dnd Vision

by AAred L. Yarbzls
Institute for Problems of Information Transmission
Academy of Sciences of the USSR, Moscow

Translated from Russian by

Cambridge, England

Basil Haigh

...' -.1 J,$

>.

,,'&

Translation Editor Lowin A. Riggs

L. Herbert BalIou University Professor
Brown University, Providence, Rhode Island

9PLENUM P R E S S . NEW YORK

1967

Al'fredLuk'yanovichYarbus was born in Moscow in 1914. He was graduated from the Faculty of Physics of Moscow University in 1941 and was a
scientificassistant at the Institute of Biophysics of the Academy of Sciences
of the USSR until 1963. He is ~resentlya senior scientist at the Institute
for Problems of Information Transmission of the Academy. In 1964 he received the degree of Doctor of Biological Sciences for his work on "The
Role of Eye Movements in Vision."

FOREWORD

The original Russian text was published for the Institute of Information
Transmission of the Academy of Sciences of the USSR by Nauka Press.
Moscow, in 1965.
Ansgjpea JZyxsaxoeuu Xp6yc
Porn

mend rnaa

npoqecee spennn

ROL' DVIZHENII GLAZ V PROTSESSE ZRENIYA

Library of Congress Catalog Card Number 66-19932

01967 Plenum Press

A Diuision of Plenum Publishing Corporation
227 West 17 Street, New York, N. Y. 10011
All rights reserved
No part of this publication may be reproduced in any
form without written perm~ssionfrom the publzsher
Printed in the United Stales of America

This book is primarily a monograph describing the original r e searches of the author. It is centered around a particular device,
originally described by Yarbus in 1956. This is an optical system in
miniature that can beattachedby suction t o t h e human eye. Many variations of the device-which we have called a "cap" f o r the sake of
brevity in this Eslglish edition-are described in detail in this book. One
f o r m of the cap is particularly useful in the recording of eye movements; this permits the use of a plane m i r r o r t o achieve an optical
lever that writes a continuous record of eye position on the moving
film of a photokymograph. Of thegreatest interest, however, a r e caps
that support an entire optical system. Attached by suction to the eyeball, such a system moves with i t and hence confronts the observer
with a stationary visual field. The consequence is a rapid fading o r
disappearance of the contours, colors, o r other features of the field
that the author has explored. Thus the book is concerned not only with
the recording of eye movements in all their various forms, but a l s o
with the consequences of eliminating the effects of eye movements on
visual perception.
A reading of this book cannot fail to i m p r e s s one with the magnitude
of the whole research project. Yarbus has shown extraordinary skill
and ingenuity in the construction of delicate optics of sufficiently light
weight t o be attached to the eye. The observer's task is not an easy
one. In most of the experiments the cornea is anesthetized, the lids
taped apart, and the subject trained to inhibit the natural tendency t o
move the eyes about. These precautions a r e necessary in o r d e r to
keep the suction device from colliding with the lids and thus inflicting
injury on the device o r on the eye itself. Although each experiment
was necessarily of limited duration, an impressive l i s t of topics was
explored. The wide coverage, indeed, testifies to the author's acquaintance with the significant problems of eye movements in their relation
to vision. In the field ofpsychology one would perhaps need t o go back

vi

FOREWORD

to the nonsense-syllable experiments of Ebbinghaus to find an experimenter so enthusiastically exploiting a new experimental technique.
In all fairness one must recogAize the limitations a s well a s the
strengths of Yarbus' work. In fact, the recording of eye movements
and the creation of a stabilized retinal image were achieved several
years earlier, a s Yarbus himself has acknowledged, inlaboratories in
America and England.
Plastic contact lenses were used in those
experiments. Poorly fitting lenses of this kind exhibit considerable
slippage over the eyeball and hence may indeed be inferior to Yarbus'
cap in experiments on eye movements and stabilized images, especially
if loaded down with relatively heavy optical equipment. While the
Yarbus cap has been used in at least one recent study in this country,
the majority of investigators now rely upon tightly fitting contact
lenses for their work. These havetheadvantage that they can be worn
with comfort for the duration of an ordinary experiment without anesthesia, taping of lids, or the risk of corneal abrasion. Furthermore,
recent evidence shows that when sufficient attention is given to the
experimental conditions, the degree of slippage of a well-fitted contact
lens is reduced to an amount too small to be of visual significance.
It may well turn out to be true that the Yarbus book has two lasting
merits: First, a s a stimulating account of what a single investigator
can achieve on the basis of an ingenious experimental technique, and
second, a s a rich collection of ideas and observations of visual phenomena that deserve to be explored by future investigators.
Lorrin A. Riggs
Providence, R. I.
June 1967

PREFACE TO THE AMERICAN EDITION

The author is pleased to have the opportunity to acquaint the reader
with the results of his work, the conclusions of which a r e of fundamental importance to the understanding of certain mechanisms of
vision.
The results of this work a r e largely due to an original
technique, described fully in the book, employing suction devices, or
"caps." In the author's opinion, this technique is suitable for use in
studying a wide range of phenomena.
It would be a source of great satisfaction if this technique were to
be adopted in the research laboratory and new and interesting results
obtained by its use. It will, of course, be realized that experiments
with "caps" a r e rather complex. They require great care and careful
preparation in each case. Often the construction of the lenses and the
accessories will have to be modified. A jeweller is required to make
the very small details of the lenses and to assemble them, and of
course this introduces considerable difficulty. However, the author
knows of no easier method by which results similar to those described
in this book can be obtained.
A. L. Yarbus
March 1967

PREFACE
This book deals with the perception of images which a r e strictly
stationary relative to the retina, the principles governing human eye
movements, and the study of their role in the process of vision.
The book is based on the results of the author's experimental
investigations.
It is intended for students and researchers in the fields of biophysics, physiology, medicine, psychology, and branches of technology
such a s television, motion pictures, and apparatus construction.
Much attention is devoted to the description of methods for recording
eye movements and methods of producing images stationary relative
to the retina. These methods may be of interest to many scientific
workers.
The investigations were carried out in thelaboratory of biophysics
of vision of the Institute of Biophysics of the Academy of Sciences
USSR, where they werediscussed by a11the staff. The author would like
to emphasize particularly the help he has received for several years
from N.D. Nyuberg and L.I. Seletskaya. The valuable advice of M. M.
Bongard, A. L. Byzov, and M. S. Smirnov was frequently sought during
these studies.
Substantial technical assistance was given by V.I.
Chernyshov, V. M. Timofeeva, P. N. Efimova, and EN. Salina.
The author wishes to take this opportunity to express his sincere
gratitude toall these colleagues and to the entire staff of the laboratory.

CONTENTS
INTRODUCTION

................................

CHAPTER I Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 Elementary Facts Concerning the Structure of the Human
Eye
2 Study of Eye Movements by Means of After.Irnages
3 Determining the Magnitude of Involuntary Eye Movements
During the Perception of Small Objects
4 Early Methods of Studying Perception of Objects Stationa r y Relative to the Retina
5 Study of Eye Movements by Visual Observation
6 Mechanical Recording of Eye Movements
7 Recording Eye Movements by a Reflected Beam of Light
8 Study of Eye Movements by Still and Motion-Picture Photography
9 Recording the Corneal Bright Spot
10 Electrooculography
11 Some Photoelectric Methods for Recording Eye Movements
12 Creation o f a Stabilized Retinal Image with a Contact Lens
13. Suction Devices (Caps)
14 Apparatus Used in Work with Caps
15 Technique of Experiments with Caps . . . . . . . . . . . . . . .
16 Course of the Rays in Recording Eye Movements by a
Reflected Beam of Light
17 Construction of Caps . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.
.

....................................

.....

............
.....................
........
............

.
.
.

.
..
.
.
.
.
.
.

.................................
................
..........................

...................................
........................
................
......................

CHAPTER I1 Perception of Objects Stationary Relative to the

Retina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 Formation of an Empty Field . . . . . . . . . . . . . . . . . . . .

CONTENTS

xii

.
.

2 . Perception of Very Bright Objects Stationary Relative to

the Retina
3 Perception of Objects of Varying Luminance Stationary
Relative to the Retina
4 Perception of Flickering Objects Stationary Relative to
theRetina
5 Perception of Objects of Changing Color. Stationary Relative to the Retina
6 Changes in the State of the Retina after Formation of an
Empty Field
7 Perception of Objects Stationary Relative to the Retina
and Occupying Part of the Visual Field
8 Delay in Seeing the Color of an Empty Field
9 Spatial Development of the Formation of an Empty Field
10 Perception of Objects Moving on an Empty Field
11 Role of Illumination of the Eye with Scattered Light
Conclusions

........................
................................
...........................

CHAPTER VI Eye Movements During Perception of Moving

Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 Involuntary Aspects of Pursuit of Moving Objects . . . . . .
2 Minimal and Maximal Velocities of Pursuit . . . . . . . . . .
3 . Eye Movements During Pursuit in Complex Conditions . . .
4 . Pursuit Accompanied by Convergence and Divergence of
the Optical Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.
.

..............................

............
.........
.
.
.......
.
.....
..................................
CHAPTER 111. Eye Movements During Fixation on Stationary
Objects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Drift of the Optical Axes During Fixation . . . . . . . . . . . .

..

CHAPTER VII Eye Movements During Perception of Complex

Objects
1 Eye Movements During Perception of Complex Objects
2 Eye Movements in Reading
3 Role of Eye Movements in Assessing Spatial Relationships
4 Optical Illusions and Eye Movements
5 Eye Movements and Perception of Movements
6 Role of Recognition in Evaluation of Spatial Relationships
Conclusions

.
.
.
.
.
.

2 Tremor of the Eyes .

.........................
3 . Small Involuntary Saccades of the Eyes . . . . . . . . . . . . .
4 . Fixation on a Point in Complex Conditions . . . . . . . . . . .
5. Nystagmus of the Eyes . . . . . . . . . . . . . . . . . . . . . . . .
6 . Factors Influencing the Movement and Contrast of the

.
.
.
.
.

.....................
...................................
..............
........
.
.................................

.............................
INDEX .....................................

................
.......................

...............

........................
....................
.................
......................
...............
.
.................................
CHAPTER V. Eye Movements During Change of Stationary
Points of Fixation in Space.....................
1. Change of Stationary Points of Fixation in Space .......
2 . Duration of Convergence and Divergence ............

..................................
..

LITERATURECITED

.............................
.......................
. . . . . . . . . . . . . . . . . . . . . . . . .'. . . . . . .

RetinalImage
7 . Role of Eye Movements
Conclusions
,

CHAPTER IV Saccadic Eye Movements

1 Duration of the Saccade
2 Development of the Saccade in Time
3 Inclmation of Saccades
4 Center of Rotation of the Eye
5 Begqning and End of the Saccade
6 Vision During the Saccade
7 Voluntary and Involuntary Saccades
Conclusions

......
.....................

3 Convergence and Divergence Preceding a Saccade

4 Scheme of Eye Movements During Change of Stationary
Points of Fixation in Space
5. Apparent Size of Object and Direction of Gaze . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.................................

.
.
.
.

CONTENTS

147
147
151

Recent developments have shown that our earlier ideas of the role
of eye movement were considerably oversimplified. As the following
facts will show, the subject a s a whole is far more complex.
In man under natural conditions the retinal image is never
tionary relative to the retina, and if a strictly stationary and uning retinal image is created artificially, the eye ceases t o see.
other words, within any- obj'ect of perception remaining strictly
ionary relative to the retina and unchanging in time, after about
sec all visual contours disappear (the resolving power of the eye
idly falls to zero).
t has long been k n o w that an observer begins to s e e the blood
els lying on the retina of his own eye when conditions a r e created
hich the shadows of the vessels acquire some degree of mobility.
riments have shown that for the conditions of perception to be
, slight but not excessive continuous o r interrupted movement
e retinal image over the retina is necessary, a s a result of which
1ight.acting on the receptors is constantly changing.
Electrophysiological studies have shown that a s a rule impulses
a r in the optic nerve of many animals only in response to a change
the light acting on the retina.
ese facts have compelled a r e a s s e s s m e n t of the role of e y e ,I
ents and have demonstrated that without an understanding of
ole it is impossible to decipher the mechanisms of vision. For
reason it has become necessary, on the one hand, to study in
il the perception of images strictly stationary relative to the
and changing in brightness or color and, on the other, to study
rious forms of eye movements.
was originally considered that the main function of the eye
ents is to retain the object of perception in the visualfield (to
the element of the object important for perception in the fovea)
to change the points of fixation, thereby widening the total angle
1

INTRODUCTION

of view. Movements of the eyes preventingthe disappearance of visual

contours in the stationary object in the process of fixation now appear
to be no less important.
The second chapter is the most important in the book. Concerned
with the perception of objects stationary relative to the retina, it
provides a new approach to certain sections of physiological optics,
and helps to establish a number of connections and analogies between
electrophysiological studies carried out on the retina in animals and
studies of human vision.
The third chapter discusses the micromovements of the eyes
accompanying the process of fixation directed towards a stationary
object. This chapter explains how the micromovements of the eyes
in ordinary conditions of perception prevent the disappearance of
differences in the object during fixation. Changing the points of
fixation, convergence and divergence of the optical axes, the pursuit
of moving objects, and some cases of assessment of spatial relation' ships a r e accompanied by macromovements of the eyes.
Chapters IV through VII explain the role of, and the principles
governing, macromovements of the eyes. Without an understanding of
the role and knowledge of the principles of not only the micromovements, but also the macromovements of the eyes, the work of the
We shall refer to eye
retina cannot be completely understood.
movements also when we speak about the structure of the eye and
when we examine the special features of perception in man. For
example, in phylogenesis the mobility of the head and eyes of some
animals made possible the appearance of a fovea and introduced
considerable refinement into the process of vision. These refinements
a r e important because objects providing essential information a r e by
no means uniformly distributed. Usually they a r e localized in small
areas of the field of vision. In these circumstances the peripheral
portion of the retina usually finds the object or element of an object
which contains or may contain essential information, and consequently
a process resembling reconnaissance takes place; this information is
perceived and analyzed in g r e a t e r detail by means of the foveal part
of the retina, when directed towards the object.
The f i r s t chapter describes various methods of studying and
recording eye movements, and methods for stabilizing the retinal
image.
When understood., the role of eye movements and the principles
governing these movements may help to solve many purely practical
problems. Functional disturbances of the central nervous Hystem a r e

INTRODUCTION

often accompanied by disturbances of various eye movements. The

centers controlling the eye movements and the pathways joining these
centers to the eye muscles a r e located in various parts of the brain
and a r e often adversely affected by pathological foci sintated near these
centers. The same is also true of disturbances in the working of the
auxiliary systems closely connected with eye movements.
Any disturbances of the visual system in diseases of the central
nervous system may help to determine both the character of the
disease and the localization of the pathologic focus. However, it is
not always simple to detect functional abnormality in an organ a s
complicated a s the eye. It f i r s t becomes necessary to know what is
normal. With respect to eye movements, this problem is far from
solution.
It is only in recent years that important data have been
obtained in this field, and these data have not yet reached the attention
of a wide range of readers for whom they would be of considerable
interest. It should be emphasized here thatdisturbances of a patient's
eye movement can be recorded objectively, and that this procedure is
particularly applicable for diagnostic purposes.
Knowledge of the principles governing eye movement inthe normal
subject may be useful also in ophthalmology. Unfortunately, no
systematic redords of eye movements have yet beenmade in the various
forms of strabismus, paralyses, and pareses of the eye muscles. It
is by no means impossible, for example, that such records could be
used to diagnose and to distinguish objectively the treatable and untreatable forms of strabismus.
Familiarity with the perception of objects stationary relative t o the
retina and changing in color or brightness undoubtedly is useful to the
neurologist and ophthalmologist. Disturbances of this perception in
patients may also give useful information regarding the character of
a disease of the central nervous system or of the eye.
In many cases when an investigator is interested'in the problem of
perception of complex objects (in normal and pathologic conditions),
records of the eye movements would be valuable. In these circumstances, by using such records itwouldbeeasy to determine the order
in which an object was examined, what elements were fixated by the
subject, how often and for how long a particular element was fixated,
and s o on. Records of eye movements illustrate the course of the
process of perception.
Knowledge of the principles governing eye movements and the
properties of perception of images stationary relative to the retina
may be used (and sometimes is used) in motion pictures and television,

"

INTRODUCTION

in apparatus construction, for the rational arrangement of instruments

on panels, for evaluating the possibilities of perception in complex
conditions, and s o on.

Chapter I

METHODS
1. ELEMENTARY FACTS CONCERNING THE STRUCTURE
OF THE HUMAN EYE*
The outer layer of the eye (Fig. 1)is formed of a tough membrane,
the sclera, consisting of firm connective tissue continuous in its
anterior part with a transparent membrane, the cornea. The sclera
enables the eye to maintain a constant shapeand protects its contents.
The same function is served by the cornea, which is also part of the
dioptric apparatus of the eye. The membranes of the eye a r e under a
certain intraocular pressure. The normal intraocular pressure lies
within the limits of 15-30 m m Hg.
The dioptric apparatus of the eye, which takes part in the formation ,
of an image on its inner surface, consists of the cornea, the biconvex
transparent lens, the transparent aqueous humor, and the transparent
vitreous, filling the eye. This apparatus also includes the ciliary body,
which allows changes to be made (by means of the ciliary muscle) in
the curvature of the lens surfaces (accommodation)andthe iris, which
modifies the diameter of the pupil (the aperture of the diaphragm).
Accommodation permits the image to be sharply focused. A change in
the size of the pupil leads to changes-in the retinal illuminance and
the depth of focus of the optical system. The optical system gives an
inverted real image of objects in front of the eye.
The diameter of the average eye in allmeridians is approximately
24 mm.
Beneath the sclera lies the vascular membrane, consisting of a
network of blood vessels supplying the eye. Next to the inner surface
of the vascular membrane lies the pigmented epithelium, containing
'Detailed information concerning the structure and function of the eye can be found in the
works of Averbakh (1940). Kravkav (1950). Tonkov (1946), and Polyak (1941).
5

CHAPTER I

Frg. I. Scl>?rneof JI lrorrzanral scctton oi the laurnan rghr cyc. 1) Sclsra; 2) cornea; 3) lens;
chlmbcr of the eye:$) v~treous:6) ms: 7) c i b r y muscle; a) conjuncrira; 9, point
nt arrarhmcnr of
recrus
101
.
. rncdl.1 .
~ muscle:
~ - .
, oornr of attachment o: lateral recNs muscle:

4 ) anterior

--

~ ~ ~. - ~ - ~ .
~~~

11) visualaxis ofthe eye; 12) opticalaxis of eye; 13) retina; 14) vascular membrane; 15) fovea
cenrralis; 16) optic nerve.

Fig. 2. R d s and cones of the retina. Left, a rod: 1) outer segment; 2) ellipsoid; 3) inner
segment; 4) outer limiting membrane; 5) rod fiher; 6) nucleus: 7) Terminal bouton. Right. a
cone:
1) outer segment; 2) ellipsoid; 3) inner segment; 4) outer limit= membrane:
5) nucleus; 6) cone fiber; 7)cone foot-plate (Averbakh. 1940).

METHODS

a dark pigment. Beyond the layer of pigmented epithelium is the innermost layer which directly receives the photic stimuli, the retina.
Schematically, the retina can be divided into two zones: a photosensitive zone facing the vascular membrane, and a neural zone, facing
the vitreous. The thickness of the retina in its central part, the macula
lutea, is about 0.1 mm.
The photosensitive cells of the retina (receptors) a r e the rods and
cones (Fig. 2). The rods a r e much more sensitive to light than the
cones. At very low levels of illumination only the rods function, and
they a r e associated with the mechanism of twilight vision. The rods
contain a photosensitive pigment with a maximum spectral sensitivity
a t a wavelength of 510 mi/ The cones contain three photosensitive
pigments with spectral sensitivity maxima a t wavelengths of 440, 540,
and 590 mp (Fig. 3). The cones a r e associated with the mechanism of
color vision.
T h e point of entry of the optic nerve into the eye has no rods or
cones. We cannot see with this part of the retina, and it is therefore
called the blind spot. The point of clearest vision is the macula lutea.
It lies on the temporal side, slightly above the point of entry of the
optic nerve (Fig. 1). The macula lutea is yellow and occupied mainly
by cones. The angular dimension of the macula lutea is approximately
6-7". Within the macula lutea lies the fovea centralis, the part of the
retina with the highest resolving power. The diameter of the fovea
centralis is about 0.4 mm, i.e., about 1.3'. The middle part of the
fovea centralis (the foveola) is pigmented less than the other parts
of the macula lutea. The fovea centralis is slightly displaced from the ,
optical axis of the eye.
In the human retina there a r e about 130 million rads and about 7
million cones. The distribution of the rods and cones in the r e t h
is indicated in Figs. 4 and 5. The diameter of the inner segment of

"'!J
Flg. 3. Curves of sensstiv~tyofhu-ndayhghreceptms-cones

(Bangard and Sm~mov.1955).

CHAPTER 1

2000

1500
I000
500

'?z 20 08 16 14 (2 I0 8 6 4 2 0 2 4 6 8 10 12 14 18
Fig. 4 . Distribution of rods and cones in the retina. Abscissa-the distance (in mm) from
the middle of the fovea centralis (the foveola) along the horizontal s e a i o n of the right eye.
Ordinate-the number of hundreds of rods and cones per mm2. The broken line represents
rods and the solid line cones (Oesterberg,1935).

the rod is about 0.002 mm. The diameter of the inner segment of the
cones varies with the position in the retina from approximately
0.002 to 0.007 mm. The cones in the central part of the retina, where
the distance between the centers of the cones is about 0.0025 mm
(0.5 minute of angle), a r e longer and thinner than in the peripheral
part. The resolving power of the eye is maximal a t the fovea centralis
and gradually diminishes towards the periphery. The relative visual
acuity in relation to the position of the image on the retina is shown in
Fig. 6.
The structure of the retina is very complex (Fig. 7). Morphologists
(Polyak, 1941) have distinguished the following 10 layers in the retina:
1, the pigmented epithelium; 2, the layer of outer and inner segments
of rods and cones; 3, the outer limiting membrane, intersected by
the rods and cones; 4, the outer nuclear layer containing the nuclei
and fibers of the rods and cones; 5, the outer plexiform layer; formed
by a plexus of endings of the photoreceptors with the fibers of neurons
of the next layer; 6, theinnernuclear layer of bipolar cells, horizontal
and amacrine; 7 , the inner plexiform layer, consisting of a plexus of
endings of the neurons of layer 6 with the endings of the ganglion cells;
8, the layer of ganglion cells; 9, the layer of fibers of the optic nerve;
and 10, the inner limiting membrane. The bipolar cells a r e of several
types, differing in morphological structure and inmode of communica-

METHODS

tion with other neurons. The same is true of the ganglion cells. The
final centers of vision a r e the occipital lobes of the cerebral cortex,
on both lips of the calcarine fissure.
The part of the optic pathway from the eye to the chiasma (the
point of partial decussation of the optic nerves) is called the optic
nerve. The optic nerve, about 5 cm long and about 4 mm2 in cross
section consists of approximately 1million nerve fibers. There are,
on the average, about 150rods andcones to each fiber. At the chiasma
the optic nerve divides into two parts (Fig. 8). The fibers running
from the nasal half of the retina proceed to the opposite cerebral
hemisphere; fibers arising from the temporal half of the retina
proceed to the hemisphere on the same side. Therefore there is an
incomplete decussation of the optic nerves in the chiasma. The optic
nerve fibers forming the optic tract then run to the subcortical visual
centers (the pulvinar of the thalamus,theanterior colliculi, the lateral
geniculate body). From the intermediate centers, nerve fibers known
a s Gratiolet's fibers run to the terminal visual centers. After reaching
the geniculate body, some fibers continue tothe temporal region of the
brain.
Injuries to the brain and corresponding disturbances of the
visual fields of the eye have demonstrated the connections between
various parts of the retina and the cerebral cortex. This projection
of the retina on the cortex is illustrated in Fig. 9.
The position of the eyes in the orbits is shown schematically in
Fig. 10. Under normal conditions eyemovements cause practically no
displacement of the center of the eye relative to the orbit. All
movements of the eye amount to its rotation about a certain center
lying inside the eye on the optical axis. The distance between the
,LO

1.0

.-

20 0.8

0.8

?i, 0.6

0.8

.-YI

'w 04

0.Y

'Z Llz

Llz

>

'
0

Fig. 5. Relative distribution of rcds andcones in the retina. 1) Mosaic of cones in the center
of the fovea; 2) rods (small dots) andcones (large dots) 0.8 mm from the center of the fovea:
3) rods and cones 3 m m from the center af the fovea (Averbakh, 1940).

60

40

20

20

40

Distance from center of

fovea, degrees

Z0

40

60

Distance from
center of fovea,
minutes of angle

Fig. 6. Relative visual acuity depending on the position of the retinal image of the retina
(Jones and Hlggins. 1947).

CHAPTER I
(---------

METHODS

11

muscles possess the thinnest fibers. The eye muscles a r e very

profusely innervated by motor and sensory nerve fibers (Duke-Elder,
1932; Fulton, 1943). The eye muscles a r e innervated by the oculomotor, trochlear, and abducens nerves. The trochlear nerve supplies
the superior oblique muscle and the abducens nerve the lateral rectus.
The oculomotor nerve innervates all the other muscles of the eye,
including the ciliary muscle and the muscles responsible for changing
the size of the pupil. All these nerves arise in the lower part of the
brain in the floor of the 4th ventricle, in the region of the corpora
quadrigemina, pons, and medulla.

2. STUDY OF EYE MOVEMENTS BY MEANS OF

AFTER-IMAGES
Modem methods of recording eye movements and of creating a
stabilized retinal image a r e by no means perfect. Frequently, when
new methods a r e being developed, long established methods a r e used.
In some cases, even complicated problems can be solved by the use
of "forgotten" techniques. This suggests thatashort historical review
of methods related to this theme would be worth while.
Flg. 7. Scheme showing the structure of the retma. I, 11, and 111. Fxrst, second, and third
neurons. A sustentacular fkber of Miiller is shown on the rlght (Kravkov, 1950).

apex of the cornea and the center of rotation of the eye is approximately
14.5 mm. Rotation of the eye aroundthis center is performed by three
pairs of muscles (Fig. 11). These muscles (named in accordance with
their position) a r e : the lateralandmedial rectus muscles; the superior
and inferior rectus muscles; and the superior and inferior oblique
muscles. The four rectus muscles a r i s e by tendons in the depths of
the orbit. They a r e all attached to the eye several millimeters from
the edge of the cornea. The inferior oblique muscle runs from the
anterior part of the orbit laterally and winds around the.eye to which
it is attached posteriorly. The superior oblique muscle arises in the
depth of the orbit, runs forward, passing over a special pulley, turns
posteriorly and laterally, and is attached to the postero-superior part
of the eye. The intermediate space between the eye and its orbit is
filled with orbital fat, on which the eye rests. In addition, the eye is
maintained in position by special ligaments.
The work o f ' the muscles during rotation of the eye is fairly
complex. The action of the individual eye muscles is shown schematically in Fig. 12. Of all the voluntary muscles in the body, the eye

Fig. 8. Scheme showing the optic pathways and centers. 1) Field of vision: 2) cornea;
3) retina: 4) chiasma; 5) subcortical visual centers; 6) Gratiolet's fibers: 7) visual cortex.
k f e c t s arising in the visual field after injuries to the optic pathways are indicated on the
right. The blind a n a inthevisualfield i s shaded. The poim of injury i s denoted by a stroke
and letter on the figure on the left (Kravkov, 1950).

Fig. 11. Scheme showing the muscles of the eye. 1) Superior rectus muscle: 2) inferior
rectus muscle; 3) Lateral rectus muscle (medial rectus lies symmetrically opposlte bur
cannot be seen on the drawing); 4) superior oblique muscle; 5) inferior oblique muscle.

Several experimenters (Dodge, 1901;Helmholtz. 1925; Duke-Elder,

1932; Barlow, 1952; andothers) have studied the character of eye movements by the use of after-images. Eyemovements were studied in the

Fig. 9. PraFcnon of visual f ~ e l don the cerebral cortex. The numbers denore degrees
(Holmes. 1918).

process of fixation, in the process of changing points of fixation, and

during examination of complexobjects. With the introduction of modern
flash lamps, this method of producing after-images has become much
more refined. The blinding brightness and short duration of the flash
(less than 0.001 sec) make it possible to produce long after-images of
great sharpness.
Let us examine a well known method of observing the movement of
the eye itself in the process of fixation. The experimenter produces
a clear after-image (reference mark) projected on the fovea and
shaped like a cross, a line, or a small triangle. The observer then
fixates on a point on a screen, which is either a sheet of graph paper
or paper on which a grid has been drawn. During fixation, the observer at the same time watches the movement of the reference mark
relative to the point of fixation and to the grid of the screen and notes
the trajectory followed by the mark during a particular period of
time.
o. inf

+-,

Right eye

Fig. 10. Scheme showing the positions of the eyes in the orbits (Duke-Elder, 1932).

Left eye

Fig. 12. Direction of action of the individual eye muscles. Broken line-veriical meridian
of the eye; black circle-pupil: r. at.-lateral rectusmuscle: r. in[.-medial rectus muscle;
r. sup.-superior
reetus muscle; r. inf.-inferior recNs muscle: 0. sup.-superior oblique
muscle: o. id.-inferior oblique muscle (Kravkov. 1950).

"

14

CHAPTER I

Since the after-image is strictly stationary relative to the retina,

the apparent movement of this image on the screen corresponds directly
to the movements of the eye. Knowing the distance between the eye
and screen, it is easy to calculatethe size of the division on the grid
of the screen in angular values, and to determine the eye movements
performed during fixation with roughaccuracy. Inthis case the sharpness of the after-image is very important. The smaller the reference
mark and the sharper it appears to the eye, obviously the greater the
accuracy with which its movement on the screen can be determined.
Another method of studying eye movements in the process of
fixation i s a s follows. The observer fixates on a point placed a t the
center of a narrow slit. A flashlamp is placed behind each half of the
slit. The flash lamps a r e switched on consecutively a t a predetermined
interval, equal to some fraction of a second. In these circumstances
after-images appear from each half of the slit. As a result of the eye
movements in the interval between two flashes, the two halves of the
after-image usually appear displaced relative to each other. From
the magnitude of this displacement, the experimenter can judge the
magnitude and character of the eye movement during fixation.
Observations on the large eye movements during examination of an
object can be conveniently made by means of an after-image in the
shape of a circle projected on the fovea. Knowing the angular size
of the circle, and memorizing its positiononthe object during perception, the observer can judge approximately with which parts of the
r e t i k he looks a t a particular element of the object and what information is oljtained in this process.
The general character of eye movements during examination of
stationary objects, the jerkiness of these movements in particular,
has been studied in the past (Landolt, 1891) a s follows. The observer
sits in a dark room and a weakly illuminated object is placed in his
field of vision, while against this background, oralongside it, is placed
a small but very bright source of light. The observer looks a t the
object or traces its outline with his eye for a certain period of time,
'after which 'the light is switched off. From the series of after-images
produced by the bright source of light, the observer can judge the
character of the eye movements. Each separate after-image corresponds t o a point of fixation; each interval between two neighboring
points of fixation corresponds to a change of the points of fixation.
Eye movements during a change in points of fixation have been
studied by several authors (Lamansky, 1869; Cobb and Moss, 1925) by
means of a bright flashing source of light. The light source, flashing
a t a frequency of several hundred cycles per second was placed

METHODS

15

between two points of fixation. During change in the points of fixation,

a s a result of the flashes of the source of light and the eye movements,
a series of images of the source of light was obtained on the retina;
subsequently, after the light was extinguished, these were perceived
by the observer as a chainofafter-images. Since the observer always
knew the flash frequency, the number of after-images and the angle
between the points of fixation, he could determine the time of alternation of the points of fixation.
Several authors (Loring, 1915) also used after-images to find
rotary movement of the eye around the optical axis. Here, the afterimage (a cross) was projected by the observer on a screen with a
grid of horizontal and vertical lines. The observer chose the point
of fixation and put his head ina pasition in which the lines of the cross
were parallel to the lines of the grid. He then performed a movement
with his eye and remembered when and approximately a t which angle
the lines of the cross changed their direction relative to the grid on
the screen. The visible rotationof the cross relative to its own center
corresponded to the rotation of the eye relative to the axis of vision.
If an observer looks a t a very bright white field through a red
filter, and the red filter is then changed f o r a violet (or the whole
procedure is performed in the opposite order), each time, shortly
after the change of filter, he can s e e the fovea of the opposite eye in
the form of a tiny star. By fixating on a point on the briglit field and
memorizing the movement of the fovea, in these circumstances, the
observer can obtain some idea of the eye movements during fixation.
As a result of adaptation, prolonged fixation onany object composed
of contrasting elements separated by sharp boundaries leads to
appreciable diminution of the sharpness of the visible differences in
color between the elements. Under these circumstances, however, a s
a result of small eye movements, sharp bands appear a t the border
between the elements of the object. Fromthe way imwhich these bands
appear and from their width, some, idea may also be obtained of the
eye movements during fixation.
Today eye movements a r e rarely studied by means of after-images
because more refined methods have been developed. This does not
mean, however, that after-images have now served their purpose a s a
method of investigation and will never again be used in laboratory
practice. Let me give a few examples.
First, after-images may he used whenever an experimenter is
interested in the perception of objects stationary relative to the
retina (after-images a r e always stationary relative to the retina), for
example, the perception of the shape or proportions of an object in

16

CHAPTER l

conditions when the subject cannotuseeyemovement,or the perception

of optical illusions in the same conditions, and s o on.
If an intracranial tumor has deprived a patient of half his field of
vision (hemianopsia), several autliors have suggested that a 'pseudofovea"' may appear within the functioningpartof the retina of each eye.
To localize this point, and to trace its evolution, it is best to use
after-images.
An experiment in this case could be conceived a s
follows: the patient i s asked to fixate on the center of a geometrical
figure (for example, several concentric circles). Next, by means of
a flash lamp the patient is made to develop an after-image of this
shape. Naturally, the patient sees only part of the shape a s the afterimage-the part on the functioning half of the retina a t the time of
fixation on the center of the figure. If the patient fixated by the use of
the pseudo-fovea, this part would be greater than half and the corresponding difference (the difference between the visible part and half
the figure, measured in angular units) would showhow far the pseudofovea was shifted relative to the center of the retina.
The apparent size of the after-image, like the apparent size of the
real object, is determined by severalfactors,andabove all by the size
of the retinal image. It changes appreciably with a change in convergence and accommodation, and with considerable deflection of the
gaze up or down. Since the retinal image corresponding to the afterimage remains strictly unchanged the whole time, the after-image is
a convenient test for the influence of various factors accompanying
perception and changing the apparent size.

3. DETERMINING THE MAGNITUDE O F INVOLUNTARY

EYE MOVEMENTS DURING THE P E R C E P T I O N O F SMALL
OBJECTS
When examining an object, determining its proportions, counting
the elements of an object, and s o on, we usually use eye movements
and voluntarily change the points of fixation. It has been found that
if an observer cannot use voluntary eye movements, the solution of
certain problems of perception (determination of proportions, comparison of areas, counting large numbers of small elements, and s o
on) becomes difficult, even when the corresponding retinal image is
completely within the .fovea and, consequently, all elements of the
object a r e clearly seen. There a r e anumber of methods by which eye
movements can be excluded artificially during perception. Under
certain conditions, a similar situation may even arise in a normal
subject.
During fixation directed towards a stationary object, the

METHODS

17

human eyes involuntarily perform small, jerky movements. If the

angular dimensions of the object a r e smaller than these movements,
the observer cannot make out the various elements separately, using
voluntary eye movements. The solution of these problems under such
conditions presents considerable difficulties. The fact that such
difficulties can arise has been used by several workers to determine
the magnitude of the involuntary eye movements during fixation.
Landolt (1891), in an attempt to determine the smallest angle of the
voluntary eye movement, i n s t ~ c t e da subject t o count a number of
vertical lines forming a regular series. The lines were placed f a r
enough away from the subject so that he could not count them but near
enough so. that be could clearly distinguish them. Landolt believed
that the difficulty experienced by the subject in counting the lines
corresponded to a situation in which voluntary eye movements were
impossible. This method cannot, of course, compete with modern
methods allowing eye movements to be recorded objectively.

4. EARLY METHODS O F STUDYING PERCEPTION O F

OBJECTS STATIONARY RELATIVE T O THE RETINA

In 1804, Troxler found that objects visible to the periphery of the

eye disappear during careful fixation onacertainpoint. This phenomenon was called Troxler's effect (Clarke, 1960). The first correct
interpretation of this effect was evidently given by Adrian (1928). By
using the results of electrophysiology (in most animals, impulses
appear in the optic nerve only in response to a change in the light
acting on the retina), Adrian suggested that the human eye also stops
working under conditions where the retinal image is unchanged and
stationary.
Adrian himself tried by very careful fixation to obtain
disappearance of visible differences within an object, on one point of
which he fixated. In certain conditions he could do this for a short
time, but these experiments were notvery convincing. However, there
a r e several well known methods by which this property of human vision
may be illustrated.
We know that large blood vessels run along the inner surface of
the retina. Although these a r e not usually seen, conditions can easily
be created in which the shadows of these vessels become visible on
the retina. If these shadows continually change their size or position,
they will be continually and clearly seen by the observer. Of the
experiments of this series, the following is the most effective
procedure.

18

CHAPTER I

An observer in a dark room takes a bright-point source of light in,

let us say, his right hand and holds it in front of him below the level
of the eyes. With his left hand, he covers his left eye and with his
right eye he looks a t a large dark screen o r a t the wall. The observer
then makes a continuous waving movement with his right hand during
which the moving source of light is always visihle t o the eye a t the
extreme periphery of the retina.
Under these conditions sharp
outlines of the eye vessels appear, and they seem to the observer to
be projected on the screen. As soonas the observer stops moving the
source of light (i.e., the movement of the shadows of the vessels), the
vessels disappear within 1-2 seconds, but reappear whenmovement of
the light source is resumed.
I will mention one further method by whichan observer can see the
vessels of his eye. An opaque diaphragm with a very small aperture
(for example, a piece of blackpaperpiercedby a fine needle) is placed
in front of the eye, near the cornea. m e observer looks through the
aperture at a bright screen and at the same time gives the diaphragm
an oscillatory movement of small amplitude. The movement of the
aperture in front of thepupil causes excursions of the rays of light over
the retina, and this inturncausesmovementof the shadows of the vessels and their appearance in the field of vision. The vessels a r e seen
much better if the pupil is artificially dilated before the experiment.
Even with the most careful fixation on a point, it is impossible to
obtain the complete disappearance of the visual contours, usually
because of involuntary movements of the eyes and head. However, there
i s one very simple experiment which demonstrates this possibility.
The observer sits on a chairnearalamp and sticks a few small pieces
of paper of different colors to the right half of his nose, to that part
which can easily be seen with the right eye. He then rests his elbows
on the chair a r m s , covers the left eye with theleft hand, and uses the
right hand a s a chin rest. He then chooses a point of fixation on the
chair s o that it lies next to the pieces of colored paper sticking to his
nose. At a certain moment,afterashort period of fixation, the visible
part of the nose and the brightly illuminated pieces of paper completely
lose their color and appear a s a uniform dark grey field. A slight
movement of the eye instantaneously restores the disappearing differences. The relative ease with which in this case success can be
obtained is explained by the fact that, because of the closeness of the
pieces of paper to. the eye, the retinal image is out of focus, and its
outlines a r e not sharp. Under these conditions, small movements of
the eyes a r e less important. The methods described can supply valuable illustrations when demonstrating the work of the eye.

A general idea of the character of the eye movements can be

obtained by direct visual observation of the eye. Some authors Uaval,
1879) used a mirror for this purpose. Observations were made on the
image of the eye in a mirror. The experimenter stood behind the
subject and did not distract him duringtheexperiment. However, with
the unaided eye, the experimenter could observe only large movements.
Rotation of the eye through I", and the corresponding movement of the
retina through 0.2 mm couldnotbenoticedby the experimenter. Later
optical instruments were used to provide a magnified image of the eye,
or of part of it, thus increasing the accuracy of the method. Lenses
giving the required magnification were used in the study of relatively
large eye movements (Newhall, 1928). and a microscope was used in
the study of small eye movements or movements duringfixation
(Gassovskii and Nikol'skaya, 1941). In the second case the experimenter made observations with the microscope on the movement of a
bifurcation of the blood vessels..
Since the ready-made optical instruments were not always satisfactory for these purposes, specially designed instruments were
created. Some authors (George, Toren, and Lowell, 1923; Park and
Park, 1933) studied the position of the center of rotation of the eye in
relation to the direction of fixation. In this work, observations were
made on the apex of the cornea by means of special optical instruments.
The center of rotation of the eye did nor remain strictly
stationary relative to the eye, for during rotation of the eye the apex
of the cornea moved over a surface slightly different from that of a
sphere. Given this difference for many directions of fixation, it was
relatively easy to determine the geometrical localization of the points,
from which the position of the corresponding centers of rotation could
be determined. Park (1936a, 193613) and Park and Park (1940) used a
special goniometer to study eye movements during fixation. The eye
movements were studied in connection with the movement of the lens.
In the opinion of these authors, fixation on a point is accompanied by
continuous movement of the eye and lens. Peckham (193'4) and Ogle,
and Prangen (1949) useda stereoscopeand telescopes to study
in convergence during fixation on an object.
Because of the creation of specially devised optical instruments
the method of visual observations is still suitable for the study of
certain types of eye movement today. However, the newer methods.
wh.ich provide more objective records, have provedmoreaccur'ate and
therefore preferable.

P
CHAPTER I

6. MECHANICAL RECORDING OF EYE MOVEMENTS

In the past several authors have used methods by which the connection between the eye and the recording instrumentwas mechanical.
Three types of these methods a r e known.
The first type utilized the convexity of the cornea: the movement
of the cornea (like the cam on a camshaft) was transmitted by a lever
and balance arm. The fulcrum in which the lever rotated was fixed
to the subject's head.
One polished end of the lever, under slight
pressure, touched the anesthetized surface of the eye. The other end
made the record on a moving paper tape. The subject's head was
usually held in a headrest. This methodwas used by Ohm (1914, 1916,
1928) and Cords (1927).
In the second type, the convexity of the cornea again was used, but
this time the. movement was transmitted, not to a lever, but to an
elastic rubber balloon filled with air. The balloon was fixed s o that it
pressed lightly against the anesthetized surface of the eye. Movement
of the eye altered the pressure inside the balloon, and this change was
transmitted along a thin tube to the recording system.
In the third type of method, small cups resembling contact lenses
In. the center of the cup was an aperture or window
were used.
through which the subject looked a t the object to be perceived. The
cup was affixed to the anesthetized eye like a contact lens. A lever or
thread was attached to the cup by means of which the eye movements
were transmitted to the recording system. Delabarre (1898) and Huey
(1898, 1900) used cups made of plaster of Paris, while Orschansky
(1899) used aluminum cups. In some experiments Orschansky fixed a
small m i r r o r to the cup; he was evidently the first to use a beam of
light reflected and projected on a screen to study eye movements.
Today mechanical methods of recording eye movements a r e outmoded. Their accuracy is low, and they a r e more complicated than
many modern methods.

7. RECORDING EYE MOVEMENTS BY A REFLECTED

BEAM OF LIGHT
By some means o r other a plane m i r r o r is affixed to the subject's
eye. A beam of light is transmitted to the m i r r o r from a source whose
diaphragm may contain a narrow slit o r a small hole. The reflected
beam is directed to photosensitive material and focused on it in the

form of a bright narrow band (if the record is made on the moving
photosensitive tape of a photokymograpb). The reflected light beam
reproduces the eye movements and records them on photosensitive
material. During the experiment the subject's head is held in a headrest. The lids of the anesthetized eye a r e taped open with strips of
adhesive plaster, or the experimenter holds them openwith his hands.
Several methods a r e known for affixing a mirror to the eye. Marx
and Trendelenburg (1911) glued the m i r r o r to an aluminum cup
resembling a contact lens. The cup together with the m i r r o r was
attached to the eye like a contact lens. Dohlrnan (1925) used a rubber
cup instead of an aluminum one, while Adler and Fliegelman (1934)
attached the mirror directly to the sclera of the eye. Riggs and
Ratliff (1949, 1950, 1951) used contact lenses to which m i r r o r s were
attached. In this case there was noneed to anesthetize the eye. More
recently, contact lenses have been used by many workers (Ginsborg,
1952; Ditchburn and Ginsborg, 1953; Riggs, Armington, and Ratliff,
1954; Nachmias, 1959, 1960; Krauskopf, Cornsweet, and Riggs, 1960;
Riggs and Niehl, 1960; and others). Contactlenses have been used for
monocular and binocular records, and other simple devices, have made
possible the simultaneous recording of the vertical and horizontal
components of eye movements. Moreover, the contactlens began to be
used a s a base unit to which other devices required for solving
particular experimental problems or for the development of new
techniques were affixed instead of the mirror. Finally, instead of
contact lenses, the present author has used small rubber suction cups
with mirrors which, by virtue of their very small mass, provided a
firm link between the mirror and the eye (Yarbus, 1954). A full
description of this method is given below.
The method of recording eye movements by a reflected beam of
light is the most sensitive of any presently known method. Its great
;drawback is that i t cannot be used when for some reason no apparatus
of any kind can be affixed to the subject's eye. Another disadvantage
of the method is the creation of distortion in the records (distortion
during recording with a reflected beam of light will be considered in
detail later). On the question of contact lenses it should be noted that
although in many cases they a r e very convenientand in fact irreplaceable, they possess two important disadvantages. The contact lens has
a definite mass which, when affixed to the eye, changes its moment of
inertia considerably. This modifies eye movements taking place with
high acceleration (e.g., voluntary saccadic movements and tremor
movements). The second important disadvantage of t h e ordinary

22

CHAPTER I

contact lens is that i t is not firmly joined to the eye. During a fast
movement, the contact lens is displaced very slightly,sliding over the
eye surface.

8. STUDY OF EYE MOVEMENTS BY STILL AND

MOTION-PICTURE PHOTOGRAPHY
Many authors have used motion-picture and still photography to
study eye movements. In methods of this type, the eye movements
were judged from the consecutive movement of images of a particular
element of the eye relative to a reference point that was firmly affixed to the subject's head. In some methods in which good fixation
of the head was achieved, the initial position of the eye itself was taken
a s the point of. reference.
Dodge and Cline (1901) were evidently the first to make photographic
studies of the eye in movement. They photographed the eyes on a still
photographic plate and obtained a series of images of the eye displaced
relative to each other and corresponding to a condition of individual
fixations. Analysis of this film gave a n idea of the character of the
eye movements. .
A more refined method was used by Judd, McAllister, and Steele
(1905). They took sequential photographs (about nine frames per second) of the eye and part of the face. A small spot of Chinese white
was placed on the cornea. The subject's head was held in a headrest,
and two small shiningballs were affixed to it; since these fell within the
exposure field of the objective lens, they were photographed in each
frame and could be used a s theoriginof the coordinates. The position
of the white spot reiative to this origin of the coordinates was determined in each frame and after suitable analysis of all the frames a
descriptive account of the eye movements was obtained.
Karslake (1940) made motion pictures of the image of the eye in a
semitransparent mirror, through which the subject looked a t the
object of perception. In this method the apparatus was behind the
subject and did not distract him during the experiment.
Barlow (1952) used the following method. A very small drop of
mercury was placed on the subject's cornea. A second drop was placed
on the forehead. By means of a microscope the images of the two
drops were projected a.nd recorded on a moving film. The combined
movements of eye and head were judged from the record of the movements of the drop placed on the cornea. The movements of the head
were judged from the record of the movements of the drop on the

forehead. During the experiment the subject's head was fixed, which
restricted its movement.
Higgins and Stultz (1953). in order to study eye movements during
fixation, photographed a magnified image of a blood vessel of the
sclera on a moving film.
The optical system of the apparatus
magnified the vessel 26 times, and the vessel was chosen s o that its
image was perpendicular to the aperture of the apparatus. The part
of the sclera containing the vessel to be photographed was illuminated
with ultraviolet light.
For the control observations of the head
movement, a marker fixed to the subject's nose was recorded at. the
same time. Haherich 'and Fischer (1958)studied the blinking movements of the eye during alternation of the points of fixation by means
n one second the time lens gave 64
of a device called a time lens. J
images of the eye on a film. The turning movements of the head were
recorded simultaneously on the film. A. R. Shakhnovich and V. R.
Shakhnovich (1961) described an apparatus in which images of both the
subject's eyes were projected in the plane of the aperture of a photokymograph. A compensating prism rotated in front of this aperture
and displaced the image in'a direction perpendicular to the aperture.
In these conditions scanning of the pupil takes place and an impression
of itsdiameter is obtainedonthefilm. Both the vertical and horizontal
diameters of each pupil may be projected in the plane of the aperture
of the optic system. Both components (vertical and horizontal) of the
movement of each eye a r e recorded on the film. The accuracy of the
method is low, and it is therefore suitable for recording the large
movements of the eyes have to be recorded. Their main disadvantage
the size of the pupil is recorded a t the same time.
Jnevaluating methods of recording eye movements based on still
and motion-picture photography, it must be remembered that these
methods can be .used successfully in many cases when the large
movements of the eyes have to be recorded. Their main disadvantage
is the relatively laborious method of analysis of the records required.

9. RECORDING THE CORNEAL BRIGHT SPOT

The radius of curvature of the cornea is approximately 8 mm, and
that of the eye about 12 mm. The center of curvature of the cornea is
displaced 3-5 mm relative to the center of rotation of the eye. The
cornea, like the convex surface of a lens, reflects part of the light
falling on its surface a s the corneal reflex (the sparkle of the eyes).
Since the center of rotation of the eye and the center of curvature of

24

CHAPTER I

the cornea do not coincide, the angle a t which a stationary source of

light is reflected in the cornea changes during a movement of the eye
s o that the corneal reflex moves whentheeye moves. Several authors
have used the corneal reflex to study eye movements. This method
appeared very tempting-eye movements could be studied without
touching the eye itself. However, every experimenter has become
clearly aware of the drawbacks of this method. Very small movements
of the reflex always take place against the background of head movements and a r e added to them, so that the accuracy of the records is
considerably reduced. In almost every variant of the method, very
reliable fixation of the subject's head or rigid fixation of optical instruments to it is required, in order to avoid artifacts caused by
movements of the head. In addition, displacements of the reflex a r e
sometimes caused by changes in the thickness of the layer of tear
fluid covering the. cornea, especially near the lids. Finally, the
relationship between the eye movement and themovement of the reflex
has been found to be complicated.
The first authors who photographed the corneal reflexto study eye
movements were Dodge and Cline (1901) and Stratton (1902, 1906).
Dodge (1907) recorded the movements of the reflex on a falling
photographic plate. He studied the eye movements in the process of
fixation, pursuit, and reading. Stratton recorded the movements of
the reflex on a stationary photographic platewhen the subject examined
complicated geometrical figures. The methods of photography of the
corneal reflexes w e r e subsequently improved, and special apparatus
was introduced.
Particular attention was paid to the creation of
apparatuses designed to record the reflexes of botheyes during reading
(Tinker, 1931; Taylor, 1937). Apparatuses were also devised to allow
simultaneous records to be made of the vertical and horizontal movements of the reflex (Weaver, 1931; Clark, 1934). Two variants of
the apparatus specially designed for studying eye movements during
fixation were developed by Hartridge and Thomson (1948). To avoid
the influence of head movements on the records, these authors constructed a special alabaster cap to which the light source, the point
of fixation, and some optical instruments were fixed. The cap was
secured firmly to the subject's bead. The light source and the corneal
reflex were photographed simultaneously by a motion-picture camera
with a frequency of 60 frames per second. In the other, more refined
variant of the method,. the subject held between his teeth a special
plate fixed to the optical system during the experiment. The optical
system rotated freely in all directions and did not encumber the
subject's head.

Besides these photographic methods, Lord and Wright have recently

developed a photoelectric method of recording the movements of the
corneal reflex (Lord, 1948, 1951, 1952a. 1952b; Lordand Wright, 1948,
1949). This method has been used mainly to study eye movements
during fixation. These authors claim that a rotary movement of the
eye amounting to only one minute of angle can be recorded by means
of this method.
In their experiments, the subject lay on his back,
his head strapped to a special headrest. The subject held between his
teeth a plate firmly fixed to the headrest. A beam of ultraviolet light
with a wavelength of 365 m p was thrown on the cornea, reflected from
it, and when it fell on a semitransparentaluminized mirror, separated
into two parts. One part of the beam was directed towards the edge of
a vertical screen and the other partto the edge of a horizontal screen.
Behind each screen and partly covered by it was a photoelectric cell.
One photoelectric cell detected the changes in the horizontal component
and the other thechanges inthevertical component of the movements of
the corneal reflex. The current from the photoelectric cells was fed
into a DC amplifier and then into a cathode-ray oscilloscope. Mackworth and Mackworth (1958) used a television technique to record eye
movements. The image of the object of perception and the corneal
reflex, magnified 100 times, were transmitted to the screen of a
television tube and coupled in such a way that the position of the reflex
on the object corresponded to the point offixation. According to these
authors, this method allowed the eye movement tobe recorded with an
accuracy of up to 1-2'.
All the methods based on recording the corneal reflex can be used
only to record relatively large movements of the eyes.

10. ELECTROOCULOGRAPHY
A definite potential difference is known to exist between the outer
and inner sides of the retina o r between the cornea and the sclera
(Mowrer, Ruch, and Miller, 1936). Thus, during rotary movements of
the eye in a horizontal plane, a change takes place in the potential dlfference between points of the skin lying to the right and left of the eye.
When the eyes rotate in a vertical plane, these changes take place
between points of the skin above and below the eye; they a r e produced
in both cases by changes in the conditions of detection of the constant
potential of the eye.
The changes in the potentials may be detected by a palr of electrodes fixed to corresponding points of the skin, and then amplified

and recorded. Since a linear relationship exists between the rotation

of the eye and the change in potential, the records obtained may easily
be used to determine the eye movenients: It i s essential in this method
that head movements do not influence the record of the eye movements
and that the record is made without contact with the eye. The main
disadvantage of the method is its low level of accuracy.
Miles (1939a. 1939b. 1940) investigated the action of various
conditions on the magnitude of the corneo-retinal potential difference
and showed that light adaptation causes an increase, dark adaptation
a decrease, in the potential difference. The potential changes recorded
were usually less than 1millivolt.
Schott (1922), Meyers (1929). and Jacobson (1930a, 1930b) were
among the f i r s t investigators touse electrooculography to study the eye
movements. Later this method was widely adopted; it has been used
by Carmichael- and Dearborn (1948), Monnier and Hufschmidt (1950),
Hodgson and Lord (1954). and others. In the Soviet Union, electrooculography was first used on a wide scale by L.T. Zagorul'ko, V.D.
Glezer, B. Kh. Gurevich, and L.I. Leushina atthe I.P. Pavlov Institute
of Physiology in Leningrad.
The best of the electrooculographic methods suggested during recent
years would seem to be that described by Ford, White, and Lichtenstein (1959). By means of this method the horizontal and vertical
movements of the eye can be recorded simultaneously.
At the present time electrooculography is usedwithfair success by
many workers when highly accurate records of the eye movements a r e
not required (where the e r r o r s in the records may exceed lo).

11. SOME PHOTOELECTRIC METHODS FOR RECORDING

EYE MOVEMENTS
Recently several authors have developed methods conventionally
known a s photoelectric for recording eye movements.
One of the first methods of this type was devised by Dohlman
(1935). The scheme of this method was a s follows. A source of light
and a photoelectric cell were affixed tothe subject's head. Next, after
fixation of the lids, a rubber cup was affixed to the subject's anesthetized eye. The experimenter applied gentle pressure when placing the
cup on the eye, so that i t easily remained in place by suction. A screen
was attached to the rubber cup, and this partly obstructed a beam of
light falling on the photoelectric cell.
The edge of the screen was situatedsothat the light was modulated
by the horizontal movements of the eyes. The amplified photoelectric

current was recorded, and the eye movements judgedfrom its fluctuations.
Drischel and Lange (Drischel and Lange, 1956; Drischel, 1958) used
the following method. Anarrowband of infrared light was projected on
the subject's eyes. The spot of light on the eye was 4 mm long and
1 mm wide. The light was directed towards the temporal side of the
eye so that one half of the band was on the sclera and the other half
on the iris. The iris absorbs light better than the sclera. Movement
o f the eyes caused modulation of the reflected light which fell on a
photoelectric cell. The quantity of reflected lightwas directly related
to the position of the eye. The current from the photoelectric cell
was fed into an amplifier and then into an oscilloscope, from which
photographic records were made.
Cornsweet (1958) developed the following method for investigating
small movements of the eyes during fixation. After clamping the
subject's head, a narrow beam of light was transmitted through the
pupil onto the blind spot of the eye. The spot of light on the blind spot
was shaped like a small rectangle. During movement of the eye, this
spot intersected the large 'blood vessels, causing modulation of the
reflected light. The reflectedlightwas transmitted to a photomultiplier
and oscilloscope, from which photographic records were made.
Smith and Warter (1959, 1960) suggested a simple variant of these
methods. After the position of the subject's head had been fixed, an
image of a small a r e a of the eye was projected by an optical system
on a horizontal slit. The center of this area corresponded to the
boundary between the cornea and sclera a t the point where the tangent
to the boundary is vertical in position. In this case the image of the
corneal border was perpendicular to the slit. The photocathode of a
photomultiplier tube was placed behind the slit. Rotation of the eye
around the vertical axis modulated the light falling on the photosensitive element of the tube. The reccrded changes-in the photoelectric
current corresponded directly t o the eye movements. Smith and
Warter also described a moving object whose speed, brightness, and
other characteristics may be modified a t will. This system was connected with the system for recording eye movements s o that the eye
movements and the movements of the object were recorded simultaneously.
Gaarder (1960) used a method with a contact lens to which a plane
m i r r o r was fixed. A beam of light reflected from the mirror was
transmitted to a photoelectric cell. By this method, both horizontal
and vertical movements of the eyes could be recorded.
Shackel (1960) described a method which couldbeused to study the
movements, not only of the eyes, but also of the head. A television

28

CHAPTER I

camera, affixed to the subject's head recorded the field visible to the
eye. The apparatus recording the eye movements transmitted to the
same screen a small spot of white light corresponding, with a certain
degree of accuracy, to the point of f w i o n a t each particular moment,
Vladimirov and Khomskaya (1961) described a photoelectric method
of recording eye movements giving adirect ink tracing. Via a n optical
system, the image of the subject's eye was projected on a piece of
ground glass divided by a vertical screenintotwo equal parts. Behind
each half of the ground glass was a photoresistor, the photosensitive
layer of which was firmly in contact with the glass. Movement of the
eye in the horizontal plane caused a displacement of its image on the
ground glass and a change in the intensity of illumination of the photoresistors. The photoresistors were included in a circuit in which a
change in the output voltage was directly proportional to the eye
movement, This change was fed into the input of an amplifier or a
loop. The horizonti1 movements of the eyes could be recorded with
an accuracy of 1-2" by this method. As these authors state, their
method is only suitable for recording the large movements of the eye.
Photoelectric methods can be used to study eye movements if an
e r r o r of 1-2" is acceptable in the work. In nearly all these methods
an increase in accuracy would require better fixation of the subject's
head. Several of the photoelectric methods enable the eye movements
to he studied without contact with the eye; this is their main advantage.

remain in the same plane during eye movements, the angle of rotation
of the reflected light is twice the angle of rotation of the eye. The
optical systemdeveloped by these authors reduces the regular displacement of the image on the screen, making it equal in the horizontal
direction to the rotation oftheeye. The image of the vertical houndary
between the two fields onthe screenmustremain stationary relative to
the retina. Clowes and Ditchburn (1959) have improved this method s o
that it is possible to compensate not only the horizontal, but also the
vertical component of the eye movements. These authors affixed a
short-focus lens together with the test object to the contact lens. The
short-focus lens was designed to produce a stationary and sharply
defined image of the object on the retina.
The main disadvantage of methods creating a stationary retinal
image by means of contact lenses is that during the experiment the
vertical border on the test field disappears only for short intervals of
time (several seconds). It then reappears for a few seconds and
disappears again, and so on. The appearance of the boundary on the
test field is perhaps due to insufficiently firm contact between the
contact lens and the eye.' Evidently, a t times when the eye makes
sudden rotations, the contact lens moves slightly and this leads to the
appearance of the border.

13. SUCTION DEVICES (CAPS)

12. CREATION O F A STABILIZED RETINAL IMAGE WITH A

CONTACT LENS
Ditchburn and Ginsborg (1952), and Riggs, Ratliff, Cornsweet,
and Cornsweet (1953) describe a method designed to produce a stabilized image on the retina.
A rod is firmly fixed to a contact lens; at the end of the rod is a
plane mirror. The rod is displaced relative to the axis of the lens s o
that the central part of the lens covering the cornea is left free and
the blinking movement of the eye hardly impeded during the experiment.
The axis of the rod, normal to the m i r r o r , and axis of the lens a r e all
parallel. After the lens has been fixed to the eye, a narrow beam of
light i s thrown from a projector ontothe mirror. The beam reflected
from the m i r r o r enters the optical system and passes from it to a
screen situated in front of the subject's eyes. The test field visible to
the subject on the screen is circular andconsists of two fields divided
by a vertical boundary. Thediameter ofthe test field is 2". When the
incident and reflected beams of light and the normal to the mirror

In this and the following sections of this chapter a description is

given of the method used by the author of this'book. The assembly of
the appropriate instruments and accessories is described in one
section.
The most important element of the apparatus used in this method is
the special suction device, hereinafter referred.to a s a "cap.' Depending on the object of the experiment, caps of different construction
a r e made. Each cap is fixedduringthe experiment to the anesthetized
surface of the subject's eye, causing no painful sensations o r undesirable aftereffects. The method by which the cap is affixed to the
eye is obvious from the description, suctiondevice. The author claims
that the cap ensures firm contact between the miniature apparatus
used and the eye. In all experiments, the eyelids a r e secured to
exclude blinking movements, which could displace the cap o r detach it
from the eye. No experiment shouldexceeda few minutes in duration.
When making the caps, several variations in shape and size a r e
permissible. In most cases, therefore, illustrations of the caps a r e
given not a s scale drawings, but a s schemes with explanations and

CHAPTER I

Fig, 13. The PI suction device or 'cap..

indications of the more important measurements in the text. Each

illustration of a cap must be regarded as a horizontal section passing
through the optical axis of the right eye.
The construction of the caps isdescribedwithall necessary details
to enable them to be made in any workshop. If the reader does not
intend to make the capand can readily grasp the idea of each construction with the description of its details, hemay ignore this description.
However, in order to understand the material in the following chapters.
i t is essential to have a general idea of the construction of each cap.
The cap shown schematically in Fig. 13 can be used to record eye
movements. To simplify the descriptions, all caps of this type will
be referred to in the future a s type PI. Any other type of cap will he
denoted by the same letter but with a different index.
The PI cap is made of rubber. It consists of a round suction part 1.
resembling an inverted dish, and a hollow cylindrical side-piece 2,
joined by a canal 3 to a recess 4. A small mirror 5 is secured to the
frame of the cap and usedforrecordingthe eye movements by reflecting a convergent beam of light. By means of the hollow side-piece, a
reduced pressure is created inside the cap, and this facilitates its
attachment to the sclera of theeye. After the cap has been fixed to the
temporal part of the sclera, the surface of the. mirror is approximately
normal to the optical axis of the eye, and the hollow side-piece does
not interfere with the incident and reflected beams of light during the
experiment. The position of rhe cap can be seen from F'ig. 13. The
diameter of the suction part of large caps of type Piis 6-7 mm, that
of small models, 3-4 mm. The size of the cap is determined by the
size of the mirror required in the experiment. A change in the s u e
of the cap is accompanied by a proportional change in all its dimensions. The diameter of the circularmirrors on the large models does
not exceed 7-8 mm-3-4 mm, on the small. The thickness of the
smallest mirrors is 0.2 mm, and that of the largest 0.3-0.4 mm. For
some experiments, a rectangular mirror (5 x 7 mm) is more convenient. The length of the hollow side-piece is approximately equal

METHODS

31

to the diameter of the suctionpart. The external diameter of the hollow

side-piece is equal to half its length and its thickness is 0.6-0.7 mm.
The weight of the large caps, together with the mirror, is 0.20-0.25 g.
and the weight of the small models 0.02-0.03g. For most purposes, it
is convenient to use a cap having a suction part 6 mm in diameter and
a mirror of the same size.
The type P2cap is shown schematically in Fig. 14; it differs from
the Pi type only in shape and s u e . The PZcap is used when the eye
has to be completely closed and at the same time its movements have
to be recorded. The P2cap consists of a suction part (the frame) 1,
the hollow side-piece 2,joinedby the canal 3 to the recess 4. A mirror
5 is attached to the frame of the cap, and its surface is normal to the
axis of symmetry of the frame and the optical axis of the eye. The
measurements of the suction part and the measurements of the recess
4 a r e such that the cap nowhere touches the cornea. The height of the
cap without the mirror is about 5.5-6 mm, and its external diameter
is 13 mm. The hollow side-piece is 7-8 mm long, its external diameter is 4-5 mm, and the thickness of its walls is 0.5-0.6 mm. The
mirror is 7-8 mm in diameter and 0.3 mm thick. The edge of the
frame of the cap touching the sclera is not more than 0.2-0.3 mm
thick. h he middle part of the frameis 1.5-2 mm thick. The weight of
the cap together with the mirror is 0.5-0.6 g.
A second variant of type Pz is the Pa cap, illustrated in Fig. 15.
The only difference between this type and type P2is that its frame is
made not of rubber but of thin duralumin. The P3model is much
lighter than the P2and causes less irritation to the conjunctive. The"
dimensions of the frame a r e given in Fig:-15; they a r e identical with
the dimensions of caps of several different types of construction. The
surface of contact of the cap with the eye is polished and corrugated.
The corrugated surface prevents the cap from slipping over the eye.
The weight of the P, model is 0.20-0.25 g.

Fig. 14. The P1 cap.

CHAPTER I

1,
Fig. 15. The Ps cap.

The Pa cap is shown schematically in Fig. 16, and this model

enables conditions of perception to be. created in which the ordinary
correlation between the eye movement and the displacement of the
retinal image is disturbed. The duraluminframeof the P4 cap has the
same dimensions a s the frame of the Pgmodel. In the top part of the
frame i s a round aperture 1, with a diameter of 4 mm. A round glass
plate 2 is fixed to the frame around its whole perimeter. To this glass
plate is fixed a rectangular mirror 3, the plane of which makes an
angle of 45' with the axis of symmetry of the frame of the apparatus.
The dimensions of the. m i r r o r a r e 7 x 10 mm. When the cap is fixed
to the eye, the subject can see an object only by means of the mirror.

Fig. 16. The P4 cap.

METHODS

33

The hollow side-piece is s o situated that the subject cannot s e e i t in

the mirror. All objects situated to the side appear to be displaced in
the frontal plane. Rotation of the eye causes rotation of the m i r r o r
and, consequently, displacement of the retinal image. The relationship
between the angle of rotation of the eye and the angle of displacement
of the retinal image is very complex and differs sharply from what
is found in normal conditions, i.e., when we look with the unaided eye.
In other words, conditions a r e produced in which the subject clearly
sees objects but cannot voluntarily choose points of fixation or use eye
movements to obtain information concerning the spatial relationships
between objects. The field of vision of an eye to which the P4apparatus
is attached is approximately 50". Theweight of this cap is 0.30-0.35 g.
If both large surfaces of the mirror of the P4apparatus a r e parallel
and polished, by removing the reflecting layer a transparent window
can be created in it. Through this window the subject can s e e objects
in front of him (practically without distortion). When the PAapparatus
with a window in the m i r r o r is fixedto the eye, conditions a r e created
in which the eye's field of vision is divided into two parts. In one part
of this field, the ordinary relationshipbetween the movement of the eye
and displacement of the retinal image is disturbed, while in the other
it is normal. To increase the sharpness of the border between the
parts of the field, the size of theaperture 1 in the P4 apparatus is r e duced to 1.0-0.5 mm.
The P6cap, illustrated schematically in Fig. 17, is used to record
the pulsations of the eye. The frame of the apparatus 2 and the hollow
side-piece 14 a r e made of rubber. The hollow side-piece is joined bT
the opening 15 to the lower chamber 3 of the cap, in which is created
the reduced pressure necessary for securing the apparatus to t h e eye.

Fig. 17. The P5 cap.

The lower chamber 3 of the cap is bounded below by the cornea of the
eye 1, at the sides by the frame of the cap 2, and above by a thin rubber
diaphragm 4 to which is secured a flat rubber button 6, lightly pressing
on the cornea. The upper chamber 5 of the cap is bounded below by
the rubber diaphragm 4, at the sides by the frame 2, and above by a
rigid plate 7, fixed to the frame of the cap. The upper chamber is
joined with the outside space througha special filter 12, fixed to a small
cylinder 13. The filter transmits a i r slowly, so that the mean pressure
in the chamber 5 is equal to atmospheric. The time taken for the
pressure in the upper chamber to become equal is much greater than
the blood pulsation period. Blood entering the eye causes pulsation of
the intraocular pressure, and this in turn leads to deformation of the
eye which is synchronized with the pulse. The deformation of the eye
is transmitted to the elastic rubber diaphragm 4, movement of which
causes a pulsation of pressure inside the upper chamber 5. The
cylinder 8 is secured to the rigid plate 7. Over the cylinder 8 is gently
stretched a very thin and elastic rubber membrane 9, several times
thinner than the rubber diaphragm 4. The cylinder 8 is joined by an
aperture with the upper chamber of the cap, so that pulsation of pressure inside the upper chamber, is transmitted to the membrane 9 and
deforms it synchronously with the pulsation of intraocular pressure.
A small mirror 10 is fixed to the edge of the membrane 9. >'he
mirror rotates during deformation of the membrane 9. Beside the
mirror 10, secured to the membrane, is a mirror 11, rigidly fixed to
the plate 7. A beam of light emerging from a slit aperture, reflected
by both mirrors, and focused on the oscillographic paper of the
photokymograph, records two lines. The beam of light reflected from
the mirror 11, rigidly fixed to the cap, records the movements of the
eye, while the beam oflight reflected from the mirror 10, secured to
the membrane, records the movements of the eye and the deformation
of the cornea caused by the pulsation of intraocular pressure. In the
course of the experiment the observer fixates on a certain point with
his free eye. To ensurethatthe scale of the records is the same in all
experiments; the distance between the observer's eye and the paper
of the photokymograph is kept constant. In addition, before the cap
is fixed in position, care must be taken that an approximately equal
volume of air is withdrawn from the hollow side-piece. In external
diameter, the frame of the Pg cap is 13 mm. In its construction it
resembles the frame of the P2 model. When the pressure in the lower
chamber 3 is equal to atmospheric and the apparatus rests with its
edges on the sclera, the distance between the rubber diaphragm 4 and
the cornea is 1.5-2.0 mm.
The thickness of the diaphragm 4 is
'

0.2-0.3 mm.
The diameter of the rubber button 6, fixed to the
diaphragm, is 4 mm and its thickness lmm. The distance between the
rubber diaphragm 4 and the rigid plate 7 is about 3 mm. The plate 7
is made of clear plastic 0.5 mm inthickness. The cylinders 8 and 13,
of the same thickness, a r e made of clear plastic and glued to the plate.
The internal diameter of cylinder 8 is 3mm and of cylinder 13, 2 mm.
The height of each cylinder and of the stand to which the mirror 11is
fixed is 3 mm. The diameters of the holes joining the spaces inside
the cylinders 8 and 13 to the upper chamber 5 are each 1mm. The
thickness of the rubber membrane 9 stretched over cylinder 8 is about
0.03 mm. The mirrors 10 and 11are square in shape, the first with
a side of 1 mm and the second of 2 mm. The thickness of mirror 10
is 0.1 mm. Filter 12 is made of a single layer of ordinary filter paper
and is glued to cylinder 13. The transmitting power of the filter is
lowered by coating part of its surface withglue. The hollow side-piece
14 is 9 mm long; the external diameter of its lower part 5 mm and its
upper part 6 mm, and the thickness of its walls is 0.7 mm. The weight
of the PScap is about 0.6-0.7g. Assembly and preliminary adjustment
of the cap are carried 'out on a rigid model of the eye, and final
adjustments a r e made after experiments on the living eye and the
necessary records have been obtained. Duringassembly it is important
to select the correct tension of the diaphragm 4 and the membrane 9
and the correct transmitting capacity of the filter 12. It is, of course,
easy to imagine a variant of the Ps cap using a piezocrystal or any
other suitable pickup. In this case, the recording of the pulsation
could he made by means of an amplifier actuating a loop oscillograph..
The scheme of the P6 cap, which creates a stationary retinal
image for the whole field ofvision~ftheeye, is shown in Fig. 18. The
duralumin frame 1 and the hollow rubber side-piece 2 of the PCcap
have the same dimensions as the frame and hollow side-piece of the
P, model. The surface of contact between theframe and the eye is
corrugated and polished. A thin duralumin cylinder 3, 0.1 mm thick,
is glued firmly to the frame. The diameter of the cylinder and its
height are about 5 mm. Inside the cylinder a r e mounted two duralumin
diaphragms 4, in firm contact with it. Each diaphragm is 0.1 mm
thick, and the diameter of the aperture is 1.5-2.5 mm. The diameter
of the aperture in the upper part of the frame of the apparatus is
2.5-3 mm. The distance between the frame and the first diaphragm
is 1mm, and that between the frame and the second diaphragm 2 mm.
A short-focus lens 5 is fixed tothe second diaphragm and the cylinder.
The focal length of the lens is 5-8 mm. Over the cylinder 3 is placed
a device conventionally called an adaptor. Theframe of the adaptor is

36

CHAPTER 1

Fig. 18. The P, cap.

made of black paper. The lower part of the adaptor consists of the
paper cylinder 6, which is firmly held by friction on the duralumin
cylinder 3. A paper square 7witha round hole 8 in the center is glued
to the cylinder 6. The side of the square 7 is a few millimeters longer
than the diameter of the paper cylinder 6, and the hole 8, in turn, is a
few millimeters smaller than this diameter, and is in fact 3.5 mm.
The hole in the square is situatedat the focus of the lens. The test field
fits into the hole. A square frosted glass (side 6 mm long), about
0.2 m m thick, is fixed parallel to one side of the square 7 and perpendicular to its plane. The mirror 10, 6 mm wide (the same size a s the
frosted glass), 9 mm high, and 0.1-0.2 mm thick, is fixed at an angle
of 45" to the paper square and the frosted glass. Triangular pieces of
black paper a r e fixed to the edges of the frosted glass and the mirror
so that the inside of the adaptor receives light only through the frosted
glass. All the slits intheadaptorare carefully glued together, and the
device is painted black.
During the experiment the adaptor occupies the position shown in
Fig. 18. Experiments with the P6capare usually carried out in a dark
room. A narrow beam of light, shown by arrows in Fig. 18, falls only
on the adaptor of the capandilluminatesthe frosted glass. The sclera
of the eye is practically in total darkness, and, consequently, light
enters the eye only through the cap, a very important factor in many
experiments with an image stationary relative to the retina.
The short-focus lens gives a magnified image of the test field which
the eye sees against the hackgroundof thefrosted glass reflected in the
m i r r o r . Adjustment of the sharpness of the image is done by moving
the adaptor along the'duralumin cylinder 3. It is clear from Fig. 18
that the lens of the apparatus is housed inside the cylinder a short
distance from the cornea and separated from it by the diaphragm.

METHODS

37

This type of construction was determined by the need for protecting

the lens against clouding, which would haveinterferedwith the conduct
of the experiment. Since the short-focus lens gives high magnification,
the visible diameter of the test field may be made greater than 50".
As mentiolied above, the test field and the projection system of the Ps
apparatus a r e firmly fixed to its frame. The frame of the cap, in
Nrn, is firmly fixed to the eye and, consequently, even during movement of the eyes the retinal imageof the test field i s always stationary
relative to the retina. The weight of the Pg cap without the adaptor is
0.15-0.20 g.
The P7 cap, by means of which two superposed retinal images may
be obtained on the retina, is illustrated schematically in Fig. 19. One
of the two images is stationary relative tothe retina while the other is
movable.
The duralumin frame 1 of the P7 model and the hollow rubber bulb
2 a r e the same size a s the frame and the hollow bulb of the Ps apparatus.
The contact surface between the frame and the eye is
corrugated and polished. To the whole perimeter of the frame is fixed
a duralumin cylinder 3; with an external diameter of 5 mm, a height
of 2 mm, and walls 0.1 mm thick. Inside the cylinder is fixed a
duralumin diaphragm 4. The thickness of the diaphragm is 0.1 mm
and the diameter of the aperture 3.5 mm. The distance between
diaphragm 4 and the frame is 1mm. The diameter of the aperture 5
in the upper part of the frame is 3 mm. A round transparent glass 6
with the lens 7 is fixed to the cylinder 3 around its entire perimeter.
The glass is 6 mm in diameter and 0.15-0.20 mm thick, The diameter
of the lens is 2-3mmanditsfocal length 5-8 mm. The cylinder 3 is so
made that it removes the glass with thelens a short distance from the
eye and thus prevents it from becoming steamed up. A round wood
T
rod 8 is fixed totheframeofthe cap and the cylinder. In turn, the test
I

CHAPTER I

..
Fig. 20. The P, cap.

field 9 is fixed to the rod and is stationary relative to the rod and to
the frame of the cap. The height of the rod is determined by the focal
length of the lens andissodesignedthat a sharp image of the test field
9 may be obtained on the retina. It is clear from Fig. 19 that lens 7
covers only the central part of the glass 6, leaving the peripheral part
open. When the cap is fixed to the eye, the peripheral part of the glass
(the circular diaphragm) enables the subject to see clearly and to
examine the surrounding object freely. The color contrast of the visible
objects is slightly below normal due to the scattered light (the out-offocus image of surrounding objects) falling on the retina through the
lens. The lens gives asharpretinal image of the test field fixed to the
rod, and since it is firmly connected to the cap, this image is strictly
stationary relative to the retina. As aresult, two images, superposed
on each other, a r e formed on the retina. One is movable relative to
the retina (movable because of the eye movements) while the other is
strictly stationary and unchanged in color (provided that the illumination of test field 9 is unchanged). By changing the area of the circular
diaphragm and the diameter of the lens, the brightness of the stationary
and moving retinal images may be varied within certain limits.
The scheme of the P, cap, by means of which images remain
stationary relative to the retinaforanypartof the field of vision of the
eye, is illustrated in Fig. 20.
The duralumin frame 1 of the P8 model and the hollow rubber bulb
2 a r e the same size as the frameand the bulb, of the Pscap. The surface of contact between theframeandtheeye is corrugated and polished. Around the whole perimeter of the frame is fixed the duralumin

METHODS

39

cylinder 3. The external diameter of the cylinder is 4.5 mm, its

height 3 mm, and the thickness of its wall 0.1 mm. Inside the cylinder
is fixed the duralumin diaphragm 4. The diaphragm 5 is 0.1 mm
thick, and its aperture is 2 mm in diameter. The cylinder is closed
on top by a second diaphragm 5, which is fixed to the cylinder around
its whole perimeter.
Diaphragm 5 is 0.1 mm thick, its external
diameter is 5 mm, and the diameter of the aperture is 1.5 mm. The
diameter of the aperture 6 in the upper part of the frame of the apparatus is 2.5 mm. The distance between the diaphragm 4. situated
inside the cylinder, and the frame of the cap is 1.5 mm. A round
transparent glass 7 is glued all around the perimeter of the second
diaphragm 5. This glass is 0.1-0.2 mm thick and 4 mm in diameter.
Jnterchangeable diaphragms 8, made of thin black paper or foil, are
glued to the glass. Depending on the object of the experiment, the
apertures of the interchangeable diaphragms vary from 0.2to 1.5 mm.
A rod 9 (a thinsteelwire), towhich is fixed the screen 10-the test field
stationary relative to the frame of the apparatus-is fixed to the frame
of the cap and to the cylinder. The height of the rod (usually about
20-25 mm) is determined bythe purpose of the experiment. Cylinder
3 and the first two diaphragms 4 and 5 a r e designed so as to keep
the glass from steaming up.
When the cap is fixed to the eye, the screen becomes stationary
relative to the retina. The small size of the aperture of diaphragm 8,
fixed to the glass, increases the depth of focus to such a degree that,
besides objects at a distance from the eye, the screen 10 can also be
clearly seen. The sharpness of the image of the screen depends on
the size of the aperture in the diaphragm. l%e smaller the aperturev
in the diaphragm, the sharper the image of the screen. However,
reducing the size ofthediaphragmaperturealsodiminishes the brightness of the visual image. For this reason, in ea experiment a
diaphragm is chosen which is satisfactory for ,the experxmenter a s
regards both the sharpness of the screen's image and the brightness
of objects visible to the eye. Some increase in the sharpness of
definition of the image of the screen can be obtained by lengthening
the rod, i.e.. by moving the screen further from the eye, but with this
manoeuvre, the rigidity of the connection between the screen and the
frame of the apparatus and between ?he frame of the cap and the eye
is rapidly lost. By giving the screen different shapes and positions
in space, it is comparatively easy to produce a stationary retinal image
of a given shape and color in any part of the retina. To increase the
brightness of the screen,a beam of light is thrown upon it, as shown
by thearrows in Fig. 20. If a filter instead of a screen is affixed to

q .

CHAPTER I

METHODS

the rod, conditions of perception a r e produced in which a definite part

of the retina is shielded by the filter. Because of the small mirror
fixed to the cylinder of the P8 cap, eye movements can be recorded
under conditions in which a given part of the retina is shielded by the
screen, i.e., is in factpreventedfromreceivingany visual stimulation.
Depending on the purpose of the investigation, the experimenter may
need to modlfy not only the construction of the caps, but also the construction of the adaptors. The descriptions in the second chapter of
certain experiments include a detailed account of several adaptors used
with the P6apparatus.

14. APPARATUS USED IN WORK WITH CAPS

A photograph of the apparatus usually used in recording eye
movements is given in Fig. 21. The apparaNs consists of a stand
(or frame), a chin rest, two light sources, and a control panel. The
frame consists basically of a large, massive stand.
Two metal uprights and the control panel, on which sockets and
switches are mounted, a r e firmly fixed to this stand. On the movable
part of the large stand is mounted a metal post ending in a chin rest.
The chin rest can be moved vertically; horizontally, it can turn about
the axis of the post, and after the desired position has been obtained,
it can be firmly fixed. In addition, the parameters of the chin rest
itself may be varied by the experimenter, depending on the size of the
subject's head. By use of this type of chin rest, the subject's head
can be securely fixed during the experiments. On each metal post
is a massive connecting rod, and a t the end of this rod a universal
stand. The light. source is fixed on ball bearings to each stand. By
means of this system the experimenter can quickly (and this is very
important) direct a beam of light reflected from the mirror of the cap
to the aperture of a kymograph or to a cassette. The switchboard
control panel permits any apparatus to be switched on and off in the
course of the experiment without interrupting the observation.
Depending on which cap is used for the experiment and the pareicular purpose of the investigation, the experimenter will need to use
different light sources and accessories. Forexample, when recording
eye movements on still photosensitive paper or film, a light source
is used which throws a spot of light not more than 1mm in diameter
onto the photosen&itivematerial. In this case the objective gives an
image of the small aperture of the diaphragm against the background
of the incandescent filament. Usually a series of diaphragms with

Fig. 21. The apparatus used in recording eye movement&

apertures between 10 and 70 p in diameter is used in an investigation.

If the eye movements a r e recorded on a photokymograph, a slit takes
the place of the diaphragm in the light source; the slits in a suitable
series vary from 10 to 70 p in width.
To illuminate the frosted glass of the P6apparatus or the screen
in the P8 apparatus, a light source is used which has an optical system
allowing a beam of light about 10-16 mm in diameter to be obtained at
any point in space, illuminating a small area of surface uniformly.
Uniformity of illumination is essential toensure thatduring eye movements, i.e., during movements of the frosted glass or the screen,
their brightness does not change within the beam of light.

In some experiments with an image stationary relative to the retina

(i.e., when the P6 and Psmodels areused), adaptors with polarized and
other different types of filters a r e attached tothe light sources. Sometimes special discs and vanes a r e used to change the hrighmess and
color of the light falling on the frosted glass or the screen in accordance with a given pattern. However, no explanation is required of the
construction of these devices.
When using the PC, P7, and Ps types of cap, the experiments are
best carried out in conditions under which the subject's face is turned
upward, and the fixation point for the second eye lies on the ceiling.
An armchair with a low back, with a controllable device for maintaining
the head in the required position, may be used for this purpose.
Depending on the object of the experiment, eye movements may be
recorded on moving or still photosensitive material. Records on still
photographic paper or film a r e made in a dark room, in which only
the object of perception is illuminated by adirected beam of light, and
placed against a black matt bacQround. Usually the photographic
material is enclosed in a cassette and opened only during actual recording. To facilitate the subsequent analysis of the records, the
photographic paper or film should be fairly large, approximately
30 x 40 cm. The cassettes. should be of corresponding size. To
shorten the exposure of the photosensitive material and the duration of
the experiment, cassettes a r e usedwhich a r e easily and quickly opened
and closed in a dark room. When records a r e made on moving photosensitive material (oscillographic paper), ordinary photokymographs
a r e used. However, in the ordinary photokymograph the speed of the
oscillog.raphic paper does not exceed 20-30 cm/sec and the length of
the slit is only 12-15 cm. However, when many eye movements are
studied, speeds closer to 5 m/sec and slits 25-30 cm in length a r e
necessary. These technical conditions are satisfied by a primitive
and somewhat modified model of a photokymograph, the making of which
presents no special difficulty and requires no detailed description.
Such a photokymograph is based on a large drum, to which wide oscillographic paper is fixed in darkness. The diameter of the drum is
50-60 cm, and its height is determined by the width of the oscillographic
paper, which should be a t least 25-30 cm wide. The axle of the drum
is connected to an electric motor throughagearbox o r reducing chain.
The whole system is covered with alightproofplywood case with a slit
surrounded by visors preventing scattered light from falling on the
drum. The gearbox allows the speed of rotation of the drum to he
changed. The linear velocitles of the oscillographic paper may be
varied in this manner from several centimeters to several meters per

second. Repeated turns ~f the drum cause the records to be superposed

on each other. Of course, the experiment must be stopped before the
whole picture becomes too confused and too complicated for analysis.
The higher the speed, the shorter the experiment must he. When the
speed of the oscillographic paper is several meters per second, in
many cases recording should not continue longer than 10-20 sec.
When it is necessary to affix apparatus essential for or facilitating
an experiment to the subject's face, he wears a Plexiglas mask, as
shown in Fig. 22. The mask fits the face and head snugly and stays in
position satisfactorily. A number of holes with a screw thread a r e
made in the Plexiplas to which the ex~erlrnentermav attach the required accessories.
Besides those mentioned above, apparatuses familiar inlaboratory
practice may also he used in work with caps.

::#
<:

15. TECHNIQUE OF EXPERIMENTS WITH CAPS

The cap cannot remain-on the subject's eye more than a few minUtes. A s soon as thecaphas been fixed to the eye, the work to he done
falls into two parts-adjustment of the apparatus, .and the experiment
itself. In many cases the adjustment of the apparatus is itself complicated and takes much time.
Very careful preparation for each experiment can minimize this
time. Preparation for working with the caps begins with the choice of
subject. The subject should have large eyes, a long palpehral fissure
and his conjunctiva should be healthy andnotirritated by amethocaine. '

Fig. 22. Plexiglas mask for attachment of accessories used in experiment8 wfh caps.

CHAPTER I

Fig.23. Position of lids held by strips of adhesive plasrer in work with the Picap.

The experimenter should f i r s t make a detailed planof the experiment,

prepare every piece of equipment required, and give the subject his
instructions. The subject sits with his chin on the chin rest or in the
special armchair. After preliminaryadjustment of the whole apparatus,
it is tested to make sure that it is ready. This is particularly
important when eye movements a r e recorded on still photographic
paper o r film. Correct positioning of the visual test object and the
cassette relative t o the subject simplifies the experiment and reduces
distortion.
The experimenter then cuts strips of adhesive plaster for holding
back the eyelids. Each s t r i p is 12-15 mm wide and about 10 cm long.
Usually two strips a r e sufficient for thelids of one eye. The prepared
strips of adhesive plaster a r e placed on a clean sheet of paper. The
whole sheet of paper with t h e s t r i p s is placed on a surface kept heated
to a temperature of 60-8O0C, where it stays until required. Heating
the adhesive plaster causes it tosticktothe skin more firmly and hold
the eyelids more securely.
Next, the experimenter wipes the subject's eyelids, forehead, and
cheeks with cotton wool lightly soaked in alcohol, making the skin dry
and clean.
He then instils two o r three drops of a l%solution of
amethocaine. into the conjunctival sac, wipes away the tears with dry
cotton wool, and after 1-2 minutes proceeds to tape back the eyelids.
This is done a s follows. The subject is asked to close his eyes and
one end of a s t r i p of adhesive plaster is pressed against the upper lid
s o that it touches the eyelashes. By pressing the skin of the eyelids
covered with adhesive plaster with two fingers, the lid is retracted
together with the plaster from the eye until a vertical fold is formed
between the fingers. By squeezing the fold,a firm connection between
the adhesive plaster and skin is obtained. By pulling on the other end
1

METHODS

45

of the strip of adhesive plaster, the lid may be raised to the required
position and held there by stickingthestripto the forehead. The lower
lid is taped in precisely the same way, except that the strip of adhesive
plaster is stuck to the cheek. If the subject wears the mask (Fig. 22),
the strips of 'adhesive plaster may be stuck to it. The position of the
lids for working with the Pi type of cap is illustrated in Fig. 23, and
the positionofthelids forworkingwitball other y p e s of cap in Fig. 24.
In the first case, a s is clear from Fig. 23, the lids a r e retracted mainly
towards the temporal part of the eye. In the second case, the adhesive
plaster is stuck to the central part of each eyelid and the lids a r e
retracted almost symmetrically relative to the optical axis.
After the lids have been fixed, the subject places his head on the
chin rest in the required position, and the experimenter applies the
cap to the eye.
To apply the cap, the experimenterholds it by the bulb with his two
fingers, squeezes the a i r from it, places it in the required position,
gently pressing the suction part t o theeye,and releases the bulb. The
re-expanding bulb lowers ,the pressure inside the cap, and the external
pressure presses i t firmly against the sclera.
Later the experimenter places the light sources in the correct
position, makes sure that all the apparatus is in working order, and
begins the experiment. The different experiments differ considerably
in their complexity, although all demand skill and precision from the
experimenter in his work. Recording eye movemenrs on still photographic paper o r film is complicated. Let us assume that the cap is
fixed to the eye. The cassette is still closed but ready to record. The
room is dark. Thevisual testobject,placed against a matt black background, is illuminated by a directedbeamof light but cavered by paper
s o that the subject cannot s e e it before recording begins. Then the

Fig. 24. Pasltlon of l ~ d sheld by strlps of adhestve plaster in work wlth all caps except
type PI.

46

CHAPTER 1

command is given, and the subject fixates on the center of the visual
test object. The experimenter moves the light source, and directs
the beam of light reflected from the mirror of the cap to the center of
the cassette. Next, the subject moves hisglance several times around
the borders of the object. The e~perimenter~watching
the movements
of the light spot over the cassette, moves the cassette (and sometimes
both the cassette and the light source) to a point in space at which the
whole record can be accommodated on the photosensitivematerial, and
distortions in the record a r e minimal. The subject then again fixates
on the center of the test object. The experimenter obtains a sharp
image of the light spot on the cassette. Before recording begins, the
subject fixates on a point situatedatadistance from the the test object
so that the light spotdoes notmove off the cassette. The experimenter
uncovers the cassette and the test object, switches on the second
counter, and tells the subject tolookatthe test object. After a certain
period of time,the experimenter switches off the light and the record
comes to an end. The cassette is then closed, the light is switched on,
the subject's face is illuminated, and the capand the strips of adhesive
plaster a r e removed. The cassette is then taken to the photographic
laboratory for processing and analysis.
The cap is removed from the eye as follows: the subject is asked
to fixate on a certain point so as to prevent eye movement, the hollow
bulb is compressed with two fingers, expelling all the air from it, and
the cap can then be removed from the eye.
When the cap is attached to the eye, the subject must restrict his
eye movements to avoid knocking the cap against the lids. Usually, at
the beginning of the experiment, the experimenter indicates the limits
of the field beyond which thesubjectmustnot choose points of fixation.
If it comes into contact with the lids, the cap may be detached from
the eye or may he displaced so that its suction part is on the cornea.
and, worst of all, it may injure theeye. This must always be borne in
mind by both experimenter and subject.
As a rule the duration of experiments with the cap should not
exceed 5 min, and only in rare cases as long a s 10-12 min. When
working with the PI type of cap, observationsmust be made constantly
on the state of the cornea, for in some subjects it begins to dry after
only 3 min. Dryingof thecornea,especially its central part, is accompanied by a sharp fall in the resolving power of the eye and is always
regarded by the subject with some alarm. In such cases, the experiment must be stopped. The cornea usually resumes its previous form
after a few minutes. The Pi cap itself cannot injure the eyes. The
worst it can do is to rupture asuperficial blood vessel in the conjunctiva. This suggests that the particular subject is not suitable. The

METHODS

47

Pi cap always causes a slight fall of intraocular pressure, but this is

restored after 1 or a few hours, and the subject feels no sensations
associated with the change of pressure. Inpeople with a normal intraocular pressure, the pressure falls approximately 1-2 mm Hg; in
patients with glaucoma, the intraocular pressure may fall by several
millimeters.
All types of cap except Pi protect the cornea against drying but
compress the blood vessels in theconjunctivaaroundthe whole perirneter of the cornea. For this reason experiments with the large caps
likewise should not exceed the time mentioned above. Usually, even
if the subject feels well, experiments should not be performed every
day, but every other day, and not more than one or two experiments
should be carried out with each eye. If all the rules are observed and
if due attention is paid to the subject, work with the caps is quite free
from risk, and the subject will feel no particularly unpleasant sensations after several experiments. The author has used a number of
subjects now for several years. No adverse results caused by work
with the caps have been observed.
The photographic material is processed by the usual methods
available in any photographic laboratory. In particular, rephotography
is sometimes used to reduce distortions, and retouching, decolorizing,
and so on may be required to increase contrast.
16. COURSE OF THE RAYS IN RECORDING EYE
MOVEMENTS BY A REFLECTED BEAM O F LIGHT

L distortions
.
We will now examine in detail the main
and errors
which may be encountered wheneye movements a r e recorded by means
of a mirror. Let us assume that:
1. The center of rotation of the eye is stationary, as is the observer's head.
2. The reflecting surface of the plane mirror lies in the center of
rotation of the eye and is firmly connected to it.
3. The surface of the mirror is always normal to the optical axis
of the eye.
4. The light source is stationary. Axial rays from the light source
pass through the center of rotation of the eye.
5. A spherical photosensitive surface is made, thecenter of which
coincides with the center of rotation of the eye.
Let us now imagine a series of planes passing through the axis of
the objective of the lightsourceandthe rotation center of the eye. For
all eye movements in which a line that is normal to the rotating mirror

CHAPTER I

Fig. 25. Scheme showing the path of the rays when reflected from a plane rotating mirror.

(and the optical axis) moves in one of the planes mentioned above, we
obtain a record onthe photosensitive surface of the sphere which differs
from the ideal only in the fact that each angle of rotation of the optical
axis is doubled by the reflected light.
In fact, let us assume that:at a certain moment of time the angle
between the incident beam of light and the line normal to the mirror
is a, (Fig. 25); the incident beam of light and the line normal to the
m i r r o r a r e situated in the plane of the drawing. Since the angle of
incidence is equal to the angle of reflection, consequently the angle
between the incident and reflected beam of light will be 2n,.
Let U s
assume further that a t some later moment the m i r r o r is turned through
a n angle fi or,whatamounts tothesame thing, a line normal to its surface is turned throughanangle /3. The angle between the incident beam
of light and the line normal to the mirror will then be a, + p = a,, while
the angle between the incident beamof light and the reflected beam will
now be Za, = 20, + 26. Hence it follows that when the mirror is turned
through an angle fi, the reflected beam of light is turned through an
angle (2a, + 2p) - Za, = 20.
Some distortion of the record is produced by the fact that the mirror
is not in the rotation center oftheeye, hut on its surface. Let us consider two cases: (a) the line normal to the mirror coincides with the
foptical axis when we use the P2o r P, type of cap (Fig. 26); and @) the
line normal to the mirror is parallel to the optical axis, but at a certain
distance from it when the Pi cap is used (Fig. 27).
Let us assume that the axis of rotation of the eye is perpendicular
to the plane of the drawing, andthatthe visual axis, which is normal to

METHODS

the mirror, and the axis of thelight source a r e also in this plane. Let
us assume that a collimated beamof lightemerges from the stationary
light source, and that its width is equal to the diameter of the eye, s o
that in both cases the rotation center of the reflected beam of light will
be shifted in space a s indicatedinFigs. 26 and 27. The rotation center
of the optical axis is stationary; the reflected beam of light, turned
through an angle twice a s great a s the angle of rotation of the eye,
receives an additional movement due to displacement of the mirror
(see Fig. 26). In the initial position, when a = 0, the beam of light reflected from the mirror situated in the center of rotation is combined
with the beam oflight reflectedfromthemirror situated on the surface
of the eye. If the eye is rotated, f o r example, through a n angle of a,
or a,, the reflected beams a r e rotated, respectively,through angles of
2a, and Zn,, and they a r e a l s o displaced through a distance of A, and A,.
A similar picture may be observed in Fig. 27. In the initial position, when a = 0, the beam of light reflected from the m i r r o r situated
a t the center of rotation is displaced relative to the beam of light
reflected from a mirror situated on the eye surface by the distance
A,. When the eye is rotated, for example through an angle of a, and a,,
the reflected beams a r e rotated correspondingly through angles of 20,
and 2,, and, in addition, they may be superposed (at a,) or displaced
'(at %, by a distance A,). However, collimated beams of light a r e not

Flg. 26. Scheme show~ngthe path of the rays when ~ f l e c t e dfrom a plane mirror rotatmg
and undergoing displacement m space, illurnmated hy a collnnated beam of light. The case
when the Pz ac P, type of cap is used.

CHAPTER I

METHODS

Fig. 29.Explanatory scheme for considering the transformationof coordinates when recording
by a beam of light reflected from a mirror.

Fig. 27. Scheme showing the parh of the rays when reflected from a plane mirror rotating
and undergoiy displacement in space, illuminated by a collimated beam of light. The case
when rhe P, cap is used.

1.k. 28.

Scheme sl~owinfthe path of the r a y s when reflected from a plane mlrror, rorarlne
2nd undergorng drsplecemcnt in space, rllumrnated by a canvereenr beam of llght.

used in the experiments. By means of the ordinary objective, the

image of a very small aperture of the diaphragm o r narrow slit is
projected on the screen, and the path of the rays in these conditions
will be a s shown in Fig. 28.
With small angles of rotation of the eye (less than 10") and a large
enough radius of the spherical surface on which the record is made,
the distortions obtained a s 'a result of displacement of the m i r r o r in
space and defocusing a r e negligibly small by comparison with the
displacemints of the image of the diaphragm aperture arising on
account of rotation of the mirror, i.e., on account of eye movement.
To represent the way in which movements of the projection of the
optical axis of the eye a r e transformed a t the plane of the object into
movements of the ray of light reflected from the mirror on to photosensitive paper, let us examine Fig. 29.
Let us assume that OA is the axis of vision, On the normal to the
mirror, OB the direction of a beam of light falling on the mirror, OC
the direction of the reflected beam of light, a the angle between the
axis of vision and the line normal to the mirror, p the angle between
the axis of the incident beam of light and the normal to the mirror,
and 0 the angle between two planes, one of which passes through the
axis of vision and the line normal to the m i r r o r and the other through
the direction of the incident beam and the normal to the mirror.
Let x and y represent the coordinates of the object to be examined,
and u and w the coordinates of the ray of light reflected on to photosensitive paper, normal relative to the ray. The axes a r e chosen s o
that Ax is the projection of the plane AOn 'on the plane of the object,
and the axis Cu is the projection of the plane nOC on the plane of the
image.
By simple geometry we reach the following relationships
between the coordinates:
u=2(xcosB-ycosasinB),

w=2(xcospsin8+ycosacos~cosB).

of the eye relative to the orbit takes place. In addition, the eye is
displaced slightly in the orbit because of the pulsation of blood, which
also gives rise to a slight pulsating deformation of the eye itself. In
most cases, however, none of these factors have much effect on the
records made by a reflected beam of light.

17. CONSTRUCTION OF CAPS

Fig. 30. Scheme for analysis of distortions obtained when recording eye movements on
plane rather than a spherical surface.

These equations a r e valid for small displacements x and y . With

larger displacements the transformation becomes more complex (not
affine, but projective).
The expressions for u and w show how distortions a r i s e when eye
movements a r e recorded by a beam of light reflected from a m i r r o r
a t certain definite values of a, P, and 8. It is easy to s e e that the.
smaller the angles o and ,8 the smaller the distortions obtained on
recording. At the limit when a = 0, P = 0, we have u = 2x and w = 2 y .
When a = Oor 6 = 0, the angle 0 is not determine. In this case we obtain
formally the transformation of the rotation throughanarbitrary angle.
This corresponds to the factthat insuch cases there is no determinacy
in the choice of the x o r u axes. By suitable choice, if a = 0 and O= 0,
the equation may be represented in the form u = 2xand w = 2y.
Usually, eye movements a r e not recorded on a spherical photosensitive surface, but on a plane surface (Fig. 30). Let the axis of
rotation of the eye be perpendicular to theplane of the drawing and let
the axis of vision, the normal to the mirror, and the axis of illumination lie in this plane. Let the reflected ray be initially a t the point A.
When the m i r r o r is rotated through an angle a, the reflected ray
moves to point B2. In this case, distortion of the recording is obtained
because the path of the ray to the plane surface (AB2)is longer than
the path traversed by t h e r a y to the sphere (ABI). It follows from this
drawing that with small angles the distortions indicated above a r e
small and may be ignored. When the record is made on a plane surface,
the distortions obtained may be calculated, and appropriate corrections
may be made when the photographs a r e analyzed.
In this discussion of eye movements, the course of the rays, and
the accuracy of the records, we have to some extent idealized the whole
picture. We have not takenintoconsiderationthe fact that some movement of the head in reality always remains. During movement of the
eye the center of its rotation is slightly displaced, i.e., displacement

The rubber parts of the caps a r e turned on a lathe with the arbor
turning at high speeds (5000-6000 rpm), from ordinary rubber (for
example, rubber stoppers). The tools required include cutting tools,
files, nail files, metal templates, emery paper, and so on. The initial
stage of work on the lathe i s done with the file, by means of which the
correct diameter of the blank is obtained. The cutting tools, the nail
files, and the metal template a r e usedtogive the part its rough finish.
Later, each rubber part is polished with emerypaper. The part itself
is continually moistened with benzene during this process, making the
rubber pliable for a short time but easily polished. In the last operation the revolving detail is cut away from the blank by a sharp scalpel.
Since the rubber details a r e turned a t high speeds, and the tool is a s
a rule unsupported except by the experimenter's hands, the appropriate safety regulations. must be strictly observed.
The holes in the rubber part of the caps a r e made either with a
very small drill (a metal tube with one sharpened end) or with a
thread. A thread is passedthroughwitha fine needle, pulled tight, and
fixed in a vertical position; the thread is soaked in benzene, and the '
rubber part is moved along i t s o that the direction of the hole and the
direction of the thread coincide. After the rubber has been moved up
and down the thread soaked in benzene several times, a hole is made
which is approxi&tely equal to thethread in diameter.
The mirrors fixed to the caps must possess good optical properties. This applies primarily to the m i r r o r s by means of which the
drift and tremor of the eye a r e recorded. T k m i r r o r s must have no
sharp edges or corners. Since the outersurface of the reflecting layer
is used in the experiments, aluminum m i r r o r s a r e more practical
than silver-coated mirrors. The silver coating quickly darkens and
loses the properties of a mirror surface.
The Pi model is assembled from, its finished part by means of
universal glue (for example, No. 88 glue) on a rigid life-size model of
the eye. Assembly on the model ensures that the m i r r o r and the
bulb of the apparatus a r e in the correct position. BF-2 glue is best
f o r affixing the paper adaptors, and for bonding glass to metal, metal

to metal, and glass to paper. No. 88 glue should he used for gluing
rubber to itself, to glass, or to metal.
The duralumin frame of the models P,-P, is turned on a small
lathe on which work can be done under a binocular loupe. Ordinary
small cutting tools a r e used. The cuttingsurfaces of these tools must
be very sharp. The sharper the tools, the thimer the parts can be
made.
It was pointed out earlier that the surface of contact between the
metal cap and the sclera is corrugatedandpolished. This corrugation
is produced on a lathe, during the first stage in manufacture of the
frame, while the blank still possesses considerable mechanical
strength. After the surface has been corrugated, the following operations a r e carried out without changing the position and angle of rotation of the support. Thirty-six marks a r e made on the cylindrical
surface of the chuck s o that the distance between two adjacent marks
corresponds t o a rotation of the arbor by 10". A cutting tool shaped
in such a way that agroove of the required shape and size is planed out
when the support is brought up is fixed i n t h e support. By using
the grooves and turning the arbor each time through lo", 36 grooves
a r e made. In the nextoperationthegrooves a r e polished, a t first each
one separately, and then all together, changing the direction of rotation
of the arbor in turn. After the corrugated surface has been polished,
the frame of the cap is then finished in theusual way.
It is sometimes difficult to obtaina short-focus lens of the required
focal length. A method of making these lenses is given below. The
wide end of a glass drop has an almost ideal spherical shape, and can
therefore be used a s the blank for making a lens. The work begins by
producing glass drops of different diameters. Thesedrops may easily
be obtained from any 'glass blower's workshop. The order of the
operations which should be followedwhen making the lenses is given
in Fig. 31. F i r s t theglass drops a r e coated with a thin layer of casein
glue, and this is allowed to dry. The glue protects the surface of the
future lens from injury. A small metal washer is then made, the hole
in which is slightly larger than thediameter of the drop, and its thickness is equal to the thickness of the future lens. A piece of glass is
glued to one side of the hole in the washer. The glass drop is placed
inside the washer a s shown in the figure, and all the f r e e space in
the hole is filled with liquid Canada balsam. The solidified balsam
holds the d r o p f i r m l y inside the washer. Next, the part of the glass
drop projecting from the hole inthewasher is ground down with emery
paper. Towards the end of this operation, emery paper of very fine
grain is used to make the plane surface of the lens smoother. A thin

'5

'7

'8

Fig. 31. Scheme toexplaintheprocesses of making short-focus lenses. Order of operations1) a glass drop i s coated with glue to protect its spherical surface from injury: 2) the drop
i s placed inside a washer and affixed to it with Canada balsam: 3) the part of the drop
projecting from the hole in the washer is ground down with emery paper; 4) part of the
Canada balsam is removed with the grains of glass; 5) the space inside the washer i s again
completely filled with Canada balsam: 6) the plane surface of the lens is polished; 7) the
lens i s carefully cleaned to remove Canada balsam and glue and then glued to a glass cover
slip with Canada balsam; 8) Canada balsam i s removed from the exposed surfaces of the
cover slip and from the spherical surface of'the lens: all surfaces are carefully rubbed
clean.

layer of Canada balsam i s then removed with a piece of cotton wool

soaked in alcohol, in order to ensure that all grains of glass rubbed
off by the emery paper a r e removed. This is necessary to prevent
scratching of the lens surfaceduringpolishingby loose grains of glass.
The space around the lens is then filled again with a drop of liquid
Canada balsam. Excess balsam is removed with cotton wool soaked
in alcohol, and the plane surface of the lens is then polished.
A piece of clean cotton cloth is draivn over the glass, it is coated
with paste, and the lens is polished by moving the washer on the cloth.
If the experimenter doesnot possess thespecial paste, paste for stropping razors may be used. After this polishing, the lens surface will
not be ideally flat, but i t should have no scratch marks or pits. After
polishing, the Canada balsam is dissolved, and the lens is taken out of
the washer. The casein glue is removed from the lens surface with
hot water. The plane surface of the lens is carefully wiped dry and in
the next operation it is glued to a previously prepared round glass
cover slip with Canada balsam. Next, the excess Canada balsam is
removed from the exposed surfaces of the cover slip and from the
spherical surface of the lens. All surfaces a r e carefully rubbed dry,
after which the lens is then ready for use.

56

CHAPTER I

It should be noted that the refractive indices of Canada balsam and

glass a r e almost equal, s o that when thelens is fixed to the cover slip
we a r e in fact replacing the not absolutely plane surface of the lens by
the plane surface of the cover slip. The lens thus obtained gives a
good image and may be used successfully in the caps.
Round glass cover slips a r e used inthe construction of some caps.
It is difficult to work with a single glass cover slip because of its
fragility. A better plan is to work with several cover slips a t once,
glued together with Canada balsam. Inthis case, the work can be done
by hand, using fine-grain emery paper. The emery paper is wrapped
around a soft pad to avoid excessive friction, which may place too
great a strain on the glass. Apaper pattern (a round piece of paper of
the required diameter) is glued tothe outer surface of one of the cover
slips, and the glass is ground down to this size.
An important detail of many of the adaptors is the very thin piece
of frosted glass which can easily be madefrom ordinary frosted glass
in any optical workshop. It may also be made by the experimenter
himself in the following manner. First, frosted glass of the required
size is fixed by liquid Canada balsam to half a glass slide. To the other
half of theglass slide is glued a rubber o r wooden block shaped to serve
a s a handle by which the material may be held during the work, and the
frosted glass is then ground on emerypaperto the required thickness.
First coarse, and then very fine, emery paper is used. The emery
paper is backed by a smooth, hard surface. It should take the experimenter no more than 20-30 minutes to grind the frosted glass.
Making some of the very small components from glass calls for
microtechniques of glass blowing, which differ slightly from ordinary
glass blowing methods. Three types of heating apparatus a r e used:
the ordinary alcohol or gas burner, a miniature alcohol burner, and a
platinum wire heated to incandescence by an electric current. The
burner in the miniature alcohol lamp is a long thin metal tube holding
a cotton wick. The tube is usually several centimeters long and 1mm
in diameter. The end of the tube is placed horizontally s o that the
glass can b e heated by placing i t not only above the flame but also
below it.
The diameter of the flame from such a burner does not
exceed 2-3 mm. If the platinum wireis being used it also is placed in
such a position that the glass can be brought near to it both above and
below. The benefits gained by this arrangement of the heating system
will be clear to t h e experimenter a s soon a s be begins to take up
microglass blowing. In many cases microglass blowing operations
should not be done by hand; it is better to employ some type of simple
micromanipulator. F o r example, to bend a very thin glass capillary

tube to a right angle, it is held with one end in a micromanipulator,

and the whole capillary tube is placed i n a strictly horizontal position.
Next, by the screws of the micromanipulator, the capillary tube is
brought smoorhly up to the incandescent platinum wire from below.
In this way. the wire and the capillary tube a r e perpendicular to each
other. For a short distance from the platinum, a small segment of
the capillary tube becomes soft and bends evenly under the weight of
the unfixed part of the capillary tube. Within a few seconds the capillary tube is bent to a right angle. This operation is much more difficult by hand. One wrong movement or tremor of the hand and the
capillary tube will be exposed to too high a temperature, or worse,
will touch the platinum. In the f i r s t case, although the capillary tube
does bend, it melts and the lumen is closed. In the second case, it
melts immediately, sticks to the platinum wire, and is converted into
a glass drop. Microglass blowing techniques demand from the experimenter knowledge of the fundamentals of glass blowing and the ability
to use simple micromanipulators.
For some experiments, diaphragms with apertures ranging from
10 to 70 p are introduced into the light source. Such diaphragms a r e
easily made from ordinary foil. Foil is placed on a sheet of glass
and gently pierced with a sharp-pointed needle. The hole thus made
is measured and examined under a microscope or binocular loupe, and
if i t is too large or too small the operation i s repeated on a fresh piece
of foil. Diaphragms withaverynarrowslit (from 10 to 70 p) a r e made
from safety razor blades. The two halves of a blade a r e fixed by BF-2
glue to a metal mount, and before the glue bas dried, the slit between
them is made the correct width using a binocular loupe.
Nearly all adaptors a r e made of paper. At first glance this work
would seem to be very simple. However, i t is one thing to glue large
pieces of paper together, but quite a different thing to glue an item
composed of very small pieces. To simplify this work, it is a good
idea to begin by soaking both sides of the paper with BF-2 glue, to
allow it to dry hard, and only then to begin forming the constructional
details. Paper previously soaked in BF-2 glue can be stuck together
by means of a smallbrushsoaked inalcohol. When the brush moistens
the point of contact between two pieces of paper, the glue on the paper
quickly dissolves and when it quickly dries it glues the paper firmly
and accurately.
With this I conclude my short list of tips for the experimenter who
wishes to make his own caps. Naturally I have omitted details of
operations and methods well known in laboratory practice, with which
the experimenter will be familiar.

58

CHAPTER 1

CONCLUSIONS
Of all the many methods whichhave beenused to record eye movements, only those which have proved most successful can now be
justified. However, none of the methods whichhave proved successful
can be regarded a s universal and perfect in all respects.
Depending on the experimenter's aims, on the conditions under
which he must c a r r y out his experiment, and finally, on the capacity
of the subject, he will select one of these methods.
Experiments recording eye movements during fixation on a point
a r e best carried out with the Pi and P3 models, for they provide very
accurate records.
If the experiment is to take some time, and highly accurate records
a r e nevertheless required, contact lenses with m i r r o r s affixed to them
should be used. and the records made a s with the Pi and P3 caps, i.e.,
by a reflected beam o f light. If, however, for some reason or other,
the subject's eyes must remain perfectly free during the experiment.
electrooculographic recording is the recommended method, despite
its low accuracy.
When the subject's eyes and head must remaincompletely free and
simultaneous records of the head and eye movements a r e required,
motion-picture photography should be used. In this case, even to
produce results of low accuracy, complicated analysis of the experimental material is required.
For producing images stationary relative totheretina,it is best to
use the PC,P1,and P8 caps.

Chapter II

PERCEPTION OF OBJECTS STATIONARY

RELATIVE TO THE RETINA
It has now been established that the provision of optimal working
conditions for the eye requires some degree of constant (interrupted or
continuous) movement of the retinal image.
This distinctive feature' of the human eye was first noted by Adrian
(1928). It was later concluded (Ditchburn and Ginsborg, 1952; Riggs
e t al., 1953) that objects stationary relative to the retina cannot be
seen by an observer all the time. Finally, it was demonstrated by the
use of the cap method (Yarbus, 1956g)that in any test field, unchanging
and stationary relative to the retina, all visible differences disappear
after 1-3 sec, and do not reappear in these conditions. The fact that
in similar experiments previous authors did not observe permanent"
disappearance of visible differences may be explained by imperfections
of their technique (incomplete stabilization of the retinal image).
In many animals, impulses run along the optic nerve in the main
only in response to a change in the light acting on the retina. If
impulses a r e regarded a s c a r r i e r s of information, it may be claimed
that in most animals the process of vision quickly comes to an end
when the retinal image is unchangedand stationary. On the other hand,
the suggestion has been made that in man the absence of change and
movement of the retinal image leads tothedisappearance o r to a sharp
reduction in the number of the impulses transmittedffom the eye to the
central portion of the visual system. Later I shall use this a s a working
hypothesis, clearly recognizing that it has not yetbeen proved because
no records have ever been made of human optic-nerve impulses.
In most experiments, the subject's second eye (theeye not directly
participating in the experiment) was covered with a black bandage not
allowing light to pass. In describing such experiments, I shall not

CHAPTER I1

mention the second eye. Cases when the subject looked a t a fixation
point with his second eye duringthe experiment or when the second eye
was illuminated with light of some kind will be examined separately.
In nearly every case, I shall give the type of cap used in the experiment. This will make it easiertounderstand the experiment, but does
demand from the reader knowledge of the construction and function of
each type of cap.
Objects whose images remain stationary on the retina whatever
movement the eye makes a r e conventionally known a s a 'stationary
test field." Objects whose images, a s a result of movement of the
objects themselves o r of the eye, a r e displaced on the retina a r e conventionally called a "mobile test field." For example, when working
with the type PC~ a p ~ i n w h i c h a s h o r t - f o c ulens
s is used, the stationary
test field is the image ofthefrostedglass of the cap, together with objects situated in its background and rigidly connected with the lens;
another stationary test field is the dark background surrounding the
frosted glass. A mobile testfieldinthis case may be an object moving
inside the cap, in front of the frostedglass, or the shadow of an object
situated between the frosted glass and the light source, and moving
across the background of the frosted glass.
When working with the type P8 cap in which, besides a lens, a
diaphragm with a n aperNre i s used, the stationary test field is the
screen o r the several screens firmly fixed to the cap. The whole of
the objectively stationary background whose image is displaced over
the retina a s a result of movement of the eyes may serve a s a mobile
test field. In addition, just a s in the experiment with' the Pscap, an
object moving over the background of the screen may be used a s a
mobile test field.
A stationary and unchanging test field in which all visual contours
have disappeared for the subject may be termed an "empty field"
arising in artificial conditions.
Often a s a result of fixationona large and uniform enough surface,
conditions of constant illumination a r i s e within the retinal image of
this surface. , If the constancy of illumination lasts more than 2-3
seconds, the inner part of 'such a surface may be called an "empty
field" arising in natural conditions.
'

1. FORMATION OF AN EMPTY FIELD

In a large s e r i e s of experiments withthe PCcap, the subjects were
shown different stationary and unchanged test fields for perception,
each of which occupied the whole field of vision of the eye. The test

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

fields differed in their angular dimensions, shape, color, and contrast

between details. The maximal visible brightness of the frosted glass
of the cap was 3000 apostilbs,* when the diameter of the aperture of
the diaphragm was 1-2 mm.
The first question to be answered was: does an empty field always
appear in these conditions, i.e., do allvisualcontours disappear? The
experiments showed that in every case, 1-3 seconds from the beginning
of the experiment and after removal of all light of varying intensity
passing through the sclera, all visual contours disappeared from the
subject's field of vision. These contours did not reappear until the
end of the experiment, i.e., for several minutes (unless something
happened to disturb the constancy and strict immobility of the retinal
image). The color of the empty field remained unchanged. Usually
the subject will call this color "black," "dark grey," "dark,' or
"twilight," a s when the eyes a r e closed.
These results have led certain investigators to consider that the
empty field should be called black. However, this immediately raises
the question: what does the observer see if, after an empty field has
been formed, a black object moves against its background? Since the
object moves, it must be visible, but a black object cannot be seen
against a black background (if in fact it i s black). To solve this
problem, a series of experiments was carried out using the Pscaps.
The test field was a bright white circle surrounded by a black
diaphragm with which it merged after theformationof the empty field.
It was found that when a black object moved across the background of
an empty field, the observer saw the object a s black (much blacker
than the empty field) and realized thatitwas a mistake to describethe
empty field a s black. Later experiments showed that an object of any
color, moving across the background of an empty field, can be distinguished from the color of this field. These results showed that the
subjective description of the color of the emptyfield (occupying the
whole field of vision of the eye) is always conventional, for by tlirect
comparison i t is distinguished from any other color. The problem of
the color of the empty field will be examined below.
Jf the cap were slightly displaced during the experiment (because of accidental contact with the eyelid o r by a deliberate gentle'
tap) all the contours of the test field reappeared instantaneously. Jf
the bandage was removed from the subject's second eye during the
experiment and he opened this eye for perception, the character of
the perception was just a s if the eye to which the cap was fixed had
been closed.
*Editors note: This is equivalent to a luminance of 300 millilamberts.

CHAPTER 11

Fig. 32. Scheme of stationary test field. The filament of an incandescent lamp seen by the
subject through an aperture in a diaphragm fixed near the lamp.

The processof formation of the empty fieldand the field itself a r e

extremely sensitive to the slightest disturbances of the strict immobility or constancy of the retinal image. Attention is particularly
drawn t o the pbssibility that lightmay enter the eye through the sclera.
Even if this light is constant, the illumination of the retina is fluctuating
because of the constant movements of the eye. It is therefore essential
to watch that during the experiment the bright beam of light falling on
the frosted glass of the cap does not illuminate the sclera. That is
why the modification in the construction of the adaptor for the Pscap,
which I have described, was introduced; this allows the frosted glass
of the cap to be illuminatedfrom thenasal side, thus leaving thesclera
in almost complete darkness.

2. PERCEPTION OF VERY BRIGHT OBJECTS STATIONARY

RELATIVE TO THE RETINA
In the experiments described in the previous section, the test
fields were of low o r average luminance. It was important to verify
whether the differences in the stationary and constant test field disappeared if isolated parts of the field a r e very bright, almost blinding.
To investigate this problem, a special adaptor was made for the P8
cap. This adaptor includeda small electric lamp connected to a source
of current by very. thin wires. The adaptor was placed in such a
position that the lens of t h e P, apparatus gave a sharp image of the
lamp filament on the retina. The image of this stationary test field is
shown schematically in Fig. 32. The thickness of the lamp filament

PERCEPTION'OF OBJECTS STATIONARY RELATIVE TO THE RETINA

subtended a visual angle of 15 minutes. The brightness of the filament could be changed during the experiment by changing the voltage of
the supply. Most of the lamp was covered with black paper, s o that no
light from it fell on the sclera, and when the experiments were carried
out in a dark room there was thus no possibility of light entering the
eye through the sclera. The wires supplying the lamp were so placed
that they did not interfere with the eye movement o r displace the cap
on the eye. Before the experiments the subject received atropine to
dilate the pupil and immobilize the iris. To reduce eye movement, the
subject was given a fixation point for his free eye.
The experiments with this adaptor showed that when the elements
of the stationary test field a r e of dazzling brightness, all visual
contours disappear from the field. During the experiment the incandescent filament of the lamp disappeared for the swbject 1-3 seconds
after the test field had become strictly constant and stationary. In
these circumstances the subject saw orlly the point fixated by the second
eye. If the lamp was switched off after the empty field had developed,
a transient appearance of the lost contours reappeared briefly, the
filament of the lamp appearing to be of dazzling brighmess.
The results of the experiments described in sections 1and 2 of this
chapter lead to the following conclusion: if a test field (of any size,
color, or brighmess) becomes and remains strictly constant and stationary relative to the retina, then all contours in the field will disappear after 1-3 seconds and will not reappear.
I claim that the contours in the test field do not reappear because
their disappearance lasts f o r the duration of the experiment, i.e., f o r
several minutes. In addition, Iam taking into consideration the result4
of the experiment described in section 4 of Chapter I. I recall, for
example, the experiment in which the vessels of his own eye become
visible to an observer during oscillatory movements of a point source
of light, i.e., when the shadows of the blood vessels situated close to
the retina a r e in motion. If the movement of the light source stops,
the vessels disappear within 1-2. seconds and do not reappear a s long
a s the light source remains stationary.
The method generally believed tocreate a stationary retinal image,
using a contact lens with a mirror attached to i t (see section 12 of
Chapter I), has not allowed experimenters to .obtainlaking disappearance of contours from a test field. Usually the contours have disappeared for a few seconds, reappeared for a few seconds, and then
disappeared again. Because of these findings, certain authors a r e
doubtful that the contours of a stationary object can disappear for
any considerable time. The experiments I have just mentioned, es-

CHAPTER I1

Fig. 33. Scheme of the stationary rest field. The black rhread is seen by the subject against
the background of the frosted glass of the cap through a round hole in the diaphragm.

pecially those with the cap, have shown that the constant appearance
of contours of the stabilized image during work with the contact lens
can be attributed only to incomplete stabilization of the retinal image,
a defect making the work of experimenters extremely difficult, and
often making definite conclusions impossible.
This repeated disappearance of visible contours of a stationary test
field (or sharp reduction in the resolving power of the eye) has been
reported by many authors (Ditchburn and Ginsborg, 1952; Riggs et al..
1953; Ditchburn and Fender, 1955; Krauskopf, 1957; Ditchburn et al..
1959; Ditchburn and Pritchard, 1960; Clowes, 1961; Ditchburn, 1961).
The important feature s o far a s we a r e concerned i s that the results
of all the investigations mentioned above confirm the' important role
of eye movements in vision.

3. PERCEPTION O F OBJECTS O F VARYING LUMINANCE

STATIONARY RELATIVE T O THE RETINA
In this section an attempt is made to establish the minimal changes
in the intensity of a light source a t which a subject begins to see
contours in a test fieldwhen the fieldis constantly stationary. The use
of the P6cap ensured that the test field was Stationary.
The stationary test field consisted of a round hole in a piece of
black paper, intersected by a thin black silk thread. The subject saw
the hole and the thread against the background of the frosted glass of
the adaptor under angles indicated on Fig. 33.
The luminance of the test field (a circle) a t which the subject looked
through the diaphragm of the cap, 1.5 mm in diameter, was measured
in apostilbs.
The frosted glass of the cap was illuminated with a beam of light
from a n incandescent lamp. The illuminance of the frosted glass was

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETlNA

65

varied (either increased or diminished) linearly withtimeby means of

a wedge placed between the light source and the cap. This adjustment
was made by a rotating disc with a wedge-shaped slit. The rate of
change in illuminance of the test field could be regulated by changing
the speed of rotation of the disc. By changing the speed of movement
of the wedge, the experimenter could produce the necessary change in
luminance of the test field.
For any initial luminance of the test field I, it was easy to select
rates of change of its luminance for which the subject saw the test
field for a certain period of time clearly or very weakly, o r did not
s e e it a t all. In these circumstances the interval during which the test
field was seen was found to be a fraction of a second o r might even
exceed one second.
With a low enough value of dl/dt, it had the
appearance of a hardly detectable circle, the visual brighmess of which
increased with an increase in the value of dl/dt, after which the individual parts of the thread and, finally, the whole thread showed up
against its background. Whenthe subject reported the appearance of all
contours in the test field, the phase during which he.distinguished only
certain parts of the thread- against the background of the circle (the
thread a s a whole had not yet appeared) corresponded to a rate of
change of luminance of the test field which I conventionally call "the
threshold speed."
Knowing the initial luminance of the test field (I,), the time of
movement of the wedge ( 0 , and the final luminance ( I ) , the rate of
change of luminance, i.e., dl/dt, could always easily be determined.
In fact, since the luminance varied linearly with time, it was always
true that (I-l,)/dt = dl/dt.
Evidently when the pupil remains unchanged, the retinal illuminance
(H) and its variation (dH/dt) bear a linear relationship to the corresponding luminances of the test field and to their variation.
As pointed out'above, contours within the test field, in response to
a variation of its luminance, were not noticed instantaneously by the
subject but after a certain time lag.(afraction of a second). This time
will be denoted subsequently by the Greek letter r. Preliminary experiments showed that the value of r is not constant but depends primarily on the value of (dH/dt)/H. However, I shall not examine this
question in detail.
Most of the measurements were made on two subjects. First I
attempted to discover how the appearance of contours in the test field
depended on the direction (the sign) of the change of luminance of this
field. The experiment showed that if a test field of a certain. arbitrary
luminance (I,) changed into an empty field a s a result of its immobility

CHAPTER 11

c
'
I

0 :

ZOO

400

600

BOO

Apostilbs
Fig. 34. Graph showing relarlonshlp between threshold rate of change of lurnlnance of a ten
fleld drldt and the luminance (I.) of thls held.

relative to the retina, i t reappearedforthesubject when the luminance

was either increased o r decreased. The threshold rates were approximately equal in their absolute value for an increase and a dec r e a s e in luminance. . During an increase in luminance the color of
the circle appeared orange to the subject; during a decrease it appeared bluish o r even blue.
Later a n attempt was made to determine whether the threshold
speed of change in luminance (from a constant I.) depends on a series
of conditions preceding the measurement-for example, whether it
changes a s a result of the preliminary action of a constant, bright
light, o r a s a result of dark adaptation. The experiments showed that
a t the time the measurement was taken, i.e., 30-40 seconds after the
formation of the empty field, these conditions hadno significant effect
on the magnitude of the threshold rate.
The aim of the following experiments was to study how the magnitude of the threshold rate of change of luminance depends on the initial
luminance of the test field (I.). T h e results of these experiments a r e
given in Fig. 34. This figure shows that over the range of luminances
from several apostilbs to 1000 (the 'diameter of the hole in the diaphragm was l.5 mm), a linear relationship exists between these values.
As the luminance I, increases, the threshold rate of change of luminance dl/dt increases proportionately. This result corresponds to the
Weber-Fechner law. It follows from the foregoing remarks that the
ratio between the threshold rate and the absolute luminance of this
field is constant, i.e.,
threshold (dl/dt)/I.

constant

The values of these constants a r e shown in Fig. 35. In the present

case these ratios, a s the graph clearly shows, were approximately
equal to 0.3 sec-I.
This means that the contours in a test field

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

67

stationary relative to the retina begin to be perceived by the subject

when the luminance of this field changes a t the rate of 30% per second.
This figure remains constant throughout the full range of investigated
luminances of the test field.
The value of the threshold rate thus determined may appear to be
contrary to everyday experience, for we can detect changes in luminance taking place much more slowly thanat the rate of 30%per second.
However, the contradiction is only apparent, for we a r e discussing
completely diEerent processes. Under ordinary conditions of ohservation, the illuminance of individual elements of theretina is continually
changing a s a result of eye movement, regardless of whether.
and a t what rate, the luminance of the observed object changes. The
figure of 30% per second is the rate of change of illuminance of an
element of the retina a t which signals appear in the corresponding
nerve fiber. The changes in the luminance of objects which we perceive
depends on how different these signals a r e when they have already
appeared.
It was next found that when the ratio (dl/dt)l, = 1 sec-', after a
period of time i, the subject saw the test field perfectly clearly (a
thread with a thickness of three minutes of angle was seen absolutely
clearly). In these circumstances this ratio remained unchanged for
the whole range of investigated luminances (from several to 1000
apostilbs).
Bearing in mind that, with the pupil unchanged, the illuminance
and the change in the illuminance of the retinal image bear a linear
relationship to the luminance and the change in luminance of the test
field, that through the investigated range of luminances the contours
in the test field were clearly seen by the subject when

and that

(&/dt)/H

dlnH/dt ,

0 0

300

400

600

800

Apostilbs
Flg. 35. Graph showmg ~elationsbpbetween the ratlo (dr/dt)/l. and luminance (5)
of the
test fleld, where d l / & a the threshold rate of change m lumrnance of the rest field.

CHAPTER 11

68

the result obtained may be written in the form of the equation:

din H/dt

= 1 sec-'

Consequently, when the absolute,value of the derivative of the

natural logarithm of the illuminance in relation to time becomes and
remains larger than unity (dlnH/dt > Isec-I), after a certain interval
r the subject clearly sees that particular (Fig. 33) test field stationary
relative to the retina.
In the present case (Fig. 33). the light circle began to appear a t
slower speeds than the thread, but in both cases the relationship
(dH/dt)/H, = const holds good. On this basis it may be concluded that
this relationship will always be observed, even though the threshold
speed may differ from one case to another.
If the change in illuminance takes place smoothly and over a long
but dlnH/dt is constant, the corresponding element of the
period (>TI,
test field will become visible when dlnll/dt I const. If, on the other
hand, the change in H occupies a timer,, which is small in comparison
with 7 , the degree of visibility of the element will be determined, not
by dlnH/dt, but by AHiH, i.e., by the extent to which H has changed.
'

4. P E R C E P T I O N O F FLICKERING OBJECTS STATIONARY

RELATIVE T O T H E RETINA
The experiments suggest that vision is possible only when light of
varying luminance or. spectral composition acts on the elements of. the
retina. We shall now try to discover to what extent these conditions a r e
not only essential, but also adequate.
Usually, in the process of vision, the image of the object is constantly moving over the retina a s a result of eye movement. Because of
this, the illuminance of the elements of theretina changes. The question arises: does this movement of the retinal image itself play a role
in the process of vision? Canwe,by excluding this movement, provide
good conditions for perception by using varying illuminance? It is
well known that during a brightflash of light, a person can detect many
of the details and shades of color of an object. In these circumstances
the duration of the flash may be soshort that the corresponding retinal
image must be practically stationary relative to the retina, and consequently, the eye movement cannot have taken part in the process of
vision. In addition, the facts described in the preceding section show
that the eye notices very small details of a stationary test field if its
luminance varies sufficiently in the course of a short enough time.

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

69

However, a single flash or a single change in the luminance of a

stationary test field (when dl/dl retains its sign) cannot provide
satisfactory conditions for perception. Inthe firstcase the subject has
insufficient time, and in the second the color of the stationary test
field is distorted, with either increase o r particularly, decrease in
luminance. For this reason, tofacilitate the perception of a. stationary
test field, the last possibility was used, illumination of the field with
flickering light.
In most experiments the visual test object was the test field shown
schematically in Fig. 36. Pieces of completely opaque black film, a
sharp image of which was seen by the subject against the bright frosted
glass, were fixed to a very thin glass cover slip placed in a PCcap:
Sometimes, instead of black film, brightly coloredgelatin (transparent)
films were used. In all cases the diameter of the test field, i.e., of
the bright round background, was 17".
The flicker frequency varied from 1 to 50 cycles. During each
cycle light and darkness were equal in duration. Each increase of the
light to a maximum. and decrease to complete darkness took not more
than 0.005 sec. The luminance.ofthefrosted glass of the cap was 3000
apostilbs. The diameter of the artificial pupil (the diaphragm of the
cap) was 1mm.
First an attempt was made to discover how the resolving power of
the eye changes with a change in the frequency of the flicker on the
frosted glass of the cap when the test field (illustrated in Fig. 36) is
strictly stationary relative to the retina, and the sclera is in total
darkness.

Re. 36. Scheme of

.rnrlon.ry

test fleld. Flve black spots seen by the sublecr agarnst a

CHAPTER I1

It was found that s o long as the flicker frequency did not exceed
four cycles, the subject observed all the black spots against the background of U l e flickering circle. When the flicker frequency was five
cycles, the subject could no longer detect the smallest black spot (with
a diameter of 1"). It had become pale and was indistinguishable from
the flickering field. With flicker of six cycles, the subject observed
only the largest or the two largest spots of the test field (with diameters
of 6and4"). Withflickeratthe rate of 7-9 cycles, the subject no longer
saw any of the dark spots on the test field, and saw only the light
flickering circle (diameter 17').
When the flicker frequency was
10-11 cycles, the subject again began to s e e the two largest (sometimes the three largest) spots, but a s the critical flicker frequency
was approached, they again disappeared.
At the critical flicker
frequency (in this case about 30 cycles), the circle disappeared, and
the whole field of vision became an empty field.
The following conclusions may be drawn from the results of these
experiments: first, with a n increase in thefrequency of the flickering
light the details of the test field, stationary relative to the retina, do
not all disappear a t once as the flicker frequency increases-the smaller
details a r e the first to disappear; second, a range of flicker frequencies
(7-9 cycles) exists in which the resolving power of the eye is a t its
lowest (disregarding the region close to the critical flicker frequency);
when the flicker frequency is above the critical level, an empty field
appears just as during continuous illumination. Inevery case (whatever
the flicker frequency), the resolving power of the eye was much below
normal.
In the f i r s t series of experiments the sclerawas in total darkness.
Let us now consider how illumination of the sclera influences the
results of the same experiments, when the light falling on the frosted
glass and that falling on the sclera flicker synchronously.
In this experiment, the part of the lightbeam not falling directly on
the frosted glass of the cap was utilized (the beam of light directed
towards the frosted glass was always wider than theglass). By
placing a sheet of white paper on the subject's temple, an illuminated
screen was obtained, scattered light from which fell on the sclera and
flickered synchronously with the flicker on the test field.
In these conditions, the black spots of the test field b'egan to disappear more quickly and a t rather lower flicker frequencies. With a
frequency of 7-9 cycles, when the screen was moved s o close to the
eye that the sclera was. well illuminated with scattered flickering
light, not only the black spots of the test field disappeared, but also
the 17" bright circle itself. The subject saw only the flickering light

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

and could detect none of the details in the visual field. The light
reflected from the paper and falling on the sclera was no brighter
than the light falling on the sclera in the room from the ordinary
daylight or artificial light. These results suggest that with adequate
brightness of scattered light, in the conditions described above, the
eye becomes incapable of perceiving any details in a test field.
Further experiments were carried out to determine the resolving
power of the eye when a flickering light is suddenly switched on (after
complete darkness), and to ascertain its changes in time if the flicker
frequency remains constant.
In these experiments, the flicker frequency was 8 cycles. The
flickering light was switched on or off by a shutter fixed to the light
source. When the flickering light was switchedon, the subject saw all
elements of the test field sharply. The circle appeared white, with
a noticeably yellowish hue, and the black spots appeared black. Next,
f o r several seconds, the black spots gradually become paler and
disappeared on the flickering circle. When the flickering light was
switched off, the subject saw the usual after-image: for a period of
1-2 seconds the circle appeared black, with a noticeably bluish tinge,
while the black spots looked grey and much lighter than the background.
These results suggest that if the retinal image is stationary, a single
illumination of the object, even for a short time, allows the eye (depending on the luminance and duration of illumination) to resolve the
small elements of the object. Periodic repetition of the same type of
illumination (particularly a t certain frequencies) will cause a sharp
decrease in the resolving power of the eye.
In the next series of experiments, the effect of constant illumination
on the resolving power of the eye was studied in conditions when the
test field flickered a t frequency of 5 cycles. Before the light was
switched on, the subject usually saw a t least four black spots against
the background of the flickering circle. Illumination was provided by
a bright beam of light, directed a t the required monient to the exposed
part of the sclera (in these circumstances orange light scattered by
the sclera fell on the retina).
In these experiments, the subject perceived a flash of orange light
when the constant illumination was switched on. If the illumination
was very bright, all the details of the test field disappeared a t once
(sometimes even the sensation of flicker disappeared). This was
followed by the appearance of the flickering backgroundand the larger
spots (three or four) of the test field. In every case after the light
was switched on, the black details were distinctly orange in color and
the flickering circle appeared bluish. By comparisonwith the moment

'

CHAPTER 11
before illumination, the resolving power of the eye was slightly lower
(it fell as the intensity of illumination increased). Switching off the
light was perceived a t once as an improvement in the resolving power
of the eye. All five black spots were seen on a white circle, and after
a few seconds, the 1" spot again disappeared.
In one experiment a n attempt was made to determine the effect of
the brightness of the flickering light on the resolving power of the eye
when the test field remained stationary and the flicker frequency
constant (5 cycles).
These experiments showed that any increase
beyond a certain luminance of flicker was accompanied by a decrease
in the resolving power of the eye. The black spotsof the test field
became lighter, and the visual contrast between spots and background
diminished.
Evidently, a s the luminance of the flicker increased, the amount of
light scattered by the optical system of the cap and the eye also increased. Jn these circumstances, part of the scattered light fell on the
retinal image of the black spots. When the scattered light falling on
the retinal image of the black spots reached a certain level of brightness, it began to have the effect described earlier in the experiments
in which the sclera was illuminated by flickeringlight. This evidently
explains the decrease observed in the resolving power of the eye.
The color o f t h e flickering light falling on the frosted glass of the
cap was changed in a series of experiments with a stationary test field.
This procedure did not produce any marked change in the resolving
power of the eye.
When the flicker frequency dropped below 3-5 cycles, conditions
frequently developed in which the test field appeared negative to the
subject. The black spots appeared light against the background of a
dark flickering circle. This phenomenon developed particularly frequently when the luminance of the flicker was reduced and when the
light thrown on the frosted glass of thecap was a saturated blue color.
If, instead of black spots on a white background, transparent
colored films were used; even with ahighsaturation and large angular
dimensions (7-a"), they disappeared for the subject sooner and faster
than the black opaque spots (in otherwise identical experimental conditions). The color of the flickering circle acquired the hue of the
films occupying the greater part of the a r e a of the circle. Switching
off the flickering light (of different frequencies) led to the appearance
of clear and saturated after-images.
In analogous tests made with the subject's second eye (without the
cap) open, the elements of the test field invariably disappeared more
rapidly and the resolving power of the eye was even further reduced.

PERCEPTION OF OBJECTSSTATIONARY RELATIVE TO THE RETNA

This phenomenon was observed evenwhen the subject fixated on a small
fixation point in a completely darkened room with his f r e e eye.
Use of a flickering light that changed smoothly from light to dark
and from dark to light on the frosted glass had no significant effect
on the experimental results.
In conclusion, we may mention again that when an empty field is
filled with a flickering environment, avery importantrole is played'by
illumination of the retina with scattered light (light scattered by the
sclera or even by the transparent media of the eye).
It is therefore clear that a constant (continuous o r interrupted)
movement of the retinal image is essential to ensure satisfactory
working conditions for the human eye, andthatthis cannot be produced
by any method of illuminating images stationary relative to the retina.
Ditchburn and Fender (1955) carried out experiments with a flickering image stationary relative to the retina.
The test field was a
flickering circle divided by a black line. These authors claim that the
most favorable conditions for perception were obtained a t the.critica1
flicker frequency. These claims do not agree with my own findings
but would appear to be attributable to the inadequacy of the methods
used.

5. PERCEPTION O F OBJECTS O F CHANGING COLOR,

STATIONARY RELATIVE TO THE RETINA
We shall now consider to what extent the apparent color appearing
on an empty field during a change in illumination depends on this
change and to what extent it depends on the color of the background
(the originally empty field), against which this change takes place.
In the experiments the test field was the frosted glass visible to
the subject in h e P6 cap against a black background. The frosted
glass was rectangular, and its sides hadangular dimensions of 14 x 7".
To the outer surface of the frostedglasswere fixed two polaroids with
mutually perpendicular planes of polarization. The border of contact
between the polaroids divided the frosted glass into two squares. By
directing light onto the frosted glass from two sources covered with
filters and suitably oriented relative to the polaroids, the experimenter
could change the color of the two halves of the test field independently
and a t will. The polarization planes of the polaroids on the frosted
glass (perpendicular to each other) were parallel respectively to the
polarization planes of the polaroids of the light sources, s o that the
color of each half of the test field was determined entirely by light

CHAPTER 11
from one source. In addition, nonpolarized light from a third source
was thrown on t o the frosted glass of the cap. By means of this third
light source, the experimentor was able to add an equal amount of
light to the two halves of the test field and then to modify i t a s necessary. Naturally, the changes in this light were always equal for the
two halves of the test field.
The maximal luminance of the frosted glass of the cap did not
exceed 3000 apostilbs when the diameter of the hole in the diaphragm
was 1.5 mm. Let us assume that the color of one half of the test field
was A, and that of the other half B. At a given moment, a s a result of
its immobility relative to the retina, the test field becomes empty.
Now to both fields an equal addition of light C is made, subthreshold
f o r both A and B, i.e., the colors A and B simultaneously become
A' = A + C, and B' = B + C. This addition is made immediately after
formation of the empty field, i.e., whenthe state of the eye has changed
only slightly on account of adaptation. Duringthis change in illumination, both fields must appear against the background of an empty field.
If their color depends entirely on the differences between A' -A and
B' - B , i t must therefore depend entirely on C, and both fields must
therefore appear t o be always equal, whatever the nature of A, B, and
C. Conversely, if the fields appearing after the addition of the third
color to two different colors A and B a r e different, then however this
difference might be expressed, it shows that the signal which is formed
depends not only on the nature of the added color C, but also on the
color to which the addition was made.
When the experiment is carried out in this way, it is very important
that the subject be asked only one question: a r e the fields appearing
after the addition the same or a r e they different? He must not be asked
to evaluate this difference. Verbal descriptions of differences between
two colors inevitably possess a subjective quality; the presence of a
difference of some s o r t a s a rule is a much more reproducible factor.
I shall describe an experiment (Fig. 37) in which one half of the
test field was saturated red and the other half saturated green. After
the test field had become a n empty field,a beam of pale blue light was
thrown in addition on both halves of the frosted glass. At the moment
the blue light was switched on, the test field appeared to the subject
to beauniformbluecolor. Next,after a period of one' to three seconds,
this color disappeared and the empty field reappeared.
In actual fact, when the blue color was switched on, although the
subject saw the test field a s uniformly blue, the red half of the test
field turned crimson and the green half bluish-green. Consequently,
a g r e a t difference in color remained between the two halves although
it was not perceived by the subject.

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

75

The following conclusion was drawn from experiments such a s

these on the discrimination between colors A, B, and C. Any subthreshold (for A and B) increase in the light falling on an empty field
reveals no differences in a test field if the experiment is carried out
in conditions in which the state of the eye has not changed appreciably
because of adaptation. At the momentthelight is increased, the whole
test field appears to be uniformly the color of the added color C. In
other words, the color seen by the subject a t the moment of an abovethreshold increase in the light falling on the empty field is determined
purely by this added light and is independent of the background against
which this addition is made.
However, the facts described above a r e valid only for an increase
in the intensity of illumination, but they donot hold good when the added
light is removed, such a s during the change from A' = A + C and
B' = B + C to A and B. In this case ("negative additionn), both fields
appeared different, and their color depended both on the initial colors
A and B and on the added color C.
m e results of a series of experiments in which the colors A, B.
and C were changed were a s follows. Any supra-threshold decrease
in the light falling on the empty field revealed contours in the test
field, but in these circumstances the visible color of the test field was
distorted, i.e., i t did not correspond to any of the colors A, B, and C.
These simple rules were verified for the different colors A, B.
and C on several subjects, but one important reservation must be
made. For a subject to s e e a certain change in brightness against the
background of a test field, the change must be above a certain
threshold itself, in accordance with the Weber-Fechner law, increases,,
with the luminance of the test field. I showed earlier (in section 3 of
this chapter) that changes in the luminance of the empty field essential
for revealing differences in the test field also correspond to the
Later I tried to observe this in conditions
Weber-Fechner . law.
analogous to those just described; the experimental results were
purely qualitative in character.
In these experiments A, B, andC had the same spectral composition
but differed considerably in brightness. If the brightness of A was
much greater than the brightness of B, the added brightness C was
chosen to be above the threshold for B but below the threshold for A.
In this case, when the addition C was switched both on and off, only
the less bright field appeared. This perfectly natural result qualifies
the conclusion that during the positive addition, the fields to which
light is added remain indistinguishable. For this reason, in describing
the e ~ p e r i m e n t I, have emphasized that the additions were deliberately
chosen to be supra-threshold for both fields.

76

CHAPTER 11

The threshold of visible changes of light depends on the background

The important feature is that this
against which they take place.
relationship is maintained even when this background cannot be perceived by the subject because of iTs immobility relative to the retina.
In the next experiment I attempted to discover whether short
flashes of light and short interruptions of light a r e perceived, and if
so, how, under the conditionsdescribedatthe beginning of this section.
Let us assume that one half of the test field is color A and the other
half color B. Next, after the appearance of an empty field, a flash
of light is thrown over the whole field, a s a result of which during the
time of the flash the halves of the test field have the color A' = A + C
and B' = B + C.
In all cases when the duration of the flashes lay
between 0.01 and 0.05 sec, the subject did not notice the differences
in the test field, and the apparent color of the flash appeared to be
color C. When the duration of the flash reached I sec, differences
again appeared between the two halves of the test field following a very
short but perceptible interval of time after the light was switched off.
Let us now assume that one half of the test field is the color
A' = A + C, and that the other half is the color B' = B + C. After development of an empty field, component C is switched off for a short
period of time. In this s e r i e s of experiments the exclusion of component C f o r some hundredths of a second was perceived by the subject
only a s some kind of change in color (he could not give the color a
name). The exclusion of component C for several tenths of a second
was perceived by the subject a s flashes of light, in apparent color
resembling supplementary color C. If component C was excluded for
1 sec, the subject then saw the difference between the two halves of
the test field.

6. CHANGES IN THE STATE O F THE RETINA A F T E R

FORMATION O F AN EMPTY FIELD
When an empty field forms, all differences in color apparent to
the subject gradually disappear until evenNally the colors a r e indistinguishable. In the beginning this diminution of all visual contours
resembles. a t least externally, the familiar phenomenon of adaptation
developing when the eye fixates onacoloredobject of one color against
the background of another color. As a result of adaptation, even with
slight eye movement, the apparent colors of the object become paler
(lose their saturation) and closer together. Admittedly, this process,

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

77

when observed without the cap, takes place much more slowly than
formation of an empty field; indeed, it hardly ever progrexses to the
complete disappearance of visual contours. But the evident similarity
between the two phenomena has led many investigators to apply the
term "adaptationn to both processes. Originally, I also considered
these phenomena to be identical, but I was puzzled by the great difference in speed of the changes observed. For this reason, I wondered
whether the process of adaptation ends when an empty field forms or
whether it continues after the field is formed.
The P s c a p was used in this series of experiments. The stationary
test field was a light rectangle on a black background (the rectangular
piece of frosted glass). The height of the rectangle was 20" and its
width 10". and the color of the black background remained constant.
The color of the rectangle was changed by means of filters.
In the first series of experiments in which different filters were
used, the bottom half of the rectangle, acting a s stationary test field,
was shielded by black paper opaque to light before each experiment.
Initially, the subject saw a bright square against a black background
(the top half of the rectangle), but an empty field developed after
1-3 sec. For the subject, the bright square was merged with the
background. The subject remained in this state for 1-2 min, after
which the black paper was removed carefully from the bottom half
of the rectangle s o a s not to disturb the immobility of the image. The
subject saw a bright square for 1-3 sec (the bottom half of the rectangle), after which the square again appeared to merge with the dark
background, i.e., an empty field developed when both halves of the
rectangle had the same intensity of illumination.
The light was
switched off after 2-3 sec, and the subject saw two quite different
after-images, belonging to the top and bottom halves of the rectangle.
In one of the experiments, s h o w schematically in Fig. 38, the
rectangle acting- a s stationary test field was screened by a red filter
and, in addition, its bottom half was covered with a neutral filter
absorbing 85%of the light. Initially, the subjects saw a red rectangle,
the top half of which was much brighter than the bottom half. Subsequently (after 1-3 sec) an empty field appeared (both halves of the
rectangle merged with the black background), and the subject remained
in this state for 2 min. At the end of this period, the red and neutral
filters were removed a t the same time, with c a r e taken not to disturb
the immobility of the retinal image, and the rectangle was illuminated
with white light. After this change, the subject saw the rectangle
reappear a s pale blue (a shade similar to the added color) and
'

78

CHAPTER I1

consisting of two sharply different halves. The top half of the rectangle, i.e.. the half illuminated previously by the brighter red light,
appeared bluer and darker than.the bottom half. In these new experimental conditions the rectangle again disappeared after 1-3 sec, i.e.,
an empty field again developed. Immediately after this, the light was
completely switched off and an after-image appeared, a dark blue
rectangle, also consisting of two sharply different halves, but this
time the top half appeared lighter than the bottom half.
Similar results a r e obtained with filters other than the red.
Naturally, the color added will be complementary to the other color.
We may conclude from these experiments that two different process e s exist: the firstis a "fast" process of disappearance of all contours
in the stationary test field, and the second anslow" easily demonstrated process-by means of after-images, for example. The next
experiment was undertaken to make long and continuous observations
on the slow process.
A red filter made of gelatin film was placed between the frosted
glass but did not touch it. It was big enough to cover the whole visual
field whatever movements the eye made. Since movements of the filter
could not be seen by the eye through the frosted glass, they could not
prevent the formation of an empty field. A hole was punched in the
gelatin film, through which a beam of white light fell on the rectangular
frosted glass, and a s a result of continuous movement of the eye, the
subject saw this beam the whole time a s a light spot (Fig. 39). Movement of this spot in the field of vision did not prevent the appearance
of a n empty field. Theemptyfieldappeared 1-3 sec after the beginning
of the experiment.
When the empty field first appeared, the spot
was seen a s a bright white against a dark grey background. Then,
because of the slow process, for a period of 30-40 sec the apparent
color. of the spot changed appreciably; a t the end of this period i t
was a saturated blue.. Similar results were obtained with filters of any
other color. The spot, white when the empty field f i r s t appeared, later
acquired a saturated color approximately complementary to the color
of the stationary test field. The results of these experiments suggest that after the appearance of an empty field the action of constant
stimuli stationary relative to the retina substantially alters the state
of the retina-the magnitude and character of its reaction to the same
radiation a r e modified.
The slow process is not observed by the
subject after appearance of the empty field, sincethere a r e no signals
in the optic nerve. When asignalappears, it is influenced by the state
of the retina, modified a s a result of the slow process.

PERCEPTION O F OBJECTS STATIONARY RELATIVE TO THE RETINA

79

7. PERCEPTION OFOBJECTS STATIONARY RELATIVE TO

THE RETINA AND OCCUPYING PART O F THE VISUAL FIELD
This section describes experiments in which the subject's eye
perceived stationary and mobile test fields simultaneously.
For these experiments a Ps cap was used, fitted with screens
(stationary test field) of different colors andsizes. Usually the screen
consisted of two halves of different colors, most frequently black and
white. The background (mobile test field) against which the screens
were seen consisted in some cases of aplain sheet of paper (all of one
color), and in other cases of a colored mosaic-a sheet of cardboard
with pictures from Ostwald's color atlas glued to it.
Just a s in the preceding experiments, all visual contours within
the stationary test fielddisappeared within 1-3 sec after the experiment
began, and it became a visually homogeneous empty field. The apparent
color of the moving test fielddidnotchange. When the angular dimensions of the stationary test field were less than those of a solid-color
moving test field, and the firstfieldlaywithin the second, the apparent
color of the empty field merged within a few seconds with the color of
the moving test field. In other words, if the shutter of the cap was
wholly against the background of a sheet of paper all one color and uniformly illuminated, the shutter regardless of color and size, merged
after 1-3 sec with the background (the paper) and was not perceived
by the subject, who saw only the sheet of paper (Fig. 40). In this case
the empty field was analogous to the blind spot, which was also filled
with the color of the surrounding background.
In one experiment the subject transferred his gazefrom a sheet of
red paper to a sheet of blue paper, so that the round shutter, consisting
of black and white halves, was seenagainstfirst one sheet and then the
other. Immediately after the change of background the apparent color
of the shutter changed within a few seconds from red to blue when the
gaze was switched to the blue paper, and from blue to red when it
was switched to the red paper.
In other words, subjectively, the color of the shutter always
changed to that of the uniform background, mergingwith it completely
within a few seconds. In fact, of course, the color of the black-andwhite shutter remained unchanged but, a s the experiments showed,
played no part in this case. This confirms the earlier hypothesis
that signals from the part of the retina corresponding to the empty
field, particularly to the empty field of the shutter, do not reach the
optic nerve.

80

CHAPTER 11

If the moving test field (the background) was a colored mosaic

(pieces of paper of various colors glued to cardboard), and the stationa r y test field (the shutter) had a much larger visual angle than the
uniform segments of the moving test field (the separate pieces of
paper); the empty field could not merge with the background, and its
apparent color remained permanently dark gray; when the other eye
was illuminated, the empty fieldassumedthe color of this illumination.
If the stationary test fieldandthe individualuniformly colored portions
of the moving test field were approximately the same size, a constant
tendency was observed for the apparent color of the empty field to
merge with the background of the parts of the moving test field. Fusion
was complete only when the empty field was entirely against the
background of one of the uniformly colored parts of the moving test
field.
In one experiment (Fig. 41). two identical black-and-white stationary
test fields Ai and Az (two shutters) were located a short distance apart
against the background of a moving test fieldBi and Bz. The diameter
of each stationary test field was 10". The distance between the edges
of these fields was 10". The moving test field consisted of two halves
(Bi and B z ) One half (Bi) of the field was checkered, and consisted of
different colored squares, each with a visual angle considerably
smaller than 10b. The second half of this field (Bz) was a solid color
(a large sheet of red paper). The subject held his eyes in such a
position that the stationary test field Al was always against the background of the checkered half'of the moving test field (B1), and the test
field Az was against the uniformly colored background B2. Our purpose
was to discover what the apparent color of each empty field would be
if, because of the experimental conditions, the apparent color of one
of them could not merge with the checkered background of the moving
test field Bi, but the apparent color of the other could merge with the
uniform background B2. In other words, is it possible, in certain
conditions, to change the apparent colors of empty fields, separated
in space, simultapeously yet independently?
Between 1 and 3 sec after the experiment began, 'the contours of
the stationary test fields disappeared, and each became an empty
field. The surface of these fields appeared to the subject'to be a
uniform dark gray. Immediately afterward, the apparent color of the
empty field lying against the uniform background B2 merged with the
apparent color of this background. The apparent color of the empty
field lying against the checkered background Bi remained dark gray.
When the second eye was illuminated in this experiment, the apparent
color of the empty field lying against the checkered background changed,

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

81

simultaneously with the color of thewholebackground (i.e., Bi and BZ),

to the color of the light illuminating the second eye. The empty field
which had merged with the uniform background Bz did not show up at
all in these conditions.
In the last experiment of this series, the stationary test field was
a black-and-white shutter seen by the subject against the background
of a screen, a uniformly and identically illuminatedsheet of red paper
(Fig. 42). The shutter was 10" in diameter. Besides the diaphragm
of the P8 cap, through which the subject looked a t the screen, a stop
was fixed to the cap to restrict the field of vision, a second round
diaphragm which left the central a r e a of the retina (50") free for
perception. The sole purpose of restricting the visual field by this
stop was so that during the experiment, with small movements of the
eye, the edge of the screen (the edge of the sheet of red paper) did not
fa11 within the subject's field of vision. During the first moment of
the experiment the subjects saw a red circle (the central part of the
screen) and a black-and-white shutter against its background.
A few seconds after the experiment began, an empty field developed
over the whoIe field of vision, and the subject said that i t had become
dark. Since the screen was stationary, its image moved over the
retina a s the eye moved. However, since the edge of the screen did
not 'fall within the field of vision (limited by the stop), and the screen
itself was illuminated uniformly, the illuminance of all points of the
retina was unchanged; this was equivalent to immobility of the retinal
image.
Next moment, four narrow strips of white paper were fixed to the
red screen to form a square situated inside the restricted field of
vision. The size and position of the square were such that the shutter
lay inside the square (within the field of vision), and with small eye
movements the square did not touch the shutter).
The subject 'immediately noticed the strips of paper and the fact
that the field restricted by these strips became red after a few seconds,
i.e., adopted the true color of the screen. If the strips of paper were
removed, the empty field again developed over the whole visual field,
and after a few seconds the apparent color of the screen changed from
red to gray; i t merged with the color of the stop limiting the field of
vision, and everything appeared dark to the subject. The important
result of this experiment from our point of view was that the blackand-white shutter, having disappeared for the subject a t the beginning
of the experiment, did not subsequently reappear. The apparent color
of the shutter always changed in step with the apparent color of the
screen.

82

CHAPTER I1

Important conclusions may be drawn from these results. First of

all, the empty field clearly has no color of its own. No change in the
luminance of any part of the field of vision disturbs the empty field
bordering on this area. A change inthe color of the region surrounding
the empty field may modify the apparentcolor of the empty field within
very wide limits.
Perception of the empty field cannot he identified with the perception
arising in the absence of active light, i.e., the perception of a stationary
black test field, before its conversion into an empty field, cannot he
identified with perceptionof theemptyfield. A black color corresponds
to a signal indicating absence of light, but an empty field corresponds
to the absence of a signal. When a signal denoting absence of light, is
present, we see black, but when no signal is present on an empty field
we may see any color.
The general conclusion may be drawn from all the preceding experiments that in ordinary conditions of perception we a r e often concerned with empty fields arising innatural conditions. when we examine
large surfaces of uniform color, for example a cloudless blue sky O r a
monochrome screen.
In fact, although the human eye is continually in movement.
frequently these movements a r e withinthe limit of a limited solid angle.
If a uniform background has angular dimensions greater than this solid
angle, there must be a region of the retina within which no changes
take place during a period of time sufficient for the formation of an
empty field (1-3 sec). An empty field will develop under these conditions on the corresponding area of the perceived object, bur it is subjectively imperceptible because it assumes the color of the surroundings.
In perception of a uniform surface, the eye extrapolates the
apparent color from the edges of a surface to its center. The absence
of signals from a particular area of the retina provides the eye with
information that this area corresponds to a uniformsurface, the color
of which does not change and is equal to the color of its edges. The
change from the state of the first moment of perception, when the eye
receives' signals from the whole surface, to the stite when extrapolation is used, takes place smoothly andinnatural conditions is never
perceived by the observer.
The apparent color of an empty field arising in natural conditions is
always equal to the color of its borders, i.e., it is always equal to the
color of the surface against which it arises.
The apparent color of the empty field arising in natural conditions
is determined by one of two conditions. In the first case, the color
depends on the color visible to the subject's second eye. When the

Fig. 37. Scheme of an experiment. In each

pair of figures, a stationary test field i s
shown on the left, and the apparent color of
this field on the right. 1) Test field and its
apparent color at the beginning of the experiment; 2) after 1-3 sec an empty field appears and all visual contours disappear;
3) pale blue light added equally to the two
halves of the test field 1rerlandvre.-n\ Th-

Pig. 38. Scheme of an experiment. In each pair of

figures, the stationary test field i s shown on the
left and the appareni color of this fieldon the right.
1) Test field and its apparent color atthe beginning
of the experiment: 2) after 1-3 sec an emom, field
-- .~
appears, and the
remains in this state for 2
min; 3) after 2 min. the colored filters a r e removed, the test field becomes white. At this moment the subject sees the test field a s consisting
of two halves (the top half darker than the bottom);
4) after 1-3 sec an empty field again appears: 5)
when the light falling on the frosted glass of the
cap is totally switched off, the subject again sees

4 FIE. 41.

4
Fig. 39. Scheme of an experiment. In each pair of
figures, the stationary test field and bright white
ohject moving against the background of this field
a r e shown on the left. The apparent color of the
test field and ohject is shown in the rlght figure of
each pair. 1) Test field and its apparent color at
the beginning of the experiment; 2) after 1-3 sec
an empty field appears (the ohject still appears
white); 3) after 20 sec (the object appears blue);
4) the picture after 40 sec (the objectappears dark
blue).

b
Fig. 40. Scheme of an experiment. In each pair of
figures, the one on the left
shows a mobile test field
against which i s seen the
stationary test field, a
black-and- white screen.
The apparent color of these
fields i s illustrated in the
right-hand figure of each ..
pair. 1) Test field and its
apparent color a t the beginning of the experiment; 2)
after 1-3 sec an empty field
is formed on the screen.
and it merges with the mobile test field; 3) the screen
against the border between
the red and blue halves of
the mobile test field.

Scheme of cxperi.
ment. 01 the left in eacl
pair of flgures is show
the mobile test field 01
whlch a r e superimpose<
two stationary test frelds
two black-and-whlte shut.
t e r s rigidly affixed to tho
cap. One half of the mo.
bile test fieldis checkered
and the other is of on,
color. The apparent coloi
of the test field is shown or
the right figure of eacl
pair.
1) Test fields an1
their apparent color at th(
beginning of the experi.
ment; 2) after 1-3 sec a1
emotv
. field i s formed 01
the shutters; 3) d t the nex
moment the apparent colo
of the right shutter merge
with the color of the uni.
form background. The a p
parent color of the lef
shutter cannot merge wit1
the checkeredbackgraund

subject through the aperture in the

diaphragm, i s shown on the left.
The diaphragm and shutter a r e
rigidly attached to the cap (stationary test fields). Theredbackground i s the mobile test field.
The appirent color of the test
field is shown by the right figure
in each pair. 1) Test field and
i t s apparent color at the beginning
of the experiment; 2) because of
the uniformity of the red background and the immobility of the
remaining pasts of the test field,
after 1-3 sec an empty field 1s
formed inside the diaphragm; 3)
when strips of white paper forming a square a r e placed on the red
background, the apparent color of
the background then becomes red,
but the hlack-and-white shutter

!
I

PERCEPTION OF OBJECTS STATIONARY RELATIYE TO THE RETINA

83

4
Fig. 43. Scheme of an experiment.

In each

pair of figures, the left one shows a red-

and-blue mobile test field on. which is a

black-and-white shutter, the stationary test

field. The apparent color of these fields i s

shown on the right figure of each pair. 1)
Test field and its apparent color at the beginning of the experiment; 2) a few seconds
later an empty field was formed on the Shutter, and i t s apparent color merged with the
color of the red half of the background:
3) when the points of fixation were changed
so that the empty field was against the background of the blue half, for the f i r s t few
seconds the apparent color of the empty field
remained red; 4) then the apparent color af
the empty field merged into the blue of the
a second change Of
5,
points of fixation for the first few secondsso
that the empty field was against the r e d half
of the background, the apparent color of the
empty field remained blue; 6) then fhe aF"
parent color of the empty field merged with
the red of the background

!
'

second eye is closed, theempty fieldis "dark," a dark gray background;

when the second eye is illuminated, the background assumes the color
of the illumination. The second case is observed when, in artificial
conditions, the empty field lies on thebackgroundof a uniform surface
and merges with this surface; in other words, the empty field arising
on the shutter ofthe cap canassume the color of the empty field arising
in natural conditions. This merger of fields can be understood if it is
remembered that in both cases empty fields develop as a result of the
absence of signals, and that the absence of signals from a given area
of the retina is interpreted in the same way by the higher levels of the
visual system. It can be said that the visual system "identifies" a n
empty field arising in artificial conditions with an empty field arising in
natural conditions.
That is why, therefore, the empty field obtained by means of caps
is always a uniform background (all visual contours with the field
disappear), and the apparent color of the field is always the color of
the surroundings, when it lies against a uniform surface. Hence the
properties of an empty field arising in natural conditions may be
studied in experiments with caps.

8. DELAY IN SEEING THE COLOR OF AN EMPTY F I E L D

b
Fig. 44. Scheme of two experiments. The
left column of figures shows stages in the
disappearance of the shutter (stationary
test field) against a background changing
color slowly and smoothly.
The even
change in the background c o l ~ rprolongs
the disappearance of the shutter and makes
the process suitable for observation. The
right column of figures shows stages in the
appearance of the shutter under similar
experimental conditions.

In this section we examine the delay experienced in seeing the

color of an Ompty field.
The P8cap was used in all experiments. The stationary test field
was a black-and-white shutter, the angular size of which was 10". The
mobile test field consisted of various paper screens.
In the first experiment, repeating one of the experiments in the
previous section, the mobile testfield wasapaper screen, half red and
half blue (Fig. 43). The essence of the experiment was that, after the
appearance of an empty field on the shutter,the subject looked a t each
half of the screen in turn and chose his posit'ion so that the shutter
was first entirely within the red half and then entirely within the
blue half. When the empty field appeared on each half of the screen,
it assumed the color of the half and merged with it completely.
We have already observed this result.
The important feature of the present experiment was a s follows.
When the subject changed his points of fixation after appearance of
the empty field on the red half so that it appeared against the blue
half, the color of the empty field was still red (when first seen against
the blue half), then, within a period of 2-3 sec, its color changed

84

CHAPTER I1

smoothly into blue, merging with the blue background. With the change
from the blue half of the screen to the red, the empty field changed
color to red in the same way and at the same speed and merged with
the red background.
In all the subsequent experiments of this series, the shutter
(stationary test field) was directed by the subject to the center of a
white screen and remained in that positionuntilthe end of that experiment. By means' of two light sources, camera shutters, and a system
of rotating polaroids in a dark room, the experimenter changed the
color of the screen (letus say, fromcolor A to color 9, and vice versa)
a t a predetermined rate. The experiment showed that for any two
colors of the screen, for example A and B, a rate of color change
from A to B and f r o m B toA can always be chosen a t which the subject
s e e s the color change clearly butdoesnotsee the shutter disappearing
against the background of the screen. In other words, during a slight
change in the apparent color of the screen, the apparent color of the
shutter can be changed along with the color of the screen.
With a f a s t e r , change in the color of the screen, the subject began
to observe the shutter. Either disappearing against the background of
a relatively unchanged screen or reappearing when the screen was
changed, the shutter always appeared a s a homogeneous circle to the
subject (although i t was in fact half black and half white). When the
shutter disappeared, the periphery disappeared first, then the center.
When i t reappeared the effect was sudden, showing the shutter intact,
with sharphy defined edges (Fig. 44). When the color of the screen
was changed s o quickly that the visible color of the shutter could not
match the color of the screen, the subject observed that the color
change of the shutter clearly lagged behind the color change of the
screen. This series of experiments very clearly illustrates the delay
mentioned above in seeing the color of the empty field.
If the changes in the screen color reached the rate of 1-3 cycles
(illumination changed in a sinusoidal pattern), the apparent color of
the shutter could no longer follow the changes in screen color, and
the shutter became a uniform hue, a mixture of thechanging colors
of the screen.
In section 4 of this chapter, we noted that a stationary object
(a stationary test field) easily disappears against a background of a
flickering environment if the retinal image of this object is illuminated
with scattered, flickering light. In the experiments in the present
section, the light o f a flickering screen, falling on the sclera of the
eye, illuminated the whole retina, including the image of the'shutter.

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

85

In accord with the results described in section 4, it was found that

with flicker rates exceeding only 3-6 cycles the differences between
the apparent color of the shutter and the screen disappeared, and the
area of the visual field corresponding to the shutter appeared to
flicker in time with the screenas awhole. It is important to note here
that a single cycle or half-cycle of changes in the screen caused
appearance on the screen of the shutter which had disappeared, but
that a continous series of thesecycles causedthe shutter to disappear,
i.e., caused the screen flicker to spread to the a r e a of the visual field
occupied by the shutter. Moreover, the flicker rate a t which the shutter
disappeared was much lower than the critical rate.
In a series of experiments, a white screen was illuminated by two
souces of light. One source threw aconstant light on the screen. The
illumination color (color A) in this case played no significant part.
After an empty field appeared on the shutter (shutter merged with the
screen), light flashes of differentintensities, spectral composition,and
duration were thrown on the screen from the other source. In some
experiments the screen illumination was completely switched off for
various short time intervals. In all cases, when the duration of the
flash or of exclusion of the light did not exceed a few hundredths of a
second, the subject observed the shutter only during this interval of
time, i.e., while t h e screen color was distinguishable from the color
A, and the apparent color of the shutter, a s a result of the delay,
remained the same a s this color. As soon a s the flash ended (if it was
not very bright) o r the light was switched off the screen resumed
color A, and the shutter with its apparent color A again became indistinguishable against its background. The apparent color of the empty
field of the shutter began to change appreciably only if the duration of
the flash or light exclusion was increased to several tenths of a second.
Then, of course, with the end of the flash or of the period of light
exclusion, the' shutter, having become a different color, remained
visible f o r a short time against the backgroundof the A-color screen.
In the last experiment in section 7 (see description of experiment
and Fig. 42), a smooth change was observed from an empty field
arising under artificial conditions into an empty field arising under
natural conditions. The black-and-white shutter on the screen, which
disappeared a t the beginning of the experiment, did not subsequently
reappear. This fact suggests that the delay in seeing is identical for
both empty fields. Differences between thesedelays would necessarily
have revelaed the disappearing shutter when the empty field changed
from one state to the other.

CHAPTER ll

Hence, the experiments described in this section have demonstrated

the definite delay in seeing the empty field. In a later section. I shall
try to show the role of this delay in the process of vision.

9.

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

87

SPATIAL DEVELOPMENT OF THE FORMATION

O F AN EMPTY FIELD

The experiments of section 6 showed that two essentially different

processes a r e part of the work of thevisual system--one "fastn and the
other "slow." Subsequently the process of perception of contours within
the frame of anunchanging and stationary retinal image will be referred
to a s the "fast" process.
This process is assumed to begin a t
the moment of the last change in the active light, and to end when
a n empty field forms. To repeat what was saidbefore, it should
he noted that the fast process may evidently be associated with the
phenomenon, known in electrophysiology, of the appearance of impulses
in the optic nerve in response to change in the intensity of light acting
on the retina (the on and off effects). The object of the experiments
described in this section was to make a more detailed examination of
the fast process.
The P6 cap and adaptors with capillary tubes (described below)
were used in these experiments. To anticipate, it may be mentioned
that these adaptors can be used to trace the spatial development of
the fast process, i.e., they enable the subject to see all stages of
this process simultaneously in different experimental conditions.
In many cases in the study of vision it becomes necessary to
displace the image of the border between two fields over the retina
a t a constant and predetermined speed. This purpose can be achieved
by means of a capillary tube fixed to the adaptor of the Ps cap and
visible to the subject against the background of the frosted glass
reflected in the mirror. One design of such a capillary tube is shown
schematically in Fig. 45. The dimensions of the capillary tube,
expressed in units of visual angle, a r e determined from its actual
dimensions and the magnification of the short-focus lens.
The internal diameter of the capillary tube is 0.03-0.05 mm. The
visual angle subtended by the external diameter of the background C
(reflected in the m i r r o r of the frosted glass) is about 45-50". The
bottom part of the capillary tube passes through the center of the
background, and its top part is displaced relative to the center by
9-11".
If the axes of symmetry of the cap and the optical system
coincide, the lower part of the capillary tube in the field of vision will

F1g. 45. Ihagram of the capillary tube fxxed to the adaptor of the P, cap (flrst variant of the
adaptor wlth the capillary Nbe).

lie at the fovea and the middle of the upper part will cross the region
with the highest density of rods. Theends of the capillary tube consist
of small curved funnels A and B, the outer diameter of which a t their
widest part i s 0.6 mm. Aplugof cotton wool is inserted into one of the
funnels but does not project above its edge. The funnels and the parts
of the capillary tube projecting outside the adaptor a r e covered with
a layer of glue, shown in Fig. 45 by a broken line; after the glue dries
it protects these parts from injury.
An alcoholic solution tinted black with aniline dye is introduced
into the funnel A with a micropipet or small brush. The black liquid
quickly soaks into the gauze and fills the whole capillary tube. As the
liquid evaporates, the meniscus always moves from funnel B to
funnel A, i.e., towards the drying pad of compact cotton wool. As the
alcohol concentration changes, the intensity of evaporationand the rate '
of movement of the meniscus of the liquid in the capillary tube also
change. The angular velocity of movement of the meniscus visible
to the subject through the short-focus lens of the P6apparatus can
easily be determined, given the length of the capillary tube in degrees
and the time taken for the meniscus to move a l o k the capillary tube,
measured by the subject with a stopwatch. The velocity of movement
of the meniscus in the capillary tube canchange from several minutes
of angle to several degrees per second. With a slight increase in the
evaporation surface of the solution in funnel A, these speeds may be
increased 10-20 times. If the inner surface of the capillary Nhe is
clean enough, the meniscus of the colored solution, a s it moves along
the capillary tubes, leaves no traces; the capillary tube is clean and
transparent, and it differs little in color from the frosted glass.
During the first second of the experiment the subject sees the meniscus
a s a sharp, moving boundary between the black liquid and the bright

CHAPTER Il

Fig. 46. Scheme to explain the first variant of the adaptor with the capillary tube. a) lmage
of part of rhe capillary visible to the subject's eye at the stam of the experiment. The black
liquid fills the whole capillary tube and moves in the direction indicated by the arrows. ln
the bottom part of the capillarymbe an air bubble can be seen moving together with the fluid;
b) subjective appearance of tfiefieldafewseconds after the beginning of the experiment, i.e..
after the appearance of an empty field.

frosted glass. The background can be colored a s desired by covering

the beam of light falling on the frosted glass of the cap with colored
filters.
As mentioned, the internal diameter of the capillary h ~ b eis approximately 0.03-0.05 mm. The mass of liquid in the capillary Nbe
is so small that any sudden movements of the eye in practice will
not produce changes in the shape of the meniscus or disturb the smooth,
uniform movement of the meniscus in the capillary tube and its image
on the retina. If the experiments a r e carried out a t the same temperature with the same solution, the variability of the velocities of movement of the meniscus from one experiment to another will not exceed
5%.
In many experiments it is useful to have two moving menisci a t the
same time. This can be done a s follows. After all the fluid from
funnel B has entered the capillary tube, a small volume of solution is
inserted into the same funnel from a micropipet, s o that a long a i r
bubble is present in the capillary tube between the first and second
portions of solution. In this case, two menisci a r e moving in the subject's field of vision-two boundaries, each of which is a boundary
between awhite or colored field and a black field. The light is turned
on a t the anterior boundary of the moving bubble and turned off a t
the posterior boundary. The subjective appearance of the field in the
first moment of the experiment and a few seconds later, i.e., after an
empty field has formed, is shown in Fig. 46.
The adaptor by means of which, a s in the preceding case, the perception of images moving a t a definite speed over the retina can be
studied, is illustrated in Fig. 47. The frame of the adaptor 1is made

PERCEPTION O F OBJECTS STATIONARY RELATIVE T O THE RETINA

89

from black paper glued together. It is a parallelepiped with one edge

cut off and covered with paper in which there is a slit. The slit is
parallel to the cut-off edge, equal toitin length, and 2.0-2.5 mm wide.
Over the slit is glued a thin piece of frosted glass 2, about 0.2 mm
in thickness.
Light can enter inside the frame of the adaptor only
through, the 'frosted glass. The frosted glass is illuminated with a
narrow beam of light so that the sclera of the eye remains in complete
darkness. The bottom part of the adaptor is crossed by capillary
tube 3, passing through the axis of symmetrv of the can and ~ a r a l l e l
to the slit covered by the frosted glass (in ~ i g 47
. it is perpendicular
to the plane of the drawing).
Two small curved funnels 4, joined to both ends of the capillary
tube, extend outside the adaptor and a r e glued to its frame. A small
pad of cotton wool soaked in a clear (without coloring matter) solution
of alcohol is placed in one funnel. The internal diameter of the
capillary tube and the width of the funnels may be close to the values
given in the description of the preceding adaptor, or they may differ
from them considerably depending on the experimental condition. A t
a distance of 1 mm from the capillary tube, a slit 5, slightly wider
than the external diameter' of the capillarytube, is situated in the base
of the adaptor parallel to the capillary tube. The slit is so placed that
the eye cannot s e e the frosted glass. The eye sees only the image of
the frosted glass in the capillary tube, magnified by the lens of the P,
cap. Usually this image appears a s a bright band, visible against a
completely black background. The position of the bright band on the
segment of the capillary tube filled with liquid and its position on the

Fig. 47. Second varlant of the adaptor with the capillary tube for the P, cap.

PERCEPTION OF OBJECTS

CHAPTER I1

Fig. 48. Scheme toexplainthe secondvariant of the adaptor with the capillary Nbe. a) Capillary tt~beand meniscus of fluid moving in it, as seen by the subject in the first seconds Of
the experiment; b) and c) the same capillary Nbe and meniscus as seen by the suhject dter
the appearance of an empty field. The arrows indicate the direction of movement Of the
meniscus.

segment free from liquid a r e displaced approximately a s shown in

Fig. 48a. In this case, movementof the meniscus in the capillary tube
takes the form of movement of two boundaries. Movement of one
boundary is accompanied by turning the light on, movement of the other
by turning it off. -The subjective appearance of the field after the
formation of a n empty field is shown in Fig. 48b and c. The width,
the number, and the position of the images of the frosted glass in the
capillary tube a r e determined by the relative positions of the frosted
glass and capillary tube, and a r e predetermined by the experimenter
a s he makes the adaptor. In some cases, it is desirable to use a
capillary tube oval, not round, in cross section. Sometimes, to simplify
the field a s it appears to the subject, the width of the slit 5 is reduced
and part of the capillary tube is covered, leaving only the brightest
band. The angular velocities of movement of the meniscus a r e determined and regulated just a s in the f i r s t variant of the adaptor with
the capillary tube. Depending on the purpose of the experiment, the
experimenter may use the f i r s t o r secondvariantof the adaptor. With
the f i r s t variant, nearlythe whole retina is illuminated with the bright

STATIONARY RELATIVE

TO THE RETlNA

91

background of the 'frosted glass, while in the second nearly all the
retina is in darkness.
After the cap had been applied to the eye and an empty field had
appeared, depending on the construction of the adaptor, the field
appeared to the subject a s illustrated in Fig. 46 or 48. As he vividly
expressed it, the subject saw apparently "two comets," moving over
the background of a n empty field. One of the comets appeared bright
and the other very black ("much blacker than the dark background of
the empty field"). Later we shall describe the bright comet conventionally a s the on-comet and the dark a s the off-comet (since the first
corresponds to the turned-on light, the second to the turned-off
light). If the fast processcanbeassociated with the on and off effects,
the spatial evolution of the fast process may be regarded a s analogous
to the trains of impulses in the optic nerve.
The anterior part of eachcomet containedanarea of uniform color.
The apparent color of these areas then changed smoothly into that of
the empty field.
To determine the time characteristics of the fast process, the
apparent length of the comets was measured. During the experiments,
a shadow from a narrow. strip of paper was thrown onto the frosted
glass. Since the shadow moved across the frosted glass, it could be
seen, although not very sharply, by the subject against the background
of the empty field. When the shadow was near the comet, the subject
compared the width of the shadow and the length of the comet. It was
always possible to choose a strip of paper giving a shadow of width
equal to the length of the comet. After the experiment, the width of
the shadow in units of visual angle was measured. Since the angular
velocity of-'the comet was known, all the information required to give
the time characteristics of the fast process was available.
The border between two fields moving over the background of an
empty field produces a sharp change in the illuminance of a certain
part of the retina and, consequently, leads to the appearance of the
fast process. Since the boundary moves at a constant speed over the
retina, conditions a r e produced. in which the subject can observe all
stages of the fast process a t the same time.
The next step was an attempt to discover how the application of
certain stimuli affects the apparent length and color of the comet in
conditions when these stimuli involve only part of the retina. Against
the background of an empty field, and very close to the on- and offcomets, objects of different color and luminance, with an angular
diameter of about 5-7". moved and therefore were visible. No appreciable change was observed in the length and color of the comets.

CHAPTER I1

When these objects coincided with the comets, the color of the comets
appeared to change only in the segments immediately next to the moving
object.
If part of the capillary tube was shaded with transverse black bands,
the subject saw the comet broken up into isolated segments when it
passed through this area. These conditions didnot appear to alter the
total length and color of the on- and off-comets. With the accuracy
attainable with this method, it was found that the length of the comets
changed proportionally to the speed of movement of the menisci.
In face of these results, the comets maybe regarded a s the evolution of the fast process in space. In fact, if the action of the stimuli
on the areas of the retina lying next to the comet in the visual field
had not influenced its apparent length and color, evidently the action of
the meniscus image could not have influenced the areas through which
it had already passed and which corresponded to the extinction of the
f a s t process. In addition, it was found that the duration of the fast
process, measured by the time required for appearance of the empty
field, and the duration of the fast process, measured by means of the
comets, coincided approximately (I say "approximately" because the
subject found it difficult to determine the moment when all contours
disappeared from the test field).
When movement of the image over the retina was fairly slow, the
menisci could not be seen, and consequently no comets appeared. For
instance, when a bright light fell onthe sclera and the luminance of the
frosted glass of the cap was 500 apostilbs, if the meniscus moved with
a velocity relative to the retina of 18-19 minutes of angle per second,
the subject could see nothing, but if the speed was 23-24 minutes of
angle per second, the subject could see the on- and off-comets. If the
experiments were carried out without illumination of the sclera,
through which scattered light usually falls on the retina, appearance
of the comets was observed bythesubjectwhen the meniscus moved a t
a speed.of 3-5 minutes of angle per second.
As mentioned above, the first part of each comet had an a r e a of
uniform color, which then appeared to change gradually. into the color
of the empty field. Evidently the presence of this area indicates that
the fast process in its initial stage shows little variation, and that
extinction begins only after a short time. The duration of the initial,
relatively constant part of the fast process was approximately 1-3 sec;
extinction of this process took 2-5 sec.
A change in. the spectral composition of the light falling on the
frosted glass of the cap caused no change in the apparent length of the
comets.
The on-comet always took the color of the frosted glass,

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

while the off-comet appeared black, tingedwitha color complementary

to the color of the on-comet. If the intensity of the light falling on the
frosted glass was increased by two orthree orders, there was a slight
increase in the apparent length of the comets.
An attempt was then made to examine the behavior of the comets
when the illuminance of the eye was sharply varied. When the cornea
is covered by the cap, it is easy to change the illuminance of the whole
retina with scattered light by c h a n g w the intensity of illumination of
the sclera. A sudden change in illuminance may cause the comets to
contract or completely disappear. Itwasalways possible to change the
luminance of the retina in such a way that objects disappearing for
the subject because of the immobility of the retinal image hardly appeared a t all, while changes in thecomets under these conditions were
clearly apparent to the subject.
I simply wish to emphasizehere the fact that a change in illuminance leads, in particular, to the complete or partial inhibition of the
fast process (the comet contracts or disappears for an instant, then
recovers). This phenomenon is evidently analogous to the process of
pre-excitatory inhibition, familiar in the electrophysiology of the retina.
Pre-excitatory inhibition is clearly seen on records of impulses
obtained from the optic nerve of animals. An example of this inhibition, in the optic nerve of a frog, is shown in Fig. 49.
The Pf cap was used (see the description of this apparatus) to
examine the behavior of the fast processes in conditions close to the
usual conditions of perception. Thestationarytestfield was a capillary
tube similar to the capillary tube from the f i r s t type of adaptor
(Fig. 45). When the cap was in position, the image of the surroundin$
objects and the image of the capillary tube were superimposed on
each other on the retina. For the first few seconds, the subject saw a
sharp image of the capillary tube a s a shadow superimposed on the
object used for fixation. Then, since the capillary tube was immobile
relative to the retina, the shadow disappeared and reappeared slightly
only if change in the points offixationwas accompanied by a considerable change in the color of the background object. When the capillary
tube was no longer seen by the s'ubject, the movement of the meniscus
of the black liquid inside the capillary tube evoked the appearance of
on- and off-comets. The subject saw the comets against the background
of the surrounding' object; the faint image of the capillary tube that
sometimes appeared during change in the points of fixation did not
prevent him from making observations on the comets.
The followi& experiment was carried out. Three screens were
placed before the subject. One was cardboard checkered with pieces

CHAPTER I1

Fig. 49. Records of optic-nerve impulses of the frog (Rone iidibundo). Impulses appearing
in response to the change from light to dark and from dark to light are shown in the bottom
photograph. k i n g the change from dark to light, the impulses continue at first, but then
are completely inhibited. This pre-excitatory inhibition is clearly seen on the records.

of paper of different colors glued to it. The two halves of the second
screen differed very slightly in color. The third screen was a sheet
of paper of uniform color and intensity of illumination. The subject
saw each screen a t such a n angle that the image of the capillary tube,
and consequently the comet, was superimposed on its background.
During the experiment the subject examined the screens in turn and
adjusted the points of fixation s o that the comets did not go outside

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE RETINA

95

the boundaries of the screen. When the subject changed the points
of fixation on the f i r s t (checkered) screen, each such change was
accompanied by a momentary disappearance of the on- and off-comets,
which then reappeared because of the movement of the head of the
comet. When the subject adjusted the points of fixation on the second
screen s o that a t the moment the change was made the background
on which the comets were seen was slightly changed, the comets
contracted, then returned to their original size. Adjustment of the
points of fixation on the third (solid-color) screendid not visibly alter
the length of the on- and off-comets.
It may be assumed that the contraction and disappearance of the
comets described above should be regarded a s indications of the
complete o r partial inhibition of the fast process, corresponding to
those parts of the retina on whichthechanges in the comets were seen
a t the moment of, and after, the change in the p i n t s of fixation. The
experiments described ahove confirm that, depending on changes in
illuminance taking place a s a result of a change in points of fixation.
the fast processes on various parts of the retina may be completely
inhibited, partly inhibited, or.not inhibited a t all.
During fixation on an object, the human eye is constantly in movement, making small but rapid rotations. These eye movements a r e
accompanied by the appearance and inhibition of a series of fast
processes, corresponding to those parts of the retina over which the
image of sharply defined elements of the object moves a s a result
of the eye movements. The apparent color of the empty field situated
inside an object of uniform color is determined by its surroundings;
the visual system extrapolates to the empty field the color visible a t
the edges of the surface. What we extrapolate with is discrete (alternation of fast processes and inhibitions a s a result of saccadic eye
movements), and what we obtain by extrapolation is continuous (the
apparent color of the empty field is unchangedduring the time that the
observer fixates on the object). However, this is understandable in
view of the existence of a delay. in seeing the color of the empty
field. This delay enables us, when extrapolating from the edges
of a uniform surface, to see it unchanged in color when the image of
the borders moves continually and saccadically over the retina, which
in turn is essential to enable us to see these borders.
At this point it should benoted that each sudden rotation of the eyes
during fixation lasts about 0.02 sec, and the duration does not exceed
3% of the total fixation time. During the sharp change in the environment of the empty field, appreciable changes inits apparent color take
place only after a delay of several tenths of a second.

96

CHAPTER I1

4
Fig. 50. Scheme of an experiment.
In each pair of figures. the stationary test field andtheredobject
moving against its badrground(the
mobile test field) a r e shorn on the
left. The right-hand figureof each
pair shows the appearance of the
test field and object in different
experimental conditions. 1) Test
field and i t s apparent color at the
herinning of the e m r i m e n t :. 2).
after 1-3 secanemptyfieldforrncd
and the color oftheobject a p p a r e d
to change; 3) when the second eye
was illuminated with red light, the
color of the empty field and of the
object appeared to change; 4) when
the secondeve was illuminated with
blue light tcenexr moment, color of
the empty field and object again
appeared to change.

10. PERCEPTION O F OBJECTS MOVING

ON AN EMPTY FIELD
I have previously described experiments in which objects moved
a c r o s s a n empty field. Let us now consider this case in rather more
detail. To facilitate the description, I shall use the word "objectsn to
describe test fields mobile relative to the retina, moving against the
background of an empty field.
In the f i r s t series of experiments in this section the P, cap was
used, and the color of the stationary test field varied widely from
experiment to experiment. In each experiment, after the appearance
of an empty field, a black object 3" in diameter was moved across it.
This caused no change in the apparent color of the empty field, hut
sometimes a small halo ameared around themovimobiect.
It mav be
supposed that these haloes appeared a s a result of extrapolations
similar to those discussed in section 7. In this case, however, the
extrapolation was directed inward, and not outward, from the border
of the moving surface. The essential feature s o far a s we a r e concerned is that no differences in the stationary test field appeared
inside the halo. The apparent color of the black object coincided with
the color of the after-image arising when the light was completely
switched off. This c a n be understood because the appearance of a
black object in the field of vision and the exclusion of light falling on
this part of the retina a r e in practice identical processes. It was then
found that the apparent color of the black object depends not only on
the color of the stationary testfield, butalso on the speed of movement
of the object. This is because the color of the after-image changes in
time, and, depending on the character of the movements of the object,
the appearance and disappearance of the after-image will differ.
At this point it is well to recall the experiments discussed in a
previous section. In these experiments, a bright white object moved
against the background of an empty field. The diameter of the object
was 3". Its luminance in all the experiments was much greater than
that of the stationary t e s t field. These experiments showed that the
color of the empty field appeared unchanged, while the white object a t
f i r s t (immediately after the appearance of the empty field) appeared
white, but changed appreciably during the next 30-40 sec, eventually
acquiring a saturated hue approximately complementary to the color
of the stationary test field.
In one experiment in which the Pscapwas used, the stationary test
field was a bright white field-the frosted glass of the cap (Fig. 50).
-

Tlg 51. hcheme of d n expcnment. In each p a r of flgures.

the red scre'n (mobile test
fleld) and the hldck-and-whtte
shutter (&ratlonary test ileld)
placed agdmst it dre shotvn oo
t h ~left.
The rlght flgure m
each p l r shows theappearan~e
of the test hela. 1) Test f ~ l d
and i t s apparent color at the
beginning of the experiment;
2) after 1-3 sec an empty field
formed on the shutter, and its
color appeared to merge with
the color of the screen. Next
the everimenter moved a r e d
circle in front of the white half
of the shutter. Theactualcolor
of the circle coincided withthat
of the screen. The subject saw
only a moving circle whose
color appeared darker and
more saturated than the,colorof
the screen; 3) when the experimenter then moved the circle
in front of the black half of the
shutter, the moving circle a p
peared to the subject to he less
saturated in color anrl mnrh

PERCEPTlON OF OBJECTS STATlONARY RELATIVE TO THE RETINA

97

!
Against the background of this fieldmovedared object, 3" in diameter,
the luminance of which was much less thanthe luminance of this field.
The experiments showed that with the appearance of the empty field,
when the color of the white test field had appeared to become dark gray,
the color of the redobject appearedtochange from red to dark red and
then remained appreciably darker than the empty field. Later the subject's second eye was illuminated with red light, a s a result of which
the color of the empty field appeared to become red, and the color of
the object appeared more saturated; the object was now "very red,"
and a s before, appreciably darker than theempty field. M e n the subject's second eye was illuminated with blue light, the color of the empty
field acquired a blue tinge, and the color of the red object appeared to
lose much of its saturation, becoming dark brown, and a s in all the
previous cases, remaining much darker than the empty field.
Although the color of the background on which the object moved and
the color of the object itself appeared to change sharply with the change
in the experimental conditions, the character of the difference existing
between the stationary test field and the object was the same a t all
stages of the experiment.
The result of one experiment in which the PBcap was used is
illustrated in Fig. 51. In this experiment, the mobile test field was a
large, well-lighted sheet of red paper, on which was placed the
stationary test field (the shutter), half black and half white. The
diameter of the stationary test field was 20". The movable object
was a small red circle, the same color a s the mobile test field and
3" in diameter. The question was: what was the apparent color of
the object moved against the empty field arising on the shutter? Not
only was the real color of the red object equal to that of the mobile
test field, but the apparent color of the empty field in these experimental conditions also became equal to the color of this same mobile
test field.
The subject could clearly see the object moving against the
background of the empty field,althoughhecouldnot s e e the empty field
itself. When against the right half of the empty field, the object appeared to the subject to be an extremely saturated red color, darker
than the mobile test field. When on the left half of the empty field,
the object appeared pink, lighter thanthemobile test field. Sometimes
small haloes wereobservedaroundthe object. Here, a s in the previous
experiment, a tendency was observed for the character of the differences between the stationary test field and the object (in brighmess
and color) to persist.

CHAPTER I1

98

11. ROLE OF ILLUMINATION OF THE EYE WITH

SCATTERED LIGHT

Experimenters investigating human vision do not usually attach

g r e a t importance to light falling on the retina through the sclera. This
is natural, because with more or less constant intensity of illumination
(and consequently constant illuminance) scattered light falling on the
retina has little effect on the results of many experiments. However,
when working with stabilized images, i t is particularly important to
exclude light or to make i t a s constant a s possible, because (see
section 5) any suprathreshold decrease in illuminance reveals contours
in a test field that have disappeared because of immobility (with any
suprathresbold addition of illuminance the subject sees only this
addition). In most cases a significant change in illuminance may be
produced not only by accidental changes in the intensity of illumination,
but also by rotation of the eye or by a shadow flickering on the sclera.
Let us first consider the phenomena arising in conditions of
flickering illuminance of great brighmess. In one of the experiments
with the P, cap, the cornea was completely covered, and light could
enter the eye only through the sclera. The eyelids were retracted a s
widely a s possible with strips of adhesive plaster and a bright flickering
light was thrown onto the sclera. Usually, with a flicker frequency
of 6 to 1 5 cycles, the subject saw bright mosaics, iridescent with all
the colors of the rainbow. These mosaics had very saturated colors;
they were small in the region of the fovea and larger a t the periphery
of the retina. The picture was particularly colorful a t the times that
/
the flicker frequency was raised or lowered.
Next, the P6cap was placed onthe subject's eye, allowing scattered
light to pass into the eye only through the cornea and pupil, i.e., only
through the transparent media of the eye (illumination through the
sclera was excluded). In this case, the bright colors of the mosaic
which had appeared immediately faded, although its faint tints were still
perceptible. However, a s soon a s an orange filter (Schott OG-2) was
introduced into the beam of light falling onthe frosted glass of the cap
the mosaic again acquired the same bright and varied colors.
These experiments reveal the role of illumination of the retina by
flickering light through the sclera, which behaves like the orange
filter. These results also demonstrate the c a r e required in interpreting results of experiments in which a fluctuating light falls on the
sclera. The actual hue of lightenteringthe eye through the sclera can
In a completely dark room the
easily be determined a s follows.
temporal part of a subject's sclera is illuminated with a bright beam
of light. Under these conditions, a s the experimenter will see, the

PERCEPTION OF OBJECTS STATIONARY RELATIVE TO THE

99

whole inner region of the eye shines with an orange light. This,
experiment clearly shows the background of illuminationagainstwhich
a retinal image is formed when the sclera is brightly illuminated.
In a series of experiments the illumination of the retina with
flickering light was carried out boththroughthesclera and through the
frosted glass of the cap. Sometimes these two forms of illumination
were synchronized for the frosted glass and sclera; sometimes they
were not. In some experiments the illumination of different parts of
the sclera and that of different parts of the frosted glass were out
of phase. In these experiments the subject often saw geometrical
figures, fantastic in their complexity, variety, and color, with their
center in the foveal part of the retina. Although much space could be
devoted to the description of these experiments we would be no nearer
to understanding these phenomena. Icanonly suggest that the complex.
unusual changes of a bright color on the retina lead to confusion of
information a t certain levels of the visual system,as a result of which
the subject sees these remarkablevisual phenomena. The experiments
of this series were a severe strain on some subjects. Subjects for
such experiments should be chosen with particular care.
I next tried to discover how much the color appearance of the test
field depends on the fact that the optic fundus is colored orange by
pigments and blood vessels. In these experiments the test field,
stationary relative to the retina, was illuminated with white light of
varying intensity. Light reachedthe retina only through the transparent
media of the eye.
The experiments showed that the white test field, having disappeared
because of its immobility, in the presence of a gradual and suprathreshold increase in the luminance of the white screen, caused a
sensation of yellow. A gradual suprathreshold decrease in the luminance of the white test field lent a n appreciably bluish color to this
field. If the disappearing test field were bright enough, the blue color
that developed would appear saturated to thesubject.
It is possible that the results obtained were due to the orange color
of the optic fundus and so must be taken into consideration in experiments with a test field of variable intensity (or color).
'To determine how the eye adapts itself to certain anomalies in
illumination, stationary relative to the retina, several experiments
were carried out using the PI and P8 caps. By means of the PI cap,
dark shadows (stationary relative to the retina) were formed on different parts of the retina. When the shadows were superimposed on the
retinal image in the process of perception, each shadow, roughly
speaking, uniformly cut off a certain constant fraction of the light.
Instead of a shutter, different neutral and colored filters were fixed

I
100

CHAPTER I1

r
I

to the P8 cap. By means of the neutral filter, the illuminance of the

retinal image could be reduced by a specified number of times on a
particular part of the retina.
These experiments showed that within a few seconds the subject
ceased to notice the presence of evendark shadows and filters (absorbing, for example, 70-90% of the light) if he looked a t uniform surfaces
o r a t objects with details possessing slight contrastof light and color.
If the subject examined very checkered or contrasty surfaces, the
shadows and filters could still be noticed to some extent. Faint shadows
and weak filters (absorbing, for example, 20-30% of the light) were
practically unnoticed by the subject even against a checkered background.
These experiments suggest that although the human eye is well
adapted to certain constant and stationary anomalies in illumination
of the retina, i t cannot become completely adapted to them (at least in
the course of one single experiment).

CONCLUSIONS
We may draw the following conclusions from the results of the
experiments described in Chapter 11. For optimal working conditions
of the human visual system, some degree of constant (interrupted or
uninterrupted) movement of the retinal image isessential. If a test
field (of any size, color, and luminance) becomesand remains strictly
constant and stationary relative to the retina, it will become and remain
a n empty field within 1-3 sec.
Very often, conditions of steady illumination arise oncertain parts
of the retina in theprocessof perception. Such conditions arise during
the perception of large and uniform surfaces and during small movements of the eyes. If the illuminationcontinues constant for more than
three seconds, a n empty field appears inside this uniform surface
(or surfaces). The empty fieldalways takes the color of the surroundings and, in ordinary conditions, is never seen by the human subject.
In other words, the visual system extrapolates the apparent color of
the edges of the surface to its center. In accordance with electrophysiological findings, I suggest that in man constancy and immobility
of the retinal image will banishimpulsesenteringthe optic nerve from
the eye o r will sharply reduce their number. In these circumstances,
absence of signals from a certa.in part of the retina gives the visual
system information that this a r e a corresponds to a uniform surface,
the color of which does not change and is equal to the color of its edges.

PERCEPTION OF OBJECTS STATIONARY RELATIVE T o THE RETINA

101

It may be concluded that the visual system "identifies " the empty field
arising in artificial conditions with the empty field arising in natural
conditions. For this reason, the empty field arising in artificial conditions always appears to the subject a s a uniform background (all
visible contours disappear inside the field), and the apparent color of
the field is always the color of its surroundings.
Two essentially different processes a r e found in the work of the
visual system:
the first, a fast process of disappearance of all
contours in a stationary test field; the second, a slow process which
usually is easily detected by means of after-images. The fast process
may evidently be associated with the appearance of impulses in the
optic nerve in response toa change in the intensity of light (the on- and
off-effects familiar from electrophysiology). Thesecond, slowprocess
is evidently associated with a change in the state of the retina-with
its adaptation.
A definite delay is found in seeing the color of an empty field.
This delay enables us, by extrapolating from the edges of a uniform
surface, to see this surface unchanged in color when the image of the
edges is continuously and saccadically displaced over the retina; this
in turn is essential in order that we can s e e these edges.

Chapter III

EYE MOVEMENTS DURING FIXATION

ON STATIONARY OBJECTS
Before describing the material contained in this chapter, I shall
try to explain the general character of eye movements during the
perception of stationary objects. Letus consider Fig. 52, which shows
a recording of the movements of a subject's two eyes when examining
a flat picture with one eye. Duringthe experiment one of the subject's
eyes was covered by a P2 cap and the other with a Pi cap. The black
dots in Fig. 52 show the points of fixationduring perception of the
picture, and the thin lines show movement of each eye from point to
point. During the recording, the picture was in a frontal plane. In this
case, change in the points of fixation was, roughly speaking, accomplished by movements of a single type-identical and simultaneous very
rapid rotation of the eyes, hereinafter conventionally termed 'saccades."
At this point, I should s t r e s s that in any situation, the saccades of
the eyes a r e of high velocity (the duration of a saccade is measured in
hundredths of a second) and uniform amplitude, and both eyes move
simultaneously. I t is clear from Fig. 52 that this last property is
preserved even when one eye is completely covered by a cap, i.e., is
excluded from perception.
Some readers may think that during the perception of stationary
objects the human eyes areable to perform smooth pursuing movements
in addition to saccades. Where stationary objects a r e concerned, this
view is incorrect. It is due to the fact that the small saccades of the
eyes a r e performed involuntarily and we a r e not aware of them. A
record of eye movements is shown inFig. 53, during which the subject
tried to follow the lines of several geometrical figures with his eyes
smoothly (without saccades). Although subjectively the trackingmovements of the eyes seemed smooth anduninterr~pted~they
were in fact,
103

104

CHAPTER 111

EYE MOVEMENTS WRING FIXATION ON STATIONARY OBJECTS

105

eyes. Records show that in fact this process is accompanied by

involuntary saccades of which the observer is unaware (sometimes
resembling spasms of the eyes). The experiment showed that it is not
only the small saccades (measuring from 2 to 20 minutes of angle)
which a r e involuntary. Large saccades (usually not exceeding 15-20")
also a r e mainly involuntary, and the observer is not aware of them
during perception.
When the object of perception is stationary relative to the observer's head, the process of perception taking place between any two
adjacent saccades in the following pages will be conventionally called
"fixation process." It has been shown experimentally that during the
perception of stationary objects, in the interval between rotations of

Fig. 52. A reproduction of A.L. Rzhevskaya's picture A

' Happy Minute. with records of the
movement of each eye during f r e e (without instruction) examination of the reproduction by
one eye for a period of 30 sec (the second eye was completely covered by a Pz cap).

a s the record shows, composed of discrete stops and small saccades.

Again in Fig. 53, we can see that the involuntary saccades arising
during the attempt to trace the lines of the figures visually did not
invariably lie along theselines.
If the observer carefully examines any point of a stationary object,
he imagines objectively that he isfixatingonthis point with motionless

Fig. 53. Record of eyemovememsdunn~examinationof geometrical figures. a) Geometrical

figures presented to the subject f a r examination; h) record of eye movements during which
the subject tried to trace the lines of the figures with his eye smoothly and without saccades;
c) record of eye movements during free (without instruction) examination of the figures for
20 sec: d) record of eye movements during examination of the figures f o r 20 sec after the
instruction 'look at the figures and cwnt the number of straight lines..

CHAPTER 111

106

EYE MOVEMENTS DURING FIXATION ON STATIONARY OBJECTS

the head and blinking movements, the humaneyes can be only in one of
two states: in a state of fixation or in a state of changing the points
of fixation. The present chapter is devoted to the study of eye movements during fixation directed to any element of a stationary object.
Accordingly, we shall examine the fixation process and the small
involuntary saccades of the eyes.
It follows from the facts described in Chapter I1 that slight movement of the eyes is essential f o r good conditions of perception, even
in the case of fixation on the elements of stationary objects. Fixation,
recorded in Figs. 52 and 53 a s black dots, is accompanied by two o r
three types of movements: drifts, tremor, and involuntary saccades.
In addition, we must always remember that some displacement of the
retinal image and change in the image itself may be caused by movement of the head, which is never absolutely still, by pulsation of the
blood, and by constant changes in the state of the lens and the size of
the pupil.
A drift is an irregular and relatively slow movement of the axes
of the eyes, in which the image of the fixation point for each eye
remains within the fovea. The drift is always accompanied by a
tremor, an oscillatory movement of the axes of the eyes of high
frequency but very small amplitude. Small involuntary saccades
usually a r i s e when the duration of fixation on a particular point of a
stationary object exceeds a certain length of time (0.3-0.5 sec) or
when, because of drifts, the image of the point of fixation becomes too
f a r removed from the center of the fovea. Not all fixations a r e accompanied by involuntary saccades. For instance, during a relatively
cursory examination of an object, most fixations a r e accompanied by
two typesof movement, driftandtremor. The most suitable conditions
for recording drift, tremor, and involuntary saccadic movements of
the eyes occur during' prolonged fixation on a stationary point.
The general (saccadic) character of the eye movements has been
known for a long time and has been studied by many authors (Miiller,
1826; Lamansky, 1869; Javal, 1879; Landolt, 1891; Delabarre, 1898;
Orschansky, 1899; Huey, 1898,1900; Judd, McAllister.and Steele, 1905;
Dodge, 1907; others).
Despite the imperfections of the methods used by these authors,
they obtained a generally correct impression of the fundamental type
of eye movement. As techniques have improved, more and more
investigations of eye movements have been made. Attention has been
directed in particular to the fixation process, in which the role of eye
movements in all its details was not clear. Very often the same
problems were studied by many investigators using differentmethods.

107

I shall not mention all these investigations, some of which duplicated

each other. My purpose is to show, using the best method available.
how micromovements of the eyes prevent the formation of an empty
field in the fixation process.

1. DRIFT OF THE OPTICAL AXES DURING FIXATION

The drift of the axes of the eyes was f i r s t discovered, and on the
whole correctly described, by Dodge (1907). He considered that there
is no constant point of fixation and suggestedthe t e r m 'fixation field."
Subsequently, nearly all authors studying eye movements confirmed
the presence of drifting movements of the eyes (Glezer and Tsukkerman, 1961). An exception in this case was Hartridge (1947) in whose
opinion fixation can take place with a n accuracy of up to one cone unit.

'

Fig. 54. Records of eye movements during fixation on a stationary point by the subject. a)
Fixation for 10 sec; b) fixation for 30 sec; c) fixation for I mi". The scale of angular measurement in minutes of angle is given in record b, and the distribution of the cones is indicated schematically on this scale.

1%
CHAPTER 111

It is clear from Figs. 52 and 53 that fixations accompanying the

perception of images a r e recorded not a s points, but a s irregularly
shaped spots. The size and shape of these spots a r e determined hy
the drift of the optical axes of the eies. To obtain a general idea of
drifts, let us first examine Fig. 54, in which the spots produced by
drifts a r e greatly magnified. This figure consists of three separate
recordings of eye movements (with exposures of 10,30, and 60 seconds)
made on stationary photosensitive paper by means of the Pi cap. Each
record shows only the drifts of the eye during fixation. In the same
figure, a small area of the retina (the fovea) i s shown on a corresponding scale. The records given in Fig. 54 (and others like them) show
that the drift of the eye axis during fixation is an irregular movement
during which, however, the image of the point of fixation always remains
inside the fovea. There a r e two reasons for this: first, there is a
definite constancy in fixation of the angle between the optical axes of
the eyes (the angle of convergence) and second, involuntary correcting
saccades of the eyes occur, returning the image of the points closer
t o the center of the fovea. Analysis of these records shows that the
drift speed varied chaotically from zero to approximately 30 minutes
of angle per second. During fixation, purely a s a result of drift, the
axis of the eye moves with ameanvelocity of approximately 6 minutes

Flg. 55. Horizontal component of the movements ofthe two eyes an a photokymagraph dur~ng
flrat~onby the subject on a statlonary polnt.
I

EYE MOVEMENTS WRING FIXATION ON STATIONARY OBJECTS

Horizontal component of the movemenrs of the two eyes an

during examlnatlan by the subject of a statlonary plcmre.

of angle per second and, consequently, moves in one second along a

path equal to 10-15 diameters of the cones of the central part of the
fovea. It i s very significant that, a s a rule, several times per second
f o r periods measured in hundredths and tenths of a second, the drift
reaches almost maximal values, i.e., about 30 minutes of angle per
second (speeds a t which no empty field can develop). Ditchburn
recently (1959) obtained a mean value of the speed of drift of 5 minutes
of angle per second.
Let us now turn to Figs. 55 and 56. Figure 55 is a record of the
horizontal component of the movements of both eyes during fixation on
a stationary point.. The recordwas madeon the vertically moving tape
of a slit photokymograph by means of a Pi cap. Tt is clear from this
figure that the saccades of both eyes were always equal, whereas the
drift was largely independent for 'each eye (the vertical lines do not
remain parallel. The scales of Figs.. 55 and 56 precluded recording
of tremor.
Figure 56 (likeFig. 55) is alsoarecord of the horizontal component
of the movements of both eyes, but during free and cursory examination of a stationary object (in this case, the nature of the object does
not matter). The records inFig. 56resemble those of Fig. 55 in many
respects. However, the saccades of the eyes recorded in Fig. 56 a r e
changes in the points of fixation and each individual act of fixation is
accompanied only by drift and tremor. It is clear from Fig. 56 that

,,

CHAPTER Ill

EYE MOVEMENTS WRING FIXATION ON STATIONARY OBJECTS

111

ments. The record was made on a photokymograph on which the tape

moved a t a speed of 20 cm/sec, and the time marker showed each
0.2 sec. In this way an accuracy of measurement of t0.005 sec was
obtained. A device mounted on the photokymograph allowed simultaneous recording of the horizontal and vertical movements of the eyes.
It may be concluded from this graph that during free examination
of stationary objects the mostprobableduration of the drifts (fixations)
was from 0.3 to 0.8 sec. In particular, the marked increase on the
graph in the number of drifts when their duration was over 0.20 sec
should be noted. This suggests that for optimal perception of stationary

Fig. 57. Horizontal component of the movements af the two eyes recorded on a phocokmograph during fixation by the subject on a stationary paint. This record clearly shows
that the drifts and the trOmor shown by one eye are independent of those shown by the other
eye.

during f r e e examination of a stationary object the drifr of each eye

is independent of the other. Finally, in Fig. 57 the independence of
and difference between the drifts of the two eyes a r e recorded on a
scale a t which the tremor is perceptible. The drift lying between two
successive saccades will conventionally be called anindependentdrift.
Many records have shown that during the free examination of
stationary objects, the overwhelming majority of fixations onelements
of the object take place without saccades (involuntary). For such
fixations the duration of the independent drifts and the duration of
fixation a r e 'identical. It is important duringperception of a stationary
object that the duration o f the independent drift (of fixation) i s
nearly always adequate to enable the eye to s e e the fixated element.
However, this time may sometimes be insufficient for the thought
process evoked by the perceived element to be completed. Usually,
our gaze i s directed towards the element about which we think; in which
case, a prolonged fixation arises, composed of independent drifts and
involuntary saccades. It may be concluded that the same element is
seen repeatedly during such a fixation, although the observer usually
is unaware of the brief interruptions caused by the involuntary saccades.
A graph of the distribution of the independent drifts of the eye in
accordance with their durationduring the free examination of a picture,
i.e., a graph of the distribution of fixations in accordance with their
duration, is given in Fig. 58. The Pi cap was used in these experi-

set

Fig. 58. Reproduction of 1.1. Levitan's picture .The Floods shown to five observers for free
examination, and graph of the distribution of 2000 drifts in accordance with their duration
Abscissas-dura~ionof the drifts; ordinate-number of drifts of approximarely equal duration.

CHAPTER 111

112

EYE MOVEMENTS DURING FIXATION ON STATIONARY OBJECTS

113

picture, depending on the size and character of the picture, saccades

may account for more than 3 4 % of the time. Finally, when the f r e e
perception of a stationary object is accompanied by a change in the
points of the fixation in space, so that convergence and divergence of
the eyes take .place, considerably more than 5%of the total perception
time may be taken up in changing the points of fixation.

2. TREMOR OF THE EYES

.$-

1 I

511
6

50
8

10
sec
Fig, 59. Diswibufion of 1000 drlfrs inrelationto their dulation during fixation on a stationary
point. The resulw for four observers aregiven. Abscissa-duration of the drms: mdinatenumber of drifts of approximately equal duration.

objects, the lower limit of duration of the fixations should be taken a s

0.25 sec. The graph also shows that, in the conditions described, the
duration of the drift (fixation) most commonly found was 0.3 sec.
The mean duration of eye drift is to some extent dependent on the
subject and his state, on the character of the visual test object, and on
the instruction which the subject receives before theexperiment. For
example, during continuous and prolonged fixation on apoint, when the
subject has been told to fixate on it continuously, the distribution of
independent drifts in relation to their duration (Fig. 59) differs appreciably from the picture seen in Fig. 58. The duration of the independent drifts is noticeably greater, sometimes amounting to several
seconds. In certain diseases, the drifts show a definite direction
instead of being irregular (sometimes in healthy observers also). In
this case, of course, a drift taking place predominantly towards one
side is corrected by small saccades in the opposite direction. This
fact has been recorded . in healthy subjects by Nachmias (1953).
Experiments have shown that when an observer tries to fixate on a
point f o r a long time about 97% of the time is given over to drifts and
only 3$ to saccades. During free examination of a flat, stationary

Of all forms of eye movements, tremor is the most difficult to

study. The amplitude of the tremor is very low and its frequency very
high; this complicates recording of the tremor because movement
of the subject's head, a s well a s the vibration of the apparatus and of
the building itself, must be taken into account during recording. In
addition, the recording of tremor imposes serious demands on the
optical system used in the investigation. The first records of tremor
were made by Adler and Fliegelman (1934); later records were made
by Ratliff and Riggs (1950,. 1951) and Ditchburn and Ginsborg (1953).
My own records of tremor date from 1956. Other authors also
have frequently recorded tremor. Most records have shown that the
amplitude of the tremor (its angular dimensions) is comparable with
the angular dimensions of the eye receptors, while its frequency varies

tic

a ' > \
Fig. 60. Photakymographic records of movements of the eye and an upper incisor tooth of a
subject during fixation on astatianarypoinr. a) Horizontal component; b) vertical of records.

114

CHAPTER 111

EYE MOVEMENTS DURING FIXATION ON STATIONARY OBJECTS

115

Fig. 61. Photokymographic record of movements of the subject's two eyes during fixation on
a stationary point a) Horizontal component b) vertical component of the record.

1:1
,I

from 30 to 90 cycles. Ditchburn (1959) notes that tremor is characterized by a continuous spectrum of frequencies up to 150 cps.
As mentioned earlier, any drift of the eyes is accompanied by
tremor, but the two forms of movement a r e independent. To obtain a
general idea of tremor, let us consider Figs. 60 and 61. These figures
show records of the horizontal and vertical components of the eye
movements on a scale at which the tremor can be seen. The records
were made on the moving tape of a photokymograph by means of the
Pi cap. Parallel and simultaneous records of the eye movements and
the movements of a n upper tooth (an incisor, to which a small mirror
was attached during the experiment), to differentiate between the.

Fig. 62. Photo!qmographic record of the horizontal component of the tremor of a stationary
hand resting on the table, and of the vibration of the chin rest (straight line).

movements of the eye and the movements of the head, a r e shown in

Fig. 60. It is clear from this figure that the high frequencies characteristic of the tremor of the eyes were not recorded with the head
movements.
Parallel and simultaneous records of the eye movements a r e shown
r
the tremor of hand and the vibration of
in Fig. 61. ~ d comparison
chin rest (Fig. 62) were recorded under thesame conditions. In these
experiments, the chin rest was fixed to a massive frame mounted on
a felt cushion. During the experiment illustrated in Fig. 62, the subject's hand rested on the chin rest. Recordings such a s those in Figs.
60, 61. and 62 show convincingly that the high-frequency oscillations recorded on the photokymograph tapes a r e not an artifact but correspond
to the eye movements customarily called tremor.
Analysis of the records I obtained of tremor yielded the following
results.
The amplitude of the tremor is 20-40 seconds of angle
(1.0-1.5 diameter of the cones in the fovea). The tremor is composed
mainly of movements whose frequency is 70-90oscillations per second
(much higher than the critical frequency of flicker fusion). As a
result of tremor, the axis of the eye moves over a conical surface,
covering each such surface in approximately 0.011-0.013sec. In other
words, i f the axis of the eye is mentally continued t o its intersection
with the frontal plane, a s a result of the tremor it will describe
elliptical figures on that plane.

3. SMALL INVOLUNTARY SACCADES O F THE EYES

It is relatively simple t o record saccades, andmany investigators
have long held the correct view that saccades for both eyes coincide in
time, in amplitude, and in direction. The small involuntary saccades
of the eyes were f i r s t discovered by Dodge (1907).
To examine in 'rather more detail the small involuntary saccades
arising during prolonged fixation on a stationary point, let us turn to
Fig. 63. This graph shows the distribution of small involuntary saccades in relation to their amplitude. The amplitude of most such
saccades lies between 1and 25 minutes of angle. The minimal dimensions of these saccades a r e 2-5 minutes of angle. The maximal
dimensions a r e approximately 40-50 minutes of angle. It should be
noted that even minimal saccades of only 2-5 minutes of angle a r e
strictly identical for both eyes (Fig. 64). Records show that the duration of the small involuntary saccades, depending on their amplitude,
is 0.01-0.02 sec. This short duration (in conjunction with their small
amplitude) is responsible for the fact that the saccades a r e completely

'

116

CHAPTER 111

10

1.5

20

25

30

35

40

45

50

55

60

Fig. 6 3 . Graph ai distribution of 1000 involuntary horizontal saccades in accordance with

their amplitude during fixation directedtawards astationarypoint. Results obtained with one
subject. Abscissa-amplitudes d saccades i n minutes of angle; ordinate-number of approximately equal saccades.

unnoticed by the observer, to whom the fixation process appears to be

continuous.
It has been found that an observer cannot voluntarily perform
saccades smaller than a certain amplitude. In one experiment the
subject was shown two fixation points separated by a distance of 8
minutes of angle. The problem was tofixate on the points alternately.
The records showed that the saccades, and the alternation of the subject's attention from one point of fixation to the other (which appeared
to the subject a s alternation of the points of fixation) in most cases did
not coincide in time. In other words, the subject could not change the
point of fixation absolutely voluntarily when the distance between them
was commensurate with the amplitude of the smallest involuntary

Flg. 64. Photokymographic record of eye movements durlng fixation on a statlonary porn.
The record sho\nrs cieeariy that even the small saccades for the two eyes colncide in amplltude and dlrect~on.

EYE MOVEMENTS DURING NGFlXATlON ON STATIONARY OBJECTS

117

saccades. The subjective assessment of the moment of alternation of

the points of fixation corresponded not to the times that saccades
occurred, but to the times that attention was switched. In this case
the subject controlled his attention but not his saccades.
Three cases may be noted when saccades a r e not associated with
shift of attention. First, involuntary saccades a r i s e when an observer
tries to cast his eye smoothly around the outlines of a stationary object.
Evidence of the existence of involuntary saccades inthis case is given
by records of the eye movements such a s those shown in Fig. 53.
Second, involuntary saccades arise duringfixation ona stationary point
when the duration of this fixation is longer than a certain time interval.
A 5-minute record of fixation made on a stationary photographic film
appears a s an oval spot (whichcanberegarded a s the projection of the
foveola on the film). The study of records within this spot shows that
many involuntary saccades occur when the image of the point of fixation lies in the center of the fovea, i.e., when there is no need for a
correcting saccade (directing the image of the point of fixation to the
center of the fovea). Finally, involuntary saccades a r i s e when correction is necessary, i.e., when,as a result of drift the image of the point
of fixation begins to move outside or has moved outside the central
region of the fovea. Such saccades a r e especially numerous in people

Fig. 65. Photokymographic record of the eye movement of a patient during fixation on a
Stationary point. Inthiscase thedrift of the patient's eye Follows a direction, andthe velocity
of the drift i s much higher than normal.

CHAPTER 111

EYE MOVEMENTS DURING FIXATION ON STATIONARY OBJECTS

119

stationary object is accompanied by movement of the head, especially

In this case, the eyes must Nm in the orbit
a ,rotary movement.
smoothly s o that, for a certain interval of time, the optical axes
intersect the point of fixation. A fairly complex situation a r i s e s if
the observer's head remains stationary, but the object of perception
moves in space, s o that the eyes must move continually to preserve
fixation. Finally, another complex case is thatwhich arises when both
the observer's head and the visual test object move simultaneously.
Figures 66 and 6 7 will give some idea of the eye movements in
complex conditions of fixation. Figure 66 shows a record of the eye
movements during fixation while the subject was continuously rotating
his head from left to right and from right to left. A record of the
movement of both eyes when the subject fixated on a point performing
oscillatory movements is shown in Fig. 67.

5. NYSTAGMUS OF THE EYES

In certain pathological cases, most frequently disorders of the
nervous system, the process of fixation is accompaniedby nystagmus.
Nystagmus consists in an oscillatory movement of theaxes of the eyes

Fig. 66. Simultaneous recording on a photalcymagraph of the movements of the eye and rorations of the subject's head during fixation on a stationary point. k i n g the experiment the
subject conriouously cotated his head from left r o right and from right to left. The bold line
i s the record d the eye movements. The thin lines running on the edge of the figure are
the records of the head movements.

with certain disturbances of the normal function of the muscular

apparatus of the eye. By way of a n example, a record made on a slit
photokymograph by means of the Picapwiththe patient trying to fixate
on a stationary point is shownin Fig. 65. The drift of this patient's eye
was directed t o the right (the thick sloping lines), and the speed of the
drift was higher than normal. Involuntary correcting saccades (thin
horizontal lines) constantly returned the eye to its initial position.

4, FIXATION ON A POINT IN COMPLEX CONDITIONS

The simplest case-when the subject's head and the object of perception a r e stationary-was examined in the previous section. Conditions a r e much more complex for the observer's eye if fixation on a

Fig. 67. Specimen of a photokymagraphic record of eye movements during pursuit of an

object by means of oscillatory movements.

120

CHAPTER 111

EYE MOVEMENTS DURING FIXATION ON STATIONARY OBJECTS

during which the amplitude of oscillation is tens of hundreds of times

greater than the amplitude of the tremor, while the frequency of the
nystagrnus correspondingly is tens of times lower than the frequency
of the tremor.
Different forms of nystagmus correspond to different forms of
diseases. The same disease in different persons may be associated
with the same o r differenttypes ofnystagmus. As an example, several

Fig. 69. Eye movements of rod monochromat T during fixation on a point. a) Horizontal
movement of a s n i p of photosensitive paper; b) horizontal component of the movements
recorded on a phqtolcymograph; c) record of vertical component of the movements.

Fig. 68. Eye movements of rod monochromats T, A, 0, and P during fixation an a point
recorded on a vertically moving strip of photosensitive paper.

records of nystagmus in rod monochromat subjects a r e shown (the

records were made .on a photokymograph by means of the PI cap). In
rod monochromats, only the rods a r e concerned in vision; the center
of the fovea, occupied mainly by cones, does not function. The records
in Fig. 68 show the eye movements of four rod monochromats a s they
attempt to fixate on a stationary point. These records show how different the types of nystagrnus present in these four cases were.
Nystagmus of the eyes of one rod monochromat is illustrated in Fig. 69.
Finally, in Fig. 70, records of the eye movements of rod monochromats
a r e given during fixation on an oscillating point.
Little study has yet been made of the various forms of nystagmus,
although it is certainly possible that knowledge of the full range of
varieties of this type of eye movement could be clinically useful. Tbe
vestibular forms of nystagmus arising in personsandin some animals

CHAPTER 111

EYE MOVEMENTS DURING FIXATION ON STATIONARY OBJECTS

123

other factors besides these influence the movement and contrast of

the retinal image.
The first point tomentionis that inmany cases the fixation process
is interrupted by blinking movements of the eye, each lasting several
tenths of a second. A blinking movement is accompanied by sharp
changes in the illuminance of the retina and by disappearance of the
retinal image for a certain period of time. While moistening the
cornea with lachrymal fluid, the eyelid completely covers the pupil.
In addition, during the blinking movement, the eyes make a small
rotation and then return to their Initial position (upward, medially,
and back again), taking about 0.1-0.2 s e c for one of these movements
(Ginsborg, 1952).
Movements and rotations of the head, even if small and compensated
to some extent by corresponding rotations of the eyes, always produce
some displacement of the retinal image.
The position of the lens in the eyeand the curvature of its surfaces
do not remain strictly constant during fixation. The continuous movement of the lens is explained by the fact that it is a component of a selfadjusting system with continuous correction. The fluctuating changes

Fig. 70. Record of eye movements dwing pursuit of a moving sphere. 1) k i n g pursuit of
a sphere swinging on a thread by a subject with normal vision; 2) during pursuit by rhe rod
monochromat person T of a sphere moving up end down: 3 and 4) during pursuit of a sphere
swinging on a thread by the rod monochromat persons A and 0. Movement of the tape is in
a vertical direction.

in response to certain types of excitation of the vestibular apparatus

have, however, been studied most intensively.

6. FACTORS INFLUENCING THE MOVEMENT AND

CONTRAST OF THE RETINAL IMAGE
As mentioned earlier,. movement of the retinal image over the
retina and changes in this image itself during fixation a r e not determined only by the drift, tremor, and saccades of the eye. Many

Fig. 71. Record of the pulsation of theeye. a) Healthy eye: h) eye of a patient with glaucoma
(increased inuaocular pressure: the higher the pressure, the smaller the amp-);
c and d) in patients with disrurhances of the vascular system of the eye. The thick lines are
records of eye movements, the thin lines record the pulsation.

CHAPTER Ul

thus produced in the parameters of the optical system of the eye lead
to an irregular displacement of the retinal image and to a change in
its sharpness. Although such changes a r e small, they nevertheless
can be detected in some experiments with a stabilized image.
The constant, although small, change in the size of the pupil which
takes place during fixation leads mainly to slight modulation of the
brightness of the retinal image and to a change in the depth of focus.
Many almost imperceptible and minor changes a r e brought about by
the pulsation of the blood. For example, pulsating movements of the
eye in the orbit (which differ in each individual) have been observed
by means of relatively simple optical instruments and accessories.
With the use of the P5 cap, the pulsation of the eye itself has been
recorded (Fig. 71). Measurements show that, as a result of pulsation,
the curvature of the cornea changes by approximately one hundredth
or several thousandths of a millimeter. Hence it is natural to suppose
that the diameter of the eye as a whole may change considerably.
Pulsating changes in the size of the eye evidently cause negligible
changes in the sharpness of the fixed object.

7. ROLE OF EYE MOVEMENTS

As mentioned above, good conditions of perception require some
degree of discrete or continuous movement of the retinal image over
the retina, and the retinal image itself must possess adequate hrightness and contrast. The experiments described in Chapter I1 showed
that with a gradual slight increase o r decrease in the luminance' of a
test field stationary relative tothe retina the eye is able to distinguish
between fairly small elements of the field, but cannot perceive color
absolutely correctly. If a test field stationary relative to the retina
is illuminated by a flickering light, the resolving power of the eye
becomes so low that details with angulardimensions of a whole degree
(and in some cases, of several degrees) can no longer be seen by the
observer. All colors become more or less distorted in these conditions. In one experiment, the test field stationary relative to the retina
was initially in complete darkness, but was later illuminated by a
bright beam of light and remained illuminated thereafter. Although the
resolving power of the eye was high immediately after the light was
switched on, an empty field formed so quickly (sometimes within 1
second) that prolonged fixation of attention, essential inmany cases of
perception, was impossible.
Modulation of illumination in these conditions candelay the forma-

EYE MOVEMENTS DURING FFlATTION ON STATIONARY OBJECTS

tion of an empty field or prevent its formation, but will always lead to
distortion of visible colors and sometimes to a decrease in the resolving power of the eye. If a test field, stationary relative to the
retina, is illuminated by a bright flash of light, the resolving power
of the eye remains high but, just as in the preceding case, prolonged
fixation of attention is impossible.
I conclude from these facts that good conditions for perception
cannot be obtained if the retinal image is strictly stationary.
Let us try to examine in more detail the types of eye movement
which are needed to create the essential conditions for perception
during fixation of an element of a stationary object. The first fact to
note is that ordinarily the end of a blinking movement or the end of any
saccade (a large voluntary one or a small involuntary one) is always
the heginning of a new process of seeing. I say "new" because I suggest that as a result of any blinking movement or saccade, certain
signals arising from the retina a r e inhibited while others reappear.
Ditchburn, Fender, and Mayne (1959) showed that even saccades of as
little as '2.5 minutes of a x l e in mamitude reveal contours in a test
field which have disappeared due to immobility of the retinal image
(the minimal amplitude of the involuntary saccades arising during
fixation is 2-5 minutes of angle).
It may be assumed that immediately after a blinking movement new
signals arise from the whole retina, and immediately after a saccade
(depending on the amplitude of the saccade and the type of object) new
signals arise from the whole retina or from certain of its parts.
Frequently situations arise when, after a saccade or series of saccades
(especially if they are small), the changes in the illuminance of the
receptors incertain parts of the retina are below the threshold required
for signals to appear, in which case these areas will correspond to an
empty field or to some stage in the formation of an empty field.
The spatial evolution of the fast process (thecomet) shows that the
initial stage of this process (immediately after a sharp change in
illuminance) corresponds to optimal conditions of perception (the front.
uniform part of the comet is always the brightest or blackest, the most
saturated in color). It may therefore be considered that the maximal
number of signals is sent to the visual centers immediately after a
blinking movement or saccade.
Formation of an empty field, which may begin immediately after
a blinkingmovementor saccade, would take place quickly and frequently
much hefore the following saccade if the retinal image remained
strictly stationary. This conclusion is basedonthe fact that the duration of the intensive part of the fast process does not exceed 1-3

CHAPTER Ill
seconds, while the duration of somedriftsduringfixation (Fig. 59) may
amount to 5 o r even 10seconds. Experiments show that the drift of the
eyes is the type of movementwhichprevents the formation of an empty
field during fixation. Because of drift', the resolving power of the eye
and the apparent color of the object show little change during fixation.
In experiments to demonstrate the spatial evolution of the fast process,
the following facts were established. If the image of the border between
black and white fields moves over' the retina a t uniform velocity, and
if this border is projected on the retina by means of an imperfect
optical system, such a s the optical system of the cap, in conditions
of very bright illumination of the eye through the sclera, the border
will be clearly visible to the subject if the velocity is not less than
23-24 minutes ofangle per second. Whensuch a n experiment is carried
out without illumination of the sclera, this border appears a t velocities
of 3-5 minutes of angle per second. On the other hand, records of
eye drifts show that, although the mean velocity of this movement is
5-6 minutes of angle per second, frequently (not less than one or
several times per second) it reaches almost maximal values, i.e., about
30 minutes of angle per second. Comparingthese results it is easy to
conclude that when sharp retinal images (sharpimages of borders) a r e
present, a single drift of the eye issufficient to prevent the formation
of an empty field over. the whole retina.
An attempt was made to investigate the role of tremor using caps
and several accessories. However, no effect of tremor on reception
could be found. This is evidently due mainly to the fact that the
principal frequency of tremor is much higher than the critical frequency of flicker fusion. If tremor has an effect on reception, i t is
only when combined with drift. The results of experiments in which I
tried to discover a s fully a s possible the role of tremor unfortunately
cannot be regarded a s completely conclusive. Forthis reason, I shall
not describe these experiments and shall leave the question of tremor
open.
Ditchburn, Fender, and Mayne (1959) found that low-frequency
tremor has a positive influence on discrimination of the elements of
a test field s&tionary relative to the retina. In this investigation
tremor was produced artificially with frequencies rangingfrom4 to 20
oscillations per second and with an amplitude ofhetween 0.05 and 1.10
minutes of angle. They found that for all the investigated frequencies
of tremor with an amplitude exceeding 0.3minute of angle, the fraction
of time for which the subject saw the test object was increased.
However, the artificial tremor created by these authors has little in
common withnatural tremor in its frequency o r character of movement,

EYE MOVEMENTS DURING FIXATION ON STATIONARY OB JECTj

127

f o r logically, tremor is best defined a s only the high-frequency (over

40 cps) component of the movement of the eye relative to the orbit.
Natural tremor is characterized by a frequency higher than the critical
frequency of flicker fusion. Low frequencies duringfixation should be
classified a s drifts.
During fixation, besides the drifts of the eye, any other factor
causing movement of the retinal image over the retina may prevent
formation of an empty field. For example, if a sharp retinal image is
shifted for several tenths of a secondthroughseveral minutes of angle
because of head movements, no empty field willform. In natural conditions of perception, the fixation process is constantly accompanied
by movements of several types. As records show, it isdifficult to
accept that 9 e sum of these movements does not exceed the frequency required for good conditions of perception even by a s little a s
one per second. Hence, under natural conditions, i t is almost impossible to make an empty field appear even when the observer wishes.
Such an attempt is usually successful only for very short periods of
time and only if the retinal image is highly out of focus.

CONCLUSIONS
Fixation of attention directed towards an element of a stationary
object is accompanied by fixation of the gaze. Subjectively, this fixation of the gaze is perceived by the observer a s fixation by stationary
eyes.
In reality, however, the eye moves in three ways during fixation:
by small involuntary saccades, equal for the two eyes; by drift, slow,
irregular movement of the optical axes in which, however, some degree
of constancy of their positionis retained; andby tremor, an oscillatory
movement of the axes of the eyes of high frequency but low amplitude.
Head movements, blinking movements of the eyes, saccades,drift,
and tremor during fixation on an element of a stationary object create
a certain mobility of the retinal image andprevent the formation of an
empty ,field.
The formation of an empty field in intervals between saccades is
prevented mainly by the drift of the eyes.

"

Chapter IV

SACCADIC EYE MOVEMENTS

As mentioned above, two features a r e characteristic of any saccadic movement of the eyes: (1) an almost perfect identity of the
movements of both eyes; and (2) highvelocity (the duration of saccades
is measured in hundredths ofa second). Under normal conditions these
features a r e constantly observed and may be clearly recorded by any
suitably sensitive method of studying eye movements.
The main function of saccades is to change the points of fixation, to i
direct the most sensitive regionof the retina (the fovea) to a particular
element of the object of perception. The nature of saccades i s r e sponsible for much of the refinement of perception. The high velocity
and correspondingly short duration of the saccades usually permit the
eye to remain in a state of fixation for about 95%of the total time.
The velocity and duration of the saccades were first studied by
Lamansky (1869) who used the method of after-images. During the
experiment, when the points of fixation were changed, a bright flickering source of light was kept in the subject's field of vision. Every
time the change of points of fixationwas complete, the light was turned
off, and the subject perceived the after-image in the form of a broken
line. Given the flicker frequency and the number of flashes imprinted
during the change in the points of fixation, the experimenter could
determine approximately the duration and velocity of the saccade.
Saccades have also been studied in 'detail by Dodge (1907) and more
recently by A. L. Yarbus (1956~). Westheimer (l958), and 8. Kh.
Gurevich (1961).

'

1. DURATION OF THE SACCADE

Let us first consider in detail the question of the duration of the
saccades. In all experiments to record saccades, the PI cap was used

CHAPTER IV

together with a special slit photokymograph, inwhich the photosensitive

paper moved a t a speed of 5 m/sec. The photosensitive paper was
fixed to the large and rapidly turning drum of the photokymograph.
During the experiments the subject was shown two points equidistant
from the axis of the cyclopic eye; he fixated alternately on the two
points. The distance between the points of fixation varied from experiment to experiment. The amplitude of the saccades was determined
entirely from the records on the photokymograph paper, since the
angular measurement of the saccades differed systematically from the
angular distance between the points intended for fixation (usually the
angular measurement of the saccade was less thanthe angular distance
between these points). Because of the high velocity of movement of
the photosensitive paper, the duration of the saccades could he
measured with an accuracy of r0.01 sec.
Some idea of the character of the eye movements during a saccade
may be gainedfrom the records shown in Fig. 72 of horizontal saccades
between two points visible to the ohserveratan angle of 8". It is clear
from Fig. 72 that during a change in the points of fixation the velocity
of the eye movement rises smoothly, reaches a maximum, and then
falls smoothly.
In natural conditions the amplitude of the saccade usually does not
exceed 20". Very often rotations of the eyes exceeding 15' may he
composed of two o r three saccades, or they may even be accompanied
by a corresponding rotation of the head. Lancaster (1941) claims that
about 99%of eye movements a r e composed of saccades less than 15' in
amplitude.
Later we shall study saccades whose amplitude does not exceed 20".
In the f i r s t series of experiments the duration of horizontal saccades
was measured. The results of these experiments a r e given in Fig. 73.
The duration of vertical 'saccadic eye movements is illustrated in
Fig. 74. The same results a r e shown in Fig. 75 f o r saccades taking
place a t an angle of 45" to the horizontal (or vertical) plane.
The following conclusions may be drawn from an examination of
Figs. 73, 74, and 75. The duration of the saccade is dependent on its
amplitude; for Saccades measuring only a few degrees, the duration
lies between 0.01 and 0.02 sec,whileforsaccades of 20" it may exceed
0.07 sec. The durationof thesaccadeis independent of o r only slightly
dependent on, the direction in which i t occurs. This conclusion is
drawn from the fact that the graphs in Figs. 73, 74, and 75 differ only
slightly from each other. It is clear from the graphs (Figs. 73, 74,
and 75) that saccades of the same size may vary in duration. The
durations of the saccades of the same amplitude may differ by as much

SACCADIC EYE MOVEMENTS

a s 0.01, o r even 0.015 second in duration. The next experiments

examined the question of whether the duration of the saccade depends
on the position of the eye at the moment the saccade begins It was
found that, apart from the most extreme positions (positions of the eye
not found in natural conditions of perception), the duration of the saccade depends entirely on its amplimde andis not appreciably dependent
on the position from which this saccadeisperfor~ned. To confirm this
conclusion, the results of one series of experiments a r e given in
Fig. 76. The graph in this figure shows the durations of the vertical
and horizontal saccades made by a subject in a number of unusual
conditions. Vertical saccades were performed when the eye was
turned downward almost to the limit, and horizontal saccades were
performed when the eye was turned to the left almost to the limit. The
distance hetween the points of fixation in hothcases was 3". It is easy
to s e e that the duration of the saccades and their scatter in this case
do not differ substantially from the duration and scatter of thesaccades
of the same amplitude illustrated in Figs. 73, 74, and 75.
Attempts were next madetodetermine the extent to which the duration of the saccade depends on the subject's will, and whether, for
example, a subjectcanperform saccades of the same angular magnitude

Fig. 72. Sample recording of saccades on photosensitive paper movmg at 5mjsec.

CHAPTER IV

SACCADIC EYE MOVEMENTS

i

sec

007

Degrees

Fig. 74. h a t i a n of vertical saccades of the eye a s a function of the angle through
the
eye turned when changing the Paints of fixalion. a) Subject P; b) subject R: I) Readings
obtained while subject's eye was moving upward: 2) the same, while the eye was moving
downward.

a t different speeds, thus changing their duration. The records showed

that a subject cannot voluntarily change the duration and character of
the saccade. Usuillly, all the subjects felt that it was relatively easy to
make the saccades faster o r slower a t will. However, the records
showed that the sensation of a fast saccade appears a s a result of a
decreased duration of fixation on the points betweenwhich the saccade
is made. All attempts to make fast saccades were accompanied by
a reduced (compared with the normal) duration of fixation. Attempts
to make slow saccades produced a very brief intermediate stop
(0.1-0.2 sec). In other words, a subject trying to make a particular
saccade slow, made instead two o r threeordinarysaccades of smaller
amplitude. The intermediate stop in this case was not noticed by the
subject, so that the changes in the points of fixation appeared to be
slow. The duration of saccades, some ofwhich subjects tried to make

CHAPTER IY

134
sec

SACCADIC EYE MOVEMENTS

80dr

0020,

Degrees

4.

Fig. 76. Duration of saccades. Fixarion points given at an angle of 3'. 1) Readings made
while the eyes were w n e d downward almost to the limit. Vertical saccades; 2) readings
made while the eyes were turned to the left almast to the limit. Horizonld saccades.

, ,,
?>,,,

.,,,

5 6 7 8

9 /J

I1 12 W 14 I5t6 77
Degrees

Fig. 75. Duration af saccades of rhe eye at an angle of 45' to the vertical as a function of the
angle through which the eye turned when changing the points of fixation. a) Subject K;
b) subject R: 1) Readings obtained when the subject's eyes moved upward and to the right;
2) when the eyes moved downward and ta the left.

very fast and others very slow, is illustrated in Fig. 77. It is clear
from this figure that for each of the three subjects some of the "fastn
saccades lasted longer than the "slow." From experiments such a s
these it is concluded that, generally speaking, after the parameters of
a saccade have been assigned and the saccade has begun, it cannot be
modified in any way.
Besides the experiments described above, other considerations
favor this view. Since saccades last only several hundredths of a
second, it i s hard to believe that any form of correction can be introduced into their course, o r that the amplitude of the saccade can be
changed during such short intervals of time. Experiments show that
any corrections to saccades, when they occur (as they very often do),
take place by means of small supplementary saccades, but only after
At the time of a saccade, the
the primary saccade is completed.

angula~velocity of movement of the eyes is so great that in perception

of stationary objects the retinal image may be regarded a s "blurred."
Because of this blurring we usually s e e nothing during a saccade, and
it i s highly improbable that in such conditions the eye should be able
to receive information necessary for introducing any form of correction.
The duration of saccades made by the eyes of several obsetvers is
shown ill Fig. 78. From such graphs an attempt was made to discover
individual variations in the saccades of different observers. It may be
concluded from an exammation of the graph in Fig. 7 3 and graph a
sec
O.050

0040 -

.6
8
Degrees

Fig. 77. h a t i o n of saccades of the eye when the observer tried to make fast o r slow
saccades. Fixation points given at an angle of 7'. Subject K: 1) 'fast" saccades: 2) "slow".
Subject P:
3) "fast' saccades; 4) "slow..
Subject R: 5) .fastm saccades; 6) "slow."

136

CHAPTER IV
sec

0.050

8
Degrees

sec
0.050

7
Degrees

Fig. 78. Duration of saccades of five subjects. Fixation points given at an angle of 7".
a-Graph of duration of horizontal saccades; b-duration of vertical saccades. 1 ) Subject K:
2) subject P; 3) subject R; 4) subject I(N, 5) subject T.

Z sec

SACCADIC EYE MOVEMEhm

137

in Fig. 78 that the duration of the horizontal saccades differs slightly

from one observer to another (by several thousandths of a second).
However, graph b in Fig. 78, showing the duration of the vertical
saccades, does not have this clear difference. Hence individual
variations in the duration of saccades in different subjects must not
be taken for granted; in many cases this variation may evidently be
disregarded.
It is clear from Figs. 76, 77, and 78 that, on the average, the
amplitude of the saccades is less than the angular distance between
the points of fixation. F o r example, when the distance hetween the
points of fixation was 7" (Figs. 77 and IS), the mean amplitude of the
saccades performed between these points was 6.5". The experiments
show that this phenomenon is always encountered if the object of
fixation consists of points. If two clearly visible and large enough
vertical lines a r e drawn through two such points, for example on the
horizontal plane, the records show that the amplitudes of the saccades
will not coincide with the angular distance between the vertical lines
in this case either.
Summarizing the data in section 1,it may be said that under normal
conditions the duration of the saccadic movements of the eyes is,
roughly speaking, only a function of the angle through which the eye
turns when changing the points of fixation. The relationship between the
duration of the saccade and the angle through which the eye turns may
be expressed by the following empirical formula (Fig. 79):

T = 0.021~~2
where T is the duration of the saccade in seconds, and a, the angle in
degrees through which the eye turns when changing the points of fixation.
The maximal scatter of the experimental points (Fig. 79) is
approximately 1-0.005-0.007 sec.

2. DEVELOPMENT OF THE SACCADE IN TIME

Fig. 79. Graph showing relationship between duration of the saccade and the angle through
which the eye turns.

Careful examination of the records of saccades in Fig. 72 will

reveal the close resemblance between these records and sinusoidal
curves. Closer examination shows that in fact the records of most
horizontal and vertical saccades not exceeding 15-20" approximate
sinusoids very nearly (oblique saccades arenotnowbeing considered).
As an example, records of horizontal saccades with the points of the
corresponding sinusoid marked directly on the paper tape of the
pbotokymograph a r e shown in Fig. 80. Records of saccades measuring

CHAPTER IV

SACCADIC EYE MOVEMENTS

139

By differentiating equation (I), we obtain a formulafor the angular

velocity (a)of the eye movement during a saccade (Fig. 82):

Fig. 80. Records of saccades. The poinrs of the corresponding sirmsoid are plotted on one
of the tracings.

15-20" show a much l e s s close approximation to the sinusoidal curve.

In the middle part of the curve, rectilinear areas appear (corresponding
to uniform velocity), and the curves become slightly asymmetrical
(the time of increase in velocity of the saccade is apparently shorter
than the time of its decrease). According to some authors (Hyde, 1959),
this asymmetry is particularly noticeable with very large saccades
(50-60"). Naturally, therefore, recordings of saccades larger than
15-20" cannot approximate a sinusoid. For saccades of less than
15-20", we can give a formula describing the change in the angle Of
rotations of the eye during the saccade in time (Fig. 81):

Naturally this formula only approximately describes the course of

saccades measuring 15-20' and it cannot beappliedto saccades larger
than 20".
Assuming that the radius of the eye is 1.2 cm, and using formula
(2), we obtain an expression for the linear velocities ( u ) of the center
of the cornea in the process of a saccade not exceeding 2Cf in amplitude
(Fig. 82).
u = 0.021 0.
(3)
It follows from equation (2) that the velocity of the saccade rises
smoothly, reaches a maximum, and then falls smoothly to zero. For
saccades smaller than 15-20", the increase and decrease of velocity
follow a sinusoidal rule (the times of increase and decrease in velocity
a r e approximately equal). For large saccades (exceeding 20") the
increase in velocity occupies less than half the total duration of the
saccade. A correspondingly longer period of time is occupied by the
decrease in the velocity of the saccade. The maximal velocities of the
saccades a r e clearly dependent on their amplitude.

where t is the time in seconds, (0 < t < T ) , a is the angle of rotation

during the saccade in degrees (0<< a < < a , ) , T is the duration of the
saccade in seconds, and a, is the amplitude of the saccade in degrees.

t,sec
Fig. 81. Relation~hip between angle of rotation of the eye and time during a saccade.

1,sec
Fig. 82. Relationship between angular and linear velocityof eye movement and time durlng a
saccade.

CHAPTER IV

140

For example, for a saccade of 5 , the eye reaches a maximal

velocity of 200 deg/sec, and for a saccade of 20". about 450 deglsec.
The study of records similar to those shown in Fig. 72 demonstrates
that maximal acceleration during saccadic movements is attained by
the eye at the very beginning of the saccade and at its end (at the end,
the acceleration has the opposite sign, slowing the saccade).
The absolute magnitudes of the twoaccelerations for small saccades
are almost identical, but for large saccades (exceeding2O0),they differ
considerably (the accelerationat the beginning of the saccade is greater
than the acceleration a t its end).
The magnitude of the accelerations clearly depends on the amplitude
of the saccade. For example, for a saccade of 5" it is approximately
15,000 deg/sec2, and for a saccade of 20" about 20,000deg/sec2.
Disregarding friction between the eye and the orbit, regarding the
eye as a sphere and the vitreous and the other intraocular media as
rigidly fixed to thesclera, and knowing theaccelerations, it is easy to
calculate the forces setting the eye in motion during the saccade.
Assuming that the radius of the eye is 1.2 cm and its specific gravity
is 1, we obtain that the maximal effort of the muscles at the beginning
of a saccade of 5" is approximately 1g, and that a t the beginning of a
saccade of 20" about 1.5 g. Whenconsidering this ideal case, we must
naturally remember that the eye does not turn in a medium of air but
in the orbit. For this reason, even assuming that muscular effort is
used up only in rotating the eye, in this case it must be considerably
larger than the forces indicated in these examples.

3. INCLINATION OF SACCADES
Many experiments have shown that saccades in a horizontal or
vertical direction a r e m most cases recorded on stationary photosensitive paper as straight lines. Saccades performed at an angle to
these two directions are most frequently recorded a s curved lines.
Two questions arise in this connection: what is the reason for this
curvature of the lines; and to what extent can we apply our conclusions
regarding horizontal and vertical saccades to oblique saccades Let
us first turn to Figs. 83 and 84.
Figure 83 shows records of saccades between all comers of two
equal squares situated in the frontal plane. The records were made
with the Pi cap on stationary photosensitive paper. One square was
placed so that two of its sides were horizontal and the other two
vertical. The other square was so placed that all its sides made an

SACCADIC EYE MOVEMEATS

141

Fig. 83. Record of saccades berween the cornersof mo squares on stationary photosensitive
paper.

angle of 45" with the horizontal. The sharp differences between the
records of the saccades in these two cases a r e clear from Fig. 83.
To determine whether the curvature of the oblique saccades is
due to rotary (around the geometrical axis) movements of the eyes, an
experiment was carried out using the Pi cap with the mirror in an
unusual position. The mirror was fixed tothe cap so that it was in the
sagittal plane during the experiment. In other words, it was parallel
to the vertical section passing through the axis of the eye, and its
reflecting surface was directed towards the temple. The source of
light and the photosensitive material (the photographic paper) were
arranged correspondingly on the temporal side.
During the experiment the subject was asked to make several
saccades with his eyes between all the corners of the two squares
situated in the frontal plane, as described in the previous experiment.
This time the position of the mirror was so arranged that only the
horizontal movement of the eye and rotary movements of the eyes
around the geometrical axis (or, roughly speaking, around the optic
axis) could be recorded on the photosensitive paper. The vertical
movements could not be reproduced. The results of this experiment
are shown in Fig. 84.

Fig. 84. Saccades between the corners of rwo squares on statlonary photosens~t~ve
paper
recorded when only the horizontal components of the movements and the rotary movements
of the eye around the optic axis could be recorded.

SACCADIC EYE MOVEMENTS

When performing saccades between the corners of thesquares, the
eye performed eight different movements: two in a vertical direction
(up and down), two i n a horizontal direction (from left to right and from
right to left), and four in various oblique directions. These results
suggest that under normal conditions, when the observer's head and the
object of perception a r e stationary, no appreciable movements a r e
observed around the geometrical axis of the eye, and, consequently,
the curvature of the lines obtained by recording oblique saccades
cannot be attributed to rotary (around the geometrical axis) movements
of the eyes. The curvature of oblique saccades may take place a s
a result of the nonsimultaneous working of the different muscles. We
may note, incidentally, that fairly considerable rotary movements of
the eyes (around the orbit) may be observed in conditions when the
subject examines a stationary object, and when so doing turns his
head around the axis of the cyclopic eye.
If two harmonic oscillations aresuperposedatanangle, curvilinear
trajectories known a s Lissajous figures a r e obtained. Since the
records of horizontal and vertical saccades- (less than 15-20") approximate a sinusoid, the possibility remains that the records of
oblique saccades consist mainly of Lissajous figures.
Naturally, when applied to oblique saccades, the equations given
in the preceding section a r e obse'rved less accurately than for horizontal and vertical saccades. However, a s Fig. 75 shows, the duration
of oblique saccades is indistinguishable from the duration of horizontal
and vertical saccades of the same amplitude. Analysis of the records
of oblique saccades (analysis of the component along the straight line
connecting the points of fixation) alsoshows thatalong the straight line
connecting the points of fixation all the characteristics of oblique
saccades a r e almost indistinguishable from the correspondingcharacteristics of horizontal o r vertical saccades. For this reason, in
calculations of practical importance, any saccades (not exceeding
15-20") of the same amplimde can usually be regarded a s equal in all
respects.

4. CENTER OF ROTATION OF THE EYE

I

George, Toren, and Lowell (1923) demonstrated that in a zone

bounded by the angle of rotation of the eye by 20" towards the temporal
side and by 30" towards thenasal side,the center of rotation of the eye
may be regarded a s a fixed polnt situated 15.4 mm from the apex of
the cornea and 1.65 mm to the nasal side of the optic axis. Outside

143

Table 1
Direction of opric axis
Distance of center of
rotation fiom apex of
cornea, mm
Distance ofcenter of
ratation from opric
axis, mm

39' towards i~ose 4' rowaids nose

i4.73

1.066

13.92

1.653

38- towards temple

12.95

0.893

the limits of this zone the center of rotation of the eye, according to
these workers, is no longer fixed. Park and Park (1933) repeated
these observations and obtained the results a s to the localization of
the center of rotation of the eye for three directions, a s shown in
Table 1.
These results (even if of doubtful accuracy) show that, for the
malority of eye movements (excluding rotations through anglesgreater
than 20-3O0), the center of rotation of the eye may be regarded a s a
stationary point. However, the possibility remains (Lord and Wright,
1950) that this center may be displaced and may even lie outside the eye
itself during certain small saccades. Such a displacement may arise,
f o r example, if a saccade is accompanied by linear displacement of
the eye relative to the orbit.

5. BEGINNING AND END OF THE SACCADE

Normally, for the perception of stationary objects the eye must be
in one of two states-in a state of fixation o r in the process of changing
the points of fixation. Let us examine the eye movements during the
transition from f k i t i o n to the process of changing the points of fixation, and vice versa. So a s not to anticipate the issue, let us exclude
from our examination the case where points of fixation a r e shifted
during pursuit o r a t moments of convergence o r divergence of the eyes.
The many records which have been made show that the transition from
a state of fixation to a saccade (for horizontal and vertical saccades
of less than 15-20') may be expressed by equation (1) in section 2. As
a rule this equation is valid not only for the main part of the saccade,
but also for its beginning, lasting for the first few milliseconds. The
transition from the state of the saccade to the state of fixation (lasting
f o r the last few milliseconds) cannot always be expressed by equation
(1)in section 2, even for small saccades. Some delay in slowing of

CHAPTER I V

Fig. 85. Record of s change m the polnrs of flnatlon on a sllr phorokymograph. The overshoa of the eye is clearly vlsiole on the records.

the saccade is frequently observed, a s a result of which the eye

apparently shoots past the point a t which it will s t o p a t the next
moment, but then returns to it (without a correcting saccade). A
s e r i e s of these overruns is shown in Fig. 85. It should be noted that
different subjects overshoot by different amounts. In the same
subject, some saccades terminate by overshooting, while others show
this tendency hardly a t all. The degree of overshoot is independent, or
almost independent, of the amplitude of the saccade. In normal Conditions, by comparison with the large saccades, the overshoot is s o
slight that it can almost be disregarded. Often, by comparison with
small involuntary saccades; the overshoot is considerable; it is especially significant for saccades resembling tremor of the eyes (tremor
of the eyes during fixation).

6. VISION DURING THE SACCADE

!,.

I1

During perception of stationary objects, a t .the moment of the

saccade no visual images a r e formed because the high velocity of the
retinal image leads to "blurring" of everythingfallingwithin the field of
vision. We can observe this "blurring," for example, if we look at the
road with a fixed glance from the window of a. fast moving automobile,
o r if we examine a rotating disk.

SACCADIC EYE MOVEMENTS

145

The question may arise: is this the only cause preventing the
appearance of a visual image during a saccade? Is there no special
type of inhibition arising concurrently with the saccade and excluding
vision during its course? Several very simple experiments give a
negative answer to this question. One such experiment is a s follows.
A large disk is placed before the observer. Along the perimeter of the
disk is drawn or glued a large, periodically repeated pattern. For
example, pieces of white paper may be glued along the perimeter of
a dark disk. The disk is then turned a t a speed at which, while the
observer fixes his gaze onthe edge of the disc, the patterns completely
merge. If the observer thenchanges his points of fixation in the direction of movement of the patterns so that the angular velocities of the
movements of the eyes and of the pattern a r e very close, for very
short intervals of time (during the saccade) he will clearly s e e the
patterns on the disk. Consequently, it can be concluded that during
the saccade the eye does not lose its PerceDtive Dower: if the eye sees
nothing, it is simply because it has to deal with a retinal image moving
a t high speed.
Examination of the evolution of the after-image in a completely
darkened room reveals the following curious fact: nearly every saccadic movement of the eyes, if large enough, is accompanied by the
temporary disappearance of the after-image, if it is weak; o r by a
change in its color if it is strong. The important thing here is that the
temporary disappearance of the after-image lasts much longer than the
saccade. Is this disappearance of the after-image due to some weak,
partial inhibition of perception coinciding with. the time of the sacc a e s ? , ( I say partial inhibition because, a s we have seen, complete
inhibition evidently does not occur.) The apparent color of the afterimage may change suddenly in response to achange in the illuminance
of the eye. In a totally dark room, however, the illuminance i s zero.
In this case, therefore, only conjecture ispossible; it could be that
during the time of the saccade the effect 1 have mentioned is caused by
a very slight dynamic shift of the intraocular media relative to the
sclera. h e shift may be too smail to evoke movement phosphene,
but sufficient to influence the apparent color of the after-image.

7. VOLUNTARY AMD INVOLUNTARY SACCADES

In the preceding chapter we discussed the small involuntary saccades accompanying fixation on a stationary object. Such small saccades a r e unnoticed and cannot be produced a t will.

146

CHAPTER IV

It was mentioned above that even large saccades, those by means

of which we change our direction of fixation, may be either involuntary
o r voluntary. There is much factual evidence to show that many large
saccades a r e involuntary. F o r example, immediately after having
examined an object we can say only approximately on which of its elements we fixed. The observer is in an extremely difficult position if
he is asked to estimate, even approximately, the number of saccades
performed in a very short period of time. He can easily perform
large, voluntary saccades if he is asked to do so, for example, in the
conditions of a particular experiment. In natural conditions of perception, even a shift of attention is not always voluntary, and the
saccades accompanying the perception process a r e mainly unnoticed
by the observer. Usually changes of attention remain in our memory.
but not changes of points of fixation.
When we have learned to walk we do not think which foot to move
first, we simply walk.. When we have learned to look we do not think
about what order or what points of fixation to choose when looking
at an object, we simply look. But both in walking and in looking at
objects, the "simplicity" of the movements is really very complex.
The change of the points of fixation and the choice of these points take
place in accordance with certain general principles which will be
examined fully below.. Here I shall merely mention that even saccades
which a r e certainly voluntary a r e not always or entirely submissive
to our will.

CONCLUSIONS
Any saccade of the eyes (a sharp rotation of the optic axes) shows
characteristically an almost ideal identity of the movements of both
eyes, and a high velocity. The main purpose of the saccades is t o
change the points of fixation, to change the direction of the most highly
developed region of the retina (the fovea) to a particular element of
the object of perception. Under natural conditions of perception, the
amplitude of the saccades usually does not exceed 20". Saccades of
minimal amplitude measure 2-5 minutes of angle. The duration of
the saccade changes with a change in its amplitude. For angles less
than 1"the duration of the saccades is 0.01-0.02 sec; for angles of 20"
it may reach 0.06-0.07 sec. The maximal velocity reached by the eye
during a saccade of 20"is about 450 deg/sec. Under normal conditions,
the duration of equal saccad-es in different observers is approximately
the same; it cannot be variedatwill by the observer and is determined
almost entirely by the amplitude of the saccade.

Chapter V

EYE MOVEMENTS DURING CHANGE OF

STATIONARY POINTS OF FIXATION
IN SPACE
If points of fixation a r e removed to various distances from an
observer's eyes, the change is accompanied not onlyby a saccade, but
by convergence of the eyes (convergence or divergence of the optical
axes). Together with accommddation (and other factors), the relative
position of the optical axes and the retinal images enables us a t moments of fixation to judge the distance and size of objects. In this
book there i s noneedtodwell indetail on questions of binocular vision;
the necessary information can be obtained from other sources-the
book by S.V. Kravkov (1950), for example. The important issue s o
far a s we a r e concerned is that the convergence and divergence of the
optical axes during a change in points of fixation differ sharply from
saccades, particularly in duration. The duration of convergence and
divergence of the eyes is approximately ten times that of saccades.
Convergence or divergence of the eyes very often is equal in duration
to fixation.
It may be assumed that the considerable difference between the
duration of saccades and that of convergence or divergence is due to
the'fact that the saccade follows a predetermined program, whereas the
program for convergence and divergence cannot be laid down beforehand.

1. CHANGE OF STATIONARY POINTS OF FMATION IN

SPACE
Let us first attempt to explain the general character of eye movements during change of stationary points of fixation in space. I have
147

CHAPTER V

EYE MOVEMENTS DURING CHANGE OF STATIONARY POINTS OF FIXATlON

149

cUA
Fig. 86. Record of divergence of the left eye of a subject on starianary photosensitive paper
during a change of stationary points of fixation. Both points of fixation are situated in the
sagittal plane. The mare distant point (6) i s higher than the nearer point (A).

already remarked that this change is composed of two types of movements-the convergence o r divergence of the optical axes and the saccade. More specifically, it is possible to change the points of fixation
in space in sucha way that there is no need to rotate the eye; the change
then will consist purely of convergence o r divergence. In the subsequent pages we shall study thegeneral case of this type of movement.
Let us examine Figs. 86-88. From the records of eye movements
shown in these figures, it is clear that a change in points of fixation in
space is composed of two types of movement, and that a saccade of the
eyes is always preceded by an initial stage of convergence or diver-

Fig. 87. Record of convergence of thelefteye of a subject on stationary photosensitive paper

during a change of stationary points of fixarion. Bothpoints of fixation are situated in the
sagittal plane. The more distant point (6) is higher than the n e a r paint (A).

Flg. 88. Record of the movements of a subject's left eye durlng alternar~onof two points of
flxar~onslmaced m the saglctal plane. The more distant pomt (6) 1s hxher
. than the nearer
pomnt (A).

gence. In the analogous experiments illustrated in Figs. 86, 87, and

8 8 , the points of fixation (A and B) were white balls arranged in the
sagittal plane relative to the subject in such a way that point B was
further away and slightly above point A. The angular dimensions of
the balls were 2". The Pi cap was used in all experiments. The
minimal distance between the subject's eyes and point A was 15 cm,
and the maximal distance to pointB 950cm. The eye movements were
recorded on stationary photosensiti\e paper in front of the subject.
One of the records from the subject's left eye during five successive
changes from point of fixation A to point of fixation B is shown in
Fig. 86; the record from the right eye (relative to the sagittal plane)
in this case was a mirror image of the record from the left eye.
All the records, which were similar to that shown in Fig. 86,
showed that a change of the points of fixation (from point A to point B)
follow the same scheme. In the first period there is very slight
divergence (AC), after which the eye performs a saccade in the
direction of point B (CD), and after the saccade the main part of the
divergence (DB) takes place. If the fixation points a r e far apart and
the saccade of the eyes is not accurate enough, often the second phase
of divergence (DB) is accompanied by additional correcting saccades
of the eyes.
Such corrective saccades can be seen in Fig. 86.
A record of the movements of the left eye during a change from
point of fixation B topoint of fixationA is shown in Fig. 87. Slight con-

"

CHAPTER V

photographic tope

50-100-50cm

Fig. 89. Record of movements of both eyeson a photokymqgraph during a change of points of
fixation in space.

vergence is observed in the first period (BE), after which the eye performs a saccade in thedirectionofpoint A (EF); and after the saccade,
the main part of the convergence (FA) takes place and is sometimes
accompanied by additional corrective saccades. In Fig. 88 a r e shown
the movements of a subject's left eye during successive changes in
points of fixation from Ato B and B to A. In Fig. 89 i s shown a record,
made by means of a slit photokymograph, of the movements of both eyes
during a change in points of fixation. It is clear from this figure that

EYE MOVEMENTS DURING CHANGE O F STATIONARY POINTS OF FIXATION

151

convergence and divergence begin and end smoothly. In most cases the
eye movements a r e different during convergence anddivergence(Figs.
89 and 90).
The convergence and divergence of the optical axes may continue
for several tenths of a second. The eyes perform smooth and slow
convergence or divergence during the initial and final segments of
this time (Figs. 89 and 90). The experiments show that a normal
process of perception is possible during these segments of time (there
is no duplication of the visual objects). In general, some degree of
convergence and divergence of the optical axes during fixation does
not cause duplication of visible objects (Kravkov, 1950); because of
this, part of the time occupied by convergence and divergence of the
eyes can at the same time be employed for perception. This fraction
or divereence
of the time is particularly large when the convergence
of the optical a i e s is very-slight.
Frequently the maximal angular velocity of the eye movement
during convergence or divergence may attain several tens of degrees
per second. Naturally, with these speeds, normal perception is no
longer possible.
Simple experiments showthat the eyes remain capable of perception at the moment of convergence or divergence (as during a saccade).

2. DURATION OF CONVERGENCE AND DIVERGENCE

In a large series of experiments using several subjects, the duration of convergence and divergence was measured. The P, cap was
used in these experiments. Records were made simultaneously from
both eyes on a slit photokymograph. The results of the measurements
a r e shown in Fig. 91. The dots inside each column denote the duration
of convergence, corresponding to a change in points of fixation situated
a t the following distances (in cm) from the subject's eyes: 25
15,
35 25, 45 -35, and so on.
The x's inside each column denote the duration of divergence,
corresponding to a change in points of fixationsituated a t the following
distances (in cm) from the subject's eyes: 15 25, 25-35, 35 -45,
etc.
The results of the experiments illustrated in Fig. 91 demonstrate
that the duration of convergence and divergence is measured in tenths
of a second. By way of comparisonit is well to recall that the duration
of the saccades is measuredinhundredths of a second. Under identical
conditions, the duration of convergencemay differ considerably (sometimes one may be twice the other). On the average, the duration of

Fu. 90. Record of movements of botheyrs on d photokymograph durlry a change of paints

of fU(&rlon m space. .\case of nsymmetrlcal eye movements during con!,ergence and diver-

152

CHAPTER V

EYE MOVEMENTS DURING CHANGE OF STATMNARY POUTS OF FIXATION

153

denote the duration of the initial stages of convergence during a change

in points of fixation situated at the followingdistances (in cm) from the
subject's eyes: 25
15, 35
25, 45
35, and so on. The x's inside
each column denote the duration of the initial stages of divergence
during change in points of fixation situated at the following distances
from the subject's eyes: 15 25, 25 -35, 35 -45 cm, and so on.
Naturally, in these experiments two adjacent points of fixation were
always slightly displaced relative to the axis of the subject's cyclopic
eye, so that a change from one to the other would be accompanied by a
saccade.
It is clear from the data in Fig. 92 that with any change of stationary
points of fixation inspace the durationof the process of convergence or
divergence preceding the saccade is between O.07and 0.2 sec, and that
this value i s roughly constant. Since the preparation for a saccade to
an object suddenly appearing in the subject's field of vision or preparation for pursuit of a suddenly appearing object takes 0.15-0.17 sec,
i t can be concluded that the onset of covergence or divergence preceding the saccade coincides with the onset of preparation for the
saccade.
In other words, i t may be concluded that a change of points of
fixation in space begins a t once by two processes-convergence o r
divergence and the preparation of the program of the saccades. The
fact that convergence o r divergence begins before the eyes rotate
towards the new object of fixation, and not after this rotation, appreciably shortens the timerequiredforchaugingthe points of fixation.

cm
Fig. 91. h v a t i a n of convergence and divergence of the eyes.

divergence is slightly longer than the duration of convergence.

The duration of convergence and divergence of healthy eyes of
different subjects is approximately equal and is not appreciably
dependent on the subject's will to make these movements faster or
slower.
In one series of experiments, conditions were created in which a
new object of fixation appeared unexpectedly before the subject. The
subject's task was to begin fixating on the new object immediately.
Measurements showed that under these conditions the character and
duration of covergence and divergence were indistinguishable from
those observed in the previous experiments.

4. SCHEME O F EYE MOVEMENTS DURING CHANGE

O F STATIONARY POINTS O F FIXATION IN SPACE
The study of many different records has shown that in any change
of the .points of fixation in space the process of convergence or disec

3, CONVERGENCE AND DIVERGENCE PRECEDING A

SACCADE
A s demonstrated above, during a change of stationary points of
fixation in space, the. saccade is always preceded by an initial stage
of convergence or divergence. In a series of experiments similar to
that described above, the duration of these initial stages was measured;
these measurements a r e shown in Fig. 92. The dots inside each column

cm

Fig. 92. Durarlon of convergence and divergence preced~nga saccade.

CHAPTER V

0.2 sec

Fig. 93. Photokymographic record of eye movements during a change of stationary points
of fixarion in space. The broken lines are copies of the recording of divergence preceding
the saccade, displaced by the magnitude of ?he saccade.

f,

*
q

!'

1'
j'
!

i'

I'

1,

1'

I'

I
j

vergence is continuous, regardless of whether it is accompanied by

saccades (when the points of fixation lie on different axes of the cyclopic
eye) o r not (when the points of fixation lie along the same axis of the
cyclopic eye). In other words, the process of convergence or divergence of the eyes before, during, and after the saccade is the same
process of continuous change in depth of the point of fixation. Any
change of stationary points of fixation in space (unless these points
a r e situated along the same axis of the cyclopic eye) consists of the
sum of two independent eye movements-convergence or divergence,
and saccade. When stationary points of fixation a r e situated on the
same axis of the cyclopic eye, the eye movements during a change of
these points may take place without a saccade.
A record of a change of the points of fixation in space (record made
from both eyes on a slit photokymograph) is given in Fig. 93. The
arrow shows the displacement of the axis of the cyclopic eye during
the saccade. The broken lines a r e copies of the recording of the
divergence preceding the saccade, displaced by the magnitude of the
saccade. It is clear' from the figure t h t the parts of the records of
the divergence flow smoothly to make a single whole. The scheme
illustrated in Fig. 94 can be constructed from the study of records
such a s these.
In the left part of the figure is shown a diagram of eye movements
during a change from a point of fixation A to a more distant point of
fixation B. In the right part is shown a diagram of the eye movements
during a change from a pointof fixation B to a point of fixation A nearer
the subject.
The eye movements during a change from point of fixation A to

EYE MOVEMENTS DURLNG CHANGE OF STATIONARY POINTS OF FIXATlON

155

point of fixation B a r e composed of: (a) divergence, preceding the

saccade in the portion AC, during which the point of intersection of
the optical axes moves along the axis of the cyclopic eye, directed to
the point A; (b) saccades of the eyes, i.e., rotation of the axis of the
cyclopic eye in.the direction of point B and accompanied by continuing
divergence; and (c) further divergence of the eyes, during which the
point of intersection of the optical axes moves along the line DB, i.e.,
the axis of the cyclopic eye directed to the point B.
During the change from point of fixation B to point of fixation A,
the movement of the eyes is composed of: (a) convergence preceding
the saccade, during which the point of intersection of the optical axes
moves alongthelineBE, i.e.,alongthe axis of the cyclopic eye directed
to point B; (b) saccades of the eyes, i.e., rotation of the axis of the
cyclopic eye in the direction of point A, with continuing convergence;
and (c) further convergence of the eyes, during which the point of
intersection of the optical axes moves along FA,i.e., along the axis of
the cyclopic eye directed to the point A. A s we have said, sometimes
a change of the points of fixation is accompanied by supplementary
correcting saccades. In this.case,thecharacter of the eye movements
remains a s indicated in Fig. 94.
In many cases a very small corrective divergence or convergence
is found a t the end of the process of changing the points of fixation.
This additional convergence or divergence introduces no essential
change into the scheme of the eye movements, but leads to considerable scatter of the durations of convergence and divergence. A more
complete idea of the character of the eye movements during changes
of stationary points of fixation in space may be gained from study of
Figs. 95, 96, and 97.

-Fig. 94, Diagram of eye movements during achange of stationary paints af fixation in space.

CHAPTER V

EYE MOVEMENTS DURING CHANGE OF STATIONARY POINTS OF FIXATION

Distance before
orbitrorily

157

+o
.
4

Fig. 97. Phorolcymagraphic record of eye movements during interchange of a series of

Stationary points of fixation in space.

Fig. 95. Sample records of eye movements (photokyrnograph) durlng repeated replacement
of one stationary fixed paint by another.

5. APPARENT SIZE OF OBJECT AND DIRECTION OF GAZE

If the distance between the eyes and the object remains constant
and only the apparent size i s considered, the size of the object will
depend primarily on the size of the retinal image and the position of

Fig. 96. Photakymographic record of eye movements during change of stationary polnts of
fixation in space siNated along the axis of the cyclapic eye.

the optical axes of the eyes. The tension of the muscles when the
optical axes a r e in a certain position and the size of the retinal image
a r e so related that, given the same retinal image with differences in
convergence, the apparent size of the object will change, decreasing
a s the convergence increases. This relationship, arising a s a result
of our experiment, is responsible forthe correct estimation of the size
and position of an object in space (the decrease in the apparent size
of an object a s covergence increases, with a constant retinal image,
can easily be determined on a Wheatstone stereoscope).
The system of the eye muscles is so constructed that, if there is
no fixation (for example, in total darkness), when the eyes a r e rotated
upward from a central position or downward, the axes of the eyes
diverge slightly. This involuntary divergence of the optical axes places
an additional load on the eye muscles during colivergence. The additional load on the muscles, in. cases when the eyes a r e strongly
deviated from the central position upward or downward, naturally
causes some decrease in the apparent size of the object. Special
experiments have revealed the distortions arising in this case. The
differences in apparent size of two equal circles (situated the same
adequate distance from the observer's eyes) a r e shown in Fig. 98;
the smaller circle corresponds to the position in which the eyes a r e
turned upward to the limit, the larger, to the position in which the
eyes a r e directed along the sagittal plane to the line of the horizon.
Ordinarily, with relatively small rotations of the eyes, these distortions

CHAPTER V

Fig. 98. Differences in the apparem size of w o equal circles. The smaller circle corresponds to the position of the eyes Nrned upward to the limit; the larger circle corresponds
to the position of the eyes directed along the sagittal plane to the horizon line.

can always be ignored (under ordinary conditions, they a r e never

noticed by the observer).

CONCLUSIONS
Any change of stationary points of fixation inspace is composed of
two independent movements-convergence or divergence, anda saccade.
The saccade is always preceded by an initial period of convergence or
divergence. The duration of this period is approximately the same in
all cases.

EYE MOVEMENTS DURING PERCEPTION

OF MOVING OBJECTS
If a visual object is inmovement,theobserver's eyes usually move
in pursuit of it. The aim of this pursuit is to make the retinal image
of the object a s stationary a s necessary relative to the retina. By a s
stationary a s necessary, we meanavelocityof movement of the retinal
image so small that the eyes retain a high level of resolving power.
The actual movement of pursuit is possible only if it is controlled by
some form of system, hereinafter termed the system of pursuit.
If the movement of thevisual test object is complex, the movements
of the eyes will also be very complex. They may be composed of
pursuit movements, saccades changing the points of fixation of the
moving object, corrective saccades appearing under particularly
complex conditions of perception, and, finally, movements of convergence and divergence of the optical axes, which may take place contiriuously a s the need arises and a r e merged with the movement of
pursuit.

1. INVOLUNTARY ASPECTS OF PURSUIT OF MOVING

OBJECTS
To determine to what extent the system of pursuit works involuntarily and automatically, a series of experiments was performed. In
one experiment, P2 caps were fixed to both anesthetized eyes of a
subject, completely covering the cornea. The subject was in total
darkness. Each cap was fitted with a small mirror. A beam of light
was projected onto the mirror and reflected onto a screen a s spots
of light; by following these, the experimenter could estimate the
movements of the subject's eyes and, if need be, record them.
159

160

CHAPTER VI

When the subject was instructed to imagine that he was looking a t

an object, the character of the eye movements coincided with the
character of the movements during the perception of stationary objects.
A change in imaginary points of fixation situated in the sagittal plane
a t different distances from the eyes alsowas accompanied by a clearly
defined convergence and divergence of the eyes. The only appreciable
difference from normal was a decrease in the accuracy of fixation.,
Attempts to pursue the movement of the imaginary object smoothly did
not have the desired result. Althoughitalways appeared to the subject
that the movements of his eyes were smooth and continuous, in fact
they consisted entirely of separate fixations, saccades, convergences,
and divergences.
In the next experiment, a smooth, uniformly illuminated screen was
placed before the subject's openeyes. A tJl cap was fixed to each eye.
The records in this case showed that the subject could reproduce a t
will all types of eye movement except smooth movements of pursuit.
An experiment described and illustrated earlier demonstrated that it
was also impossible to obtainsmoothpursuingmovements wht..~tracing
the outline of astationary figurewiththeeye. It may.thus be concluded
that in normal conditions the system of pursuit of the eyes cannot be
operated a t will without the presence of an object moving in the field
of vision.
Usually we can s t a r t or stop completely at will the pursuit of an
object moving in our field of vision. If the object is large enough,
the system of pursuit is often set in motion involuntarily. Moreover,
if the moving object occupies the whole field of vision or a large enough
part of that field, in some conditions, once put into operation. the
system of pursuit cannot be stopped until the object stops moving o r
the subject closes his eyes..
In one experiment the P4cap (see its description) was used. After
the cap had been affixed to the eye, the subject saw objects only by
means of the small m i r r o r attached to the apparatus, i.e., by means
of a m i r r o r moving together with the eye. Under these conditions,
eye movements evoked movements of the retinal image. The relationship between the angle of rotation of the eye and the angle of rotation
of the retinal image was always very complex. The subject saw the
objects clearly but could not choose his points of fixation a t will, i.e.,
he could not use eye movements to obtain information on the spatial
relationships between objects.
The fact that objects visible to the subjectthrough the cap could not
be examined a t will created an unpleasant sensation for him. This
unnatural state frequently had the resultthatthesystem of pursuit was

EYE MOVEMENTS DURING PERCEPTION OF MOVING OBJECTS

161

s e t in operation involuntarily, and the eye began to make fruitless

oscillatory searching movements, accompanied by oscillations of the
visual images. In such conditions the subject could not stop his eye.
i.e., he could not prevent his system of pursuit from functioning, and
the oscillations of the eye continued until the end of the experiment.
If the image of an object moving with constant velocity appeared
in the field of vision (in the mirror of the P4cap), this image was
always pursued with acceleration. As soon a s the eye began to follow
the movement of 'the image of the object, the m i r r o r of the cap turned
also, and the apparent velocity of the moving image was increased
(the velocity of the retinal image relative to the retina was increased).
The eye, a s it moved, made anappropriatecorrection, and this in turn
led to a still greater increase in the apparent velocity. A series of
these corrections accelerated the movement of the eye, and this acceleration continued a s long as the visual image of the object remained
within the field of vision. The subject's attempts to control their
system of pursuit and make it more "intelligent" in these conditions
likewise proved unsuccessful. Consequently, when the usual relationship between the movement of the eye and the displacement of the
retinal image is disturbed, 'the system of pursuit cannot perform its
function, although it tries to do so.
If a moving object is present in the field of vision, the observer
can start and stop pursuit of the object a t will. However, an observer
can never (at least, without special training) interferevoluntarily with
the actual process of pursuit and change its speed deliberately, making
it greater or less than the speed of the moving object.

2. MINIMAL AND MAXIMAL VELOCITIES O F PURSUIT

We next attempted to determine the range within which the system
of pursuit works if the object of perception movesat a uniform speed
along a horizontal straight line lying in the frontal plane. For this
purpose two series of experiments were carried out.
In most experiments, the movements of the visual test object were
recorded a t t h e same time as theeyemovements. A spot of light from
a slit source was divided intotwoparts s o that one part, a s a n illuminated point, moved over a screen a t which the subject looked, serving
a s the visual test object, while the other, in the form of a narrow
vertical band, fell on the horizontal slit of a photokymograph and was
used to record the movements of the object. From the beam of light
reflected from the mirror affixed to thecapon the eye, which also fell

CHAPTER VI

on the slit of the photokymograph, the eye movements were recorded.

Neither of the last two beams could be seen by the subject. Given the
appropriate distances, the experimenter could always calculate the
relationship between the recorded amplitudes of object movement and
eye movement. From the simultaneous records of the movements of
the eye and object, the experimenter could determine very accurately
the degree to which the movements of the eye and of the object corresponded under different experimental conditions.
In the first s e r i e s of experiments, the objects pursued by the
subjects moved with a verylowangularvelocity. The records obtained
showed that smooth pursuit begins when the speed of the object equals
the speed of the irregular drift of the eye (always present during
fixation).
The smooth pursuing movement develops a s a result of
change of the irregular drift into a regular drift, i.e., a drift with a
dominant direction. When the speed of the object is equal to one
minute of a r c per second, some degree of order can already be
observed against the background of the drift, and this may be taken a s
smooth pursuit. When the speed of the object reaches 5 minutes of
a r c per second, the smooth pursuit is now perfectly clear, although the
drift still causes marked distortion of this movement (Fig. 99). Not
until the speed exceeds 10-15 minutes of a r c per second is the drift
hardly noticeable in the records. Pursuit of low velocity (measured in

I;I

162

Fig. 99. Record of eye movements during pursuit of an abject moving from Left to right at a
speed of 5 minutes of arc per second. The record clearly shows the drift of the eye and the
small saccades.

.j

iI

ii

:\

9
$

,,,

I!

EYE MOVEMENTS WRING PERCEPTION OF MOVING OBJECTS

163

units of minutes of a r c per second) is usually accompanied by small

involuntary saccades.
The amplitude of these saccades and the
frequency with whichthey appear differ only slightly from the amplitude
and frequency of the saccades arising during fixation of a stationary
object.
In the second s e r i e s of experiments, the subjects pursued objects
moving a t high angular velocities. In this case, the ability t o pursue
was determined not only by the angular velocity with which the object
moved, but also by whether this object had been present all the time
in the subject's field of vision or had appeared unexpectedly and then
rapidly disappeared from the visual field.
A certain period of time is required for pursuit to develop; if the
moving object is present in the field ofvision for less than this period
(0.15 sec), pursuit is impossible. On the other hand, if the moving
object remains in the field of vision,theconditions for pursuit become
more favorable. The eye then has sufficient time to prepare for each
individual pursuit or saccade, coinciding in direction with the moving
object.
The saccadic movements. serving to change the points of fixation
last for hundredths of a second. During a saccade the eyes move in
accordance with a definite law, and this movement differs sharply from
the movement of pursuit. Nevertheless, saccades enable the details
even of very fast moving objects to be examined, if they follow the
direction in which the object is moving. In this case the object may
attain a velocity of 400-500 deg/sec. Somewhat smaller velocities of
movement of the object (350-400 deg/sec) may be accompanied by very
brief movements of pursuit.
During free examination of stationary objects, the duration of the
shortest, and a t the same time the commonest, fixations lies within
0.20 and 0.25 sec. This duration of fixation evidently corresponds to
satisfactory conditions of perception. There is therefore reason to
believe that satisfactory conditions for perception of moving objects
a r i s e if pursuit movements of a durationof 0.20-0.25 sec become possible. When a moving object is seen continuously, a s the experiments
show, such satisfactory conditions of perception a r e possible if the
speed of the object does not exceed 2OOdeg/sec. If the object appears
every time unexpectedly, such conditions a r i s e only a t speeds of
150-100 deg/sec.
Although we have considered eye movements only, it must be
remembered that under ordinary conditions of pursuit of moving
objects the observer's head also turns and that this essentially
facilitates his task.

CHAPTER Vl

EYE MOVEMENTS DURING PERCEPTION OF MOVING OBJECTS

165

shown by the fact that when the oscillations of the object a r e of high
frequency, the amplitude of the oscillations of the eye may be considerably less than the amplitude of oscillations of the object (Fig. 100).
This i s because before one movementcanbecompleted, the eye begins
to take part in a second movement, following the object. The attempt
of the eyes to pursue the object is often accompanied by corrective
saccades, a s a result of which the retinal image of the object falls on
the fovea.
It may be concluded from a study of these records that the eye
obtains the information essential for pursuit in two states-a state
of fixation and a state of pursuit. In the first case, the essential
information is given by evaluation of the angularvelocityof movement
of the object, and in the second by evaluation of the difference between
the angular velocities of the object and the moving eye. The information thus gained may be used in two ways. Sometimes the eye changes

Fig. 100. Sample simultaneous records of movements of the object and eye. Smooth linesmovements of the object. Movements of the eyes are characterized by corrective saccades:
a ) unsuccessful attempt at pursuing; b) shift in phase clearly seen on the record (the result
af delay of the eye), amplitude of the ascillarions of the eye diminished by comparison with
the movement of the object; c) more accurate pursuit.

3 . EYE MOVEMENTS DURING PURSUIT IN COMPLEX

CONDITIONS
In many experiments the object moved with an acceleration which
was a complex function of time. Here,the records clearly showed how
the system of pursuit works and what its possibilities a r e under
conditions of varying complexity.
The records shown in Fig. 100 will perhaps give some idea of the
system of pursuit under complex conditions.
It will be noted that during the attempt to pursue the object, the
observer's eyes attempt to repeat the movement of the object, but do
s o only after a slight delay (0.1-0.2 sec). This delay can easily be
detected from the shifts in phase on some of the records in Fig. 100,
which reflect attempts made by the eye tofollow the oscillating movements of the object.
When trying to follow the movements of the
object, the system of pursuit continually introduces corrections, a s

__H_

Fig. 101. Movement of both eyes recorded during pursuit of a pendulum oscillating in the
frontal plane.

CHAPTER VI

EYE MOVEMENTS DURING PERCEPTION OF MOVING OBJECTS

Fig. 104. Record of movements of the eye during pursuit of an abject moving at uniform
velocity. and appearing unexpectedly in the right peripheral part of the subject's field of
Vision. Pursuit at the necessary speed begins instantaneously following a saccade.

Fig. 102. Eye movements recorded during pursuit of an object taking part in a complex
movement. It i s clear from the records that the sharp change in the velacity and direction
Of movement of the eye is coordinated with the correcting saccade.

1,
!'

I'

Fig. 103. Records of movements of an object (thick lines) and the eye, showing clearly the
delay in the movements of the eye initiated during pursuit.

its velocity smoothly (in magnitude and direction), followingthe movement of the object (Fig. 101). Sometimes, however, it begins to move
a t once at a certain velocity or sharply changes its velocity (in
magnitude and direction). This second case is always preceded by a
saccade, after which a movement follows immediately (withinmilliseconds) at a velocity equal to o r close to the velocity of the object. In
other words, the sharp change invelocityis always timed to take place
during the correcting saccade (Fig. 102).
If a moving object appears suddenly in the foveal region of the
subject's field of vision, naturally pursuit cannotbegin instantaneously.
For pursuit to begin, some time is required for the system of pursuit
t o become operative; the records show this time to be approximately
0.1-0.2 sec, usually 0.15-0.17 sec (Fig. 103). If the mov&g object
appears suddenly and the subject sees it with the peripheral part of the
retina, preparation for pursuit (requiring 0.1-0.2 sec) takes place on
the basis of data obtained from the periphery, and is accompanied by
simultaneous preparation for a saccade to the object. It is most
important that the saccade always take place after the preparation for
pursuit, i.e., when the system of pursuit "knowsn approximately with
what angular velocity it must carry out its pursuing function. For this
reason, after the saccade has taken place and the image of the object
lies in the central part of the retina, pursuit begins practically
instantaneously, within milliseconds (Fig. 104).
Records such a s those in Figs. 102and 104 suggest that the system
of pursuit is prepared for its pursuing function o r for a sharp change
in its velocity and direction not after but before the saccade, i.e.,
during the fixation preceding the saccade, or during the pursuit
preceding the saccade. Experiments show that participation of the eyes
in the process of pursuit begins not a t the end of the saccade, but a t
its beginning. In other words, the instructions to begin the saccade
and the new act of pursuit take place simultaneously or almost simultaneously. Since the velocity of the saccade is large (and its duration
small), displacement of the eyes resulting from premature pursuit

"

CHAPTER Vl

EYE MOVEMENTS DURING PERCEPTION OF MOVING OBJECTS

169

may be taken a s negligible, but the positive effect is substantial, and

there is no pause in the work of the eyes between the end of the saccade
and the beginning of pursuit. If the eyes do not begin their pursuing
function until after a saccade, a s a result of inertia they would not be
able to begin pursuit quite s o instantaneously (within milliseconds).
If a moving object pursued by anobserverdisappears unexpectedly and
instantaneously, the eyes cease their pursuit,slowdowngradually,and
eventually stop completely (after approximately 0.1 second).

4. PURSUIT ACCOMPANIED BY CONVERGENCE AND

DIVERGENCE OF THE OPTICAL AXES
When an object of perception, moving inspace,moves nearer to or
further from the observer, pursuit is accompanied by convergence or
divergence of the optical axes and is the most complex case of eye
movement.
I mentioned earlier that the pursuing movement of the eyes and
convergence or divergence of the optical axes, althoughdifferent types
of movements, take place in such close cooperationduring pursuit that
they cannot be differentiated on records. This merger is evidently
explained by the fact that convergence and divergence have much in
common with pursuit and may, insomecases,completely take over its
function. For example, if a moving object shifts along the axis of the

Fig. 105. Records of eye movements during pursuit of a pendulum oscillatiy in the sagittal
plane passing through the right eye (cop figure) and between the eyes (bottom figure).

I
I

1
,IIi

Fig. 106. Records of eye movements during pursuit of objects accompanied by complex
movements. a) OscilIations; b) irregularmovement in space; c) during perception of stereoscopic picture.

cyclopic eye, pursuit is reduced to movements of convergence o r divergence and correcting saccades.
It may thus be considered that the convergence or divergence of the
optical axes is a .movement during which the point of fixation slides
along the axis of the cyclopic eye. The intersection of the optical axes
may be displaced (as a result of convergence or divergence) along the
axis of the cyclopic eye relatively quickly and suddenly during a change
of stationary or moving points of fixation, or, on the other hand,
steadily and'slowly during pursuit of a moving object. The axis of the
cyclopic eye may itself be displaced in space,making steady rotations
during pursuit and gudden movement during saccades.
Sample records' of eye movements when the object of perception
performs complex movements in spaceareshowninFigs. 105 and 106.
These records demonstrate to some extent the possibilities open to the
system of pursuit and the system of the eye muscles.

CHAPTER VI

CONCLUSIONS
The system of pursuit of the eyes cannot be put into action voluntarily, in the absence of an object movingin the field of vision. Under
normai conditions, no smooth pursuit movements a r e possible without
the presence of an object moving inthe field of vision. Smooth pursuit
can begin when objects move a t speeds equal to those of the drift of
the eye arising during fixation. Satisfactory conditions of perception
a r e possible when the speed of the object does not exceed 100-200
deg/sec.
The eye must obtain the necessary information for pursuit while
in the state of fixation or in the state of steady pursuit. The information obtained can be used in two ways. In some cases. the eye
changes its speed (in magnitude and direction) steadily, and follows
the movement of the object. In other cases, the eye begins to move a t
once a t a definite velocity o r changes its speed suddenly (in magnitude
and direction). In these circumstances, the instantaneous beginning
o r sudden change in velocity is coordinatedwitha saccade of the eyes.

Chapter VII

EYE MOVEMENTS DURING PERCEPTION

OF COMPLEX OBJECTS
In this chapter we shall try to determine how the human eye
examines complex objects and what principles govern this process.
For example, it may seen to some people that when we examine a n
object we must trace its outlines with our eye and, by analogy with
tactile sensation, 'palpate" the object. Others may consider that, when
looking a t a picture, we scan the whole of its surface more or less
uniformly with our eyes.
The problem of eye movements during perception of pictures
during reading, during the perception of optical illusions, and during
comparison of distances is investigated below. Toround out our study
of the role of eye movements, we shall also consider the question of
perception during which the subject cannot voluntarily use his eye
movements or choose points of fixation.

1. EYE MOVEMENTS DURING PERCEPTION OF COMPLEX

OBJECTS
Before reading the text, the reader should quickly look over the
records of eye movements shown in Figs. 107-124. Study of such
records a s these suggests first of all that, when examining complex
objects, the human eye fixates mainly on certain elements of these
objects. Any picture (unless it is auniform background or a repetitive
mosaic) contains different elements; the eye r e s t s much longer on
some of these than on others, while someelements may receive lkttle
o r no attention. What distinguishes the elements particularly attracting
the observer's attention, and what a r e the characteristic features of
those elements which do not draw his attention ?
Analysis of the eye-movement records shows that the elements
attracting attention contain, in the observer's opinion, may contain,

CHAPTER VII

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

108. Seven records of eye

freely wirh both eyes. The recor

interval between records was 1 or 2

174

CHAPTER VIl

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

175

Fig. 110. Record of the eye movements for 3 minutes during free examination, divided into
seven consecutive parts. The duration of each part is about 25 seconds.

Fig. 109. Seven records of eye movements by the same subject. Each record lasted 3
minutes. The subject examined the reproduction with both eyes. 1) Free examination of
the picture. Before the subsequent recordingsessions, the subject was asked to: 2) estimate
the material ~Pcumstances of 'the family in the picture; 3) give the ages of the people;
4) surmise what the family had been doing before the arrival of the 'unexpected visitor":
5) remember the clothes worn by the people; 6) remember the position of the people and
objects in the room; 7) estimate how long the "unexpected visitor' had been away from the
family.

information useful and essential for perception. Elements on which

the eye does not fixate, either in fact or in the observer's opinion, do
not contain such information.
Let us now try to explain and prove this statement. F i r s t we will
note that special attention o r indifference to the elements of a picture
is in no way due to the number of details composing the elements.

CHAPTER VII

Fig. 111 i s a record of the eye movements f o r 35 seconds during free examination. The
recording i s divided inro seven consecutiveparts of 5 seconds each. The number of fixations
in each part of the record is: 1. 18: 2, 16: 3, 18: 4. 14: 5, 17; 6, 13; and 7, 15. In Fig. 112
each point of fixation of the eye movements illustrated in Fig. 111 i s covered with a black
circle. The size of the circles corresponds to the size of the central fovea of the subject's
eye (1.34. This figure shows which elementsof the reproduction were fixated by the central
fovea and in what order during examination of the picture f a r 35 seconds.

EYE MOVEMENTS WRING PERCEPTION OF COMPLEX OBJECTS

CHAPTrR Vll

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

For example, the figure of the hunter in Fig. 118 particularly attracts
the observer's attention although it is nomore detailed than many other
elements of the picture. Although the bears in Fig. 113 have fewer
details than the branches of the trees, they attract most attention. It
is clear from the records of Fig. Illithat the eye fixates mainly along
the line of the horizon and on the trunk of the birch trees, although in
this picture the details composing the grass and leaves a r e especially

Fig 114. Photograph df a girl's face. Record

o the eye movements during free examination
of the photograph withbotheyesfor oneminute.

Fig. 113. Reproduction of 1.1. SNshkin's p i c m e .Morningin the Pine Forest." The records
of the eye movements were made during free examination of the picture with both eyes for
2 and 30 minutes.

CHAPTER VII

Fu. 115. 1 he flrst p q e of rho m a g z m e "Ogonek.. No. 23 (1959): 'Girl from the \'olg.s"
(photoc~pliby S. FrldlyanJ). Record oi the r y e movemcnrs during free exarnlnarion of the
photograph wlrh both eyes for 3 mtnutas.

EYE MOP'EMENTS DURING PERCEPTION O F COMPLEX ORJECTS

Fig. 116. H ead ai the Egyptian QueenNefertiri(I6thcenturyB.C.).

Record of eye movements
during free examination af a photograph of the sculptured head with both eyes ior 2 minutes.

CHAPTER VII

EYE MOVEMENTS WRING PERCEPTION OF COMPLEX OBJECTS

For many observers one color is more pleasing than another, and
sometimes the expression 'my favorite colorn is heard. Most of the
reproductions used in my experiments for free examination by the
subjects were colored. Sometimes both colored and black-and-white
reproductions of the same size takenfrom the same picture were used.
However, in,,.np_,c.a_see.didthe corresponding records reveal any appreciahle!:influence
. .-.-- of
.. - .
the distribution of the points of fixation.
It must be concluded that if the color of an element has no special
significance and is irrelevant to the meaning of the picture under
examination, it will have no effect on the character of the eye movements. The results describedin Chapter 11 show that an important role
in the process of vision is played by the outlines of objects perceived.
The question arises: to what extent does this importance of outline

cola-n

Fig. 117. Reproduction of 1.l.Leviran'spic~ure~TheBirchWood.. Record of eye movements

during free examination of the reproduction with both eyes for 10 minutes.

numerous. Any record of eye movements shows that, per se, the
number of details contained in an element of the picture does not
determine the degree of attention' attracted to this element. This
is easily understandable, for in any picture, the observer can obtain
essential and useful information by glancing a t some details, while
others tell him nothing new or useful.
Similarly, the brightest or darkest elements of a picture are not
necessarily those which attract the eye (if the brightness of these
elements is considered alone). All records of the eye movements
show that neither bright nor dark elements of a picture attract the
observer's attention unless they give essential and usefulinformation.
For example, in the picture in Fig. 113the eye rests most on the dark
brown figures of the bears; in the picture in Fig. 117 the observer's
attention is attracted by the white trunks of the birch trees; in Fig.
118, by the figureofthehunterwhichis almost completely merged with
the background; and so on.

Fig. 118. Reproduction of 1.1. Shishktn's picture "In the Forest.' Record of the movements
of one eye during free examinarion of the p i c w e with both eyes for 10 minutes.

CHAPTER VII

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

3

Fig. 119. Drawlng by V.Surkov. Recordof the movements of one eye durlng free evamlnatlon
of picme w ~ t hboth eyes for 3 mmutes.

Fig. 120. Drawing by V.A. Vatagin. Record of the movements of one eye during free
examination Of the picture with both eyes for 2 minutes.

CHAPTER VII

Fig. 121. V.A. Vatagm's sculpture 'Gorilla.' Record of the movements of one eye during
free examination of a photograph of the sculpture by bath eyes for one minute.

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

Fig. 122. G.L. Peuushavich's sculpture 'My Child.' Record of the movements of one
eye during free examination of a photograph of the sculpture w ~ t hboth eyes for 2 minutes.

CHAPTER Vll

i
4

Fig. 123. Photograph ofthefriezefromthe

Pergamon altar. Head of Clitus. Record
of the movements of one eye during free
examination of the photograph with hoth
e y e s for one minute.

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

Fig. 124. "Ogonek: No. 21 (1959): photograph of agirl by

S. Orlina. Record of themovements of one eye during free
examination of the photograph with both eyes for one
minute.

.j

190

I,
!i

CHAPTER VII

influence the eye movements and the distribution of the points of

fixation?
Analysis of Figs. 107-124 shows that the 0u$.@~e_s_th_em_s_e1w.have
no effect on the character of the eye movements. In the movements
of the eye we have no analogy with the movements of the hand of a blind
person, tracing the outlines and contours. Outlines and contours a r e
important f o r the appearance of the visual image, but when the image
has appeared and is seen constinously, the observer has no need to
concern himself specially with borders and contours. Borders and
contours a r e only elements from which, together with other no less
important elements, our perception is composed and the object recognized. Clearly the outlines of an object will attract an observer's
attention if the actual shape of the outline includes important and
essential information. F o r example, when examining the sculpture of
Queen Nefertiti's head (Fig. 116), the observer directs nearly all his
attention to the profile o f the head, to the outline of the sculpture. It
is easy to s e e that the main features of the sculpture a r e concentrated
in these places, thus enabling the observer to form a complete representation of the head. On the other hand, the record in Fig. 119
which shows the eye movements during f r e e examination of a picture
that is purely in silhouette is completely indistinguishable from the
record of ordiniry picwres.
All the records given in this chapter, a s well a s others not mentioned here, show conclusively thatthe chara~ter_of&e,ye movements
....,.
is either completely independent of or only very slightly dependent on
the material of the picture
and how it was made, provided that it is
. . ~
Examination of ~ & ~ . , . . s i t u a t e din a room or
flat
nearly
examination of a sculpturedheadinwhicha
. .. . ..
change is made from points
of fixation situated a t different distances from the observer is accompanied by convergen-. and .divergens of the optical axes, and
differs from the perception of flat objects only. in this new type of
movement (i.e., convergence of the eyes).
Records of eye movements show that the 0 b S e ~ e r ' S&Eenil,onis
usually held only by certain.elements of the picture. As already noted,
the study of these elements shows that they give information allowing
the meaning of the picture to be obtained. Eye movements
, .
reflect
.-..-.-.
the
human
.
thought
.
. . .--..p
.,-....r
o
c
e
s
s
e
s
s
o
the
observer's
thought
may
be
followed
?.
,.
t o some extent from records of eye movements (the thought accompanying the examination of the particular object). It is easy to
determine from these r e c o r d s which elements attract the observer's
eye (and, consequently, his thought), in what order, and how often.

or

,)

iI
'1

43

1
:I
I

.Y

6
.s

.a
.j

'-&

-~

#8

EYE MOVEMENTS WRING PERCEPTION OF COMPLEX OBJEC7S

If we look a t the records in Figs. 107 and 108, made during free
examination of Repin's picture "An Unexpected Visitor," we can see
that in all 14 records the faces of the people shown in the picture
attract the observers' attention much more than the figures, and the
figures more than the objects in the room,and s o on. Moreover, even
the faces of the different people attract attention unequally, depending
on the place they occupy in the theme of the picture.
When looking at a human face, an observer usually pays most
attention to the eyes, the lips, and the nose. The other parts of the
face a r e given much more cursory consideration. Looking at the
record in Fig. 114, we s e e that in this portrait attention is drawn
almost exclusively to the girl's beautiful expressive eyes. The
observer pays much less attention to the lips and nose. In the photograph of the smiling girl in Fig. 115, the lips attract far more
attention; the reason f o r this will be clear to the reader when he looks
a t the photograph. It is curious that, when examining the picture of
the lion's head (Fig. 120) and the sculpture of a gorilla (Fig. 121),
most of the points of fixation of attention a r e found on the eyes, nose,
and mouth of the animals.
The human eyes and lips '(and theeyes and mouth of an animal) a r e
the most mobile and expressive elements of the face. The eyes and lips
can tell an observer the mood of a person and his attitude towards the
observer, the steps he may take next moment, and so on. It is therefore
absolutely natural and understandable that the eyes and lips attract
the attention more than any other part of the human face.
During the examination of the photograph of a sculpture (Fig. 122)
tl+e observer pays special attention to the face of the sleeping child
(the closed eyes, lips, nose), the mother's smiling face (the lips
attract special attention), and the mother's hand holding the child's
It is easy to see that the child's face reflects blissful sleep
iiead.
and the mother's f a c e and her tender smile express the happiness
of motherhood, while the mother's hand expresses: care and affection.
In this case, then, the observer fixatedonthe elements which revealed
t o him the main theme of this picture.
The, observer's attention is frequently drawn to elements which
do not give important information but which,inhis opinion, may do so.
Often an observer will focus his attentiononelements that a r e unusual
in the particular circumstances, unfamiliar, incomprehensible, andso
on. For example, many of the points of fixation in the recordings in
Figs. 117 and 118 fall on the horizonline. Evidently the observer was
examining the horizon in the hope of finding something important. In

CHAPTER VII

192

Fig. 118 he was interested in the gap between the trees. In Fig. 124
he spent considerable time examining the amusing tuft of hair on the
child's head, and s o on. A careful examination of the two series of
records in Figs. 107 and 108 showed, that despite the great similarity
among all 14 records, the similarity is greater in the seven records
of Fig. 108, which reflect the process of perception and thought of a
single observer, than in the seven records of Fig. 107, which were
obtained from seven different ohservers. It may be concluded that
individual observers differ in the way they think and, therefore, differ
also to some extent in the way they look a t things.
All the observers whose eye movements a r e shown in Figs. 107
and 108 were well educated people and were well acquainted with
Repin's picture. This evidently accounts for the generally considerable
similarity between all the records.
Perhaps the records would
have been much less similar if the observers had differed more in
cultural level and education. Undoubtedly, observers'familiar both with
the picture and the epoch represented in it would examine the picture
differently from people seeing it for the first time and unfamiliar with
the epoch it represents.
In just the same way it would be natural to assume that acomplex
object of perception understood by a physicist but unfamiliar to a
biologist (or vice versa) will be examined qnitedifferently by a physicist and a biologist.
Depending on the task in which apersonis engaged, i.e., depending
on the character of the information whichhemust Obtain, the distribution of the points of fixation on an object will vary correspondingly.
because different' items of informationare usually localized in different
parts of an object. This is confirmed by Fig. 109. This figure shows
that, depending on the task facing the subject, the eye movements
varied.
For example, in response to .the instruction "estimate the
material circumstances of the family shown in the picture," the obs e r v e r paid particular attention to the women's clothing and the
furniture (the armchair, stool, tablechloth, and s o on). In response to
the instructibn give the ages of the people shown in the picture," all
attention was concentrated on their faces. In response to the instruction "surmise what the family was doing before the arrival of the
'unexpected visitor,'" the observer directed his attention particularly
to the objects arranged on the table, thegirl's and the woman's hands,
and to the music. After the instruction "remember the clothes worn
by the people in the picture," their clothing was examined. The
instruction 'remember the position of the people and objects in the

'

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

193

room,' caused the observer to examine the whole room and all the
objects. His attention was even drawn to the chair leg shown in the
left part of the picture which he had hitherto not observed. Finally,
the instruction 'estimate how long the 'unexpected visitor' had been
away from the family ," caused the observer to make particularly
intensive movements of the eyes between the faces of the children and
the face of the person entering the room. In this case he was undoubtedly trying to find the answer by studying the expressions on the
faces and trying to determine whether the children recognized the
visitor or not.
Records of the eye movements after an instruction a r e interesting
because they help in the analysis of the significance of eye movements
during the free examination of a picture; they show clearly that the
importance of the elements giving information is determined by the
problem facing the observer, and that this importance may vary within
extremely wide limits.
It is clear from Figs. 107 and 108 that the observer's attention
rested on the faces and figures depending on their significance for the
picture a s a whole. It should benoted that the scale of the figures and
objects and their position in the picture-in other words, everything
which we call the composition of thepicture-alsohas definite importance. The figure of the woman, being the largest and being centrally
situated, attracts more attention than any of the other figures. Hence
composition is the means whereby the artist to some extent may
compel the viewer to perceive what is portrayed in the picture.
If the eye movements a r e recorded for several minutes during
perception of an object, the record obtained will clearly show that, .
when changing its points of fixation, the observer's eye repeatedly
returns to the same elements of the picture. Additional time spent
on perception is not used to examine the secondary elements, but to
reexamine the most important elements. The impression is created
that the perception of a picture is usually composed of a series of
"cycles," each of which hasmuchincommon. For example, it is clear
from Fig. 114 that the examination of the portrait in fact was reduced
to the repeated alternate fixation on f i r s t one and then the other of the
girl's eyes. The same can be seen in Fig. 115. The records in Fig.
113, one of whichcontinuedfor2minutes,and the other for 10 minutes,
generally speaking, differ only very slightly. The distribution of the
points of fixation and the character of the eye movements were almost
identical. It should be noted here that, although the records were
obtained from the same observer, the interval between the experi-

194

CHAPTER VII

ments was a whole month. The reiterativemovements of the eyes a r e

seen particularly well in Fig. 116. Inthe 2 minutes, the observer cast
his eyes several times around the profile of the sculptured head.
Analysis of these complex records of eye movements shows that
the duration of a cycle during which the observer's eye can cover the
whole picture amounts sometimes to several seconds, sometimes to
several tens of seconds. The more complex an object and the more
associations it arouses, the longer the duration of this cycle.
A reproduction of Repin's picture "An Unexpected Visitor" is
shown in Fig. 110 with a 3-minute record of the eye movements during
f r e e examination of the picture byanobserver. The record a s a whole
has been divided into seven consecutive parts, the duration of each
part being 25 s e c (during the experiment the sheet of photosensitive
paper on which the record was made was changed very quickly every
25 sec). Analysis of these separate records shows that each of them,
roughly speaking,, corresponds to a cycle during which the eye stops
and examines the most important elements of the picture. In each
part, the observer's eye examines the faces of all the people shown in
the picture.
In other words, during a 3-minute examination of the
.picture, the observer directed his attention a t least seven different
times to each face.
The records in Fig. 109 show that this cyclical pattern in the
examination of pictures is dependent not only on what is shown on the
picture, but also on the problem facing the observer and the information
that he hopes to gain from the picture. F o r example, the record made
following the instruction "estimate the ages of the people shown in the
picture' shows relatively few reiterated movements.
After the
instruction "estimate how long 'the unexpected visitor has been away
from the family." there were several times more reiterated movements.
This cyclical pattern followed in the examination of objects
evidently reflects some special features of our perceptionand thought.
In this connection I shall donomore than mention the purely conjectural
hypotheses which have been put forward.
In order t o observe the special features of the eye movements in
the initial stage of examination of Repin's familiar picture "An Unexpected Visitor," an interrupted 35-second record was made of the
eye movements of an observer freely examining the picture. The
record of the eye movements a s a whole was divided into seven
consecutive records, each 5 seconds long (the sheets of photosensitive
paper were again changed quickly, every Sseconds in this experiment).
The results areshown in Figs. 111and 112. In Fig. 112 all the points

EYE MOVEMENTS WRING PERCEPTION OF COMPLEX OBJECTS

195

of fixation of the observer's attention were superposed on the picture

and covered with black circles. The circles corresponded in size to
the angular magnitude of the central foveaof the observer's eye (1.3").
In other words, the circles may be regarded a s the projection of the
central fovea of the observer's eye onto the picture in the conditions
of this particular experiment.
Analysis of the experimental results is facilitated by Fig. 112.
This figure shows thatduringthe first 5seconds the observer examined
the man entering the room. The observer's attention was focused
mainly on the man's face and theupperpart of his body. By the end of
this period, the observer had examined the visitor's boots and the faces
of the women standing by the open door. Altogether, during this first
5-second period, the observer changed points of fixation 18 times.
Apparently 16 of the fixations enabled the observer to gain a t least a
general impression of the identity of the "unexpected" person. Two
fixations only were devoted to the study of the faces of the women
standing by the open door (secondary characters). In the second
5-second period, the points of fixation were mainly situated along the
line of sight of the man entering the room (directed a t the face of the
elderly woman) and along the.line of sight of the woman (directed a t the
man's face).
During this period, the observer was apparently
investigating the relationship between the two main characters in the
picture. It is possible that during this period the observer also to some
extent formed a clear idea of the main theme of the picture. During
this second 5-second period, the observer changed points of fixation
16 times. The third 5-second period was mainly devoted to the study
of the elderly woman (her face and figure) and of the woman sitting by
'the piano. Characteristically, this study, like the preceding periods.
was accompanied by fixation on the face of the man entering the room.
.The third period included 18 fixations. During the fourth 5-second
period, the observer continued to study the maincharacters and began
to examine the faces of the children sitting by the table. The fourth
5-second period included 14 fixations.
During the fifth 5-second
period, all the active characters were examined, but the main attention
was given to the children. The fifth period included 1 7 fixations.
Apparently, the fifth period concluded the first cycle of examination
of the picture; the eye movements in the sixth and seventh periods
largely repeated those made in the previous periods.
The results of t h e experiment illustrated in Figs. I l l and 112
suggest that the observer's eye examined all the more important
elements of the picture during the first 25 seconds, and the observer
himself evidently obtained a general idea of the theme of the picture.

196

CHAPTER VII

Characteristically, during each 5-second period the observer directed

his attention at least once to the man entering the room. During perception of the picture, the observer's thought was constantly directed
to the "unexpected visitor," with which figure all the remaining
elements of the picture were linked and compared. In the course of
25 seconds, the observer changed points of fixation 83 times (an
average of more than three points offixationper second). The number
of points of fixation, and their distribution in space and time give the
reader some idea of the perception of so complex an object a s the
picture "An Unexpected Visitor."
In conclusion, I must s t r e s s once again that the distribution of the
points of fixation on an object, the order in which the observer's
attention moves from one point of fixation to another, the duration of
the fixations, the distinctive cyclic pattern of examination, and s o on
a r e determined by the nature of the object and the problem facing the
observer a t the moment of perception.
The material in this section clearly can be regarded a s merely
the beginning of the study of the perception of complex objects by
recording the eye movements.
The many pictures included in this section should be regarded by
the interested reader not merely as an illustration to the written text,
but also a s material for study. I hope that some of these pictures
will be used by other authors. On many records taken from these
pictures we can s e e that during perception many of the elements of
the picture a r e not perceived by foveal vision. This is illustrated
particularly clearly in Fig. 112. Foveal vision is reserved mainly
f o r those elements containing essential information needed by the
observer during perception.
In this connection, we cannot help thinking how important and
biologically desirable is this heterogeneous structure of the retina,
particularly, the fact that a fovea is present. By means of the fovea
centralis, man sees many details only aroundthe point of fixation, i.e.,
around a point which, a s a rule, provides essential information. The
lower resolving power of the eye periphery is useful because it
enables l e s s essential information to be obtained and facilitates the
differentiation between the useful and useless information.

2. EYE MOVEMENTS IN READING

The first studies of eye movementsinreadingwere evidently made
by Java1 (1879). His results were obtained by visual observation of the

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

197

subject's eyes. The f i r s t photographic record of eye movements

during reading was obtained by Dodge in 1899 (see Taylor, 1957).
Several other authors have subsequently studied this problem (Buswell,
1937; Gilbert, 1953; Taylor, 1957). Since I have not personally made
any detailed study of eye movements during reading, the data given
in this section a r e mainly compiled from other sources.
Some idea of the eye movements during reading is given by the
record in Fig. 125, made by means of the Pi cap. The subject in this

I.
mu.

"0;

crwx

P..*c4p...r*
no*..c..

npx..ol

A*

no.*..

He

6rewr "a.#moE.

me*..*
**r r m .

*enunn*i,
r**i.

c o . . r . ~ ~ a rreA**r.?

*-

Fig. 125. Record of the eye movements of a subject reading a Shahzespeare sanner. Record
on sratimary photosensitwe paper (a) and on moving phototape of a photokymograph (b).

'$:!
{ ',-+ >:.,

'?,;,

.-

^.$

CHAPTER VII

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

199

i,
$ i

\.-x

experiment was a student with average reading ability. One of the

records was made on stationary photographic paper, the other by means
of a photokymograph on moving oscillographic paper.
Examination of the second record shows clearly that during reading
the duration of the fixations usually lies between 0.2 and 0.4 sec, with
a mean value of 0.3 sec for this particular subject. During reading the
character of the eye movements remains the s a m e a s during examination of other stationary objects if the natural regularity and succession
of the eye movement along the lines of the text a r e now counted. The
same record also shows that the reading of each line ends with a prolonged fixation (or two fixations), almost a whole second in duration.
These prolonged fixations correspond to the long process of analysis
of what has been read, in this case the line of the poem. Prolonged
fixations a r e found during the reading of any text; the more complicated
the text and the more numerous the thoughts, associations, and ideas
evoked by the word o r line read, the longer the fixation. These results
indicate that the required resolving power of a printed text is maintained entirely by the foveal and parafoveal region of the retina.
A detailed study of the eye movements duringreading, wirh a large
number of students used a s subjects, yielded results which, when
presented a s a table, could be used to evaluate methods of teaching
pupils in various grades to read.
Records of the eye movements of 5000 students were used by
Taylor to compile this table (Table 2). The table gives the mean
values of the elements constituting the art of reading, which can be
studied by suitable experimental techniques.
The mean range of
recognition per fixation by the children studyingin the first six grades
' was less than one word (if a word is regarded a s consisting of ten'
printed symbols). The mean range of recognition per fixation by the
college students was 1.33 words. It was found that students trained by
means of special tachistoscopes have an increased range of recognition
per fixation, although no significant changes were observed in reading
speed. As the students continued to develop, the number of fixations
(per 100 words of text) fell to one-third the original number, while the
number of retraces during reading fell to one-fifth The duration of
the fixation showed little change and in general remained the same a s
during free examination of any stationary object. The rate of understanding (the number of recognized words per minute) increased
fourfold.
Investigations have shown that a person who reads aloud badly
usually looks a t a word when pronouncing it and, in doing so, makes
two or more fixational pauses. Suchaperson does not run ahead of the
word being pronounced with his eyes. A person who reads aloud

CHAPTER VII
fluently does not dwell on the word being pronounced, sometimes running along the line with his eyes for several words in advance.
Although some people can be taught to read a t a speed of 1000 or
more words per minute, the average reader cannot attain such a
result even with training. Usually people who can be taught to read
very quickly readvery quickly before training. There is a considerable
individual scatter in relation to speed of reading. Records have shown
that teachers usually read a t a speed of 350-500 words per minute,
i.e., that most of them do not read better than students in Grade 9.
Most adults read a t a speed less than 300 words per minute. Many
students read a t the same speed a s grade school pupils.
In 1935, Taylor examined a student who was a very fast reader.
He read a t the speed of between 600 and 2200 words per minute.
When reading a t speeds of between 1000 and 1500 words per minute,
he could recite word for word extracts from high-school textbooks.
When reading faster than 2000 words per minute, the boy guessed most
of the text just a s a high-school pupil does when reading a t the rate of
500-600 words per minute. The records of the eye movements of this
boy during reading were the most unusual of any Taylor observed during
his investigation of about 10,000 subjects. Inthe course of one fixation
the boy perceived several words o r even a whole phrase. Taylor
emphasizes that this young reader was very gifted. At the age of 20
years he obtained a doctoraldegreeandsoonafter became a teacher at
one of the leading universities.
Reading speed is determined not by the properties of the muscular
apparatus of the eyes, but by the capacity and the special features of
the individual's higher nervous activity. For this reason, when teaching
pupils to read, attention should be concentrated on the accuracy with
which the student reads and not onincreasingthe speed of his reading.
The necessary speed will appear without special training, and usually
a t the end of his studies the pupil will be able to grasp the material in
accordance with his capabilities.

3. ROLE O F EYE MOVEMENTS IN ASSESSING SPATIAL

RELATIONSHIPS
Let us now consider a series of experiments showing the role of
eye movements when the observer needs to a s s e s s proportions and
to compare lengths, a r e a s , and angles.
In one experiment in which a bright after-image, rectangular in
shape, was created for a subject by means of a flashing light in a
totally dark room, the subject was told to determine the ratio between

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

201

Fig. 126. Drawings for solving certain visual problems mentioned in the text.

the sides of this rectangle.

Although this task seems simple at
first glance, it was really quite complicated becausethe subject could
not use his eye movements for comparing the size of the rectangle.
The subject's records show that the attempt to discover the ratio
between the sides was accompanied by rotation of the eyes, the head,
and even of the trunk (sometimes, when turning the trunk, the subject
turned the chair on which he sat). After these attempts had proved
unsuccessful, the subject decided to attack the problem by means of
complicated calculations. This second method of solution appeared
much more complicated than the method using eye movements.
For some idea of the two ways of solving problems analogous to
thatjust mentioned, look at Fig. 126. On the left, a spiral and a short
horizontal straight line a r e seen. The problem is to determine the
length of the spiral using the straight line a s a unit of measurement.
This problem is relatively easily solved with the use of eye movements
(applying the unit of measurement along the spiral). If, on the other
hand, the center of the spiral is fixated, continuously and the subject
meanwhile attempts to determine its length visually, the problem"
becomes very complex, although during fixation the whole spiral and
the horizontal line a r e clearly and simultaneously visible to the observer. On the right in Fig. 126 is a drawing of a rectangle and a
small square by means of which the reader is ipstructed to measure
the area of the rectangle. Again, it is clear that the problem becomes
very complex if one tries to solve it under conditions of continuous
fixation; if eye movements can b e used, however, the problem is
relatively simple. Much may be learned fromproblems such a s these
(problems in two and three dimensions).
A curious experiment was carried out by means of the P4 cap (see
the description of the cap). The cap was affixed to one eye, and the
second eye was covered with a bandage. The reader will remember
that the perception of surrounding objects through such a cap placed
the observer in a situation in which he saw objects perfectly clearly
but could not use eye movements a t will. For example, change by

CHAPTER Vll

practically stationary. At the same time the brighmess of the flash was
s o great that a clear and prolonged after-image of the drawing appeared.
When perceiving the drawing, the subject could judge the
presence or absence of a corresponding illusion.
I pointed out earlier that evenlarge saccades a r e often involuntary.
In many cases, saccades, sometimes even a group o r a series of
saccades apparently voluntary in nature, a r e not entirely under the

Fig. 129. Drawing and record af eye movemenrs of a subject told t o .follow with your eye
the horizontal segments, then the sloping line, and continue the direction of the sloping line
heyond the horizontal line:

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

205

observer's control. This "disobedience" is particularly obvious in

records of eye movements accompanying the perception of optical
illusions. A well known illusionarisinginthe assessment of distances
between the edges of objects and a record of the eye movements
accompanying the assessment of the drawing a r e shown in Fig. 127.
It i s clear from this figure that saccades of different amplitude correspond to objectively equal segments. Here the visual evaluation of
length and the amplitude of the saccades a r e in mutual agreement.
Many experiments such as that illustratedinFig. 127 show that the
subject's subjective evaluation may always be judged from the eye
movements made during comparison of distances. Subjective evaluation of distance in many cases depends on the shape and position of the
objects the distance between which is being assessed. At first glance
this may appear strange, but, for example, objectively equal distances
between two vertical lines and two ends of a horizontal line a r e assessed differently. The length of a horizontal line appears much less to
some observers than the distance between the vertical lines, and this
difference is reflected in the eye-movement records.
Since illusions similar to those illustrated in Fig. 127 continue to
occur both in experiments with the P4capand with a stationary retinal
image, when the drawing is illuminated by a flashing light, we may
conclude that eye movement has no appreciable effect on the presence
of many illusions. Onthe other hand,as the records show, the presence
of illusions appreciably influences the amplitude of the saccades
accompanying the evaluation of distances.
Let us consideranother example. The familiar illusion of displaced
segments of a straight line is shown inFig. 128. The record shown in
Fig. 129 illustrates that during the visual extension of a straight line,
the observer slightly changes its direction, increasing the angle between

Fig. 130. Illusion of perspective.

CHAPTER VII

Fig. 131. The deceptive spiral.

the prolonged straight lineand the intersectingline. As in the previous

case, this and similar illusions a r e nor due to the movements of the
eyes, but themselves bring about a change in thedirection of visual
pursuit.
I have given only two cases demonstrating how optical illusions
may influence eye movements. Many types of optical illusions a r e now
known. Some of these a r e easily explained, for example, illusions of
a change in the apparent size of objects depending on their brightness
(we s e e bright objects bigger than dark objects o f the same size). Such
illusions a r e due to an effect of irradiation Kravkov, 1950). Some
illusions have a definitely central origin, and these also a r e easily
explained. The origin of many illusions isnot yet known, however; the
explanations advanced to date cannot be regarded a s convincing.
Different illusions influence eye movements todifferentdegreesand in
different ways. At the same, time, some optical illusions have no influence whatever on eye movements. As an example of this last case,
the reader should carefully consider the two strong illusions illustrated
in Figs. 130and 131. Forinstance,the illusion in Fig. 130 has no effect
on the movements of the eyes, but ifan observer attempts to follow the
outline of the deceptive spiral in Fig. 131 with his eye, it will frequently jump from one turn of the spiral to the next.

!
I

I
I

5. EYE MOVEMENTS AND PERCEPTION OF MOVEMENTS

When we look a t stationary objects around us, every shift of the
Points of fixation i s accompanied by a displacement of the retinal image
over the retina. Although the perception of surrounding objects is

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

207

constantly accompanied by these displacements of the retinal image,

we s e e all stationary objects a s stationary. On the other hand, when
a moving object appears in the field of vision of our eyes, the movement of its image over the retina is such that moving objects a r e seen
a s moving. Then, under conditions of pursuit, (when the retinal image
is slightly mobile relative to the retina), theeye continues to see it as
a moving object.
Comparison of these two facts shows that types of retinal-image and
eye movements exist during which, in accordance with objective
reality, perceived objects appear to us to be moving, and that types of
retinal-image and eye movements exist during which perceived
objects appear to us to be stationary. Let us try to determine the
combinations of eye and retinal-image movements (or immobility)
with which objects a r e seen a s moving and the combinations with
which they a r e seen a s stationary.
As a result of a saccade, and synchronously with it, the retinal
'image of a stationary object moves across the retina through an angle
equal to the angle of rotation of theeye, in a direction directly opposite
to the direction of movement of the retina. The visual object is then
seen a s stationary, and a large enough saccade is always perceived a s
a change of the points of fixation (transfer of attention) on a stationary
object. In some experiments with the P4 cap, the retinal image of the
object moved in step with the saccades in a given direction so that the
angle of rotation of the eye andtheangle of displacement of the retinal
image always differed in magnitude, and the direction of the displacement of the retinal image was directly opposite to the direction of
movement of the retina, a s under normal conditions. In this case, the
visual object always appeared to the subject to be moving jerkily.
Naturally, the smaller this difference, the smaller the movement of
the object appeared to be; when the difference became on the order of
involuntary saccades during fixation (5-15 minutes of angle), the object
appeared to the subject tobestationary. If the direction of the retinalimage shift in step with the saccade was not directly opposite to the
direction of movement of the r e t i a , the visual object also appeared
to the subject to be moving jerkily.
Some particularly interesting experiments have !+en carried out
in which a visual object occupying a comparatively small part of the
visual field remained stationary relative to the retina a t the moment
of a saccade (or any other eye movement) and a t the moment of rest.
Such conditions may be regarded a s an ideal which the system of
pursuit strives to reach, although it never succeeds. P1and P8caps
were used in the experiments. The test field, stationaryrelative to
the retina, was, in some experiments, a transparent shadow visible to

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

the subject against the background of surrounding stationary objects.

In some experiments the stationary test field was a n opaque shutter,
also seen by the subject against a background of stationary checkered
objects.
These experiments showed that any test field, stationary
relative to the retina, visible to an observeragainst the background of
stationary objects, appears to be moving duringmovements of the eye,
and these movements coincide in direction andvelocitywith the movements of the eye. In this case the perception of movement of the
object is preserved even when a n active system of pursuit is no longer
needed, and either has ceased to work or is idling.
If the eye i s a t rest, movement of the retinal image, not too fast and
not too slow, over the retina is always perceived by the subject a s
movement of the object. We s e e moving objects a s moving not only
when we pursue them, but also when our attention is stationary. This
conclusion is valid even when the retinal image of an object moves
against a n empty field occupying the whole field of vision (as in the
experiments with "comets" and other experiments in Chapter 11).
In a series of experiments using the P4 cap, holes were made in
the cap m i r r o r s o that the subject's field of vision appeared to be
split into two different parts (one field will appear inside the other).
The visual objects in the part of thevisual field perceived by means of
the m i r r o r were in a state of constant and irregular movement, for
the subject could not choose his points of fixation a t will, and his
attempt to do so led to the result mentioned earlier. In the second
(smaller) part of the field of vision (corresponding to the hole in the
mirror), perception was a s usual and the subject could use eye movements a t will. Since the second (inner) field of vision was surrounded
by a field in which everythingwas moving about a t random, the question
arose whether or not these ,surroundings influence the perception of
objects visible through the hole in the m i r r o r of the cap.
The experiments showed that when the fields were related in a
certain way, when the outer field was large enough and the inner field
small enough, some influence of the surroundings was observed. The
objects in the inner field appeared displaced towardsthe side opposite
to the apparent movement of the objects in the outer field.
This fact suggests that the evaluation of the whole situation in
which a person finds himself in each concrete case plays a role in the
evaluation of the mobiiity o r immobility of surrounding objects.
Mistakes in the evaluation of mobility and immobility of surrounding
objects a r e observed in many pathological cases. Such disturbances
may be caused by a discrepancy between the muscular effort and the
actual movement of the eyes. By interfering with the free movement

209

of the eye or byexertingpressureuponit, we may observe an apparent

displacement of visual objects.
'
In conclusion, we repeat once again that under normal conditions.
with the head stationary and under ordinary conditions of perception, an
object appears stationary to us first, if the eye and the retinal image
of the object a r e simultaneously stationary (the process of fixation on
a stationary object) or if theretinalimageof the object, synchronously
with the eye movement during a saccade, moves relative to the retina
through the angle of rotation of the eye with the angular velocity of
the eye in a direction directly opposite to the retinal movement. In
all other cases, the observer sees the object a s moving in the visual
field. An object is also seen a s moving when the eye is stationary,
but the retinal image is moving over the retina, and also when the
retinal image is stationary (andseenagainstabackground of stationary
objects), but the eye is moving.

6. ROLE O F RECOGNITION IN EVALUATION O F SPATIAL

RELATIONSHIPS
The perception of surrounding objects, the evaluation of distances,
and the determination of the relative arrangement of objects a r e accompanied not only by eye movements, but also by rotation of the
head function of the vestibular apparatus, and the use of all our ontogenetic experience towards constancy of perception and recognition.
It is evident that during examination of objects turns of the head
play the same role a s eye movements. The immobility of the sur- .
rounding objects apparent to an observer when he turns his head is
maintained in perception by the vestibular apparatus. In this section
I would emphasize that recognition also plays an important role in
evaluation of the spatial arrangement of surrounding objects. The
following case is an instructive illustration of this.
Near a small wood, on the bank of one of the Volga reservoirs, an
old wooden barge was moored against the bank to serve as a landing
stage. A little two-room house had beenbuilt on the barge; an elderly
couple (employed by the river steamship line) lived there in the
summertime and worked on the landing stage. The landing stage
was a busy place. The house and its surroundings were very familiar
to many visitors. One day, however, the water level in the reservoir
fell by almost a meter and a half below the normal level. The barge,
one end of which rested on the bank, sloped steeply when the water
level fell, approximately by 1.5'.
Naturally the house on the barge,

210

CHAPTER VIl

and everything in it, sloped a t the same angle. The perception of anyone
who, when walking along the sloping surface of the barge, glanced or
walked through the open door of the house was very interesting. The
furniture and other things in both rooms hadbeen fastened in their old
places, s o that everything appeared a s usual. Every visitor looking into
the room involuntarily placed his body in theposition perpendicular to
the floor of the house (i.e.. at a n angle of approximately 75' to the
horizontal); he would either fall to the ground or land on one of the
walls of the room, which he would hold on to for some time until he
regained his balance, and began to move about the room unsteadily,
holding on to the walls and furniture.
The view from the window of the little house, if part of the room,
o r a t least the window frame, were included within the field of vision,
was quite fantastic. The horizon, the surface of the water, and the
surrounding neighborhood were all seen a s sloping. Moreover, the
large Volga steamers were plying on this slopingsurface of the water,
themselves a t an improbable angle.
There was a telephone in one room. When speaking on the
telephone, the visitors usually overcame their unsteadiness and stood
with body vertical. To another visitor sittingat the table, the position
o f a person speaking on the telephone appeared completely unnatural
and very funny. It was quite incomprehensible how he managed to
stand on his feet when his body was inclined to such a degree. A body
moving upright in these conditions looked l i k e a circus trick. Many
other details were equally unnatural. For example, it was amusing
to s e e the line of tea in a c u p o r the weights of the clock, which somehow seemed to have come away from the wall and to be hanging in space
because of mysterious forces.
This example illustrates the important role of recognition of a
situation a s a whole in our perception. By recognizing a familiar
environment and using this a s a basis, we thereby evaluate all the
secondary elements of what we see. Ifwe a r e mistaken in the evaluation of our basis, the second-degree elements may appear distorted to
us, and as our example shows, these distortions remain even if they
a r e contrary to common sense and to the indications of the vestibular
apparatus.
This example suggests that what is in general a very important
and desirable feature of our perceptionmay leadunder certain unusual
and r a r e conditions to distortion, o r in other words, to optical illusions.
An illusion of this type was shown inFig. 130. Here we recognize and
s e e a road going into the distance, andcolumns standing a t the edge of
the road. In natural conditions in accordance with the laws of per-

EYE MOVEMENTS DURING PERCEPTION OF COMPLEX OBJECTS

211

spective, the size of the retinal image of the columns should diminish

with an increasing distance from the observer (if all the columns
objectively a r e of the same height). If, on the other hand, the retinal
images of the columns a t different distances from the observer a r e
equal, the columns a r e in fact of different sizes. This is the case we
recognize in Fig. 130, so that the .objectively equal columns in,the
drawing appear to be different sizes.

CONCLUSIONS
The human eyes voluntarily and involuntarily fixate on those
elements of an object which carry or may carry essential and useful
information. The more information i s contained in an element, the
longer the eyes stay on it. The distribution of points of fixation on
the object changes depending on the purpose of the observer, i.e.,
depending on the information which he must obtain, for different information can usually be obtained from different parts of an object.
The order and duration of the fixations on elements of an object a r e
determined by the thought process accompanying the analysis of the
information obtained.
Hence people who think differently also, to
some extent, s e e differently.
Normally, reading speed depends not on the muscular apparatus of
a person's eyes, but on his higher nervous activity.
Optimal conditions for the solution of certain problems (visual
evaluation of proportions, estimation of length, comparison of areas,
angles, and so on) require macromovements of the eyes (large saccades). Without these, many visual evaluations a r e impossible or
attended by great difficulty and considerable waste of time.
In ordinary conditions of perception,an object is seen a s stationary:
first, if the eye and the retinal image of the object a r e stationary (the
process of fixation on a stationary object); second., if, synchronously
with the movement during a saccade, the retinal image of the object
moves relative to the retina through the angle of rotation of the eye
with the angular velocity of the e y e , in a direction opposite to the
retinal movement. In all other cases, the observer sees the object
a s moving in the field of vision. Specifically, an object is seen as
moving: first, when the eye is stationary, but the retinal image moves
over the retina; and second, when the retinal image is stationary (and
is seen by the observer against the background of stationary objects),
but the eye is moving.

LITERATURE CITED
Adamson, J. Ocular scanning and depth perception. Nature 168(4269):345 (1951).
Adler. F.H.. and Fliegelrnan. M. Arch. Ophthalm. 12:475 (1934).
Adrian. E.D. The Basis of Sensations. London, 1928.
Averhakh. M. I. Ophthalmological Sketches. Moscow-Leningrad, 1940.
Barlow. H.B. Eye movements during fixation. J. Physiol. (Engl.) 116(3):209 (1952).
Boeck, W. Die Physiologie der Augenbewegungen. Deursche Opt. Wochenschr. 6:48 (1951).
Bongard. M.M., and Smirnov. M.S. Doklady Akad. Nauk SSSR 102(6):1111 (1955).
Buswell. G.T. Haw adults read. Suppl. Educ. Mongr. (45). Chicago. 1937.
Byford, G. H. Eye movement recording, Nature 184(suppl. 19):1493 (1959).
Carmlchael. L.. and Dearhorn, W.F. Reading and Visual Fatigue, Londm. 1948.
Clark, B.C. Amer. J. Psychol. 46:325 (1934).
Clarke. F. J. Optica Acts (7):219 (1960);
Clowes. M.B. Some factors in brightness discrimination with constraint of retinal image
movement, Oprica Acta 8(1):f.1 (1961).
Clowes, M.B., and Ditchburn. R. W. Animproved apparatus for producing a stabilized retinal
image. Optica Acta 6(3):252 (1959).
Cobb. P. W., and Moss. F. K. J. Franklin Inst. 200:239 (1925).
Cords, R. Graefes Arch. Ophthal. 118;771 (1927).
Camsweet. T.N. New technique for the measurement of small eye movements. J. Opt. Soc.
Amer. 48(11):808 (1958).
Crawford. W . k Visual accuirj, eye movements, and head-eye movement integration, J.
Physiol. (Engl.) 145(2):50 (1959).
Delabarre. E.B. Amer. J. Psychol. 9:572'(1898).
Ditchbum. R. W. Problems of visual discrimination, 20th Thomas Young oration, delivered
to the society November 12. 1959.
Ditchburn. R.W. Eye movements in relation to perception of color. Visual Problems Colour
2 6 1 (1961).
Ditchburn. R. W. and Fenaer. D.H. Thestabilizedretinalimage. Optica Acta 2(3):128 (1955).
Ditchburn, R. W. Fendei. D. H., and Mayne. S. Vision with con+olled movements d the
retinal image, 3. Physiol. (Engl.) 145:98 (1959).
Ditchburn. R.W., and Ginshorg, B.L. Vision with a stabilized retinal image, Nature 170
(4314):36 (1952).
Ditchburn. R. W. and Ginsborg. B.L. Involuntary eyemovements during fixation, J. Physiol.
(Engl.) 119(1):1 (1953).
Ditchburn. R. W., and Pritchard. R.M. Binocular vision with ma stabilized retinal images,
Quart. J. EW. Psychol. 12(1):26 (1960).
Dodge, R. Psychol. Monogr. 8(4). 1907.
Wdge, R., and Cline. T.S. Psychol. Rev. 8:145 (1901).
Dohlman. G. Acta 070-Lame." Suvol. 5:78 11925).
Dohlman, G. Acta 070-Larpg. Suppl. 23:50(193?5).
Drischel. H., and Lange. Uber den Frequenzgang der horlzontalen Folgebewegungen des
meuschlichen Auees. Pflixer's Arch. Ges. Pmisiol. 26214):307-333 (1956).
Drischel, H., pfliigerss&ch. &s. Physiol. 262(1):34 (1958).

..

INDEX
Abducens nerve, 11
Accommodation, 5, 16
Adaptation
and stabilized image perception, 77
and visibility of contours, 15
Afterimages
apparent size of, 16
and convergence, 16
and empty fields, 77-78
and eye movements, 13-15, 129
Amacrine cells, 8
Amethocaine, 44
Anesthesia, see Amethocaine
Anterior colliculus, 9
Apparent size
of afterimages, 16
of objects, 156-158
Aqueous humor, 5
Blind spot, 7
see also Optic nerve
Blinks, 123
Blood vessels, 17-19
Ciliary body, 5, 6, 11
Color
delay in seeing, 83-96
fading of, 76
induced, 78, 96
Comets
of, 91-95
on, 91-95, 125
Cones
diameter of, 8
function of, 7
number of, 9
structure of, 6
see also Visual pigments

Cone foot-plate, 6
Conjunctiva, 6
Contact lens
and eye movements. 21
and stabilized images. 28-29, 63-64
see also Stakilized image perception
Contours
and adaptation, 15
disappearance of, 86
moving, 86-97, 91, 207
Convergence
and afterimage size, 16
and apparent size of objects, 156158
duration of, 151-154
preceding saccades, 147-149, 152156
and pursuit movements, 167
velocity of, 151
Cornea, 5, 6, 11
changes of curvature of, 124
and mechanical recording of eye
movements, 20
and visual observations of eye movements, 19
Corneal bright spot, 23-25
Corpora quadrigemina, 11
Divergence
and apparent size of objects, 156158
duration of. 151-154
and pursuit movements. 167
and saccades, 147-149,152-156
velocity of. 151

Drifts, 106
duration of, 111-112
function of. 126
and instruction to the subject, 112113
speed of, 108-109
Electrooculography. 25-26
Ellipsoid, 6
Empty field, 60-62
delay in seeing color in, 83-86
fading of color in, 76
and induced colors, 78, 96
and moving contours, 91, 96-97
and neural signals, 82
and scattered light, 98-100
spatial development of, 86-96
see also Stabilizedimageperception
Eye movements
and afterimages, 13-15
during complex object perception,
171-211
and corneal bright spots, 23-25
and electrooculography, 25-26
during fixation of a point, 104, 112,
118-119, 147-151
during fixation of stationary objects,
103-127
function of, 1. 2, 124-127
following geometric forms, 103-105
involuntary, 16, 17
macromovements, 2, 19, 103-127,
129-211
mechanical recording of, 20
micromovements, 2, 15, 17, '19, 21,
103-127
during movement of an object. 119,
159-170
during movement of the head, 118119
during movement of the head. and
object, 119
and photoelectric methods of recording, 26-28
and photographic methods of recording, 22, 23
during reading, 24, 196-200
and reflected beams of light. 20. $1,
47-53

Eye movements (continued)

during spatial relationship assessment. 200-202
and suction devices. 29-40
visual ohsewation of, 20
see also Drifts, Nystagmus. Pursuit movements. Saccades,
Tremor
Eye muscles
inferor oblique, 10, 13
inferior rectus, 10, 13
lateral rectus. 6. 10, 11, 13
medial rectus, 6, 10, 13
superior oblique, 10, 11, 13
superior rectus, 10, 13
Fast visual process, 78, 101
and disappearance of contours, 86
inhibition of, 93-95
time characteristics of, 91-92
Fixation
eye movements during, 104, 112,
118-119
field, 107
process, 105
Flicker. and stabilized image perception, 68-73
Fovea centralis, 6, 7
see also Retina
Ganglion cells, 8
see also Optic neNe
Head movements, and eye movements,
118-119,163
Hemianopsia, 16
Horizontal cells, 8
Induced colors, and stabilized image
perception, 78, 96
Inhibition. of the fast visual process,
93-95
Inner segment, 6
see also Cones, Rods
Iris, 5, 6
Lateral genicnlate body. 9
Lens, 5, 6, 123
Limiting membrane
inner. 8
outer, 8
Macromovements, see Eye movements
Macula lutea, 7
Medulla, 11

INDEX
Micromovements, see Eye movements
Mobile test field, 60
Natural viewing conditions, and stabilized image perception, 82-83,
85
Nuclear layer
inner, 8
outer. 8
Nystagmus, 119-122
amplitude of, 120
defined, 119-120
frequency of, 120
in rod monochromats, 120-121
vestibular, 121-122
see also Eye movements
Occipital lobes, location of, 9
and projection of visual fields, 12
Oculomotor nerve, 11
Optical axis, 6, 15
and drifts during fixation, 107-113
Optic chiasma, 11
Optic nerve, 1, 6, 8, 9
activity dur@gstabilizedimageperception, 79
Optic tract, 9
Optical illusions
and eye movements, 202-206
and stabilized image perception, 202
Outer limiting membrane, 6, 8
Outer segment, 6
see also Rods, Cones
Pareses, 3
Pigmented epithelium, 5, 8
Photography, of eye movements, 22,23
Plexiform layer
inner,8
outer, 8
Plexiglas mask, 43
Pons, 11
Pulsations of the eye, and suction devices, 33-35
Pulvinar, 9
Pupil, 5, 18, 124
Pursuit movements, under complex
conditions, 164-168
convergence and divergenceaccompanying. 168
of fast-moving objects, I63
and head movements. 163

Pursuit movements (continued)

of an imagined object, 160
involuntary aspects of, 159-161
and saccades, 167
of slow-moving objects, 162
velocity of. 161-164
see also Eye movements
Reading, and eye movements, 24, 196200
Retina, 6, 7, 11
Retinal image, 1 , 16
see also Stabilizedimage perception
Rods
diameter of, 8
function of, 7
number of, 9
structure of, 6
Saccades
acceleration of, 139-140
afterimages used in studying, 129,
145
and center of rotation of the eye,
142-143
defined, 103
development in time of. 137-140
direction of, 130
duration of, 129-137
function of, 129
involuntary, 106, 108-110, 115-118
large, 105
oblique, 140-142
overshoot of, 143-144
and position of the eye. 131
and pursuit movements, 167
and retinal image, 135
and rotary eye movements. 141-142
sinusoidal rule of, 137-139
small, 105
subject's control over, 133-134
vision during, 144-145
voluntary versus involuntary, 145146
see also Eye movements
Sclera, 5, 6
Sinusoidal rule, of saccades, 137-139
Slow visual process, 78, 86, 101
Stabilized image perception
and adaptation, 77
of bright objects, 62-64

Stabilized image perception (continued)

and color of test field, 61, 72-79
and contact lens method. 28-29, 6364
disappearance of contours during.
59, 61, 63
and displacements of test fields, 125
of flickering test fields, 68-73
and illumination of the sclera, 62,
70-71, 84-85, 92-93, 98-100
and increments and decrements of
light, 73-76
and luminance of test field, 72
luminance variation during, 64-68,
99
and moving contours. 86-97, 207
and natural viewing conditions, 8283, 85
nonstabilized image, influences
upon, 79-85
and optical illusions, 202
and optic nerve activity, 79
physiological explanation of, 59, 79,
82-83
and retinal blood vessels, 18-19
retinal changes during,. 76-78
and suction devices, 35-40
and vision through the second eye,
61, 12
Stationary t m t field, 60
see also Stabilizedimage perception

Strabismus, 3
Suction devices, 21, 29-40
apparatus used with, 40-43
application of, 45
construction of, 53-57
and eye movements (PI,P2,PS.P4).
30-33
and experimental technique, 43-47
and pulsations of the eye (Pd,33-35
removal of, 46
and stationary retinal images (P,,
P,, Pd, 35-40
Terminal bouton, 6
Tremor. 106
amplitude of, 113, 115
frequency of, 113-115
function of, 126
see also Eye movements
Trochlear nerve, 11
Troxler's effect, 17
Vascular membrane, 5
Vestibular nystagmus, see Nystagmus
Visual axis. 6
Visual cortex, see Occipital lobes
Visual following, see Pursuit movemencs
Visual pigments. 7
Visual receptor cells, see Cones, Rods
Vitreous humor, 5, 6
Weber-Fechner Law. 66, 75