Type for printing consists of many little prisms made of a certain metal, at the end of [each of] which there is, in relief, a letter or some other character used in the printing of books. The surface of these characters, which, carrying a film of ink – black red or of some other colour – and pressed firmly by the printing press on to prepared paper leaves its impression there. See Ink, Printing, Printing press.

It is apparent that the character that is to leave its imprint on the paper must be the mirror image of its imprint. For example , in order that the character B should give the imprint B the character must have this form :
since, if one imagines a piece of paper applied to this shape so that it receives its image, it is evident that when the paper is turned over to view the imprint the left side will now be on the right and the right side on the left and one will see not
but B. It is as though one looked from behind at the character printed on transparent paper. That is why reading a [printer’s] forme is difficult for those who are not accustomed to it. See Printing, Forme.

One may imagine that, if one were to have as many of these little characters in relief as could be found in writing and if one possessed the means of arranging them in the required order to represent writing and of coating them with some colouring material then applying them to paper so that the paper received only the images of the imposed characters, one would be in possession of the most desirable of all arts, that of multiplying at little cost and in an infinite numbers of copies all those good books for which this art ought to be reserved; for it would appear that, since the art of Printing puts the fruits of the human spirit into the hands of everyone, only those books should be printed whose reading can damage no one.

This art [pre]supposes the arts of making the characters and of using them : the art of making them divides into two others, that of making the punches needed to cast the characters and that of casting them using the punches  [1].

The art of printing may thus be divided into three : first, cutting the punches  [2], secondly, typefounding  [3], and thirdly the use of the type which has been cast, for which we reserve the name printing .

Here we shall describe the methods of cutting the punches and of casting type [typefounding]. As for their use, that will be found in the article Printing together with a detailed history of the whole art.

Punch cutting . One may regard the punchcutters as the original creators of all the moveable type  [4] with which printing has been carried out since its beginning. It is they who invented moveable type, improved it and brought it to perfection by long and painstaking stages until it reached the state in which we now see it.

Before the invention of printing, text was cut on a wooden block with one block corresponding to each page or to a single sheet; but the difficulty of correcting the errors that slipped into the blocks and the encumbrance of these blocks which multiplied to infinity gave rise to the idea of making the characters moveable and of having as many pieces as there are distinct signs used in writing  [5].

This discovery was made in Germany about 1440 and was found so useful that its success was very rapid. Several individuals were involved simultaneously in improving it, some of them cooperating with the inventor, others, it is claimed, stealing part of the secret to form their own association, and enriching the infant art by their own experiments to the extent that we do not know who was the true inventor of the admirable art of cutting punches and casting type since several people were involved more or less at the same time. However, this honour is most commonly attributed to John Guttemberg, a German gentleman  [6]. See the article Printing.

The punchcutters are little known in the republic of Letters. Unjustly, - there are glaring examples of this - reputations have been attributed to those printers who produced the finest editions which should at least have been shared with their skilled workmen who cut the punches from which the type was cast  [7]. Without the difficult art of typefounding  [8] it would be as though one were to attribute the renown of a good engraving to the copperplate printer who bought the plate and sold the images that he had printed from it to the public.

Much has been said of the Plantins, the Elzevirs, the Estiennes and other printers celebrated for the beauty and clarity of their types without noticing that they were not the types’ creators and, to speak exactly, only showed off the work of others for all that they added value to it by their sophisticated and clean impressions  [9].

We do not intend here to denigrate the art properly called Typography : it has its rules that are not always easy to observe correctly and its difficulties can be overcome only by long experience of working on it. This work divides into several branches each of which demands a particular skill. But is the praise which he attracts for the mechanics of composition, the cleanness of impression, the accuracy of correction, not sufficient without transferring to him also that due to men left in obscurity although one owes to them all that is finest in printing? For it is an astonishing fact that those authors who have written histories of printing at various times, who have followed its progress and shown themselves the most learned on this subject, have dilated greatly on the merits of the printers almost without saying a word about the cutters of characters though the printer, or, rather, typography is only to the punchcutter what a skilled singer is to a good musical composer.

It is so that these artists should have the recognition that they deserve that M Fournier the younger who is, himself, a skilled punchcutter and typefounder in Paris, has discussed them in his book of models of type faces published in 1742. He has included in those who were distinguished in the art Simon de Colines born in the village of Gentilly near Paris, who in 1480  [10], cut roman faces similar to those we have today. Aldus Manutius did the same in Venice at the same time. In 1510, Claude Garamond, a native of Paris, carried this work to the highest level of perfection that it has ever reached, both in the shape of the characters and in the quality and precision of their execution.

Near the beginning of the present century some letters were improved but the exactness and uniformity introduced by Garamond into punch cutting was not surpassed. It was Garamond who, at the command of François I, introduced the face that brought so much renown to Robert Estienne  [11]. Robert Granjean, also a Parisian, son of Jean Granjean printer and bookseller, cut very fine Greek and Latin [roman] type and particularly excelled with his italic faces. He moved to Lyon in 1570 and worked there for eight years after which he went to Rome at the invitation of Pope Gregory XIII.

The faces of this cutter [Granjean] have been more praised than those of any of his contemporaries ; they were similar in style but more refined. The strikes or matrices were distributed across Europe and are still used in many places  [12].

The taste for his italics began to wane about the beginning of the eighteenth century : this typographic revolution, so to speak, was led by MM Granjean and Alexandre, punchcutters to the king, whose faces are used in the Imprimerie royale. In 1742 M Fournier the younger, whom we have quoted with much approval, brought them [italics] closer to our style of writing by the form and the shapes of the solid and white areas that he applied to them. See the article Italic  [13].

Guillaume le Bé, born in Troyes in the Champagne about the year 1525, cut several faces and worked particularly on Hebrew and Rabbinical characters. He worked first in Paris from where he went to Venice, Rome etc. then returned to Paris where he died. Robert Estienne made frequent use of his faces in his Hebrew editions.

Jacques de Sanlecque, born in Cauleu in the district of Boulogne in Picardy, began in early youth to practise punchcutting. He worked around 1558 and was very successful.

His son, [also] Jacques de Sanlecque, born in Paris, began with literary studies in which he made good progress and he became a worthy successor to his father as a punch cutter. In 1614 the Senlecques, father and son, were the only punchcutters in Paris. The son cut some very fine type for Plainchant and music and some fine typefaces among which are included the smallest there was in Paris, which we call Parisienne. See Parisienne.

M. Fournier the younger, a most reliable authority because of his competence in his craft and knowledge of its history, makes the severe judgment that, from the time after the younger Senlecques until the beginning of the eighteenth century, there was no French punchcutter worthy of the name.

With regard to making the distinction between i and u as consonants and as vowels, there was not a single punchcutter able to cut acceptable punches; the old punches that are found from time to time demonstrate how much the craft had degenerated. This happens with all arts whenever those who practise them are but rarely employed. Equestrian statues are only rarely cast; punches for type-casting last almost for ever, thus it must follow that the ways of going about the manufacture of such things, and excelling therein, become largely forgotten.

The cutting of letter-punches is the secret of printing; it was this art that had to be invented for it to be possible to produce infinite numbers of letters and thus to make it possible for printing to produce as many compositions as a language has words, or as the imagination can conceive ideas, and as many as the symbols that Man can invent to express them.

The character is cut in relief on one end of a piece of steel about two inches long and of a thickness proportional to the design that is to be formed on it. It must be cut with the greatest possible perfection according to the principles of Art and the proportions of each letter; for it is on the perfection of the punch that depends the perfection of the type to which it gives rise.

Punches are made from the best steel available. One begins by establishing the form of the letter: this is a matter of taste and we have seen over time how letters vary – not in their essential shape but in the mutual inter-relationships of the components of the form. Take the design of the capital B that we took as an example; this letter is made up of white regions and black regions; the first are hollow and the latter are raised.

To produce the hollow areas one makes a counter-punch in the shape of the white areas See Plate III of typefounding, fig 52  [14] which is the counter-punch for the letter B . When the counter-punch has been smoothed well and tempered to hardness then annealed a little so that it will not fracture, it is ready for use.

With the counter-punch ready, the punch must next be made. To do this good steel is taken and a piece of suitable size is prepared and brought to red heat in the fire to soften it then cut into pieces of the length we have described. One end, which will become the top, is rounded and the other end is carefully filed so that its face is exactly perpendicular to the [long] axis of the punch. This is achieved by putting it in the square  [15] and rubbing it on the oilstone, as will be explained later. Care must be taken to smooth two of the long faces of the punch which are to be placed against the internal edges of the square. A reference mark is made on one of these faces, which serves two purposes 1, to mark the top or the bottom [side] of the letter, according to which face is marked; 2, to ensure that the same faces are set against the square each time as were used the first time. This precaution is absolutely essential; without it one could never have a satisfactory little face on which to cut the letter  [16].

When the punch has been prepared as just described, if it is large it is made red hot, if it is not [large] it suffices just to reheat it, to receive the impression of the counter-punch. It is clamped up in a vice with a suitable opening  [17]. It [the hot punch blank] is secured [in the vice] by two screws with the face perpendicular to the axis at the top. The counter-punch is presented to the face and driven in to the body of the punch blank to a depth of about one line by a blow of the hammer so that it receives the imprint of the hollow areas of the letter.

The counter-punch is then removed and the punch taken out of the vice and filed both on the face perpendicular to the axis [which now carries the impression of the counter-punch] and on the lateral faces, then it is dressed on the stone using the square. Some punchcutters trace around the outer contours of the masses forming the parts of the letter in relief using a very sharp steel point, but, if the counter-punch has been well cut the workman can just by guided by the shape [of the counter-punch’s impression]. The metal outside the tracing - if one has been drawn - is removed with the file; a tracing is always used in the cutting of type ornaments  [18]. Care must be taken not to spoil the contours of the letter by removing too much metal. Next, the letter is smoothed on the oilstone to remove the burr that is left by the file, then the letter is finished with the file or sometimes with a graver [19] so that, on the end of the punch, only the letter is left as is seen for the letter B in figs.3 and 4 of the same plate  [20]. These figures show the finished punch; it can be seen that the excess metal has been removed leaving a slope from the letter down to the shank  [21] .

The dressing square [jig] seen in figs. 6 (lower) and 7  [22] is a piece of wood or copper formed by two rectangular blocks DB, BF that enclose a right angle at B  [23] so that the line descending from B to the stone is perpendicular to the horizontal. The bottom faces of the square, which rest on the stone, are shod with steel or some other metal that has been smoothed perfectly flat on the oilstone, which must itself have a perfectly flat and level surface. The punch is placed in the angle of the square and held in place by the thumb then, with the rest of the hand which holds the sides of the square, the whole assembly is slid across the oilstone which has been spread with a little olive oil. The stone grinds both the sole of the square and the face of the punch. Since the punch remains in contact with the angle of the square at B and, because the wide surface of the square ensures that its walls remain vertical, this ensures that the punch also is kept orthogonal to the stone and thus that the plane of the [face of] the letter is perfectly perpendicular to the [long] axis of the punch.

Once the punch has been dressed it is tempered to harden it. Then it is annealed a little so that it does not fracture when it is used to stamp the matrices; on its solid structure depends its durability and its quality. If it is too hard it will break easily, too soft and the angles of the letter will be blurred and it will have to be re-cut and filed.

All the punches for one body [size] must be the same height measured from the face. All the capitals must be of equal size and the ascenders of the lower-case letter s b, d, l etc . and others that have them, and the descenders of p and q must be of equal length. The small letter must also be of equal height among themselves but shorter than [the letters with ascenders or descenders], for example, m, a etc .  [24] The letters sizes are checked with a gauge which is a flat piece of lead with three recesses, the first for the largest letters such as the long j , capital Q etc. the second for long letters such as capitals and lower-case letters with ascenders or descenders like d, b, p, q, and the third for the other lower-case letters such as m, a, c, e . The letter on the punch when presented to the gauge must fill the appropriate recess exactly so that when the letters are cast their tops and bases will be exactly aligned as can be

seen in this example :
etc. etc.

When the punches are finished they are passed to the Caster who must make certain that the punches he has bought or made have the [letter] face well finished and of sufficient depth and that the bases and tops of the letters fit properly between parallel lines. The letter M is usually dealt with first and acts as a guide for the other letters.

The casting of type . Casting of the type follows the cutting of the punches. The term Typefounding has several meanings  [25]: it can mean either a complete set of punches, matrices and letters, signs, figures etc. used in printing together with the moulds, furnaces and other tools needed for the casting of letters. It can also refer to the workshop where type is cast or to the place where the metal for type is prepared, or, finally, it may refer to the technique of casting type and it is in this last sense that we are going to treat of it here.

Typefounding is a free craft. Its practitioners are not subject to masters, conditions of admission  [26] or inspection. Nevertheless, they the enjoy privileges, exceptions and immunities allowed to printers and are regarded as part of the printing fraternity.

The art is not well known since the vulgar make no point of distinction between Typefounding and printing and regard print as the creation of the printer in the way that a picture is the work of a Painter. There are few places in which the art is practised; there are scarcely a dozen type foundries in France and, of that dozen, more than half are in Paris.

The first Typefounders were punch-cutters, casters of type and printers. That is, the people who made the punches struck the matrices and cast the type  [27] from them, set up the type in formes and printed from it. But the art was divided into three parts because it was difficult [for one person] to practise all of them equally successfully.

One may make the same observation of those who only cast type as of those who only print, that both of them make images, the one group in metal, the other on paper. Whether the letters are agreeable or ugly they are neither to be praised or blamed; both contribute to the quality of the edition, the Printers through the composing [of the type] and the printing, the Typefounders by their pains to ensure that the letters are cast precisely according to the requirements of the Art; that is to say that all the letters of each body [size] are of equal height  [28] and height-to-paper, that all their shapes are properly aligned and equally spaced, that all the roman letters are upright and completely vertical, that the italics are of uniform inclination and similarly for other letters according to their characteristics: we shall deal with all these matters in detail.

Once the Typefounder has procured the best punches his first task is to prepare the matrices. He takes the best copper  [29] available and files from it little lozenges between fifteen and twenty-eight lines long and of a width and thickness appropriate for the letter at the size to be cast. These copper pieces, smoothed and annealed  [30], are placed one at a time on a little anvil and the engraved end of a punch is applied to a suitable place [on the copper blank] then with one or more blows from the hammer the punch is drive in to a predetermined depth of between half-a-line and a line-and-a-half.

In this process, the copper accepts the exact shape of the [letter on the] punch and becomes, in effect, a mould for the body of letters  [31] the same as that on the punch; for this reason it [the stamped copper] is called a matrix . The word mould , on the other hand, is reserved for a whole assembly of which the matrix is only the most important part.

But the newly-struck matrix is imperfect with respect to the shape of which it carries the imprint. The upper face which carries the stamp of the punch must be exactly parallel to the [shape of the] letter stamped on it and the lateral faces must be absolutely perpendicular to the upper face. fig. 1.3. de la Fonderie en caracteres, Planche  [32] The first condition is brought about by filing away the metal of the matrix that projects beyond the plane parallel to the letter face and the second is fulfilled using the file and the square [to file the sides].

This done, the recesses a, b, c visible in figs. 12 and 13 are used. The two recesses a, b one on the upper surface, the other on the lower, (fig 13) at the same level, serve to fix the matrix to the mould. The third recess, c receives the end of the arc or archet [a spring] which pushes the matrix against the mould, as we shall explain.

The mould  [33] is an assembly built of many parts of which one can consider the total to be divided in two  [34].

All the parts of each half of the mould are held in contact with each other [the halves of the mould are pushed together by] by nuts and bolts; the parts are all well smoothed and polished metal except for the two outermost pieces which are of wood and so have the name the woods of the mould  [35] . This [wooden] covering protects the hands of the workman from the heat that is transmitted to the mould by the molten metal which is repeatedly poured into it.

The first two parts of the mould to be considered are those seen in figs 20 and 21 of Plate III of typefounding .Figure 20 shows the platen seen from inside and fitted with all its component parts: in Figure 21 is the same – or a similar – platen seen from the opposite side. To the platens are attached all the other components and it serves as their support as is apparent [from the figures]. The first component to be added to the platen is B, figs. 1. 2. 3. 17. 20., called the longue piece [long piece] since it and its partner are, in fact, the longest components of the mould. ( Note that a given component carries the same identifying letters in the different figures ). This long piece is ten lines in width and is as thick as desired; it is forked at one end X figs. 17. and 18  [36] and thus receives the head of the potence  [37] of the other half [of the mould] for which it acts as a slide. It must be kept in mind that the two halves of the mould are similar in most respects and that all the parts we have described and are about to describe are double [i.e. paired] with one in each half of the mould.

The long piece is fixed to the platen by a round-headed screw b, fig 18 which, after passing through the hole b, fig. 21  [38] screws into the tapped hole in the long piece level with the fork X  [39]. This tapped hole does not penetrate right through  [40] the long piece at the end of which is a square hole dfigs. 17 and 18 which receives the square tenon  [41] of the potence, figs. 9 and 10 .

Before the potence D is put in place one of the blancs C is applied; figs. 14 and 15 show the blanc assembled with the potence. These blancs are the same width as the long pieces. They are a little less that half the length of the long piece

and the same thickness as the body of the type to be cast in the mould.

The blanc attached to the long piece, as fig. 20 shows, is pierced by a square hole similar to that shown in fig 7 . This hole receives the square tenon X of the potence, figs. 9 and 10.  [42] The tenon passes through the blanc and the platine and fixes these parts together. The end D of the potence projects on the side nearest to the long piece. The end m is threaded and carries a nut which is marked d in fig. 21  [43].

The nuts, which are facetted, are turned with the nut-spanner shown in fig. 26  [44].

The blanc may also be attached to the platen by a screw with no visible head which goes through the platen; its shaft is screwed into the wall of the blanc where it ends; however this [type of fixing] is no longer in use.

On top of the long pieces are placed the jets A, figs. 5 and 6  [45] seen in fig. e o  [46]. These jets form two halves of a pyramidal funnel whose external faces are perpendicular to each other. These faces that abut on the platen, the blanc and the long piece, must fit exactly against them. When the two parts of the mould are fitted together it is clear that the jets from a chute with the smaller opening to the bottom. The sloping faces, A, fig. 20 must pass a little beyond the surfaces of the long piece and the blanc, to restrain the path of the molten metal that will be poured into the mould. At the same time they determine the boundary where the excess metal will be broken off from the cast type, and make this fracture easier. See figures 2. 3. 20  [47] where the projection of the inclined faces of the jets is clearly visible.

Each jet carries a screw visible in fig. 6  [48] which, with a nut, fixes it to the platen as is seen in a fig. 21  [49]. The part of the shaft of this screw within the plate of the platen is square and passes through a hole of the same shape; the prevents the jet from rocking as also does the firm contact of one of the faces of the jet with the platen and the other face which rests against the long piece and the blanc.

Below the square hole d in the long piece there is a screw f attached to a dovetail in the long piece  [50]. A nut F, fig 20 on this screw attaches the piece E, fig. 19  [51] called the register. The part of the screw, or screwed tenon, f that fits into the register is square and goes into a socket that is longer than it is wide; this allows the register to be moved out or in as desired.

Each platen has, on its back, a screw G as seen in figure 21 ; this screw passes through a wooden plate, the wood of the mould , which is the same shape and size as the platen to the back of which it is held by a nut. To make the platen and the wood fit together precisely the wood has cavities excavated in it to accommodate the screws, nuts and other projecting parts that are visible on the back of the platen in fig. 21 .

The two halves of the mould, built as we have just described and seen [separately] in figs. 1 and 2 fit together exactly to form a whole as shown in fig. 1  [52].

The potence of one [half] fits into the forked recess in the long piece of the other side and, since the recesses point in the same direction as the potences, they both serve as slides and thus the two blancs are able to move back and forward in relation to each other when one half of the mould is moved over the other.

It is also clear that the space between the jets forms a truncated pyramid with the part which is between the long pieces and the blancs forming a quadrangular prism about ten lines high and of constant thickness. The separation of the blancs is adjustable by moving them together or apart by the action of the registers that are moved back or forward as required, as we have explained. The spaces within the jet and the prism communicate and, in effect, form a single cavity.

This, then, is the assembly of the parts; however, the mould is not yet complete and still lacks its essential component, the reason for the devising and arranging of all the other components, that is, the matrix . The matrix is fixed between the two registers in M as is shown in fig. 2  [53], one end rests against the platen of the other side and the other end is fixed to the attache . The attache is a little piece of sheepskin glued to one of the woods of the mould which passes between the jimblet and the wood. The name jimblet is applied to a little piece of iron driven into the wood of the lower half of the mould that, by restraining the attache, prevents the matrix from moving out of place  [54].

The matrix is held in place between the registers and against the long pieces and the blancs by the spring D C E, fig. 1 which is called the arc or archet. The end E of the spring fits into the recess C of the matrix figs. 12, 13 and acts to push the matrix against the opposite platen and the heurtoir, the part shown in fig 22 . This part [heurtoir] is placed against that visible at m in fig. 21  [55] riveted to the back of the platen; it acts to raise or lower the matrix from the main part of the mould as required and [so] to adjust the position of the letter on the body of the type; to carry out this adjustment [heurtoirs] of various thicknesses are used.

To prevent the matrix from falling out from between the registers a little hook is put between the platen and the wood holding the attache as shown in fig. 23 ; this hook is called the jobet. The ring of the jobet is threaded over the projecting screw G of the platen, fig. 21  [56]; its hook passes under the matrix and supports it but without getting in its way, as can be seen in x fig. 2  [57].

In addition to the components we have just described, each half of the mould carries a hook a b, figs. 1. 2. 3. whose purpose will be explained later.

Before the mould is closed a half-cylinder should be noted, as shown in fig. 17 , a b , on the upper part of the long piece, placed about two lines below its upper surface. This half-cylinder, the cran , is a locating piece with its exposed face rounded, running along the long piece. But since this raised portion would prevent the blanc of the other side from making complete contact with the one carrying the ridge, the opposing blanc has an appropriately-sized cavity excavated in it. The hemi-cylindrical cavity receives the hemi-cylindrical ridge. The cavity is shown in b a fig. 15 .

That completes the description of the structure of the mould, one of the most ingenious devices that could be imagined as will be borne out by what we are about to say about [the process of] casting [type].

The mould is made up of twelve principal components as has been described. All the parts are of well-smoothed iron, fitted together exactly, and together they form a whole from two to four inches long, according to the size of the type, by about two inches wide and containing within its horizontal dimension at least forty separate parts. The two parts into which the mould is divided are called respectively the upper and the lower parts . The one with the archet is the lower part.

Once the mould has been built and made ready the next operation is the preparation of the material from which the type will be cast. To do this, lead and regulus of antimony  [58] are taken and melted separately; then the molten material is mixed in the proportion of four parts of lead to one part of antimony which gives a mixture suitable for type casting.

Alternatively , take crude antimony, an equal quantity of potin  [59] and mix with molten lead and you will have another alloy.

The former alloy is preferable to the latter which it seems has been abandoned in France for the last twenty years because it was found that potin and antimony  [60] produced a great deal of slag which made the alloy pasty and requiring much more heat [to liquefy it].

Thus, we can say that, in general, type is cast from a mixture of lead and regulus of antimony, in which the latter corrects the softness of the former.

The metal is melted in a furnace similar to that shown in the picture [at the tope of] the first Plate of typefounding; it is divided into two parts, both built of brick. That in fig. 4 is a furnace on which is placed a cast-iron cauldron in which is molten lead. This cauldron is heated by a wood fire, as can be seen, and the smoke escapes by an opening at the back and is carried into a chimney common to both furnaces.

The second furnace, fig. 3 of the same image, is a proper furnace; its upper part is the opening of the fire chamber, the lower part is the ash box; these two pars are separated by a horizontal grate. The grate supports a crucible containing the regulus of antimony and the hot charcoal which melts it. The fire is blown by the draught through the grate. The workers who tend the furnace are advised to avoid the fumes from the regulus which are believed to be poisonous: but this is just prejudice, the use of regulus does not expose the Founders to any particular disease. The fumes are not deadly - particularly to cats: the first time they are exposed to them [the fumes] they are affected by such strange vertigo that, after being tormented by them for some time in the room in which they were forced to inhale them they throw themselves out of the windows; I have seen this experiment twice in the course of a single day. But, having escaped, and the first exposure not having been fatal, they enter the fumes that affected them so violently on first exposure and live quite happily in foundries.

The molten regulus in the crucible is poured into the lead in the cauldron in the required amount. The workman 4 [i.e. in fig. 4] ladles out the mixture and pours it into moulds or pottery vessels at his feet. On the floor can also be seen the tongs for the crucible, its lid, a ladle and other tools used in the foundry  [61].

The proportions of antimony and lead [in the metal] are not the same for all kinds of type: since the action of antimony is to give body to the lead [harden it] more or less of it is added according to the size of the type. Since smaller type sizes are less resistant to the action of the press than larger ones, they are cast from metal that the workers call strong and the larger ones from weak metal  [62].

When the metal or alloy has been prepared thus and cast into ingots it goes to the Founders’ furnace. See the furnaces in the picture, fig. 2 and 2 at the right and left  [63].

The furnace is made of the [type of] clay used by furnace-makers for crucibles, but less fine; it is made from powdered broken pots ground up and china clay, beaten together with water. The furnace is some eighteen to twenty inches high, ten to twelve inches in diameter and two and a half feet long; it is divided horizontally by a grate made either of iron or pottery. The wood is placed on top of the grate, the lower chamber D is the ash box  [64]; the upper part has a round opening B about six inches in diameter surrounded by a sort of circlet that supports the iron vessel A fig. 9, called the furnace ladle  [65]. This furnace ladle is divided into two or three portions as can be seen [in fig. 9]; these divisions hold metal of different strengths or qualities, according to the [needs of] the workers using it, and each workman dips into  [66] the division holding the metal of the composition that he needs.

The furnace also has another opening H to which is fitted a pottery pipe that carries the fumes out of the workshop, as can be seen in the image [top of Plate I]. The whole furnace is supported by a table F G G G with a shelf F at half its height which serves to hold various tools [fig. 10].

Beside the furnace several benches are placed as can be seen in the image and, below, in the Plan fig. 11 ; these are sorts of tables with their top at working-height; they are surrounded by an upstand and must be two or three inches lower than the top of the furnace and have one end placed close to it as shown in the image [Plate I, top]. A plate of pottery or iron is arranged with one end resting on the furnace and the other on the table. This plate serves to collect the drops of molten metal that fall from the casting ladle or that the workman shakes out of the mould when it is overfilled.

When the workman is ready to cast a letter, he takes the mould, prepared as we have described and as shown in fig. 1 in his left hand; he puts the end of the arc or archet in the recess that we have explained is in the lower part of the mould so that it bears exactly against the long pieces and the projecting parts of the blancs. He then pushes the two halves of the mould so that the registers are well placed against the lateral surfaces of the matrix and he lutes the bottom of the jet lightly with a little ochre dispersed in cold water, if the letter is very thin. This lute makes the metal flow freely and directs it into the bottom of the hollow pyramid before it has time to cool and solidify through contact with the surfaces of the components that form the walls of this space. The same method is used with the moule à reglet of which we shall speak below. Since the metal in this mould is often thicker and has a long distance to flow it is even more likely to solidify and not reach the bottom of the mould so, in this case, the whole of the inside surface of the mould is coated with ochre and not just the jet, using the thinnest coat possible; but let us return now to the casting of letters .

When everything is ready, the Founder takes up in the casting ladle shown in fig. 13 a quantity of molten metal and throws it into the sort of funnel that we have explained is formed by the jets. The liquid metal descends into the hollow prism between the surfaces of the long pieces and the blancs and spreads out over the surface of the matrix filling all its contours so that, when it is taken out, it is exactly similar to the punch that formed it [ i.e. that stamped the matrix]. It also carries as a hollow the impression of the half-cylinder a b attached to one of the long pieces as we described previously. The resulting hollow is called the nick  [67] and must always be placed on the side that corresponds to the top of the letter  [68]; it allows the printers to detect whether the letter is the right way up or if it is upside down. See the article Printing. The two operations of taking up the metal from the ladle  [69] and pouring it into the mould are shown in figs. 5 and 6 of the image.

There is another important action; at the instant that the metal is poured into the mould the latter must be jerked upwards so that the metal which is flowing downwards strikes the surface of the matrix more forcefully and takes up its stamp.

When the workman has poured his metal he returns his casting ladle to the furnace and prepares to open the mould; to do this he first pushes aside the arc or archet from the recess of the matrix and puts it into a nick in the wood under the heurtoir. He opens the mould by separating its halves and, if the letter is stuck to one half [of the mould] he detaches it with the hook that is fixed to other half; this is called decrocher . This is what is being done in fig. 8 of the image [there is no figure 8!]. After this, the mould is closed again, the arc put back under the matrix, metal poured in and the whole operation repeated as many as three or four thousand times per day.

It must not be thought that the letter coming out of the mould is finished, as least in respect of its body. The letter face is finished - be it elegant or ugly accordingly as the punch that stamped the matrix was well or poorly cut.

Whatever the letter on the face, the counterpunches, punches, matrices etc. and the casting are the same; all the operations are identical for Arabic, Greek, French, Hebrew etc.

As it comes out of the mould the letter carries a pyramidal excrescence attached by its top to the foot of the [shank of] the letter. The piece, called the jet is formed by the excess metal necessary for the casting of type, that was poured into the mould. It is easily separated from the body of the piece of type because of the narrow neck produced by the inclined planes of the components of the mould called the jets , as was explained above and is shown in fig. 16 of Plate II.  [70] Also the brittleness of the typemetal produced by the addition of antimony, almost like tempered steel, makes the separation easier; the piece of jet broken off is called the rompure .

After all the letters are broken-off, that is the jet has been separated from all of them - the jets are returned to the molten typemetal - the letters are rubbed on a piece of millstone as seen in fig. 2 of Plate I  [71]; the stone is known as the rubbing stone . This stone is between fifteen and twenty-five inches in diameter; it is the same kind of stone as Cutlers use for sharpening. To prepare the stone for use by the type founder, one takes two of them placed one on the other with river sand between them, and grinds them circularly, replacing the sand from time to time, until all the little lumps are ground away and the surfaces are flat and uniform. The sand grinds the millstones and does not polish them; little granules remain that remove the excrescences that the casting leaves on the type.

Not all letters can be rubbed down; there are some - more frequently in italic than in roman faces – in which part of the character protrudes beyond the side of the body. If these were ground it is obvious that the stone would carry of this piece and deform the letter. So one begins by scraping gently and taking a small amount of metal off with a penknife, so that it [the piece of type] can fit properly into the space next to the neighbouring letter. This operation to remove the excess metal with the penknife is called cerner  [72].

After the letter is kerned it is scraped and all the excess metal is removed with the penknife from the face right down to the foot. These two operations replace rubbing; kerned and scraped pieces of type meet and fit together just as well as those that have been rubbed. The two faces of the type that are rubbed on the stone are those that were against the blancs of the mould when the metal was poured; these surfaces [of the type] are rubbed to remove the swarf or granulation produced either by the surface of the blanc of one side or the long piece of the other side [of the mould].

Once the pieces of type have been rubbed or kerned and scraped, they are arranged in a [compositor’s] stick – the stick seen in fig.5 Plate III of Typefounding [no such figure!! But see Planche IV fig 2] is a wooden rule with an upstand on which the type is put with the letter at the top and all the nicks turned to the same side in such a way that one has all the letters a arranged thus: a, a, a, a, a, a, etc. and not thus: ava , vav , and so for the other letters; this is made easy by observing the nicks  [73]. The pieces of type so arranged in the stick are moved to the iron rule of the justifier A B fig. 3 of the same Plate [no figure 3 see figs 1 & 2] they are placed with the foot uppermost and the letter against the horizontal face of the justifier, which is just, as can be seen, a composing stick made of metal. To this rule another, C D is, applied which has a shoulder at C similar to that at B on the first piece [fig. 3]; this rule also has at C and D little projections that enter the mortises a and b of figure 3 in such a way that when the two rules figs. 3 and 4 are applied one to the other, they enclose the whole row of type placed on the first rule; thus only the feet of the letters project about one line above the iron rules of the justifier.

The justifier loaded thus with a row of type is placed between the two jaws A B, C D of the cutter shown in fig. 1 Plate IIII [fig 1 no 2?]. The cutter is a kind of heavy solid table to which are fixed the jaw A B which is a plank a good inch thick, and the iron bar F E which has a hook E at one end and a hook F at the other. The hook F is threaded and receives a screw by which the second rule of the justifier, described above, can be moved forward.

The two rules of the justifier are held together by the other jaw C D shown as the lower part in fig. 2. A B, C D are two stout iron bars whose hooks, A, C, are attached to the table of the cutter; B D is another iron bar carrying a nut which receives the screw F E that is turned like a vice by a handle F G . The whole assembly is fixed to the table of the cutter in such a way that only the jaw C D pushed or pulled by the screw E F fig. 2 is able to move.

It follows from this description that, if the screw E F of fig. 2 is turned, the jaw A B moves towards the fixed jaw C D, fig. 1 and that thus the two rules of the justifier are forced against the row of type that it contains. However, to force the pieces of type together in the row, the screw F f must be turned. This screw pushes the second rule of the justifier along the [side of] the row of type until its shoulder C fig. 4 meets the end of the row and pushes them together and towards the shoulder B of the first rule fig. 3 [fig. 2] until they are all pushed against each other. When this has happened, it is clear that the type forms a kind of solid bar contained between the shoulders of the two parts of the justifier, while the whole length of the sides of the row is compressed between the two jaws.

However, before tightening the screws to compress the type in this way, a strip of hard wood is pushed against the projecting feet of the type to force the letter-faces against the horizontal plank of the justifier.

When all is arranged, the type is cut with the plane as we shall now explain.

The tool shown in fig. 6 of Plate III of typefounding [in fact, ie Suite de la Pl III  [74]] is called the plane . It consists of an iron chassis fig. 10 . Under the part N O of this box are screwed two guides Ce, Df [fig. 2 e Suite de la Pl III , top]. This assembly is surmounted by a piece of wood P Q seen in fig. 8 which serves as the handle of the plane. It is attached to the part N O, fig. 10 as can be seen in fig. 6. The iron A B of the plane is placed on the inclined face of the chassis held by the two tapped screws G H in the transverse iron, which are themselves fixed to the chassis as shown in R. All these parts assembled form the plane of fig. 6. The screws are tightened with the screwdriver of fig. 16 of the same Plate.

When type is to be cut, the plane is placed on the justifier so that the projecting parts of the type are between the guides of the plane. The iron is raised or lowered - its cutting edge is slightly rounded – so that it shaves off as much material as one wishes to remove.

There are directives about the height of letters; it is ordained that the height  [75] of the piece of type from the face to the bottom of the foot shall be ten and a half royal lines. This height is not the same everywhere; in Holland it is almost one line more than in Paris, while in Flanders, and even Lyon, it is more than ten lines. When printers, without regard to the regulations, want type more or less than ten and a half lines high the founder has little pieces to add to the letter mould between the jet and the long pieces [to adjust the height].

The pieces are called hausses  [76]; according to the thickness of these the same mould can cast type with greater or lesser height to paper as this dimension is called.

Because the plane iron is convex, the cut pieces of type will all have a little hollow [in the foot] so that, when they are standing on their foot they will bear only on two lines, so to say, instead of on the whole surface [of the foot]. The hollow in the foot of the type is intended so that the pieces of type will be easier to fit on the stone of the press because their bearing surface is small and less subject to disturbances by irregularities [in the surface of the stone].

But this is not the only surplus material to be planed off; some must be cut from the height of the face as is shown in fig. 14 . This cut is applied to both sides of characters that have neither ascender nor descender but only to the side opposite to the ascender or descender if there is one. Fig. 14 shows that the letter B projects more than the letter A even though they were cast in the same mould.

The device shown in fig. 14 holding the letters A and B mentioned above is called the justification ; it is used to ascertain, by means of the little rule seen in fig. 13 called a jetton , if the outlines of the letters are correctly aligned. To do this, once the letters m m, that we have said are the first to be cast, have been justified, one places an a , for example, between the two m s thus: mam , and tests whether the edge of the jetton rests equally on all three characters.

The piece of glass, fig. 12 and its jetton fig. 11  [77] are used to test the thicknesses in the same way and both devices serve in the same way, to test using the jetton, whether the outlines of the characters lie in the same straight line, as we have just explained.

By a fount  [78] of characters is understood, in Printing, a complete set of all capitals, lower case letters, accents, punctuation marks, numbers etc. needed to print a work, all cast on the same body .

The body is of a fixed and determined thickness  [79] for each individual face it is this thickness that determines the spacing between the lines of a book and that gives its name to the face and not the form of the letter on the type. However, to explain further, one speaks of casting a Cicero on a St Augustin body as is done as a means of increasing the white space between lines of type.  [80]

To form a clear understanding of the meaning of what are called in Typefounding and Printing the body, face-design and white space  [81] consider any line [of letterpress print] and imagine it divided into seven equal parts by [horizontal] parallel lines then imagine printed between these parallel lines one of the letters that printers call short  [82] such as a, c, m etc . to distinguish them from those with ‘tails’ such as p, q, d which they call long . Imagine that the [short] letter is traced between the parallels in such a way that its base rests on the third parallel counting from the bottom and that its top reaches the third parallel down from the top  [83] or, to put it another way, that it occupies the space of the three middle bands into which the parallel lines divide the line. Clearly, there remain two unoccupied bands above the three occupied by the short letters, and two below them. This understood, it is not difficult to understand the distinctions between face-design , body and white space . The body corresponds to the whole [height of the] line, the face-design fills the three bands in the centre and is the same height as the letter, and by the white space one understands the two empty bands above and below the face-design .

On the printed page, C D represents half the white space between a line and that above and E F half the white space between the same line and the one below.

Some letters fill the whole height of the body, for example, j complete with its dot, as the example shows, capital Q in roman faces and
 [84] and f in italic as well as the signs ( , § , [ , etc.

For the long letters such as d and q two parts must be distinguished, the body and the tail; the body fills the three middle bands in the same way as the short letters, and the tail occupies the two white bands either above [an ascender] or below [a descender] according to whether the tail points up or down. See , in the example, the d and the q . If there happens to be a q in one line and a d exactly underneath in the line below, there is no space at all between the tails; the ends [of the descender of the q and the ascender of the d ] touch. Thus is the relative height  [85] of the letter body and that of the face determined. Thus, is it necessary that Typefounding and Printing be subject to regulations? If not, [how] should the size of the letters or characters be determined relative to their height: no one has yet tried to regulate this.

We invited M. Fournier to whom we owe the table of the relative sizes of type bodies to provide us with a table of the relative proportions of the letter sizes for each body. The latter is as important to the art of engraving letters [punch-cutting] as the former is for the perfection and convenience of the art printing.

To this end one might consult the rules that the great writing-masters have prescribed and those that the best punch-cutters have followed as a matter of taste.

An observation that presents itself naturally and which it will not be inappropriate to include here is that there is a relationship between printing and the characteristics of a particular language. For example, German is quite extraordinarily verbose, [in printed German text] there is almost no white space between the lines and the letters are very compressed together on the line; by this means the Germans try to recover the space on the page that the prolixity of their speech demands.  [86]

But the expressions face-design, white space, face cast on the body of a different face etc. need no longer cause confusion.

The terms light body and heavy body were used at the time when it was unknown what proportion the face-design should bear to the body and what the relation should be between [type of] other bodies and face-designs. This state of ignorance persisted among us until the year 1742 when M. Fournier le jeune, punch-cutter and typefounder, published his table of the relations between the various bodies of printing type . We shall not delay in describing it. But, before we do so, let us observe that before the appearance of this table there was no reliable rule for making type, each printer ordered type according to that which he already had in his possession, or that he thought he would like to have. No one had any [fixed] idea of either body size or letter size; for example, a Cicero face had as many heights of body and different face-designs as there were printing houses; and one which was called light in one place might be heavy in another, a large-face in one place might be small-face in another.

One spoke of a fount of Cicero , of Petit-Romain etc . each time type was cast on a body so named. Type faces are bigger or smaller according to the needs - or the means - of the printer who orders them by weight  [87] or by the sheet  [88]. When a printer orders type by the five hundred pounds he means that a set of properly assorted characters of all kinds should weigh five hundred pounds. When he orders ten sheets he means by this the amount needed to set up ten sheets, thus twenty formes, for printing without needing to distribute the type  [89]. The typefounder then does his calculations; he allows one hundred and twenty pounds weight for each sheet, including the quads and spaces, or sixty pounds per forme which corresponds to half a sheet. This does not mean that the type for a sheet always weighs a hundred and twenty pounds nor that for a forme sixty; that all depends on the size of the forme and there is always some type left over in the compositor’s cases.

Although all formes do not contain the same number of letters, nor the same varieties of letter, it should be remembered that, since a language has some sounds that occur more frequently than others, and thus some symbols that must recur more frequently than others in their use in printing, a fount does not contain the same number of letters a as it does b or c and so on. The determination of the proportions that there should be of the various characters making up a fount is called the police  [90]. Clearly the police may vary from one language to another but it is the same for all the varieties of face or size of type for a given language. To give an idea of the police in our French suppose that a fount of one hundred thousand characters [pieces of type] is required. To make up this number of a hundred thousand the following numbers of each character are required. This information has solved a problem for the Typefounders for which it would have been difficult to propose an exact solution in the past. I hope that the Philosophers and Grammarians will look with approval at the following table and ask for similar ones for Latin, Greek, English, Italian and most of the known languages. To find such tables they only need to address themselves to the Typefounders of the countries in which the languages are in use.

Police for one hundred thousand letters intended for ordinary printing in French .

The reader will notice that the table contains only grammatical symbols and that we are concerned here only with these. Consequently, this police is not intended for a book on algebra or arithmetic or chemistry but only for works of literature, poetry and so on.

While it is clear that one police does not apply to different languages, it is equally evident that it does apply to all type faces [in printing a given language] whatever their body [size].

There exist in Printing, or, rather, in Typefounding twenty different bodies.

Each of these has its own unique name applied to faces cast on that body. The smallest is called Parisienne and in passing from Parisienne to the largest faces one has:

Nompareille, Mignone, Petit-Texte, Gaillard, Petit-Romain, Philosophie, Cicéro, Saint-Augustin, Gros-Texte, Gros-Romain, Petit-Parangon, Gros-Parangon, Palestine, Petit-Canon, Trismégiste, Gros-Canon, Double-Canon, Triple-Canon, Grosse-Nompareille.

See the articles on these faces under their proper names and, below, samples of these faces in the Plates at the end of this article. These plates are composed from the faces of M. Fournier le jeune from whom we have also taken all the material in the article and in the others on Typefounding. We may say with assurance that our Work [the Encyclopédie] would have left nothing to be desired about the Arts if we had always found persons as devoted to the progress of their art, as learned and as open as M. Fournier le jeune. One observation that we have had occasion to make a hundred times is that, of the artisans of a particular craft, it is the ignorant, and of all the different crafts, it is those which are the least developed and the poorest, that have shown themselves the most obscure; that stands to reason.

These bodies [type sizes] follow each other in regular progression; the base-size, its double, triple quadruple and so on, and the others, made up of two or more in combination, always make up the [size of the] large body that is at the head of the combination; this regular relationship is essential to printing.

However, to bring into existence this regularity to which we have just referred, and that will be well demonstrated in the table of size-relations below, M. Fournier was obliged to create a special body [size] called Gros-Texte equivalent to twice the body of Petit-Texte and to revive [the use of] two others that were unknown - or hardly known – Palestine and Trismégiste . The first is equivalent to twice the body size of Cicéro , the most commonly used type in printing; and the second is twice Gros-Romain.

Without these three body sizes the series would not be continuous. In the following table the first column contains the name of the body and in the middle column the bodies which are its equivalents.

The sign || in the text of the middle column indicates that the number of pieces of type that make up the size of the face named in the margin is about to change and that a different body size [type name] is about to follow; the number required to make up the size of [one piece of] the face named in the left column is given.

But it was not enough to decide on the fixed number of twenty for the body sizes and to establish the size-relationships that must be maintained between the members of this twenty. It was also necessary to fix the absolute size of one of the bodies, no matter which. To do this, M. Fournier le jeune created a scale in consultation with the most experienced Typefounders.

This scale is made up of two parts called inches ; these ‘inches’ however, are not the same length as those of the Royal Foot  [91]. We shall explain later the relation between the inch of the scale and the standard inch [one twelfth of the Pié de Roi]. He [Fournier] divides his inch into three lines and his line into three points  [92]. This scale is shown at the top of the table that follows  [93].

This table has four columns:

The first gives the order [serial number] of the types.

In the second column are the names of the type and its equivalent size in terms of other types.

In the third and fourth [columns] are the sizes  [94] in divisions of the scale.

M. Fournier has thus set up quite a simple system by building his table of the relation between type sizes. One printer orders  [95], for example, a Cicéro from the Typefounder and sends a sample in letters  [96] [of the body] on which he wants the Cicéro to be cast. A different printer also orders a Cicéro and, since the type that both want carries the same name one would suppose that the type itself would be the same. Not at all; the sample sent by one of the printers is either bigger or smaller than that sent by the other and the Founder finds that he needs to adjust all his moulds, or even to make new ones. This whole business can run to extremes in the same way as other fantasies. It seems that there was more agreement among writers than there was among printers on the subject of the size and width of characters. They began by agreeing on the dimensions of the point of the pen, then they fixed the number of pen-points  [97] for each sort of character.

In constructing his table of size relationships it would seem that M. Fournier followed the spirit of the Royal edict of 28 February 1723 for the regulation of printing that seems to have supposed the existence of [such a] table. For example : when the edict ordains that Gros-Romain should be equivalent to [the sum of] a Petit-Romain and a Petit-Texte, what is this intended to mean? Which Petit-Romain and which Petit-Texte should one choose? They are of different sizes everywhere. In setting out this regulation one would imagine that either there was an established table of relative sizes or that one was to be set up. But when a fixed and determined size for type had been set one would not yet have reached the state of perfection that one might have sought, for, to have a convenient equivalent to Gros-Romain it was not at all a Petit-Romain and a Petit-Texte that were needed. According to M. Fournier the body sizes of type had to progress by doubling or halving, for reasons that we shall explain, from which it followed that Gros-Romain must be equal in size to two Gaillardes and not to the sum of a Petit-Romain and a Petit-Texte.

By establishing the sizes of type bodies M. Fournier has enabled Printers to calculate exactly how many more or fewer pages one face would require compared to another. For example, he can work out how many lines for a duodecimo (12 ^mo ) to be set in Petit-Romain instead of in Cicéro or St. Augustin and how many pages per sheet one would gain or lose and hence how many more or fewer sheets a volume would contain printed in such and such a face.

The [relative] sizes being established and known makes the mechanics of printing more reliable and certain; the workman, aware of the space occupied by his type, is able to fill the empty spaces in his works exactly whether in the setting of ornaments or in carrying out any other difficult task that needs judgment. If he has, for example, a space of six lines of Nompareille to fill he knows at once that he can substitute for it four lines of quads of Gaillarde, or three of Cicéro or two of Gros Romain or a single line of Trismégiste. He has a choice and all these will fill and make his white space without care or difficulty.

By this means confusion in Printing is avoided, especially with faces called ‘ two-line letters’  [98] these letters must be cast so that they have exactly twice the body size of those whose ‘two points’ they are; see Two-Line Letters  [99]. But the bodies, whether Petit-Texte, Petit-Romain, or Cicéro being indeterminate in size, larger in one print-shop and smaller in another, it follows that these ‘two-line letters’ having no fixed relationship to their large bodies, form a multiplicity of different thicknesses or bodies in Printing; in spite of this, however, they have no other name than ‘two-line letters’.

To use with these two-line letters there must be quads and spaces made expressly and adjusted to be of the same body height. The table simplifies this; two-line letters of Petit-Texte are cast on the Gros-Texte body; those of Petit-Romain are cast on the Petit-Paragon body, those of Cicéro are cast on Palestine and so on. Thus it is no longer necessary to cast special quads and spaces for two-line letters since those already in existence for the face of twice the body size can be relied upon to fit  [100].

We observed at the beginning of this article that Typefounding has owed much for its progress since its birth until our own times to Simon de Colines, Claude Garamond, Robert Grandjean, Guillaume le Bé, and Jacques de Sanlecque, in the 15 ^th , 16 ^th and 17 ^th centuries, and in the 18 ^th century to MM Grandjean and Alexandre who devoted their efforts to the Imprimerie du Roi.

But justice requires that we do not pass over in silence the contributions of M. Fournier le jeune to the Art since the times of these gifted men. He began by making the important table of the size relations [of various types] that we have discussed above. Then, in search of further useful innovation, he realised that printing lacked large capitals for posters, notices and titles. Those used before his work were too small and of outdated style; the wooden letters were frequently distorted and liable to breakage and rot etc. He cut letters of fifteen lines height thus much larger than the cast letters previously in use; then he continued the series from this size right down to the smallest.

Then he doubled his efforts by making italic letters of the same size [range]; there were previously no letters of this kind available for printing. The largest [italic letters] were two-line Saint-Augustin or Gros-Romain – and these were skinny and ill-cut. It must not be denied that very well-designed letters of this kind were used in the Imprimerie Royale, but only up to a certain size; and for printers in other parts of the kingdom they might as well not have existed.

The [existence of] these large capitals has almost extinguished the custom of printing notices and frontispieces in red and black. The words to be emphasised are brought out sufficiently by the use of combinations of lines of roman and italic whose design renders them sufficiently distinct. By this means double pulls in red and black  [101] are avoided and the elegance of title-pages is improved.

Printing also hardly possessed the little cast ornaments called vignettes. Those that did exist were so old and outdated in style that they could hardly be used. M. Fournier, in imitation of masters Grandjean and Alexandre who had made very fine ones for the Imprimerie du Roi, has created more than one hundred and fifty varieties that he has cut in proportion to the type body size. For example, an ornament cut to be cast on a Cicéro body at half the body thickness only has to be turned over to serve for Nomparaille; another is square and equivalent to Cicéro in all dimensions; another of a Cicéro and a half corresponds to Gros Romain body, one of two Cicéros is for Palestine, and so on. Cast on a body of one size they are suitable for use with other body sizes according to how they are rotated; they present predetermined sizes whose space is exactly filled by [types of] other body sizes, larger or smaller.

In this way, using combinations of the little objects, cast ornaments can be constructed larger or smaller according to need and wider or less wide according to the whim of the type compositor. See examples of these ornaments in the plates of faces at the end of this article.

M. Fournier has made alterations to the cutting of punches for Plainsong music for which he has the gratitude of the printers of the various dioceses whom he has supplied. Sharp and flat signs etc. were cut and cast of various different sizes according to their shapes so that, to compose the music and align the lines, spaces of all sorts of sizes had to be cast – some of them extremely thin. These spaces carried four [parallel] lines; when assembled they produced the same number of breaks in the lines of the note because the alignment was never as good enough for [the defects in alignment] not to be visible; particularly when the type was somewhat worn these breaks became more apparent and could not but be disagreeable. In addition, the workman was always required to justify his line by packing it out as is done using spaces in a line of ordinary letters. To avoid these difficulties M. Fournier cut the punches for the notes, sharps, flats, guidons  [102] , rests etc . all of exactly the same size and the spaces with their four lines of the same thickness, or twice, three, four or five times as large; the thinnest [spaces] are half the thickness of the note. Thus, all the thicknesses being the same and known, when the printer has decided the length of the line, all the lines align automatically; all that need be done is to use the same number of notes or their equivalent in spaces, which is easily done. Once at the end of the line, a half-note is inserted, or its equivalent [space] or the equivalent of a note, or a space equivalent to several notes, according to the void to be filled and the line will be found to be justified. Errors in composition are not difficult to correct because one can always insert the exact equivalent of whatever has been removed. Since there is no need to justify using thin spaces, there are fewer breaks [in the stave] and the result is better.

To bring all this about it is only necessary to make the lines attached to the note all the same length and to put the correct size of note, or other sign, on them as in the following example :

M. Fournier removed from the notes used previously a useless duplication of eight kinds which were ugly as was apparent in the custom of putting the tails of the notes pointing downwards where they became mixed up with the notes below. To avoid this what had to be done? Simply to point the tails in the opposite direction as is done in [writing] Music. This expedient also has the advantage that one finds already in the set of notes the necessary material to do this without having to make it specially.

For example :

turn the notes upside down, and you will have,

that is, the result desired, with less expense, without difficulty and with a better effect.

See the example in the tables of characters that follow.

Much use is now made in printing of rules, single, double or triple, which were not in use ten years ago, thanks to M. Fournier who invented a mould to cast them. They were previously made of copper or brass; they were expensive and never right. It would have taken too much time, and perhaps was not possible, to make flat strips of brass of the exact thickness of the several bodies of type [faces]. The only thing available was [to use] several strips of different thicknesses that were adjusted as best one could. M. Fournier’s mould remedies all this; it is a simple and convenient device fourteen or fifteen inches long by about an inch wide in which are cast strips of fourteen inches long and of the body size [height] of a given type face. The [one] mould serves for whatever height is desired; it only has to be adjusted - which can be done in a moment; the strips are put into the cutter, and with the plane used for letters, and some specially made tools, one cuts whatever pattern one wishes on one of the faces.

This mould for rules was so well received that within two or three months of its use the other Typefounders rushed to imitate it: but their device was larger, more complicated and less convenient to use; master Fournier had not published his design and had reserved it for use [only] in his own Foundry. See the article Rules, for the explanation of the device and, in our Plates of Typefounding , an illustration of it and its details  [103].

To add some variety to an impression and for use in some particular works, M. Fournier has just cut a new face in his series; it is in two parts and on two different bodies. The first, cast on the body of Grand Parangon is called bâtard coulee ; the other which has a larger letter-design is cast on Trismégiste and is called bâtard .

These faces, which carry an alphabet of decorated and festooned letters to be used in place of small capitals, are made to work together and form a set called caractere de finance because they imitate writing. See the sample in the plates that follow.

The most useful of M. Fournier’s contributions to printing, and which does him the most credit, after his table of corresponding sizes, is his modification of italic faces to which he has given a more firm shape in which he has made the open and the black areas more distinct and more similar to handwriting.

At the beginning of this century, masters Grandjean and Alexandre made some changes in the italics that they cut for the Imprimerie du roi; their example encouraged M. Fournier. To help the reader to appreciate his work here are some lines of italics as he found them, and those with which he replaced them.

Former Gros Romain Italic:

New Gros Romain Italic:

To measure the required proportions of the characters and to confirm their size one must carefully measure, on the text, a length of forty lignes  [104] according to M. Fournier’s scale : that is to say thirty-seven lignes in ordinary measure  [105] [and count the number of characters this length contains]. This number of lines should contain : 48 Parisiennes, or 40 Nomparailles, or 36 Mignones and 1 Gros Texte, or 30 Petits Textes, or 26 Gaillardes and 1 Nomparaille, or 24 Petits Romains, or 21 Philosophies and 1 Gaillarde, or 20 Cicéros, or 16 Saint-Augustins and 1 Gros Texte, or 15 Gros Textes, or 13 Gros-Romains and 1 Nomparaille, or 12 Petits-Parangons, or 10 Gros-parangons and 1 Petit-Parangon, or 10 Palestines, or 8 Petits-Canons and 1 Gros-Texte, or 6 Trismégistes and 1 Palestine, or 5 Gros-Canons and 1 Petit-Parangon, or 4 Double-Canons and 1 Gros-Texte, or 3 Triple-Canons and 1 Palestine, or 2 Grosses-Nomparailles and 2 Palestines.

Should one find some large or small faces that are not mentioned in the table of sizes or the preceding list, this is because these large faces are not [now] cast and the small faces such as perle and sédanoise etc. do not have the correct proportions [to fit the scheme], even though the are [still] cast. In any case it is to be hoped that they would be adjusted to the proportions given in the table which would improve the quality of printing and make its practice easier.

It remains only to say a word about the regulations to which Typefounders are subject.

Before they start exercising their craft, Founders are required to present themselves to the syndics and adjoints d’Imprimerie and to be registered in the community register as Fondeurs de caracteres ; this registration is required to be carried out without any charge.

Nevertheless, they are forbidden to work as Boooksellers [publishers] or Printers.

They must reside and work in the University district.

We have seen previously the details of the edict about the regulation of the proportions of type faces. The Founders are required to cast their faces of good and friable metal ( see above for details of the type metal ): to give priority to work for Parisian printers over those in the provinces: not to export any face without first reporting to the office of the Community its details, weight and quantity: to cast foreign faces to the dimensions of the Parisian faces: not to deliver type to anyone except printers.

These, then, are the principal regulations from which one can see how imperfect they are and how doubtful it is that there has been any real advantage in the separation of the professions of Engraver [punchcutter], Typefounder and Printer.

I have spared no effort to explain clearly everything about the first two which are essential preliminaries to the third. I hope that men of letters, who, through their [written] works have some claim to immortality, do not accuse me of being too prolix; as to the opinion of others, I care little. I would have written at much greater length had I not taken it upon myself to pass lightly over the less important procedures. On the other hand, I have attempted to describe the others in a way that respects the craft, and would help to preserve it were it ever in danger of being lost. For the next stages see the article on Printing. Do we owe less to Typefounding through which works of great genius are multiplied and preserved than we do to casting in bronze that celebrates heroes and their deeds? Se e Casting in bronze in the article Bronze.

There follow examples of all the Type Faces in use: they come from the print house of M. le Breton , our printer, and the type Foundry of master Fournier, except only for Perle and Sédanoise , which are found only in the Imprimerie Royale, and which M. Anisson, director of that printing house, was anxious to include.

We direct the reader [also] to our engraved Plates of the alphabets of most peoples, both ancient and modern.

Examples of all the type faces roman and italic in use in printing .

Notes

1. les poinçons This description is, of course, an abbreviation. The type is cast from matrices which in turn are stamped out by the punches (les poinçons).

2. graver les poinçons ,

3. fondre les caracteres

4. caracteres mobiles I shall now use ‘type’ to describe ‘moveable characters’ in general and reserve ‘character’ for a single piece (stick) of type or for the image in relief it carries.

5. There is no reason to suppose, that, as Diderot claims, it was contemplation of wood-block printing that inspired the development of moveable type – except possibly in the very general sense of the realization that it is possible to apply an image to paper using an inked carving in relief. The wood-block books were generally of dismal quality and were not at all attractive in a market accustomed to manuscript books.

6. Jean Guttemberg, [sic] gentilhomme Allemand

7. Once again, this is misleading – though the sentiment is just, that punchcutters and typefounders have not enjoyed the public reputation accorded to printers. After the very early days of printing it was rare for punchcutters to be employed within print shops; they rapidly became a set of specialists on their own. Type was sometime cast by the printers themselves but even in the sixteenth century was often supplied by specialist typefounders who did, usually, employ punchcutters.

8. The text has l'art typographique. But Diderot must mean punchcutting and type casting and not typography which he describes below, correctly, as an art on its own. The example of the engraver versus the copperplate printer makes the point, but is unduly dismissive of the poor printer and his pressmen. Though, indeed, the printer does not engrave the plate any more than the framer paints the picture, his skill in printing from it is very far from negligible in the quality of the result - as is painfully apparent when one compares good and bad pulls from the same plate.

9. This is a painfully simplistic view. My admiration for the early punchcutters is unbounded. But, surely, to praise magnificent printers is not to denigrate the cutters and founders of their type? The achievements of the French punchcutters and typefounders of the sixteenth century can only be described as magnificent. They, more than any others, laid the foundations of modern type design and converted most of Europe from heavy ‘gothic’ to elegant and much more legible ‘roman’ faces. But without the skill of the great printers their magnificent type would have gathered dust unseen - indeed it would never have been cast, for it was those famous printers who ordered it and caused it to be created to their requirements - and paid for it - often long in advance of its delivery.

10. In fact, Simon de Colines was born in 1480 and was not active as a printer until about 1520.

11. Robert Estienne the first, 1503-1559. The typeface in question was the Grecs du Roi cut by Garamond and used by Estienne to print Greek texts.

12. Diderot appears to be speaking of Granjean, and the next paragraph makes it clear that it is specifically his italic faces that are in question. Granjean’s italics were - are - indeed, justly renowned.

13. There does not appear to be an article on Italic type.

14. The reference to the plate is incorrect. It should be to figure 1 on the first (unnumbered) plate in the series on typefounding labelled Fonderie en Caracteres Planche . The link in the text is to the correct plate.

15. l'équerre See Fonderie en Caracteres Planche figs 6 and7 on the lower row of figures.

16. .... sans elle on ne parviendroit jamais à bien dresser la petite face du poinçon, sur laquelle la lettre doit être pour ainsi dire découpée. The point is that the face carrying the letter in relief must be absolutely perpendicular to the sides of the shank of the punch which must also be parallel to each other so that they will lie squarely against the jig or ‘square’( l'équerre ) that holds the punch as it is ground on the oilstone.

17. See plate Fonderie en Caracteres Planche fig 5 (incorrectly labelled 6) at the right of the top row of figures.

18. Vignettes So named because ornaments were originally confined to drawings of grapes and vine leaves.

19. Burin

20. Again, the plate reference in the French text is incorrect; it has been corrected in the translation and link.

21. la lime a enlevé en talud les parties qui excédoient les contours de cette lettre.

22. Once again, reference to plate corrected.

23. The text clearly does not refer to the figure that was published. To add even more confusion, the figure legend refers to letters not present in the figure, and Fig 6 appears twice on the plate – the first time erroneously for Fig. 5! I have corrected and rewritten the text to make sense in relation to the figure 6 (on lower row) and 7 of Fonderie en Caracteres Planche in which the shape of the square or jig and its relation to the stone is perfectly clear.

24. In summary, the heights of the letter bodies must be equal and, for the lower case, the ascenders and descenders of all the letters must be equal in length. This is very important - imperfect sizing of the punches, and thus the matrices and the type cast from them, leads to type that can never be set to look properly aligned.

25. Fonderie en caracteres The multiple meanings deriving from the use of Fonderie both to mean ‘casting’ and as a concrete noun, ‘foundry’ do not arise in English. Typefounding or Typecasting is the term for what Diderot calls l'art même de les fondre the ‘technique of casting letters’ The place where they are cast is, of course, a Type Foundry.

26. réception Presumably, being allowed to enter the craft - for example by formal apprenticeship.

27. tiroient les empreintes des matrices – ‘extract the images from the matrices’ - a somewhat picturesque expression which the following sentence explains. Diderot means simply ‘cast the type’.

28. The text says d'une épaisseur & d'une hauteur égale but type of one body size is not all the same width - this varies with the character; it is - it must be - all the same height (vertical dimension of the body parallel to the paper) and of the same height-to-paper (the distance from the bottom of the slug of type to the printing face of the character it carries).

29. cuivre de rosette Pure copper. In French cuivre rouge is copper, cuivre may be either copper or brass ( cuivre jaune ).

30. Recuit(s) Copper hardens very quickly when it is worked in any way (‘work hardening’) and becomes very brittle. Thus it has to be softened by heating and slow cooling ( recuit ) - annealing - before it can be worked again. If this were not done the punch would not stamp an impression of the letter, the copper would fracture. The verb recuire is used because, in common practice, tempering - of steel for example - by heating to high temperature and cooling very rapidly - usually by immersion in water – (‘ rougir et tremper’ ) which hardens the metal, is usually followed by controlled softening - annealing - for which the metal is re -heated - ‘re-cuit’ - then allowed to cool slowly. Hence , recuire came to mean the process of annealing.

31. moule de corps de letters This seems to suggest that the matrix forms the body of the type; this is inaccurate; the function of the matrix is to shape the letter in relief on the end of the stick of type. The body - the ‘stick’ is formed by other parts of the mould, of which, as Diderot has just said, correctly, the matrix is only one part.

32. The figure is that of the seated woman on the left of Figure 1 of the un-numbered plate that begins the series on typefounding. According to the figure legend she is filing the la partie extérieure de la lettre ; The details of her work are not really visible in the plate but she appears to be filing a punch rather than a matrix. I suspect there is no illustration of the filing of the matrix. Indeed, the next sentence would make more sense applied to the punch than to the matrix – though it is certainly necessary to adjust the newly-stamped matrix – a process called justification – not to be confused with the use of the same word in printing!

33. The mould for casting type was always called the hand-mould in English; for brevity I shall simply call it the ‘mould’ here.

The figures of the parts of the mould, which is a complicated device containing about 50 components, are necessarily not very easy to understand. But, to make confusion worse confounded, the legend for Plate III is attached to Plate II, the legend for Plate II is attached to the plate labelled Iere - which is, in fact the second plate, and the first plate, instead of being numbered 1 has no number. The third plate is labelled Suitte de la Pl. II. in the top right corner . In addition (see note 23) there are errors in the numbering of the figures on the first plate, and the numbers in the text do not always correspond to those in the figure or in the legend. I have done my best to make sense of the resulting chaos....

34. That is, most of the components are paired with one in each half (upper and lower) of the mould.

35. They are named thus by Moxon

Moxon, J. Mechanick exercises or the doctrine of handy-works . London. Printed for Joseph Moxon. 1677. vol. 2 pp 131; 141ff and fig. 18.

36. The French text erroneously says 20 instead of 18.

37. Potence an L-shaped piece, see figs 9 and 10 of Plate II.

38. The head of the screw is visible in Plate II fig 21.

39. This is the round hole under the letter b of fig 17 in Plate III; the hole is also visible from the other side in fig 18.

40. Figures 17 and 18 show the hole penetrating the long piece completely!

41. D in figs 9 and 10, Plate II

42. There is no X in figs. 9 and 10 but the square end of the potence is clear.

43. There is no d in fig. 21. The nut on the threaded end of the potence is probably that at the bottom left, with the letter f below it, of the left of the two figures 21 on plate III. This plate is labelled Suitte de la Pl. II. in the top right corner .

44. There is no figure 26! see Fournier plate 5 number 4 for an illustration of the nut-spanner.

Fournier, P.S. Manuel typographique. Vol. 1, Paris: Imprimé par l'Auteur, rue de Postes, et se vend Chez Barbou, rue S. Jacques; 1764

45. Figs. 5 and 6 of plate II

46. There does not seem to be any figure corresponding to e o.

47. The lower Fig. 2 and Fig. 3 of Plate II and figure 20 of Plate III

48. Fig. 6 of Plate II.

49. The nut is that marked a on the left fig. 21 at the bottom of Plate III.

50. See figs 17 and 18 of Plate III. It seems that the screw has a shaped tenon that slides into a dovetail - - queue d’aronde - cut in the long piece.

The rather charming French name queue d’aronde ‘swallow tail’ reminds us of the old name arondelle of the bird now called hirondelle , – as, for example, in the Rue de l’arondelle of the old Quartier Latin in which the famous surgeon Ambroise Paré had his house in the mid sixteenth century.

51. There is no figure 19 but the register, E, and the nut that fixes it, F , are visible at the bottom of the right fig. 20 of Plate III.

52. Figures 1 and 2 of Plate II.

53. The matrix is not visible in Fig.2 and there seems to be no part marked M. However, the matrix illustrated in Figs. 12 and 13 does carry the impression of the letter m – apparently a lower-case m.

54. attache and jimblet These components are described more fully by Fournier Tome 1 p. 196. See note 43 for details of Fournier’s Manuel typographique .

55. The left figure 21 of Plate III.

56. Shown in place in the right figure 21 of Plate III.

57. The hook is not visible in figure 2, Plate II.

58. Now realised to be only partly purified, regulus of antimony was probably the purest form of the metal available in the 18 ^th century. Fournier le jeune, (Pierre Simon Fournier) the famous Parisian type founder describes in great detail the manufacture of the regulus from crude antimony. He insists on the requirement for well-purified antimony to make good type metal.

Fournier, P.S., Manuel typographique . Paris: Imprimé par l'Auteur, rue de Postes, et se vend Chez Barbou, rue S. Jacques; 1764 p. 115 ff.

59. potin I have left this word untranslated because its definition is very imprecise; according to some definitions it is a mixture of cuivre rouge (copper) and cuivre jaune – ‘brass’; but brass now refers only to an alloy of copper and zinc. Fournier – an expert – describes potin as ‘ du cuivre cru dit potin’ . I suspect that the metal to be used was, in fact, tin but doubtless in an impure form. Tin certainly is a component of modern typemetal which is an alloy of lead, antimony and tin in various proportions according to its exact use. Fournier regarded the typemetal containing potin as much inferior to that made from lead and regulus of antimony. See Diderot’s next sentence....

60. Diderot has not made clear that, in the second kind of typemetal, the potin was to be melted separately with antimony and the mixture added to the lead. See Fournier p. 109 ff for an account of this alloy and its disadvantages.

61. In fact, no tools are to be seen on the floor!

62. Matiere forte, matiere foible

63. There is a casting furnace with three workmen on the left of the picture in the first plate, but none on the right.

64. Plate I upper image shows the furnace on the left. The details are shown in the lower drawings of Plate I. See fig. 10.

65. cuilliere I have called this vessel the ‘furnace ladle’ to distinguish it from the much smaller ‘casting ladle’ cuilliere à verser while still preserving the notion of a spoon-shaped receptacle.

66. puise dans

67. cran the nick, see also note 67.

68. French type was eccentric in using the nick to mark the top of the letter. In England and most other countries the nick was placed on the side of the shank corresponding to the bottom of the letter.

69. The text says puiser dans le moule but this is clearly an error.

70. The text has, erroneously, fig. 2 of Plate II

71. The girl - marked fig. 2 on the right of the image nearest the hearth is rubbing the type on a grindstone. The text refers, erroneously, to fig 7 of Plate III.

72. Note here on kerns and kerning?

73. Presumably the sequences ava, vav are intended to indicate mixtures of letters correctly and incorrectly arranged. The point is that all the letters are arranged the same way up and the same way round - as the nicks will show.

74. There are three plates missing from the series on typefounding. {they will eventually be added} The parts of the plane are on a missing plate labelled at top right ie Suite de la Pl III and it is to this plate that Diderot’s description refers.

75. The ‘height to paper’. For obvious reasons it is essential that all the type on a page has the same overall height to paper or it will be impossible to pull an even impression from the page.

76. Hausses shims would be an appropriate translation.

77. Text has, erroneously, fig. 1 .

78. There is now considerable confusion between fount ( font in American English) and face . This is mainly due to the careless use of ‘font’ in computer terminology to mean sometimes face , sometimes fount and sometimes both at the same time. In modern usage before widespread use of word processing by computer, the distinction was clear. A type face is set of type of a single design. Each face contains an arbitrarily large number of founts where is fount is a set of type all of the same face and all of the same size. Thus Times New Roman is a face; Times New Roman 10 point is a fount of Times New Roman. Other point sizes of the same face comprise other founts. The distinction between face and fount is well worth making. Note that the italic styles of a type design are different faces to the romans.

An example of the confusion: To say ‘use a different font’ may mean use another design of type (a different face) or use another size of the current design (another fount of the same face) or use a different size of a different design (another fount of a different face). Nor is : ‘use a bigger / smaller font’ clear. It may mean a bigger or smaller point size, but it need not mean this since different faces often have different sizes as well as weights of letter in the same point size. To describe particular type, a face and fount must both be specified.

In the translator’s opinion the word font in English English should be used only in connection with baptism. In American English, font should be used to mean what it used to mean before it entered computer jargon, and should be carefully distinguished from face . But I do not suppose many will follow this advice.

79. ‘thickness’ not the width of a letter which varies between different characters, but the vertical dimension of the type body parallel to the paper. As the text explains, this dimension determines the spacing between lines of print - provided, of course, the type is ‘set solid’ that is, provided no extra spacing (leading) is inserted by the compositor between each line of type. See also note 79 below.

80. For the modern reader this is all rather confusing.

In the eighteenth century, and later, type was described by a name that indicated both the face design and the body size which was, in principle, fixed for type of that name. When line spacing was to be varied this could be done by casting the named face, for example Cicero, from its usual matrices but changing the body size by adjusting the hand-mould to a larger body size such as that normally used for a larger type such as St Augustin. The way in which this was done was described earlier in the article. Hence ‘Cicero on St Augustin body’ .

We now name faces according to the design of the letters on the type and not the size of the type. Sizing, is of course expressed by the body size measured in points and, when using hand-set metal type, white space between lines was increased by adding shims – leading – between the lines of type as it was being set.

This has all changed with the advent of digital type in which glyphs – vector representations of the shape of the letter, that is, effectively a set of instructions for constructing the shape from a series of lines, are used. These are almost infinitely scalable, so 10, 12, 14 etc. point sizes are generated by multiplying the vectors of the glyph by variable factors and the leading is adjusted by altering the position of the digital images of the letters relative to the paper. So in computer typography one can generate an almost infinite number of founts from the glyph set of a single face . But it is still convenient to describe a block of type as, for example, Times New Roman 12 on 14 point meaning that the final effect is that of setting 12 point Times Roman metal type with a leading of 14 points.

81. corps, oeil,& blanc

82. Letters without ascenders or descenders.

83. If the parallel lines are numbered 1 to 7 starting at the bottom the ‘short’ letter fits exactly between lines 3 and 5.

84. s the long s, somewhat similar to a lower case f, which is no longer is use in contemporary type.

85. NB the height of the letter body – the x-height of the letter and not the body height of the type! Note the confusion caused by the use of caractere in French to mean both type face and also letter!

86. Diderot writes as a good Frenchman.... It is true that German text printed in one of the so-called gothic faces has a very heavy appearance; the page looks very black. Not for nothing were gothic faces called black-letter! But the lack of white space is really due to the design of the gothic faces rather than the prolixity of the language, though it is also true that the German of Diderot’s time was rather prolix compared to contemporary French.

87. par cent pesant by the hundred pounds, that is by weight.

88. The amount of type needed to print one sheet.

89. distribute To break up the forme and replace each sort - piece of type - in the correct pigeon hole of the compositor’s cases. In the hand-press period printers did not usually have enough type to keep more than a few formes set up. The required number of sheets was printed from each then they were cleaned and broken down to free the type for reuse once it had been distributed to the cases.

90. police This derives from the usage of police to refer to the regulation of a state of affairs and has no criminal connotation. Thus the police for type is the set of rules based on the relative frequency with which each of the various characters is required and so the proportion they should form of the total number of pieces of type.

91. pié de Roi The standard foot (pied, pié) of the Ancien Régime was divided into twelve inches (pouces).

92. The definitions of the line and point in the text do not correspond to their values in the table; the text is simply wrong! Even when the correct values of six points per line and twelve points per inch are substituted for Diderot’s values, the table makes a perfectly simple concept quite unnecessarily obscure. But Fournier’s admirable tables in his Manuel Typographique make all clear. See note 43.

The scale of two inches’ length is divided into 144 points; one inch is thus 72 points. The body sizes are defined so that each can be made up as a multiple of a smaller body or as the sum of multiples of smaller bodies. For the larger type sizes many such combinations are possible and the centre column lists these, or some of them.

In the rightmost columns the body size of the type named in the leftmost column is expressed as a sum of x Lines + y Points, where a line contains 6 points. Note that the size is given opposite the last line of text in the middle column relevant to that type. It is very easy to read the wrong size by taking the figures from the line below which nearly aligns with the next index size. The key to understanding the table is to convert the Lines + Points sum to points and to use only this number in making comparisons. A few examples may help:

The smallest type, Parisienne has a body of 5 points (0 lines, 5points)

Nompareille is 6 points (1 line, 0 points as expressed in the table)

Cicéro is 12 points (2 lines, 0 points) and is equivalent to (middle column) 2 Nompareilles ( 2 x 6) OR 1Parisienne + 1 Mignone (5 + 7)

And so on with the number of possible combinations increasing with increasing type sizes.

Anyone who wishes to do calculations with this system would be well advised to ignore Diderot’s table and to work from Fournier’s tables which simply express the size of each body in points. The relationships between the type sizes are then transparent. According to his own account Fournier invented the points system in 1737; in fact, Truchet had proposed a (different) points system earlier. Fournier did an enormous service to typography though his point size was later modified by basing it on the inch of the Pied de Roi - the official French foot - rather than on Fournier’s inch. The point varied in size in different countries at different times though it was always approximately 1/72 of an Imperial inch. The point has now been standardized as the DTP (desktop publishing) point with 1 point = exactly 1/72 inch in British and US units, and this is now the standard measure for type size.

93. In the scale at the top of the table each inch is divided into 12 divisions; these are the ‘lines’ each of which is further divided into 6 points. The small divisions at the right of the scale are nominally 2 points wide but the execution and printing of the scale is not good enough for this to be of any practical use.

94. leurs hauteurs en parties de l'échelle This refers to the body ‘height’, that is, in English usage, the body or the point-size.

95. demande Present tense, but this is all clearly to be read as applying before Fournier’s revolution.

96. un échantillon en lettres This reads oddly; from Diderot’s usage one would expect en caracteres but he must mean that physical type was supplied by both printers since there was previously no accepted way of describing type size.

97. becs de plume

98. letters de deux points These are called two-line letters by Smith in his Printer’s Grammar of 1755. The second edition (1787) pp 113-115 includes a fuller description of these large letters. See note 98 below.

99. There appears to be no article Lettres de Deux Points . However, Fournier explains these types ( Manuel Typographique Tome 1 p.168) :

Les lettres de deux points , ainsi nommées parce qu'elles occupent l'espace de deux lignes au commencement du texte de l'ouvrage où on les emploie, ont la figure des capitales, et sont fondues de façon que leur corps soit justement le double du caractère dont elles font les deux points, et avec lequel elles se trouvent en ligne par le bas.

Letters of two points are so called because they occupy the space of two lines at the beginning of the text of the work in which they are used. They are shaped as capitals and are cast in such a way that their body is exactly twice that of the face for which they form the ‘two points’ and with which their bottoms align..

These letters could also be used as drop capitals. Further explanation follows in Fournier (pp 168-171) and they are illustrated in his second volume. See note 43 for details of Fournier’s Manuel.

The whole subject of large letters and their use in eighteenth century printing and distinctions such as those between Double Letters and Two-line letters is confused and confusing. For some sense of the confusion that reigned before the introduction of the points system for type measurement one need only dip into the two editions of Smith’s Printer’s Grammar .

Smith, J. The printer's grammar: wherein are exhibited, examined, and explained, the superficies, gradation, and properties of the different sorts and sizes of metal types, ... sundry alphabets, ... figures of mathematical, astronomical, musical, and physical signs; ... With directions to authors, compilers, &c. . .. By John Smith, London: printed for the editor; and sold by W. Owen; and by M. Cooper, 1755.

Smith, J. The printer's grammar: containing a concise history of the origin of printing; also, an examination of ... the different sizes of types ...With directions to authors, compilers, &c. How to prepare copy, and to correct their own proofs. Chiefly collected from Smith's edition. To which are added directions for pressmen, &c. ... London: printed by L. Wayland; and sold by T. Evans, 1787.

100. Thus, for example, a two-line face based on (12 pt) Cicéro would use quads and spaces of Palestine (24 pt).

101. Printing in two colours - usually read and black - was a tricky business (see Brullé’s article on printing) that required two separate impressions, one for each colour, and the results often left much to be desired. In spite of this, some printers, particularly in the early hand-press period, achieved beautifully aligned and strikingly effective blocks of text in contrasting red and black.

102. guidon A sign placed at the end of a line of music to indicate the position of the first note of the next line. By analogy with catchword at the end of a page one might call this a ‘catchnote’.

103. Plate IV Moule à Reglets & à Interlignes.

104. I have left the French lignes because it is not a number of lines of text but a vertical height across the lines of text, of 40 lignes length, that is to be measured! The French use of ligne to mean both a line (of text) and a unit of size invites confusion; since we also have here two different measuring standards for length - Fourier’s pouce and the standard pouce (de Roi) confusion is worse confounded. See note 104.

105. lignes géométriques The Paris line was approximately 1/12 inch; (it was 1/12 pouce which in turn was 1/12 pre-revolutionary pied (pié) . The modern standardized conversion is 1 ligne (line) = 2.2559 mm. Fournier’s pouce, which contained 12 of his lignes or 72 points, was apparently smaller that the standard pouce of the time since the height of text to be measured was 40 Fournier lignes or 37 standard lignes .

Diderot does not really make clear what was to be measured. In fact it was a height, that is to say it was a distance to be measured parallel to the vertical edge of the page, from the bottom of a line of text downward. The required height of text to measure was 40 x 6 = 240 of Fournier’s points. Since his table gives his body sizes in points it is easy to calculate that this distance should contain, for example 20 Cicéros, that is to say, 20 lines of text set in the Cicéro face .

The operation is the inverse of the modern method of estimating the body height of a given type by measuring the vertical height of 20 lines of the type rather than measuring a fixed distance and counting the number of lines it contains.