r
Z
2S
SCHOOL
— iPtTBLICATIOJNTS
OJ^THE
Clark University Library
WORCESTER, MASS.
Editkd by JLOtTIS N. WILSON, Litt. D., LroBAHiAN
Vol. 3.
January, 1912.
No. 1.
The Relative Legibility of Different
Faces of Printing Types
By
BARBARA ELIZABETH ROETHLEIN, A. M.
Clark University
WORCESTER. MASS.
f _ • •
THE RELATIVE LEGIBILITY OF DIFFERENT
FACES OF PRINTING TYPES
By Barbara Elizabeth Roethlein, A.M., Clark University
Communicated by John Wallace Baird
I. Introductory I
II. Historical 2
III. Experimental 5
a. Materials and Apparatus 5
b. Method of Procedure ...... 7
c. Observers ........ 8
d. Results 9
IV. Discussion of Results ....... 10
V. Conclusions 33
VI. Bibliography 34
I. INTRODUCTORY
The aim of the investigation which is here reported was
to determine the relative ease or difficulty with which various
'faces'^ or forms of printed letters can be read; and to
discover what relationship obtains between legibility and cer-
tain definite modifications of 'face.' A number of typically
different faces of type were selected for investigation; and
our experimental procedure consisted essentially in determin-
ing to what extent the legibility of each face was affected
— both when the letters were presented in isolation and in
groups — by the introduction of unfavorable conditions for
reading. The present paper will deal only with 'type-faces' ;
the question of the part which is played by printing papers
and printing inks will be discussed in a later paper.
^ The term ' face ' is employed here and throughout this paper in
the sense in which it has come to be used by printers and type-
founders. It is customary to group the numerous variants of letter-
form into families, — Caslon, Cheltenham, Jenson, Ronaldson, etc., —
and to speak of each family or typical variant from the common letter-
form as a ' face.'
247891
2 ROETHLEIN
At the time when the art of printing from individual or
moveable types was first introduced, the forms of the letters
of the alphabet were few in number and exceedingly crude
in design. The use of 'black letter' had been popularized by
generations of manuscript writers ; and the early printers
were content to appropriate those letter-forms which they
found to be in current use. But within a few decades
designers and draughtsmen set themselves the task of simpli-
fying and improving the existing forms of letters ; and, indeed,
it seems probable that the Roman 'faces' were introduced
about the year 1465, and the Italic 'faces' some thirty-five
years later. The ingenuity of many generations of mediaeval
and modem designers has produced a multitude of variants
of letter- forms, many of which are familiar to every reader
(a, a, a, a, a; g, g, g, g, g). Hundreds of different 'faces'
of type have been designed and ptit upon the market; and
while it is true that certain of these 'faces' are employed
only for purposes of ornamentation and display, yet an enor-
mous variety of letter-forms is to be found in our books and
magazines. Every reader has observed that all of these
variants of letter-form are not equally legible — an obser-
vation which raises the theoretical question: What are the
factors upon which legibility depends? And the practical
question : How should one proceed if one set out to improve
the legibility of printed letters?
II. HISTORICAL
Psychologists have been engaged these many years in an
investigation of the act of reading in its various aspects. But
there is a singular dearth in the literature so far as the
specific topic of the present investigation is concerned.
More than forty years ago, Exner (17) and Baxt (2)
undertook to measure the brief period of time which is
necessary for the perception of visual objects (letters, words) ;
and in 1885 Cattell (6, 7, 8) continued the investigation
of the same problem. Baxt had reported that, under his
most favorable conditions of illumination, it was possible to
read two or three letters of a total group of seven when
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 3
they were exposed for one one-hundredth of a second.
Cattell devised a falHng-screen apparatus which enabled him
to vary and to measure his exposure-times. He found that
the differences in the times which are necessary for the
recognition of isolated letters, whether upper case or lower
case,^ whether Latin or German, are of negligible magnitude.
When the exposure-time was very brief, it was found that the
letters were not always read correctly. A record of the
percentage of correct readings of the various letters, when
presented under constant and uniform conditions, enabled
Cattell to determine the relative legibility of the letters.
The order of legibility (descending) was found to be:
WZMDHKNXAYOGLQISCTRPBVFUJE
and dkmqhbpwuljtvzrofnaxyeigc s, — W being
read correctly in 89 per cent, of the trials, E in 23 per
cent. ; d in 87, and s in 28 per cent.
Sanford (35) employed a similar method, but obtained a
somewhat different result; his order was (for Snellen type,
lower case) :
mwdqvyjp kfblighgrxt ouanescz
Sanford also determined the relative legibility of the same
letters by a distance method, and obtained the following
result :
wmqpvyjf hrdgkbxlnu atizocse
Sanford also tested alphabets representing two other letter-
forms, — a modern face, and a bold oldstyle face. The modern
face letters were recognized in the following order (distance
test) :
dpqmyknw ogvxhbjlia tuzrscfe
The oldstyle letters fell into the following order (tested
by the method of brief exposure) :
mwpqvykb djrlonighu atfsxzce
Finzi (18) employed the method of brief exposure, pre-
senting a group of nine letters at each exposure. From
' The term ' upper case ' will be used throughout to designate the
capital letters, and ' lower case ' to designate the small letters.
4 ROETHLEIN
the percentage of misreadings of each letter he computed
the order of legibility as follows :
P U A Q X T D S E W AI V Y Z H C N F L R G B K O I,
the percentage of errors ranging between .8 for P, and 7.8
for I. The particular type or letter-form which Finzi em-
ployed is not specified in his paper.
Griffing and Franz (21) investigated the influence of size
and form of letter upon legibility. Their experiments com-
prised a fourfold test : What is the difference in the rapidity
with which small print and large print can be read (five-
point and tw^elve-point, both Roman) ? How many letters
of each of these two sizes can be read in a single brief
exp: sure? How long must letters of each size be exposed
in order to insure their correct reading? What intensity of
illumination is necessary for the recognition of letters of
various forms and sizes, — Roman letters, .8 mm. and 1.6 mm.
high; Gothic letters, .9, 1.6, 3.1, and 6.0 mm. high? These
investigators found that their larger types were, in every
instance, more legible than their sm.aller types ; and that
Gothic letters were m.ore legible than Roman letters.
Besides these experiments which have just been described,
numerous attempts have been made to investigate other factors
which have to do with the act of reading. The nature and
the extent of the eye-movements by means of which the reader
follow^s the printed line have been examined and measured
by Huey (22, 23, 24), Dodge (13, 14, 15), Erdmann and
Dodge (16), Dearborn (10), and others. It has been estab-
lished that the movement of the reader's eyes does not pro-
ceed gradually and continuously across the page, in any
such fashion as, for example, a meteor moves across the
sky. Typical eye-movements consist of a succession of
alternate leaps and pauses ; nor is the movement always in
a forward direction, because it frequently happens that one's
eye-movement proceeds backward, i.e., to the left, from
an intermediate fixation-point. The number of pauses may
vary from two to seven in a line whose length is twenty
centimeters ; but the usual distance between successive fix-
ation points or pauses is approximately 2 cm. It seems
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES $
probable that no words are seen while the eyes are in move-
ment; and that the act of reading a printed line consists of a
series of interrupted glimpses, during each of which one
reads a small section which extends to the right and to the
left of the fixation-point.
This discovery makes it seem probable that a considerable
part of the printed line is imaged, not upon the fovea, but upon
para-foveal regions of the retina. And the capacity of these
paracentral regions to distinguish the forms of letters be-
comes an important topic for investigation. This problem
has, indeed, been attacked by Kirschmann (28) and by
Dockeray (12) ; but further investigation is needed before
one can make any definite statement regarding the legibility
of letters in indirect vision.
Numerous other investigations of the problems of reading
have been made by Babbage (i), Becher (4), Goldscheider
and Miiller (19), Javal (26, 27), Maire (31), Messmer (32),
Pillsbury (33), Quantz (34), Schumann (t,/), and Zeitler
(40) ; but it seems more appropriate to discuss their results
in connection with our own findings, than to summarize them
in this section.
III. EXPERIMENTAL
A. ISOLATED LETTERS
a. Materials and Apparatus
The materials which were employed in the present in-
vestigation were sheets of printed letters ; and the apparatus
consisted of a mechanism by means of which these letters
could be presented at a variable distance from the observer.
When the investigation was first undertaken, we made
a careful examination of the various faces of type which are
listed in the sample-books of the American Type Founders
Company.^ We selected fifty faces of type, — comprising
' The author is indebted to the American Typefounders Company,
Jersey City, for a liberal donation of types and of prints, without
which the investigation would have been impossible. Especial thanks
are due to Messrs. Frank B. Berry, L. B. Benton and Morris Benton
of that firm for valuable suggestions regarding letter-forms and
6 ROETHLEIN
some thirty ordinary faces, together with such variants as
italic, bold, condensed, expanded and various combinations
of these variants.
In these earlier experiments, it was decided to employ the
method of brief exposure; and an apparatus was devised
which provided for a succession of exposures, each one one-
thousandth of a second in duration. The series of exposures
of any given letter was terminated by the depression of a
key by the observer; and an automatic counting device re-
corded the number of exposures which had been necessary
for the reading of the letter. For certain technical reasons
this apparatus and mode of procedure were abandoned after
a number of preliminary experiments had been made; and
the distance test was substituted.*
The apparatus, by means of which we obtained the results
on which this paper is based, consisted of a long bench along
which moved a sliding carriage containing the letters to be
read. The bench was 440 cm. long and 15 cm. wide. Its
proximal end was 78 cm., and its distal end 63 cm. above
the floor, so that the observer was able to assume the primary
position of regard throughout. The higher end of the bench
was provided with a vertical support, which carried a head
rest, — the hood of a stereoscope. This device enabled the
observer to assume and to maintain a constant position in
relation to any point on the scale upon the side of the bench.
regarding the interpretation of our results. We are also under obli-
gation to Dr. H. L. Koopman, of Brown University, and to Mr.
C. Chester Lane of the Harvard Press; Mr. L. D. Evans of the River-
side Press, Cambridge, Mass., has furnished us with valuable data
concerning compositors' errors and proof-readers' errors.
* It is a well-known fact that concentration of attention has a very
pronounced efifect upon reaction-time, and that the duration of the
reaction varies with variations in degree of concentration. It seemed
difficult, even impossible, to maintain the same degree of concen-
tration through thousands of readings of letters ; and, in the absence
of a control of this exceedingly influential factor, the results of our
tachistoscopic experiments seemed wholly unreliable. For this reason
the tachistoscope was abandoned, and a method which consisted in pre-
senting the letters at variable distances was substituted.
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 7
The carriage which moved along the bench consisted
of a box 55 cm. high, 40 cm. wide, and 25 cm. deep.
The front of the box had been cut away ; and the back
consisted of a wooden wall against which the sheet of
printed letters was attached and held in place by a
sheet of glass. The back wall of the box was illuminated
by a number of electric lamps, so arranged that the whole
surface of the sheet of letters was uniformly illuminated.
To the side of the carriage was attached an indicator, which
just cleared a metric scale upon the side of the bench ; this
device enabled the experimenter to ascertain the distance of
the sheet of letters from the eye of the observer at any given
setting of the carriage.
The sheets upon which the isolated letters were printed
were 21 cm. wide and 36 cm. long; and the paper of all
of the sheets was of the same quality and texture.^ Each
sheet contained twenty-eight letters, all of the same face
and of the same case, — the complete alphabet, with two of
its letters repeated. The letters were arranged in random
sequence, in four lines ; they were so spaced that each letter
stood at a distance of 3.7 cm. from its nearest neighbors
on the same and on adjacent lines. Fifty-two different
sheets of letters were investigated, representing the following
twenty-six faces of type, both lower case and upper case.
All of our letters were of the size which is technically
described as ten-point ; the reader will find them illustrated
in Tables I and II. (See inserts between pages 8 and 9.)
American Typewriter
Bold Antique
Bulfinch
Caslon Oldstyle No. 540
Century Oldstyle
Century Oldstyle, Bold
Century Expanded
Cheltenham Oldstyle
" This paper is technically described by the manufacturers as a white,
coated book-paper, 25 x 30 — 80.
a ROETHLEIN
Cheltenham Bold
Cheltenham Bold, Condensed
Cheltenham Italic
Cheltenham Wide
Clearface
Clearface Bold
Clearface Italic
Clearface Bold Italic
Cushing No. 2
Cushing Oldstyle No. 2
Cushing Monotone
Delia Robbia
DeVinne No. 2
DeVinne No. 2, Italic
Franklin Gothic
Jenson Oldstyle No. 2
News Gothic
Ronaldson Oldstyle No. 551
b. Method of Procedure
All of the readings were made in a semi-darkened room,
in order that the (artificial) illumination upon the sheet of
letters might be controlled and kept constant throughout.
The experimental procedure was as follows : After the
observer had become adapted to the illumination of the room,
a sheet of letters was placed in position in the carriage and
the series of readings began with the carriage at the farther
end of the bench. The observer had been instructed to read
the letters at a uniform tempo, substituting "blank" for the
name of any letter which was not easily decipherable. This
precaution seemed necessary to prevent the observer from
giving an undue amount of attention to any one letter of the
series at the expense of the other letters, i. e., to prevent him
from puzzling longer over one member of the series than
over any other member. After he had thus attempted to
read through the complete list of letters, the carriage was
moved to a point twenty centimeters nearer his eye than
the initial setting; and he made a second attempt to read
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 9
the letters. In the first and alternate readings of each series
he began at the upper left-hand corner of the sheet and
proceeded from left to right along each line, taking the lines,
in order, from the top downward. In the second and alter-
nate readings, he began at the lower right-hand corner of
the sheet and proceeded from right to left, and from below
upwards. This procedure was continued, the carriage being
advanced step by step, until every letter upon the sheet
had been identified. The experimenter was provided with
a duplicate sheet of letters, similar to the one which had
been inserted in the carriage ; and upon this duplicate sheet
she recorded the misreadings, and the farthest distance at
which each letter was read.
Each of the fifty-two sheets of letters was read twice
by each of the six observers. Before being inserted in the
carriage for its second reading, each sheet was cut along
its longitudinal and its transverse diameters ; and the four
quarter-sheets v/ere reassembled in such fashion that those
letters which had formerly appeared upon the marginal regions
of the original sheet now appeared upon central regions of
the reconstructed sheet, and vice versa. A period of several
days always elapsed between the first and the second reading
of any sheet.
c. Observers
The observers were instructors or students in the depart-
ment: Messrs. R. Acher, J. W. Baird, E. O. Finkenbinder,
F. A. Lombard, H. B. Moyle, and C. W. St. John ; they all
possessed emmetropic or adequately corrected vision. Each
observer gave a complete series of one hundred and four
readings.
B. GROUPED LETTERS
In the second group of experiments the apparatus and
the method remained unchanged, but here the letters were
presented in groups instead of singly. In these later experi-
ments only lower case letters were employed. Eight faces
were elected from the twenty-six which had already been
lO ROETHLEIN
used, and Scotch Roman was added, — the complete list of
nine faces being as follows :
Bulfinch
Caslon Oldstyle No. 540
Century Oldstyle
Century Expanded
Cheltenham Wide
Cushing- Oldstyle No. 2
Cushing A-Ionotone
News Gothic
Scotch Roman
Each group of letters formed a nonsense combination;
and the groups of each face were arranged in three lines
upon sheets of the same size as had been employed in the
former experiments. Sets of grouped letters were printed,
in duplicate, upon coated book paper, of the good quality
and upon an antique laid book paper of the same quality
and weight as the paper of this page.
Our mode of grouping the letters aimed to introduce as
many difficult and confusing combinations of letters as pos-
sible. We were guided in the combining of the letters into
groups by data furnished by Mr. L. D. Evans, an expert
proofreader, and by the confusions which had been recorded
in our earlier experiments. The combinations of letters which
are here appended illustrate a typical series of groups ; they
also illustrate the Scotch Roman face.
ksitugy cdzxpbj ftoceygqa wrvlindh
hknurfkxzqg munimm bhwvjyst oceo
wvxarlizxp ybhdonactilsf dnupqcetrlj
Only three observers, Messrs. Baird, Finkenbinder and
St. John, took part in these experiments. Each observer
gave two readings of each face, — an average of twenty-four
readings of each letter of each face.
TABLE I. UPPER CASE. ISOLATED LETTERS
Showing the Average Distance, Expressed in Cm., at Which Each Letter op Each Face was Read (Twelve Readings, Six Observers). The First Column in Each Division of the Table Shows the Actual Size and Form op the
Letter Which was Presented for Identification; and the Number Indicates the Average Distance, from the Eye, at Which the Letter was Correctly Identified
American
Typewriter
Bulfinch
Caslon 0. S.
Century 0. S.
Century
Expanded
Cheltenham
O.S.
Cheltenham
Wide
Clearface
dishing 0. S.
Cushing No. 2
Cushing
Monotone
Delia Robbia
De Vinne
No. 2
Jenson 0. S.
News Gothic
Roualdson
O.S.
Average
A
221.7
A
280
A
270
A
300
A
300
A
291.7
A
281.9
A
257.5
A
241.7
A
234.2
A
293.3
A
265
A
300
A
286.7
A
263.3
272 A
B
176.7
B
231.7
B
206.7
B
221.7
B
213.3
B
210
B
221.7
B
204.2
B
190.8
B
179.2
B
211.7
B
211.7
B
229.2
B
230.8
B
195
208.9
C
195
c
283.3
C
278.3
C
285
C
271.2
C
290
C
288.8
C
259.1
C
243.3
C
251.7
C
282. 5
C
252.5
C
310
C
276.7
C
259.2
265.1
D
1S3.3
D
275
D
265
D
281.7
D
273.3
D
270
D
251.7
D
255.8
D
218.3
D
236.7
D
243.3
D
249.2
D
300
D
265
D
246.7
254.3
E
193.3
E.
230.8
E
213.3
E
241.7
E
240.8
E
259.2
s^H
E
229.6
E
198.3
E
209.2
E
192.5
E
228.3
E
226.7
E
245. S
E
240
E
208.3
223.9
F
191.7
F
254.2
F
210.8
F
265
F
265.8
F
241.7
I5
F
270.7
F
234.2
F
208.3
F
217.5
F
244.2
F
243.3
F
276.7
F
253.3
F
241.7
241.6
G
169.2
G
276.7
G
256.7
G
265
G
255
G
275
0^
G
258.6
G
220
G
197.5
G
214.2
G
256.7
Q
1S4.2
G
273.3
G
247.5
G
213.3
237.6
H
190
H
266.7
H
230
H
250
H
239.2
H
256.7
H
261.2
H
213.3
H
202.5
H
214.2
H
283.3
H
233.3
H
273.3
H
253.3
H
240
240.5
I
230
1
300
I
285
I
295
I
275
I
268.3
I
327.2
I
290
1
244.2
I
251.7
1
301.7
I
273.3
I
296.7
1
296.7
I
270
280.4
J
226.7
J
290
J
285
J
300
J
301.7
J
[281.7
P
J
323
J
295
J
271.7
J
260
J
315
J
275.0
J
320
J
295
J
281.7
287.5
K
177.5
K
262.5
K
220.8
K
256,7
K
241.7
K
254.2
S 5-
K
236
K
228.3
K
205.8
K
200
K
217.5
K
243.3
K
270
K
231.7
K
229.2
231.7
L
248.3
L
305
L
281.7
L
308.3
L
301.7
L
300
sa
L
315.1
L
296.7
L
261.7
L
266.7
L
296.7
L
290
L
311.7
L
300
L
283.3
291.1
M
189.2
M
316.7
M
290
M
313,3
M
298.3
M
303.3
Oo
M
323
M
315
M
273,3
M
275
M
321.8
M
293.3
M
318,3
M
308.3
M
268.3
293.8
N
184.2
N
246.7
N
244.2
N
258.3
N
241.7
N
261.7
r^^O
N
247
N
221.7
N
201.7
N
215
N
258.3
N
225
N
270
N
243.3
N
225
235.5
186.7
O
293.3
O
267.5
293.3
265
O
278.3
274.5
230.8
215.8
219.2
275
226.7
296.7
250
253.3
254.0
P
190
P
301.7
P
245
P
263.3
P
281.7
P
270.8
^ c?"
P
286
P
255
P
238.3
P
229.2
P
252.5
P
263.3
P
296.7
P
278.3
P
229.2
257.9
ft
214.2
Q
308 . 3
Q
270
Q
268,3
Q
275
Q
291.7
Q
285
Q
255.8
Q
220
Q
245
Q
246.7
Q
240.9
Q
295
Q
278.3
Q
231.7
261.7
R
186.7
R
227.5
R
200.8
R
210.7
R
235
R
255
~
R
216.9
R
198.3
R
196.7
R
180
R
228.3
R
193.3
R
234.2
p
240
R
205.8
214.0
S
168.3
5
236.7
S
197.5
S
225
S
209.2
S
215
^
S
211.2
S
210
S
182.5
S
183.3
S
203.3
S
200
S
223.3
s
216.7
S
203.3
205.7
T
220.8
T
276.7
T
280
T
305
T
280
T
271.2
I'
T
273.9
T
280
T
219.2
T
237.5
T
295
T
251.7
T
308.3
T
280.8
T
251.7
268.9
U
200.8
U
268.3
U
247.5
U
256 7
U
276.7
U
266.7
3
U
269.1
U
233.3
U
220
U
232.5
U
275
U
240.9
U
270.8
U
273.3
U
237.5
251.3
V
205
V
275
V
255.8
V
281.7
V
278.3
V
2SS.3
V
277.6
V
246.7
V
237.5
V
236.7
V
291.7
V
248.3
V
310
V
264.2
V
256.7
263.5
w
165
w
310.7
W
291.7
w
306.7
w
316.7
W
306.7
«)
w
318.2
W
331.7
W
288.3
W
285
w
343,3
w
315
W
306.7
w
316.7
w
301.7
300.2
X
192.0
X
260.8
X
237.5
X
257.5
X
250
X
263.3
?•
X
247
X
240
X
212.5
X
222.6
X
252.5
X
236.7
X
256.7
X
240
X
228.3
239.8
y
221.7
Y
268.3
Y
260
Y
271.7
Y
261.7
Y
249.2
c_
Y
266
Y
250
Y
241.7
Y
233.3
Y
263.3
Y
225
Y
261.7
Y
258.3
Y
250.8
252.1
z
187.5
z
261.7
Z
228.3
Z
250
Z
237.5
Z
261.7
Z
240.7
Z
218.3
Z
204.2
Z
207.5
z
256.7
Z
216.7
Z
270
Z
255.8
Z
210.7
233.8
Average IQQ. 8
273.8
250.7
270.4
264.8
268.5
269.3
247.6
224.8
228.4
266.8
243,2
281.7
264.6
241.7
252.8
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES II
d. Results
A. ISOLATED LETTERS
The experimenter's record-sheet contained a statement of
the distance at which each letter was read, together with a
list of the confusions or misreadings, and an introspective
description of the procedure which the observer had followed
in deciphering the letters. The numerical data have been
compiled and tabulated in various ways.
Tables I and II show the averages of the numerical results,
arranged in order of faces. These two tables report the
data for 'ordinary' faces only, — the italic, the bold and the
condensed faces not being included here. The numbers which
appear in these tables indicate the averages of the extreme
distances at which the letters were read, — hence the larger
the number appended to any letter, the greater the legibility
of that letter. Thus, in the first column of Table I, "H 190,
I 230" may be taken to signify that the upper case I of the
American Typewriter face is considerably more legible than
the upper case H of the same face.
Each vertical column of these two tables contains, there-
fore, a statement of our findings regarding the relative legi-
bility of the various letters of a given face; and the number
at the foot of the column indicates the average legibility
of the twenty-six letters of that face. The numbers in each
horizontal line of the tables indicate the relative legibility
of each of the sixteen variants of each of the twenty-six
letter forms.
Table III presents the grand averages of the sixteen faces;
it also contains similar data for the bold and for the italic
faces and for two extra-bold faces, Franklin Gothic and
Bold Antique ; while Table IV shows the grand average
distance at which each letter of the alphabet was read, the
data being here compiled from the readings of the complete
set of sixteen faces.
The efifect of certain definite modifications of a given letter-
form is shown in Tables V and VI. These two tables are com-
piled from results which were obtained with Cheltenham Old-
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES II
d. Results
A. ISOLATED LETTERS
The experimenter's record-sheet contained a statement of
the distance at which each letter was read, together with a
list of the confusions or misreadings, and an introspective
description of the procedure which the observer had followed
in deciphering the letters. The numerical data have been
compiled and tabulated in various ways.
Tables I and II show the averages of the numerical results,
arranged in order of faces. These two tables report the
data for 'ordinary' faces only, — the italic, the bold and the
condensed faces not being included here. The numbers which
appear in these tables indicate the averages of the extreme
distances at which the letters were read, — hence the larger
the number appended to any letter, the greater the legibility
of that letter. Thus, in the first column of Table I, "H 190,
I 230" may be taken to signify that the upper case I of the
American Typewriter face is considerably more legible than
the upper case H of the same face.
Each vertical column of these two tables contains, there-
fore, a statement of our findings regarding the relative legi-
bility of the various letters of a given face; and the number
at the foot of the column indicates the average legibility
of the twenty-six letters of that face. The numbers in each
horizontal line of the tables indicate the relative legibility
of each of the sixteen variants of esch of the twenty-six
letter forms.
Table III presents the grand averages of the sixteen faces ;
it also contains similar data for the bold and for the italic
faces and for two extra-bold faces, Franklin Gothic and
Bold Antique ; while Table IV shows the grand average
distance at which each letter of the alphabet was read, the
data being here compiled from the readings of the complete
set of sixteen faces.
The effect of certain definite modifications of a given letter-
form is shown in Tables V and VI. These two tables are com-
piled from results which were obtained with Cheltenham Old-
12 ROETHLEIN
style presented in ordinary, in bold, in bold-condensed, (in
wide,) and in italicized form.
Tables VII, VIII, IX, and X show the five variants of each
letter which proved to be most legible, and the five which
proved to be least legible.
B. GROUPED LETTERS
The results of our second series of experiments are pre-
sented in Tables XI and XII. Table XI shows the average
distance at which each letter of each of the nine faces was
read when presented in groups. The reader is warned against
drawing any conclusion from this table regarding the relative
legibility of the various letters of any given face, e.g.,
regarding the relative legibility of the Caslon m and the
Caslon, or any other, k. In other words, the numbers v/hich
appear in the horizontal lines are comparable with one an-
other; but the numbers which appear in the vertical columns
are incomparable with one another. Table XII presents
the same results as are contained in Table XI, but they are
now arranged in order of magnitude in order to show the
(descending) order of legibility of the several variants of
each letter-form.
IV. DISCUSSION OF RESULTS
A. ISOLATED LETTERS
The problem with which we are here concerned may be
given the following general formulation : Which of several
geometrical figures is most clearly perceptible, and most
readily distinguishable from other geometrical figures? But
while the twenty-six letter-forms which constitute the alpha-
bet may be regarded as a series of geometrical figures of
dififerent shapes and of different degrees of complexity of
detail, and while any twenty-five of the different twenty-six
different faces of type which were employed in the present
investigation may be regarded as variants from the twenty-
sixth or common letter-form, yet, as a matter of fact, such
a simple mode of envisagement of our problem does not do
justice to the complex conditions which are found to be
N IN Each Division op the Table Shows the Actual Size and Form of the
ICH THE Letter was Correctly Identified
a Robbia
De Vinne
No. 2
Jenson 0. S.
News Gothic
Ronaldson
0. S.
Average
157.5
a
174.2
a
183.3
a
190
a
169.2
177.0
223.3
b
195.8
b
220
b
245
b
219.2
217.8
181.7
c
182.5
c
197.5
c
221.7
c
190.8
193.8
276.7
d
253.3
6
246.7
d
256.7
d
245
254.3
175
e
170
^
191.7
e
191.7
e
165
173.5
260
f
205
f
225
f
263.3
f
235
233.0
218.3
g
196.7
g
202.5
g
258.3
g
198.3
220.8
236.7
h
205
h
220
h
235.8
h.
232.5
222.7
219.2
i
214.2
i
231.7
i
249.2
i
204.2
224.1
210.8
1
236.7
210
J
280.8
3
247.5
239.4
206.7
k
215
fc
230
k
233.3
k
224.2
216.9
250
1
190.8
I
251.7
1
226.7
1
243.3
236.3
295
m
315
m
291.7
m
326.7
m
300
296.8
203.3
n
197.5
n
206.7
n
220
n
190
195.7
193.3
o
170.8
o
188.3
207.5
o
180
190.1
239.3
p
225
P
255
p
248.3
P
217.5
236.1
238.3
q
212.5
q
220.8
q
257.5
q
217.5
226.4
190
r
198.3
r
196.7
r
237.5
r
181.7
203.6
138.3
s
154.2
s
150
s
177.5
s
141.8
152.6
191.7
t
191.7
t
188.3
t
223.3
t
202.5
199.6
195
u
194.2
«
180
u
215
u
185.8
193.1
205.8
V
199.2
V
219.2
V
235.8
V
219.2
213.1
280
w
265
w
270
w
305
w
263.3
261.6
171.7
X
185
X
192.5
X
196.7
X
170
181.7
226.7
y
208.3
y
241.7
y
246.7
y
210
224.6
175
z
170
z
171.7
z
199.2
z
185
171.6
214.2
204.8
214.7
236.4
209.2
213.7
TABLE II. LOWER CASE
Showing the Average Distance, Expressed in Cm., at Which Each Letter of Each Face was Read (Twelve Readings, Six Ob,server8). The First Column in Each Division of the Table Shows the Actual Size and Form op the
Letter Which was Presented for Identification; and the Number Indicates the Average Distance, From the Ete, at Which the Letter was Correctly Identified
American
Typewriter
Bulfineh
Caslon 0. S.
Century 0. S.
Century
Expanded
Cheltenham
0. S.
Cheltenham
Wide
Clearface
Gushing 0. S.
Cashing No. 2
Cushing
Monotone
Delia Robbia
De Vmne
No. 2
Jensen 0. S.
News Gothic
Ronaldson
0. S.
Average
a
b
d
e
184.2
208.3
168.3
230
162.5
a
b
c
d'
e
185
233.3
207.5
258.3
181.7
a
b
c
d
e
170.8
190
185
253.3
165.8
a
b
c
d
e
197.5
245
230
199.1
a
b
c
d
e
196.7
233.3
273.3
187.5
a
b
c
d
e
155
221.7
180.8
261.7
169.2
a
b
c
d
e
190.8
245
200.8
276.7
185.8
a
b
c
d
e
182.1
223
235.9
283.4
179.8
a
b
c
d
e
173.3
206.7
179.2
260
164.2
a
b
c
d
e
155.8
171.7
174.2
205. S
140
a
b
c
d
e
155.8
203.3
171.7
233.3
146.7
a
b
c
d
e
167.5
223.3
181.7
276.7
175
a
b
c
d
e
174.2
195.8
182.5
253.3
170
a
b
c
d
e
183.3
220
197.5
246.7
191.7
a
b
c
d
e
190
245
221.7
256.7
191.7
a
b
c
d
e
169.2
219.2
190.8
245
165
177.0
217.8
193.8
254.3
173.5
f
g
h
i
3
241.7
203.3
207.5
206.7
256.7
f
i
i
261.7
251.7
241.7
243.3
228.3
f
S
h
i
i
225
182.5
210.8
209.2
235
f
r
1
i
J
256.7
245
235
245
240
f
g
h
1
J
256.7
241.7
238.3
215.8
283.3
f
I
i
i
235
195.8
211.7
235
234.2
i
S
h
j
241.7
216.7
238.3
251.7
248.3
f
i
j
241.7
250
247
213
264.4
f
e
h
i
j
232.5
231.7
211.7
231.7
227.5
f
g
h
i
J
184.2
225.8
190.8
200
218.3
f
g
h
i
j
178.3
215
200
215
208.3
f
i
j
260
218.3
236.7
219.2
210.8
f
g
h
i
j
205
196.7
205
214.2
236.7
f
i
i
225
202.5
220
231.7
210
f
g
h
j
263.3
258.3
235.8
249.2
280.8
f
g
h
i
j
235
198.3
232.5
204.2
247.5
233.0
220.8
222.7
224.1
239.4
k
1
m
n
188.3
230
195
195.8
188.3
k
1
m
n
o
236.7
256.7
336.7
213.3
220
k
1
m
n
197.5
219.2
291.7
190
180.8
k
1
m
n
240
245
310
213.3
211.7
k
1
m
n
235
252.5
286.7
199.2
213.3
k
1
m
n
o
211.7
240
295
181.7
181.7
k
1
m
n
o
224.2
254.2
315
208.3
189.2
k
1
m
n
223.2
280.2
325.1
213.7
199.5
k
1
m
n
221.7
232.5
315
172.5
172.5
k
1
m
n
183.3
208.3
273.3
162.5
170
k
1
m
n
200
199.2
276.7
163.3
175
k
1
m
n
o
206.7
250
295
203.3
193.3
k
1
m
n
o
215
190.8
315
197.5
170.8
fc
1
m
n
230
251.7
291.7
206.7
188.3
k
1
m
n
233.3
226.7
326.7
220
207.5
k
1
m
n
224.2
243.3
300
190
180
216.9
236.3
296.8
195.7
190.1
P
q
r
s
t
238.3
233.3
210.8
149.2
198.3
P
q
r
s
t
271.7
240
216.7
167.5
230.8
P
q
r
s
t
220.8
221.7
203.3
145
180.8
P
q
r
s
t
240
233.3
216.7
167.5
206.7
P
q
r
s
t
240
233.3
225
165
221.7
P
q
r
s
t
223.3
221.7
187. 5
138.3
183.3
P
q
r
s
t
247.5
219.2
214.2
160
190
P
q
r
s
t
259.7
234.4
215.4
150.4
228
P
q
r
s
t
245
235
205
145.8
208.3
P
q
r
s
t
205
196.7
175.8
145.8
169.2
P
q
r
s
t
201.7
206.7
183.3
145.5
179.2
P
q
r
s
t
239.3
238.3
190
138.3
191.7
P
q
r
s
t
225
212.5
198.3
154.2
191.7
P
q
t
s
t
255
220.8
196.7
150
188.3
P
q
r
s
t
248.3
257.5
237.5
177.5
223.3
P
q
r
s
t
217.5
217.5
181.7
141.8
202.5
236.1
226.4
203.6
152.6
199.6
u
V
w
y
z
198.3
195
183.3
180.8
230
159
u
V
w
X
y
z
225
240
271.7
211.7
245
196.7
u
V
vv
X
y
z
185
180.8
243.3
171.7
221.7
162.5
u
V
w
X
y
z
212.5
230
266.7
200
230.8
182.5
u
V
w
X
y
z
198.3
217.5
268.3
195
218.3
171.7
u
V
w
X
y
z
177.5
205
263.3
169.2
222.5
164.2
u
V
X
y
z
197.5
223.3
293.3
184.2
236.7
179.2
u
V
w
X
y
z
207.4
225.4
296
178.1
237.5
171.4
u
V
w
X
y
z
175.8
225.8
281.7
187.5
227.5
157.5
u
V
w
X
y
z
176.7
183.3
201
157.5
194.2
156.7
u
V
w
X
y
z
165
205
233.3
155
195.8
143.3
u
V
w
X
y
z
195
205.8
280
171.7
226.7
175
u
V
w
X
y
z
194.2
199.2
265
185
208.3
170
o
V
w
y
2
180
219.2
270
192.5
241.7
171.7
w
X
y
z
215
235.8
305
196.7
246.7
199.2
u
V
w
X
y
z
185.8
219.2
263.3
170
210
185
193.1
213.1
261.6
181.7
224.6
171.6
Average201.7
233.6
201.7
228.0
226.7
206.4
224.3
229.5
212.6
185.6
190.6
214.2
204.8
214.7
236.4
209.2
213.7
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES I3
TABLE III. ISOLATED LETTERS
A Comparison of the Average Legibility of Various
Faces. Order of Legibility
UPPER CASE
Lower Case
The Sixteen Roman Faces
JENSON
281.7
News Gothic
236.4
BULFINCH
273.8
Bulfinch
233.6
CHELT. W.
268.5
Clearface
229.5
CENTURY 0. S.
270.4
Century 0. S.
228.0
CLEARFACE
269.3
Century Exp.
226.7
CHELT. 0. S.
268.5
Chelt. W.
224.3
DELLA ROBBIA
266.8
Jensen
214.7
NEWS GOTHiC
264.6
Delia Robbia
214.2
CENTURY EXP.
264.8
Gushing 0. S.
212.6
CASLON 0. S.
250.7
Ronaldson
209.2
GUSHING 0. S.
247.6
Chelt. O. S.
206.4
DE VINNE NO. 2
243.2
De Vinne No. 2
204.8
RONALDSON
241.7
American Typewr.
201.7
GUSHING MON.
228.4
Caslon O. S.
201.7
GUSHING NO. 2
224.8
Gushing Mon.
190.6
AMERICAN TYPEWR.
196.8
Gushing No. 2
185.6
Average
252.8
Average
213.7
Bold Faces
CENT. O.S. BOLD 296
CHELT. O.S. BOLD 286.2
CLEARFACE BOLD 273.7
Average (Bold) 285.3
Cent. O.S. Bold
Chelt. O. S. Bold
Clearface Bold
Average (Bold)
255.1
233.4
230.5
239.7
Italic Faces
CLEAR. ITALIC
274.3
Clear. Italic
231.2
CHELT. ITALIC
259.6
Chelt. Italic
203.8
DE VINNE ITALIC
235.5
De Vinne Italic
201.9
Average (Italic)
256.5
Average (Italic)
212.3
Average (of same
Average (of same
faces, Roman)
260.3
faces, Roman)
213.6
Bold Italic Face
CLEAR. B. ITALIC 265 . 4 Clearface B. Italic 213.2
Extra Bold Faces
BOLD ANT. 307.4
FRANK. GOTHIC 284 8
Bold Antique 260 . 5
Franklin Gothic 245.2
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES I3
TABLE III. ISOLATED LETTERS
A Comparison of the Average Legibility of Various
Faces. Order of Legibility
UPPER CASE
Lower Case
Tht
I Sixteen Roman Faces
JENSON
281.7
News Gothic
236.4
BULFINCH
273.8
Bulfinch
233.6
CHELT. W.
268.5
Clearface
229.5
CENTURY 0. S.
270.4
Century 0. S.
228.0
CLEARFACE
269.3
Century Exp.
226.7
CHELT. 0. S.
268.5
Chelt. W.
224.3
DELLA ROBBIA
266.8
Jensen
214.7
NEWS GOTHIC
264.6
Delia Robbia
214.2
CENTURY EXP.
264.8
Gushing 0. S.
212.6
CASLON 0. S.
250.7
Ronaldson
209.2
GUSHING 0. S.
247.6
Chelt. O. S.
206.4
DE VINNE NO. 2
243.2
De Vinne No. 2
204.8
RONALDSON
241.7
American Typewr.
201.7
GUSHING MON.
228.4
Caslon O. S.
201.7
GUSHING NO. 2
224.8
Gushing Mon.
190.6
AMERICAN TYPEWR.
196.8
Gushing No. 2
185.6
Average
252.8
Average
213.7
Bold Faces
CENT. O.S. BOLD 296
CHELT. O.S. BOLD 286.2
CLEARFACE BOLD 273.7
Arem^e (Bold) 285.3
Cent. O.S. Bold
Chelt. O. S. Bold
Clearface Bold
Average (Bold)
255.1
233.4
230.5
239.7
Italic Faces
CLEAR. ITALIC
274.3
Clear. Italic
231
2
CHELT. ITALIC
259.6
Chelt. Italic
203
8
DE VINNE ITALIC
235.5
De Vinne Italic
201
9
Average (Italic)
256.5
Average (Italic)
212
3
Average (of same
Average (of same
faces, Roman)
260.3
faces, Roman)
213
6
Bold Italic Face
CLEAR. B. ITALIC 265.4
Clearface B.Italic 213.2
Extra Bold Faces
BOLD ANT. 307.4
FRANK. GOTHIC 284 8
Bold Antique 260.5
Franklin Gothic 245.2
14
ROETHLEIN
TABLE IV. ISOLATED LETTERS
The Relative Legibility of the Letters of the
Alphabet. (Average of Sixteen Faces;
192 Readings of Each Letter)
UPPER CASE
Lower Case
W
300.2
m
296.8
M
293.8
w
261.6
L
291.1
d
254.3
J
287.5
j
239.4
I
280.4
1
236.3
A
272.4
P
236.1
T
268.9
f
233.0
C
265.1
q
226.4
V
263.5
y
224.6
Q
261.7
i
224.1
P
257.9
h
222.7
D
254.3
g
220.6
254.0
b
217.8
Y
252.1
k
216.9
U
251.3
V
213.1
F
241.6
r
203.6
H
240.5
t
199.6
X
239.8
n
195.7
G
237.6
c
193.8
N
235.5
u
193.1
Z
233.8
190.1
K
231.7
X
181.7
E
223.9
a
177.0
R
214.0
e
173.5
B
208.9
z
171.6
.s
205.7
s
152.6
Average
252.8
Average
213.8
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES I 5
TABLE V. UPPER CASE
The Effect of Various Modifications of a Given Face
Cheltenham
RnlH
Ordinary
^
Bold
Condensed
Wide
Italic
A 291.7
A
303.3
A
190.0
A
273.3
B 210.0
B
236.7
B
204.2
Co
B
220.0
C 290.0
C
295.0
C
253.3
C
283.3
D 270.0
D
284.2
D
260.0
"S
D
261.7
E 259.2
E
273.3
£
225.8
"1
E
237.5
F 241.7
F
281.7
F
239.2
F
235.0
G 275.0
G
280.0
G
232.5
G
240.0
H 256.7
H
295.0
H
213.3
H
263.3
I 268.3
I
306.7
I
273.3
I
270
J 281.7
J
295.0
J
290.0
~ r
0.1 t^
J
276.7
K 254.2
K
277.5
K
234.2
. Widi
of thii
K
240.8
L 300.0
L
311.7
L
280.0
L
288.3
M 303.3
M 321.7
M
295.0
i S
M 301.7
N 261.7
N
270.0
N
216.7
•li
15
TV
256.7
278.3
O
295.0
243.3
O
280.0
P 270.8
P
293.3
P
240.8
P
286.7
Q 291.7
Q
300.0
Q
258.3
^■l
Q
286.7
R 255.0
R
253.3
R
205.0
R
244.2
S;i 215.0
S
228.2
s
187.5
? lettef
Style
S
206.7
T 271.2
T
306.7
T
271.7
T
258.3
U 266.7
U
298.3
U
240.0
^1
U
263.3
V 288.3
V
305.0
V
254.2
V
240.8
W 306.7
W 323.3
w
300.0
^^
W^ 278.3
X 263.3
X
270.0
X
255.0
X
242.5
Y 249.2
Y
273.3
Y
256.7
Y
245.0
Z 261.7
Z
263.3
Z
235.0
Z
268.3
Average 268.5
286.2
248.2
^
259.6
i6
ROETHLEIN
TABLE VI. LOWER CASE
The Effect of Various Modifications of a Given Face
Cheltenham
"RnlH
Ordinary
Bold
Condensed
Wide
[talic
a
155.0
a
195.8
a
167.5
a
190.8
a
159.2
b
221.7
b
221.7
b
209.2
b
245.0
b
195.0
c
180.8
c
198.3
c
185.8
c
200.8
c
192.5
d
261.7
d
285.0
d
275.0
d
276.7
d
263.3
e
169.2
e
185.8
e
170.8
e
185.8
e
165.0
f
235.0
f
244.2
f
214.2
f
241.7
/
261.7
g
195.8
g
222.5
g
202.5
g
216.7
g
216.7
h
211.7
h
251.7
h
238.3
h
238.3
h
241.7
i
235.0
i
226.7
•
I
206.7
i
251.7
i
251.7
i
234.2
•
J
280.0
•
J
246.7
J
248.3
J
222.5
k
211.7
k
225.8
k
244.2
k
224.2
k
217.5
1
240.0
1
270.0
I
231.7
1
254.2
I
230.0
m
295.0
m
301.7
m
273.3
m
315.0
m
273.3
n
181.7
n
217.5
n
175.8
n
208.3
n
198.3
o
181.7
o
192.5
160.0
o
189.2
169.2
p
223.3
p
255.0
P
225.8
p
247.5
p
223.3
q
221.7
q
255.0
q
205.8
q
219.2
q
196.7
r
187.5
r
223.3
r
199.2
r
214.2
r
205.8
s
138.3
s
140.8
s
135.8
s
160.0
s
131.7
t
183.3
t
222.5
t
208.3
t
190.0
t
192.5
u
177.5
u
217.5
u
170.8
u
197.5
u
179.2
V
205.0
V
263.3
V
208.3
V
223.3
V
158.3
w
263.3
w
300.0
w
261.7
w
293.3
tt
245.0
X
169.2
X
190.8
X
185.0
X
184.2
X
174.2
y
222.5
y
245.0
y
224.2
y
236.7
j;
174.2
z
164.2
z
175.8
z
154.2
z
179.2
z
160.0
Average 206.4
233.4
205.9
224.3
203.8
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 17
t^ t>;
C^ CO t> l>
CO l>;
CO CO CO
.-; CO t^ »0 !>; 00
1— J T-H lO
-*' CO CO CO
CO rH »0 t^ 10
00 00 00 lO o
05 CO CO (M' i-I ic
05 (M 00 I> '^
»0 ^ lO (32
'^f i-H 'rtH I^
!>. t^ (N '-H 00
CO t^ i-H ut) CO lO
<N (M (M (M (M
CM (M (M (M CO
CO CO CO (N (M
(M (N (M (M C-1
<M CSI CO (M C^ (N
cooi
c»
DeVinne No. 2
Century Exp.
Cushing 0. S.
Clearface
Delia Robbia
cj cj .3 OD
Ci o .3 <^
^66^''S
^ d
3S:i«^d
X-p Xi Irt
6 >:>p6S
-g^^cb
o o o -:
0)^00 o
2 bOp^ c^O
i5 -2 3^ g
CU el c3 oj :S
0000:2;
Bulfin
Clearf
Chelt.
Jenso]
Centu:
News
News
Delia
Chelt.
Cushii
Clearf
Centu
News
Delia
Jenso:
News
<QQUOui
CO i>. (Z) CO 00
u-on:^H-,
iiijSZO
clOq^c/^H
D>^X>HM
C^ !>■ !>.
(M !>. CO CO
1> (N t^ 00
00 t^ !>; t>; CO I>
CO r-I 06 CO
10 >0 i-H CO i—H
T^' lO CO 00 00
T-i LO Tt^ CO CD
C5 i-I CD CO CO CO
Oi (M 00 t^ 'Tt^
CO CO CO Gi
lO O T-H lO t^
CO 00 CO r-< CO
t^ 00 T-i «J0 CO to
(N Ca C^ C^ (M
(M (M (M (M CO
(M CO CO (M (M
(M (M <M <M (M
(N (N CO (M (M <M
S o^d,
COM dn
OQ
ci( o o
02 d, .^ .5
^ XX
. • -^ X
•
X •;:3-p
.X JDX!
^ WW
00 , :SW
CO o
f=^ n. ^^
OH ^^
O)
cd S " b b
•^ C3 C3 S-H !-i
fcH ^H CS ^ S-,
o-^-g >o
b^ ^op
o o
Delia
Clearf
Clearf
Centu
Centu
Centu
Centu
Clearf
NeAvs
Centu
Chelt.
Bulfin
Bulfin
Centu
Chelt.
Centu
Clearf
Jensoi
News
News
g5| g^^
c! « g d-ajlJ
<CQOQW
Ul^OX — *^
^^jS^O
pHac^cA)h-
^>^X>N
(N !>.
00 CO t^
t^ CO CO CO CO
!>. CO
CO CO Ca »0 !>;
ci lO T-H
10 CO CO »o
CO 00 00 00 CO
CO T-H LO CO "jO
CO 00 00 J> CO CO
(M C5 t^ "^
CO t^ CO rH
lo o '— 1 lO cri
00 O CO C^ C5
1^ CO >— 1 ^ CO i— 1
CO <N <N (M C<)
(N (M (M CO CO
(N CO CO c^ csi-
(M (M (N (N (M
C^ (>] CO (M <N <N
CO
Ph .<S
0202 .^
si .2
O CO
^ 6
X ^
do 5
„»a S
News Gothi
Chelt. 0. S.
Clearface
Century 0.
Clearface
Bulfinch
Jenson
Jenson
Chelt. 0.
Bulfinch
Century
Century
Jenson
Bulfinch
Bulfinch
Delia Ro
Century
Century
Jenson
Delia Ro
Bulfinch
Clearface
Chelt. 0.
Century
Jenson
Delia Ro
-^(fiUC^U^
feOXl-.-o
^hJ^ZO
Onap^c/ii-
3>^X><^s^
CX) CO !>. c»
!> CO !>.
»0 !>; 00 1> CO
t^
!>. t^ 00 CO t>.
CO CO 1-H 10
>0 CO '-^
C^' r-H r-H ,-H CO
CO lO O »0 lO
»0 1— 1 '— 1 O 00 r- 1
CO C^ GO ^
t^ t^ t^ (N
CO '-< (M CO C5
C2 Ci 'f <M O
t^ Ol CO O CO CD
CO cs (M <M cq
(M (M (N CO CO
CS| CO CO (N (N
(M Oa <M <M CO
C^ (M CO CO (M (N
^.2 "^
o3
.S CO
002 02
.^.^02
d:S d
> 1
:i^d
|dd
^■^6 ^
tury
vs Go
finch
tury
son
Clearface
Chelt. 0.
Jenson
Delia Ro
Jenson
finch
son
laRo
It. 0.
tury
o o 03 --^ r^
CO M >- "t; ■^
la Ro
la Ro
tury
finch
.finch
!lt. 0.
<U ^ ^ OJ 0)
w^Ooo
£ g S g g
(U <u S ^ ^ ^
QQ^Wpqo
<qqUQw
fcOffi-^
^^^Zzo
aHOQ::c/}H
D>^X>-N
1>. (N
l>- !>. CO (M
T-H !>.
i> CO •>: <^
t>. CO CO !>.
i-H ci
CO CO CO t^ CO
O »^ CO O CD
T-I 00 >0 CO 00
CD O CO CO T-H O
CO '— 1 >o
t^ t^ 00 Oa (M
!>. rH O^ t^ C5
O O to CO o
t^ ^ ^ CO t^ t^
CO (M CO CO C^
(M (M <N CO CO
C^ CO CO (M C<l
CO CO (M CSI CO
c^ CO CO c^ c^ oi
pH
o3
d, S3 CO
•r— 1
02
& :i«^d
Century
Bulfinch
Jenson
Jenson
Chelt. 0.
Jenson
Bulfinch
Delia Ro
Clearface
Clearface
Jenson
Clearface
Clearface
Jenson
Jenson
Bulfinch
Bulfinch
Chelt. 0.
Bulfinch
Jenson
U Century
V Jenson
W Delia Ro
X Chelt. 0.
Y Century
Z Jenson
<iicQUQuJ
feOX^^
t^jSlZiO
cuOoCcoH
I8
ROETHLEIN
CO 00
t>. !>. t^ CO
CO !>. CO to
to 'i^ CO
CO '^ tq !>; l>.
lO CO O O to
CD i-H CO lO 00
TjH i-i to CO CJ
1>I to to O GO
00 to O cm' CD T-H
00 Ol O O 00
to ^ CO CO "^
CM to 1-1 o c:5
^+1 CO i-H CD O
Ol CM 00 C5 CO t^
T-H (M (N (M 1-1
iM <N (M CN (M
C^l CM CO CM rt
CM
,; 6
CM CM CM 1-1 CM
T-l CM CM rH CM 1-1
a5
CZ2 d, g
02 02
d .s^ .
d&^^ .
^^
• d -d
X IS02 02
W 3 • . .
l-H
pqoooo
bbf§o^
OJ OJ <D H rj
OOQOO
CI f2 ^
^ bJD 1 ^ bX)
O- W °5 CD tB
OOOOO
_g T! o3 o -t^ o
1-3
(O-O O 'C u
<+H tJO-C .
^« S c o
D< O* !« M +->
3 > ^ « >> »4
CO LQ CO lO
t>. CO CO t^
CO to I> to
CO CO !>. 1>
'^ oo J>; CO tq CM
W
o CO t^ CO t^
CD LO CO CO CD
CO cm' to CO i>^
CO (Z5 CO CO r-J
i>^ to i-J CO i> oi
i-q
Oi (M O t^ (X)
lO "* CO '^ to
CO to T-l T-H O
Tt^ CO 1-H O C-1
O CM 00 O CO t^
fa
T-H (N (M (M rl
(>J (M <M <M (N
CM CM CO CM CM
CM CM CM i-H Ol
CM CM CM T-l CM 1—1
o
a ?3 &( ci
doi ^•
u d, a
o S 02 d d
^ ^ .2
o
O
M
Q
O
o
X . >
WO . s
^, dd^ ,, .
O O C ^ ^
op^l bb
o O) ;3 J3 S
-+J -(^ ;ii cC .
ooom<5
ob^^o
<;oooi2;
ps S C c ri
;2;Gooo
1^ o <^ t>-
OoO^OQ
(OjD O'XS O)
«w bjO-fl.--'-3
^^ S c o
Q. O" V-l M 4-3
3 > ^ X >» N
02
00 !> t^ !>.
CO TtH
CO CM --H CO t^
t^ tq CO
tq CO t^ tq
O lO i-H O >— 1
O O 00 LO '#'
CO •^' to CO i-H
to O CO t^ CO
CM O CO to 1-5 CM
Q
CI '^i <M t^ Ci
CD to CO ^ CO
CO to CM '— 1 r-H
to to ^ CO CM
1-1 CO Oi C5 '^ oo
--• <M <M CSJ T-*
(M (M (N (M (N
CM CM CO CM CM
CM C>1 C<l ^ CM
CM CM CM ^ CM T-(
O O .^ O
.2 doi
0202
CO o
0202 d, .02
<1
. -tJ -^J X! -t^
-o WO
Ol . .
• . oo
02 .-2
• 0:5
X02 .
OO -W -o
g
> O O O O
o oi o
H-l C3 ^ fH CS
^^1 bo
-+J -(^ --; +f tn -*J
P5 g <u rt c C
O a> rd O (T) o
s
^ O) <D O 0)
^ CO _( _i 03
Qoooo
O ^ B ^ ai
Scd cd -^ 1
g'3'^ g|
O
<
(0J2 OTP <U
<4- "(Mr^.^.^
v^^ S fi o
a cr t- co+->
D > ^ X >«N
1> !>. !>;
I-^ !>. !> (M CO
!>; !>. CO CO
t^ j:^ to
00 t^
CO »0 O CO '— <
1— i 1— 1 1— I Ci o
to GO CD CO CO
O O CO t^ 00
to to CD O to CO
P4
Ci ^ CO 1> C5
O to ^ '^i O)
CO to O-l i-H .— 1
to to CM CO CM
1— 1 CO o o Tti oi
^ (M C^ iM ^
(N (M <M (M (M
CM CM CO CM CM
CM CM CM --1 CM
CM CM CM CM CM T-l
02
idn 02 .
O O
d o d,
d d
Q o 02
1— 1
X . CO
^ ^ fi
• f-H tH
X •■P, X
W :S H
X X
WW
o
js jzi ^ o 9
blo-^ b
551^^1
o
-1^ ;=; +.5 ;=H CO
C 2 rt a) cj
sis P P
3 3 3 53 li
ll il-2
g3 S 3 g
•♦-I 3 3 C H-<
-3 S g ^-S
OOOPQO
gg'slll
.^^-2 o == ^
!2;^Qc3pqpQ
c«J2 0'T3 (J
V4- -OOJ^ .
^_- E c o
Dh C? ^ t/5 -M
=5 > ^ X >.N
lO Ci ^ lO
CO CO t> CO
I>- O-l CO CO
CM to to to oo
t^ I>. iq
o
t^ uo lO CO l>I
CO 00 t^ i-I CO
CO O CO O 00
i-H t^ l> I>I O
to O to T— J o l>^
o
o "* CO CO Oi
CD to Tt^ to 00
CO 00 CO CM ^
l> to CO 1> CO
CM '^ O 1— 1 m OS
r-l (M (M C^ ^
(M CM (M (M (N
CM CM CO CM CM
CM CM CM T-l CM
CM O-l CO CO CM 'th
W
OJGC CO
_o _o &I
o
_o .2 .2
^o _o _o
^:S .W
do d
13
3 S S
-1^ -tj -t-i
'S 'S 'S
>-H ?H ^ Co ?-4
CO 5^^ b
-1 ?J rH O ,H
-3 ^30-3
_-. O O 0_cH
oOOO o
_rJ _c! O r-. O O
o oOoOO
ooooo
M cc 't; ^J _g
t^ !^ 5^ 3 c
£ rfS Ps
3 M M W a
ss is P P
3 Ji :3 o 3
13 O O OJ Ci
3 C! O 3 OJ O)
l2;:2;ooo
pqoPQ^W
pq^^^W
pqW^pQlz^lz;
C^^ (JTS 0)
H- ClD^ .rt-r-s
_i^^ E c
CXCTi- trt-^
3 > 5 X >»N
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES
19
Eh
[a
hH
^
CJJ
>H
H
e!
hH
P
1-J
l-H
C)
PQ
M
S
la
u
y^
Fi
M
ta
H-I
Ph
b
U)
w
«
Q
W
<)
1^
p:;
<ii
W
w
'Z
hH
15
n
U
H
K
C/2
t^
W
;^
^
a
<J
K
^
h-I
<
(^
J
<
eq
^
Ki
a
Eh
Eh
rr
^
H
c
i-q
ffi
(yj
H
m
^
H
>l
p
cc
is
CO
12:
p^
IS
02
W
w
&H
CO t^ C<1 !>. <N
(M C^ CO !>
CO CO 00
CO t^
»0 00 1> !>■ CM !>.
CO CD C5 CD 05
-^ rtn CO T-H
06 CO »o
10 10 06 i-H
t-^ 06 i-H CO cri CO
CO 10 Tt"
CO rH CO t^ 00
(M CO C3i C^ CO
TfH TtH Oi 10
CO 'rt^ CO Tfl i-H
(M (N (M (M (M
(M iM (M <M (M
(M (M (N (M C^
(M (M i-H (N (M
CM CM CO CM 05 CM
C^
(M
(M
<M
CM CM CM
CO d 6
oi d CO
. d ai d
6 6 6
Ison
0. S
Ison
Ison
g No
. y^
^'^ (D
fl '^ P3
^ (D ^
nO. S
ng Mo
ngO.
mne N
Idson
OJ -CD
bjO bC Pi .iS ^
bJO^ P3 ^ bX)
^ Pi ^ fi
^ rt ^^ P3
P3 rt P3^.ii
C '-' P3 ^ p3
2 i=i ^ P3 ^
c3 o3 c3 -q
la ;5 > "^ =^
^1> 1^
CO P5 g CO
rn ^ !n tl
o3 ^ o3 ^r; -^^ »C
pi^ Pi^-o;^
^ c3 9 ^
3 -s CD q
::; CD q 3
i 13 pi 0^ ^
a; jH <i>
C^OP^P^O
OOQP^P^
u^p\^6
OOOPPh
P^QP^QOQ
<!pqoQ[ij
feOE hHH-.
Wh4^^0
O^CP^ CAjfH
p>^X>Htsj
LO (M (M CO CO
iC CO (N CO (M
CO !>; !>;
CO !>; I>; 10 1>
CO l>- !>; CO »>• 1>.
(>; '^^IH as CO 06
i>l CO tH oo T-H
CD l-H 10 CD
CO »-H CD t^ i-H
CO CO i-H 06 i-H
10 10 '*
T— 1 1— 1 1— 1 CD !>•
C^ 00 t^ C^ (M
CO CO (Ti 02 10
CO Th Oi CM "* 1-1
(M (M (M (M <M
(M CM oq <M (N
C^ (M (M (N CM
(M (M rH 1—1 C<l
CM CM CM CM CM CM
iM
(N
^. (N ^.
. . CM
02 02 02.^
C fl
. . gcd d
. .
mm .
. . . jD .
ng Mo
Idson
ng ]\Io
. 0. S.
ng No
MCO ^
gNo
Ison
gNo
0. S
ne N'
. . m
ngO
ngO
ngO
Rob
Idson
00 • SIz; a
^ ^0 '^
biD iJO^ ^ bjO^
p: d G c3 a
Pi jrd Pi rt Pi
c fl fl2 o2
3 ;^^ ^ ^
M c« tn r— 1 g
•3 03 -jZl -M -pi
CO g M CD CO
^^:a:a>
:a 2 :a >
CC g CO M '^
•p 'Pl 03 ^3 03
CO CO CO '-' CO i-l
CO CO CO 02 ,,
=1 :3 =:^
:3 13^ ;3
OPhOOO
03 03 ;3 li 0^
Pi3 P2 03^
OPhOOG
pi P3 03 j3 q
OOOp:?OP^
<:3pqoOW
Pr^OW .
^ J^^O
0. OQi coH
tD>^ ><1>-N
I>. !>. i>. t^ CO
I> 10 00 t^ !>.
»0 I>. CO (M
i>. 05 CO CO >o
10 >0 CO 10 CO to
i-H CD T-H CD CO
r^ CO CD CD
I>^ CD CO Id C5
CO T)H CO CO t^
CM t^ 06 CM CO r^
Ttl C^ LO CO
r-H C3 »— 1 "O CD
1 1 CO t^ ' 1 T— (
CO (M Oi 00 CO
CO CO 00 0-1 CO
C^ T-H (M (M .-H
C^ r-H (M C^ (M
(M (M C^ (M <M
(M (M ^ i-H (M
CM CM CM CM CM CM
<N ...
^ . .^.
^ • <^ •
(M (M
. CM CM . . .
fi C CO
. .m a
.^ Pi . Pi P5
Pi d d Pi Pi
Pi . . Pi Pi Pi
00.
^ bC bX) bC C
_3
>-?- 15?-
1^ <H 1^ 1^ 1^
000000
!2^ g^SO
I Robl
ing M
ing N
ing M
ing M
g:2;^ggg
bC^ bJD iaC bO
bJD Pi Pi be bJD
bC bC bjO bC bC bC
fl ^ C fl fl
CO g tn m 03
c c g c S
^ B B ^ ^
Pi Pi g Pi Pi Pi
^^^^>
^ ^ ^ ^ ^
..C>->^^
^ ^ ^ ^ ^ ^
02 CO M 02 ^,
r-H m M m CO
C» ^,, ^^ CO CO
M CO CO M CO CO
:= :3 3 s
c3 P3 P3 P3^
^ ^ ^ ^ ^
J- OJ (D ^j ^
uQQoo
2 P3 =) Pi Pi ;3
000000
<:fqun w
t^ Wi-H «-?
^^Szo
pL, acsf 02 H
;d>^x!>m
(M_ (M_ CO CO CO
CO c^ 10 (N
!>. CO !>.
(M_ t^ ^ 00
l> 10 CM
'*' oi CO 06 CO
(X) '^ C^' Tj^
1-H 06 T-H 10
oi CD od cr>
CD 10 CM "O '^
CO t^ '^ '-H 02
OO ^ CD
CD CD --H
(M (N CO CO (M
CM CO 00 1-1 CM
C^ '-H <N (M i-c
C<J 1-H (M (M (M
(M (M (M (M (M
(M CM T-i ,-( (M
CM CM CM CM CM CM
. . (M (M .
cq "^ c^ c<i .
. (M (M (M
(N . (N .
CM . . CM '^ CM
fl C . . ;-,
. . • e
Pi .
. ^H . ^
. Pi Pi . d •
k
1;^
00 00
fe fe
b?-
faJD bJO bJO bJD >,
§^ §^^
Ronaldson
Gushing N
Am. Type^
Gushing N
Am. Type''
bC n bO bO bjO
bO bC^ bC bJO
be bO bfi bJO Pi bjD
.s.a.s.SH
q _g c g g
Pi s j=; Pi Pi
CO CO g CO CO
g Pi fl pi.g Pi
^ ^ ^^ .
;5 j>i3ia la
;j:i "^ ^ S >■ s
CC 02 M W rH
CO CO CO CO
CO CO CO CO , CO
;= =5 q 13 G
oooo<l
P3 US ;3 13
oPooo
=3 P3 q j3 pi
OOPhOO
P3 pi ::3 pi 0^ p
ooooPo
< pq QH
U.OX>-.H^
M jSZO
P^Opc; c/)EH
D>^X>.N
t> t^ CO 10
!>. <N l>.
10 CO Ol (M_ I>
»0 CO CM
00 10 t>. iq
T-H CD >0 CO (N
'-H ci CD
t^ 06 c^i ^' CD
I> CO oi
C5 10 10 cm' i-H t^
(M t^ C35 (X. Oi
C3i CD C5 CO (M
t^ ■* 00 CO CO
CJi l-H CiO CO l-H
CO C32 CM 00
C^ 1— 1 1— ( r— 1 1 — 1
-— 1 --H r-H (M (M
1 — 1 C<) 1 — 1 1 — 1 1 — 1
l-H CM l-H 1—1 CM
CM CM i-H i-H CM 1— 1
',^ u u ;h C3
U U U U lA
;^ ;^ ^^ !L^ ;^
. . • . <M
U t-i ^ t^ .
tJ, !^ t^ ^ %^ ^^
^ ^ ^ ^^
^ p: ^ ^ ^
fe: p: ^ ^ ^
<D CD S CD Z
;^ ^ ii '^ ^ ^
CD (D O) CD S
CD <X) CD Qi
(D 03 CD CD
a; CD (D (D Q.) 4)
a CI. &H a
Oh p., Ph Ph &,
Ph Ph P. Cu a
^ i^tl f=^Tl
&, Ph P^ a P, Ph
>.>>>.>.«)
>>>.>.>.>.
>.>>>>>> >^
>, t>,bC>,bX)
>> >-. >^ >-. >i >^
a s a s 1
EhHHHH
HHHHH
HH.an.a
HHHEhHH
a a a a a a
a a a a a
a a a a a
pi P 2 pi CO
5-1 I-H ;3 K ^
<5<:<<<i^o
<ji<ii<:<<^
<<<<<3l<t1<
<<<lo<^o
<1 <t^ <^ <; <j <j
< PQ 0. W
[xi c!3 K M t-a
Wh^SI^O
Ph(3?P^C/2 H
tD>^}><!>HlSl
20
ROETHLEIN
^
H
,
^
U
u
g
M
jyj
(H
Fh
Ph
t-l
&
1-1
O
P5
HH
o
i=(
1— 1
p^
^
o
w
p^
Q
w
<J
Q
>i,
Pi
CJ
H
«
^
(H
Q
Q
hJ
;?;
hJ
n
U
C)
O
w
w
CO
o
M
ffi
^
^
<1
<;
W
Ph
^
ij
<1
pq
CJ
HH
O
W
W
W
H-1
H
F-i
cc
-<
H
O
K
xn
W
w
H
CO
H-)
^
o
o
o
Z
o
P
u
CO
o
w
o
w
i:^
H
(M l>. l> l>.
to t^ (M to
1^ I> !>.
CO CM to CO
CO i>- CO to
05 CO 1-5 CD to
to CM o ^' t^
CO 1—1 1— 1 o o
CO cJ o to o6
O to 00 1-^ 00 CM
CO O 00 rfH CO
CI O 1-H i-H C^l
O C^l Ci 05 00
CM T-i O '^i 00
00 o -^ t^ 1-1 CO
i-l (N t-H (M i-H
CM CM CM CM CM
CM CM CM 1—1 1—1
CM CM i-< r- 1 T-4
r-H CM C^l 1-1 CM T-I
02 aj
CM
m
CM
CM
Idson 0.
ng 0. S,
Robbia
n 0. S.
Idson 0.
nO. S.
n 0. S.
nO. S.
nne No.
ng 0. S,
Robbia
Gothic
nO. S.
nO. S.
Idson 0.
. 0. s.
Robbia
ng No. !
nO. S.
nO. S.
. 0. s.
nne No.
iry Exp,
nO. S.
Rona
Cushi
Delia
Jenso
Rona
Casio
Jenso
Casio
DeVi
Cushi
c3 g^^ %
"^ >T^ m m ^
<a ^ <A <A o
q;z;oop^
Chelt
Chelt
Delia
Cushi
Jenso
Jenso
Chelt
DeVi
Casloi
Cent I
Casio
aJ.O OT) a>
u- b*0j2 — .^
.bi^ £ c o
D< cr* V- w -to
3 > ^ X >J tsj
»0 CO (N M
CO to C-1 CO
CM !>. t>.
00 to CM CO CO
to CM CO t^ CM
i>^ CO oj to ^'
to cc i>l oi 00
O oi CO i-i to
O t^ O CO CO
!>; ci CO o o ai
CO O t^ Tt< CO
CM Oi O O ^
o 1-1 oo 00 i>
CM 1-H 00 Tf< 00
i> o; ■* i> 1— 1 to
r-H (M i-H (M 1-1
CM T-i CM CM CM
C^l CM CM i-H 1-1
C^l CM 1-1 rH T-(
1—1 1—1 CM r^ CM 1—1
. .^ .
«? CM
CO
<M 02 02
.g n c/2 oco
.^O b • •
^ . d^ d
• d
d -OO '-^
Sl'^^a
l"?[3»l
^ floj^M
«^^^^«l
o ^^ o ^^
bC*^ >,0 bJD
d gcjdo
• a; ^ o o Q,
O fi -^^ ^ >;
shin
slon
ntur
elt.
shin
^2 . rt .
— ^ «2 W g CO
Jenso
Rona
Am.'
Casio
Cushi
o 03 +2 c> ;i^
;ii JC o c: C3 "^
«^ Ts "m S S d
CD 13 -3 O 3
QOOPhO
:i 03 S -d 13
ooooo
OPhOOO
^ ^ c3 Q o c
OPOp:;p:?<
«5 ^ o Ti o
M-l fajOrCj .- ._
^ _ S c o
Ol O* W, M .faj
3 > ^ X l>.N
GO 00 (M_ CO to
00 t^
to CO t^ to to
to to CO 00 00
1> CO CM CO to
to id '^ CO c<j
to to lO CO o
I> 00 CO CM CM
l>^ CM CO i-I O
CD to CO Ci 0C> t^
to C5 t^ CO CO
O Ci O O 1-H
C5i O t^ t^ I>
T-I l-H (^ Th 00
i> cn CO CO o to
1— 1 T— 1 1 — 1 (N i-H
CM >-< CM CM CM
1—1 CM CM 1-1 1-H
CM CM 1-1 1-; 1-1
T— 1 T— 1 CM 1— < CM 1— <
.^^1 . .
CM CM
CM . . .
CO <^. .C/2
^ . ^. .
Co . C !->
d d ;-*' •
. . flCCCC
Ison
ne No
g Mon
[son
0. S.
. f^ a c o2
s '^ >; ^ CD
hJ3 d bJD bC >^
neN
0. S.
neN
ypew
0. S
0)00..
^!2;^oo
bjj >, bX)0 G bX)
Cushin
DeVin
Cushin
Cushin
Am. T;
DeVin
Chelt.
DeVin
Am. T;
Jenson
«.s .s.s.s
^ ^ !5 S ^
CO CO CO 02 CO
nald
:Vin
shin
nald
slon
shin
.1. T;
shin
elt.
Vin
shin
c3 ;3 d d S
ooooo
O 0^ ^3 O 03
p:5P0P40
o<:ooPo
cj .D o t::! d)
<^ OOJS-H .^
j^^ S d o
0< O" n C/2 +-I
P > ^ X >> N
00 !>; t^
CM t^ CM !>;
CO 00 CO CO 00
t^ t^ CM
00 00 GO to GO 1>
to O i-H O CO
"^ i-H O Tj^" O
00 cJ CO ■*■ o
to CO i-J "^ ci
to O O t>^ to CD
to Oi t^ CO '^l
00 C5 o o ^
00 C5 t^ CO t^
o o 00 00 t^
t^ (X O to Cj: to
1 — 1 1— ( r— 1 C^ T— 1
1-1 1— 1 CM C-1 CM
T— i 1 — 1 CM »— 1 1— 1
CM CM 1—1 CO 1-1
1—1 1—1 C-i 1—1 1—1 1—1
(M
(N^ M_ .
•^(M .^.
<^^ ■^■'2 •
CM CM .CM
. . c ^ c
ing No.
inne No.
ing Mon
Mson 0,
1 Robbia
^* d .Co
. co.^ fi
CO . . . g .
oco o ^ O
bJo'O bC >" bD
>. fl biD bJO fl
ing No
ing Mo
ildson <
. Robb:
ing Mo
. CO c o o O
bjo'^' bC bJO bJD bC
•S g.SH.S
H.S.S.S.S
c c f^ c c c
rd ^ rd . ^
W CO M rH 03
13 ci ;3 G ;3
CO "^ 03 K —1
S a; g ^ a;
CO -f: e '— < ^
03 03 CO CO CO 02
^ O j3 O O)
3 US q <u s
;; c3 P 3 ^ J3
ooo<ic:)
oPop^Q
-<pu6q
^UP^QZJ
UOOOOO
rt _Q o^d <u
^ b«-E| .-H —
^ — S fl o
G. U t-( t/3 ■(->
;r > ^ X >, N
i>. CO 00
CO to 00 CO
CO to tq
00 !>. 00 CO CM
CO CO CM CO
to i-H a5 to o
00 CM* O O CO
CO t^ to c^i o
<d CO to 00 ci
to CO CO to Tt^' CO
to 1-^ CO O "*
t^ «j o o o
00 C5 Oi CO i>
O C5 t^ CO CO
CO CO X to o -^
1 — 1 1 — 1 1 — 1 O'l 1— 1
1-H r-( 1-1 CM C^l
1 — 1 1 — 1 1 — 1 r— 1 1— 1
CM »— 1 1 — 1 1 — 1 1— 1
T-H 1-H T— t 1— 1 1—1 1-H
(M . <N CM
CM CM .
CM . CM CM
. CM CM CM
• CM . -CM .
• rH • •
d . . . c
. <A i^ . .
C . .
C . ^H g . g
.-^' o ^ o o
O r/2 O O O
O O fe: O O
bC bJO >, bC bJO
O O O ./ o
c o ^ o o o
O bjD >^ bC biD
S .^^S
Slz; |SJz;§
bc'^ bJO bC bC
bC bC bcO bJD
bC bC >^ bC bC b£i
^.SH.S.S
.S rt.S.S s
_g S H .S .S
C C C . C
c c ^ c c g
;i^xj .AA
rC^-d^^
IS ^ . ^ !5
'^'^'rs:i;i={'^
s;h .is:s;j3
<P W r^ to 03
CO 03 CO CO 02
CO 02 c; CO CO
:3 :3 c 13 13
CO O' CO (LI 03
CO 02 C CO CO W
;3 ir; c =3 13 s
^ 3 C r3 ;=
d Oj 3 3 ^S
p q; 3^ 13
oo-<oo
ooooo
oo<:oo
oouoo
oo<ooo
(0 J3 O T3 <U
.« tJDJlS — ••-.
^ ^ S G O
P< a* k. u5 -»-•
J3 > ^ X >» N
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 21
TABLE XI. THE AVERAGE LEGIBILITY OF GROUPED
LETTERS
The Numbers in this Table Indicate the Average Distance
AT Which Each Letter of Each Face Was Read. Nine
Face's Were Selected for this Experiment; and only
Lower Case Letters Were Employed. The Averages are
Compiled from 'Internal' Letters Only, i. e., the Initial
Letters and the Final Letters of the Groups are not
Included in these Averages
a;
o
d
o
-^
o
e
(i,
!Xl
(^^
c
ai
o
c3
X
6
a
o
6
s
c
c^
-2
1-1
CI
a
bJO
fcJD
O
O
m
»— «
"m
c
C
o
m
2
>
o
^
a
QJ
o;
I-*
3
r-"
<a
m
O
o
O
O
O
o
"^
CO
a 140
a
137
a 150
a 145
a 142
a 123
a 139
a 160
a 135
b 138
b
155
b 158
b 155
b 152
b 138
b 146
b 170
b 182
c 147
c
130
c 140
c 143
c 152
c 131
c 142
C 167
c 123
d 142
d
133
d 154
d 150
d 158
d 123
d 140
d 163
d 127
e 124
e
127
e 135
e 145
e 135
e 120
e 138
e 149
e 115
f 160
f
175
f 193
f 178
f 168
f 120
f 150
f 174
f 170
g 185
g
190
g 205
g200
g210
g 195
g 213
g218
g 215
h 163
h
184
h 170
h 160
h 165
h 150
h 172
h 113
h 150
i 143
i
143
i 156
i 168
i 153
i 152
i 167
i 162
i 149
j 140
J
175
j 170
j 180
j 165
3 175
j 187
j 155
j 160
k 183
k
183
k 160
k 185
k 175
k 150
k 195
k 168
k 185
1 136
1
137
1 153
1 162
I 163
1 141
1 147
1 138
I 148
ml48
m
141
ml52
ml44
ml52
ml47
m 165
mlG5
ml48
n 136
n
129
n 140
n 145
n 155
n 150
n 145
n 146
n 127
o 143
o
139
141
155
o 160
128
138
160
o 120
p 158
P
195
p 190
p 168
p 195
p 195
p 220
p 169
p 185
q 150
q
163
q 177
q 190
q 175
q 147
q 173
q 173
q 164
r 146
r
141
r 158
r 165
r 141
r 141
r 157
r 158
r 141
s 165
s
122
s 135
s 152
s 130
s 120
s 133
S 175
s 129
t 136
t
125
t 149
t 158
t 123
t 128
t 173
t 171
t 158
u 153
u
123
u 140
u 145
u 150
u 133
u 153
U 166
u 136
V 153
V
154
V 157
V 153
V 158
V 138
V 153
V 154
V 148
wl88
w
170
wl87
wl95
wl85
w 188
w203
Wl98
wl85
X 166
X
133
X 158
X 161
X 144
X 134
X 158
X 167
X 137
y 153
y
142
y 160
y 162
y 170
y 155
y 195
y 164
y 158
z 163
z
126
z 135
z 149
z 147
z 120
z 130
Z 165
z 125
Aver. 150
149
159
162
159
144
163
166
151
T
2 2 ROETHLEIN
TABLE XII. THE RELATIVE LEGIBILITY OF
GROUPED LETTERS
The Letters in Each Horizontal Line of this Table are
Arranged in Descending Order of Legibility — the
Vertical Column at the Left Containing the Most
Legible Alphabet. The Data from which this Table
Was Compiled did not Include the Initial Letters
nor the Final Letters of the Groups
a
a
a
a
a
a
a
a
a
b
b
b
b
b
b
b
b
b
c
c
c
c
c
c
c
c
c
d
d
d
d
d
d
d
d
d
e
e
e
e
e
e
e
e
e
f
f
f
f
f
f
f
f
f
g
g
g
g
S
g
g
g
g
h
h
h
h
h
h
h
h
h
•
1
i
i
i
i
i
•
I
i
i
•
J
J
j
j
•
J
i
j
j
•
)
k
k
k
k
k
k
k
k
k
1
1
1
1
1
1
1
1
1
m
m
m
m
m
m
ra
m
m
n
n
n
n
n
n
n
n
n
o
o
o
o
P
P
p
P
P
P
P
P
P
q
q
q
q
q
q
q
q
q
r
r
r
r
r
r
r
r
r
s
s
s
s
s
s
s
s
s
t
t
t
t
t
t
t
t
t
u
u
u
u
u
u
u
u
u
V
V
V
V
V
V
V
V
V
w
w
w
w
w
w
w
w
w
X
X
X
X
X
X
X
X
X
y
y
y
y
y
y
y
y
y
z
z
z
- z
z
z
z
z
z
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 27,
present in such a problem as this. The factors of size, of
geometrical figure, and of heaviness or lightness of the lines
which constitute the figure also play a part ; and these two
latter factors are neither constant nor uniform in their opera-
tion, as the reader may infer from an examination of the
faces which are illustrated in Tables I, II, III, V, VI,
VII, VIII, IX, X, XI and XII.
It is true that all of the letters employed in the investi-
gation were printed in ten-point size. But the type-founder's
unit of measurement refers to the size of the block of metal
upon whose base the letter is cast, and not to the size of
the letter itself. It is, of course, true that the size of the block
of metal sets a limit which the size of the letter cannot exceed ;
but the size of the letter may be, and usually is, less than
the size of the block of metal which supports it and which
determines the number of 'points' which shall technically
describe the size of the letter. Indeed, it is not unusual to
cast, say, an eight-point letter upon a ten-point body ; and
a much greater difference between size of letter and size
of body is possible, although greater divergences are neither
customary nor commendable.
Hence, the fact that all of our letters are technically
described as being of ten-point size does not guarantee that
they shall all be of uniform width, nor even of uniform
height. There are certain letters, the w's and the m's, which
must, in the very nature of the case, be wider than others,
the i's and the I's ; and even the widths of the m and the w
may vary, from face to face, if only the draughtsmen who
design them choose to have it so. Nor are the heights of ten-
point letters necessarily constant and uniform. Moreover, the
lines which constitute the letter- form are themselves subject to
variation ; they may be faint, they may be heavy, they may
be bold, they may contain both hair-lines and heavy lines,
or they may contain lines of only one width or thickness
throughout. Hence, any discussion of our results must con-
sider the possible operation of three factors which may have
to do with the legibility of letters : letter-form, size of letter,
width of lines of letter. And when we come to the dis-
24 ROETHLEIN
cussion of the results obtained from groups of letters, we
shall see that yet a fourth factor — the extent of white margin
around the letter — plays an important role,
A survey of Tables I, II and III shows that, as a matter
of fact, all three variables are represented in the letters
which were employed. But if, for the moment, we ignore
these details, and seek to determine which of these sixteen
faces, as actually employed in the art of printing, is the
most readily legible when presented singly, we find an ex-
ceedingly interesting state of affairs.
Confining our attention first to the upper case letters, we
discover that the Jenson face has a considerable lead over
all of its competitors. The letters of the Jenson face were
read, on the average, at a distance of approximately 282 cm.,
while the average for the sixteen faces is ca. 254 cm. The
least legible face is American Typev/riter, — which, indeed,
is much less legible than any other face of the sixteen. The
other fourteen faces fall into three groups. The most legible
group contains (besides Jenson) Bulfinch, Century Oldstyle,
Clearface, Cheltenham Oldstyle, Delia Robbia, Century Ex-
panded, News Gothic; Caslon Oldstyle, Cushing Oldstyle,
DeVinne No. 2 and Ronaldson constitute the second group ;
while Cushing Monotone and Cushing No. 2 come next in
order, with American Typewriter standing in a class by itself,
as the least legible face.
Jenson was much less successful in designing a legible
lower case face ; his lower case characters are indeed scarcely
more legible than the median of the sixteen faces. News
Gothic proved to be the most legible of the lower case faces ;
and Cushing No. 2 is the least legible face of this series.
Here the sixteen faces fall naturally into three groups :
I. News Gothic, Bulfinch, Clearface, Century Oldstyle, Cen-
tury Expanded and Cheltenham Wide; 2. Jenson, Delia
Robbia, Cushing Oldstyle, Ronaldson Oldstyle, Cheltenham
Oldstyle, DeVinne No. 2, American Typewriter and Caslon
Oldstyle; 5. Cushing Monotone and Cushing No. 2. The
difference between the least legible face and the most legible
face is much less in the lower case letters than in the upper
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 25
case letters, — twenty-one per cent, as compared with thirty
per cent.
The briefest examination of the letters which were em-
ployed in our investigation is sufficient to show that they
differ not only in their form, but also in their size, and in
the thickness or heaviness of the lines which constitute them.
In order to obtain a clearer insight into the relative signifi-
cance of each of these variable factors as determinants of
legibility, we made accurate measurements of the height and
the width of certain lower case letters of each face, as well
as of the breadth of the lines which go to make up the
letters. Our measurements were made by means of a micro-
scope which was equipped with a micrometer scale. Three
letters, m, o and z, were selected for measurement ; and it is as-
sumed that the average size of these three letters, chosen from
any face, may be regarded as being typical or representative
of the relative size of all of the letters of the complete
alphabet of the face to which they belong. In order to
make it possible to institute a comparison between the size
and the "blackness" or "heaviness" of any face of type,
on the one hand, and its degree of legibility, on the other,
the following table is appended. The first column contains the
list of sixteen faces, arranged in order of legibility. The num-
bers in the second column indicate the average width, ex-
pressed in microns, of the m, o and z of the face whose name
appears upon the same line in the first column ; and the
third column specifies the average height of the same three
letters. The numbers in the fourth column indicate the
thickness of the stem of the i of each face, also expressed
in microns. Each number in the fifth column is obtained by
multiplying the height of the letter by its width ; each of
these numbers may be assumed to represent the coefficient
of area of a typical letter of each face. The numbers in
the sixth column are obtained by multiplying the number
which represents the area-coefficient by the number
which represents the breadth of the principal lines which
constitute the letter: these numbers in the sixth column
may, therefore, be regarded as area-breadth coefficients which
ROETHLEIN
TABLE XIII. ISOLATED LETTERS
Showing the Width and the Height of Letter, and the
Thickness of Line, Employed in the Various Faces,
and their relation to legibility.
1. News Gothic . .
2. Bulfinch
3. Clearface ....
4. Century O. S..
5. Century Exp. .
6. Cheltenham W.
7. Jensen O. S. . .
8. Delia Robbia. . .
9. Gushing O. S..
10. Ronaldson O. S
Ghelt. O. S..
II.
12. DeVinne No. 2.
13. Am. Typwr. . .
14. Gaslon O. S.. .
15. Gushing Mon..
16. Gushing No. 2.
Wide
High
Thick
hXw
hXv.-Xt
Aver-
age dis-
tance
1933
1982
333
3831
1277
236
2033
1800
328
3659
1200
234
1840
1750
333
3220
1073
230
1944
1780
300
3460
1038
228
1813
1766
300
3202
961
227
1833
1420
317
2603
823
224
1812
1560
3^7
2827
896
215
1757
1602
275
2815
774
214
1553
1700
333
2640
880
213
1613
I481
243
2389
580
209
1523
1440
316
2193
693
206
1853
I5IO
450
2798
1259
205
1544
1820
2(X)
2810
562
202
161I
1420
292
2288
668
202
I47I
1464
220
2154
474
191
1493
1480
200
2210
442
186
represent the relative size of letter and breadth of line which
appear in the several faces.® Now an examination of the
numbers in the fifth column shows that there is, in the upper
part of the column at least, an approximate correlation be-
tween size of letter and degree of legibility. And when
"heaviness" of line is also taken into consideration, as is
done in the sixth column, one finds that the correlation is
even more striking, for the first eight faces of the list. The
fact that degree of legibility is correlated, at least in this
approximate fashion, with size of letter and breadth of line
' The numbers in the fifth and the sixth columns are intended to
represent relative magnitudes only. In order to facilitate comparison
we have omitted several of the right-hand digits, — three from the
numbers of the fifth column, six from the numbers of the sixth
column.
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 27
of letter, leads one to suspect that, in these eight faces at least,
the size of the letters and the amount of ink imprinted upon
the paper by the lines of the letter were the chief determinants
of legibility ; and that such variants of letter-form as were
represented in these eight faces played no essential part in
determining the legibility of the letters.
The less legible faces of the list do not show such a perfect
correlation. The lack of correlation is most evident in the
case of DeVinne No. 2, which apparently should stand near
the head of the list if size of letter and breadth of inked
lines vvere the chief determinants of legibility. But it will
be noticed that this face contains by far the heaviest lines
of the series ; and its relative illegibility is probably due to
the fact that the optimal width of line is here exceeded,
and that, in consequence, the white spaces within the letters
have been encroached upon in so great degree as to pro-
mote illegibility. It would appear that the optimal breadth
of line for lower case letters of this size may be in the
neighborhood of 275 to 333 microns. It seems to be evident
from our table that, when the breadth is reduced much below
250, or is increased in the neighborhood of 450, ten-point
letters become relatively illegible.
It is scarcely possible, from the data at hand, to make any
definite general statement regarding the relative significance of
the three factors in question. Javal and others have insisted
that the width of the letter is of prime significance as a deter-
minant of legibility. And the fact that Cheltenham Wide,
lower case, is read at an average distance of two hundred
and twenty-four centimeters, while Cheltenham Oldstyle,
which differs from the former scarcely at all save in its
lesser width, is read only at two hundred and six centimeters,
seems to confirm Javal's statement. But our experiments
have convinced us that width of letter is but one of several
factors which contribute to legibility ; and one could assert
with quite as much justice that the breadth of the lines which
constitute the letter is the essential determinant of legibility.
This latter statement, however, is true, if at all, only within
relatively narrow limits.
28 ROETHLEIN
The results which appear in Tables V and VI are of interest
in this connection. Our experimental material here consisted
of alphabets which represented a series of modifications of
a given face. Cheltenham Oldstyle may be regarded as our
standard face; Cheltenham Bold, Cheltenham Bold-Con-
densed, Cheltenham Wide and Cheltenham Italic represent
modifications of our standard letter-form in the directions
of greater heaviness of face, lesser width of letter-form,
greater width of letter-form, and inclined or italic letter-
form.
A survey of these results reveals the fact that legibility
is very much increased by increased heaviness of face. In
the upper case letters the increase in heaviness or breadth
of line amounted to approximately forty-seven per cent.,
and the increase in legibility amounted to twelve per cent. ;
while in the lower case letters the corresponding data are
forty-seven and fourteen per cent. When the Cheltenham
Bold face is condensed by about twenty-three per cent., as
is illustrated in the letters which appear in the third column
of Table VI, its increase of legibility is lost, — being re-
duced by twelve per cent. This finding indicates that what-
ever advantage might have been derivable from increased
heaviness of face, as compared wnth Cheltenham Oldstyle,
is neutralized by a disadvantage which is due to a narrowing
of the internal spaces within the letters, and a consequent
sacrifice of detail. When the letter-form is broadened by
eighteen per cent., its legibility is increased by nine per cent.
The Italic letter-form proves to be but slightly less legible
than the Roman face. In both the upper case and the lower
case letters, the bold face is the most legible member of the
series of modifications.
The data which are presented in Table IV show the rela-
tive legibility of each letter of the alphabet, both upper case
and lower case, when presented as isolated letters. The
numbers which appear in these columns represent the aver-
ages of the readings of the sixteen faces of each letter.
The upper case letters are, of course, legible at a greater aver-
age distance from the observer than the lower case letters. But
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 29
the average legibility of both cases varies between wide
limits; and it is a significant fact that certain of the latter
letters are more legible than many of the former. The
progressive decrease of legibility from the upper limit to the
lower limit is fairly regular and uniform throughout, in both
cases, with the exception of m and s. Indeed, our results
show that the lower case s is by far the least legible letter
of either alphabet. The upper case S also stands at the
foot of its class, but the relative inferiority is here much
less than in the lower case letter.
Any movement which plans to improve the forms of the
letters of the alphabet must properly begin with the letters
which appear in the lower half of this table. And, indeed, it
might well confine itself, at the outset at least, to the lower
case letters, for the twofold reason that they are, in general,
less legible than upper case letters and are therefore most
in need of reformation, and that our reading has to do, in
the main, with the recognition of words which are com-
posed of lower case letters exclusively. It is not the pur-
pose of this paper to recommend or even to suggest the pro-
cedure by means of which an increased legibility of printing
types is to be attained. But it may, at least, be mentioned
in passing that letters are made more legible by an increased
heaviness of printing-face; and that defective letters may
be made more legible by simply extending their width. Both
of these innovations would meet with opposition, however,
from those readers who demand aesthetic beauty and grace
of form as well as legibility. But it seems possible that all
demands could be met and all interests could be safe-guarded
if only sufficient skill and ingenuity were devoted to the task.
B. GROUPED LETTERS
The object of this part of the investigation was to de-
termine how and to what extent the legibility of letters is
affected by the presence of adjacent letters. It was to be
expected that the isolated letter should possess an advantage
over the member of the letter-group in point of average dis-
tance at which reading is possible. Is the disadvantage which
30 ROETHLEIN
results from grouping equally great for each face? What
is the relation between legibility and position (initial, final,
intermediate) within the group of letters? What light is
thrown upon our problem by the number and the nature of
the misreadings or confusions which are due to the grouping
of letters?
Isolated letters are invariably read at a greater average
distance than are those letters which occur in groups. Among
our material were eight faces which had been employed in
both of our series of experiments, i. e., they had been presented
singly in the earlier experiments, — grouped in the later ex-
periments. It is therefore possible to make a direct com-
parison between the degrees of legibility which were revealed
under these two conditions. (The Scotch Roman face can
not be included in making this comparison, because it was
employed only in the later series of experiments.) The aver-
age distance at which these eight faces were read in the
earlier experiments was 232 cm. ; in the later experiments,
157 cm. The varying degrees of legibility tend to be re-
duced to a common level as a result of grouping, i. e., differ-
ences in the legibility of different faces tend now to disappear.
While the legibility of the eight faces, presented as single
letters, varied between the limits 191 cm. and 2^6 cm., they
varied only between the limits 144 cm. and 166 cm. when pre-
sented in groups. And the general reduction of legibility
was so great that the most legible face of grouped letters
proved to be wholly undecipherable at a distance where the
least legible face of isolated letters was clearly and unmis-
takably legible.
The gross results of the experiments with grouped letters
are presented in Table XL The average distances are tab-
ulated in more significant form in Table XIII, which also
contains a statement of the relation between decreased legi-
bility and size-heaviness of face.
News Gothic is still the most legible face, and Gushing
Monotone the least legible. Gentury Oldstyle, Gentury Ex-
panded, Gheltenham Wide and Gaslon also maintain the same
positions in the series which they held when isolated letters
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 3 I
were employed. But Gushing Oldstyle, which occupied the
sixth position in the former series, now moves up to the
second position ; and Bulfinch drops from the second to the
sixth position. Scotch Roman, the face which had not been
tested in the first part of the investigation, stands seventh
in descending order of legibihty.
Gushing Oldstyle and Bulfinch are the only faces which
show any decided loss or gain in relative legibility, when
submitted to the group-test. But one is led to inquire why
even these two faces should fail to maintain the same degree
of relative legibility when presented in groups which they
established when presented as isolated letters.
A comparison of the amounts of reduction in legibility
which the various faces suffered from being grouped reveals
the fact that these amounts vary from twenty-three per cent,
to thirty-seven per cent. It is inevitable that certain faces
should be subject to a relatively great disadvantage as a
result of being combined into groups, and from being printed
"solid," as upon the body of an unleaded page. The letters
of certain faces are so wide and so high that a comparatively
narrow margin is left upon the block of metal which con-
stitutes the body of the type, while in other faces the margin
is relatively wide. For example, our measurements show
that the Bulfinch m is 2560 microns in width and 1828
microns in height, while the corresponding measurements of
the Gheltenham Oldstyle m are 1948 and 1440 microns. Vari-
ations in width of letter need not be considered here, because
there need be no corresponding variations of width of blank
space between adjoining letters upon the printed page. But
the difference in height between the letters of these two
faces is so great that adjacent lines of the two, if set "solid,"
would show an intervening space of white paper which would
be 380 microns wider in the case of Gheltenham than in
the case of Bulfinch. It follows then that while the ten-
point letters of greatest average height are most legible when
presented singly, — because they subtend a larger visual angle,
— they tend to lose their advantage when combined into
groups and lines, — because the lines are separated from one
32 ROETHLEIN
another by narrower intervening spaces of white; and the
reader finds that the letters of adjoining Hnes tend to merge
into one another.
The effect of breadth of interlinear space could be shown
quantitatively by correlating the average height of each face
with the relative amount of decrease in legibility which it suf-
fers as a result of grouping. It would then be found that
those faces whose average height is greatest have suft'ered
most from being presented in groups, while those faces whose
average height is least have suffered least. The correlation
becomes more striking however when thickness of line as well
as height of letter is taken into account. But even here, News
Gothic and Gushing Old Style appear to be exceptions to
the general rule.
Now it is a remarkable fact that these two types are by
far the heaviest faces of the series. In selecting an arbitrary
method of giving quantitative expression to the combined
influence of the factors of size of letter and of heaviness of
line, in order that seriation and comparison might thereby
be facilitated, we have chosen to multiply the height of the
letter by the thickness of the line. It is evident, however,
that these two factors are wholly different in their mode
of operation and in their relation to legibility. The legibility
of a face undoubtedly increases directly with the increase
in the size of the face; and the ratio between the two is
capable of simple quantitative expression. But the relation
between legibility and heaviness of face is far from being
a simple one. Beginning with the faintest monotone, legibility
increases with increase of thickness of line until an optimal
thickness of line, for a given size of letter, is reached ; from
that point onward legibility decreases with increase of thick-
ness of line, because now the internal details of a, c, e, o, etc.,
are being narrowed down to mere points, and discrimination
of detail soon ceases to be possible. We are not in possession
of any data which furnish a basis for estimating the point
at which the optimal thickness of line is reached, nor for esti-
mating the relative significance of size of letter and thickness
of line. And until such data are obtained it will remain
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 33
impossible to seriate the numbers in the fifth column in any
just or appropriate fashion.
Our results, however, indicate that the variations of letter-
form which were present in our nine faces did not vary-
between sufficiently wide limits to have any considerable
effect upon the relative legibility of our letters. Such differ-
ences of legibility as were revealed are, we believe, refer-
able solely to variations in size of letter (and of interlinear
space), and to variations in heaviness of face.
If legibility is to be our sole criterion of excellence of
type-face, 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 aesthetic
factor must always be taken into account, however, here
as elsewhere. And the reader who prefers the appearance
of Gushing Oldstyle or a Gentury face may gratify his
aesthetic demands without any considerable sacrifice of
legibility.
An examination of our records of the misreadings of letters
throws an interesting light upon the question of similarities
of letter-form, and upon the question of the criteria employed
by the reader in deciphering letters. When the letters first
appeared at or beyond the extreme limit of reading distance,
the reader was usually conscious, from the outset, of a more
or less definite Einstelhing. If they were lower case letters,
and especially if they were presented in isolated form, the
reader first set out upon a more or less definite search for
the "broad" letters, — m and w. He seldom failed to discover
these tvv'o letters in the earlier stages of the settings of the
carriage, although he was seldom able to discriminate be-
tween m and w until the carriage had been brought nearer to
his eye. His next endeavor was to find the "narrow" letters,
— i, 1, j ; another category was " letters which are wider at
the top than at the bottom," — v, r; V, Y, T, F, P. Here
again he almost invariably succeeded in referring the letter-
form to its appropriate category some little time before he
was able to distinguish its details and name it with any de-
gree of certainty. Other general groups were "circular
34 ROETHLEIN
letters,"— o, e, c; O, C, G, Q; "square letters,"— H, K, E,
B, D. Among the lower case letters, "ascenders" and "de-
scenders" were almost invariably recognized as such, before
their details became clear. While the reader was still
struggling with one or other of these categories, the "un-
classified" letter-form began to emerge in an order which
can not well be formulated in a general statement.
This description of the reader's procedure indicates the
type of confusion or misreading which proved to be most
frequent. Letters which may be regarded as members of a
common category were especially likely to be confused with
one another; r, v; o, c, e; x, z; u, n; b, h, k; q, y; i, 1,
j, t, f; M, W; H, K, E, B, D; O, Q, C, G; V, YET;
I J, L.
Several wholly new types of misreadings made their ap-
pearance when the letters were presented in groups. These
may be described as i. Combinations, 2. Separations, and
3. Elisions, i. It frequently happened that adjoining letters
were blended together by the reader, and that a single letter
was constituted by this combination of parts, or wholes, of
different letters, thus: Ic (k) ; Is (k) ; Ix (k) ; li (h) ;
cl (d) ; cf (d) ; un (m) ; in (m) ; vr (w) ; js (p) ; vj (y) ;
cj (q); hj (ly); bj (Iq) ; chn (dm); ck (dx). 2. The
separation of letters into their parts has been illustrated in
the above group. It also ocurred in such cases as wm (vun) ;
Id (bl) ; bj (Iq) ; hj (ly) ; kd (hcl). 3. Perhaps the most
striking misreading was due to the complete elision of cer-
tain letters of the group. This phenomenon was especially
common with such slender or narrow letters as i, j, t, r, v, s.
When i or t was interpolated between ascending letters, when
j occurred betv/een descenders, and when r and s were present
in any context of whatever sort, they were likety to be wholly
overlooked. And it frequently happened in such cases that
the observer read and reread the group of letters, specifying
correctly every letter which was present excepting these
narrow letters. VvHien the carriage had been brought so close
to his eye that he finally detected their presence, he usually
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 35
expressed surprise that these "indifferent" letters had escaped
detection so long.
The position of the letter in the group — initial, final, in-
termediate — is an important factor in determining the legi-
bility of the letter. When a group of letters first comes into
view, it almost invariably happens that certain of the letters
stand out more prominently than their neighbors. The part
which these "dominating" letters play in our recognition of
familiar words has been discussed at length by Zeitler, Mess-
mer, and others, who have pointed out that the role of the
"indifferent" letters in our recognition of familiar words is
relatively insignificant. Pillsbury and others have shown that,
all other conditions being equal, the initial letter or letters
in the word are of most significance in the act of reading.
In compiling the results obtained from the reading of our
letter-groups, we have made three distinct tables of data, —
for the initial letters, for the final letters and for the inter-
mediate letters of the groups. The latter results have already
been presented, in Tables XI and XII. A comparison of
these three series of average distances shows that the initial
position constitutes an optimal condition of legibility ; the
final position comes next in order, and the intermediate posi-
tion is least favorable, — the general averages being 196 cm.,
185 cm., and 156 c, respectively.
It must, of course, be borne in mind that all intermediate
positions are not equally disadvantageous, because the form of
the adjacent letters is really an essential factor which aids
or hinders legibility. For instance, an m which stands be-
tween u and n is much less legible than an m which stands
between o and c; and the helpful or harmful influence of
ascending and descending letters is also very great. The
significance of the initial position or the final position is also
subject to variation. It is evident, for instance, that those
letters whose "loop" is upon the right, — b, h, p, — will derive
more benefit from appearing in the final position than will
those letters whose "loop" is upon the left, — d, q, Yet the
superior advantage of the initial position is so great that
36 ROETHLEIN
b, h, and p frequently proved to be read at a greater dis-
tance when they appeared as initials than when they appeared
as final letters, notwithstanding the fact that the character-
istic parts of these letters appeared in close proximity to a
neighbor in the former case, while in the latter case they
jutted out into the blank spaces between the groups.
Lower case c was read at the following average distances :
Initial 155 cm., final 163 cm., intermediate 144 cm. The
corresponding data for several other letters are: e, 140, 147,
123; o, 158, 153, 143; b, 187, 198, 155; d, 200, 165, 143;
h, 185, 200, 162; k, 195, 198, 176; p, 201, 207, 186; q, 209,
194, 168; f, 196, 202, 176. The "symmetrical" letters show,
in more accurate quantitative form, the relative advantages
of the initial and the final positions; v, 185, 172, 152; o, 158,
153, 143; m. 163, 156, 151; w, 210, 199, 189.^
The Relation Between Legibility and Quality of
Paper-Surface
As already indicated, our groups of letters were printed
in duplicate upon a coated book-paper, and upon an antique
laid book-paper. A comparison of the average distances at
which the letters were read under these two conditions should
throw light upon the general question as to the dependence
^ The reader is again warned that these data do not warrant a
comparison as to the relative legibility of different letters of the
alphabet. The writer does not guarantee that the juxtapositions of
letters which occurred in our groups were equally disadvantageous
for all of the letters of the groups. It is obviously an exceedingly
difficult task to so arrange a series of groupings that all of the letters
in the groups will be treated with even-handed justice in so far as
advantageous and disadvantageous collocation with their neighbors is
concerned. A juxtaposition which is relatively advantageous to 1 or h
may be relatively disadvantageous to o or v. This condition we could
not hope to fulfil ; our aim was rather to treat all of the faces with
equal fairness, — to present no combination of letters of one face
which did not recur in the letters of every other face; our data there-
fore warrant a comparison of the relative legibilitj^ of the various
faces, but not of the relative legibility of the various individual letters
of the same or different faces.
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 37
of legibility upon the quality and texture of paper employed
by the printer.
This question has been raised in various forms, in the
history of the art of printing. Many years ago Babbage
recommended the use of slightly yellowish paper for the
manufacture of books ; and indeed, he succeeded in per-
suading a publisher to produce a book of logarithms in
accordance with this recommendation. More recently Javal
has advocated a similar innovation, urging that such a plan
would minimize the disadvantage to which the reader is sub-
jected on account of the intensive contrast between the black
ink and the white paper which are in general use.
It is impossible, from data available in the literature of this
and cognate topics, to determine whether this Babbage-Javal
suggestion is really of value. But such results of the present
investigation as have a bearing upon the question would seem
to indicate that little or no improvement of legibility is to
be expected from progress in this direction.
The results which we obtained in those experiments which
consisted in presenting groups of letters of nine faces which
had been printed upon coated white paper and rough-finished,
very slightly yellowish paper, show a surprisingly slight differ-
ence of legibility in the two cases. The average distance at
which the letters were read in the former case was 144.9 cm.,
and in the latter case, 145.0 cm. Altogether 234 letters were
employed in these experiments, and 2,808 readings were taken.
The experimental conditions were identical in the two cases,
excepting the difference in the quality, the color and the
texture of the paper upon which the letters were printed.
Occasionally we found individual differences among our ob-
servers, — a greater efificiency when a particular face of type
appeared upon the one or other of the papers ; but in not a
single instance does the individual difference in legibility
amount to more than three per cent. ; and in no instance did we
find a unanimous preference of either paper with any face of
type. Not only then are the individual variations so slight
as to be negligible, but they are so irregular and inconstant
as to be subsumable under no general principle. And the
38 ROETHLEIN
only conclusion which they warrant is that even such a con-
siderable difference as was represented in our two papers
has little or no significance for legibility.
V. CONCLUSIONS
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 : i. the form of the
letter ; 2. the size of the letter ; j. 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 sur-
rounds the letter ; 5. the position of the letter in the let-
ter-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, i.e., in the t3^pe-
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 Gushing Mono-
tone.
5. The initial position in a group of letters is the most advan-
tageous 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.
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 39
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.
VI. BIBLIOGRAPHY
1. Babbage, C. Table of the Logarithms of the Natural Numbers.
(Introd.) London, 1838.
2. Baxt, N.. Uber die Zeit welche notig ist, damit ein Gesichtsein-
druck zum Bewusstsein kommt. Arch. f. d. ges. Physiol.
(PMger), IV, 1871, 325-336.
3. Beer, M. Die Abhangigkeit der Lesezeit von psychologischen
und sprachlichen Faktoren. Zeitschr. f. Psychol., LVI, 1910,
264-298.
4. Becher, E. Experimentelle und kritische Beitrage zur Psycho-
logie des Lesens bei kurzen Expositionszeiten. Zeitschr. f.
Psychol, XXXVI, 1904, 19-74.
5. Brandi, K. Unsere Schrift; Drei Abhandlungen zur Einfiihrung
in die Geschichte der Schrift und des Buchdrucks. Gottingen,
1911, 80 pp.
6. Cattell, J. McK. The Inertia of the Eye and Brain. Brain,
VIII, 1886, 295-313.
7. Cattell, J. McK. Ueber die Zeit der Erkennung und Benennung
von Schriftzeichen, Bildern und Farben. Phil. Stud., II, 1885,
634-650.
8. . Psychometrische Untersuchungen. Phil. Stud., Ill, 1886,
305-335; 452-493-
9. CoHN, H. Hygiene of the Eye in Schools. London, 1886. 236
pp.
10. Dearborn, W. F. The Psychology of Reading. N. Y., 1906,
134 PP-
11. deVinne, T. L. The Practice of Typography: Plain Printing
Types. N. Y., 2nd ed., 1902, 403 pp.
40
ROETHLEIN
12.
13-
14-
15-
i6.
17-
i8.
19.
20.
21.
22.
24.
25-
26.
27-
28.
29.
30.
31.
32.
DocKERAY, F. C. The Span of Vision in Reading, and the Legi-
bihty of Letters. Jour, of Educat. Psychol., I, 1910, 123-131.
Dodge, R. Five Types of Eye-movements in the Horizontal Mer-
idian Plane of the Field of Regard. Amer. Jour, of Physiol.,
VIII, 1902-3, 307-329-
. The Illusion of Clear Vision during Eye Movement.
Psychol. Bull, II. igos, 193-199.
. Eine experimentelle Studie der visuellen Fixation.
Zeitschr. f. Psychol., LII, 1909, 321-424.
Erdmann, B. and Dodge, R. Psychologische Untersuchungen
iiber das Lesen. Flalle, 1898, 360 pp.
ExNER, S. Uber die zu einer Gesichtsv^rahrnehmung notige Zeit.
Sitsungsber. d. Acad. d. Wisscnsch. in Wien, LVIII, Abth. II,
1868, 601-632.
FiNZi, J. Zur Untersuchung der Auffassungs- und Merkfahigkeit.
Psychol. Arheiten (Kraepelin's), III, 1900, 289-384.
GoLDSCHEiDER, A. und MuLLER, E. Zur Psychologic und Patho-
logic d. Lesens. Zeitschr. f. klin. Med., XXIII., 1893, 131-167.
Gress, G. E. The Art and Practice of Typography. N. Y., 1910.
160 pp.
Griffing, H. and Franz, S. I. On the Conditions of Fatigue in
Reading. Psychol. Rev., Ill, 1896, 513-520.
Huey, E. B. Preliminary Experiments in the Physiology and
Psychology of Reading. Amer. Jour. Psychol, IX, 1898, 576-586.
Huey, E. B. On the Psychology and Physiology of Reading.
Amer. Jour. Psychol, XI, 1900, 283-303; XII, 1901, 292-313.
Huey, E. B. The Psychology and Pedagogy of Reading. N. Y.,
1908. 469 pp.
Hylan, J. P. The Distribution of Attention. Psychol. Rev., X,
1903, 373-403; 498-533-
Javal, E. L'evolution de la typographic. Revue scientiUque,
XXVI, 1881, 802-814.
Javal, E. Physiologic de la lecture et de I'ecriture. 2me edition.
Paris, 1906. 296 pp.
Kirschmann, a. Ubcr die Erkennbarkeit geometrischer Figuren
und Schriftzcichen im indirekten Sehen. Arch. f. d. ges.
Psychol, XIII, 1908, 352-389.
Koopman, H. L. Scientific Tests of Types. The Printing Art,
XIII, 1909, 81-83.
Koopman, H. L. Types and Eyes. The Printing Art, XII, 1909,
359-361.
Maire, a. La technique du livre. Paris, 1908. 388 pp.
Messmer, O. Zur Psychologic des Lesens bei Kindern und
Erwachsenen. Arch. f. d. ges. Psychol, II, 1903, 190-298.
LEGIBILITY OF DIFFERENT FACES OF PRINTING TYPES 41
33. PiLLSBURY, W. B. A Study of Apperception. Amer. Jour.
Psychol, VIII, 1897, 315-393.
34. QuANTz, J. O. Problems in the Psychology of Reading. Psychol.
Rev. Mon., II, 1897, 1-51.
35. Sanford, E. C. The Relative Legibility of the Small Letters.
Amer. Jour. Psychol., I, 1888, 402-435.
36. ScHULTZ, A. J. Untersuchungen iiber die Wirkung gleicher
Reize auf die Erfassung bei momentaner Exposition. Zeitschr.
f. Psychol, LII, 1909, 110-148; 238-296.
37. Schumann, F. Psychologic des Lesens. Bericht iiber den II.
Kongress f. exp. Psychol. {Wursburg, 1906). Leipzig, 1907,
153-186.
38. ScHWENDER, J. Die wichtigsten Ergebnisse der experimentellen
Untersuchungen iiber das Lesen. Zeitschr. f. exp. Pddagogik,
IX, 1909, 169-224.
39. Shaw^ E. R. School Hygiene. N. Y., 1901, 255 pp.
40. Zeitler, J. Tachistoskopische Untersuchungen iiber das Lesen.
Phil Stud., XVI, 1900, 380-465.
1
14 DAY USE
RETURN TO DESK FROM WHICH BORROWED
This book is due on the last date stamped below, or
on the date to which renewed.
Renewed books are subject to immediate recall.
viuH? y-j^^'
j
General Library
LD 21-50m-8,'57 University of California
(,C8481sl0)476 Berkeley
Stockton, Calif.
T.M. Reg. U.S. Pat. Off.
z
1-1^ '
^J--J
THE UNIVERSITY OF CALIFORNIA LIBRARY