Skip to content

Instantly share code, notes, and snippets.

@unruthless
Created May 26, 2010 01:31
Show Gist options
  • Save unruthless/413930 to your computer and use it in GitHub Desktop.
Save unruthless/413930 to your computer and use it in GitHub Desktop.
CSS for <sub> and <sup>
sub, sup {
/* Specified in % so that the sup/sup is the
right size relative to the surrounding text */
font-size: 75%;
/* Zero out the line-height so that it doesn't
interfere with the positioning that follows */
line-height: 0;
/* Where the magic happens: makes all browsers position
the sup/sup properly, relative to the surrounding text */
position: relative;
/* Note that if you're using Eric Meyer's reset.css, this
is already set and you can remove this rule */
vertical-align: baseline;
}
sup {
/* Move the superscripted text up */
top: -0.5em;
}
sub {
/* Move the subscripted text down, but only
half as far down as the superscript moved up */
bottom: -0.25em;
}
@unruthless
Copy link
Author

This is tested to not break line-heights in WinXP/IE6, WinXP/IE7, WinXP/IE8, Mac/FF 3.5.9, Mac/Chrome 5.0, Mac/Safari 4.0.4, assuming a base font size of 14px and a line-height of 21px, or 1.5em. Poke this, try to make it break!

@paulirish
Copy link

Great work, this. :D

@ryanflorence
Copy link

Glad this made it in.

@tchalvak
Copy link

Nice.

@michaeladamek
Copy link

Having a problem in webkit using a ® in an anchor link... Line-height breaks. Suggestions?

@unruthless
Copy link
Author

michaeladamek: I'm unable to replicate this issue -- http://jsfiddle.net/ksZrh/ works fine in Safari 5.0.5. Can you link to a test case that shows the breakage you're talking about?

@tchalvak
Copy link

Not able to duplicate in google chrome 14 or it's so subtle that I can't tell, even when the original font size is at 500%.

@michaeladamek
Copy link

michaeladamek commented Sep 22, 2011 via email

@tchalvak
Copy link

Err, so you're saying the underline breaks, not the css line-height, right? That I can confirm.

@michaeladamek
Copy link

michaeladamek commented Sep 22, 2011 via email

@unruthless
Copy link
Author

Thanks for reporting the potential issue! I don't think the link made it in to your comment. Can you re-post just the link?

[EDIT] Ah, link's there now. Thanks, will take a look at it shortly.

@michaeladamek
Copy link

michaeladamek commented Sep 22, 2011 via email

@michaeladamek
Copy link

Sorry. I don't have a live link to share because the superscripted (R) didn't make it into production because of the bug.

@tchalvak
Copy link

tchalvak commented Oct 4, 2011

Seems like the best solution here would be to simply lose the underline style on link sup/sub blocks, though when I tried that out earlier I couldn't get that destyling to apply. Anyone else able to make the fiddle display that way?

@michaeladamek
Copy link

michaeladamek commented Oct 4, 2011 via email

@unruthless
Copy link
Author

Folks, I threw together a test case with all of the HTML entities from http://en.wikipedia.org/wiki/List_of_XML_and_HTML_character_entity_references, not including those that are just spaces or otherwise non-rendering characters.

As michaeladamek points out, the glyphs for the (R) and (C) symbols aren't superscripted by default like the (TM), degree, spanish ordinal operators (a) and (o), and some other glyphs are.

Turns out that if you apply an underline to any glyph that isn't superscripted by default, it will render inconsistently between Webkit and non-Webkit browsers, because Webkit browsers superscript the underline and the glyph, while non-Webkit browsers superscript only the glyph. Whether the underline is applied via a border-bottom or text-decoration:underline declaration doesn't seem to make a difference, unfortunately.

Here's the full test case if you'd like to fork it: http://jsfiddle.net/ksZrh/9/

I'm going to keep thinking about this, and welcome suggestions for normalizing the underline placement cross-browser.

@unruthless
Copy link
Author

Interesting: a potential approach to getting underlines by using gradients rather than border-bottom or text-decoration. Lost the link to the tweet, but credit goes to @chikuyonok -- http://jsfiddle.net/yyHNp/5/

@debjitbis08
Copy link

Having a strange issue with IE7. I have a large content (Terms etc.) inside a fixed height div with scroll: auto. There are a lot of sup tags inside. These show up below the div and for the full length of the content.

@rposky
Copy link

rposky commented Dec 19, 2013

This is the only consistent solution I have found to the superscript underline problem outlined above. Thanks for sharing @unruthless.

@GigaGerard
Copy link

Disclaimer: I don't know what this GitHub is and they are not allowed to bill me for anything, I just want to comment on this excellent post.

I use sub and sup a lot and specially with the old Explorers and now with the new Firefox 27 you run into troubles that only the here proposed CSS can save you from. Well done!

In special cases a subscript or superscript can be given
margin-left: -1px;
margin-right: 1px;
depending on the context.
Also a paragraph can be given some extra line-height when sub and sup appear in it.

@tellmethetruth
Copy link

Nice! It's work for me.. 👍

@bheitkem
Copy link

Thank you! Great work!

@imrehg
Copy link

imrehg commented Jan 30, 2015

@unruthless Trying the full test linked above, with Firefox 35.0.1 (current), now it behaves like Chrome as well, breaking the text decoration with superscipt.

@vfonsecad
Copy link

Amazing, thank you!!

@kaidjohnson
Copy link

@unruthless @michaeladamek -- Not sure of the complete x-browser implications (will test IE/Safari tomorrow), but one way to remove the text-decoration from <sub> and <sup> is to display: inline-block. http://jsfiddle.net/kaidjohnson/89jdwp0f/ Seems cleaner than adding more selectors, but full browser testing is still needed on this example.

@o-o-ZEPHYROS-o-o
Copy link

o-o-ZEPHYROS-o-o commented Nov 22, 2016

This is great! Since I was already using the SUP tag, I just added your code to my CSS and voila. Well done!

@sjorsvanheuveln
Copy link

Thanks this implementation is fantastic. This helped me to solve some alignment problems that were caused by superscript for mathematical notations.

@Tzikas
Copy link

Tzikas commented May 12, 2017

+1

@Disney-Banje
Copy link

Thanks, man this helped

@christoph-werker
Copy link

Thanks man. Good work 👍

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment