Both serif typefaces and sans serif typefaces have a long history, going back to inscriptions in Ancient Rome and probably earlier (see Sects. 1.2 and 1.3). Historically, serif typefaces were used in printing earlier than sans serif typefaces, and they have usually been preferred to sans serif typefaces for use in formal documents. Section 1.2 described a number of theories aimed at explaining why serifs should have survived in modern typography and indeed at explaining why serif typefaces should be more legible than sans serif typefaces when reading from paper: (a) that serifs constitute additional visual cues to support the reader’s gaze; (b) that they overcome the harmful effects of irradiation; and (c) that they facilitate the operation of line detectors in the human visual system. However, none of these explanations has proved especially convincing in the light of subsequent arguments and evidence. In the case of reading from screens, it was argued that serifs and other details might be lost when reading from low-resolution monitors, but this is not a plausible argument now that high-resolution monitors are widely available.

A different approach is to claim that serif and sans serif typefaces do not differ in their legibility because of inherent properties of serifs themselves, but that the presence or absence of serifs serves as a proxy for some other property of typefaces. This approach was adopted by Wilkins et al. (2020; see Sect. 11.2), who suggested that serifs tended to accentuate the spatial periodicity of letter strokes (i.e., the vertical “stripiness” of words), as measured by the first peak in a word’s horizontal autocorrelation. Wilkins et al. showed that words of low vertical stripiness are read more quickly than words of high vertical stripiness, and that serif typefaces tend to have higher stripiness than sans serif typefaces, but they did not show that this leads to variations in how quickly words in different typefaces are read. For the moment, the theoretical and practical implications of vertical stripiness must remain unclear and contentious. More important, the arguments put forward by Wilkins et al. entail that serif typefaces should be less legible than sans serif typefaces both when reading from paper and when reading from screens, which is not position that has been adopted to date by any other researchers.

In contrast, there is a good deal of evidence about the negative effects of horizontal stripiness—the extent to which successive lines of text tend to resemble horizontal stripes. As mentioned in Sect. 4.1, horizontal stripes are known to induce eye strain, visual illusions, headache, and even seizures. Wilkins et al. (1984) used the Michelson contrast to measure the horizontal “stripiness” of visual displays: this is defined as the difference in the luminances of the light and dark sections of a pattern divided by the sum of their luminances. In the case of horizontal stripes, Wilkins et al. found a clear relationship between this measure and the probability of illusions and seizures. Wilkins and Nimmo-Smith (1987) applied the measure to lines of printed text and found that it was inversely related to readers’ ratings of the clarity and comfort of the material. More recently, Wilkins et al. (2020) applied Fourier analysis to text displayed on an LCD screen and obtained similar results. In both studies, the clarity and comfort of the text appeared to be improved by increasing the spacing between the lines of text relative to the x-height of the typeface. However, Wilkins and Nimmo-Smith (1987) had compared the mean Michelson contrast of samples of 24 serif typefaces and seven sans serif typefaces: they found a highly significant variation across the 31 typefaces, but they did not find any systematic difference between the serif typefaces and the sans serif typefaces. While clarity and comfort may be very important characteristics of text presented on paper or on screens, there seems to be no difference between serif and sans serif typefaces in these characteristics.

In fact, the copious evidence that has been reviewed in this monograph leads to the conclusion that there is no difference in the legibility of serif and sans serif typefaces either when reading from paper or when reading from screens. This contradicts various assertions made over the last 100 years by typographers, designers, and other authority figures. What this means is that assertions to the effect that “everyone knows” that such-and-such should not be accepted on the basis of the authority or status of the people or organisations making them but should be regarded merely as conjectures that might well be subject to refutation through carefully designed empirical research (cf. Popper, 1959, 1962.) Of the large number of studies that I have reviewed in this monograph, some have been more carefully designed than others, and I have identified at least some of the design flaws that are apparent in previous research. Nevertheless, the overwhelming thrust of the available evidence is that there is no difference in the legibility of serif typefaces and sans serif typefaces either when reading from paper or when reading from screens. Typographers and software designers should feel able to make full use of both serif typefaces and sans serif typefaces, even if legibility is a key criterion in their choice.