Title: Retinal receptor orientation in amblyopic and nonamblyopic eyes assessed at several retinal locations using the psychophysical Stiles-Crawford function
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00098644/00001
 Material Information
Title: Retinal receptor orientation in amblyopic and nonamblyopic eyes assessed at several retinal locations using the psychophysical Stiles-Crawford function
Alternate Title: Retinal receptor orientation in amblyopic and nonamblyopic eyes ..
Physical Description: x, 259 leaves : ill., diagrs., graphs ; 28 cm.
Language: English
Creator: Bedell, Harold E., 1948-
Copyright Date: 1978
Subject: Amblyopia   ( lcsh )
Retina   ( lcsh )
Photoreceptors   ( lcsh )
Psychology thesis Ph. D   ( lcsh )
Dissertations, Academic -- Psychology -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Statement of Responsibility: by Harold E. Bedell.
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 245-258.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00098644
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000066136
oclc - 04394312
notis - AAH1351


This item has the following downloads:

retinalreceptoro00bede ( PDF )

Full Text








It is a pleasure to acknowledge the considerable contri-

butions which a large number of individuals have made to the

success of this research.

Dr. Jay M. Enoch, an outstanding teacher and scholar,

provided the author with many of the technical skills re-

quired in this research. He has fostered, by his excellent

example, a "first things first" attitude toward research.

Serving more than ably as an advisor, he has offered encour-

agement and astute criticism as needed. His contributions

to this dissertation are truly appreciated.

My wife, Niki, made enormous contributions to the com-

pletion of this dissertation. The least of these, but the

most tangible, were endless statistical calculations, aid

with several of the figures, assistance in proofreading,

and compilation of the bibliography. Daniel and Anna have

also contributed, perhaps more than they ought.

The patience, diligence and endurance of the observers

who participated in this research cannot be adequately ex-

pressed. They receive both my admiration and gratitude.

Virtually the entire Opthalmology Department at the

University of Florida, and especially the members of the

Eye Clinic, aided this research in some fashion. Special

contributions were made by Dr. C. R. Fitzgerald, M. D., and

Dr. I. [4. Rabinowicz, M. D., who kindly examined the observers

of this study. Dr. William Dawson and his clinical techni-

cian, Ms. B. Rhodes, performed the electrophysiological

measurements which are reported in this dissertation.

Dr. Emilio Campos, M.D., instructed the author in the

techniques of the strabismological examination and also

helped in examining some of the observers. Discussions

with Dr. Campos during the germinal phases of this research

were invaluable.

Dr. Frank Tobey provided useful suggestions as to in-

strument design and calibration. Mr. Larry Brock and Mr. Dan

Capehart assisted in the design and construction of electronic

and mechanical components of the apparatus.

Dr. John Flynn, M.D., and Dr. W. J. Knauer, M. D., pro-

vided clinical records of observers participating in this


Statistical procedures employed in this dissertation

were discussed with members of the J. Hillis Miller Health

Center Biostatistics unit. The advice of Dr. James Boyett

and Dr. Randy Carter was especially valuable.

Computing was done utilizing the facilities of the

Northeast Regional Data Center of the State University

System of Florida located on the campus of the University

of Florida in Gainesville. Mr. Bruce Wilhite assisted in

programming. The Physiology and Neuroscience Departments

graciously permitted the author to have access to remote


An excellent job of typing this dissertation was per-

formed by Ms. Diane Fischler.

This research was supported in part by National Eye

Institute Research Grant EY-01418 and Training Grant EY-07046

(to Jay M. Enoch) NIH, Bethesda, Maryland, and in part by

National Eye Institute Training Grant EY-00033, William

Ruffin, M.D., Project Director (to HEB).



AKNOWLEDGMENTS . . . . . . . . . . ii

ABSTRACT . . . . . . . . . .. viii

CHAPTER 1: INTRODUCTION . . . . . . . 1

Description of the Stiles-Crawford Effect. 5
Techniques of Measurement. . . . . 8
Psychophysics . . . . . . 8
Objective Measures . . . . .. 12
Summary and Comparison . . . .. 15
Basic Experimental Data. . . . .. 15
Retinal 1]luminance/Photopic vs.
Scotopic Adaptation . . . .. 15
Comparisons Between Different Retinal
Locations . . . . . . .. 17
Wavelength and Absorption of Ocular
Media . . . . . . . 21
Factors Influencing the Stiles-
Crawford Effect . . .. . . 26
Additivity of the Stiles-Crawford
Effect within the Pupil . . .. 29
The Stiles-Crawford Effect in Visual
Pathology . . . . . . . . 31
The Transient Stiles-Crawford Effect and
Directional Light Adaptation. . . . 35
Stiles-Crawford-like Effects in Infra-
human Species . . . . . . . 38
Theory . . . . . . . ... 42
The Stiles-Crawford Effect of the Second
Kind (SCE 11) . . . . . . .. 47
Basic Experimental Data. . . .. 47
The Stiles-Crawford Effect of the
Second Kind in Anomalous Vision . 51
Stiles-Crawford Effect of the Second
Kind: Theory . . . . ... 53
Significance of the Stiles-Crawford
Effect. . . . . . . . .. 54








Observers . . . . . . . .
Evaluation . . . . . . . .
History . . . . . . . .
Visual Acuities . . . . . .
Refractive Status . . . . .
Examination . . . . . .
Tension and Slit Lamp, Ophthalmoscope
Examination . . . . . .
Selection Criterion . . . . .

SCE Apparatus and Testing Procedures. .
Interferometric Resolution Testing
Apparatus and Procedures . . ..

RESULTS . . . . . . . ..
Stiles-Crawford Effect (SCE) Function
Peak Locations . . . . . . .
Control Observers . . . . .
Amblyopic Observers . . . ..
Within Eye Comparisons of Peak
Locations . . . . . .
Directionality of SCE Functions . . .
SCE Function Measurements for Observer
PMC . . . . . . . . .
SCE Function Peak Locations .. ..
SCE Function Directionality . . .
Other Measurements Obtained on
Observer PMC . . . . . .
SCE Function Results for Observer JEC .
Visual Resolution Thresholds . . .
Control and Amblyopic Observers .
Resolution Thresholds for Observer
JEC . . .. . . . . .
Visual Resolution Thresholds at the
Locus of Fixation and at the Posi-
tion of the Entoptic Fovea . .























DISCUSSION . . . . . . . 173
Patterns of Inferred Retinal Receptor
Alignment. . . . . . . . 173
Inferred Retinal Receptor Alignment
in Observer PMC. . . . . . .. .176
Central and Peripheral SCE Function
Directionality . . . . . . 180
The Relationship Between Inferred Retinal
Receptor Orientation and Amblyopia in
the Present Sample of Amblyopic Eyes .. 181
Inferred Receptor Orientation and Visual
Resolution Thresholds. . . . .. .184
Visual Resolution Measurements in
Amblyopic Eyes . . . . . .. 185

CONCLUSIONS . . . . . . . . 189



Key . . .


. 205
. 204





Abstract of Dissertation Presented to the
Graduate Council of the University of Florida
in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy



Harold E. Bedell

June 1978

Chairman: Jay M. Enoch
Major Department: Psychology

Retinal photoreceptor orientational tendencies were as-

sessed within both eyes of samples of control and selected

functional amblyopic observers, using the psychophysical

photopic Stiles-Crawford effect (SCE) function as an indi-

cator. SCE function determinations were made at visual field

testing locations spanning 30 of the horizontal meridian

of the visual field.

One aim of this research was to help define the role

of possible receptor alignment anomalies as a potential con-

tributing factor to the visual resolution deficits within

amblyopic eyes. A second aim of these studies was to iden-

tify individuals having anomalous patterns of inferred reti-

nal orientation and to characterize the nature and extent of

these anomalies. The nature of disturbed retinal receptor

alignment within such eyes can provide information about the

characteristics of hypothesized retinal photoreceptor align-

ment mechanisms within normal eyes.


In order to define the extent of visual resolution

deficits in the sample of amblyopic eyes tested, resolution

thresholds for flashed grating targets, formed by two beam

interference, were determined at several of the visual field

locations at which SCE function determinations were made.

The SCE function peak locations for the range of visual

field locations tested were found to cluster within a sub-

region of the pupil for all but one of the eyes tested.

Hence, in confirmation of previous such studies, retinal

receptors within these eyes are presumed to tend to align

toward the exit pupil of the eye.

The single exception to this generalization was found

for the nonamblyopic eye of one of the amblyopic observers.

Within this eye, SCE function peak locations, estimated

across 45" of the horizontal meridian of the visual field

and 150 of the vertical field, indicated that photoreceptors

at these testing locations tended to align more closely

toward the center of the eye than toward the exit pupil.

Photopic increment and dark adaptometric thresholds, deter-

mined at peripheral visual field locations, were found to

be elevated within this eye, consistent with predictions

based upon the SCE function results. Preliminary electro-

retinographic results indicated disturbances in the late

components of flash initiated electroretinograms in this

eye. This eye represents a possibly unique opportunity to

understand the mechanisms and consequences for vision of

retinal photoreceptor alignment in humans.

Within the sample of amblyopic eyes tested, inferred

retinal receptor alignment did not markedly differ from the

pattern found within control eyes, either at foveal or other

tested visual field locations. Thus, within this sample of

amblyopic eyes, visual resolution deficits seemed not to be

related to retinal photoreceptor alignment anomalies.

A nonamblyopic observer was identified, having a dis-

placed foveal SCE function peak in one eye and a more nearly

centered SCE function peak in the second eye. Foveal visual

resolution results for this observer revealed modestly in-

creased resolution thresholds for low photopic luminance

level targets within the eye having the displaced SCE func-

tion peak and no apparent differences for higher photopic

luminance targets. These data suggest that the modest amount

of foveal receptor tilt present in this eye has only a slight

effect upon visual resolution.

Within four of the sample of five amblyopes tested,

resolution thresholds within the amblyopic eyes were found

to be poorer than those of the corresponding nonamblyopic

eyes at both central and near peripheral (to 100 visual

field) locations. Additionally, in two amblyopic observers

having eccentric monocular fixation positions, best resolu-

tion thresholds were determined at the region corresponding

to the location of the fovea, rather than at the locus of

eccentric fixation.

The results of this dissertation have implications for

both the nature of retinal photoreceptor alignment mechanisms

and for the understanding of functional amblyopia.


This dissertation is an attempt to define the extent to

which possible anomalies at one level of the human visual

system contribute to the pathophysiology of functional am-


Functional amblyopia is characterized by a diminished

visual acuity, usually in one eye only, which is not resolved

by optimal refractive correction. Gross indications of pa-

thology are either absent or are of insufficient magnitude

to account for the acuity loss.*

A rather extensive body of evidence indicates that the

psychophysical Stiles-Crawford effect (SCE) function reflects

the orientational properties of retinal photoreceptor groups.

This evidence, as well as a fuller description of the SCE,

is presented in Chapter II. Since grossly maloriented recep-

tors have been observed to degrade retinal resolution capa-

bility (see Chapter II), a direct relationship is indicated

For the purposes of this argument, any ocular or sys-
temic condition which impairs vision and is localizable in
its effect is deemed to be pathological. Thus, for example,
opacification of the lens, which may simply be related to
theageing process, and various avitaminoses which impair
photopigment production, transport and/or regeneration, are
here considered to be pathologies.

between possible photoreceptor orientation anomalies and at

least a modest amount of visual resolution impairment.

As the major portion of this dissertation research,

the orientational tendency of groups of photoreceptors at a

number of retinal test locations within amblyopic eyes, for

the most part having only limited visual loss, and non-am-

blyopic eyes has been inferred from SCE function determina-

tions. In order that inferred receptor alignment might be

compared with visual resolution capability in the same eyes,

minimum angles of resolution for an interferometrically

formed grating target were measured at a number of locations

at which SCE function measurements were carried out.

This research was approached from a second and comple-

mentary point of view as well. In the eyes of all species

so far examined histologically, photoreceptors at all regions

of the retina tend to align toward the anterior part of the

eye and presumably toward the pupil, which is the source of

relevant visual stimuli. This analysis has been extended to

human observers utilizing the SCE function as an indicator

of retinal receptor orientation. The results from the eyes

of normal human observers indicate that the photoreceptors

tend to align toward a region near the center of the pupil,

for test locations as far as 350 in the peripheral visual

field. Additionally, a limited number of cases have been

documented, in which the SCE function, and hence presumably

receptor alignment, was disturbed during the active phase

of retinal pathology but subsequently recovered in conjunction

with remission of the pathological condition. All of these

data are more fully reviewed in Chapter II. Taken together,

they strongly indicate that an active process governs photo-

receptor orientation. The mechanism or mechanisms by which

receptors obtain and maintain their alignment with respect

to the pupil remain to be clarified.

Previous SCE function measurements in amblyopic eyes

(reviewed in Chapter 1I1) indicate that, in some cases, re-

ceptor alignment at the locus of fixation is disturbed. Such

disturbances might reflect a general alignment tendency for

all receptors within such eyes toward an anomalous alignment

centrum displaced from the exit pupil center. Alternatively,

photoreceptor alignment at more peripheral retinal locations

might show a tendency for orientation toward the pupil center.

In either case, measurements which assess photoreceptor align-

ment within these anblyopic eyes at a number of retinal loca-

tions would presumably give information concerning the nature

of the mechanism or mechanisms by which receptors maintain

their alignment. For example, one might distinguish whether

alignment is governed by retina-wide (global) or by more

local control mechanisms. Thus, the nature of receptor

alignment in amblyopic eyes was approached with the idea of

elucidating a possible model in which normal receptor align-

ment mechanisms might be studied.

As a part of this investigation, a unique individual was

identified. Photoreceptor orientation within one eye only

of this individual was determined to be directed more closely


toward the center of the retinal sphere than toward the eye

pupil. Studies of this individual are presented in Chap-

ter VII. The identification of this individual has important

implications relative to mechanisms subserving receptor

alignment and,moreover, promises to permit the consequences

for vision of anterior (pupil) pointing of photoreceptors

to be assessed.

Thus, this study is an attempt to characterize possible

photoreceptor alignment anomalies within amblyopic eyes with

relation to the visual loss sustained in these eyes and,

moreover, to evaluate the implications of these findings for

hypothesized receptor alignment mechanisms within normal eyes.


Description of the Stiles-Crawford Effect

The discovery of the Stiles-Crawford effect arose from

an attempt to determine pupil size from visual photometric

matching data. The technique employed by Stiles and Craw-

ford assumed that a homogeneous beam of light, limited by

the pupil, should change in apparent brightness exactly with

the area of the pupillary opening. By making a visual bright-

ness match between such a homogeneous beam, for which the

pupil forms the aperture stop, and a second beam which passes

through the pupil in sufficiently small cross section so that

retinal illuminance is unaffected by the pupil size, one

ought to be able to determine pupil area, and hence its di-

ameter, from the luminance ratio of the two beams at the sub-

jective match point. When Stiles and Crawford (1933) per-

formed this experiment, calculated pupil sizes consistently

deviated, in the direction of underestimation, from reported

objective (photographic) results. By then employing two

beams, both of which entered the artificially dilated pupil

in small cross section (using Maxwellian view), and by making

brightness matches when these beams were separated by vary-

ing distances in the pupil (one held at the pupil center for

convenience), it was found, under conditions of foveal view-

ing, that a beam entering the periphery of the pupil had to

have a luminance of from 5 to 10 times that of a beam enter-

ing the pupil center to provide the same subjective bright-

ness. That is, a beam passing through the periphery of the

pupil was judged considerably less bright than a physically

equal beam entering at the center.

Stiles and Crawford plotted the parameter 7, defined

as the ratio of the luminance of the standard beam (at the

pupil center) to that of the displaced beam at the photo-

metric match point, against the position of entry of the

displaced beam in the pupil. The resulting curves were es-

sentially symmetrical about a maximum value at or near the

pupil center and decreased monotonically to the pupil margins.

Stiles (1937) empirically fit the psychophysical retinal

directional sensitivity function, or SCE function, with the


log = log max r2

Throughout this discussion luminance will be used to
refer to the physical properties of the light stimulus, as
corrected for the spectral response characteristics of a
standard observer, whereas brightness will be reserved for
the subjective impression occasioned by the stimulus. Thus,
the Stiles-Crawford effect represents a situation in which
luminance and brightness, as defined here, are dissociated,
in that a source of constant luminance is judged to change
in brightness depending upon the region of the pupil through
which light from the source enters the eye. Some authors
have attempted to incorporate a correction for the SCE into
the definition of retinal illuminance (Wyszecki and Stiles,
1967; LeGrand, 1968, see below).

The parameter p is as defined above and 7 max is the value

of q at the function maximum. Distance within the entrance

pupil from the function peak in mm is represented as r. The

shape parameter, provides an index of the directionality,

or the spread, represented by the function. Enoch and Bedell

(1978) have suggested the use of the half sensitivity half

width, i.e., the distance within the entrance pupil, in mm,

from the SCE function peak to the position at which 7 falls

to one-half of 7 max, as a more readily interpretable indi-

cator of SCE function directionality.

Stiles and Crawford (1933) calculated the differential

light absorption which would occur within the eye media for

beams entering the pupil centrally and peripherally, based

on published data for ox eye media. Since the calculated

difference in absorption was small, they concluded that the

effect they described was not preretinal in origin. Craik

(1940) determined that preretinal absorption in cat and frog

eyes was insufficient to account for the SCE; however, Stiles-

Crawford-like effects had not been measured for either animal.

Good evidence that the SCE is, in fact, retinal in origin

derives from Crawford's (1937) discovery that the phenomenon

is greatly reduced in going from photopic to scotopic vision.

Presumably the change in adaptation state represents a change

of the retinal receptor type mediating brightness rather than

of preretinal absorption. Moreover, the SCE has been shown

to be greatly altered, and in some cases to subsequently re-

cover, in some retinal and pigmented epithelial pathologies

which do not alter preretinal absorption (see below).

The underlying basis of the SCE is currently thought to

be the differential sensitivity of the photoreceptors and as-

sociated morphological elements to light striking them at

different angles; one mm in the pupil corresponding to nearly

2.50 of angle at the fovea (O'Brien, 1946). Light entering

the receptors parallel to their long axes is presumably the

most effective for producing excitation. Hence, the SCE

function apparently samples the orientational properties of

groups of retinal photoreceptors.

Techniques of Measurement


In their original measurements, Stiles and Crawford em-

ployed both direct photometric matching and flicker photome-

try, both techniques giving essentially the same results.

Flicker matches were reported to have the advantage of reduc-

ing the disturbing effects of ocular aberrations for beams

displaced toward the periphery of the pupil and hence imaged

along an off-axis optical path within the eye. Enoch (1958),

who also compared direct matching and flicker photometric

techniques for Stiles-Crawford effect curve determinations,

found that blurring introduced by the peripheral pupil path of

the displaced beam caused a brightness decrement for beams

passing eccentrically in the pupil in addition to that pro-

duced by the SCE itself. Flicker photometrically measured SCE

curves were found to be less sensitive to optical blur. Ad-

ditionally, when SCE functions were determined with narrow

band spectral stimuli, rather than with white light, the ef-

fect of chromatic aberration upon beams which enter periph-

erally in the pupil, and which disturbs both flicker and di-

rect matching judgments, was minimized.

Wright and Nelson (1936) were among the first authors

to confirm the basic SCE phenomenon, using Wright's binocular

colorimeter to obtain measurements. A field produced by a

moveable beam entering one pupil was matched to a closely

apposed field seen by the other eye and fixed in its entry

position at the pupil center.

Crawford (1937) measured SCE functions by finding the

incremental threshold for a small test field. The test beam's

entry position in the pupil was displaced for successive

threshold determinations whereas the background beam always

entered at the pupil center. The SCE curves determined by

the increment threshold procedure were comparable in all re-

spects to those determined by direct photometric matching.

Some of Crawford's measurements were made with the background

field extinguished, i.e., as the absolute threshold for the

test beam as a function of its pupillary entry position.

For test fields confined to the foveal region, absolute and

increment threshold determinations of the SCE gave similar

results. Crawford's experiments indicated that the foveal

SCE is relatively unaffected by the condition of light adap-

tation (but see below) and, moreover, provided the empirical

justification for the use of sensitive threshold techniques

in SCE measurements. [At about the same time, Goodeve (1936),

who was investigating the visibility of stimuli in the far

red region of the spectrum, also utilized a threshold proce-

dure to indicate the existence of an SCE at wavelengths of

about 830 nm.]

Flamant and Stiles (1948) used the increment threshold

technique to determine peripheral photopic and scotopic SCE

functions. In the procedure they adopted, the test beam

was fixed at the pupil center and the surround field beam

was displaced across the pupil. This modification of the

earlier increment threshold technique has the advantage that

the larger surround field, rather than the smaller test field,

is subjected to the off-axis optical aberrations of the eye.

Thresholds for small fields are highly dependent upon blur,

as evidenced by the use of stigmatoscopy for visual refrac-

tion (Ames and Glidden, 1928), whereas larger fields are less

so (Ogle, 1961a, b). Additionally, since the surround field

beam is displaced across the pupil, its image undergoes what-

ever shifts of position on the retina may accrue from spheri-

cal and other aberrations, instead of the test field's image.

This assures that thresholds are always determined for the

same retinal locus with respect to a stationary fixation


The particular increment threshold technique employed

by Flamant and Stiles (and others) measures the SCE indi-

rectly, however. What is actually determined is the threshold

change in the incremental field produced by the change of

brightness of the surround field due to the SCE. Fortunately,

insofar as the Weber relation holds, i.e., A L/L = constant,

which is over a considerable range of luminances in normal

observers, the change in the surround field brightness, in

logarithmic units, is accompanied by an equal change in the

log luminance of the increment at threshold:

AL/I = constant (Weber relation)

thus A Lo/Lo = aLx/Lx

where AL and L are the luminances of the incremental and

surround fields, and the subscripts refer to the reference

(o) and displaced (x) beams.

Therefore, log (A Lo) log (Lo) = log (ALx) log (Lx)

hence, log (Lo) log (Lx) = log (A Lo) log (A Lx)

(or change in log surround = change in log


Moreover, since

S= Lx/Lo

and log 7 = log (Lx) log (Lo)

log 4 may be plotted as the differences in log increment

threshold values between the displaced and reference beams.

Thus, over the Weber range of the increment threshold curve,

measurement of changes in increment threshold is equivalent

to measurement of changes in effective background luminance.

It is apparent, however, that for regions of the increment

threshold curve other than the Weber portion, increment

threshold changes are not equal to changes in the effective

luminance of the surround field; hence, increment thresholds

as a function of the pupil entry position of the background

beam do not adequately represent the SCE.

In practice, an increment threshold function is obtained

with the entry position of the background beam held fixed

in order to determine a background field luminance for which

a valid SCE function may be obtained.

Blank et al. (1975) described a rapid method for de-

termining the location of the peak of the SCE function (also

see Enoch, 1959a). Two beams are made to enter the pupil at

a fixed separation (2 mm in practice); light from each beam

forms one-half of a bipartite photometric field. The pupil-

lary entry positions of the two beams, which move in tandem,

are adjusted to give a photometric match. If symmetry of

the SCE function about its peak is assumed or has previously

been determined, then at the match point the entry positions

of the two beams in the pupil straddle the position of the

SCE peak in the tested pupillary meridian. In principle,

this peak finding technique might be modified to generate

entire SCE functions as well.

Objective Measures

In 1937, Crawford, using a subjective psychophysical

technique, measured the recovery of visual sensitivity in

the dark following a light preadaptation mediated by a beam

either entering at the pupil center or near its margin. As

expected, higher thresholds were evident during the early

phase of the recovery of sensitivity (the cone portion) when

the adaptating stimulus entered at the pupil center, i.e.,

close to the peak of the SCE function.

Presumably, the recovery of visual sensitivity after

exposure to bright light depends in large part upon the re-

generation of bleached pigment (Rushton, 1972). Crawford

measured the recovery of sensitivity in time, from which the

pigment bleaching effectiveness of adapting beams entering

the pupil centrally or eccentrically could be inferred. With

the development of funds reflectometry techniques, by which

retinal photopigment could be measured directly before and

after exposure to pigment bleaching adapting lights, pigment

bleaching efficacy could be directly measured as a function

of the pupil entry position of the adapting beam. [Goldmann

(1942) performed an experiment, similar in principle, which

anticipated funds reflectometry. He noted ophthalmoscopi-

cally that retinal images formed by central and peripheral

pupillary beams were unequally bright when the observer

whose retina he was observing had photometrically matched

the two fields.] Thus, Ripps and Weale (1964) determined

that the relative efficacy of monochromatic lights in bleach-

ing foveal photopigment when these entered either at the

pupil center or 3 mm displaced was comparable to the rela-

tive efficacy of these lights in producing brightness. Coble

and Rushton (1971) found adapting light intensities which

produced equal photopigment bleaching for entry positions

spaced across the eye pupil. The obtained "equal bleaching"

function was in excellent agreement with a psychophysically

determined SCE function for the same observer.

Spring and Stiles (1948) changed the pupil entry posi-

tion of a large (520) Maxwellian view adapting field and

measured the consensual pupillary reflex in the contralateral

eye. They observed a relative dilation of the contralateral

eye when the adapting beam entered peripherally in the pupil

as compared with central entry; however, the effect was con-

siderably less than was expected on the basis of the psycho-

physically determined SCE. It seems probable, however, that

the large white field used by these investigators favored

peripheral rods rather than central cones, in which case a

much reduced SCE would be expected (see below). Alpern and

Benson (1953) repeated the Spring and Stiles study using a

centrally viewed 10 red adapting field in order to favor

cone vision. Under these conditions, a considerable change

in contralateral pupil diameter was recorded, consistent with

the psychophysically determined SCE, when the entry position

of the adapting light was displaced in the ipsilateral pupil.

Both Armington (1967) and Sternheim and Riggs (1968)

employed electrophysiological methods to determine SCE func-

tions in human observers. In both cases constant luminance

but pattern modulated stimuli were used to evoke electrore-

tinographic (ERG) responses favoring cone activity. ERG

and psychophysically determined SCE functions were highly

similar. Sternheim and Riggs also measured fast flicker

(25 Hz) ERGs for different pupil entry positions of a homo-

geneous stimulating beam which yielded a function quite like

the psychophysical SCE as well. Slow flicker (4 Hz) ERGs

showed much less amplitude change as pupil entry position

was altered; presumably scotopic activity was preferentially

tapped with this procedure.

Summary and Comparison

The SCE is a remarkably robust phenomenon in terms of

its insensitivity to the technique by which it is measured.

Thus, psychophysical, electrophysiological and other objec-

tive techniques all yield very similar functions, although

the variability of data and potential artifacts encountered

with some techniques make them less attractive than others.

[The techniques described here by no means exhaust the pos-

sible alternatives. For example, at the suggestion of

Dr. Michael Halasz, of the National Eye Institute, an SCE

function has been successfully determined using a supra-

threshold magnitude estimation direct scaling paradigm.

Stiles (1959), Westheimer (1968), and Heath and Walraven

(1970) all describe procedures by which the SCE can be

visualized entoptically. Local variations in the SCE may

also be seen entoptically (see below.)] In general, psycho-

physical threshold measures show the least variability; when

coupled with criterion independent (signal detection) para-

digms a remarkable degree of sensitivity can be achieved.

Basic Experimental Data

Retinal Illuminance/Photopic vs. Scotopic Adaptation

Crawford's (1937) study indicated that the foveal SCE

function was essentially independent of the background

luminance. Both Stiles (1939) and Crawford (1937) showed a

similar independence of background luminance for the extra-

foveal SCE provided the level was relatively high (greater

than ca. 1.2 log photopic td at 5). At low background in-

tensities or at absolute threshold, extrafoveal SCE functions

were reported to be much more nearly flat (Crawford, 1937;

Stiles, 1939; Flamant and Stiles, 1948), indicating that in

scotopic vision, brightness is much less dependent upon entry

position of the beam in the pupil. Using a sensitive signal

detection psychophysical procedure, vanLoo and Enoch (1975)

definitively demonstrated the existence of a scotopic SCE on

the order of 0.2 0.3 log units which had been previously

hinted at in Stiles (1937) data and occasionally in subjects

whose SCE functions were decentered in the pupil (Flamant

and Stiles, 1948; Daw and Enoch, 1973; Bonds and MacLeod,

1978). When correction was made for the differential path

lengths of central and peripheral beams through absorbing

lens pigment (see below), the magnitude of the scotopic ef-

fect was found to be quite similar at all wavelengths tested.

Moreover, both photopic and scotopic SCE functions taken at

the same retinal locus had common axes of symmetry with re-

ference to the pupil. This finding is highly significant in

terms of the retina] receptor orientation properties which

the SCE apparently reflects.

In the mesopic range, the SCE makes a gradual transi-

tion between the small scotopic effect and the larger photopic

effect. This so-called "rising-sun" phenomenon was first

reported by Crawford (1937).

At very high photopic levels, the magnitude of the SCE

at the fovea increases, i.e., there is more difference in

brightness between beams passing through the center and pe-

riphery of the pupil (Stiles, 1937; Miller, 1964). Walraven

(1966) has also confirmed this effect for midspectral mono-

chromatic lights and proposed that it may be due to cone pig-

ment bleaching (see below). In an often cited article Le-

Grand (1948) reported that the peak of the SCE curve shifted

with respect to the position of the corneal reflex (produced

by the photometric field) in an orderly fashion as the testing

luminance was changed. This intriguing finding is not re-

plicated in either Crawford's (1937) or Miller's (1964) data.

LeGrand's results may reflect a change in position of the

corneal reflex with luminance changes, due perhaps to altered

fixation (Simon, 1904), rather than of the position of the

SCE peak in the pupil.

Comparisons Between Different Retinal Locations

Westheimer (1967) argued that if the SCE reflected re-

tinal photoreceptor optical properties, and if these were

physical optical rather than geometrical optical in charac-

ter, due to the small size of receptor apertures (Toraldo

di Francia, 1949), then foveal and extrafoveal photopic SCE

functions might be expected to differ, reflecting anatomical

differences between foveal and extrafoveal cones. In fact,

he found that the foveal SCE was of lesser magnitude than that

measured 3-3/40 parafoveally. Similar results can also be

seen in the data of Vos and Huigen (1962) taken at 00 and 40,

although these authors failed to comment on this finding.

Stiles' (1939) results also indicate a larger SCE parafoveally

(50) than foveally for long wavelength stimuli; the same re-

lationship is not seen in his short wavelength data (but see

below). Enoch and Hope (1973) measured SCE functions for

orange test and surround fields at 0, 20, 3-3/40 and 100

and found that the change in magnitude of the SCE occurred

between 00 and 20 with little change thereafter out to 10.

Recently,Bedell and Enoch (1978) determined that the SCE at

350 is similar in magnitude to that at the fovea.

Histoloyical investigations in many species have shown

that photoreceptors do not point toward the center of the

eye but rather toward some anterior location, presumably

within the pupil (Laties et al., 1968; Laties, 1969; Laties

and Enoch, 1971; Enoch, 1972; Enoch and Horowitz, 1974;

Baylor and Fettiplace, 1975). Webb has drawn similar conclu-

sions based upon the X-ray diffraction patterns produced by

photoreceptors in intact eyes (Webb, 1972; 1977). Thus,

only in the vicinity of the posterior pole are receptors

perpendicularly aligned with respect to the underlying pig-

ment epithelium substrate.

Inasmuch as the SCE represents the orienting properties

of the photoreceptors (Enoch and Laties, 1971), one would

expect that peaks of SCE 'functions measured at different

retinal loci would mirror these histological findings, i.e.,

SCE peaks should maintain an approximately constant relation

to the pupil center. In the earliest study in which the SCE

was determined over a considerable range of retinal locations,

Aguilar and Plaza (1954) found that SCE peaks up to 370 from

the fovea remained within the dilated entrance pupil. Enoch

and Hope (1972a) carefully measured SCE functions from 50

in the temporal visual field, approximately the posterior

pole, to 200 in the nasal visual field. Their findings were

that both horizontal and vertical traverses of the pupil

yielded SCE peaks clustered about a point, slightly different

for each of their three observers but for each near the cen-

ter of the entrance pupil of the eye. They confirmed these

results for three more observers at loci between 00 and 100

in another study (Enoch and Hope, 1973). Bedell and Enoch

(1978) extended this analysis to 350 in the temporal periph-

eral field and found that receptor alignment within the pupil

is maintained. Because the optic disc is interposed between

this peripheral region and the fovea, forces which act on the

retina, such as occur in accommodation and during rapid eye

movements (see below) will have somewhat different effects

in these two areas. Thus, receptor alignment toward a region

within the eye pupil on both temporal and nasal sides of the

optic disc is quite remarkable.

VanLoo and Enoch (1975) found that photopic and scotopic

SCE functions had a common axis of symmetry when measured at

60 in the nasal visual field. Results of Bedell and Enoch

(1978) for one observer at 350 in the nasal retina are also

indicative of common alignment of photopic and scotopic

receptors (see also Flamant and Stiles, 1948; Daw and Enoch,

1973; Bonds and MacLeod, 1978).

Given the consistency of their results, Enoch and Hope

(1972b) sought to determine whether the foveal SCE peak

better aligned to the center of the dilated or constricted

pupil, since the pupil center shifts slightly nasally from

complete dilation to complete constriction (Gullstrand, 1962;

Enoch and Hope, 1972b). Moreover, normal SCE functions have

a slight bias toward peaking on the nasal side of the dilated

pupil center (e.g., Dunnewold, 1964). Enoch and Hope measured

SCE curves with respect to the corneal reflex, since the

pupil center obviously could not be used. They employed a

"peak finding" technique, in which observers adjusted two

equal luminance beams having a fixed separation in the en-

trance pupil, to that position of pupil entry which rendered

both fields equally bright. Although for two of three ob-

servers SCE peaks were nearer to the constricted than to the

dilated pupil center, the result must be taken to be incon-


While the overall pattern of receptor pointing, as in-

ferred from the SCE is remarkably constant across at least

a considerable region of the retina, there is apparently

some degree of variability, as would be expected in any

biological system, in the alignment of groups of neighboring

receptors. This was first recognized by O'Brien (1950) who

constructed an aperture which corrected for the photopic

SCE but saw residual bright and dark patches in what was

expected to be a uniform field. Brighter and darker regions

shifted with respect to one another as the entry position

of the compensated beam in the pupil was altered. The ef-

fect rapidly faded when the beam was not moved, presumably

due to image stabilization. The same phenomenon has been

reported by Enoch (1967) and by Heath and Walraven (1970)

for short flashes. Enoch et al. (1978) have observed that

a similar, if somewhat more difficult to observe, effect

also occurs in scotopic vision, apparently reflecting local

variations in orientation of groups of rods. All observa-

tions to date have been qualitative only.

Heath and Walraven (1970) proposed that in the central

4 of the retina, receptors are parallel rather than point-

ing toward a common point in the eye pupil. Their evidence

consists of slight shifts of SCE function peaks measured in

this region. Since the shifts involved are rather small and

since these results have not been replicated (Enoch and Hope,

1972b), this proposed departure from anterior pointing of

photoreceptors must be regarded as questionable.

Wavelength and Absorption of Ocular Media

Stiles (1937) reported that the magnitude of the foveal

SCE changed systematically with the wavelength of the test

stimulus. He found the smallest SCE in the green region

and larger effects at both ends of an equal luminance spec-

trum. In later work (Stiles, 1939), he confirmed this gen-

eral wavelength dependence; however, he measured a smaller

SCE at all wavelengths using different measurement techniques.

Both sets of measurements were on Stiles' own eye, which

during this period also evidenced a shift in the peak of the

SCE (see below).

Foveal SCE data as a function of wavelength have also

been published by Enoch and Stiles (1961) for two subjects,

as raw data only, also for two observers, by Safir and Hyams

(1969), and for one observer by Wijngaard and van Kruysbergen

(1975). In all cases, a larger SCE is observed for long and

short wavelength than for mid-sprectral monochromatic targets.

In 1939, Stiles also measured the wavelength dependence

of the photopic SCE for parafovea] targets. Here the SCE

was largest at the red end of the spectrum, falling to a

rather constant (non-zero) value in the green and blue re-

gions. This experiment has apparently never been repeated.

It is curious that the wavelength dependence of the SCE for

the foveal and parafoveal photopic retina should differ.

Although the parafoveal curves are not at all scotopic in

appearance, a possible rod contribution at the more eccen-

tric pupil entry positions needs to be considered. This is,

in fact, suggested by discrepancies between SCE functions

determined by increment thresholds in which test and surround

fields were of the same wavelength and functions determined

with test and surround of different wavelengths. (Compare

Stiles, 1939, Fig. 13 with Fig. 20, 23.) It can thus be

said that the wavelength dependence of the photopic SCE out-

side of the fovea is an open question. On the other hand,

the magnitude of the scotopic SCE seems to be less dependent

on wavelength, after correction has been made for absorption

by the ocular media (Flamant and Stiles, 1948; VanLoo and

Enoch, 1975).

Returning to the foveal data, the function of magnitude

of the SCE vs. wavelength has been suggested to represent a

composite of the somewhat different directional sensitivities

of three cone types posited for trichromatic vision (Stiles,

1937, 1939; Enoch and Stiles, 1961; Walraven and Bouman, 1960;

Safir et al., 1971). Stiles (1939) approached this problem

by determining increment threshold functions for a variety

of test field, surround field wavelength combinations and

either central or peripheral pupil entry of the test beam.

For some combinations of test and surround wavelengths, the

difference between increment threshold curves for central

and peripheral pupil entry of the test beam changed more or

less abruptly as a function of the background luminance, in-

dicating a change in the directional sensitivity of the de-

tecting mechanism. Stiles interpreted these data in terms

of three cone mechanisms, the blue cones being more direc-

tionally sensitive than either the red or green cones, and

the green cones slightly more so than the red cones for wave-

lengths under 620 nm.

Enoch and Stiles (1961) calculated the directional

sensitivities of three cone types based on color matching

functions for central and peripheral pupil entry of test

fields. The results indicate that all three receptor types

have directional sensitivities which change as a function of

wavelength; the blue receptors were found to be more direc-

tionally sensitive than either the red or green receptors.

The effect of strong light adaptation on the magnitude

of the SCE was briefly alluded to above. Stiles (1937) noted

a marked increase of the SCE in midspectrum for high lumi-

nance targets. Walraven (1966) also found evidence for an

increased SCE after strong light adaptation for "green" sensi-

tive elements but not for long wavelength sensitive receptors.

Additionally, he measured a recovery period for this increased

SCE, with a time constant of ca. 30 sec, consistent with that

of cone photochemical regeneration following bleaching.

Brindley (1953) noted that the Stiles-Crawford hue change

(SCE Il; see below) effectively disappeared when the eye

was first exposed to a bright adapting stimulus (but see

also Wooten et al., 1977). Earlier, Wright (1946) had de-

termined that the foveal luminosity curve for a Maxwellian

view beam entering the pupil 3 mm from its center was most

similar to the luminosity curve for a higher luminance tar-

get entering at the pupil center. Taken together, these re-

sults suggest that eccentric pupil entry and light adapta-

tion have comparable influences on (at least midspectral)

luminosity. Brindley (1953), following a proposal made by

Stiles (1937), suggested that this similarity might be due

to reduced effective receptor photopigment concentration in

both instances; due to a partial bleaching from light adapta-

tion and to a diminished effective absorbing path length

within the receptor in the case of oblique incidence. These

ideas will be further treated below.

Corrections for ocular media. While the SCE would seem

to be primarily retinal in origin, the "retinal function" is

modified somewhat by differential absorption of beams passing

through the center and periphery of the pupil within the eye

media. The bulk of the absorption within the eye is attribut-

able to the lens pigment which most effectively absorbs light

at the short wavelength end of the spectrum. Macular pigment

absorption is relatively unimportant since the path length

differences involved are only about 2 per cent.

Weale (1961) corrected the magnitude of the SCE as a

function of wavelength for lens absorption, assuming homo-

geneous pigment concentration throughout the lens. Since

the path length through the lens is longer for beams passing

through the center than through the periphery, the correction

results in a modest increase in the magnitude of the SCE at

middle and long wavelengths; at short wavelengths, the in-

crease in the (presumed retinal) SCE is quite substantial.

Weale also noted that after lens path correction, Stiles'

1939 data indicated a small SCE in scotopic vision; this ef-

fect was later measured by VanLoo and Enoch (1975).

Mellerio (1971) investigated Weale's assumption that

lens pigmentation was homogeneous, finding it to be correct.

His own results indicated that changes in lens absorption

with age reflected increased lens thickness rather than in-

creased pigment density. Hle calculated a correction for the

SCE based on his data, indicating an even larger SCE at

short wavelengths than Weale.

Vos and van Os (1975) challenged the Mellerio correc-

tion for short wavelengths, arguing that he failed to account

for the slight decentration of the SCE peak in the pupil

generally found in normal observers. They argued that the

effect of lens pigment at short wavelengths is to displace

the psychophysically determined SCE peak and distort the

function's shape, rather than to diminish its magnitude by

as much as the Weale and Mellerio corrections suggest. It

would seem that the validity of the various lens corrections

could be addressed experimentally in a aphakic observer, cor-

rected with a corneal contact lens, as suggested by Bailey


Factors Influencing the Stiles-Crawford Effect

Richards (1969) hypothesized that the rise in visual

threshold which accompanies rapid eye movements, or saccades,

may be due, in part, to forces acting on the retina as the

result of the eye movement. In particular, he proposed that

the retina might be subjected to a sheering force relative

to its substrate during the rapid acceleration and decelera-

tion phases of saccades. If such sheering occurred, it

might be expected to transiently alter receptor orientation.

Such an effect would presumably be reflected in the SCE, if

measured directly after saccades. In fact, foveal SCE func-

tions determined 40 msec after 50 amplitude saccades showed

a 0.6 mm shift in the peak toward the temporal edge of the


More recently, Blank et al. (1975) showed that foveal

SCE function peaks shifted between one-half and one and a

half mm nasally in four eyes under conditions of marked (9 D)

accommodation. Marked accommodation had previously been

shown to result in an advance of the retinal boundary toward

the ciliary body, which had been calculated to increase reti-

nal area by over 2 per cent (Moses, 1970; Enoch, 1973).

Ronchi (1954, 1955) reported that mydriatic agents,

both a parasympathetic blocking agent, and hence cycloplegic,

and a sympathomimetic, affected the magnitude of the foveal

SCE. Both the direction and amount of change in the SCE

(assessed as the change in log ) depended upon the drug

used and the time since instillation. Effects lasted for

several hours. No influence of the drugs on the location

of the SCE peak was described; however, since most of the

measurements were a comparison of the relative efficacies

for two points in the pupil, rather than determination of

entire SCE curves, influences upon the peak cannot be dis-

counted, and may, in fact, represent much of the described

effect. It is not inconceivable that such drugs might in-

troduce forces upon the retina which would alter the posi-

tion of the SCE peak. Retinal distortion, tearing, and de-

tachment are known, for example, to occasionally accompany

the use of powerful miotic agents such as diisopropyl fluoro-

phosphate and phospholine iodide (Lemcke and Pischel, 1966;

Moses, 1970).

In 1939, Stiles reported the results of six years of

SCE determinations on his own eye. During that period the

peak of his foveal SCE function had shifted from 0.2 mm

nasal of pupil center (Stiles and Crawford, 1933) to 0.9 mm

temporal in 1939 (Stiles, 1939), giving a total shift in

the horizontal meridian of just over 1 mm. Very little or

no change occurred in the position of the SCE peak for a

vertical pupillary traverse. Crawford's eye showed essen-

tially no change in the position of the horizontal peak

from 1933 to 1937 (Stiles, 1939). Safir et al. (1970)

failed to find any horizontal shift in the SCE peaks of two

observers over two years. Bedell and Enoch (1978) measured

both vertical and horizontal foveal SCE functions for two

observers and compared these with earlier data. In one

case, no change in either peak could be detected after a

lapse of five years; in the second, a 0.4 mm horizontal

shift was detected over a 17 year interval. (As a point of

reference, a 1 mm shift in the position of the SCE peak in

the entrance pupil corresponds to a change in angle of ap-

proximately 2-1/2' at the posterior pole.)

With the exception of Stiles' eye, then, normal SCE

function peak positions within the pupil are remarkably

stable over time considering the stresses to which the

retina is constantly exposed. In contrast, cases of active

retinal pathology may show changes in the position of the

SCE peak over the course of months (e.g., Campos et al.,

1978). Transient changes of the position of the SCE peak

can result from forces acting upon the retina, such as ac-

company rapid eye movements or accommodation. It is pos-

sible that a shifting of the SCE peak over time in a normal

eye may reflect the action or relaxation of such forces.

Additivity of the Stiles-Crawford Effect within the Pupil

In the paper in which Stiles and Crawford first reported

the SCE, they also investigated whether the relation which

described the relative brightness of beams entering at two

points in the pupil also described the brightness of a beam

which filled the pupil (Stiles and Crawford, 1933). They

found that when they graphically integrated the data for

horizontal and vertical traverses of the pupil across the

entire pupillary aperture, their predictions for the apparent

brightness of extended beams were good to 6 per cent or less.

Since this first study others have reported slight departures

from perfect additivity in one direction or the other (e.g.,

Ercoles et al., 1956; Ronchi, 1955; Enoch, 1958). Enoch's

study demonstrated that some and perhaps most of the depar-

tures from additivity resulted from the blurring of off-

axis images. lie found much closer approaches to perfect

additivity when blur was corrected or when brightness was

determined by flicker photometry rather than direct matching.

What has perhaps not been adequately appreciated is that all

tests of additivity rest upon the assumption of a perfect

radial symmetry of the psychophysically determined SCE

function as well as a perfect description of the SCE by the

mathematical function chosen to represent the data. Clearly

neither assumption is warranted.

Drum (1975) sought to determine whether the SCE oc-

curred for non-Maxwellian view stimuli, i.e., whether the

SCE was an artifact of Maxwellian view optics, a problem

which Enoch had also addressed earlier. Drum found that

the SCE did occur with Fraunhofer images of his stimuli

(which corresponds to normal viewing circumstances) and

that the brightness of a stimulus which passed through the

pupil in an annular beam was virtually exactly predicted

from the average of the brightnesses of four quarter annuli.

Corrections of retinal illumination through an ex-

tended pupil for the SCE to yield an effective retinal il-

lumination have been offered (Moon and Spencer, 1944; Wyszecki

and Stiles, 1967; LeGrand, 1968). Not only do these correc-

tions assume additivity but also a standard observer with

an SCE centered in the pupil. Moreover, they are predicated

upon foveal viewing, a moderate luminance and white light.

It is seen that such corrections are useful as a general ap-

proximation in photometry.

Enoch and Iaties (1971) addressed the question of the

effect of a decentered SCE within the pupil upon perceived

Even if the retinal SCE were perfectly symmetric about
its peak, which is not established, differential lens absorp-
tion for different optical paths would tend to distort the
psychophysical function for a peak not centered in the pupil
(see above).

brightness for viewing with natural pupils of various sizes.

They employed an analog device (an aperture incorporating

the SCE) and assumed perfect additivity. The results indi-

cate that decentering of the SCE peak in the pupil has only

a moderate effect upon "integrated brightness"; a decentered

SCE more reduced "brightness" for smaller than larger pupil


The Stiles-Crawford Effect in Visual Pathology

Clearly the precise receptor alignment which is seen

in careful histology, and is presumably reflected in the

normal SCE, may be altered or disrupted by the action of

pathological or physiological challenges to the retina or

pigmented epithelium. Thus, Fankhauser et al. (1961) docu-

mented anomalies of the SCE in patients with retinal detach-

ment, angiomatosis of the retina and retinoschisis. In cases

in which photocoagulations were performed, marked alterations

were observed in postoperative SCE functions. Abnormal SCE

functions have been noted in other types of retinal pathology

as well (Fankhauser and Enoch, 1962; Dunnewold, 1964; Enoch

et al., 1973; Pokorny et al., 1977; Smith et al., 1977;

Campos et al., 1978; Fitzgerald et al., 1978). However,

not all retinal disease, nor even all affecting the photo-

receptor and pigment epithelial layers, disturbs retinal re-

ceptor alignment. For example, an essentially normal SCE

function has been measured in a case of Best's disease

(Benson et al., 1975) in which a large egg yolk-like lesion

apparently lifts the retina from the pigment epithelium.

What is more instructive from cases of pathology is

that following disturbance, a realignment of photoreceptors,

as reflected in the recovery of the SCE function, may occur

in tandem with the resolution of the pathological condition.

The first indications that recovery was possible were seen

in a case of retinal degeneration in which photocoagulation

was performed (Fankhauser et al., 1961) and,more clearly,in

a case of serious retinal detachment (Fankhauser and Enoch,

1962). Virtually complete recovery of SCE functions has

been documented in cases of total retinal detachment (Enoch

et al., 1973) and of subretinal fluid subsequent to trauma

(Campos et al., 1978). In the latter case, there was evidence

that receptor reorientation occurred within the macular region

at the same time that the SCE function of a paramacular loca-

tion, apparently outside the region of the primary lesion,

showed deteriorative changes. Subsequently, this paramacular

test location also showed the recovery of a normal SCE func-

tion. Marked improvement of a severely disturbed SCE func-

tion has recently been seen in a case of senile macular de-

generation as well (Fitzgerald et al., 1978).

Dunnewold (1964) reported a patient with an iris co-

loboma, resulting in a displaced pupil, in which the SCE

function peaked near the center of the displaced rather than

physiologic pupil. Bonds and MacLeod (1978) have documented

a similar case, in which a displaced pupil resulted from

insult to the eye; both photopic and scotopic SCE curves

were symmetrical about a point near the displaced pupil

center. It is possible that both of these instances also

represent a "recovery" of receptor orientation to abnormal

situations of retinal illumination.

It is significant that in eyes in which anomalous SCE

functions have been demonstrated to accompany observable

retinal pathology, other visual functions, including visual

acuity, have also shown adverse changes (Fankhauser and Enoch,

1962; Enoch at el., 1973; Campos et al., 1978; Fitzgerald

et al., 1978). In cases in which the SCE function has been

found to recover, visual acuity has also shown improvement.*

Ohzu et al. (1972) and Ohzu and Enoch (1972) observed

that the modulation transfer function, indicative of the

optical resolution capability, of isolated retinas is

markedly inferior in areas in which receptors are poorly

oriented as compared with regions of well oriented receptors.

Such observations have been made in rat retinas and in squir-

rel monkey and human foveas. Unfortunately, a characteriza-

tion of the extent of malorientation of receptors in poorly

oriented areas could not be specified quantitatively. This

is a very complex problem.

Campbell (1958) reported psychophysical evidence for de-

creased resolution of grating targets entering the pupil at

peripheral locations and presumably imaged obliquely onto

It is clear that visual acuity changes in these cases
may also represent other aspects of the pathological process
than the inferred changes in receptor orientation.

foveal photoreceptors. Resolution was poorer for targets

oriented perpendicularly to the pupillary test meridian than

for targets oriented parallel to it. Later work indicated

that much of the effect could be attributed to aberrations

suffered along peripheral optical paths in the eye (Campbell

and Gregory, 1960; Green, 1967; Enoch, 1971). However, a

small effect was found to persist even with interferometri-

cally formed targets (Campbell and Gregory, 1960; but also

Green, 1967). Enoch and Glismann(1966; Enoch, 1971) saw a

reduced resolution capability in rat and monkey isolated

retinas for changes in angle of the incident light. Both

meridional and non-meridional effects seemed to contribute

to the measured resolution decrement. The change in observed

resolution for an 80 shift in angle of incidence was on the

order of 10 to 20 per cent, much smaller than the resolution

losses reported by Ohzu ot al. (1972; Ohzu and Enoch, 1972)

in areas of poor receptor orientation. However, the latter

represented alterations in alignment in excess of 8.

SCE anomalies have been demonstrated in some amblyopic

eyes, from which gross pathology of the retina can be ex-

cluded [Enoch, 1957; 1959a, b; 1967a; Dunnewold, 1964; Mar-

shall and Flom, 1970; Bedell, 1974 (reviewed in Chapter III)].

Photoreceptor malorientation, which may be inferred from SCE

abnormalities, has been suggested to play a part in the de-

creased visual acuity of amblyopic eyes (Enoch, 1957; 1967a).

However, the reduction in resolution which could be expected

to result from photoreceptor orientation anomalies is modest

in terms of that found in profound amblyopias.

It is uncertain whether the receptor orientation anoma-

lies, which have sometimes been detected in amblyopes, repre-

sent a sign of microscopic pathology in the outer retina, a

failure of the normal receptor alignment mechanism, or the

sequelae of transient retinal pathology during infancy, which

has otherwise resolved (Burian and von Noorden, 1974). To

confound the situation even further, SCE functions which are

significantly decentered with respect to the pupil are oc-

casionally reported for presumably normal eyes (Flamant and

Stiles, 1948; Westheimer, 1968; Wijngaard and van Kruys-

bergen, 1975). Based upon Enoch and Glismann'sresults,

significant resolution losses would not be expected for a

modest "simple tilt" of the receptors, i.e., the maintenance

of good alignment between neighboring receptors but an over-

all orientation tendency toward a noncentral region of the

entrance pupil. The histological evidence presented above,

the results in retinal pathology, and microwave simulation

studies (Enoch, 1960), all suggest that a general malorienta-

tion of the receptors, a loss of alignment between neighbor-

ing elements, would more significantly disturb resolution


The Transient Stiles-Crawford Effect
and Directional Light Adaptation

Makous (1968) reported that if two fields, which had

been equated for the SCE and which entered the eye at oppo-

site sides of the pupil, were interchanged, this exchange

was marked by a sudden increase in brightness, followed by

a slow decline to a lower steady state value. When these

two fields were alternately employed as the surround upon

which a small incremental test field appeared, a sudden rise

in the increment field threshold, on the order of 0.7 log

units, was found to occur when the surround beams were ex-

changed. Thereafter, the increment threshold returned to

an asymptotic value with an apparently exponential course

and a half time of 10 25 sec, depending upon the surround

field luminance. Curiously, the recovery was slower for

dimmer background fields. The increment threshold results

seemed to mirror the subjective perception in all respects.

Confirmation of this so-called transient SCE appeared

in the work of Heath (1970),Bailey (1974), and Sansbury et

al. (1974). Heath noted that fields compensated for the

SCE and entering the pupil at disparate points demonstrated

flicker when interchanged at a moderate rate. Bailey pur-

sued this finding in his dissertation and was able to define

directional sensitivity functions much narrower than the

standard psychophysical SCE function, utilizing the criterion

of critical flicker frequency. Sansbury et al. replicated

Makous' original experiment using monochromatic, rather than

white, test and surround fields and confirmed his report in

all particulars. They also demonstrated, as had Makous,

Makous (1977) has recently reported that a transient
reduction in visual resolution capability occurs under simi-
lar experimental conditions, lie reports the magnitude of
this reduction to be about a factor of two.

that both the magnitude and time course of the transient SCE

were independent of the pupil entry position of the incre-

mental test beam. Finally, they showed that in the steady

state condition, the threshold raising properties of surround

fields entering at opposite sides of the pupil were linearly

additive. Both Makous and Sansbury et al. argued on the

basis of their experiments that the transient SCE indicated

the existence of receptors or channels with a directional

sensitivity smaller than the aperture of the dilated pupil.

However, they concluded that the effect was not the result

of a preferential light adaptation of different groups of

receptors by background beams incident at the retina at dif-

ferent angles. Coble and Rushton (1971) measured the frac-

tion of cone pigment bleached in the fovea densitometrically,

using a measuring beam which was varied in its pupil entry

position, and also concluded that bleaching did not occur

differentially in groups of cones with dissimilar orienta-

tions. While MacLeod (1974) also failed to find psychophysi-

cal evidence of directionally selective light adaptation at

the fovea, lie did elicit such an effect at 6 in the para-

fovea. In contrast to the foveal results, he found that the

entire SCE function shifted slightly depending upon the entry

position of a background beam in the pupil; for peripheral

pupil entry positions, the difference between the curves was

approximately 0.3 log units.

Both the transient SCE and the apparent entoptic visuali-

zation of receptor subgroups (see above) seem to indicate

the existence of at least modest variation in receptor ori-

entation. This implies that the directional sensitivities

of individual receptors might well be narrower than the SCE

curve, also indicated by microspectrophotometric analyses

of single receptors in vitro (see Enoch, 1975). However,

there is as yet no adequate explanation for the transient

SCE. Makous noted that the transient SCE was similar in

time course, magnitude and the effects of varying luminance

to a phenomenon described by Baker (1949) in conjunction

with rapid light adaptation. It would seem that contrary

to Makous' argument that the two phenomena are unrelated,

sufficient similarity exists to actively pursue a connection.

In particular, both effects may reflect processes initiated

by a transient signal from (groups of) photoreceptors but

which occur at a more proximal site.

Stiles-Crawford-like Effects in Infrahuman Species

There have been relatively few attempts to identify

Stiles-Crawford-like effects in animal models. The existing

data are in the form of retinal electrophysiological responses

to different angles of incidence of a test illumination.

Donner and Rushton (1959) recorded ganglion cell re-

sponses to a flashed incremental field presented against a

steady adapting background from a frog eyecup preparation.

Ganglion cell threshold was found as the angle of incidence

of the incremental beam at the retina was varied. The authors

were careful to direct the test field image onto a portion

of the ganglion cell receptive field remote from the record-

ing electrode, in order to avoid potential shadowing arti-

facts. The test illumination, necessary to produce a cri-

terion ganglion cell response, changed 0.4 0.8 log units

over a 150 range of angles of incidence when the preparation

was photopically adapted. Scotopically, changes in the angle

of test beam incidence had no measurable effect upon ganglion

cell threshold.

Tobey et al. (1975) report similar unpublished experi-

ments by Reynauld and Laviolette, in which a directional

sensitivity was found under photopic conditions for goldfish

retina, as assessed by ganglion cell responses.

Baylor and Fettiplace (1975), using the turtle eyecup,

recorded slow potentials intracellularly from photoreceptors

while changing the angle of the incident light. The authors

report clear sensitivity changes in "red" and "green" cones

as a function of the angle of incidence and possibly smaller

effects in "blue" cones and rods, which were encountered

less frequently. Pautler (1967) had earlier found evidence

for directional sensitivity of turtle receptors in an iso-

lated retina preparation. He noted changes in the magnitude

of the S-potential when the angle of incident light was varied.

Stiles-Crawford-like effects in this preparation are

modified by the presence of high refractive index, absorbing

oil droplets between cone inner and center segments. These

have differential absorption of incident light depending upon

its angle, due to changes in path length through the oil

droplets. However, the oil droplet absorption should tend

to reduce rather than enhance measured Stiles-Crawford-like

effects in turtle cones. Baylor and Fettiplace report that

the magnitude of the directional sensitivity effect, for a

given cell, depends upon the wavelength of the incident light,

directional sensitivity becoming less or "inverting" (i.e.,

a greater sensitivity to presumed off-axis light) for wave-

lengths away from the cell's sensitivity maximum. These

curious wavelength effects may be largely or wholly attribut-

able to the directionality of the oil droplets, that is,

their lesser absorption of obliquely incident light.

Baylor and Fettiplace conclude that the directional sen-

sitivity of individual turtle receptors is fairly broad, cor-

responding well with the angular subtense of the turtle eye

pupil at the retina. This conclusion is in contrast to op-

tically determined directionalities of single photoreceptors

in other species. Enoch and coworkers have determined the

angular radiation patterns of single goldfish, frog, and rat

photoreceptors when retroilluminated in isolated retina

preparations (e.g., Tobey et al., 1975). On the basis of

Helmholtz's theorem of optical reciprocity, i.e., the equiva-

lence of forward and backward passage of radiant energy

through an optical system, Enoch has argued that the angular

radiation pattern specifies the angular acceptance pattern,

or the optical directional sensitivity. Directional sensitivi-

ties of single receptors, derived in this way, are considerably

narrower than either the angular subtense of the pupil or of

the optical directional sensitivities of large groups of

receptors. Differences between rods and cones are quantita-

tive rather than qualitative. Enoch (1961a; 1967b) has also

made similar observations when light is passed through re-

ceptors in the physiologic direction, i.e., from inner retina

to outer retina. That is, varying the angle of incidence of

a light beam on the retina by only a few degrees greatly

alters the transmissivity of individual rods and cones as ob-

served in numerous mammalian species, including humans. It

should be stressed that these optical measurements and ob-

servations have been made in species without receptor oil


Electrophysiological determinations of Stiles-Crawford-

like effects in animals are potentially a very powerful tool

for understanding basic receptor directional sensitivity

processes. Thus, electrophysiologically determined direc-

tional sensitivities of single receptors and of ganglion cells

in the same preparation might indicate the role of variations

of individual receptor alignment in the psychophysical SCE.

The use of preparations with noncircular pupils, measuring

directional sensitivities for two or more pupillary axes,

could suggest mechanisms involved in maintaining receptor

alignment. The difficulties involved in such studies are

also tremendous, however. In an eyecup or isolated retina

preparation, specification of receptor alignment relative to

the eye's pupil is difficult. The effects of the electrode

itself upon the stimulating light are uncertain. This is

an especially difficult problem when intrareceptor recordings

are attempted since the mere presence of the recording elec-

trode in proximity to the photoreceptor must perturb light

propagation along the receptor. Obviously, electrode pene-

tration can also influence cellular orientation. It is not

clear what kind of controls are possible to avoid potential

artifacts from such sources. However, the direct observa-

tion of the optical properties of the receptors studied is



The first important theoretical issue concerning the

SCE was whether it is primarily retinal in origin or an ef-

fect of preretinal absorption. Stiles and Crawford's (1933)

own calculations indicated that estimated preretinal absorp-

tion (and reflection) differences between beams passing

through the center and periphery of the pupil were small

compared to the magnitude of the psychophysically measured

effect. Experiments by Craik (1940) and Goldmann (1942) fur-

ther indicated that the phenomenon was probably retinal in

origin. Weale's (1961) analysis of the effect of lens ab-

sorption indicated that this, in fact, reduced the magnitude

of the psychophysical SCE. Differences between normal pho-

topic and scotopic SCE functions (Crawford, 1937; VanLoo and

Enoch, 1975) and changes seen over time in the SCE of ob-

servers with retinal pathology (Fankhauser et al., 1961;

Fankhauser and Enoch, 1962; Enoch et al., 1973; Campos et al.,

1978; Fitzgerald et al., 1978) also indicate a retinal basis.

Wright and Nelson (1936) proposed, virtually in passing,

that the directional sensitivity of the retina might derive

from the light trapping properties of retinal receptors.

That is, light incident upon receptors normally to their

long axis, or at very shallow angles to the normal, might be

expected to suffer total internal reflection and hence re-

main within the receptor for its entire length. On the

other hand, light incident at more oblique angles to the

normal would no longer be contained by internal reflection

but rather be refracted out of the receptor into the inter-

cellular space. Rays leaving the receptor would, of course,

have a lower probability of encountering, and therefore ex-

citing, visual pigment rendering these rays visually less

efficacious. The same concept had been proposed by Bricke

(cited in Helmholtz, 1924) some 90 years earlier.

O'Brien (1946, 1951) further developed the concept of

total internal reflection within receptors as a possible

basis for the SCE, stressing the importance of the ellipsoid,

coupling the inner and outer segments, and its taper angle.

Within his framework, light losses from receptors occurred

primarily within or near the ellipsoid region. O'Brien's

model brought attention to the possibility that photoreceptors

might concentrate light from the larger diameter inner seg-

ments into the narrower outer segments. In fact, light col-

lection has been directly observed in retinal photoreceptors

(Tansley and Johnson, 1956; Enoch, 1961a). Psychophysical

evidence, which suggests that human receptors concentrate

light, has been provided by Brindley and Rushton (1959) and

Brindley (1959).

Toraldo di Francia (1949) pointed out that the dimen-

sions of retinal receptors are on the order of the wavelengths

of visible light; hence, receptor optical characteristics

could not be adequately specified by the geometrical opti-

cal approaches such as Wright and Nelson and O'Brien had

proposed. Toraldo drew an analogy between photoreceptors

and dielectric antennae, suggesting not only the application

of physical wave optics but also the usefulness of the opti-

cal reciprocity theorem, i.e., that angular radiation and

acceptance patterns of receptors are identical.

Further analyses of the SCE have generally begun with

the presumption that receptors act as optical waveguides.

Waveguide properties are ascribed on the basis of the dimen-

sions of receptors, which in cross section are on the order

of wavelengths of light, and on the higher refractive indices

measured for receptors than for surrounding extracellular ma-

terial (Barer, 1957; Sidman, 1957; Enoch and Tobey, 1978).

Energy passes along waveguide structures in characteris-

tic interference patterns or modes, which depend upon the

receptor geometry (morphology), size and refractive index

relative to the surrounding material as well as upon the wave-

length and angle of incidence of the exciting light (Enoch,

1961a, b, c; Enoch and Horowitz, 1974). Such modal patterns

have been observed in the receptors of many vertebrate spe-

cies (Hannover, 1843, cited in Enoch and Tobey, 1973; Enoch,

1961a, b, c; MacNichol, 1967; Enoch and Tobey, 1973; Tobey

et al., 1975). Moreover, observed modal patterns may be

altered by a change of wavelength or of the angle of inci-

dence of the exciting light (Enoch, 1961a, b, c; Enoch and

Tobey, 1978). The modal patterns seen are those which would

be predicted on the basis of photoreceptor dimensions and in-

dices of refraction.

Waveguides evidence directionality in the sense that

energy incident at some angles is accepted preferentially

over that at others. As noted above, optical studies of

vertebrate receptors indicate a high degree of directionality,

both in rods and in cones. Lesser directionality is found

for groups of receptors than for single elements (Enoch,

1975). Waveguides also tend to concentrate light, in that

their effective light capturing area is typically larger

than their geometrical cross sectional dimension (Snyder

and Hammer, 1972). Psychophysically determined retinal di-

rectionality is thus seen to be in good qualitative agree-

ment with the optical waveguiding properties of retinal re-

ceptors based upon their physical structure.

Snyder and Pask (1973, also Pask and Snyder, 1975)

and Wijngaard and van Kruysbergen (1975, also Wijngaard et

al., 1974) have proposed models of the SCE based on physical

wave optical treatments. In such analyses, receptor dimen-

sions, refractive indices' of the receptors, of their subre-

gions, and of the extracellular interstitial matrix and as-

sociated structures are important parameters. These models

require specification of the physical parameters noted above

with high accuracy as well as simplifying assumptions as to

receptor geometry and homogeneity. Unfortunately, the physi-

cal measurements made to date of photoreceptors and their

natural retinal milieu fail to reach the accuracy demanded,

essentially because of limitations inherent in available

measurement techniques (but see Barer, 1957; Sidman, 1957;

Enoch and Tobey, 1978). This makes an evaluation of such

models difficult, since all of them can be made to reason-

ably fit the psychophysical data by suitable adjustment of


An important issue for all models of the SCE is the

extent to which the psychophysical function represents the

directionality of single receptors, rather than an averaged

directional tendency of a population of small acceptance

angle receptors with interreceptor scatter in orientation

(Crawford, 1937; Safir and Hyams, 1969; Safir et al., 1971).

Present evidence seems to favor the latter view, since (1)

optical specification of acceptance angles of individual

receptors of several species indicates small (20 4') half

angles (Enoch, 1975); (2) apparent local variations in recep-

tor pointing can be demonstrated entoptically (O'Brien, 1950;

Enoch, 1967a; Heath and Walraven, 1970; Enoch et al., 1978);

(3) the transient SCE and directional light adaptation demand

at least a modest nonuniform directional sensitivity at some

level in the visual system [although the nonuniformity might

lie within individual receptors (see King-Smith, 1974;

Sansbury et al., 1974; Crawford, 1972)]; (4) Bailey (1974)

has measured narrow directionality SCE functions using a

critical flicker fusion technique. On the other hand, funds

reflectometry indicates no measurable receptor scatter at

the fovea in terms of bleaching (Coble and Rushton, 1971)

and the single electrophysiological study of the directional

sensitivity of single retinal receptors (which is subject to

the limitations described above) finds broad acceptance angles

which fill the pupil (Baylor and Fettiplace, 1975).

Laties and Enoch (1971) have demonstrated that retinal

histology can be of little help in assessing the small angu-

lar differences in neighboring receptor orientation which

may be involved. Further intrareceptor electrophysiology

and optical studies of single receptors and groups of recep-

tors are important. Psychophysically, an understanding of

the transient SCE might aid in the resolution of this ques-

tion. A note of caution: It is not unreasonable to assume

that, in different species or at different retinal locations

within the same species, contributions of the acceptance

angles of individual receptors, of interreceptor orienta-

tional scatter and of neural summing across groups of recep-

tors to the SCE might differ in importance.

The Stiles-Crawford Effect of the Second Kind (SCE II)

Basic Experimental Data

When the SCE is measured for monochromatic lights, using

a direct photometric matching technique, a comparison field

which enters the pupil displaced from the center may undergo

not only a change in brightness but in perceived hue as well.

This hue shift has come to be known as the Stiles-Crawford

effect of the second kind (SCE II).

The first detailed description of the SCE II was by

Stiles (1937). He presented observers with a bipartite

field which was to be matched for both hue and brightness.*

Standard and comparison half fields derived from two mono-

chromators, the exit slit of the first imaged at the center

of the observer's pupil. The comparison field beam, from

the second monochromator, was imaged successively at half mm

intervals across the width of the dilated pupil. The ob-

server controlled both the luminance and wavelength of the

comparison beam, thereby permitting him to match the standard

half field, which was set at one of 11 wavelengths spanning

the visible spectrum.

The SCE II functions reported by Stiles (change in ap-

parent wavelength vs. pupil entry position of the comparison

beam) are complex in shape. Despite a general resemblance

In measuring the SCE II, a brightness match between
comparison and standard half fields, in order to compensate
for the SCE (of the first kind) is required in order to
avoid contamination of the SCE II by the Bezold-Brhcke hue
effect. The latter is the change in apparent hue which oc-
curs with changes in luminance. In general, as luminance
is decreased, hues shift toward a greener or redder appear-
ance (see, e.g., Boynton and Gordon, 1965). The magnitude
of the Bezold-Bricke effect, for a 1.0 log unit change in
luminance, is comparable to that of the SCE II. The two
effects are dissimilar in many particulars, the respective
hue shifts being in opposite directions in many regions of
the spectrum.

to SCE curves, several of the functions seem not to be sym-

metrical about the peak of the simple SCE function. Since

all lights were essentially monochromatic, lens or other

preretinal absorptions cannot be responsible for these asym-


With a few exceptions in midspectrum, most of the hue

shifts are toward longer apparent wavelengths. This is in

contrast to the Bezold-Brdcke hue shifts which are toward

both shorter or longer wavelenghts in various regions of

the spectrum.

Walraven and Bouman (1960) presented SCE II hue shift

data for comparison beams entering the pupil 3.5 mm from the

SCE peak. Wijngaard and van Kruysbergen (1975) also present

SCE II data for a single observer. Neither study gives any

particulars whatsoever as to measurement techniques; however,

the results agree qualitatively with both Stiles (1937) and

later work by Enoch and Stiles (1961).

Stiles (1937) noted that for most wavelengths, compari-

son and standard fields differed not only in (brightness and)

hue but in saturation as well. In the blue-green region

of the spectrum, a beam passing through the periphery of

the pupil appeared more saturated than one entering at the

center. Enoch and Stiles (1961) thus addressed the problem

of specifying a complete color match (hue, lightness, and

saturation) between fields entering at the center and pe-

riphery of the pupil. Standard and comparison monochromatic

beams, of the same wavelength, were brought to match by the

addition of small amounts of fixed primary lights to one or

both fields, i.e., by trichromatic colorimetry. Standard

fields were adjusted in luminance to an equal luminance

spectrum. Enoch and Stiles' results, specifying the com-

plete SCE II color change, are given in terms of Wright's

u, v, w color coordinate system. From these data, the ap-

parent shift of the spectrum locus for the principle observer

for a beam 3.5 mm from the simple SCE peak could be specified.

The calculated shifts of dominant wavelengths coincide well

in a qualitative way with SCE II hue shifts reported for

other observers. Note that the SCE II color shift is speci-

fied in stimulus terms; an indication of the perceptual color

shift as a function of wavelength might be gotten bya trans-

formation of observer JME's data (Enoch and Stiles, 1961)

from Wright's u, v, w color space to one of the more nearly

uniform chromaticity spaces (Wyszecki and Stiles, 1967).

Brindley (1953) investigated the effects of intense

adapting lights on color matching. In one of his experiments,

he matched a monochromatic yellow with a mixture of red and

green lights, all beams entering at the center of the pupil.

When the pupil entry position of all stimuli was shifted to

3 mm from center, the match no longer held. (Unfortunately,

the spectral characteristics of Brindley's matching lights

were shaped by relatively broad band filters and hence, dif-

ferential lens absorption as a function of wavelength for

the two pupil entry positions may have played some role in

his results.) This phenomenon was, of course, much more

extensively studied in the later Enoch and Stiles work. How-

ever, Brindley further observed that after adaptation to an

intense (104 td) yellow field, an SCE II hue shift was no

longer seen for a monochromatic yellow light, when its entry

position was shifted in the pupil. The observer could still

discriminate actual wavelength changes of the stimulus after

the intense light adaptation, however. Brindley's results

indicate that, at least for a single wavelength, the SCE II

hue shift could largely be nullified by intense light adapta-


Wooten et al. (1977) reported in a preliminary communi-

cation that an SCE effect does persist at high bleaching

levels. The high intensity SCE II apparently differs both

qualitatively and quantitatively from that reported for

lesser retinal illuminances.

The Stiles-Crawford Effect of the Second Kind in Anomalous

Walraven and Leebeck (1962) measured the SCE II hue

shift in single deuteranomalous and protanomalous observers.

The hue shifts of both observers were larger than those ob-

tained in color-normals; this only emphasizes the necessity

of expressing SCE II hue shifts in perceptual rather than

stimulus terms, i.e., the color anomalous' hue discrimination

functions were presumably not as acute as those of color nor-

mals (Wright, 1946). The color anomalous observers reported

increased saturation as well as hue changes for some wave-

lengths entering peripherally in the pupil. The most curious

aspect of the data is that for wavelengths greater than 500 nm

the hue shifts were pronouncedly toward shorter wavelengths

for both anomalous observers, in contrast to the shift toward

the red described by color normals above ca. 560 nm.

Since receptor malorientation has been described in cer-

tain visual pathologies, including some amblyopic observers

(see above), Alpern et al. (1967) looked for a change in

anomaloscope matches in strabismic amblyopic eyes as an in-

dicator of retinal receptor tilt. (Earlier observations of

anomalous SCE functions in some amblyopic eyes had been made

on nonstrabismic amblyopic subjects having limited acuity

decrements.) The rationale was that if cones in his sample

of amblyopic eyes were tilted, one might expect a slight hue

shift toward the red in the amblyopic eyes' anomaloscope

matches, corresponding to the SCE II hue shift for yellow

light. No clear evidence of a hue shift was found; however,

a number of technical objections to the experimental proce-

dure can be made. On the other hand, Pokorny et al. (1977)

not only found disturbed simple SCE curves in several cases

of senile macular degeneration, but also noted abnormal

anomaloscope matches for various matching field sizes in

the same eyes. The authors have suggested that these re-

sults may be understood in terms of an SCE IT-like hue shift

toward longer wavelengths as the result of maloriented re-

ceptors in their subjects. This interesting hypothesis

would seem to require additional verification, such as de-

terminations of SCE II functions in these anomalous eyes.

Stiles-Crawford Effect of the Second Kind: Theory

Very early on, Stiles (1937) entertained the idea that

SCE II changes resulted from different directional sensitivi-

ties of the fundamental cone mechanisms. Although this must

certainly play some role in the SCE II, Stiles' calculations

indicated that another factor or factors must also contribute

to the color changes.

One possibility is contained in the so-called self-screen-

ing hypothesis. Briefly, this is based on the observation

that pigment absorption characteristics change with the effec-

tive density of the pigment in its substrate medium, or alter-

natively, with the path length through a solution of given

pigment concentration (e.g., Walraven and Bouman, 1960).

Stiles, and later Brindley (1953), Walraven and Bouman (1960),

and Enoch and Stiles (1961) noted that light entering the

pupil away from the center would strike well-oriented recep-

tors at an oblique angle and hence, rather than pass down the

whole absorbing length of the outer segment might "leak" out

after traversing only some fraction of this distance. Hence,

for nonsmall pigment densities, the spectral absorption of

such receptors would depend upon the angle at which incident

light strikes the receptors.

Enoch (1961a, b; 1963) reported that the outer segments

of bleached photoreceptors illuminated with white light ap-

peared as a multicolored'mosaic of modal patterns. Changes

of the angle of the incident light not only altered the modal

patterns but also the distribution of hues observed (Enoch,

1961a, 1963). In general, increased obliquity of incident

light resulted in an increased transmission of long and short

wavelengths and lesser transmission of middle wavelengths in

bleached receptors. Enoch's observations indicate that recep-

tor waveguiding properties must be taken into account not

only in the SCE brightness phenomenon but in SCE II color ef-

fects as well.

Walraven and Bouman (1960) proposed a theory of both

the SCE II and the wavelength dependence of the SCE of the

first kind based upon the self-screening hypothesis and geo-

metrical optical considerations. Later Wijngaard et al.

(1974) revised this model in light of the physical optical

characteristics of retinal receptors. Enoch and Stiles (1961)

showed that self-screening did not adequately account for all

aspects of their data, assuming geometrical optics and light

"leakage" from receptors independent of wavelength. It is

not clear that physical optical considerations, as applied

by Wijngaard ct al., provide a superior fit to the data.

While such models appear to be on the right track, current

knowledge of the physical constants involved is insufficient

to permit meaningful evaluation of such modelling attempts.

Significance of the Stiles-Crawford Effect

The SCE or Stiles-Crawford-like effects have been demon-

strated in frog (Donner and Rushton, 1959), goldfish (Rey-

nauld and Laviolette, cited in Tobey et al., 1975), turtle

(Pautler, 1967; Baylor and Fettiplace, 1975) and humans.

Optical studies of isolated retinas of goldfish, frog, rat,

and humans have indicated directional sensitivity of indivi-

dual receptors or of small groups of receptors in these

species (Enoch, 1975). Directionality has shown for both

rod and cones. It thus seems that directional sensitivity

may be ubiquitous in vertebrate photoreceptors.

Laties and coworkers found that photoreceptors at all

positions of the retina tended to align to a common region

at the anterior part of the eye, presumably near the center

of the exit pupil. Such anterior pointing of photoreceptors

has been seen in fish, amphibian, reptile, bird, mammal, and

primate retinas (Laties et al., 1968; Laties, 1969; Laties

and Enoch, 1971; Enoch, 1972; Enoch and Horowitz, 1974; Baylor

and Fettiplace, 1975). The peak of the SCE function of human

observers has been shown to remain near the center of the exit

pupil of the eye for test locations up to 20 in the nasal

visual field and 350 in the temporal visual field (Enoch and

Hope, 1972a, Enoch and Hope, 1973; Bedell and Enoch, 1978).

The psychophysically determined SCE thus apparently reflects

the retina] photoreceptor orientation and complements the his-

tological findings. Additionally, in some species with re-

flecting tapeta, the tapetal reflecting surfaces have been

demonstrated to be oriented approximately perpendicularly to

a hypothetical ray passing through the center of the exit

pupil (Denton and Nicol, 1964; Nicol, 1969; Enoch, 1972).

The rather precise alignment of retinal structures

with what is apparently the exit pupil of the eyes of di-

verse species raises two obvious questions. The first is

teleological: What is the purpose of such alignment? The

second question is how this alignment is established and

maintained. Answers to both must be somewhat speculative.

The SCE has often been suggested to serve a contrast

enhancing function (e.g., Beck, 1950; Enoch, 1972). Whereas

the eye pupil is the source of relevant visual signals, such

signals may be degraded by the presence of contrast reducing

scattered light within the eye. Directionally sensitive

receptors which are aligned toward the region of the exit

pupil of the eye are presumably maximally sensitive to light

entering through the pupil and less sensitive to more ob-

liquely incident light scattered from within the retinal

sphere. In this regard, it has been noted that human pho-

topic and scotopic SCE contours correspond to the relatively

constricted and dilated pupils typical of the respective adap-

tive conditions (e.g., Stiles, 1962). Thus, rods are rela-

tively more sensitive than cones to light entering near the

edge of the dilated pupil, a region which can only be a

source of stray light under photopic adaptation. In species

in which the pupil performs an apodizing function, by virtue

of an asymmetric or exotic shape, a correspondence between

Stiles-Crawford-like contours and the pupillary aperture is

more questionable,albeit experimentally testable.

A tendency for photoreceptors at all regions of the

retina to align toward the center of the exit pupil permits

the optimal utilization of the receptors' directional sensi-

tivity properties across the entire retina. Were the recep-

tors perpendicular to the pigment epithelial substrate


throughout the eye and yet directionally sensitive, at reti-

nal locations remote from the posterior pole, sensitivity to

stray light would be enhanced at the expense of sensitivity

to visual signals entering through the pupil. Moreover,

photoreceptor alignment toward the exit pupil of the eye

presents the greatest density and path length of photosensi-

tive molecules, aligned in transverse stacks of outer segment

discs, to light entering the eye pupil.

In terms of reducing sensitivity to intraocular stray

light, the effect of the lens upon the SCE may also be con-

sidered. Lens pigments tend to reduce the magnitude of the

SCE measured psychophysically with respect to the presumed

underlying directional sensitivity at the retina. Thus, the

retinal sensitivity to intraocular stray light is actually

less than that indicated by the magnitude of the psychophysi-

cal SCE. The greater magnitude of the SCE in the blue and

red regions of the spectrum may be understood in terms of a

further reduction of sensitivity to the stray light within

the eye. Intraocular scattered light is presumably greater

at short wavelengths, due to Rayleigh scattering, and at

long wavelengths, as the result of reflections from vascular

tissues, than at middle wavelengths.

As recognized by Brbcke (cited in Helmholtz, 1924), the

orientation of tapetal surfaces perpendicularly to rays enter-

ing the center of the exit pupil of the eye tends to pass

reflected light from the tapetum through the same photorecep-

tors that were traversed during the forward passage through

the retina, potentially aiding resolution. Additionally,

light which is not absorbed within the receptors after tapetal

reflection has been redirected toward the pupil, and hence

tends to leave the eye rather than increase stray light

levels within the eye.

An obvious consequence of the SCE is that the effective

stimulus for visual response cannot be simply specified in

terms of retinal illuminance. The distinction between distal

and proximal stimuli (Riggs, 1965) becomes rather complicated.

LeGrand (1968) has proposed the effective troland as a unit of

retinal illuminance, in which compensation for the SCE has

been incorporated. Some of the pitfalls inherent in defining

and applying a standard SCE function integrated across the

pupil were considered above. Additionally, markedly differ-

ent SCE corrections are necessary for photopic and scotopic

vision and the change from scotopic to photopic values occurs

over a considerable range of mesopic illuminances, ca. 1-1/2

to 2 log units (Crawford, 1937; Stiles, 1939). It is not

immediately apparent how SCE corrections might be incorporated

into photometry in a standardized, and as yet meaningful, way.

Vos (1960, 1966) has pointed out that an SCE peak which

is decentered in the pupil tends to shift the effective pupil

center, i.e., the location of the center of gravity of all

the light entering the pupil when individual rays have been

weighted for their luminous efficiencies, from the geometric

pupil center toward the location of the SCE peak. In fact,

perfect centration of the SCE is not often found in normal

observers. Vos has argued that individual differences in

the magnitude and the direction of chromostereoscopic ef-

fects may be explained when the disparity between the effec-

tive and geometrical pupillary centers is considered. He

has presented SCE data for several observers which tend, at

least qualitatively, to support these conclusions.

Since visual resolution is more dependent upon the

presence of optical aberrations than upon the small changes

in target brightness which might occur on either side of the

SCE peak, maximal resolution is found for light entering at

the pupil center (Campbell and Gregory, 1960; Enoch, 1971).

Thus, to a limited extent, the effective pupil centers for

brightness and for resolution may differ in position. In one

case of a displaced pupil, the best visual acuity was achieved

for target beam entry 2-1/2 mi! from the SCE peak (Bonds and

MacLeod, 1978). Clearly, such a disparity may be of impor-

tance in the use of subjective alignment to artificial pupils

or to a Maxwellian view system.

The mechanism or mechanisms by which photoreceptor (and

tapetal) orientation are established and apparently maintained

are unknown. Precise photoreceptor orientation has been his-

tologically demonstrated to exist prior to birth in chick and

monkey eyes (Laties and Enoch, 1971; Enoch, 1972). In the

latter instance, the eye was not exposed to light prior to

histological processing of the retina. Demonstrations that

the SCE can recover after disturbance (Fankhauser and Enoch,

1962; Enoch et al., 1973; Campos et al., 1978; Fitzgerald

et al., 1978) and that it apparently "recovers" toward the

center of displaced pupils (Dunnewold, 1964; Bonds and Mac-

Leod, 1978) suggest the operation of a postnatal mechanism

which maintains receptor alignment. The recovery of human

receptor orientation subsequent to pathological disturbance

has been demonstrated to occur in as short a period as three

weeks (Campos et al., 1978). In this regard, the continuous

renewal of photoreceptor outer segments, with a time course

of 9 12 days in rhesus monkey rods (Young, 1976), is pro-

vocative. However, Enoch (1972) has noted that receptor

orientation at retinal locations away from the posterior

pole seems to result from a relative bending of the inner

segments at the external limiting membrane rather than at the

connecting cilium joining the inner and outer segments.

It is seen that the orientation of photoreceptors and

associated structures with respect to the source of relevant

visual stimuli is apparently a common characteristic of ver-

tebrate, and also of a large number of invertebrate (Snyder

and Menzel, 1975) visual organs. On the basis of their

physical properties, the photoreceptors of these diverse

species apparently uniformly exhibit a directional sensitivity

to incident light as well. The SCE is a sensitive psycho-

physical function which, according to current evidence as

presented here, reflects the underlying directionality and

orientation of retinal photoreceptors.

Teleologically speaking, the prevalence of photorecep-

tor orientation and directionality across numerous species


indicates a highly significant role for these specializations

in the visual process. At this time, neither the mechanisms

which establish and maintain directionality and orientation

nor their complete role in vision are well worked out. Be-

cause the orientation and directionality of photoreceptors

must intimately bear upon the primary phases of the neural

visual response, these factors and their significance mustbe

understood in order to meaningfully evaluate their influence

upon subsequent visual processing, both in normal and in

pathological conditions.


The nature of the anatomical and physiological changes

which occur in functional amblyopia are as yet unknown. The

search for such changes is hampered by the high probability

that amblyopia is not a unitary syndrome with a single under-

lying pathophysiology. On the contrary, anomalies at any

number of sites within the visual system might give rise to

a decreased acuity.

In the current literature, interest focuses on the

dorsal portion of the lateral geniculate nucleus (LGN) and

even more so, on the primary visual cortices. Wiesel and

Hubel (1963a, b) demonstrated morphological changes in

dorsal LGN neurons and alterations of ocular driving pat-

terns of visual cortical neurons in monolaterally sutured

kittens. More recently, functional neuronal connections

have been shown to be altered in monkey striate cortex fol-

lowing unilateral eyelid suture or experimentally induced

strabismus or anisometropia (Hubel and Wiesel, 1977; Baker

et al., 1974; von Noorden and Crawford, 1977).

See footnote, Chapter I.

Suppression of the amblyopic eye during nonmonocular

("binocular") viewing is a common clinical and psychophysical

finding in apparent accord with the cortical neurophysiologi-

cal data. However, there are indications that changes may

occur more distally in the amblyopic visual system as well.

For example, amblyopic eyes are reported to evidence an ex-

cessive areal summation (or diminished inhibition), a func-

tion generally ascribed primarily to the outer retina (Flynn,

1967; Danis and Meur, 1967; Lawill et al., 1973). Ikeda and

Wright (1974, 1976) measured electrophysiological contrast

sensitivity functions for single neurons in the kitten LGN;

responses presumably indicated contrast sensitivities of

retinal ganglion cells. Neurons driven from the deviated

eye of experimentally esotropic kittens showed losses of

contrast sensitivity, especially sustained units (X cells)

with receptive fields in or near the area centralis. These

data not only implicate the retina as at least one of the

sites of anomaly in this animal model of strabismic amblyo-

pia, but also conform to the typical clinical and experi-

mental finding of a deficit primarily localized to central

vision in amblyopic eyes. That is, monocularly measured

functions, including visual acuity, typically indicate the

greatest disturbance at the center of the amblyopic eye

visual field and approach (or attain) normal levels of func-

tioning in the near periphery (Meur and Conreur, 1968; Kandel

and Bedell, 1973; Kirschen, 1977).

It is the author's working hypothesis that functional

changes in amblyopia may occur at virtually any level of the

visual system. Because amblyopia is probably not unitary

in its pathophysiology, discrete subgroupings may be identi-

fiable, based upon the site or sites of the primary lesion

or the evolution of the pathological process during develop-

ment. Since workers have seemed reluctant to recognize the

possibility of functional subgroupings of amblyopia (but see

Burian and von Noorden, 1974), there is little indication

in the present literature as to what sort of divisions may

be useful. Currently, functional amblyopias are distinguished

on the basis of their inferred etiology, e.g., strabismic,

anisometropic, deprivation, etc.

Information processing in the human visual system is

primarily centrifugal, at least in its early stages. It is

apparent that psychophysically or neurophysiologically demon-

strated abnormalities, which are sampled at proximal sites

within human amblyopic or experimentally induced amblyopic

visual systems, might reflect either anomalies at that level

of processing, or anomalous input from more distal sites,

or both. It therefore seems reasonable to approach the

pathophysiology of atmblyopia in a disto-proximal direction,

For example, Sherman and coworkers (Sherman et al.,
1972; Norton et a]., 1977) reported selective loss of transi-
ent (Y) cells and apparent integrity of the X cell system in
an animal model of form deprivation amblyopia. These re-
sults are in contrast to those of Ikeda and Wright, who used
an animal model of strabismic amblyopia (see above).

i.e., to ascertain the status of the information passed along

to subsequent stations from each "processing" center in the

amblyopic visual system.

Advances have already been made in this line of attack.

In post hoc analyses of a limited number of cases, the opti-

cal transfer properties (imaging capability) of amblyopic

eye media were found not to be impaired relative to normal

eyes (Fankhauser and R5hler, 1967; also Burian, 1967a). It

is clear that this statement niust be qualified in cases of

amblyopia with anisometropia.

On the other hand, SCE function measurements have indi-

cated the existence of anomalies in some amblyopic eyes at

the next stage of the visual system, namely at the level of

the photoreceptors. Based upon the evidence presented in

the previous chapter, it is probably reasonable to suggest

that psychophysically determined SCE functions represent

(1) the directionality of individual receptors and their re-

lated structures, (2) some distributive orientation factor

between receptors and between groups of receptors, and (3)

possible neural integrative processes by which the outputs

of groups of receptors are combined. Additionally, one may

posit perceptual, criterion and/or judgmental factors, in-

herent in the nature of psychophysical measurements, which

might derive from possible differences in the appearance of

stimuli entering the pupil at different locations. It is

likely that the contribution of some or all of these factors,

as well as their relative importance, changes with position

on the retina and perhaps with other observer or stimulus

variables as well. Such changes, should they occur, presum-

ably modify the shape of and/or the directionality represented

in psychophysically determined SCE functions. However, the

location of the peak of the SCE function would seem to be a

valid indicator of the overall alignment tendency of a group

of retinal photoreceptors with respect to the pupil. SCE

functions which show no clear peak, or multiple subpeaks, are

assumed to indicate a "general malorientation" of the photo-

receptors with respect to one another, i.e., the disruption

or lack of a well-defined alignment tendency within the

group of receptors sampled.

The SCE function has been found to be disturbed in some,

but not all, amblyopic eyes. Enoch (1957, 1959a, b) found

clearly anomalous SCE functions at the locus of fixation in

two of six amblyopic observers tested. In these two observers,

amblyopic eye SCE functions demonstrated marked departures

from the typical paraboloid shape, appearing instead rather

flattened and asymmetrical. Similarly disturbed SCE func-

tions have been seen in cases of active retinal pathology,

in which visual acuities have also been found to be reduced.

It is clear, however, that the reduced visual acuities in

these cases of retinal pathology might derive from factors

in addition to inferred receptor malorientation. Moreover,

the anomalous SCE functions determined for these amblyopic

eyes and those determined in cases of observable retinal

pathology cannot be assumed to have a common etiology. In

three other observers of Enoch's sample, amblyopic eye SCE

functions had maxima, estimated from the positions of the

peaks of horizontal and vertical pupillary traverses, dis-

placed from the pupil center by between 1-1/2 and 2-1/4 mm.

The nonamblyopic eyes of these observers all demonstrated

normal appearing SCE functions, with peaks within 1 mm of

the pupil center. For one amblyopic observer, normal ap-

pearing SCE functions were measured in both eyes.

Dunnewold (1964) measured SCE contours at the point of

fixation in both eyes of each of two mildly amblyopic ob-

servers. He employed a clinical testing instrument (Vos

and Huigen, 1962) without a biteplate and using a subjective

alignment procedure. These factors, as well as his failure

to show raw data or indicate error variance, render his re-

sults somewhat suspect. In one of his cases, the SCE func-

tion peak was displaced approximately 3 mmn from the pupil

center; however, the maximum of the SCE function of the non-

amblyopic eye was itself 2 mm displaced. Other than decen-

tration of the peaks, the functions of both eyes appeared

normal. Dunnewold's second case was found to have normal

SCE functions and centered peaks in both eyes.

Marshall and Flom (1970) measured SCE functions, in the

horizontal meridian only, in four moderate and severe am-

blyopes. In two of these observers, both of whom fixated

extrafoveally under monodular viewing conditions, SCE func-

tion peaks, determined at the locus of fixation, were sub-

substantially displaced from the pupil centers. The SCE

functions of the nonamblyopic eyes all had peaks near the

pupil centers and were normal in appearance. One of the ec-

centric fixators subsequently recovered a centric fixation

in the amblyopic eye. At that time, a more centered SCE

function was measured in this eye. Marshall and Flom, ap-

parently assuming that photoreceptors align toward the center

of the globe, rather than toward the exit pupil of the eye,

argued that decentered SCE functions were to be expected in

cases of eccentric fixation and might be an indicator of

such noncentric fixation rather than of tilted foveal photo-

receptors. As reviewed in Chapter II, recent psychophysical

studies of normal eyes indicate that receptors at both cen-

tral and peripheral retinal locations tend to align toward

the eye pupil, contrary to the assumption of Marshall and


Bedell (1974) determined SCE functions in a single

severe strabismic amblyopic observer, with a large eccentric

fixation, at the locus of fixation and at several points

nearby. All curves showed slightly displaced peaks and a

lack of symmetry about the maximum value. Steadiness of

fixation, which is of considerable concern when measuring

SCE functions for strabismic amblyopic observers, was evi-

dent from the small variability of SCE function measurements

and from the consistency of the data obtained over several

months. SCE functions in the nonamblyopic eye were nearly


Alpern et al. (1967) attempted to detect anomalies of

the SCE functions of three strabismic amblyopic observers

utilizing the SCE II effect (see Chapter II). Negative re-

sults were reported for all three observers. However, de-

tails of the testing procedure and certain questionable as-

sumptions which the authors made regarding the SCE II make

interpretation of this ingenious experiment difficult.

SCE functions with peaks displaced from the pupil center

have also occasionally been measured for presumably normal

observers (Flamant and Stiles, 1948; Westheimer, 1968;

Wijngaard and van Kruysbergen, 1975). The literature cited

above suggests that anomalous or displaced SCE functions may

more often be found in amblyopic than in nonamblyopic eyes.

Unfortunately, the samples of both amblyopic and nonamblyopic

eyes are rather small. Moreover, observer alignment and cen-

tration within the testing instrument was not adequately

monitored in all of these studies. It is presently not clear

to what extent visual acuities are affected in normal eyes

having displaced SCE function peaks. Evidence which bears

upon this question is presented in Chapter VII.

On logical grounds, disturbed receptor orientation

could degrade visual resolution in several ways (Enoch,

1957; 1959a, b; 1967a): (1) a decrease in brightness of the

resolution target because of the SCE, (2) a locally effec-

tive meridional increase of the retinal receptor mosaic di-

mension for sheered over receptors, (3) light leakage or

cross-talk between neighboring receptors (Enoch, 1960),

(4) reduced contrast because of increased scattering from

the receptors, and (5) increased sensitivity to stray light

within the eye, also reducing contrast. As discussed in the

last chapter, in vitro studies of the optical properties of

both animal and human receptors confirm that poorly oriented

receptors are poorer light collectors and have poorer opti-

cal transfer capabilities than well oriented receptors. The

evidence (see Chapter II) suggests that a modest amount of

"tilt" of the receptors, in which orientation within groups

of receptors is maintained, but the overall alignment tendency

is toward a noncentral region of the exit pupil, should en-

tail a lesser degradation of retinal resolution capability

than malorientation in which there is a loss of alignment

between neighboring receptors.

However, hypothesized photoreceptor alignment anomalies

within amblyopic eyes might contribute to amblyopia other

than by a direct degradation of resolution. For example,

some amblyopic subjects report that stimuli viewed with the

amblyopic eye appear less bright than when viewed with the

fellow, nonamblyopic eye (Grosvenor, 1957; Burian, 1967b;

Flynn et al., 1971; Bedell, 1974). These results are quali-

tatively consistent with the presence of certain forms of

anomalous receptor alignment. In particular, alignment of

the receptors toward a significantly decentered region of the

dilated pupil would be expected to cause a reduction of the
apparent brightness of stimuli viewed with the natural pupil.

Such a mechanism could be implicated in the suppression
of the amblyopic eye under conditions of "binocular" viewing,
for example.


However, quantitative data for such a relationship are

lacking. Moreover, the reduction of perceived brightness

in these amblyopic eyes might well be attributable to changes

at a site or sites other than at the photoreceptor.


The receptor alignment anomalies, inferred from SCE

function measurements, which occur in some amblyopic eyes,

represent the first level in these amblyopic visual systems

at which a discrete pathology is indicated. The contribu-

tion of such anomalies to the status of visual functioning

in affected amblyopic eyes therefore warrants extensive

investigation. The extent to which such anomalies occur

in amblyopic eyes as well as in nonamblyopic and presumably

normal eyes also requires clarification.

In the past, the assessment of receptor orientation in

amblyopic eyes has been confined to the fixation area (Enoch,

1957, 1959a, b; Marshall and Flom, 1970; Dunnewold, 1964)

or to a small region around it (Bedell, 1974). Since the

SCE function peaks of normal observers have been shown to

cluster near the center of the pupil when measured up to 350

in the peripheral field, it seemed reasonable to ask whether

the evidence of photoreceptor malorientation found in some

amblyopic eyes at fixation also exists at other retinal loci.

The visual resolution reduction caused by preretinal
media anomalies and pathology is not here categorized as
amblyopia. See footnote, Chapter I.

In particular, since the visual functioning of amblyopic eyes

apparently tends to approach that of normal eyes in the am-

blyopic eye peripheral visual field, it seemed appropriate

to investigate whether receptor orientation conforms to this

same trend. That is, it was sought to determine whether, in

those amblyopic eyes in which receptor orientation anomalies

could be demonstrated, such anomalies were confined to a cen-

tral retinal area in which acuity deficits were found. Al-

ternatively, receptor orientation anomalies might extend be-

yond the region of visual acuity impairment and characterize

such retinae as a whole.

This research, then, sought to utilize the psychophysical

SCE function to characterize the nature of possible receptor

orientation anomalies within the eyes of functional amblyopic

observers at several locations spanning the central and near

peripheral retina. Since the number of investigations which

have examined the SCE function in amblyopic eyes is small,

this research, even though limited in its sample size, was

also expected to contribute to the question of the extent to

which receptor alignment anomalies occur in functional amblyo-

pic eyes exhibiting limited visual decrements.

As discussed in the Introduction, the pattern of inferred

receptor alignment across central and near peripheral retinal

locations in amblyopic eyes might be expected to provide in-

formation concerning the nature of the hypothesized mechanisms

controlling such alignment as well. In order for a precise

control of receptor alignment to occur, a signal (or signals)

must exist, based upon which departures from correct align-

ment are determined. Thus, receptor alignment mechanisms

presumably contain both afferent, error-signal-detecting,

and efferent, receptor-alignment-correcting, components.

Various forms of receptor alignment anomalies might reflect

disturbances of either of these components or of the align-

ment signal itself.

Thus, if amblyopic eyes were identified, in which the

inferred receptor orientation across a wide retinal area

indicated that receptors converged toward an anomalous align-

ment centrum, displaced from the pupil center, then a global

or retina-wide disturbance of alignment mechanisms within

such eyes might be entertained. One might hypothesize that

the alignment signal, or its detection, was altered in such

cases. Were evidence of gross receptor malorientation found

at all sampled retinal locations, a global disturbance of

the alignment mechanism, probably of a different sort, would

be indicated. On the other hand, were evidence of receptor

alignment anomalies found only within circumscribed retinal

regions, and evidence of normal alignment toward the central

area of the exit pupil found elsewhere on the same retina,

then a local disturbance of receptor alignment mechanisms,

and not of the alignment signal, might reasonably be inferred.

Having identified retinal regions or entire eyes in which

receptor orientation, and presumably receptor alignment

mechanisms, were apparently disturbed, then subsequent studies

might attempt to identify electrophysiological, metabolic,


biochemical or other anomalies within such retinal regions

or eyes which correlated with the apparent malfunctioning

of the hypothesized receptor alignment mechanisms.

Thus, one of the aims of this dissertation research was

to identify individuals in whom receptor alignment anomalies

existed and to characterize such anomalies with respect to

location on the retina. Subsequent investigations of the

eyes of such individuals could then reasonably be directed

toward understanding the nature of hypothesized receptor

alignment mechanisms.



Amblyopic and nonamblyopic control observers were so-

licited from the University of Florida campus and J. Hillis

Miller Health Center populations by means of posted advertise-

ments and by word of mouth. None of the observers of this

study were referred directly from clinical sources. Poten-

tial observers were screened for possible amblyopia with a

portion of the clinical evaluation described below (history,

monocular visual acuities for Landolt targets, cross cover

test, entoptic projection of the fovea). Informed consent

was obtained in writing from observers agreeing to partici-

pate in the study. The observers were compensated for time

spent on the project.

One of the control observers (SBS) had participated in

a previous SCE function study (Bedell and Enoch, 1978).

All other observers were naive to visual psychophysical



Both the amblyopic and nonamblyopic observers were

evaluated with the same program of clinical tests. These

included (1) a history, (2) best corrected and pinhole visual

acuities for Landolt targets, (3) refractive status assessed

by retinoscopy, (4) strabismus/binocularity/motility examina-

tion, (5) ocular tensions,and (6) slit lamp and ophthalmo-

scopic evaluation of the eyes.


The following information was obtained from potential

amblyopic observers: (1) age of onset of amblyopia, (2) cir-

cumstances surrounding onset, (3) treatments and attending

physicians, (4) presence of reduced vision or strabismus

along relatives, and (5) present visual status. The last

item inquired as to the observer's subjective impressions

concerning monocular and "binocular" vision under everyday


All observers were asked about any personal or family

history of eye disease, and specifically whether there was

a history of glaucoma or of ocular hypertension. Observers

were asked to give known drug reactions and to describe

their last visit to an eye care professional, including

whether dilation was performed at that time. This portion

of the history was intended to elicit information concern-

ing possible contraindictions for dilation of the eyes.

Dilation was deemed necessary for the SCE function measure-


Visual Acuities

Monocular visual acuities for 8 position double break

Landolt C targets were obtained with best correction. Targets

were presented on a commercial chart (Bausch and Lomb, Roches-

ter, N. Y.) viewed at 20 ft. Chart luminance was 1.93 log

cd/m2 provided by white fluorescent lighting. Target con-
trast was approximately 80 per cent. Acuities were checked

for improvement with a pinhole placed over the observer's


Refractive Status

Visual corrections were initially estimated from lensome-

ter readings of observers' spectacle correction, if any. Ad-

ditionally, retinoscopy was performed on all observers at

the locus of fixation, in most cases by Dr. Jay Enoch.

Dr. Enoch or the author performed retinoscopy for peripheral

visual field test locations in order to estimate the lens

corrections necessary for SCE function testing at these loca-


Strabismus/Binocularity/Motility Examination

In most instances, the strabismus examination was per-

formed by the author. The cross cover test was used to

screen for tropia (misalignment of the visual axes of the

two eyes under "binocular" viewing conditions) on the initial

visit to the laboratory. When a positive result was found,

the prism necessary to neutralize the deviation on the cover

test was determined. Tropias were evaluated both at far and

Chart luminance Target luminance
Contrast Chart luminance + Target luminance

at near, with correction, and in up and down gaze to deter-

mine A or V pattern of any deviation. A 9 gaze position

muscle field screened for incomitant deviations. Lateral

phorias in orthotropic observers were evaluated at both far

and near by finding the prism necessary to vertically align

a target seen in binocular diplopia as the result of a 6

diopter vertical prism placed before one eye.

Retinal correspondence under "binocular" viewing condi-

tions was assessed by the afterimage and striated glasses

tests. In the former, horizontal and vertical line after-

images were formed respectively in the nonamblyopic and am-

blyopic eyes of amblyopic observers, by sequential monocular

inspection of appropriately oriented illuminated slits. There

was no preferential order of afterimage formation for the

control observers. Under "binocular" viewing conditions and

in the absence of monocular eccentric fixation, alignment of

the centers of the horizontal and vertical afterimages, to

form a cross, indicates normal retinal correspondence. Mis-

alignment of the afterimage centers is indicative of an anoma-

lous retinal correspondence.

The striated glasses test was performed with Bagolini

striated lenses (House of Vision, Chicago) placed obliquely

at 450 and at 1350 before the right and left eyes respec-

tively. Intersection of the streaks, formed by the lenses

over each eye, at a small, illuminated fixation target is

indicative of anomalous retinal correspondence in hetero-

tropic observers. Misalignment of the streaks equivalent

to the angle of deviation indicates a normal retinal cor-


The afterimage test was conducted at far and the stri-

ated lenses test at both far and near. Disagreement between

the two tests in heterotropic individuals is not uncommon,

apparently because of the more dissociating (less like normal

seeing conditions) aspects of the afterimage test (Bagolini,


Monocular fixation positions were estimated using one

or more of three entoptic techniques (c.f. Moses, 1970).

These were the location of the Maxwell spot centroid with

respect to a fixation target, the location of the Haidinger

brush with respect to a fixation target and the location of

the avascular zone of the entoptic Purkinje retinal vessel

pattern with respect to a fixation target. All measurements

were converted to units of visual angle.

The Maxwell spot is a pinkish or lighter splotch of

color seen within a purple field (other colors of Maxwell

spots are seen against other color fields). The spot sub-

tends about 2' for most individuals and apparently results

from the selective absorption of short wavelength light by

the macular pigments. In many individuals, substructure is

evident in the spot. Observers viewed the Maxwell spot

monocularly on an alternately purple and neutral (Edmund

Roscolene filter #827 + #846 for purple, #883 for neutral)

fluorescent back-lighted screen of 1.4 log cd/m2, measured

with filters in place. Alternating purple and neutral lights

were employed to avoid fading of the spot. The observer

localized the centroid of the spot with respect to a black

fixation target by superimposing a movable and adjustably-

sized circular cursor light upon the Maxwell spot while fixat-

ing the target (Kandel and Bedell, 1972).

The Haidinger brush is another entoptic phenomenon also

apparently related to the macular pigment and in particular

to its presumed dichroic absorption characteristics. In

bluish light, the brush is seen as a propellor shaped object

through a linear polarizer. If the polarizer is rotated, the

propellor appears to spin. Using a movable cursor light,

the axis of rotation of the propellor may be located with

respect to a fixation target. This test was performed using

the same equipment as for the Maxwell spot test, but substitut-

ing a blue (Wratten #34) filter and rotating linear polarizer

for the alternating purple and neutral filters. Screen lumi-
nance was 0.6 log cd/m under the actual viewing conditions.

The shadows of the retinal blood vessels may become ap-

parent when the angle of the incident light upon the retina

is changed, causing the vessel shadows to move. Since the

foveal region is characterized by the absence of capillary

support from the inner retinal circulation, it appears as a

"hole" in the entoptically viewed retinal vessel pattern.

Purkinje retinal vessel patterns were generated by rotating

an eccentrically located '1.8 mm diameter pinhole in front

of the dilated eye (Boer and Hofstetter, 1972). The observer
viewed a 3.1 log cd/m luminance white screen through the

rotating pinhole. The avascular region of the vessel pattern

was located with respect to a fixation target using a cursor


Vessel patterns could also be generated within the SCE

testing apparatus (see Chapter VI). In this way SCE function

and interferometric visual resolution targets could be as-

sured to be centered within the avascular region.

Binocularity was evaluated using three separate tests.

The first of these was the Titmus vectograph test which

presents a series of graded horizontal disparity targets,

seen in depth by binocularly normal individuals, through

polaroid lenses (Titmus Optical Co., Inc., Petersburg, Va.).

This test has been critized because of the presence of many

monocular cues in the design. Binocularity was also as-

sessed by the ability to appreciate the Pulfrich stereo-

phenomenon, which is the appearance of depth in the orbit

of a swinging pendulum viewed binocularly, but with a neutral

filter over one eye (c.f. Gregory, 1973). Finally, observers

were presented with pairs of random dot stereograms (Julesz,

1971) in a Clement Clarke synoptophore. Targets had a hori-

zontal disparity of approximately 600 sec of arc. Binocularly

normal individuals appreciated a figure or figures either

above or below the plane of the target background when

stereogram half-pairs presented to each eye were brought into

register. No figure is available in either stereogram half-

pair alone.

In addition to the above examinations, observer SSD was

referred to Dr. Matthew Rabinowicz of the Opthalmology Depart-

ment, J. Hillis Miller Health Center for further strabmis-

mological evaluation.

Tension and Slit Lamp, Ophthalmoscope Examination

Intraocular pressures were measured using applanation

tonometry at the time of the slit lamp and ophthalmoscopic

examinations. These examinations were performed by Dr. Con-

stance R. Fitzgerald, of the Department of Ophthalmology,

J. Hillis Miller Health Center. Slit lamp examination as-

sessed the cornea, anterior chamber, lens and anterior vitre-

ous body. In particular, the examination sought small opaci-

ties in the optical media which might interfere with SCE func-

tion measurements. Gonioscopy was performed to evaluate the

depth of the angle between the iris and limbus. An ophthal-

moscopic examination of the funds was performed in order

to rule out observable pathology in amblyopic (and nonamblyo-

pic eyes) as a cause for possible anomalous SCE functions.

Selection Criterion

Amblyopic observers were expected to have a difference

of one line or more in their best corrected, monocular visual

acuities for Landolt ring targets. The acuity deficits were

expected to be of long standing, i.e., dating to childhood,

as revealed by the history. Furthermore, some contributory

history of strabismus, anisometropia or early abnormal visual

experience was anticipated. Anterior chamber and funds

examinations were expected to reveal no abnormalities which

might be responsible for the acuity findings. All amblyopic

observers met these criteria. The results of the clinical

examinations of the amblyopic observers are summarized in

Appendix A.

Control observers were expected to have at least 20/20

best corrected monocular visual acuities for Landolt targets

in each eye and minimal between eye acuity differences. An-

terior chamber and funds examinations were expected to re-

veal no abnormalities. The control observers were also ex-

pected to be orthotropic. Observers SBS and MAP met these

criteria. Observer JEC, who was originally recruited to be

a control observer, revealed possible funds abnormalities

within both eyes on the ophthalmoscopic examination. The

pattern of SCE functions measured in both eyes of this ob-

server also departed from results obtained for other control

observers. Thus, observer JEC will be treated as a special

case. The results of the clinical evaluations of JEC and

of the control observers are presented in Appendix B.


SCE Apparatus and Testing Procedures

The instrument used to determine SCE functions appears

schematically in Fig. 1. A cube beam splitter (BS1) divides

the collimated beam (lens LO) from a tungsten ribbon fila-

ment source (S) into test (A) and surround (B) beam chan-

nels. Within both channels, the ribbon filament is imaged

by lenses L1A and LIB at four times lateral magnification

onto approximately 0.30 mm diameter round apertures (APA,

APB), mounted on mechanical stages, which serve as secondary

sources. A portion of the beam in channel B is diverted by

a pellicle (PL) and falls on a photovoltaic cell (V) the out-

put of which is monitored. The apertures are collimated by

lenses L2A and L2B and channels A and B are rejoined at a

second beam splitting cube (BS2), after passing through ad-

justable field stops (FSA, FSB). After passing through beam

splitter BS3, lens ,3 forms unit lateral magnification images

of apertures APA and APB in the observer's entrance pupil

(EP). Movement of either aperture by means of its mechani-

cal stage mounting causes an equal and opposite movement of

its image in the entrance pupil. The filament images were

found to be homogeneous to 0.10 log unit through 7 mm


P2 8 5T LA
--B -- ---iA




Schematic Diagram of Stiles-Crawford Function Apparatus

horizontally and to less than 0.10 log unit through 8 mm


At BS3, a portion of both test and surround field beams

are deflected and imaged by lens L3' onto a first surface

mirror conjugate with the observer's entrance pupil. Re-

flected images pass backward through lens L3' and join infra-

red radiation reflected from the observer's eye, provided

by tungsten infrared sources (IRS), at cube BS3. Lens L4

forms an image of the test and surround field beams and of

the observer's pupil on a reticle (R), marked in concentric

circles. The reticle is retroilluminated by infrared source

S' reflected in cube beam splitter BS4. The reticle and

images of the entrance pupil and both the test and the sur-

round field beams as they enter the pupil are viewed by the

experimenter (E) in an infrared image converter system

(IRC, RCA #6914A).

Provision for filtering of both test and surround field

beams is made in the collimated portions prior to lenses LIA

and LIB. The test field beam passes through a Kodak neutral

wedge and balance filter (W) and is interrupted twice per

second by an episcotister (EP'). Calibrations of the wedge

and neutral filters were carried out in situ and with chro-

matic filters in place using a model 1980 Spectra Pritchard

photometer (Photo Research, Burbank, Ca.) fitted with a 40X

objective attachment and placed at EP. Calibrations were

verified, also using the Pritchard photometer, by the method

of Westheimer (1966) for Maxwellian view systems.

Field stops FSA and FSB are adjustable toward and away

from lens L3, thereby functioning as Badal optometer systems

with a range of approximately 2.0 diopters. Supplemental

lens corrections, based on retinoscopic examination, could

be placed (at RX) close to the observer's eye. Vertex dis-

tances were carefully measured and corrections centered on

the optical axis of the system. A correction for the spec-

tacle RX was applied to the displacements of the surround

beam in the entrance pupil. For data taken at other than

fixation, the observer's gaze was directed to a dim red col-

limated fixation source (FX). At the locus of fixation, the

centered test array itself served as a fixation target.

The observer was held in position by means of a dental

impression and forehead rest, both attached to a frame which

is adjustable in the x, y, z directions. The experimenter

positioned the observer using these controls, while observ-

ing the image of his entrance pupil upon the reticle in the

IR image converter. Pupil position was monitored continuously

and adjusted during experiments to maintain proper alignment

with respect to the test and surround field beams and the

exit pupil plane of the instrument.

Observers' eyes were dilated with 10 per cent phenyleph-

rine hydrochloride (Neo-Synephrine) or 1 per cent tropicamide

(Mydriacyl) after ruling out ocular pathology. Photopic SCE

functions were determined by an increment threshold procedure,

the test beam being fixed at the pupillary center and the

surround beam displaced in successive steps across the pupillary

aperture. The observer viewed a 0.500 test field limited

by aperture FSA, superimposed upon the center of a larger

(424') surround field defined by FSB. Since both FSA and

FSB are mounted on mechanical stages, aperture FSB could be

shifted to compensate for changes in position of its retinal

image, as the result of occular aberrations, when the sur-

round beam was displaced from the pupillary center. The test

and surround fields were thereby maintained in concentric

alignment for all pupil entry positions of the surround field


Discounting the SCE function itself, background field

luminance was 3.04 log photopic trolands (1100 trolands).

Both test and surround fields were orange (Kodak Wratten

#23A filter). Increment thresholds with both test and sur-

round beams at the pupil center were determined over a 4 log

unit range at each visual field test location for each ob-

server. The increment threshold data indicated that all SCE

tests were conducted on the Weber portion of the increment

threshold curve (see Chapter II, also Enoch and Hope, 1972a,

Appendix 2).

SCE functions were determined at a number of visual

field locations spanning central and near peripheral retinal

positions. The test locations chosen were (1) the locus of

fixation, (2) the fovea, if different from (1), (3) 5 nasal

visual field (NVF), (4) 1i0 NVF, (5) 200 NVF, (6) 50 temporal

visual field (TVF), and (7) 100 TVF. All visual field test-

ing locations were not examined for all observers. For some

observers, SCE functions were determined at test locations

in addition to those listed above.

In order to determine whether testing at the fixation

locus also included the fovea, the Purkinje retinal vessel

pattern was generated within the SCE testing apparatus. In

this way, the entoptically viewed avascular zone of the ves-

sel pattern could be located with respect to the test field

under the actual testing conditions. A 2 diopter prism,

placed between aperture APB and lens L2B was rotated by a

variable speed motor, causing the image of APB to rotate

through a circle of approximately 2 mm radius in the ob-

server's entrance pupil. Observers who could appreciate the

pattern saw a slightly wobbling retinal vessel pattern within

the surround field and noted, when fixating the test field,

whether the avascular zone of the vessel pattern was concen-

tric about the test field. If not, the observer located a

variable position fixation target, produced by back reflec-

tion of an attenuated laser beam from the rear surface of

field stop FSA, in the position which brought the avascular

area of the vessel pattern to surround the test field. Dis-

tance between the fixation light and the center of the test

aperture was determined by micrometer and converted to visual

angle. "Foveal" SCE functions were determined with the ob-

server fixating either this laser fixation target or one of

the movable collimated fixation sources (FX) placed at the

appropriate location.

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs