I Love Typography

Inconspicuous vertical metrics


There are generally taken to be five vertical measures of note in type design (from bottom to top): descender, baseline, midline*, caps-height, and ascender.

Vertical Measures in Minion

But if you delve into the minutiae of font design, you soon discover that there are a slew of important vertical metrics that aren’t much talked about. In this article, I will take a look at several of these metrics, and how they are used in font design.


Take a look at the basic alphabet from the venerable Minion, with the top three measures highlighted across each glyph:

Minion alphabet

You’ll note that the lowercase ‘t’ sticks out like a proverbial sore thumb. The crossbar aligns with the font’s x-height, but the top of the glyph is on a vertical plane all its own. In my research, I find no accepted terminology to measure the top of the ‘t’, and so I propose we give a name to this vertical metric (simply enough): t-height.

Minion t

The tradition for serif types is, like with Minion, for the crossbar of the ‘t’ to be at the font’s midline line, and for the top stem of the ‘t’ to come up somewhere midway between the midline and the caps-height.

t-heights compared

Typically, sans serif faces adhere to the same rule, as do slab serifs. However, gothic and geometric sans faces often break with tradition and elevate the ‘t’ to the ascender. Handwriting fonts often do the same.



Looking at the string ‘xoXO’, you might think that there are only three of the standard five vertical metrics in play: baseline, midline, and caps-height.


But if you view these glyphs at a larger size, you can see that the bottoms and the tops of the ‘o’ and ‘O’ overhang the baseline and their respective constraining heights. This is standard practice in type design; the reason being that if the ‘o’ and ‘O’ came exactly to the baseline and x-height/caps-height, it would appear to the naked eye as if they were smaller than other surrounding glyphs. This goes for any glyph rounded on the top or bottom. Here are some typical lowercase overshoots, top and then bottom:

lowercase overshoots

And typical uppercase overshoots:

uppercase overshoots

In principle, the overshoots should not be apparent to the eye at normal text sizes. The whole idea of overshoots is that they make their glyphs appear to be the same height as flat-topped and flat-bottomed glyphs.

e-bar height

Typically, the bar of the ‘e’ rests somewhere between halfway and two-thirds up from the baseline to the midline. (It’s extremely rare to find a typeface where the bar of the ‘e’ is below the halfway point between the baseline and the x-height.) Let’s call the height of the center of the bar of the ‘e’ the e-bar height. Here are Minion, Helvetica, Times New Roman, and Museo, and how their e-bar heights compare to the A-bar height and the x-junction height:

e-bar heights

A-bar height

The A-bar height, taken as the measure from the baseline to the center of the bar of the ‘A’, typically ranges from around 30-40% of the distance from the baseline to the caps-height.

A-bar heights

Incidentally, as you can see from the above image, the H-bar height is generally higher than the A-bar height. This makes sense architecturally, if you think of the crossbars of these glyphs as providing a sort of balance. The ‘A’ crossbar, if it were up at the height of the ‘H’ crossbar, would make the ‘A’ rather top-heavy.

R-Junction, P-Junction, B-Junction, and K-Junction heights

When I first started designing fonts, I thought, naively, that the junction points in the middle of the vertical bars of the ‘R’, ‘P’, ‘B’, and ‘K’ were all at the same height. Wouldn’t it be easy if you could just design a ‘P’, and then just stick a leg on it for an ‘R’, and another hoop on it for a ‘B’? I soon learned better. Here, once again, is Minion to show us the subtle variations in vertical metrics that make a font interesting:

K R B P Junction heights

Upon closer inspection, we can really see the different heights of these junctions. Here’s a close-up of the ‘R’ next to the ‘P’. Note that the P-junction height is slightly lower than the R-junction height.

R P Junction heights

The B-junction height is much higher than the R-junction height:

R P Junction heights

E-Junction and F-Junction heights

You might also suspect that the middle bars of the ‘E’ and ‘F’ would be at the same height, but this is not always the case.

E F Junction heights

In Minion, the E-junction height is slightly higher than the F-junction height. You can see, also, that there is a wide variance of heights across the middles of some glyphs:

A E F G H Junction heights

Look how strange Minion would look if all of these middle heights were the same:

A E F G H Junction heights equalized

And then some

We’ve really just scratched the surface. Well, I suppose we’ve dug in a little more deeply than the surface, but there are still a great many more interesting vertical metrics we could explore. G-bar height, briefly alluded to in the images above; foot serif height, and head serif height; two-storey a-junction height, or where the upper stroke of the bowl intersects with the vertical stem; two-storey g-bowl height, or how high the upper bowl of the ‘g’ starts above the baseline.

If your interest is piqued, I highly recommend Karen Cheng’s Designing Type, which is filled with wonderful analyses of minutiae like these.

* midline should not be confused with x-height; strictly speaking, the x-height is the distance between baseline and midline.


  1. sjaq

    Great post, thanks! Now I have some more typo terms I can throw at my typography teacher at school ;).

  2. Very good read! It’s a nice reminder of all the things I studied when I was taking Typography classes. This makes me wanna get into type design of my own, haha.

  3. a person can get dizzy looking at typefaces from all those typographic heights ;) Great article! Thank you, Alec. At least, while ‘falling’ I’ve learned a lot about the passing heights ;)

    John, what happened to those distinctly iLT curly braces around the comments? :) I see much more “Lucida Grande” has been injected into the style. I like it, the site appears much cleaner now. But I’m not crazy about the comment/date styling.

  4. Great read as always, I wonder how much differences in these values would affect how a font is percieved. perhaps by lining up the R, K, F etc functions perfectly, you’d create an entirely differently effect. I do wonder…

  5. If the x-height refers to the space between the base and mid-lines, does the Cap-height have a similar reference to space? If so what is the propper term for the line which the 1st illustration refers to as a Cap-height?

    Wow, fantastic article! I never considered all of these factors. Thank you so much for sharing!

  6. Intriguing, Alec. I appreciate these insights, and thanks for the book recommendation. I’ll definitely take a look.


    I clicked-through from my feed reader, and was greeted with a ‘StumbleUpon’ welcome message at the top of your post. Might want to look into that.

    Bye for now.

  7. Travis,
    x-height and cap-height are the universally used terms, no one actually says midline.

  8. Davide

    Very insightful Post. Thanks :-)

  9. also,

    It is nice to see these details highlighted in this way. Most people do not realize the subtle differences to create optically even and ‘normal’ looking designs.

    However, I have never really thought of any of these as vertical metrics – with the exception of overshoot. Overshoot should certainly be in the same category as the “general 5 measurements” you mentioned. Do not forget though, there are 4 or overshoot lines: two each for upper and lowercase letters and maybe another for the descenders depending on the design. But overshoot is generally not mentioned because it’s value is numerically so close to the other lines, it is difficult to show.

    These other ‘measurements’ are very important for designers to consider, but fundamentally they are a different category. Those general 5 measurements are all guidelines that affect many or most letters – they travel through the whole design. These others measurements you mention are only used on a per glyph basis. It does not make sense to create the titles t-bar height or E-junction height etc.* They are optical and proportional relationships and are not really in the same category as the general vertical metrics measurement lines.

    *these terms are useful for discussing and critiquing typefaces though.

  10. Great post! Thanks a lot for this lesson—I’m looking forward for more :-)

  11. Wow, and here I’ve actually been thinking that P, R, B had the same juncture hight as well as E and F. It looked like it made sense.

  12. Rob’s post +1. Well said. While all the nomenclature may not be necessary, the awareness of all these nuances is especially invaluable to fresh eyes. Cheers.

  13. amon

    what an inspiring piece - thankyou

  14. Rob Keller is right on. He speaks the truth!

    While I certainly applaud this article, I should add that the title is misleading. In type design, the term “Vertical Metrics” refers to something else.

    This article seems to me to be more about “Vertical Alignments.” In fonts, “Vertical Metrics” refer more to cetain values that will determine line spacing in applications.

  15. This is great! Thanks a lot.

    I’m sure that, like others, I’ll find an appropriate time to point some of this information out in the company of non-designers. And I’m also sure they’ll tut and give each other the “who invited the typographer” look!

    Keep the excellence coming!

  16. Many thanks for this article.
    I like details in typography (as well as in life).


  17. Thanks, all!

    Rob, et al: I agree that these measures (would it have been better if I had called the article “Inconspicuous Vertical Measures”?) could be in a different ontological category from the standard five vertical metrics. If they weren’t, I suppose they would already have names and would be discussed more often. However, that said, they are measures, and a synonym for ‘measure’ is ‘metric’, so I don’t feel too bad in dubbing them as such. And even if each of these metrics only applies to one glyph, it still affects a typeface’s design and can be fruitfully compared against the same metric in other faces. Rob, you say “It does not make sense to create the titles” of these metrics. But how do you propose that we then talk about it? “The height of the t”? Is that really any better than saying “t-height”?

    In any event, the piece was not meant to be the beginnings of a nomenclature revolution, but just an analysis of vertical measures that don’t generally get much attention.

    Thanks for the comments, all. Keep ‘em coming!

  18. Leandro

    I liked the post and also the comments. It’s nice to have a space of discussion like this one to learn some fundaments about typefaces.

  19. Great read. I teach a type 1 class and we’re just now starting to hit on these topics. Reinforces the idea that type design is a lot of balance and a lot of adjustment to make something look right.

  20. Thank you for this in-depth analysis of my favourite typeface. I can also recommend Karen Cheng’s Designing Type (German: Anatomie der Buchstaben).

  21. Troy H.

    Once we get done painting the exterior of the house, I know where I’m coming for address numbers (selection and placement) advice.

  22. Thank you for your hard work. Really. I’ve learned a lot from your site.

  23. Great article. I learned a lot about type metrics. Thanks!

  24. great post! now I have more things to look at when I get bored in english lessons! and more terms to confuse my mates with =P

  25. Yay! A learning article :) Alec, you are an awesome teacher. There is so much that goes into designing a typeface. It’s amazing. Now I’ll get even more eye-rolling from my husband and designer friends. Perfect ;)

  26. Craig

    Aren’t t and J [in cases where they curve at the baseline] also typical overshoots at the bottom?

  27. Jelmar

    When I learned to write, we always learned the t is supposed to be shorter than the ascenders. So this is a clear remnant of handwriting. Therefore any t’s that extend to the ascender height look silly to me.

    The crossbar of the F is usually placed slightly lower to optically compensate for the large open space below it.

    In FontLab, the only lines that fall under Vertical Metrics (i.e. appear/disappear when you show/hide them) are the Descender, Baseline, x-height, cap-height and Ascender lines.

    FontLab (and such) has even more lines! ;)
    They’re called ‘Alignment zones’, which are used for (auto)hinting.
    They can be set around all the Vertical Metrics I mentioned above.

  28. Pedro Alberto

    Awesome article.
    I’m a beginner type-student (15 years old), so this is really handy.

    Thank you for this awesome blog.

  29. @Craig: Yup. ‘t’ and ‘J’ often overshoot the baseline.

  30. Alec,

    I suppose I would have framed the article more like “inconspicuous vertical relationships” or something more of that nature.

    The nomenclature of many of these elements is not new, and as I mentioned earlier, they are useful in discussing type. But when describing a glyph it really doesn’t make much difference saying the “the height of the t” or “the t-height.”

    Dan pointed out rightfully that the metrics are primarily for vertical spacing of the fonts in applications, and they are also for things such as hinting. Overshoots would be in the “metrics” category as well, but these other measures are something else entirely.

    Again, good to point out the finer points of type design that don’t get enough attention. These are the details that can really make or break a design – slight adjustments of only a few units one way or another can make dramatic changes (in text typefaces at least).


  31. Sebastian Nagel

    small caps height!

  32. Flavio Hebaru

    very nice post!

  33. Great Article Alec. I never knew about e-bar and A-bar. Thanks for the information, and keep it comin’.

  34. Mr. Smith

    Although I already knew a lot of this, there’s things you mentioned I’ve never really thought about. This was a good read, and I’m going to have to check out that book. Thanks!

  35. So I’ve always been curious if there is any sort of “rule” for overshoots, like how mush larger to make them. I know a steady rule is impossible, but a jumping off point would be nice.

  36. Great post Alec!

    Joshua: There isn’t one rule. The overshoots of a display font need to be smaller as the ones for a text font. The overshoots also depend on the shape of the bowls. The flatter they are, the less overshoot is needed. Only trust your eyes ;-)

  37. @Sebastian: good one! Small caps generally are right around a font’s x-height. See this article for more than you ever wanted to know about Minion’s small caps.

  38. … do not forget the figure set hights, old style figures, lining figures, tab figures, small caps figures, and small caps hights from text to display faces … however i am still working on an old typeface that also has different hights in the characters «t» and «f» and «o» or «e» … but it is good to see that people are intrested in «hight definitions».

  39. reggyR

    A lot to take in but a great article. Thanks Alec and Johno!

  40. jac

    wow! Great post! I do notice these little differences, but for you to blog about it and compare these fonts is true passion :)

  41. syahrulfikri

    amazing. this is really great post!!!

  42. @Rob Keller I am aware that they are the common terms, but I was intrigued by the concept of the term x-height measuring space and i was wondering how far that thought could be followed.

    I remember when I learned that the aperture of a camera was not the little bit that opened up to let the light in, but actually the space that was created by that little bit when it opened. Similar revelation here. It struck me.

    So, does any one know; “strictly speaking,” if the x-height refers to the space between the base and mid-lines, does the Cap-height have a similar reference to space? If so what is the propper term for the line which the 1st illustration refers to as a Cap-height?

  43. I am interested in knowing where the term ‘midline’ originated. I am suspicious of it’s invention and more importantly its use here…

    I believe you are correct Travis about it’s technical nature, although again I will reiterate that I don’t consider it to be the correct term for ‘everyday use.’ X-height is the standard definition for measuring the space between the baseline and the (relative) height of the lowercase characters. In my library, I only have two references to midline in my collections. The first is in Typo Magazine from July 2005 from an article by Krisztina Somogyi (and I believe it to be the improper use in this case). The second is from the Panose system specs. Their definition is “The Midline digit describes the placement of the midline and the treatment of diagonal stem apexes.” Honestly, I don’t fully understand that definition or how it is calculated. They also have a separate definition for x-height which defines the relative height of the lowercase to the uppercase.

    You need to bear in mind that most of these typographic terms have old roots and were not so technical. Strictly speaking, a comparable term for cap-height, that is more technical, does not exist.

    Maybe someone can correct me on the correct use of midline. But as I see it, there are not many reasons to ever use it in general discussions – the correct term is always x-height.

  44. Maybe Johno can weigh in here about the midline controversy. In the first draft of this article, I used “x-height” everywhere, but Johno suggested that there were places where “midline” was more appropriate.

  45. Sama

    great article! Aside from overshoot, all of it was new to me. keep em’ coming.

  46. Thank you so much for this article. It made a dull day brighter :)
    Always feels nice to know that someone else has gone into the same minute details that one thought abut the whole day !

  47. Rob Keller
    Alec’s original draft used x-height throughout the article. In most places I replaced it with mid-line; the footnote is also mine.

    Bringhurst in Elements defines the x-height thus:

    The distance between the baseline and the midline of an alphabet,…

    I don’t have a problem with x-height being used as a synonym for midline, but thought it was important to point out the distinction in this article. I still prefer midline. Thanks for your input. You should consider writing something for iLT.

    What you write about x-height and cap height highlights problems with the nomenclature of type design. But as Rob says, there’s really no alternative to cap height.

  48. For what it’s worth, Karen Cheng does make some mention (and illustration) of the variations in junctions between P-R-B, t-height, and some other often overlooked metrics in Designing Type. I’m rereading it right now and finding it a much better resource than I first thought.

  49. Taylor

    Bend a few of those serifs in a bit more on the last image showing what it would be like if they were all lined up and you would almost have a Helvetica Serif haha.

  50. Taylor
    Well, there is HelSerif (bad idea); and for a truly dreadful Helvetica Serif see this (please sit down first).

  51. nils

    I suppose being from Sweden with all the åäö I think the vertical alignment of the dots (whatever they are called) deserves a mention. Not only to add a measure but because a lot of type designers seem to haphazardly throw them in there in the end when the typeface is already done. I know it’s not a very big market but for me a typeface with ugly åäö characters (and ÅÄÖ of couse) simply is unusable. so that’s atleast two vertical measures that needs some more focus, and yes, maybe some sort of word for it.
    Probably the same for a lot of other languaages with less common letters?

  52. nils
    Yes, the placement (& design) of diacritics (accents) is very important. Good timing, because next week we have an article devoted to the topic.

  53. nils

    Sounds great!
    I’ll be looking forward to that. And I agree, the design/shape is absolutely as important.

  54. julia h

    thank you so much, that you share your knowledge with us :)

  55. Johno,

    X-height terminology is apparently more varied and divided than many, including myself, might have guessed. In my experience, the expression “x-height” has served well nearly single-handedly. It is by far the most frequently used of all these terms; therefor I would not say that x-height is a synonym for midline, rather it is the other way around. Maybe someone has references to these terms being coined?

    However, the use of midline does occasionally pop up – I did not notice it in Bringhurst the other day… I was looking for a definition of midline, of which there was none. You will notice that Bringhurst also uses the term torso (sparingly) for the same thing. Furthermore, other posts from this site use the additional terms “mean line” and “waist line” (w/ high and low waist lines). Torso and waist line feel a bit conflicting, but that is something else. The up side this plethora of terms is that since the the height of the x is such a common typographic feature to discuss, writers can have a wide palette of idioms to describe it.

    As for the footnote: midline should not be confused with x-height; strictly speaking, the x-height is the distance between baseline and midline., it is still rather ambiguous to me. I appreciate fine distinctions, but this one has been lost on me.

    I am curious about your affinity for the use of midline. Is that coming from somewhere specific, or simply personal preference?


    PS– we can certainly talk about future articles sometime

  56. From a purely etymological standpoint, “midline” sounds more like an even division between two measures. A horizontal measure that cuts equidistant between the baseline and cap height (or even the ascender and descender) would not have a consistent use or significance across every design. Correct me if I’m overlooking something from a typesetting/line-height perspective.

    “X-height”, on the other hand, is a fairly universal term. It’s the height of the lowercase because the lowercase ‘x’ is the most consistent character in defining this measure across a majority of styles (roman or italic or handwritten). You could introduce “lower-height” as it’s replacement, but why call it anything different at this point in history? And why try to assert something as especially ambiguous as “midline”?

    Let sleeping dogs lie, especially if they’re not biting the hand that feeds them. (Apologies for conjoining metaphors with such zeal. They just fit so well.)

  57. beautiful post, thanks for sharing this with us, i will be linking it from my website :)

  58. This was very informative, thanks very much for taking the time to write this. All these different heights, it just goes to show what a true art form typography is.

  59. Wow! Very well written! I couldn’t stop reading it!


  60. Discussions involving Fonts always attract folks with a great insight into their importance and placement in everyday life.

    Not enough importance is placed on Typography in education, that way more of us could be involved at a higher level.

  61. Attention to detail can make or break a font! Thanks for showing the depth and thought that goes into font design.

  62. It’s nice to see such an indepth article on the creation of type.

  63. Dann

    Great post. I’m really going to read more about this.

    There is something I can’t find too much to read about: accents. There was a comment from nils on Swedish accents. As I’m Hungarian I’d like to learn more about ó ú ö ő ü ű á é í, I’m also interested in the Italian or French ones where these “long” accents are reflected. Nils’s very good point was their vertical alignment but their angle, shape and horizontal alignment are very interesting too!

    I hope it will give you some headache :D

  64. @Rob Keller & @Johno
    You know, in any other subject I would be put off by the seeming inconsistancies in the naming conventions and ways they are used. But in this case, the case of typography in general, there is such a romanitc quirkyness to it, it becomes not an annoyance but another reason for endearment.
    It’s really the make-it-up-as-we-go-along attitude to the science of type that makes it so personal and at the same time so specialized.

  65. Thanks Alec for sharing this. I will be linking to this post from my site. This is awesome.

    - Harnish

  66. Gumboot

    The dot heights on ‘i’ and ‘j’. Will they be addressed in the diacritics article?

  67. Filip Blažek’s Diacritics Project at http://diacritics.typo.cz is a great resource for some of the subtler points about the use and abuse of diacritics for type designers and typographers.

  68. nutty t

    This is a really cool web site. I’m so glad i fount it.

  69. I simply love this kind of stuff. Thanks for writing and making those clear explanations. I already have Karen Cheng’s Designing Type. Love it and recommend it.

  70. Thanks for this really awesome post. I love how you went into detail about a very particular aspect of type and explained it so thoroughly. I hope you don’t mind that I will be linking it from my blog! :)

  71. What an intriguing read. I appreciate the time that went into detailing the minutiae and unique character traits of each font.

  72. Throw another “great post” on the pile. This is a must read for anyone in love with the craft of typography. Thank you.

  73. Uau!!!!!!!!!!!!!!!!!!!!!!

    this site is everything that i need!!!!!


  74. Kim Kozak

    I’m an HR Consultant, not a graphic designer, and I’ve got a question. I found this article looking for a font to use in my logo. (Currently unemployed, going to try consulting) This is overwhelmingly complicated for a non-designer but I thought you may know of one.

    I just want my logo to be my name, modernly set without a space between. like KimKozak……with Kim being a color and Kozak being black. However, I’d like the Z to go below the line, like a cursive Z does, but don’t want to use cursive font. That way, I can snuggle a small “.com” after the Z.

    Origionally I wanted to use Minion (the best I could find in Word 2007) because it is modern, yet looks a little like Times, and I do a lot of law-related work. But I think finding a cursive serf Z will either be impossible or ugly.

    So, can you or anyone else reading this recommend a font for me? I can’t afford a designer with now and I would appreciate it SOOOOOO much!!

    Kim.kozak [at] gmail.com

  75. Thank you for this in-depth analysis of my favourite typeface. I can also recommend Karen Cheng’s Designing Type (German: Anatomie der Buchstaben). :)

  76. This was a great article. I am a designer by profession and work with large font list daily. I have always wondered why these verticals varied so much and this article was great in realizing the visual impact of the verticals vs. making every character conform to strict boundaries. Its also interesting how the subtle changes have been made over the years… very enlightening and a good read for anyone that works with type.

  77. I’d like to remark that just as there may overshoot lines in addition to caps-heights and x-heights, it should be noted that there may be different t-heights as well: a geometric t-height that touches the topmost point of the letter; and an optical t-height that goes slightly lower, especially where the tip of the t is rounded or pointed.

  1. | Clay Carson—Jan 29, 2009

previous post: To a typetastic new year

next post: Greener type

May Fonts April Fonts March Fonts February Fonts January Fonts December Fonts November Fonts October Fonts September Fonts August Fonts July Fonts June Fonts May Fonts April Fonts March Fonts February Fonts January Fonts January Fonts November Fonts October Fonts September Fonts August Fonts July Fonts June Fonts May Fonts April Fonts March Fonts February Fonts January Fonts december Fonts October Fonts September Fonts August Fonts July Fonts June Fonts May Fonts April Fonts March Fonts February Fonts January Fonts December Fonts November Fonts October Fonts September Fonts August Fonts July Fonts May Fonts April Fonts March 2011 Fonts February 2011 Fonts January 2011 Fonts December 2010 Fonts November 2010 Fonts October 2010 Fonts September 2010 Fonts August Fonts July Fonts June Fonts May Fonts April Fonts March Fonts February 2010 featured fonts December Fonts November Fonts October Fonts September Fonts August Fonts July Fonts June Fonts May Fonts April Fonts March Fonts February Fonts January Fonts December Fonts November Fonts October Fonts September Fonts August Fonts July Fonts June Fonts May Fonts April Fonts March Fonts February Fonts January Fonts December Fonts