January 7, 2002: Someone wrote me asking for evidence to support my assertion that serif vs sans was of limited use in predicting legibility on screen. I can't say I found a lot, but here's some: HFI reviews a CHI study comparing Times and Verdana (negligible difference) Swedish study (PDF in Swedish) comparing Verdana and Georgia (slight advantage to Georgia, maybe) CHI '95 study comparing Arial, MS Sans and MS Serif (size matters, Arial sucks, otherwise negligible) Usability News compares Arial and Times New Roman (scroll down to find that other factors matter more than serif vs sans) and again, with more fonts (century schoolbook, times, georgia, arial) US Gov says it doesn't matter, but don't skimp on size (thanks jcdriscoll!)