•INSTITUT imtnnRiiuimu
I
D ' O . S- T |No
CM
•:•

H$

INTERNATIONAL LABOUR OFFICE
STUDIES AND REPORTS
Séries J (Education) No. 3

MOTION STUDY
IN

TYPEWRITING
A RECORD OF EXPERIMENTS
BY

J. M. LAHY
Head of the Laboratory of Experimental Psychology,
Ecole des Hautes Etudes, Paris

GENEVA
1924

PREFATORY NOTE

The report on ability in typewriting published by the International Labour Office in October 1923 gave a practical example
of the application of experimental methods to vocational
guidance. The author, Mademoiselle Dora Bieneman, gave
first a brief analysis of the work of typewriting, and then
endeavoured to arrive at methods of detecting the abilities
required. Her report both emphasised the complexity of the
problem and described the mechanism of tests.
The present report is of quite a different character. The author
set out to discover by scientific investigation the best methods
for typewriting. Hitherto the technique of typewriting has been
mainly a matter of rule of thumb. Mr. Lahy, on the other hand,
in a series of experiments conducted from 1912 to 1914, and by
control experiments since the war, made a detailed investigation
of the mechanism of touch, and he hopes to arrive at genuinely
scientific rules by graphic motion study. As a result of his
experiments he draws many conclusions relating to teaching
which should be of value, not only in the selection and training
of typists, but also in constructing a scientific keyboard.
The value of the investigations, however, does not end with
these practical applications. .Mr. Lahy does not pretend to give
a final answer to every prol k m which arises. His object is
rather, as he says himself, to m ike known " a method of studying
occupational problems". It is indeed a problem of methods of
organisation which is involved, going beyond the boundaries of
the occupation of typewriting and applying to almost every
form of human work. In the words of the author, " co-operation
1
Ability in Typewriting
in Relation to Vocational Guidance, by
Mademoiselle Dora B I E N E M \ N . S¡ . I tics and Reporls, Series J (Education), No. 2. Geneva, 1923, 47 pp. s.; 25 cents.

— 4 —

between manufacturers and scientists should become the basic
principle of social progress". Mr. Lahy's work has thus a very
wide application, since it gives a concrete illustration of the
manner in which psycho-physiology can be of service to industry. This, as much as the strict scientific accuracy of the
experiments described, is the reason for which it is now
published.

-

5 —

CONTENTS
Page

INTRODUCTION.
I.

Technical Progress and Scientific Method . . .

M E T H O D OF R E C O R D I N G

Pneumatic Recording
Electric Recording
Electric Recording
Electric Recording
Electric Recording
Combined Methods of
II.
III.
IV.

V.

by Spring Contacts
by Mercury Contacts
by Contact with the Escapement . .
Recording

7
11

11
13
14
15
16
16

CHOICE OF SUBJECTS AND T E S T

19

ANALYSIS OF THE GRAPHS

21

RESULTS

22

Period of Depression of Keys
Duration of Intervals. Necessity of Alternating the Hands
Confirmation of Rule of Alternation.
Non-Alternating
Phrases and Isolated Letters
Sentence Typed with the Right H a n d
Sentence Typed with the Left Hand
Typing Two Isolated Letters
Typing Short Isolated Words
Relation between Period of Contacts, Period of Intervals,
and Force of Stroke
Physiological Reasons for Alternation of the Hands . . . .
Limits to the Rule of Alternation
Use of Fingers of One Hand. " J u m p i n g " and Alternation
R h y t h m in Typewriting

22
27
31
31
32
33
34
36
39
51
51
53

T H E TEACHING OF T Y P E W R I T I N G

57

Selection of Typists and Methods of Training
Scientific Arrangement of the Keyboard

57
58

(i

TABLES

I.
II.
III.
IV.
V.
VI.
VII.
VIII.
IX.
X.
XI.

XII.
XIII.
XIV.
XV.

Period of contacts and intervals, graph 84
Period of contacts and intervals, graph 224
Relative period of depression of different keys
Duration of strokes with the left and right hand respectively
Duration of strokes and intervals for four subjects . .
Analysis of graphs of the touch of good typists
Results of typing sentence with the right hand only . . .
Results of typing sentence with the left hand only. . . .
Results of striking isolated keys
Results of typing isolated words
Variations in speed index, contacts, and intervals, with
complete absence of alternation (test 1), occasional
alternation (test 2), and equal distribution of alternation and non-alternation (test 3)
Results of tapping tests
Relation between period of contacts and intervals with and
without alternation of the hands
Results of " j u m p i n g " and alternating fingers of the same
hand . .'
Average deviation of alternating intervals

Page
24
25
26
26
28
30
32
33
34
35

37
42
47
52
54

FIGURES

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.

Diagram of apparatus for pneumatic recording
Marey drum
Desprez recorder
The first Desprez recorder (on the right) records strokes electrically, the second (in the middle) records hundredths of
seconds, and the Marey drum records strokes pneumatically
Diagram of appliance for electric recording by spring contacts
Diagram of contacts for electric recording by mercury
contacts
'Map ' typewriter fitted for electric and pneumatic recording
Graphic record of the sentence: Nous sommes en possession
General view of the apparatus
Diagram showing numerical results of graphic record of
one subject's strokes
Graphic record of the word monsieur
Distribution curve of reaction times, subject A
Distribution curve of reaction times, subject B
Chronological record of reaction times, subject A
Chronological record of reaction times, subject B

Page
11
12
13
14
14
15
17
18
20
29
38
49
49
50
50

INTRODUCTION

Technical Progress and Scientific Method
The efforts which men have made throughout the centuries
to perfect the technique of the ancient trades almost pass the
imagination. Innumerable blind attempts, experience lost and
regained, sage observations forgotten and repeatedly revived,
have gone to build up the tradition of the carpenter's or blacksmith's movements, and make of it a rule of work which is
religiously handed on from master to apprentice.
In primitive civilisations the technique of many trades, as
sociology and ethnology have proved, had their origin in magical
beliefs. The effects of movements which are today attributed
to the. physical constitution of the worker were believed by primitive man to be the work of a supernatural power in the tools
themselves. The primitive beliefs of the first artisans have
given its strength to the tradition which compels a young
worker to adopt the movements evolved by age-long experience.
Excellent though the final results of such methods may be,
it is inconceivable that they should be applied to the many new
callings which have appeared in modern society. Nevertheless,
one of the newest occupations, that of typewriting, followed this
outworn method until psycho-physiology took its rightful place
in the sphere of human labour and offered, in place of the primitive rule of thumb, a technique based on scientific study of
the movements required for the typewriter. This is as yet still
a recent development, as to the writer's knpwledge his are the
first experimental investigations undertaken.
Although typewriting is barely forty years old, it has no rules
of work except those due to haphazard individual discoveries,
sometimes handed on, more often lost, and in no case absolutely

— 8 —
valid, as they were not based on scientific investigation. The
keyboard, for instance, was invented in America for purposes of
correspondence in the English language; the letters are grouped,
according to the frequency with which they occur, on those
portions of the keyboard which seem easiest to reach. The
frequency with which letters occur varies, however, in different
languages; and thus English-speaking typists are exceptionally
favoured in comparison with their colleagues in other parts
of the world.
But there is a still more striking fact to be noted: as the
letters were arranged solely according to their frequency of
occurrence, no attempt was made to determine their relation
to the ease or otherwise of different finger movements. No one
tried to discover the easiest combination of movements in order
to use them for the combinations of letters which occur most
frequently. This study will show that the so-called "universal
keyboard " is not really the best, even for Americans who gave
it to the world.
In trying to frame rules for teaching typewriting, false
conceptions of human work were adopted. Instead of utilising
observation and experiment, the so-called "ten-finger method "
was imposed on all pupils, regardless of the fact that this method,
although excellent for exercising the muscles of the hand, had
absolutely no experimental basis and was in fact used by a very
limited number of persons only. Needless to say, skilled psychophysiologists were never invited to assist by studying the
movements involved in order to determine scientifically the
laws governing the adaptation of movements to the purposes
of the occupation, i. e. accuracy and rapidity. The legitimate
criticisms of the ten-finger method subsequently to be advanced
show the utility of rules based on graphic analysis of movements.
Readers of this study should certainly be convinced that by
the adoption of scientific methods humanity could gain years
and even centuries in establishing the best rules for executing
occupational work. Co-operation between manufacturers and
scientists should become the basic principle of social progress.
The experiments described in this volume were begun in 1912
and continued until 1914, when the results now published had
already been obtained. They were not then made public, as it
was desired to carry out a series of control experiments with

— 9 —
professional typists whose technical opinions did not agree with
these results. The control experiments were resumed after
the war. In July 1921, for example, the presence in Paris of a
large number of first-class typists who were competing in the
international championship organised by the Chamber of
Commercial Organisation (Chambre syndicale d'organisation
commerciale) offered an opportunity for further experiments.
The time is now ripe for publishing the results of these experiments, with the certainty that they will provide professional
typists with positive data enabling them to improve or modify
their methods of work.

I
Method of Recording
The method employed for analysing touch was to record the
movements of the keys of a typewriter when struck by typists
of all grades of skill; these movements are measured by the
graphic method usually employed in physiological laboratories.
Three methods of transmitting the movements of the keys to a
recording cylinder were employed, and will be described in
succession.
PNEUMATIC

RECORDING

As the keys of a typewriter are rigidly fixed to the levers (key
bars) which move the type bars, a graph of the striking movements can be obtained by recording the movements described
at any point on the mechanism. The. point chosen was towards
the front of the key bar, which is depressed when the key is
struck.

d
/

TTTTTTTTTTTT
t

c
s

c
FIG. 1. Diagram of apparatus for pneumatic recording.

In the pneumatic mechanism adopted (Fig. 1), a flexible
indiarubber tube t, sufficiently rigid to revert immediately to
its normal position when compressed by a blow from a hard

— 12 —
body, is fixed under the machine. This tube is held in position,
perpendicular to the key bars, by a metal trough c, c, c. At one
side of this trough there is a hole through which passes a fine
indiarubber tube. The other end of the tube is connected to a
Marey drum (Fig. 2). The Marey drum, or lever drum, consists

FIG. 2. Marey drum.

of a metal cupola, the mouth of which is covered by a thin
sheet of rubber m. On the centre of the rubber sheet is fixed
a very light metal disc d, carrying a lever hinged at a which
causes a light pen or needle inserted at p to oscillate about a
fixed point /. The movements of this pen are registered on a
revolving cylinder e (Figs. 1 and 4). The whole apparatus is
air-tight, the air inside it being maintained at atmospheric
pressure by a valve s (Fig. 1).
When the typist strikes a key, the key bar is depressed and
comes into contact with the tube t. This drives the air towards
the Marey drum and distends the rubber sheet. A movement
is thus imparted to the recording needle, causing it to deviate
from the position of equilibrium as long as the key bar presses on
the tube. When this contact ceases the needle returns to its
original position. As each key is struck, the corresponding
letter is marked on a sheet of paper inserted in the machine, and
simultaneously the duration and range of the lever's movements
are marked on the recording cylinder. The range is measured
by the distance from its original position traversed by the
recording needle, and the duration by the time which elapses
between the departure of the pen from and its return to its
original position.

— 13 The duration of the movements can be measured by recording
time divisions simultaneously. There are several methods of
doing this. The electric diapason was chosen for the present
experiment, as it gives absolutely accurate readings if the
instrument is graduated. The diapason, for instance, gives
100 double vibrations per second; each vibration makes and
breaks an electric circuit, and the current synchronously moves
the armature of an electro-magnet (Fig. 3) in a regular rhythm

FIG. 3. Desprez recorder.

of 100 movements per second. A recording needle fixed to the
armature impinges on the recording cylinder parallel to the pen
of the lever drum and on the same generator of the cylinder
(Fig. 5).

ELECTRIC

RECORDING

If it is desired to study the duration of each striking movement without regard to the force employed, electric recording
is preferable. The typewriter is placed in an electric circuit,
with a battery and a Desprez recorder. A movable part of the
typewriter, set in motion every time the typist strikes the keys,
acts as circuit breaker, its movements being transmitted to
the recorder which registers them on the cylinder.
The problem of electric recording is therefore to find what
part of the machine can be used as circuit breaker when the
keys are struck. Three different systems were used: spring
contacts; mercury contacts; contacts on the escapement.

— 14 —

FIG. 4. The first Desprez recorder (on the right) records strokes electrically, the second
(in the middle) records hundredths of seconds, and the Marey drum records strokes
pneumatically.

Electric Recording bij Spring

Contacts.

In the first system a horizontal metal rod, resting on two small
hollow supports, is placed under the key bars of the typewriter
and at right angles to them (Fig. 5). A vertical rod t, sliding

FIG. 5. Diagram ot appliance lor electric recordingby spring contacts.

in each of the supports P, with as light friction as possible, is
connected with the end of the horizontal metal rod T at A.
A spiral spring in each support keeps the rod at a fixed level
corresponding to the spot where each key bar touches the horizontal rod. The apparatus is fixed to the typewriter, but is

— 15 —
insulated from it by sheets of ebonite, so that, when the machine
and the apparatus are placed in an electric circuit, it is closed
every time a key is struck and the corresponding key bar comes
into contact with the horizontal rod. This rod must be perfectly rigid, to avoid accidental contacts through vibrations
caused by the violent shock imparted to the keys when struck.
The springs in the two supports ease the working of the apparatus,
while ensuring permanent contact.
The body of the machine is connected to one of the poles of an
electric battery. The horizontal rod is connected to the other
in such a manner that the circuit is closed when the key bar
comes into contact with the horizontal rod. A Desprez recorder
similar to that described above for measuring time is placed in
the circuit; and each contact is thus registered on the recording
cylinder. It was found that this arrangement did not hamper
typists in their work, and proved satisfactory to the subjects
examined, although they were, very particular as to the adjustment of their machines.
Electric Recording by Mercury Contacts.
This system was employed until 1914. After the war,
however, as the investigations progressed, it was desired to
repeat the experiments and make them more accurate. A
new mercury contact appliance was adopted. A tank of
mercury B, 26 by 4 centimetres, was placed under the key bars
of the machine and at right angles to them (Fig. 6). Needles
of galvanised brass were attached to each key bar in the same
vertical plane, almost touching the surface of the mercury.

^
\
B
FIG. 6. Diagram of contacts for electric recording by mercury contacts.

The machine as a whole is connected with one pole of an
electric battery and the mercury in the tank with the other.

— 16 —
When a key is struck, the point of the needle enters the
mercury and closes the circuit, a Desprez recorder registering
the breaking and closing of the circuit. Here also the time was
recorded in hundredths of a second at the same moment and on
the same generator of the recording cylinder. The duration of
each stroke and the interval between strokes could thus be
measured automatically. with great accuracy.
The very accuracy of the apparatus, however, rendered it
difficult to operate. It was necessary to adjust the level of the
mercury so that the keys most rapidly struck did not overlap,
even when the operator was an extremely rapid typist. Once
this adjustment had been made, however, it suited all the subjects.
It was also necessary to ensure that the duration of contact
was the same for every stroke of equal duration. To ensure
this result, the points of all the needles were carefully fixed in
the same horizontal plane, and the level of the mercury and the
body of the machine itself were both kept perfectly horizontal.
The time required to lower the point of the needle to the level of
the mercury was insignificant, a space of one or two millimetres
being sufficient for adjustment.

Electric Recording by Contact with the Escapement.
As machines of different makes were used in the investigations
a simpler apparatus was adopted. The movements of all the
keys of a typewriter are transmitted to the escapement, which
then moves the carriage. This part of the machine is easily
accessible, and is subject to alternating movements corresponding to the movements of the keys. A record of the electric
contacts effected by this part of a typewriter therefore reproduces
the movements of the keys.

C O M B I N E D M E T H O D S OY R E C O R D I N G

The escapement can also be used for pneumatic recording.
For this purpose all that is needed is to place the end of the lever
of a Marey drum in contact with the movable portion; this

— 17 —
drum is connected with a second one placed on the recording
cylinder, so that all the movements of the escapement are transmitted to the needle of the cylinder.
On the French ' Map ' typewriter, now being used for the
investigations, another part of the machine, the universal bar,
was selected, as the range of its movements is greater and
consequently easier to record. A lever drum, placed under the
machine, receives the movements of this bar and transmits
them to the recording drum. As the space bar does not move
in relation with the universal bar, another lever drum is adjusted
to it, and connected with the other drums; its movements are
thus transmitted to the recording drum (Fig. 7).

F I G . 7.

' Map * typewriter fitted for pneumatic and electric recording.

A typist's stroke, recorded pneumatically and electrically,
and the time in hundredths of a second, can thus be recorded on
the same chart (Fig. 8) and analysed with great accuracy.
The machines used for the experiment were a Smith Brothers
No. 5, a Remington with non-visible writing, one of the latest
Remington models, and the French Map machine, with which
experiments are at present being pursued. The graphs reproduced in this volume were all made with Map machines.
2

^í^

1
2
3
1
2
3
F I G . 8. Grapiiic record of the sentence: Nous sommes en possession...
(1) Record of strokes b y p n e u m a t i c transmission (Marey d r u m ) .
(2) Record of hundredths of seconds (electro-diapason and Desprez recorder).
(3) Record of strokes by electric transmission (Desprez recorder).

— 19 —
The apparatus as a whole is shown in figure 9, where the
various parts described can easily be recognised: recording
cylinder, recording signals and drums; and the - typewriter
equipped for the experiment.

II
Choice of Subjects and Test
The persons and the test to be employed for the experiments
were carefully selected. The typists were all professionals of
the highest class, as the difficult and obscure problem of the
abilities of average workers and the conditions required for
doing good work can only be solved by studying the movements
of the best workers.
It was essential that all the subjects should have sufficient
experience to type an ordinary sentence automatically. Several
sentences of this kind were used, but as all the results agreed
those given here are all based on one sentence : Nous sommes en
possession de votre lettre (We are in receipt of your letter), in
order to limit this study to essentials.
This test sentence was chosen in order to obviate any mental
effort, and in fact none of the subjects paid any attention to his
movements. Most of them replied, when questioned concerning
the test : " My fingers hit the keys of their own accord ", and they
all stated that they typed it absolutely mechanically. Some
even said that they "thought on the machine", presumably
meaning that their mental images were exclusively those of
typing movements. and not those of handwriting.
A psychological examination of this kind was necessary to
ensure that all the subjects could be compared with each
other; otherwise the experiment would have had to be carried
out on completely different lines. Another method (I have
recently adopted it for further experiments, the results of which
will be published separately) would have been to follow the
progress made by one subject during the course of his training,
with a view to ascertaining the aptitudes most easily acquired
and his mental losses, or again to take subjects who had attained

F I G . 9. General view of the a p p a r a t u s .

— 21 —
varying degrees of skill according to their period of training
but had all been trained in the same way. Of all these methods
the one provisionally selected was that which seemed best
calculated to give general results.
All the subjects seemed to be keenly interested in the investigations and took an evident pleasure in submitting to the
experiments, thus making it possible to obtain a considerable
number of graphs, which constitute important documentary
evidence. Typists at the top of the profession very kindly gave
up their time and spent long days in the laboratory, then at
some distance from Paris. They made valuable observations
on their individual technique, the analysis of their movements,
sensations, and ideas. They thus provided material for extensive psycho-physiological research and opened up hitherto
unexplored spheres.
Mr. Charles Dellion proved a valuable and untiring collaborator. He was present at many of the experiments and used
the results in working out his method of typewriting 1 . This
method constitutes the first attempt at adapting the data of
experimental science to professional technique. It will mark
an epoch in the history of modern methods of training, as it
shows, supported by practical results, what can be achieved by
co-operation between laboratory and workshop 2 .

Ill
Analysis of the Graphs
When the graphs have been recorded by the methods
described above, they have to be analysed mathematically.
Perpendiculars are dropped from points at the beginning and
end of each movement on the line recording strokes on to the
time scale, from which the duration of each movement can be
read off in hundredths of a second (see fig. 8). The period
1

Ch. DELLION : Tu seras dactylographe. Paris, Editions « Graphica »,
24, rue de Bondy.
2
A recent application of these laboratory researches has been made
by Mr. J. H. Estoup, who in his treatise on typewriting evolves from
these experiments a system of fingering and a method of training.
(J. H. E S T O U P : Métatype. Paris, 20, rue Gassendi.)

— 22 —
of contact with each key, thus found, is entered in column 2
of the tables, and the period of each interval, if between two
strokes with opposite hands, is entered in column 4, and if
between two non-alternate strokes (i.e. with two fingers of the
same hand) in column 5. The average deviation (cols. 6 and 7)
is then calculated by taking the arithmetic average of the differences between the duration of each interval and the average
duration of that class of interval (i. e. alternating or nonalternating). The average deviation is an index of regularity.
Needless to say, this analysis proved very difficult, owing
to the strain on eyesight and attention caused by the number
and minute detail of the signs recorded. The procedure adopted
in order to avoid errors was as follows: the investigators first
made the calculations, which were then done twice independently
by outsiders. The results were compared, and only when all
the figures arrived at agreed were they accepted as correct.
This method of recording and calculation makes it possible
to measure both the period for which the key is depressed and
the period elapsing between the striking of two keys. Each
of these elements in the manipulation of the keys will be considered separately.

IV

. Results
P E R I O D OF DEPRESSION OF K E Y S

When a first-class typist is working at full speed, every key
is not held down for exactly the same period ; it varies from two
to six hundredths of a second. The standard sentence employed
for this study, Nous sommes en possession de votre lettre, contains
41 signs, including the space bar for separating words, and 40
intervals.
Table I gives the analysis of the graph of one of the subjects,
table II that of a graph recorded under the same conditions
eight years later for the same subject, who had not pursued
his occupation for several years, having been called up for
military service.

— 23 —
It is unnecessary here to reproduce the actual graph recorded,
as was done in the preceding experiment (Fig. 8); throughout
the remainder of this report only the numerical analyses will
be given and conclusions based on them. The unpublished
graphs are preserved among the records of the Laboratory of
Experimental Psychology at the "Ecole des Hautes E t u d e s " ,
and may, if required, be published.
The results shown in the two tables are similar, except for
a difference of 32 hundredths of a second in the total duration
of the test, and variations in the length of the strokes and intervals. Variations of the same character had already been noted
with the same subject in successive tests made in 1912 and
quite recently. The increased length of the period of contact
of each stroke, however, is probably due to an improvement
in the method of recording. As the resistance of a brass rod is
by no means negligible, whereas that of mercury is, it is not
surprising that the duration of contact should be slightly
longer and the intervals slightly shorter, as they proved to be.
Careful analysis of the data contained in tables I and II
reveals facts which will subsequently form the subject of more
minute experiments. In the first test, the distribution of strokes
by period was as follows:
N u m b e r of strokes

2
15
16
8

Duration ('/ioo second)

5
4
2
2

Average
3.26
In the second experiment it was as follows:
Number of strokes

7
11
11
7
5
Average

Duration (Vioo second)

6
5
4
3
2
4.2!)

The period of depression of certain keys varied from the
average by four and three hundredths of a second in one case
and three and five hundredths in the other, and it was desired
to find the reason of this. The keys struck were therefore
grouped according to the period of depression (table III).
It was expected to find a relation between this period and the
position of keys on the keyboard. No definite answer to this

TABLE I. PERIOD OF CONTACTS A N D INTERVALS; GRAPH 84
Intervals
Key
struck
(1)

n
0

u
s
Space bar
s
0

m
m
e
s
Space bar
e
n
Space bar
P
0

s
s
e
s
s
i
0

n
Space bar
d
e
Space bar
V
0

I
r
e
Space bar
/
e
1
1
r
e

Period of
contact
C/.00
second)
(2)

4
3
4
3
4
4
3
3
4
3
4
4
3
3
4
4
4
3
5
3
3
2
2
4
3
5
4
3
4
2
4
2
2
4
3
3
3
2
2
2
3

Hand
used
(3)

i

R
R
R
L
R
L
R
R
R
L
L
R
L
R
R
R
R
L
L
L
L
L
R
R
R
R
L
L
R
L
R
R
R
L
R
R
L
R
R
R
L

Period ('/ ,oo second)
Opposite
hands
(4)

—
2
7
8
4

—
—

•

3

—
4
9
2

—
—
—
3

—
—
—
—
5

—
—
—
6

—
3
8
3

—
—
4
4

—
4
5

—
—
3

Deviation C/,00 second)

Same
hand
(5)

Opposite
hands
(6)

Same
hand
(7)

15
11

—

4.10
0.10

—
—
—
—

2.57
2.43
3.43
0.57

—
—
3.10
1.90

—

—
—

1.57

—

1.90

—
—
—

0.57
4.43
2.57

0.90
1.10
5.90

—

—
—
—

1.57

2.90
3.10
3.10
2.90

—

—
—
—
—

0.43

1.90
0.90
1.90

—

—
—
—

1.43

2.10

—
—
—

—

1.57
3.43
1.57

1.10
2.10

—
—

—
—

0.57
0.57

—

3.10

—
—

0.57
0.43

2.90
1.10

—

—
—

1.57

14
9
9

10
12
5

8
14
14
8
9
10
9

13

12
13

14

8
12

—
Total
87
134
229
Duration of
the test
4 50
Average
period of the
41 contacts
3.26
Average of the 19 intervals
between strokes with opposite
hands
4.57
—
Average of the 21 intervals
between strokes with the same
hand
10.90
—
General average of intervals
7. 90

—

—
—
—

—

—

—

—
—
—
—
—

—
—

—

31.85

48.10

—

—

1.67

—

—
—

2.29

—

,

TABLE II. PERIOD OF CONTACTS AND INTERVALS ; GRAPH 224
Intervals
Key
struck

Period of
contact
C/.00

Hand
used

second)

a)

(2)

(3)

71
0
11
S

6
5
2
6
5
6
3
3
4
5
6
4
5
3
5
4
5
4
6
5
4
5
3
4
2
5
6
6
4
3
4
3
2
5
4
5
4
2
3
2
4

R
R
R
L
R
L
R
R
R
L
L
R
L
R
R
R
R
L
L
L
L
L
R
R
R
R
L
L
R
L
R
R
R
L
R
R
L
R
R
R
L

Space bar
s
0

m
m
e
s
Space bar
e
n
Space bar
P
0

s
s
e.
s
s
i
0

n
Space bar
d
e
Space bar
V
0

1
r
e
Space bar
/
e
t
I
r
e

(Period »/ ,oo second) Deviation (i/ioo second)
Opposite
hands
(4)

Same
hand
(5)

Opposite
hands
(6)

3
6
9
2
—
—
2
—
4
8
2
—
—
—
4
—
—
.—
—
.4
—
—
—
6
—
4
6
3
—
—
4
3
—
2
5
—
—
3

15
10
—
—
.—
—
13
12
—
8
.—
—
—
10
14
7
—6
12
13
6
—
11
10
11
—
8
—
—
—
13
10
—
—
18
—
—
12
13
—

—
1.21
1.79
4.79
2.21
—
—
2.21
—
0.21
3.79
2.21
—
—
—
0.21
—
—
—
—
0.21
—
—
—
1.79
—
0.21
1.79
1.21
—
_.
0.21
1.21
—
0.21
0.79
—
—1.21

.—
—
—
2.17
1.17
—
2.83
—
—
—
0.83
3.17
3.83
—
4.83
1.17
2.17
4.83
—
0.17
0.83
0.17
—
2.83
—
—
—
2.17
0.83
—
—
2.17
—
—
1.17
2.17
—

27.27

44.51

—

—

1.43

:—

—
—

2.02
—

—
80
227
Total
173
Duration of
the test
48 0
Average
period of the
41 contacts
4.29
Average ol the IS) intervals
between strokes with opposite
hands
4.21
-—
Average of the 21 intervals
between strokes with the same
hand
—
10.80
General avera ge ot inte rvals
7.É /

Same
hand
(7)

4.17
0.83

— 26 —
question was obtained, but one fact was noted, namely, that
strokes with the left hand were longer than those with the right.
TABLE I I I . RELATIVE P E R I O D OF DEPRESSION
OF D I F F E R E N T K E Y S
PERIOD
(seconds)

1

KEYS STRUCK

6/1OO
5/1OO
4/IOO
S/100

s- LU © 0

© ® (2

0

S O S

2 (H S

2/1OO

Underlined

®

D

1

=

struck with right hand.
struck with left hand.
alternation before and alter,
alternation after.
alternation before.

Subject: D ; test given in 1920.

Table IV, containing the results of 14 tests with six subjects,
tends to show that this fact may be treated as a generalisation.
TABLE IV. DURATION OF STROKES W I T H T H E L E F T AND RIGHT
HANDS

RESPECTIVELY
Contacts

Subject

Number
Right |

A

—
—
B
C
B
A

—
—
D
E

—
—
F

20
20
20
20
25
20
20
20
20
25
20
20
20
20

Total period
(V100 second)

Average per stroke
(V100 second)

Left

Right

Left

Right

21
21
21
21
16
21
21
21
21
16
21
21
21
21

82
87
96
86
116
92
100
97
102
93
99
99
102
147

105
106
128
106
77
107
122
120
1.17
80
139
1.44
136
184

4.10
4.35
4.80
4,30
4.64
•1.60
5.00
4.85
4.90
3.72
4.95
4.95
5.10
7.35

Number
of graph

Left

5.00
5.04
6.09
5.04
4.81
5.09
5.80
5.71
5.76
5.00
6.61
6.85
6.47
8.76

238.1
238.2
234.3
238
90
238
234.1
234.4
234.2
214
234-B
224
224-B
303

— 27 —

DURATION OF INTERVALS
NECESSITY OF ALTERNATING THE HANDS

The analysis of graphs given in tables I and II shows definitely the conditions determining rapidity in the intervals
between strokes; this point is brought out in columns 4 and 5
of the tables.
It will at once be seen that when two letters are struck
with different hands the interval is much shorter than when
they are struck with the same hand. The averages bring this
out very clearly.

First test . . . .
Second test . . .

Alternating
strokes

Non-alternating
strokes

General,
average

4.52
4.21

10.90
10.80

7.71
7.67

These experiments, frequently repeated with very different
subjects, always produced similar results. Take, for instance,
three other typists, varying widely both among themselves
and from the first (D), whose performance has already been
recorded. C is a well known champion typist, who uses only
one finger of each hand. G is a woman, slightly less skilled
than the two preceding, but an excellent typist. H is also
an expert, and though not a champion has the reputation of
being one of the best typists in the profession. Detailed analysis of their tests, on the lines shown in tables I and II, confirms the fact noted above.
If the results of the analyses are grouped in a single table,
it will at once be seen that the intervals are always shorter
when the hands are used alternately (table V).
For some of the subjects the results of two tests are given,
in order to show (1) that the same experiment repeated under
absolutely identical conditions — as in the case of G — does
not give the same numerical results, a fact to be explained
by the instability of the human psychomotor system, which
in itself constitutes a capacity for improvement; (2) that,
whatever numerical variations may be noted with the same
subject, the rule as to alternation of the hands is always confirmed.

— 28 TABLE V. DURATION OF STROKES A N D INTERVALS, IN H U N D R E D T H S
OF A SECOND, FOR FOUR

SUBJECTS

Average period
Subjects
in order of
speed

C
D

(1913)
(1921)
G (1913)
(1921)
H

Period
of test

Strokes

Intervals
between
alternating
strokes

411
450
480
536
525
550

4.58
3.26
4.29
6.24
6.04
3.12

4.26
4.57
4.21
6.00
6.38
7.00

Average deviation

Intervals
between
Nonnon-alter- Alternating
alternating
nating
strokes
6.76
10.90
10.80
7.47
7.25
13.17

1.06
1.67
1.43
1.07
1.48
1.52

1.70
2.29
2.02
1.82
2.82
2.09

The facts mentioned above are even more strikingly shown
by another graph (Fig. 10), in which the strokes and the intervals between them are the abscissae and the time in hundredths
of a second the ordinates. Two curves are then drawn, one
(continuous line) denoting the fluctuations in the length of
the intervals, the other (dotted line) the fluctuations in the period
of the strokes. Intervals between two non-alternating strokes
are marked on the intervals curve by small circles. Strokes
with the left hand are marked on the other curve by a cross.
It will be noted that the two curves differ, that of the intervals
fluctuating more than the other. The greatest fluctuations
are due to non-alternation, and the contact curve shows less
dispersion than the former. Strokes with the left hand generally occur on the upper portion of the curve.
My most recent enquiries, using the most accurate methods
with a large number of some of the most rapid professional
typists, have served to extend and confirm the first results
obtained. It should be noted that the subjects, though all
first-class typists, employ different methods, using 4, 6, or 10
fingers. The results are shown in table. VI, to which are added
data supplied by the tests with a subject who uses the tenfinger method almost exclusively, and who. though not a frequent competitor in championships, may be regarded as a
first-rate typist. On the other hand, subjects like C, who has
often won championships, and H, who enjoys a great reputation in the profession, both use two fingers only, and yet appear
equally efficient, although using different methods. It would

__

^>

<D
t_

\s

^- .'•
^

-»i

é

,

7*"n
^ ***» ~,
/
•

O

—
^g.

>

Sui
TJ

e

\

—•—•

**~"»-^

r> )

.,''

<¿. —>—. *-

• ^ ^

-'•*

N

-C

o

p*^'
\

e —
. ——
—
>-~ "—

r
s

/

'

^ >

?

o
a

s

v

-v

">.,

^

=-

\

-^
:?*

=-

\
=~

\

^

1^: - ,
^ .-'
/
( O h ( 0 i / l ' t i O t \ | - O Û ) i O l

> o m m e í

\

s

pi

.«¿=

0

v

i D i n ^ l O N

o
•e

TABLE VI. ANALYSIS OF GRAPHS OF THE TOUCH O

Subject

Period
of test
C/,00

second)
(1)
(2)

A
B
A
C
A
D
E
G
H
F
H

386
387
395
401
404
407
411
417
417
450
480
510
510
525
536
550
570
589

Average t o t a
Average
duration of
duration
contact and of contacts
interval :
C/.oo
second)
index of speed
('/«»second)
(4)
(3)

9.35
9.38
9.75
9.69
9.77
9.92
10.09
10.14
10.14
11.16
11.96
11.96
12.37
12.97
13.17
13.20
13.39
14.18

4.53
4.70
4.31
4.16
5.39
5.39
4.55
5.46
5.34
3.26
4.29
5.56
5.80
6.00
6.24
3.12
8.07
3.02

Average duration of interval
C/,w second)
Alternating
strokes
(5)

3.64
3.58
3.88
3.70
3.17
3.47
4.26
3.76
3.88
4.57
4.21
4.76
5.11
6.38
6.00
7.00
3.31
8.11

Non-alternating
strokes
(6)

6.00
5.78
6.60
6.56
5.60
5.60
6.76
5.60
5.73
10.90
10.80
8.04
8.04
7.25
7.47
13.17
7.17
14.21

Ave
(7

4
4.
5.
5.
4.
4
5
4,
4
7
7
6
6
6
6
10
5
11

— 31 —
seem, therefore, that the common elements in the touch of all
these subjects, as recorded by my system, is in fact characteristic of the vocational physiology of a typist.
At the beginning of this report it was affirmed that the data
obtained from first-class typists could be generalised. This
hypothesis had to be confirmed; subjects with a good knowledge
and practical experience of typewriting, though slow, were
therefore tested. The results confirm those of previous tests.
One subject, for example, was naturally slow in movement,
his slowness being almost pathological; he strikes the keys as
slowly as it is possible to conceive. This person is fifty years
of age and has been a typist for twenty years; though such a
slow operator he is useful in the office where he is employed
owing to his scrupulous accuracy. Even in a semi-pathological
case like this, the rule of alternation is confirmed.
Average time of contact
Alternating intervals
Non-alternating intervals

Hundredths
of a second
4.75
24.90
27 .90

All the results obtained therefore confirm the theory that
alternation of the hands promotes speed in typewriting.

CONFIRMATION OF R U L E OF ALTERNATION
NON-ALTERNATING PHRASES AND ISOLATED LETTERS

A rule to which there are no exceptions in ordinary sentences
nevertheless has to be checked by control experiments. For
this purpose special tests were given, consisting of sentences
typed without alternating the hands, i. e. entirely with one
hand.
Sentence Typed with the Right Hand.
The sentence chosen was: CO, hop là, mon joli moujik, which
all typists, whatever method they follow, type entirely with
the right hand. The test was given after the subjects had mastered this somewhat peculiar sentence, and the unfamiliar

— 32 —
one-hand method involved. The results of analysing the graphs,
as was done for the other sentences, were as follows:

TABLE VII. RESULTS OF TYPING SENTENCE
WITH T H E RIGHT H A N D ONLY

Subject

Average
period of
contact
(Vioo second)

Average
period of
intervals
("1W second)

Total:
index
of speed
( V i c second)

Number
of
graph

(1)

(2)

(3)

(4)

(5)

B
A
D
—
G
F

4.60
5.68
3.76
6.07
6.92
6.64

8.25
9.00
14,51
14.07
15.95
16.33

12.85
14.68
18.27
20.14
22.87
22.97

235
235
214-4
87-1
88
303

All the subjects were excellent typists, the first two of the
highest class. They were ranked according to their score in
column 4 (the sum of the averages in columns 2 and 3). It will
be noted that this ranking corresponds exactly to one made
on the basis of the average period of the intervals, which would
seem to show that speed is attained rather by shortening the
intervals than by shortening the period of contacts.

Sentence Typed with the Left Hand.
A similar method was adopted with the sentence Restez vers
cette terre, which is ordinarily typed entirely with the left hand,
except for the space bar, which is always struck with the right.
The results, which are given in table VIII, were what was expected.
Here also the ranking according to the sum of the average
interval and contact coincides with one based on the average
period of intervals alone, except for the two best performers,
who were always very close together and considerably ahead
of their competitors, although the latter were also excellent
typists.

— 33 —
TABLE VIII.

RESULTS OF T Y P I N G

WITH THE LEFT

Subject

B
A
G
D

HAND

SENTENCE

ONLY

Average period of intervals
(V100 second)

Average
period
of
contact
O/ioo
second)

Nonalterna ting

Alternating

Average

5.21
5.86
5.86
7.08
5.04

6.06
5.87
6.93
9.87
1-1.81

5.50
4.83
8.00
8.66
13.16

5.76
Ó.35
7.46
9.26
14.36

Average
total
period:
index of
sjiced

('/,..

Number
of
graph

second)

10.97
11.27
13.32
16.34
19.40

235
235
88-D
88-C
214-3

The experiments might be completed by tests with a sentence
in which the hands are used in regular alternation, but it is
difficult to find one complying with these conditions. The
space bar, which is always struck with the right hand, makes
complete alternation impossible — a difficulty which also
arose, as mentioned above, with the sentence typed entirely
with the left hand. "With the alternating sentence it was thought
that the difficulty could be ignored, since the alternation
involved in striking the space bar in itself promoted speed,
and to prove the greater is to prove the less. In this case,
however, the interruption of the regular alternation frequently
lessened speed, or, in other words, was unfavourable to the
subject, and diminished the value of the tests. For this reason
it was considered preferable to neglect it.

Typing Two Isolated Letters.
The rule of alternation was verified even more simply and
accurately. Supposing this rule to be correct, if a single key,
such as /, be struck repeatedly and as rapidly as possible and
then another key, say /', in the same way, the speed so achieved
will be. less for either key than if thekeys were struck alternately.
Experiment on these lines fully confirmed the rule, as is shown
below.

3

— 34 —
TABLE

IX.

RESULTS

OF STRIKING

ISOLATED

Average period
(Vioo second)
Key struck and finger used

KEYS

Total
second)

Number
of
graph

15.04

215-1

17.68

215-2

7.67

216-1

(7.00

Contacts

Intervals

f (first finger, left)
j (first finger, right.)

5.60
2.40

9.14
15.28

Alternation :
/•(first finger, left)
j (first finger, right)

3.60

4.07

/'(first finger, left)
Space bar (first finger, right)

5.04

3.59

8.63

215-3

Non-alternation:
j (first finger, right)
Space bar (first finger, right)

3.66

11.35

15.01

215-4

j (first finger, right)
Space bar (right thumb)

4.84

17.88

22.72

215-5

The first three lines in the table show that alternation shortens
the duration of the movements by 50 or even 75 per cent. The
size of this proportion, greater even that than produced by previous experiments, is due to the fact that this test brings into
play only the most elementary forms of nervous activity.
The fourth line of the table shows that the letters / and / do
not constitute a special case, since alternation with the space
bar (a part of the machine offering greater resistance) also
produces greater speed. The figures, in the fifth and sixth
lines confirm these data, showing also that it is definitely
better to strike the bar with the index finger than with the
thumb, as generally recommended. This, however, needs to
be verified by more numerous experiments.

Typing Short Isolated Words.
These experiments can be extended to groups of letters
forming ordinary words, which occur very frequently in typewriting. Table X, for instance, gives the results of analysing
a graphic record of typing the word avec.

— 35 —
The first line gives the result where the word was typed on the
ten-finger system, i.e.
a — fourth finger, left.
v — first finger, left.
e — second finger, left.
c — second finger, left:
space bar — right t h u m b .
a — fourth finger, left, etc.

There are two alternate for every three non-alternate movements; and the rule is verified. A more rational method of
typing, based on the principle of alternation, would be as follows :
a — fourth finger, left.
v — first finger, right.
e — second finger left.
c — first finger, right.
space bar — first finger, right.
a — fourth finger left, etc.

This gives four alternate movements to one non-alternate;
the results given in the second line of the table show that speed
is increased 50 per cent.
TABLE X. RESULTS OF T Y P I N G ISOLATED WORDS

Word and method
of t y p i n g

Contacts
C/.00
second)

Intervals ('/ioo second)
Alternating

Nonalternating

C/,00
second)

Number
of
graph

Total

avec (ten fingers), alternating with space
bar only

5.60

12.77

15.73

28.50

218,1

avec, maximum alternation

3.69

4.77

9.28

14.05

218,2

vers (ten fingers), alternating with space
bar only

5.46

8.42-

12.08

20.50

218,3

vers, maximum alternation

3.35

4.35

13.28

17.63

218,4

lettre (ten fingers), 2 alternations with space
bar only

4.71

6.87

15.42

22.20

218,5

lettre, alternating

4.07

5.80

11.18

16.98

218,6

— 36 —

If another word, vers, is typed according to the two methods
(table X), the totals show the superiority of the alternating
method. In this experiment the total length of each test was
compared taking into account the same number of signs in
each case; the first test took 621 hundredths of a second, the
second only 359 hundredths.
Finally a longer word, lettre, typed according to the two
methods (table X), again provides confirmation of the rule.
These experiments were made with a subject, who uses the
six-finger method, but had previously practised a long time
before, typing with ten fingers. For definite conclusions, however, the experiments would have to be repeated.

RELATION BETWEEN PERIOD OF CONTACTS, PERIOD OF
INTERVALS, AND FORCE OF STROKE

The period of contacts and that of intervals have hitherto
been studied separately; it now remains to consider the relation
between these two factors in the speed of typewriting. If, in
table VI, the. figures for the period of contacts are subtracted
from those for intervals, both positive and negative results
are obtained in a single test.
A

B
D

+0.29
— 0.02
— 1.01
— 0.86
— 0.78
— 0.54
+ 0.93
+ 0.97
+4.64
+ 3.38

E
G
H
C
F

— 0.84
— 0.79
+ 0.97
+ 0.53
+6.95
+ 8.13
+ 0.93
— 2.61

Other experiments produced similar results, but those given
above are sufficient for present, purposes. It will be. noted that
the subjects fall into three classes:
(1) those whose intervals are longer than their contacts,
i.e. B, D, G, H, C;
(2) those whose intervals are the shorter, i.e. E, F ;
(3) mixed types, such as A.; these are rare.
There, are clearly marked individual differences in each group,
for example, among B, D, and H, from which a motor characteristic can often be deduced, though it cannot be related to

— 37 —
special efliciency. Though D, B, and H employ different
methods, all turn out excellent work. Investigations now in
progress, however, may, it is hoped, establish a relation between
this characteristic and the rhythm of strokes which could be
used in training typists.
The question must also be studied from another point of view,
to ascertain how these two factors vary when speed increases
through alternation of the hands. When there is no alternation
whatever (see table VII), the speed index of a single subject
increases by one-third, if not more (table XI), wliich denotes
a corresponding decrease in typewriting speed. The next
question is to what extent the two factors in this index — period
of contact and period of interval — are modified by alternation
of the hands.
Roth factors are decreased. If a new element of comparison
be introduced, consisting of a left-handed test in which the
necessity for striking the space bar with the right hand causes
occasional alternation, it will be noted that the period of the
intervals varies with changes in the index of speed, but that of
the contacts does not do so.
TABLE X I . VARIATIONS IN S P E E D I N D E X , CONTACTS, AND I N T E R VALS, WITH COMPLETE ABSENCE OF ALTERNATION (TEST 1 ) ,
OCCASIONAL ALTERNATION (TEST 2 ) , AND EQUAL D I S T R I B U TION OF ALTERNATION AND NON-ALTERNATION (TEST 3 )

Subject
1

Speed index
C/ioo second)

Period of intervals
0/,oo second)

Period of contacts
(7,00 second)

Test

Test

Test

2

3

1

2

3

1

2

3

B

12.45 10.97

9.69

8.25

5.76

5.13

4.60

5.21

4.31

A

14.68 11.27

9.35

9.00

5.35

4.82

5.68

5.86

4.53

D

20.14 19.40 11.16 14.57 14.36

7.90

3.76

5.04

3.26

6.79

6.92

7.08

6.24

G

22.87 16.34 12.97 15.95

9.26

The above table, which gives a few figures illustrating these
points, shows that in the second test the period of the contacts
even varies in inverse ratio. This cannot, however, be treated
as a generalisation, as opposite results were obtained in several

— 38

-

cases. It can simply be stated that
speed in typewriting is mainly
achieved by shortening the intervals;
shortness of contacts would seem to
be an effect rather than a cause of
speed. This point required further
investigation.
The third factor in touch, namely,
the force used in striking a key, is
also related to the period of the contacts and intervals.
Alternation of the hands immediately reduces all the factors in movement: period of contact, intervals,
and strength of touch. This is best
shown by pneumatic recording machines.
One of the subjects, C, having
stated that he had been struck by
the ease with which he typed the
word monsieur, a graph of this word
as typed by him was accordingly
made with a pneumatic instrument.
This graph (Fig. 11) shows that the
first three letters, which are struck
with the same hand, are typed relatively slowly, both contacts and
intervals being comparatively long;
the last five letters, for which the
hands alternate regularly, are typed
with great speed; the intervals are so
short that the curves for the strokes
almost run into one another, and the
contacts so short that the recording
needle runs less than half as high as for
the first three letters, which shows
that the force of the strokes decreased
by half. The first three letters were
typed in 40 hundredths of a second,
end the last five in only 29hundredths.

— 39 —
PHYSIOLOGICAL REASONS FOR ALTERNATION OV THE HANDS

The standard technique of typewriting when investigated was
observed to conflict with the result of laboratory experiments,
which should scientifically provide the. basis of all training.
Enquiry into the cause of this error showed that it arose from
the false analogy commonly made between the technique of the
piano and that of the typewriter.
The pianist produces a melody by contacts and intervals
of varying length which is strictly determined by the composer.
A typist, on the other hand, should not consider such matters;
his movements must be rapid and as uniform as possible. The
keys of a piano may be played either legato or staccato, whereas
in typing each stroke must always be distinct. In piano
playing harmonies are formed by striking several notes simultaneously, the left hand being used largely to play chords as an
accompaniment to the air. The technique of typewriting is
entirely different, for in this case to strike two keys at once is
a serious fault. The keyboard of a piano is long and narrow
and the hand must therefore be spread as widely as possible.
The keyboard of a typewriter, consisting of four parallel lines
of keys, is much more compact; it does not call for spreading
the hand, but in strict logic should require, lengthening the
fingers, which is absurd.
For all these, reasons, there is a need for a technique of typewriting entirely different from that of the piano.
It is unnecessary here to pass any final judgment on the tenfinger method, which is based on a false analogy between the
typewriter and the piano, but it may be noted that the majority
of typewriting pupils practise for months and years in order to
achieve difficult feats, for which they have, no aptitude, simply
because this method has become orthodox.
Moreover, the same fate befalls typewriting methods as any
other form of constraint on the working of the human organism
imposed either by prejudice or in the name of some half-science:
physiology always reasserts its rights. When, at. the height of
the craze for Taylorism. I ventured severely to criticise the work
of the celebrated American engineer 1 , I encountered more
1
J. M. L A H Y : Le système Taylor et la physiologie
sionnel. Paris, Gauthier-Villars.

du travail

profes-

— 40 —

hostility than experimental criticism of an unduly rigid method
of typewriting would be likely to meet with today. The human
organism tends to free itself from external constraint; defective
training alone does not account for the fact that the majority
of typists taught on the ten-finger system later automatically
take to using six, four, or even two fingers, as best suits their own
organism. This was attributed to deliberate slackness on their
part, and pupils were still compelled painfully to learn the
orthodox system of fingering. Yet it would have been easy
to make the simple experiment of tapping rapidly with one
hand or the other, and then with the two hands alternately;
this would have shown that much greater speed was attained
when the hands were alternated.
This phenomenon has long been studied in laboratories of
experimental psychology by means of tapping tests. There
are two methods of recording taps, pneumatic and electric.
With the pneumatic method the subject taps with one or
more finger tips on a semi-rigid membrane covering a closed
wooden box. In a small opening in the box is inserted a rubber
tube-connected with a Marey recording drum; thus each tap can
be recorded on a revolving cylinder.
With the electrical method the subject wears a metal thimble
or holds a special pen with which he taps as rapidly as possible
on a copper plate. The plate, the pen (or thimble), and a
Desprez signal are placed in an electric circuit. When the disc
is tapped with the pen the circuit is closed, and it is broken as
soon as contact ceases, the Desprez signal recording each tap
on a revolving cylinder. The time in fifths of a second is also
recorded with a Jacquet chronograph.
The typewriter equipped for experiments as described above
also provided a convenient tapping test and was used for this
purpose.
It is in many respects better than other appliances,
as alternation of the fingers as well as the hands can be recorded.
It should be noted, however, that the results obtained by
different methods cannot be compared, particularly those
obtained with a typewriter and those with the two tapping
appliances described above, owing to the resistance of a mechanical system of transmission, which is a negligible factor in_electric
or pneumatic tapping instruments. Except where otherwise

— 41 —
stated, the results quoted here were obtained with electric
tapping instruments.
With these appliances graphs are obtained from which can
be calculated motor rapidity, expressed by the number of taps
per second during a 30-second test, and the coefficient of liability
to motor fatigue, expressed by the formula
n1 + n2

~

r?

where n 1 represents the first five seconds' tapping and n 2 the
last five.
Table X I I shows all the results obtained by analysing the
tapping tests of a group of typists who assisted in the experiments during the International Championship at the Grand
Palais in Paris. The table gives the figure for motor rapidity
and the coefficient of liability to motor fatigue of each subject.
It will at once be noted that alternation promotes speed.
The figures given in previous tables and in table XII do not
refer to exactly similar phenomena, as the former indicate the
length of the intervals between two strokes in typewriting, the
latter the number of experimental taps per second ; nevertheless
the direction in which the figures vary indicates the relation
between them. In elementary movements such as tapping
(table XII) alternation increases motor speed by about onethird; this increase corresponds to that already noted in actual
typewriting. It would appear, therefore, that alternation in
typewriting is a primary phenomenon, since it is based on a
physiological characteristic as elementary as that involved in the
tapping test.
If this be the case, the speed attained by experienced operators
constitutes a limit which they cannot exceed. Several facts
can be cited in proof of this. When a subject is asked to type
as fast as possible, but to maintain a regular rhythm, he instinctively bases this rhythm not on the shortest intervals,
though these appear to favour automatic action, but on the
longest.
One subject who was accustomed to work rhythmically, as he
has to cut stencils, involving greater pressure on the keys in

^ <¡ w o a H v> wo is o z § r ^ > » ^
Subject

r+
S

-S

H
O
G
M

>
Z
Ö
SB
S»
Rapidity
(taps per second)

H

>
•0

Alternate

(-'
M
H' l-l
t 0 s ] 0 0 O 0 0 ! 0 00 0 l ( C K H 0 ! 0 O 0 3 O I » h i t D
OCDCOOOOii-iis3 0 i t O C 5 < 1 0 0 0 * - i - * W * ' a 5
tsj<iüia:oi~aOH^5j2wi-'^o>t^OHJooi-'

g

O
H
O
S)

HI

Left hand

Or<IO^OO'<?>CnOi{J50iOiC50^05(rft05^.10i
C0Or'r'*.O<lr»^C0(»HMÜ\h'0>0l0l
OltúO'tv>Oo4'-—lOiO**t\3>-'1í>-Ul>->OlOdi

M

*]

Right band

>^.<K100
UiO*-i-'>^-ÜiOO*-WttiOOOO^.oo
4^<lüil-iV<r^ts500-^il-'OWUiC^4i-OtOO

Liability to fatigue

+

O O O O O O O O O O O O O O O O O O
MlOMOl-'IOkJOK'MKi-'OllOIOOoO

>
a

Alternate

+

H

Left hand

+
O O O O O O O O O O O O O O O O O O
i-'OriOO-'-'OK'LOMOOIO-'OlOh'
05MO)010MlOO)r'OPO)ai<IWÜiO)M

Right hand

O O O O O O O O O O O O O O O O O O
b b ^ b b i o b b b ^ K o b H b H a o
00<130C3500*-00000^102tsSOr^OOOOooOO

Z

M
o;

Aptitude

Zf —

alternation

4^OOOOOC0OOOOi-1O<IOOOO
MíiWOriüJOl^MUOOiOWOlOWO
O i > a o i - ' W B # . i \ : u o * . c o o M w o

Method

! ooi r 1 OiOöaioooMfflSsoisi

fingering

!

en

— 43 —

order to pierce the wax used for making copies, typed the
standard sentence, Nous sommes en possession de votre lettre,
with the following results:
Hundredths
of a second

Non-alternating strokes
Alternating strokes

12.52
10.52

The average deviation of alternating and non-alternating
strokes, previously 6.38 hundredths of a second, fell to 2 hundredths. This shows a clearly marked tendency towards regular
rhythm if the longer intervals are taken as basis. The figures
are as follows:

Non-rhythmic periods
Rhythmic periods
Increase in periods

Non-alternating
strokes
(Via, second)

Alternating
strokes
C/ioo second)

10.90
12.52

4.52
10.52

1.62

6.00

Difference
C/ioo second)

6.38
2.00
4.38
(tendency to rhythm)

What is here called the tendency to rhythm, i.e. the difference
between the figures expressing the variation in the period of the
strokes (alternating or non-alternating), is merely a convenient
method of ranking subjects from this point of view.
An experiment with another subject gave the following
results:
Non-alterna ting
strokes
C/ioo second)

Alternating
strokes
('/ice second)

C/ioo second)

Non-rythhmic test
Rhythmic test

13.04
14.13

6.82
7.94

7.22
6.19

Increase in period

1.09

1.12

—

Tendency to r h y t h m

Difference

1.03

In comparing the results for these two subjects, it will be seen
that the ability to make typing movements rhythmic is not
found to an equal degree in all typists; it is an individual
characteristic which should be utilised in specialising for copying
work.

— 44 —

Be this as it may, the periods in rhythmic tests were much
the same as the longest ones in non-rhythmic tests, which shows
that the greatest speeds are independent of deliberate rhythm
and measure elementary physiological capacities.
This hypothesis is confirmed by another fact. With the
typewriter used as a tapping apparatus a record was taken of
taps on two adjoining keys: d and ¿>, struck first with the left
hand, then alternately with each hand. The subject of the
experiment (D) tapped 6.13 times per second with one hand,
and 11.66 times per second alternating the hands. Table X I I
shows that in ordinary tapping the same subject tapped 6.12
times per second with the left hand and 10.33 per second with
alternate hands. The figures recorded are thus of the same
nature, and the phenomenon which they indicate appears to be
practically incapable of reduction to more elementary factors.

Without embarking on investigations or explanations in^
pure physiology, one is led to enquire what mechanism determines the limit of typewriting speed.
The nervous and muscular factors involved when two successive strokes are performed with the same hand are probably
the same in each movement. As these movements are very
rapid, not exceeding a few hundredths of a second, the internal
changes in the tissues which determine these movements occur
during the periods of inertia in each movement of the hands
when used alternately.
All movements, however simple, are the result of a chemical
process inside the nerve cells, to mention these only. When a
cell functions it expends part of its reserves, and simultaneously
a process of assimilation takes place to reconstitute these
reserves. There is an incessant interchange of nutritive products and waste products; this is indicated by a substance called
chromophil found in the interstices of the network formed by
the cytoplasm of the nerve cell. When the balance of nutrition
is disturbed, this substance disappears progressively, beginning
at the nucleus, which itself moves from the centre of the cell
towards its circumference, while the cell diminishes in volume.
While the balance of nutrition in the cell is destroyed by
excessive activity, the destruction of this minute organism is

— 45 —
hastened by another phenomenon. The waste products, the residue
of the chemical processes which have taken place, are not
eliminated sufficiently quickly and paralyse the cell's action like
a poison.
These two causes together will stop the action of a cell if
there is not a sufficient interval between two successive stimuli
or two reactions. There are several proofs of this. When
the brain cells of a dog which has been trepanned are stimulated
electrically, the motor reactions to this stimulus are only
obtained if there is an interval of at least one-tenth of a second
between each stimulation; the refractory period in animals is
thus one-teuth of a second. The same is true of human beings.
The special cells forming the retina under the action of light
waves undergo a chemical process which produces visual
sensation. If the successive sensations are to amalgamate, the
intermittent stimulations causing them must strike the retina
at intervals of less than one-tenth of a second. If the intervals
are greater, the sensory cells start a new process and separate
instead of fusing the successive images, for as each refractory
period ends a new image is distinguished.
As is well known, if a subject wishes to produce a series of
voluntary muscular reactions, the maximum number is nine
or ten per second for the most gifted persons under the most
favourable conditions. But this limit is not the same in
reactions produced by exciting the" muscle electrically without
recourse to the nervous system ; in this case 40 separate shocks
per second can be produced. The refractory period of the
nervous cells is thus approximately one-tenth of a second, which
means that if two stimuli are applied to one of these cells within
a shorter period, the same phenomenon will occur as in the
fusion of luminous impressions: the cell will not react to the
second stimulation.
Two things must at once be noted without which it would
be impossible to explain individual differences in typewriting
speed. In the first place, the figure of one-tenth of a second
is not strictly accurate; there are probably individual variations
which, though slight, are nevertheless characteristic. Secondly
— and this is important — the movements of the various parts
of the hand used to strike the keys involve more than one cell.
It is probable that groups of cells fulfil this function, and though

— 46 —

they act almost together substitutions may be effected and
increase the muscular action of the hand. Moreover, though
the same hand may be used for striking, the same finger is not
always used. The reciprocal action of adjacent cells in the
same area of the nervous system is not known in detail, but it
may legitimately be supposed that the use of different fingers
involves different though adjacent cells1.
Take as examples our two best subjects, A and B. A rearrangement of their results, as is done for one of them in
table XIII, gives the following figures, which may be compared
with their results in the tapping test.
Time
C/ioo second)

Unit measured
A

Complete stroke, not alternating
Tapping movement, not alternating
Complete stroke, alternating
Tapping movement, alternating

10.48
14.44
8.16
12.03

B

10.82
13.29
8.17
8.71

It will be seen that both subjects come very near the maximum
rapidity which physiologists have endeavoured to define. The
favourable effects of the position assumed in typewriting and
of the subject's constant practice in this position are clearly
shown by the fact that here' the electric tapping is slower than
the typewriting movements. The tapping test clearly shows
that striking alternately with either hand, using nerve cells
situated on opposite sides of the system and acting independently
of each other, produces very much greater speed in the movements as a whole. This explains the benefit derived from
alternating the hands in striking the keys.
There is too often a tendency to explain apparently simple
physiological phenomena by facts which themselves are too
simple. It is my aim to avoid this mistake. To attribute the
.*. An example of substitution is afforded by the muscles used in
striking the keys of a typewriter. The movement of the finger muscles is
localised to varying degrees in different subjects ; often the entire hand
and sometimes the arm are involved. W h a t happens in the muscles
between the shoulder and the finger tips must also happen in the nerve
cells involved to varying degrees in the act of striking the keys.

— 47 —

TABLE

XIII.

RELATION

BETWEEN

PERIOD

OF

CONTACTS

AND

INTERVALS WITH AND WITHOUT ALTERNATION OF T H E H A N D S 1

Successive
strokes

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

Total
Average

Total
Contacts

Intervals

4
4
5
5
4
5
5
4
5
6
4
5
5
4
3
4
4
4
5
5
5
5
3
4
4
4
6
5
5
3
4
4
5
7
5
2
6
3
5
5

10
5
4
4
6
2
3
7
1
9
3
7
3
5
5
3
3
6
9
8
6
2
8
6
9
1
7
3
6
1
7
4
4
3
6
6
2
8
4
4

180

200

—

—

« Subject: A (grapli.No. 238).

Alternating

Nonalternating

—

14
9

9
9
10
7

—
—

6

—

7
12
8

—
—

—

7

—
—
—
—
7

—
-—
—
5

—

8
11
4
11

—
—
10

—
—
—
—

8
11

—
15

—
—
—
9
8
7

—

10
14
13
11

—
It
10
13

—

13

—
—
—
—
8
9

—

—
—
—
—

11
—
8
11
9
9

139

241

—
8

8.16

10.48

— 48 —
superiority of certain typists over others entirely to the rapidity
of their nervous processes is, I am well aware, an easy solution.
I have pointed out elsewhere that memory, spelling, and various
other mental functions are factors of superiority which, combined
in varying degrees in different typists, lead to superior efficiency1.
But apart from the fact that such functions are very different
from the motor functions just discussed, they do not in themselves explain why highly skilled operators are not exactly the
same in their exercise of the profession. Reverting to the
examples of A and B, who take the highest places in international
competitions, it will be noted that A shows greater regularity
than B in the same tests. Can experimental psychology furnish
an explanation of this difference ?
A partial explanation is already given by the tapping tests
of these two subjects. In simple measurement of motor rapidity B is superior to A, but when the tapping test is in typewriting, and not on experimental appliances, Ais more rapid. To
explain this the reaction times of both subjects were carefully
analysed.
Their reaction times were measured in a series of tests, which
gave the following significant figures.

Significant figures

A

B

Calculated from all the observations.
(1) Arithmetic average (A)
(2) Mean deviation (D)
(3) Index of variability given by the formula

15
1.4

15.33
2

100 1)
A

9.3

Calculatedfrom observations between the quart iles,
i.e. excluding extremes.
(4) Median (m)
(5) Semi-interquartile -i(6) Index of variability given by the formula
i _
2 X

100
m

15
1
6.8

13

15 ( + 0.03)
1.5
9.3

1
J. M. LAHY : " La méthode à suivre pour la sélection des travailleurs :
les dactylographes ", in Anals de l'Institut d'orientacio
professional,
Vol. II, No. 3, pp. 65-78. Barcelona.

— 49 —
These figures show that t h e two subjects are normal and of
average speed, but A is the quicker of the two. The most
typical differences are shown by Galton's ogive (Figs. 12 and 13).

S

c

r1

s
-S

Vlédian ,14 7

1

s

L

Moyenne >15
i l l
1
N u m b e r of reaction times in srder of magnitude

Frequency

F I G . 12. Distribution curve of reaction times, subject A.

1

s
n •

r
,-J

1 u
r
Frequency

¿

Médian M. .3

Moyenne = 15.33 l 1I
1
N uml j e r of reeict ioi i t im es in 01 de r (>f nai mit lid e
F I G . 13. Distribution curve of reaction times, subject B .

With A the mode is clearly defined, there being little dispersion.
The series is homogeneous, the averages are definite, and outlying observations rare; the curve shows that the reactions are
regular and homogeneous. With B the mode is less clearly
defined, the curve has a greater range, and outlying observations are more frequent. A is less variable than B, their
indices of variability being in the proportion of about 7 to 10
for the normal range and 9 to 13 for the total range. A is also
superior in motor ability, as she taps more quickly and varies
less. These facts are brought out even more clearly by the
chronological record of the observations. (Figs. 14 and 15).
In spite of his greater speed in elementary functions (tapping),
B is less well equipped in motor ability for typewriting owing
to the greater instability of his motor functions. The results

— 50 —
of testing reaction times thus agree with those of professional
efficiency.

« Mean

F I G . 14. Chronological record oí reaction times, subject A.

22
21
20
19
18

j i\

/

• 7 -C )
16
15 — \
14
13
12
'1

{^

|
H?F

s1

\

IL
v

V

11

\
-X.
\1
\

r\\
\
\

1

H

1 f

/

—(t-tl-

-j /

/

'"

— Vlean
M )-<.>

V• >

F I G . 15. Chronological record of reaction times, subject B .

These remarks, which might appear in a study of professional superiority, are inserted here to show that, though
physiology can furnish explanations for the necessity of alternatingthe hands, it also teaches that its rules are not yet as rigid
as the laws of physics.
Physiology is characterised by the diversity and multiplicity
of its working in any given subject; the application of rules
based on physiological observation therefore produces less
uniform results than that of rules based on observations in
physics.
The individual differences noted between A and B thus
explain why they were not always ranked in the same order,
though their general efficiency was the same.

— 51 —
LIMITS TO THE R U L E OF ALTERNATION

There would seem to be no exceptions to the rule as to the
alternation of the hands; it is obvious, however, that when
this involves moving both hands towards the extremities of
the keyboard it is desirable to use the fingers of one hand, to
avoid moving the other a considerable distance.
Another reason for limiting the alternation of the hands is to
avoid crossing them. For instance, in the word compartiment,
if the hands are used in strict alternation, for the first syllable
the left hand after striking the letter c would have to pass under
the right hand, which strikes o, in order to type m; such movements would be complicated.

U S E OF FINGERS OF ONE H A N D : "JUMPING " AND ALTERNATION

"When alternation of the hands is limited as described above,
various questions arise. How should the keys be struck with
fingers of one hand ? Is it better to "jump " from one key to
another with a single finger, or to alternate the fingers ?
Using the method of recording already described, these
questions may be answered by analysing the resulting graphs.
The subject is first asked to strike, two adjacent keys in the same
row, / and d, for example, as quickly as possible, with the same
finger; he is then asked to strike the same keys very rapidly
using two fingers of the same hand alternately. The results,
given in table XIV, show that it is better to alternate the fingers.
With two other keys, d and e, on different rows though
adjacent, these results are confirmed (see third and fourth
lines in the table). A new factor may be introduced by the use
of the space bar, compelling alternation of the hands; it will be
seen (line 5) that the speed is scarcely increased at all, in spite
of the added assistance given. By combining alternation of the
fingers with alternation of the hands, however, it will be seen
(line 6) that the maximum speed is attained by combining the
signs involved. If two keys in different rows and distant from
each other, c and c, for example, are struck first by "jumping "
the second finger from c to e, and then by using the second
finger for e and the first finger for c, comparison again shows
the superiority of the alternating method.

— 52 —
These experiments tend to show that for the left hand the
fingers should be alternated. If this rule appears so clearly
for the left hand, one might suppose that it would apply even
more strictly to the right hand in right-handed persons, as the
right hand is generally more dexterous, more frequently
employed, and is often stronger. A similar experiment, however, gave a different result, as shown in the table. Alternation
of the fingers of the right hand is thus not favourable to speed.

TABLE

XIV.

RESULTS

OF " J U M P I N G "

AND

ALTERNATING

F I N G E R S OF T H E SAME HAND

K e y struck and finger used

Period
ot
contact
C/lOO

Period
ot
intervals

CU

Total
C/.00

Number
of

second)

second)

second)

graph

5.81
7.37

15.06
7.60

23.87
14.97

216,2
216,3

5.43

13.66 .

19.09

217,3

8.12

7.06

15.18

217,4

5.19

9.65

14.81

217,5

5.66

5.70

11.36

217,6

4.50

19.73

24.23

217,1

5.50

11.93

17.43

217,2

3.00
4.00

13.40
12.46

16.40
16.46

216,4
216,5

Keys in adjacent rows:
j , u: jumping
alternating

7.03

12.59

19.62

216,9

Keys in distant rows :
ñ, a: jumping (first finger)
alternating: n first, u
second finger

2.68

11.93

14.61

217,7

3.12

15.93

19.05

217,8

Left Hand.
Keys in one row:
/', d: jumping
alternating
Keys in adjacent rows:
rf, e : jumping (second finger)
alternating: d first, e
second finger
jumping, with space bar
alternating
alternating: d first, e second finger (left), bar
right t h u m b
Keys in distant rows:
c, e: jumping (second finger)
alternating : c first, e second finger
Right Hand.
Keys in-one row:
u, i: jumping
alternating

— 53 —
These results proved somewhat surprising, and enquiries
were made among a large number of typists to ascertain whether
they regularly used methods involving the same movements for
both hands. Here again, as with the. ten-finger method, it was
found that hard and fast rules are not followed in practice.
It was also noted that typists alternate the fingers of the right
hand less than those of the left. This is perhaps the reason
why my first enquiries into vocational efficiency tended to show
that ambidexterity is a sign of greater efficiency. It may be
noted in this connection that the greater number of keys most
frequently struck are in the left half of the keyboard.
There thus appear to be two kinds of movements, the efficiency
of which varies. Hand movements are quicker and more skilful
with the right hand than with the left; finger movements are
simpler and quicker for the left hand than the right. These
facts are confirmed by the prevailing practice. Typists questioned on this point agreed that it was easier to strike keys
in the left half of the keyboard with the right hand than to strike
a key in the right half of the keyboard with the left hand.
The following rule may be deduced from these observations:
with existing keyboards alternation of the fingers of the left
hand and " j u m p i n g " the fingers of the right promote speed.

RHYTHM IN TYPEWRITING

The two time factors in striking the keys, period of contact
and alternation, having been analysed, the function of rhythm
in typewriting may now be considered. The solution of this
problem would require a systematic study of touch, with
variations in rhythm. This has not yet been undertaken, but
an idea of the results obtainable may be formed by using the
information already gathered in the course of my investigations.
In typewriting with a good rhythm, the intervals should be
as nearly equal as possible; the rhythm would be further improved if the period of contacts were also equal.
Table XV shows the average deviation in the intervals
recorded by a number of the subjects. It may be recalled
that this figure is obtained by taking the differences between
the general average of the intervals and the actual length of
each interval, and calculating the average of these differences.

— 54 —
This average indicates the stability of the phenomenon under
consideration. The nearer the average deviation is to zero,
the greater is the regularity of typewriting and consequently
the better the rhythm.
TABLE XV. AVERAGE D E V I A T I O N OF A L T E R N A T I N G AND N O N ALTERNATING

INTERVALS

Average deviation
Subject

A
—
—
—
B
F
E
D
C
G
H

Number
of

Alternating
(Vioo second)

Non-alternating
C/.oo second)

graph

1.10
1.16
1.21
1.77
1.66
1.19
1.18
1.17
1.24
1.54
1.38
1.67
0.06
1.26
1.48
1.52

1.00
1.72
1.88
1.37
1.86
2.13
1.82
1.33
2.58
1.58
2.10
2.19
1.70
1.72
2.82
2.09

234-3
238-6
234-4
234-2
238-5
234-1
238-1
238-4
303
224
224-5
84
90
88
89-2
85-1

It will be noted that the average deviation of the alternating
intervals is less than that of the non-alternating, which means
that alternation promotes regularity. But the question then
arises whether it follows from this that in a scientific system
typists should be required to adopt a rhythm in order to increase
their speed - - a conclusion which seems warranted by the
facts stated above. Although no very exhaustive enquiries
were undertaken, a more or less fortuitous experiment may be
cited, namely, cutting stencils.
When a good typist cuts stencils, i. e. when each type, instead
of being impressed on a sheet of paper, penetrates the waxed
sheet used for the duplicator, he types more regularly but also
more slowly. It appears that in any one subject speed is incompatible with regularity of rhythm. This can easily be

— 55 —
explained. In ordinary typing, speed is the main object; the
typist makes the slight movements required in the minimum
time necessary for repairing the wastage of the tissues where
the changes on which the movements depend occur, as explained
on an earlier page. In cutting stencils the typist directs his
attention to each of his movements, watching them and regulating their force as he does not in ordinary typing. Part of his
available energy is thus deflected by this additional effort. His
movements are less automatic, every part of them is consciously
observed, and they are harmonised in the simplest possible
manner, by equalising their values. If two graphs for the
same subject are compared, one made where the subject typed
without considering the force, employed, the other when considering it, it will be observed that the rhythm is far more
regular in the second case.
The second fact to be noted is that, when two good typists
are cutting stencils, the average deviation of the intervals,
which, as already stated, indicates the regularity of their
movements, is as follows:
Average deviation of intervals
Subject
Alternating
('/,oo second)

D
G

1.52
2.35

Non-alternating
(Vioo second)

1 .80
2.01

When the same subjects typed a sentence at full speed
without stencils, the results were, as follows:
Average deviation of intervals
Subject

D
G

Alternating
(VIM second)

Non-alternating
(Y,oo second)

1.67
1.26

2.19
1.72

The complete reversal of the relative position of the two subjects calls for explanation. Before attempting to answer the

— 56

-

question, the average deviation of the reaction times of the
two subjects was measured with a d'Arsonval instrument, with
the following results:
Subject

Period
('/.oo second)

Deviation
(Vioo second)

D
G

13.90
18.53

1.30
2.27

It is probable that the subject with the best psychological
equipment, i. e. whose reactions are quicker and more regular,
loses his regularity in trying to work at high speed. His colleague, who works more slowly, is more stable. Experiments
with reaction times call for very simple responses to sensory
stimulation; but typing is différent. In this case the psychomotor system called into play is more complex, and owing to
this complexity some individuals are able to evolve forms of
reaction peculiar to themselves.
In the first experiment the keys were struck as hard as
possible in order to cut the stencil; in the second no force was
required, and the subjects therefore worked as quickly as possible. Thus speed decreases as rhythm improves.
It would be desirable to ascertain the relation between
speed and force in striking the keys. This is at present the subject
of investigation, which should provide fresh data for a scientific
technique of typewriting. What is already known concerning
rhythm in relation to the length of contacts and intervals
justifies the assertion that rhythm imposed from without
hampers speed. But it does not follow that, if a typist works
in regular rhythm, either naturally or as the result of training,
he cannot work quickly. On the contrary, rhythm added to
other capacities may make its possessor a better typist than
his fellows.
The facts noted appear contradictory at first sight; there is
a spontaneous tendency to regular rhythm in alternating
strokes, which are the quickest; on the other hand, speed
decreases when the typist is required to achieve regularity as
well as speed. But these facts can easily be explained. In the
first case, speed in typewriting tends to approach the physiological limits of speed in simple movements, which are essentially

— 57 —
regular; in the second, rhythm in more complex factors is sought:
alternating and non-alternating strokes, contacts, intervals, and
force. This involves considerable mental effort, which naturally
limits speed.
These observations do not solve the problem of rhythm in
typewriting; they simply show that rhythm is not so important,
for speed at any rate, as has been thought. At the very least
the problem has been stated.

V
The Teaching of Typewriting
SELECTION OF TYPISTS AND METHODS OF TRAINING

All the investigations, both those described in this report
and the much larger number of control experiments, have
furnished two classes of data. There are first those which
agree in all the experiments and may therefore be regarded
as absolute; results of the second class, on the other hand,
though frequently agreeing, are subject to certain exceptions
and can only be regarded as of more or less general value.
This should cause neither surprise nor scepticism concerning
the value for professional purposes of the data supplied by
scientific research. On the contrary, data of this kind should
strengthen our convictions, as they confirm the common impression of the extreme diversity of human types. The conclusion to be drawn is that, though there are a few absolute
rules which are of use in occupational work, there are many
which can only be applied after more or less exhaustive examination of the subject — the typewriting pupil in this instance.
The practical application of the results of these investigations
will be based on this principle.
The two factors in touch have been analysed by the graphic
method; these are the period of contact with the keys, and the
length of the intervals between these contacts — factors which
were formerly indistinguishable. The relations between the two
groups can now be established.

— 58 —
of Alternation ^ fundamental law, which is so general that it should govern
of Hands improvements both in typewriting machines and in typewriting method, is the law of alternation of the hands. Both
makers of typewriters and typewriting experts should endeavour
to make alternation of the hands as frequent as possible. This
law, it should be emphasised, is based, not on an individual
opinion or on any technical necessity in the construction of
typewriters, but on the working of the human organism.
It cannot be ignored that, even though every attempt to
improve typing technique should tend towards alternation of
the hands, such efforts are necessarily limited. However the
letters and signs are arranged on the keyboard of a typewriter,
successive letters in any given word cannot always alternate,
since the keyboard is fixed. No system of typing can secure
perfect alternation of the hands. Consequently the fingers of
the same hand must be employed successively.
and'^jum'in " H ° w should the keys be struck with the fingers of the same
of Fingers hand ? Should the fingers be alternated, or should one finger
be " j u m p e d " ? Contrary to the prevailing belief, the answer
is not the same for the two hands; it is a curious fact that
the fingers of the more dexterous hand (usually the right) are
less independent. In this case, therefore, " j u m p i n g " is preferable, whereas the fingers of the left hand should be alternated.
The second fundamental rule of typing touch may therefore be
stated as follows : alternation of the fingers of the left hand and
" j u m p i n g " with those of the right promote speed.
SCIENTIFIC ARRANGEMENT OF THE KEYBOARD

The keyboards at present in use have been arranged on the
basis of statistics as to the frequency with which letters occur
in the words most used, the letters struck most frequently
being placed in the most accessible parts of the keyboard.
This principle is erroneous. The rule of alternation of the hands
proves that the statistics used should be those of the. letters
most frequently alternated and that these letters should then
be placed in opposite parts of the keyboard.
The space bar should no longer be regarded as a purely
accessory appliance, placed outside the keyboard because it

— 59 —
works neither a letter nor a sign. On the contrary, it should
be treated as an essential part of the keyboard; it should be
included in statistics of the frequency of alternation and placed
in the keyboard accordingly.
The rectangular form of the keyboard also requires investigation. It might be advisable to make it in the form of a curve,
more adapted to the fan-like arrangement of the fingers. Graphic
analysis of the movements involved in typing would make a
valuable contribution to the solution of the problem.
Typing methods are based on the number of fingers used. of ,¿!Ie,(
The theories formulated are generally based on certain facts
which, though not always in themselves inaccurate, suffer
from the common defect that they are based solely on observation uncontrolled by laboratory experiment. As general
rules have been based on a single set of observations, it follows
that they are not applicable to the majority of cases; hence
the controversies between the partisans of the 2, 4, 6, 8, or 10
finger methods.
It is a curious fact that one of these, the ten-finger method,
is based on erroneous observations, and yet enjoys the greatest
popularity and is the most frequently taught in typewriting
schools. The error involved is the analogy drawn between
typewriting and piano playing. The falsity of this analogy has
already been explained, and it is unnecessary to revert to it.
Graphic analysis of typewriting, as carried out above, leads
to conclusions incompatible with the ten-finger method.
According to physiological data, a scientific method should
favour alternation of the hands; but the use of all ten fingers,
in positions fixed with reference to each key, keeps the hands
more or less motionless over each half of the keyboard (the
fingers being used to strike the keys nearest their tips only)
and decreases the frequency of alternation of the hands.
The ten-finger method, having become the standard, is more
minutely regulated than any other. Consequently it is less
flexible and less adaptable to individual characteristics. The
rules of this system, which are based on reasoning not always
in harmony with the results of practical experience, cause
waste of time in typewriting. For example, the two letters
c and e, when situated in different rows of the keyboard, have

—

60 —

to be struck by "jumping " one finger. As shown in table XIV,
this is an error, and it is better to alternate the fingers, striking c
with the first finger and e with the second. It can also be shown
that in the combination br it is better to strike b with the right
first finger, and r with the left first finger, than to strike them
both with the right first finger, "jumping ".
One is therefore led to enquire whether the method which
most favours alternation of the hands, i. e. the two-finger
method, is not the best, as may logically be deduced from the
first rule established above. Here again, however, practical
experience is better than logic. It was shown that this method,
rigidly applied, would cause waste of time owing to the excessive
range of movement involved in striking two adjacent keys at
one end of the keyboard after one hand had been used to strike
a key at the other end.
These statements may seem to make any positive practical
conclusions impossible, but on the contrary they help to solve
the problem of teaching typewriting. The ten-finger method,
which is the least favourable to alternation of the hands, should
not be imposed on pupils, but neither should the opposite
two-finger method, for, though it gives the maximum of opportunity for alternating the hands, this advantage is limited by
the arrangement of the keyboard. Tt must also be admitted
that, though there are codified rules for the ten-finger method,
there are no principles for teaching the two-finger system;
this method is not taught, it is merely used by a few typists.
The 4, 6, and 8 finger methods might also be criticised in
the light of many experiments with individual cases, if their
rules were as well defined as those for the ten-finger method.
The impossibility of establishing a rigid method applicable to
all typists has led to the general conclusion that typewriting
touch is a personal matter, just like ordinary handwriting.
There are very few people whose handwriting is the same;
similarly typists are not built on identical lines, and cannot
therefore produce identical results with the same methods.
It has often been noticed, though hitherto insufficiently
emphasised, that typists taught by the ten-finger method
nearly always end by abandoning it. An enquiry conducted in
a large typewriting school showed that only one-tenth of the
pupils practised the ten-finger system, strictly though it was

— 61 —
taught. The teachers had to admit that after laborious training
each pupil adopted his own method.
Among typewriting champions are to be found some using
two fingers, such as C, others six fingers, such as Y, A, and B,
and very few, for example Z, ten fingers. These facts, however,
must be accepted with reserve until the cinematographic
investigations now in progress are concluded, as they alone can
determine accurately the precise methods used by champions
during record performances.
The adoption of one or other method for teaching typewriting
should depend on many factors. One of these is the anatomical
structure of the hands and fingers (the ten-finger method
presumably requires a large hand); the functional independence
of the muscle groups which determine the agility of the fingers
is also an important factor. But the essential factor in vocational
characteristics, in my opinion, is to be sought in the more delicate
anatomy, in the nervous system which determines motor
rapidity. Take, for instance, an excellent subject, E, who uses
the ten-finger method with great ease : the various tapping
tests which he performed gave the following results:

Fingers used

Speed

Liability.
to
fatigue

I'irst finger, right
First finaer, left
First fìnger, left and right
Left h a n d :
first and second fingers
first nnd third fingers
first and fourth fingers

6.75
6.15
8.95
5.71
6.15
7.G8

0.18
0.16
0.24
0.10
—
0.08

Right hand: first and second fingers
first and third fingers
first and lourth fingers

9.70
10.07
9.80

0.22
0.19
0.25

Aptitude
for
alternation

—
0.3.3
0.33
0.11
O.'M
0.11
—

It will be observed that the motor rapidity of this subject
is greater when the various fingers of the right hand are alternated with the first finger than when the hands are alternated.
This is not so with the left hand, but the superiority of the right
hand is enough to give this typist an advantage and confidence
in the method he uses. It should also be noted that, thougli his

— 62 —

speed is less when the fingers of the left hand are alternated,
the liability to fatigue is much less than with the right hand,
or even with alternation of the hands. E is therefore well
equipped for the ten-finger system.
When subjects are not so equipped, it is useless to compel
them to learn a method which they will not use in professional
practice.
Ea

shouidbbeCt Ifc d o e s n o t follow that teaching should be left entirely to
Taught the chance. To shorten the rperiod of trainingö and make it as
Method

.

.

effective as possible, the physiological characteristics of each
candidate should first be determined. Tapping tests will be
very useful here; the candidate's ability to alternate the various
fingers can easily be ascertained and will show which method
he should use. Everyone wishing to become a typist should
therefore undergo an examination of this kind. When candidates
have been selected by methods I have indicated elsewhere,
they should be advised to adopt the method best suited to their
natural abilities.
tfaTypis°t"
More exact knowledge of each typist will in future make it
to the Machine possible to decide not only what method he should be taught,
but also what machine he should use. As is well known, all
typewriters are not equally reliable or equally rapid. As the
greatest speed is attained by alternating the hands, it is simply
necessary to make the typist strike one key with one finger of
either hand alternately and as quickly as possible, and then to
note if the letter is typed for every stroke. When a typist
'jams ' a given machine, his motor rapidity is thus known by
means of this tapping test. By grading machines, as it were,
learners could be given the type of machine best suited to them
before they even began their training. The slowest subjects —
and they too may make good typists — will achieve the maximum speed of which they are physiologically capable with
slower machines.
My investigations, which show that educationalists should
adapt pedagogic rules to the individual conditions of the human
organism, provide data applicable to the teaching of any kind
of technique. To take one example: since alternation of the
hands should be encouraged, pupils should no longer be required
to treat the keyboard as divided in two portions, right and
Suited to Him

— 63 —
left, but to endeavour to alternate the hands in typing every
word. Mr. Dellion has already put this idea into practice in his
rational typing system.
Even the manufacture of typewriters should benefit from the
data supplied by these experiments. Study of the period of contact of the fingers with the keys shows that this contact should
be as brief as possible : for this purpose the .key should oñer the
minimum of resistance, unless it also guarantees regularity of
touch. Unlike the pianist, the typist should always strike the
key rather than press it.
The methods of applying the results of these investigations
to the improvement of typing technique are far from exhausted,
but my object is less ambitious. I have aimed rather to describe a method of studying occupational problems than to
supply definite answers to every question which arises. This
method may be defined as follows: every movement of the
occupation is analysed by graphic, photographic, or cinematographic experiment, and from this analysis are deduced the
best rules which workers should be advised to adopt in order
to obtain the maximum output from any given tool.
In conclusion one advantage of such a method may be indicated. It is easy to conceive the advantage which typewriting
schools would derive from the possession of an apparatus like
mine for investigating the touch of each of their pupils. The
defects revealed by the graphic method could at once be rectified, and teaching methods could, when necessary, be modified
to suit the physiological aptitudes thus revealed. The matter
is of the utmost importance for recruiting and teaching typists,
to judge by the number of these workers, the part they play
in business life, and the importance of the industries which
have been called into existence by typewriting.