Our modern world submits with an ill grace to the nuisance of
spectacles, but flatters itself that after all they afford a measure of
civilization. Thirty-five years ago Dr. Émile Javal, a Parisian oculist,
contested this self-complacent inference, believing the terrible
increase of near sight among school children to be due rather to a
defect than to an excess of civilization. He conceived that the trouble
must lie in the material set for the eye to work upon, namely, the
printed page. He therefore instituted a series of experiments to
discover its defects from the point of view of hygiene. Being an
oculist, he naturally adopted the test of distance to determine the
legibility of single letters at the limit of vision, and he employed the
oculist's special type. His conclusions cover a wide range. He decided
that paper with a slightly buff tint printed with an ink tinged with
blue was the most agreeable combination for the eye, though in absolute
clearness nothing can surpass the contrast of black upon white. He held
that leading is no advantage to clearness, and that it would be better
to print the same words on the page in a larger type unleaded. He found
the current type too condensed; this is particularly a fault of French
type. But he favored spacing between the letters of a word, a conclusion
in which he has not been followed by later investigators. He found
shaded type a disadvantage and advocated a fairly black type in which
all the lines are of uniform thickness. But most interesting are his
conclusions regarding the letters themselves. He found that the eye in
reading follows a horizontal line which cuts the words just below the
tops of the short letters, the parts of the letters being indistinct in
proportion as they are distant from this line. It is chiefly by their
individuality on this line that letters acquire distinctness. But just
here he found that an unfortunate tendency towards uniformity had been
at work, flattening the rounded letters and rounding the square letters.
In a series of articles he gives exhaustive studies of the various
letters, their characteristics, and their possible reform.
[Illustration: These ten-point lines in Della Robbia of the American
Type Founders Company include the principal elements of reform advocated
by Dr. Javal, as well as others mentioned below]
A few years later Dr. Cattell, now a professor in Columbia, but then an
investigator in Wundt's psychological laboratory in Leipsic, made a
series of studies on brain and eye inertia in the recognition of
letters. Like Dr. Javal he found some alphabets harder to see than
others and the letters of the same alphabet different in legibility. He
saw no advantage in having a mixture of capital and small letters. He
condemned shading in types and opposed all ornament as an element of
confusion. He regarded punctuation marks as hard to see and proposed
that they should be displaced, or at least supplemented, by spaces
between the words corresponding to the pause in the thought or the
utterance.
He tested the letters by their legibility when seen for a small fraction
of a second through a narrow slit in a falling screen. Beginning with
the capitals, he found that out of two hundred and seventy trials for
each letter, W was recognized two hundred and forty-one times and E only
sixty-three times, the former being much more distinct and the latter
much less distinct than any other. Some letters, like S and C, were
found hard to recognize in themselves, and certain groups of letters,
such as O, Q, G, and C, were constantly confused with one another. Said
Dr. Cattell, "If I should give the probable time wasted each day through
a single letter, as E, being needlessly illegible, it would seem almost
incredible; and, if we could calculate the necessary strain put upon eye
and brain, it would be still more appalling."
In regard to the small letters he found a like difference in legibility.
Out of one hundred trials d was read correctly eighty-seven times, s
only twenty-eight times. He found s, g, c, and x particularly hard to
recognize by reason of their form; and certain pairs and groups were
sources of confusion. The group of slim letters, i, j, l, f, t, is an
instance. He suggested that a new form of l, perhaps the Greek [Greek:
l], should be adopted; and he advocated the dropping of the dot from the
i, as in Greek. He made experiments upon the German as well as the Roman
alphabet, but he found the former so bad that he could only advise
giving it up altogether.
Somewhat later, in 1888, Mr. E. C. Sanford, now president of Clark
College, published in the "American Journal of Psychology" an exhaustive
study on "The Relative Legibility of the Small Letters." He studied
simply the letter forms, to determine the order of legibility in the
alphabet and the groups most liable to confusion, in order to discover
what letters most need improvement and upon what clearness depends. He
too employed a special type. He found the order under the distance test
to be w m q p v y j f h r d g k b x l n u a t i z o c s e, and the order
under the time test m w d q v y j p k f b l i g h r x t o u a n e s c z.
It will be noticed that of the seven letters most largely represented in
a full font of type, e t a i n o s, all fall in the last third of one or
the other of these two groups, four are there in both groups, while e,
the letter used most of all, stands at the very foot of the list in the
distance group. Could there be any clearer call for the reform of our
letters?
Mr. Sanford enters at length into the question of the points that help
and hinder legibility and that should therefore be considered in
reforming the shapes of letters. Enlargement of size and increase of
differences are obvious aids to clearness. Simplicity of outline and
concentration of peculiarity upon one feature are important elements of
legibility. Even a letter of small size, like v, is brought into the
first group by a combination of these two qualities. Serifs are
necessary to prevent irradiation, or an overflowing of the white on the
black, but they should be stubby; if long, they take on the character of
ornament and become confusing. The letters g and a are complicated
without being distinctive and are therefore continually confused with
other letters. The c e o group of much used letters can be made less
liable to confusion if the gap on the right of the first two letters is
made wider and the line of the e slants downward as in Jenson. Another
group, a n u, are confused together. To avoid this the top and bottom
openings of n and u should be made as open as possible and the a should
go back to the old script form =a= as in the Humanistic type. The letter s
is a source of great difficulty, being either not recognized at all in
the tests or confused with other letters. It will be remembered that
Franklin greatly deprecated the giving up of the long f, and a return to
this form is now suggested, care being taken, of course, to
differentiate it from f, especially by carrying it below the line. The
dot of the i is of no use when the letter stands alone, but it is an
important element of distinctness in words like "minim." The dot, as Dr.
Javal suggests, should be set on a level with the top of the l rather
than on a level with the top of the t. A reduction of serifs would
lessen the confusion of x and z and of s and z.
But it is unnecessary to trace these studies in all their minutiae. In
the twenty-eight years that have followed the appearance of Mr.
Sanford's article work along the same lines has been done by many
investigators in various countries. Some of the conclusions that we have
noticed have been sustained, others have been discredited. The most
important conclusions of the investigators down to 1908 will be found
scattered through the pages of Huey's "Psychology and Pedagogy of
Reading," which appeared in that year. Such matters as the normal length
of a line of print, the size of type appropriate to schoolbooks for
children of different ages, the possibilities of future type design with
reference solely to the reader's needs, are among the many subjects
there set forth in an interesting fashion.
In all these studies one obvious subject of investigation appears to
have been overlooked, and that is the actual types of everyday print. Do
they vary greatly in legibility? Are some of them so bad that they ought
to be rejected _in toto_? On the other hand, have the designers of
certain types attained by instinct or by happy accident a degree of
legibility that approximates the best to be hoped for? If so, can we
trace the direction to be followed in seeking further improvement? To
answer these questions an extended investigation was undertaken at Clark
University in 1911 by Miss Barbara Elizabeth Roethlein under the
direction of Professor John Wallace Baird. Her results were published by
Clark University Library in January, 1912, under the title "The Relative
Legibility of Different Faces of Printing Types." The pamphlet abounds
in tables made clear by the use of the very types under consideration.
The following are the conclusions reached:
1. Certain faces of type are much more legible than other
faces; and certain letters of every face are much more
legible than other letters of the same face.
2. These differences in legibility prove to be greater when
letters are presented in isolation from one another than
when they are presented in groups.
3. Legibility is a product of six factors: (1) the form of
the letter; (2) the size of the letter; (3) the heaviness of
the face of the letter (the thickness of the lines which
constitute the letter); (4) the width of the white margin
which surrounds the letter; (5) the position of the letter
in the letter group; (6) the shape and size of the adjacent
letters. In our experiments the first factor seemed to be
less significant than any of the other five; that is, in the
type-faces which were employed in the present investigation
the form of any given letter of the alphabet usually varied
between such narrow limits as to constitute a relatively
insignificant factor in the determination of its legibility.
4. The relatively heavy-faced types prove to be more legible
than the light-faced types. The optimal heaviness of face
seems to lie in a mean between the bold faces and such light
faces as Scotch Roman and Cushing Monotone.
5. The initial position in a group of letters is the most
advantageous position for legibility; the final position
comes next in order of advantage; and the intermediate or
internal positions are least favorable for legibility.
6. The size and the form of the letters which stand adjacent
to any given letter play an important role in determining
its legibility; and the misreadings which occur in the case
of grouped letters are of a wholly different sort from those
which occur in the case of isolated letters. When letters of
the same height or of similar form appear side by side, they
become relatively illegible. But the juxtaposition of an
ascender, a descender and a short letter tends to improve
the legibility of each, as also does the juxtaposition of
letters which are made up wholly or chiefly of straight
lines and letters which are made up wholly or chiefly of
curved lines.
7. The quality and the texture of the paper is a much less
significant factor than has been supposed, provided, of
course, that the illumination and the inclination of the
paper are such as to secure an optimal condition of light
reflection from its surface.
8. There is an urgent need for modification of certain
letters of the alphabet.
Contrary to previous results with special types, these tests of
commercial types represent the capitals as more legible, by about
one-fifth, than the lowercase letters; but, in view of the much greater
bigness and heaviness of capitals, the earlier judgment would seem to be
supported so far as the letter forms of the two classes are concerned.
The order of each class, taking an average of all the faces, is as
follows: W M L J I A T C V Q P D O Y U F H X G N Z K E R B S m w d j l p
f q y i h g b k v r t n c u o x a e z s. Considering only the lowercase
letters, which represent nine-tenths of the print that meets the eye, we
still have four of the most used letters, s e a o, in the lowest fourth
of the group, while s in both sizes of type and in all faces stands at
the bottom. The average legibility of the best and worst is: W, 300.2;
S, 205.7; m, 296.8; s, 152.6.
The tests were by distance; the letters were all ten-point of the
various faces; and the figures represent the distance in centimeters at
which the letters were recognized. There is a satisfaction in being
assured that the range between the best and the worst is not so great as
had been estimated previously, the proportion being in the one case not
quite 3:2 and in the other not quite 3:1.5. The following twenty-six
widely different faces of type were studied:
American Typewriter
Bold Antique
Bulfinch
Caslon Oldstyle No. 540
Century Oldstyle
Century Oldstyle, Bold
Century Expanded
Cheltenham Oldstyle
Cheltenham Bold
Cheltenham Bold, Condensed
Cheltenham Italic
Cheltenham Wide
Clearface
Clearface Italic
Clearface Bold
Clearface Bold Italic
Cushing No. 2
Cushing Oldstyle No. 2
Cushing Monotone
Della Robbia
DeVinne No. 2
DeVinne No. 2, Italic
Franklin Gothic
Jenson Oldstyle No. 2
News Gothic
Ronaldson Oldstyle No. 551
Of these, omitting the boldface and italic types, as well as all
capitals, the six best text types, ranging in average distance of
recognition from 236.4 to 224.3, are News Gothic, Bulfinch, Clearface,
Century Oldstyle, Century Expanded, and Cheltenham Wide. The six worst,
ranging from 206.4 to 185.6, are Cheltenham Oldstyle, DeVinne No. 2,
American Typewriter, Caslon Oldstyle, Cushing Monotone, and Cushing No.
2. The author says, commenting on these findings:
If legibility is to be our sole criterion of excellence of
typeface, News Gothic must be regarded as our nearest
approximation to an ideal face, in so far as the present
investigation is able to decide this question. The esthetic
factor must always be taken into account, however, here as
elsewhere. And the reader who prefers the appearance of
Cushing Oldstyle or a Century face may gratify his esthetic
demands without any considerable sacrifice of legibility.
To what extent these conclusions may be modified by future experiments
it is, of course, impossible to predict, but they clearly point the way
towards definiteness and boldness in the design of types as well as to a
preference for the larger sizes in their use. All this, as we shall see
in the next chapter, is in harmony with what experience has been
gradually confirming in the practice of the last generation.