was looking for. In the second he cites something that Griffith had sent him in which a trend is signalled heading in the opposite direction from types such as Cloister and Caslon, and leading towards firmer, more precisely drawn and more scintillating letter forms.19 In the third memorandum Dwiggins came up with a brilliant invention, the ‘M-formula’, in which he combined his love for puppets with his love for type.

MEMORANDUM 3. In cutting marionette heads in wood I came up against the problem of projecting the face of a girl—so that the doll would really look like a girl of 18—subtly modelled features, delicate, springlike, young—to the people in the back row. (Aged folk like us are easy to carve, and project). I started by making delicately modelled heads. [exhibit A] These were charming at arms length, but the girl quality did not carry to the back benches. Then I made a discovery. Instead of soft curves for the cheeks, etc., I cut flat planes with sharp edges [exhibit B]. These sharp cut planes, when viewed on the stage, by some magic transformed themselves into delicately rounded curves and subtle modellings; and the faces looked like young girls from clear across the room, as well as from the front benches.

Adstract decorative ornament

‘Geometrical spinach’, a stencilled ornament from ‘Memorandum 4’ of July 1937

In this way Dwiggins responded to the phenomenon that light, and even more distance, influence the manner in which objects are perceived. He rightly concluded that these influences on his puppets were comparable to the influence of a marked reduction in letter size and the distance from which they were read. In a later explanation of the ‘M-formula’ he stated: ‘The said M-formula applies to small sizes. It is, as discussed before, a method to trick the eye (in viewing objects much reduced) into seeing curves that aren’t there. Naturally its most effective work would be in a news letters of 7 point or less.’ The curves ‘that aren’t there’ are the curves on the inner side, where he introduced sharp angles. At the same time the free-hand drawn curves led to a slightly old style effect, which he refrained from saying anything about at this time. He did write, however: ‘My scheme of attack in general is, a more nearly monotone letter with heavier serifs and a higher body.’

The fourth memorandum in the document of 7 July 1937 concerns the writer’s wish to combine taut curves with sharp angular elements. His inspiration for this came partly from his marionettes again, and partly from his ‘geometrical spinach’—the ornaments which he made with the help of stencils. He concluded: ‘These letters are “classical” anatomy processed à la marionette. You will see the method in the drawings. It is more evident in the lower case than in the capitals, but that is OK because most of the character of any font is in the l.c.’ It was after this that the stream of working drawings for the first complete font in 7 point began to pour forth.

Drawings for the letters a, g, o, t, e, f, h, r, n, l, n, i, m, u, p, d

The first sketches showing the M-formula applied to type, an illustration also from the document of 3 July 1937

Technical drawing of a letter ‘a’, rendered in reverse, surronded by notes and a small illustration of an American flag

One of the thin paper drawings to guide Mergenthalers drawing- and matrix-departments. (Orig. size of the drawing 180 × 300 mm) Dwiggins has dated this drawing 1.1.39. Received by Griffith on 9 May 1939. A note made by Griffith on 16 January 1941 says to use a third drawing of this character for cutting. This is WAD’s first large drawing of the 223 a, with his comment: ‘To see how two tricky letters would work on the M-formula and 223 weights’ (a and g). In triumph he put the American flag on top

The document of August 1938 (3) became a reservoir of ideas. WAD proposed backing up the ‘fusing quality’ with a well-balanced fitting. According to WAD it must be possible, depending on the relative width of the signs, to arrive at a formula for the distances between verticals, between verticals and curves, etc. Griffith didn’t believe this. In his opinion an appropriate fitting could only be achieved by taking account of the idiosyncracies of each separate design. This ruled out the possibility of a fixed formula. In the course of the development of Falcon this discussion became much more intense, but no solution was found. Dwiggins also re-emphasized his opinion that short ascenders and descenders did not detract from readability. Opinions about this matter differ. Allen Hutt, an authority on newsfaces, states in his book: ‘… the heavy reduction of the upper part—the ascender—produces a certain effect of confusion, or lack of distinctiveness, which tends to check the eye.’20 There simply is no definite answer to this question either. Personally, I go along with Dwiggins. Unfortunately the drawings that should accompany the document of ’38 cannot be found. It is unclear, therefore, what the significance is of a number of references to 5 different sketches.

The document of May 1939 is entitled ‘Action’. (4) Dwiggins made use of this concept on various occasions in his writings, for example in the ‘Newsface’ document. Generally speaking, what he was trying to do with it was to describe a particular treatment of the curves. In the document we are dealing with he wrote: ‘In calligraphy it would mean the result of the combined motions of hand and flexible pen. […] If you think of type-letters as descendants of pen-letters (Stan. Morison thinks we ought to, apparently) this attribute would occur in type letters also. […] To build up out of angles and straight lines the whip-lash “action” that makes a freely drawn pen-stroke crackle with vitality […] (and which is so damn hard to draw deliberately and slowly) is a triumph of impressionistic art—and for the M-formula! No?’ And he adds in an aside: ‘You once spoke about my trade mark trick of having the arches take away from the stems in a point. Why trick, and trade mark? Isn’t it just the way nice letters behave?’

Drawing of the letter g

An inked version of the trial thin-paper g of 1.1.39, titled by Dwiggins: ‘Püterschein Patent Go-get-em g’

Drawing of the letter h

Caledonia h

Dwiggins also wrote about this sort of ‘action’ and a phenomenon closely resembling the M-formula in the specimen accompanying the publication of the Caledonia in 1939. ‘About the “liveliness of action” that one sees in the Martin letters, and to a less degree (one modestly says) in Caledonia: that quality is in the curves—the way they get away from the straight stems with a calligraphic flick, and in the nervous angle on the under side of the arches as they descent to the right.’ Dwiggins worked on Caledonia from 1932 to 1939, and during the last years of this period he also plunged into № 223. It is not unlikely, therefore, that there was a mutual interchange of ideas between the two designs. As it happens the M-formula was applied somewhat more circumspectly in Caledonia than in № 223. This is possibly the reason why in Caledonia it turned out to be successful.

Finally, the document of 30 September 1940 (5) put forward a completely new approach. This is the document in which Dwiggins expresses his preference for a ‘thundering big character’. At its close WAD makes this proposal: ‘It would be interesting, says I, to see what would happen if you put an old style action through the Excelsior course of sprouts, aiming at a heavier page, say, […] The TRAFFIC drawings (not a bad name by the way) get something interesting in the l.c. The caps not so good.’ Apart from the two drawings included there appear to be no further examples of this design. With these remarks Dwiggins dropped the M-formula and in so doing arrived at what is virtually a new design.

Letter drawings, shown in the text “TRAFFICrmviejof”

The Traffic drawing (Orig. drawn as 10 × 7 point)

Bible page with the heading “GENESIS 12”, showing a dense setting of tightly spaces type

№ 223, specimen in 7 point, dated 13 April 1942

From the beginning of 1942 till 1946 Dwiggins went on working at another type which can be regarded as a newsface. This was Experimental № 267D, designated in an early sketch as a ‘Times Roman-Bookman-Caledonia composite’. He also compared this design with Plantin. The roman lower case and capitals were cut on a 9-pt body, and also a number of italic letters and a couple of trial-letters on 12-pt. This type is clearly old-style in character. Dwiggins failed in this design to forge the variety of different starting-points into a unified whole and thus to give it an unmistakable character of its own. In my view he came much closer to this goal with № 223.

One of the surveys that Mergenthaler sent to Dwiggins and in which the ‘status of characters’ of № 223 is given, shows that the 7-pt normal roman and italic were cut between May 1941 and 1943. In January 1943 a number of bold-face characters were tried out, and the roman lower case on a 10-pt body was cut between November 1943 and February 1944. The 10-pt version shows clearly that the ‘M-formula’, in a dosage as strong as it is in № 223, loses its magic vanishing properties in sizes larger than 7-pt or 8-pt. and in fact becomes irritatingly visible. This rendered the design unsuitable as a newsface. But in a 4 to 7-pt series it could have had its uses for specialized ends (bibles, timetables, even telephone-books). The Mergenthaler management may well have considered this too limited a market to justify starting on the production of matrices.

In his file dealing with the 10-pt version Griffith added the note about the influence that this version had on the Corona. But the 10-pt dates from ’43 and the Corona came out in ’41. What Griffith had in mind was possibly № 223 as a total project. He does not mention any details. Corona makes a narrower impression than Excelsior and its serifs are shorter and more rectangular. It is also rather heavier than Excelsior—its x-height is pretty much the same but the general effect is of a greater openness.

Four-line font sample, with text that begins “How is one to assess and evaluate a type face in terms of its esthetic design? …”

№ 223, specimen in 10 point, dated 3 January 1944. The capitals shown here are Excelsior

Since then, various new developments have taken place in the field of newspaper type, both in the ‘old-style’ direction and in that of the ‘modern’ faces. The most important old-style newspaper fount appeared on the scene as early as 1932 in England: the Times New Roman. Times Roman is a much subtler creation than any of the members of the legibility group or any of Dwiggins’s attempts at improvement. But the Times was developed for totally different conditions of production. At the date of its appearance the Times newspaper was being printed at low speed and on better paper and with better inking than any other newspaper in the world. Later developments, such as Walter Tracy’s Jubilee (1953, English Linotype), Matthew Carter’s Olympian (1970, Mergenthaler) and Times Europa (the Times Roman replacement designed for greatly changed conditions of production, 1971, English Linotype), another of Tracy’s designs, have demonstrated that old-style contained a great potentiality for development. It ought to be added, however, that Times Europa is really more a transitional than an old-style fount.

A lot has been happening with the moderns as well. Jakob Erbar’s Candida of 1937 headed off in quite a different direction from the Ionic-Corona series. Hermann Zapf’s Melior (1952), too, represented a radically new approach with its much more rectangular letter-design. Edwin Shaar’s Imperial (1954), American product as it was, remained much closer to the legibility group. The most striking design in the light of Dwiggins’s № 223 is certainly Tracy’s Telegraph Modern of 1969. This face—although made for still more drastic conditions of production, such as thinner paper, thinner ink, higher printing speeds—displays at a number of points an accentuation of the letter shapes which reminds one strongly of the M-formula, despite being created with quite other ends in view. In his talk at the A.Typ.I congress in London in ’71 this is what Tracy himself had to say about it: ‘So what seemed to be needed for modern newspaper work was a type in which the profiles of the letters are so clear in the original state that they can survive the destructive effect of production and still present a crisp appearance at the end.’21 This led to a crispness of cut that could be called Dwigginsian.

Since the 1970s countless newspaper firms have changed over from the use of metal type and letterpress-printing to photo-composition and offset printing. As a result the conditions which led to the features we have been concerned with no longer exist. But there was a period prior to this change-around—say from 1965 to 1975 roughly—when the newspaper production situation mirrored in an intensified form—and particularly in Europe—conditions governing the period when Dwiggins was working on № 223. This experimental—possibly in a somewhat modified form à la Caledonia could have performed a very useful function during that period. But in America this development was much less noticeable and, moreover, the switch from hot metal and letterpress to photo-composition and offset took place there years before it did in Europe. Since photo-composition, like so many new techniques, first had to undergo an imitative phase, Excelsior and its successors were coldbloodedly copied. This meant, of course, that the traditional image of the newspaper face was perpetuated, although the quality was certainly not improved. It would have been much better if authentic newsface designs could have been created to match the new composing techniques.

Font sample of the letters a, b, e, n, s

Some characters from Tracy’s Telegraph Modern

If Mergenthaler had persisted with № 223 and Griffith had asked Dwiggins to adapt the M-formula on the basis of his experiences with a 10-pt body, the traditional image of newspaper type would have been altered before 1950. Dwiggins would probably have succeeded in changing the image of a useful but somewhat clumsy workhorse into something more subtle, more lively and more brilliant—a face, that is to say, of a very different colour. But the whole question obviously remains one of pure speculation. The legibility group was from the very beginning such a success that there was simply no motivation for Mergenthaler to bother deeply about such matters. This is presumably why, one must conclude, № 223 was never brought to completion.

  1. Dorothy Abbe, The Dwiggins Marionettes (New York 1970). 

  2. Dorothy Abbe, William Addison Dwiggins, A Talk Delivered to the Bookbuilders of Boston (Boston 1974). 

  3. Dorothy Abbe, Stencilled Ornament and Illustration, compiled and arranged by D.A. (Hingham, Massachussets 1979). 

  4. W.A. Dwiggins, Layout in Advertising (New York 1928). 

  5. (Anon.), CHG to WAD, The Type Director to a Type Designer (Atlanta, Georgia 1955). 

  6. W. Tracy, R. Ruzicka, et. al., Letters and Nils Larson (New York 1959). 

  7. R. Ruzicka, C. H. Griffith 1879–1956 (Lancaster, New Hampshire 1972). 

  8. George W. Jones (1860–1942), printer and scholar, adviser to British Linotype from 1921–42. Responsible for the design of e.g. Granjon, Estienne and Georgian. 

  9. Winchester Uncial, viz. ‘Postscripts on Dwiggins’, in: Typophile Chapbook, № 36, vol. 2 (New York 1960), pp. 232 ff. 

  10. Arthur W. Rushmore, former Harper vice-president in charge of design and production, later ran the Golden Hind press. 

  11. W.A. Dwiggins, Caledonia, a new printing type designed for Mergenthaler Linotype (New York 1939). 

  12. D. B. Updike, Printing Types (Cambridge, Massachussets 1962), pp. 193 f. 

  13. J.E. Allen, Newspaper Make Up (New York 1936). 

  14. Much of the information on Century Expanded and Century Schoolbook, as presented here, was provided by Professor G.W. Ovink. To this he adds:

    Without Century Schoolbook no Ionic № 5! But Griffith was undoubtedly reluctant to admit any indebtedness to the ATF, Linotype’s arch-enemy. Hence, and by no means unjustifiably, the name Ionic, even if Ionic № 5 is more Schoolbook than it is Miller & Richard’s Ionic № 1. The italic of Ionic № 5, on the other hand, is not that of Century Schoolbook but (following the Schoolbook procedure) an adaptation of another ‘modern’ newspaper italic. (There was no matching italic as yet with Miller & Richard’s Ionic № 1.) Personally, therefore, I disagree with the idea that Griffith borrowed more from Ionic № 1 than he did from Century Schoolbook. 

  15. A. Hutt, Newspaper Design (London 1967), pp. 54 f. 

  16. The high-shrinkage matrix was developed for economic reasons. First the flong (or matrix) is moistened, to make it pliable enough for moulding. Then the impression of the form is imbedded into the swollen fibres. As the flong is scorched (dried out), prior to the casting of the metal plate from it, the cellulose fibres contract considerably more in width than in length, thus narrowing the type. (After John E. Allen.) 

  17. Hutt, op. cit., pp. 55 f. 

  18. (Anon.) American Type Designers and Their Work (Chicago, Illinois 1947). 

  19. Here WAD quoted an article entitled ‘What will be Tomorrow’s Type’, by Raymond Hopper (source and date of publication unknown to author) sent to him by Griffith. 

  20. Hutt, op. cit., p. 56. 

  21. Interpress Grafik, № 7 (Budapest 1972). 

This article is typeset with a single style of Marionette, a typeface currently in development, inspired by a handful of glyphs Dwiggins sketched in 1937 to illustrate his M-formula (specifically, those shown in figure 12 above). This page and the typeface are made by Nick Sherman at HEX. Images from the original article were provided by Letterform Archive.