All measurements of type originally referred to the dimensions of a cast piece of metal bearing a single character. One pound of type, in any point size, covered about 3 square inches.
The height, called type height, is the distance from the face that touches the paper to the feet. It is important only to the letterpress printer, to whom it is all-important, and is dealt with at the end of this article.
The width depends on which character the piece of type will print. A width with a special name is the em, the width of a square piece of type, one whose face is as wide as it is long. The name comes from its typically being the body of the capital letter “M.” When type was set by hand, in America the quantity set was measured in ems. In England it was billed in ens, an en being half the width of an em. The exact dimensions of an em or en depend, of course, on the length of the type. The length of the type determines how high the printed character can be, and this is the body size.
The size of type is usually measured in points, a unit of length in use since 1735, with various values. In the United States it was formerly 0.013837 inch (72 points = 0.996″), but with the rise of digital typesetting the value 0.013888…, that is 72 points = 1 inch, has become more usual. Twelve points = 1 pica; 6 pica = 1″.
Type sizes were originally named; catalogs with such names appeared as early as 1592.
Some of the names came from the type of book typically produced in that size. “Cicero” was a size used to print editions of classical authors; “Primer” was used to print religious books ordered by Henry VIII.
Another class of names were boasts of the type's beauty, such as “Paragon” and “Nonpareil.”
The problem with using names was that there was no clear relationship between the names, and they were not related to well-defined linear units like the inch. In the nature of printing by letterpress, to get the type on the chase to lock up properly it all has to fit together.
Another disadvantage was that sometimes a name used to describe a body size was also used as the name of a typeface. “English,” for example, meant a typeface in the style called blackface as well as approximately 14 point type.
The French addressed the problem in 1723 with a royal order that the sizes of type be fixed. 1735 Pierre Simon Fournier points. first in 1737, and final version in 1764. In , 72 points to the pouce (1 pouce was then about 2.707 centimeters or 1.066 inches). By 1764, however, Fournier had dropped the pouce and instead defined his point by a (badly) printed scale. It seems likely, as Theodore de Vinne speculates in The Practise of Typography (footnote, page 141), that in the intervening years Fournier had adjusted his point so that it would fit existing sizes of type as well as possible. In Fournier's system the cicéro (which plays the same role on the continent that the pica does in English typesetting) is 12 pts. As it was used in France in the 19th century, one of Fournier's points was approximately 0.35 millimeters.
The biggest shortcoming of Fournier's system was that it was not related to any other system of units. To remedy that, around 1785 François-Ambroise Didot, a well-known Parisian typefounder, established a new system which really returned to Fournier's 1737 definition of 72 points to the pouce, so one Didot point was approximately 0.376 millimeters. Didot abolished all names, replacing them with numerical sizes. In doing so he was forced to make the basic size, cicéro, 11 points–which may have been what Fournier was trying to avoid. Twelve is a very convenient basic size, since it is easily divided into halves, thirds, and quarters that can be built up from 1-point pieces. Even half of eleven is a special, non-integer size. Nevertheless, after many decades in which both Fournier's and Didot's systems were used side-by-side, Didot's prevailed, and is currently in use in Europe (except Belgium) and certain other countries.
Meanwhile, across the English channel and across the Atlantic, type sizes continued to be known by names, not numbers. Many attempts were made to introduce a system but no system was accepted by a majority of type founders and printers. Perhaps the most interesting failure was that proposed George Bruce of New York in 1882. In his system sizes increased by the sixth root of 2, so that each size was 112.2462% of the size before it and double the size seven sizes down.
The Great Chicago Fire of 1871 destroyed the premises and all the equipment of a prosperous type foundry, Marder Luse, and Co. In rebuilding, the foundry decided to adopt as its pica size the “Johnson pica,” used by seven of the largest American foundries, including the most successful, MacKellar, Smiths, and Jordan of Philadelphia. Richard Hopkins3 has described the results.
Marder Luse had an outpost in San Francisco run by an inventor named Nelson C. Hawks. Hawks was forced to stock type and supplies from many foundries, noted that pica type was ¹⁄6 of an inch high. Nonpareil was half the size of pica. In between the two were 5 other sizes. What whole number has 6 whole numbers between it and its double? The answer is 6 with 7, 8, 9, 10, 11 before 12. So Hawks said, let's call Nonpareil 6 point. From that all the other named sizes can be given a point size very close to their actual size.
Hawks persuaded Marder, Luse to adopt his system and then, with the enthusiasm of an inventor, sold his interest in the firm and devoted himself to proselytizing for his system. In 1886 the U. S. Typefounder's Association adopted it, with a twist. They wished to tie any new system to the metric system, and so they resolved that 83 picas would equal 35 centimeters exactly. This means 1 pica = 4.217 millimeters and 1 point is approximately 0.35 14 millimeters (approximately 0.01383486 inches), so that 6 picas = 0.996 inch, to the confusion of generations of fledgling typographers. The system became the standard in the United States and Great Britain.
But there is more to the story. The Johnson pica came from the Lawrence Johnson foundry, whose assets MacKellar, Smiths and Jordan had purchased. The Lawrence Johnson foundry, in turn, had been the successor to Binny and Ronaldson, America's oldest type foundry. The standards Binny and Ronaldson used in sizing their molds came from typefounding equipment Benjamin Franklin purchased from Fournier in the early 1700s. In Fournier's system the size Franklin called pica would have been cicéro, 12 points, 6 of which would have been 0.992 inches (instead of 0.996). But in a century and a half of reproduction wear on the moulds could easily have increased the size of the type by 0.004 inch. The American Point System, which Hawks believed he had invented, was probably Fournier's system plus a little wear.
Recently the programmers who devised the digital typesetting systems, like Didot unwilling to perpetuate so mindless a standard, have rationalized the point, so that it is now ¹⁄72 inch in software the world over. In Canada, 1 pica = 4.0894 millimeters.
Until 1886, in North America the height of the type or plates, from the surface of the platen to the face of the type, was 11⁄12 inch = 0.9166…. In that year the U. S. Type Founders Assn. adopted a new type height when it adopted the point system. Having discovered that 83 pica = 35 centimeters, it decided to let 15 type heights = 35 centimeters. So type height became 23.333… millimeters, approximately 0.918 inches. Although only 0.002 inches higher than the old standard, this difference is great enough that old and new type could not be mixed in the same line.
This value was adopted in Britain in 1898.
The French regulation of 1723 fixed type height in that country at 10.5 ligne, (de Vinne, pg 133; Fournier vol 1, page 125) approximately 23.686 millimeters. In modern times it has been 25.56 millimeters. Germany standardized on the French standard, thanks to Heinrich Berthold, a Berlin type founder who at his own expense distributed precision standards to other German foundries. In the Netherlands, type height is 24.85 millimeters; and in the former Soviet Union, 25.1 millimeters.
Pierre Simon Fournier.
Fournier on typefounding; the text of the Manuel Typographique (1764-1766) translated into English and edited with notes by Harry Carter. New ed. with a foreword and supplementary bibliography by the author.
New York: Burt Franklin, 1973.
Reprint of the 1930 edition published by the Soncino Press, London. A translation of Manuel Typographique, Utile aux Gens de Lettres, & à ceux qui exercent les différentes parties de l'art de l'Imprimerie., Paris, 1764.
Richard L. Hopkins.
Origin of the American Point System for Printers' Type Measurement.
Terra Alta, West Virginia: Hill & Dale Press, 1976.
At present writing (2004), this book is available from the author at P. O. Box 263, Terra Alta, WV 26764. firstname.lastname@example.org
For all columns but the last, ems to the foot have been converted to points at 72 points = 1 inch, and rounded to the nearest tenth of a point. Sources for the data are given below the table.
3 London foundries,
3 American Foundries, 1856
|+||after U. S.
|English, or 2-line Minion||same||13.1||13.5||13.5||13.5||13.4||13.6||14||14|
|Paragon, or 2-line Long Primer||same||—||19.4||19.4||19.4||—||19.2||20||20|
|Double Small Pica||Double Pica†||22.7||20.8||20.8||20.8||21.1||21.6||22||22|
|Double Pica||2-line Pica†||—||24.2||24||24||24||24.2||24||24|
|Double English||2-line English||26.2||27||27||27||26.8||27.2||28||28|
|Double Great Primer||2-line Great Primer||—||33.9||33.9||33.9||33.2||36.3||36||36|
|Meridian, 4-line Small Pica or Canon||2-line Double Pica||—||41.6||41.6||41.6||42.1||43.2||44||44|
*Moxon mentions small pica, but says it differs so little from pica it is too dangerous to use, that workmen will accidentally mix the fonts.
†In most type size names “2-line” and “double” are synonymous, so that 2-line english is the same size as double english. But the introduction from Holland of small pica around 1700 led English foundries to treat “double pica” and “2-line pica” as distinct sizes, the former twice as large as small pica, and the latter twice as large as pica.
(1) Joseph Moxon, Mechanick Exercises, London 1683, pages 13 and 14.
(2) Smith, Printer's Grammar, 1755.
(3) William Savage, Dictionary of Printing, 1842, page 802. “A” is Caslon; “B,” Figgins; and “C,” Thorowgood & Besley.
(4) Printer's Miscellany, New York, July 1857. Theodore de Vinne, The Practice of Typography, 2nd ed. Century, NY, 1902.
Copyright © 2000 Sizes, Inc. All rights reserved.
Last revised: 16 August 2004.