The Measure of Type

Agates. Ciceros. Nuts. Even people who use type every day may not know these weird and obscure measuring terms for the job of setting type. But once you understand them, you can tap into their power to make your work better and easier.

Relatively Speaking
One thing that makes type measurement unique is that it uses two parallel systems: absolute measurements and relative measurements.

Absolute measurements always have the same, fixed values. Inches and feet are absolute measurements. For type in the English-speaking world, the basic absolute measurements are picas and points.

Relative measurements have no fixed value — the dimensions they express vary according to the situation. The fundamental relative measurement is the em. The size of an em depends on the size of the type you’re using. In 10-point type, an em equals 10 points. The practical result of this is that everything measured in ems or fractions of an em always remains in constant proportion to your type even though the size of the type itself might change.

Although its name derives from the letter M, the width of an em is not related to the width of that letter. As shown below, an M can have number of widths, and it’s not normally a full em wide.

Figure 1. Not all Ms are created equal, or equally wide. The red lines indicate the full width of an em for each column of characters.

The only typographic element that you can count on to be exactly one em wide is an em space.

Measurements expressed as fractions of an em have scads of uses. The widths of characters in a font, for example, are commonly expressed in thousandths of an em. (The TrueType and OpenType specifications allow for character widths to be expressed in units as small as 1/32,678 em, but that’s overkill by most measures.) By comparison, many venerable typefaces — such as Times New Roman — that were designed for hot-metal machine composition have character widths based on only 18 units to the em.

Figure 2. Typefaces created for the Monotype composing machine were designed with character widths expressed in increments of one eighteenth of an em. Below are two characters superimposed on an 18-unit wide em square. This shows that in the Monotype system, the W was 17 units wide and the y, 9 units wide.

Kerning adjustments, which are also stored within fonts, are also expressed in thousandths of an em. Most typesetting programs follow suit for manual kerning adjustments, as well as for tracking values.

The charm of relative measuring units in all these cases is that the same values apply no matter what the size of the type is. A character that’s 500 units (half an em) wide will be 6 points wide at 12 point, and 36 points wide at 72 point — always half an em. Likewise, you can kern the numerator and denominator of a fraction, raise its point size by a factor of 10, and the positions of the elements of the faction will remain proportionally constant and correct.

Figure 3. Because the kerning adjustments used to position the characters of the fraction on the left were expressed in relative units — fractions of an em — they have the same net effect even when the fraction is tripled in size, on the right.

Traditionally, ems have also commonly been used to define paragraph indents, because this automatically reflects a proportional relationship between the size of the type and the indent that goes with it. No popular word-processing or page-layout program lets you do this, though. For them, indent values have to be explicit numeric values. That’s too bad.

Ems Divided
A half of an em is an en. If you consider the shape a lowercase m, this is a very clever name. But because em and en sound so much a like, an en was often called a nut in type shops of yore. This gave its name to certain other characters whose widths were half an em wide, such as the nut fractions shown in Figure 4:

Most spaces (thins, hairs, etc.) are based on the em. They have no fixed width, but they’re nevertheless called fixed spaces. This term doesn’t relate to their absolute size, but to the fact that their widths are never altered during justification. Only word spaces can be flexed that way, and along with figure spaces (the same width as the numerals in a typeface) are the only commonly used spaces that are not derived from the em.

The names for em-based spaces smaller than an en are not standardized. A thin space, for example, has traditionally been defined as either a quarter or a fifth of an em. InDesign and QuarkXPress define it as an eighth of an em.

The Absolutes
For centuries after Gutenberg invented his system for making and using movable type, there were no standards for how type was sized. Type sizes tended to be named after famous works they had been used for, such as Cicero in France, and pica in Britain (a pica being a collection of church regulations; it comes from the Latin for magpie, perhaps because of the black-and-white outfits of both cleric and bird). Both of these, as it turns out, correspond more or less to what we today would call 12-point type.

The French — ever the rationalists — tried to organize things, through a series of regulations that found life in the hands of Pierre Simon Fournier, who created types built on a standardized measuring system of pouces (inches), lignes (lines), and points. The famous Didot printing family thought they had a better idea, and prevailed in redefining these values to a somewhat larger size, and their cicero (.1776 inch) and Didot point (.0148 inch) became European standards. So entrenched did it become in the publishing industry that it withstood the onslaught of the metric system, to which France converted in 1801.

The American point system was finalized only in the late 19th century, creating in North America a competing standard, with a new value for the pica (.1660 inch) and for the point (one-twelfth pica, or .0138 inch). By the early 20th century, it had been generally adopted in Great Britain as well.

These standard stuck until the invention of PostScript by Adobe systems, which rounded out the value of the American pica to exactly one-sixth inch and one point to exactly 1/72 inch. While the rest of the world may be going — or have long gone — metric, typographic systems are still based on the inch.

In most of the English-speaking world, legacy documents that predate PostScript are based on the American system. In Europe, older documents are based on the Didot point system. That both are yielding to the PostScript point system is a good thing, because a single standard is better than two.

That said, both InDesign and QuarkXPress can work in any of the three standard measuring systems, either for those who need to adapt legacy documents or those who simply prefer to work with older standards. In both programs, the choice is effected through their Preferences. In XPress, go to the Print Layout/Measurements dialog box and make your selections from the pop-up menus for the Horizontal and Vertical fields. If you choose Picas, you can assign the value you prefer: PostScript or American (but you have to remember the numeric value — they don’t give you any hints). In InDesign, go to the Units & Increments dialog box and choose Picas or Ciceros. A pop-up menu handily lets you choose between American or PostScript points.

Figure 5. In the InDesign Units & Measurements preferences dialog box, you can easily choose between PostScript points and traditional American points.

In those Preferences dialogs, you’ll see another choice: agates. Agates are an old measuring system used for setting classified ads, which were set 14 lines to the vertical inch. This translates into a leading value of 5.143 PostScript points. The type designed to go with it, is also called agate and is the equivalent of 5.5-point.

While we’re on the subject of minutiae, so to speak, here’s a list of other traditional names for types of specific sizes:

Pearl, 5-point
Nonpareil, 6-point
Minion, 7-point
Brevier, 8-point
Bourgeois, 9-point
Long Primer, 10-point
Small Pica, 11-point
Pica, 12-point
English, 14-point
Great Primer, 18-point
Paragon, 20-point
Canon, 48-point

Splitting Hairs
Picas and points have survived the centuries and their industrial and technological revolutions because they’re the right tools for the job. For all of the technical changes in how type gets on paper, the type doesn’t look fundamentally different than it did 500 years ago. Type used for text is about the same size, as are the leading and the line measures we use. So it’s only logical that our measuring systems should stay more or less the same as well.

Nevertheless, for computerized typesetting systems, there’s no physical limit on how type can be sized and positioned. In the old days of metal type (and we’re only going back 50 years or less here), the physical nature of the print media made minute sizing and positioning adjustments impossible.

But just because our program can move or resize type in increments of a thousandth of an em or a point, is that a useful thing to do? Rarely. Over the last few of my columns, you’ve seen how minute tracking variations can make big changes in how type composes. That’s because the tiny tracking changes you apply are multiplied by the number of characters you apply them to. The cumulative effect can be significant, which is what makes tracking adjustment such a powerful tool.

But do you get the same benefits from minute changes in point size or leading? The answer is almost always no.

Over the centuries, divinely beautiful books and broadsheets have been made with systems that could only adjust leading in full-point increments, using typefaces that were available in at best a dozen point sizes.

I use half-point leading increments, but never any finer that. The only exception is when using what programs call baseline shift — a sort of leading adjustment — in which you’re obliged to position characters vertically using explicit point values. (More on this in the next installment of this column.) And rarely if ever do I find myself using fractional point sizes.

I don’t want to make a big deal about this, but usually looking for typographic solutions or refinements in minute changes of point size or leading is a waste of time. Differences in result are generally too minor to notice or matter.

Go to page 2 to read Jim’s argument for the metric system.

Posted on: February 17, 2010

James Felici

James Felici has worked in the publishing industry—in both editorial and production—for more than 30 years. A veteran journalist and former managing editor of Publish magazine, he has set type by hand as well as on systems from IBM, Linotype, Compugraphic, CCI, and Magna. His books include The Complete Manual of Typography (Peachpit Press), The Desktop Style Guide (Bantam/ITC), How to Get Great Type Out of Your Computer (North Light), and contributions to The Macintosh Bible (Peachpit Press). He has written for numerous publications, including PC World, Macworld, and The Seybold Report, and has been a featured speaker at Seybold Seminars, Macworld Expo, and other events worldwide.

1 Comment on The Measure of Type

  1. Here it is April 2011 and I’m seeing Mr. Felici’s February 2010 lecture for the first time. Thank you!

    His final paragraph may miss the point of “a big deal” about type but in the old days of type shops, this detail wasn’t so much about the look of he end result as it was to keep up knowledge and interest in what was a much more difficult profession with physical font racks and linotypes. Keeping the details that were only occasionally called to duty was necessary to keep that fine-tuned quality in mind for when it was needed.

    At 75, agate was about the only frequently used description when I first got into typesetting. The ’em’ and ‘en’ was mostly heard to define the appropriate ‘—’ or ‘–’ to differentiate dashes before typewriters forced ‘–‘ where ’em dashes’ weren’t available.

    Good to look back and appreciate the abilities we have today.

    An old man, a writer who likes people, living in the middle of the Pacific ocean near volcanoes, in tradewinds and soft bird songs.

Leave a comment

Your email address will not be published.