See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/279643659

Visual Acuity

Article · May 2004

CITATIONS READS

8 16,447

2 authors:

Andrew Jonathan Jackson Ian L Bailey

Belfast Health and Social Care Trust University of California, Berkeley
120 PUBLICATIONS 1,983 CITATIONS 219 PUBLICATIONS 10,599 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Optimal dynamos View project

Investigation of Ocular, Heart and Lung Diseases and Household Fuel Use in Kaski District, Nepal View project

All content following this page was uploaded by Ian L Bailey on 03 December 2015.

The user has requested enhancement of the downloaded file.

Optometry in Practice Vol 5 (2004) 53–70

Visual Acuity
AJ Jackson1 PhD MCOptom and IL Bailey2 OD MS FCOptom FAAO

1
Department of Ophthalmology, Royal Group of Hospitals, Belfast and 2School of Optometry, University of Berkeley,
CA, USA
Accepted for publication 14 January 2004

Almost 150 years after Herman Snellen (Snellen 1862,

1864), the renowned ophthalmologist from the University
of Utrecht, published his table of optotypes for assessing
the visual status of the eye, we continue to utilise optotype

Relative Brightness
test charts as the mainstay of our clinical evaluation of
vision. Although the term ‘visual acuity’ is almost
synonymous with the assessment of central visual status,
using optotypes, it is however more specific and refers to
the ability of the visual system to resolve spatial detail. It
is a measure of the angular size of detail that is just
resolvable by the observer and its limitations are imposed
1.22λ/p Distance
by a combination of optical and neural factors.
1.64λ/p

2.23λ/p

Optical Limitations 2.67λ/p

When the emmetropic eye is in optimal focus, the image Figure 1 Diagrammatic illustration of the diffraction
of a point object, formed at the macula, is not a perfect phenomenon described by Sir George Airy in 1835. The
point, but a blur circle surrounded by a series of faint graph illustrates the relative brightness and position of
concentric rings (Figure 1). This is a diffraction surrounding rings in relation to the central disc (Table 1)
phenomenon, which was first satisfactorily explained by (courtesy of Mr C Wolsey).
the Astronomer Royal Sir George Airy, in 1835, and
applies to all aperture-limited optical systems (Airy 1834,
cited in Jenkins & White 1976). The central circle is
known as the Airy disc, and its peak intensity is 57 times

Table 1. Relative intensities and sizes of the light distribution in the Airy disc and surrounding rings for the
image of a point source

Relative intensity Relative intensity Radius of ring

(Lmax) (Ltotal) (radians)

Airy disc 1 1 –

First dark ring 0 0 1.22 λ/p

First bright ring 1/57 1/12 1.64 λ/p

Second dark ring 0 0 2.23 λ/p

Second bright ring 1/240 1/30 2.67 λ/p

Lmax, relative brightness of surrounding rings at their brightest point, in comparison to the central point in the Airy disc which has been
allocated an arbitrary value of 1; Ltotal, relative brightness of each ring in total compared to the brightness of the central disc.

Address for correspondence: Prof AJ Jackson, Low Vision Clinic, Department of

Ophthalmology, Royal Group of Hospitals, Belfast BT12 6BA
53
© 2004 The College of Optometrists
AJ Jackson & IL Bailey

brighter than that of the first surrounding ring. The angular For light of different wavelengths, there will be a
size, or radius, of each of the surrounding dark and light longitudinal difference in the position of best focus. Short-
rings is a function of the wavelength of the incident light (λ) wavelength light will have its best focus about 0.5mm
and the pupil diameter (p). In the case of the first dark ring anterior to the best focus for long-wavelength light. This
the diameter, expressed in radians, is 1.22 λ/p, where λ and represents about 1.5D of longitudinal chromatic
p are normally expressed in millimetres. The diameter of aberration. The light spread in the retinal image depends
the first bright ring is 1.64 λ/p radians. In converting the on pupil size. Light spread due to diffraction increases as
measurement from radians to degrees and subsequently to the pupil becomes smaller. On the other hand, light spread
minutes of arc, one needs to multiply the figure by 57.29 due to monochromatic and chromatic optical aberrations
and thereafter by 60 (Table 1). is reduced as the pupil becomes smaller. The optimal pupil
size required to balance these effects best is
If the eye were perfectly focused and free of aberrations, approximately 2.5mm (Westheimer 1964).
the only optical factor limiting the resolution of the eye
would be diffraction. Applying the Rayleigh criterion, the
amount of overlap between adjacent blur circles, to render Neural Limitations
resolution impossible, would be half the diameter of the
blur circle. At this point, the centre of one blur circle A significant neural limitation on visual resolution is
would be aligned with the edge of its neighbour. An eye imposed by the anatomical structure of the retinal
with a pupil diameter of 4.6mm would thus have a photoreceptor layer. In the foveal region, cones are packed
minimal angle of resolution (MAR) of 0.5min of arc closely together and cones are known to have an average
whereas an eye with a pupil diameter half that size diameter of about 2µm. It might be expected that, in order
(2.3mm), would have an MAR of 1min of arc (Bailey for images of two-point objects to be resolved, they should
1998). The relationship between pupil size and image fall on two individual cones separated by a cone which is
quality is, however, more complex because superimposed not stimulated. The two-point images should thus be
on diffractive blur are degradations in image quality due to separated by a distance of 4µm. Assuming the nodal point
optical aberrations. These optical aberrations result from of the eye is 16.67mm from the retina, then the neural
the shape, alignment and positioning of the various optical limit to resolution becomes 0.82min of arc. Other neural
surfaces within the eye and variations in the refractive limitations of the system may also be imposed by the
indices and the chromatic dispersion of the optical media. various complex interconnections and interactions
When the eye views a point source, it receives a cone of between neurones within the retina and visual pathways.
diverging rays and the diameter of the incident cone is
limited by pupil size. Because of aberrations, the cone of
rays that converge to form the image on the retina has an Defining Resolution
apex that is imperfectly defined. Ideally all rays would
converge as a cone to a single point, and with the eye in Resolution can be defined as minimum detectable,
focus, the image point at the apex of the image cone would minimum separable or minimum recognisable.
be on the retina. In reality, for light of a given wavelength,
rays from different parts of the pupil deviate from the ideal Minimum detectable resolution
cone pattern, so the image on the retina is a patch rather
than a point. The light distribution across the patch is Minimum detectable (or minimum distinguishable)
called the point spread function and the distribution is resolution is the minimum angular size of a spot or the
not necessarily symmetrical. In general, rays from the minimum angular width of a line necessary for it to be
more peripheral parts of the pupil are more aberrated. detected against its background. In the case of a spot,
Retinal cones are, however, directionally sensitive, being when the diameter of the image formed on the retina
less sensitive to rays that fall upon them obliquely (Stiles becomes smaller than the surface area of a single
& Crawford 1933). The Stiles–Crawford effect thus photoreceptor, detection remains possible as long as the
reduces the impact of the aberrational effects associated visual system can recognise that adjacent photoreceptors,
with rays from more peripheral parts of the pupil. If the one receiving the stimulus and one not, have responded
object is a fine line, the cross-sectional distribution of light differently. Threshold detection is achievable when the
in the image has a bell shape. This is called the line spread spot is large enough or has sufficient contrast to create a
function. This profile differs from the point spread detectable elevation or reduction in illumination on the
function in that it is narrower and it does not have the photoreceptor, in comparison to its neighbour. Under
noticeable rings or fringes. photopic conditions, the differential contrast threshold

54
 Visual Acuity

remains a constant ratio, at approximately 1%,

irrespective of overall field luminance. This phenomenon
forms the basis of Weber’s law (Weber 1834). As lighting
levels drop, the threshold contrast ratio between target
spot and its surround must increase by up to 20 times in
order to enable target detection to take place.
Detectability will also vary as a function of target size,
wavelength and the distribution of luminance within the
visual field. The minimal dimension of a high-contrast
point object required for detection is reported to be
between 10 and 35 seconds of arc, whereas for a high-
contrast line target, the width only needs to be 4–10
seconds of arc (Duke-Elder 1937).

Minimum separable resolution

Minimum separable resolution is a measure of the

minimum separation between adjacent spots or lines
required for the observer to identify the objects as
separate. This form of resolution is usually assessed using
grating targets, the dimensions of which may be expressed
in cycles per degree (cpd). In a grating with a spatial Figure 2 Resolution and spurious resolution
frequency of 30 cpd, the width of each dark stripe and An illustration of the light distribution within a three-line
each bright stripe will be 1 min of arc. The limit of an eye’s target when the spacing is close to the resolution limit. The
ability to resolve a high-contrast grating, and distinguish it figure depicts the light distribution on the retina viewing an
object that consists of three fine white lines. Due to
from a uniform field of the same average luminance, is a diffraction and aberrations, each individual fine line creates
measure of its threshold spatial frequency for minimum an image with a bell-shaped light profile, and each of these
separable resolution. An eye that can just resolve a 30-cpd light distributions is called a line spread function (LSF).
grating can be said to have a MAR of 1 min of arc. Diffraction and aberrations affect the width and shape of the
LSF.
In measuring resolution with grating targets, there may be Bottom: the light distribution of the retinal image when there
spurious resolution, in which the presence of a grating can is only marginal overlap of the adjacent LSF. The observer
be recognised, even though the perception of the number should easily be able to tell there are three lines present.
of stripes in the grating is erroneous. This can be seen most
Middle: the lines have been brought closer together (spatial
easily in resolution tests that use sets of three line gratings frequency has been increased) so there is substantial overlap
that have different spatial frequencies (Figure 2). between the three LSFs. In the total light distribution of the
image, there are three peaks corresponding to the positions of
Hyperacuities the three lines in the object. There is only a relatively small
difference between the retinal illuminance at the three peaks
and the two troughs. If this difference (contrast) is sufficient,
To this point we have described visual acuity as a measure the observer will perceive a low-contrast grating pattern from
of the ability of the eye to detect spatial separation the three lines. As the lines are brought a little closer together,
between details, and that the magnitude of visual acuity is the difference between the retinal illuminance at the peaks
closely related to the angle between adjacent retinal and troughs will be further reduced.
receptors. There are, however, some tasks in which Top: the lines are even closer together. The overlap between
information about location is pooled across many adjacent LSFs is more substantial. Now, summation between
receptors, allowing spatial discriminations that are finer the sloping sides of adjacent LSFs gives rise to peaks in the
than the minimal separable resolution by a factor of about total light distribution. Here, the sum of the three LSFs gives
10-fold. The term ‘hyperacuity’ is used to describe these two peaks with an intervening trough. The observer may
perceive a striped pattern of low contrast, but it will appear
ultrafine spatial discriminations. Vernier acuity and that there are only two lines present . This is called spurious
stereoacuity are the two best known of the hyperacuities. resolution. The observer can tell the object has a grating
When assessing Vernier acuity, the task is to judge structure, but has an erroneous perception of the number of
whether two line segments are in line with each other. component lines.
When assessing stereoacuity, the task is to judge whether
one line segment is closer or further away than another in

55
AJ Jackson & IL Bailey

three-dimensional space. Thresholds for these

hyperacuity tasks are commonly in the order of 5 seconds Serif E 5 × 5 Landolt ring
of arc. Other hyperacuity tasks include recognising that a
line is curved rather than straight, discriminating
deviations from vertical or horizontal, discerning
deviations from parallelism, detecting a bump in a line and
judging equality of lengths in a bisection task.

Minimum recognisable resolution

Minimum recognisable (or minimum legible) is the term

given to the measurement of acuity using optotype
symbols. Whereas grating acuity is often used to
Sans serif H 5 × 5 Sans serif E 5 × 4
determine optical performance under laboratory
conditions, most clinical tests of visual resolution utilise
symbols or optotypes as targets (Figure 3). The optotypes
chosen may, however, be simple or complex. Landolt
rings, which represent the simplest of targets, were
adopted in 1909 by the XI International Ophthalmological
Congress as the standard test object (Landolt 1899,
Woodruff 1947). Each ring is constructed so that its
external diameter is five times greater than its stroke
width, which in turn, is equal in size to the gap in the ring.
Most vision tests using Landolt rings simply require the Figure 3
observer to identify the location of the gap which may be An illustration of the font styles and grid structure of serif
in one of four positions – to the right, left, up or down. and sans serif letters, the Landolt ring and tumbling E.
Sometimes obliquely oriented gaps (45º, 135º, 225º and
315º) are included, giving a choice of eight gap
orientations. When testing with Landolt rings, the patient 5 × 4 grid, in which stroke width measures one-fifth of the
need not be familiar with any particular alphabet. The test grid height, or a 5 × 5 grid. Snellen’s original charts used
is therefore well-suited for testing persons who are letters the stroke width of which were equal to one-fifth of
illiterate or who cannot name letters that might otherwise the letter height. Although most letters were constructed
be used as the optotype. For research purposes, Landolt using a 5 × 5 grid, some 5 × 6 letters were used (Snellen
rings may be the preferred optotype because the 1862, 1864). His letters had serifs, which are short lines
resolution task is essentially the same for each of the added to the ends of the limbs of the letters. Since Snellen,
alternative gap positions. When using a combination of there have been numerous variations in the design of the
letters as optotypes, the observer’s task and decision- letters and the chart layout. Bennett (1965) reviewed the
making strategies might be quite different from one letter history of visual acuity chart design at about the time the
to another. Tumbling E (or ‘illiterate E’) charts have been British Standards Institute chose to adopt a set of 10
constructed using an E optotype with four alternative letters based on a 5 × 4 grid format as the standard
orientations. Typically, the E is constructed on a 5 × 5 grid optotypes for clinical testing of visual acuity (BSI 4274-1
with each of the three parallel limbs being one grid-unit 1968). The UK standards have recently been updated in
wide. As for the Landolt ring test, the observer’s task is to such a manner as to introduce log MAR notations and
identify the orientation of each optotype by indicating if several other features consistent with log MAR testing
the ‘legs’ of the E point right, left, up or down. Forced- strategies (BSI 4274-1 2003).
choice methodology is often used in the assessment of
threshold acuity using tumbling Es or Landolt rings,
especially when used for research purposes. Optotype Test Chart Design

More traditional visual acuity tests utilise charts with sets Since the time of Snellen, the practice of expressing
of alphanumeric symbols as optotypes, as this makes the central visual status using measurements obtained from
process of testing simpler and quicker. Individual letter, or optotype charts has been confounded by the proliferation
number, optotypes are usually constructed using either a of charts utilising different design features. The

56
 Visual Acuity

assessment of visual acuity using optotype targets has Symbol selection

become almost synonymous with clarity of vision (Figure
4). Bennett (1965) in his extensive review on ophthalmic Not all letters are of equal legibility and legibility will differ
test types classified the three principal problems according to font style. Snellen’s charts (1862) used
associated with standardisation as: ‘Egyptian paragon’ styled serif letters, each of which had
prominent ornamental cross-strokes at the end of each
1. Selection and style of symbols limb. Green (1868, 1905), who favoured simpler font
2. Progression of sizes of symbols styles, initially used a sans serif Gothic design and
3. The notation used for expressing acuity excluded letters with oblique lines (Z, N, S etc). Hartridge
and Owen, who compared the legibility of Snellen and
Green style letters, recommended that acuity should be
assessed using a series of nine letters (D F H N P T U X Z)
selected from the Green series (Hartridge & Owen 1922).
By 1872, Green’s thoughts on the matter had changed in
favour of a modified serif design as sans serif letters had
been criticised as being ‘unfinished in appearance’ (Green
1872). Dennett (1885) adjusted individual letter sizes in
an attempt to compensate for relative legibility whereas
others, including Cowan (1928) and Monoyer (1875),
emphasised the importance of individually selecting font
styles and linear dimensions (ie 5 × 5, 5 × 4 or 5 × 3) on a
letter-by-letter basis. As suggestions concerning letter
shape and size proliferated, reverse contrast was proposed
as a method of compensating for legibility differences
(Walker 1942). Much of the original work on relative
legibility was undertaken by Woodruff (1947) and Coates
(1935) who accumulated considerable normative data on
legibility. The effect of font style on relative legibility has
Figure 4 In popular perception the test letter chart quickly been reviewed by Borish (1970). Bennett (1965)
became synonymous with the sight test, as demonstrated by illustrated how different letter design styles could affect
the cartoonist Sidney Strube (1891–1956) in this newspaper legibility, serif letters in particular, introducing
drawing from 1938. The Prime Minister, Neville characteristic identification errors. In his comprehensive
Chamberlain, is portrayed as an ophthalmic optician whose review on the subject, he outlined the rationale used by
sum total of equipment consists of an electrically various test chart designers for selecting letters of various
illuminated screen chart, a simple drop-cell trial frame and
case of trial lenses. The inability of the governing MPs styles. British Standard BS4274-1, which was published in
(portrayed as the British lion in a school cap) to read the 1968, recommends 10 (5 × 4) sans serif letters giving
smallest lettering is intended as a symbol for their failure to legibility ratings ranging from 0.92 (R) to 1.10 (Z) (D E F
focus on what supporters of Appeasement felt to be a more H N P R U V Z) (BSI 4274-1 1968). The most recent
important issue in a country still recovering from the effects version incorporates log MAR notation, recommends a
of the great economic slump. Mother, in the form of Winston 5 × 5 letter grid format, a modification to row progression,
Churchill, dressed somewhat ridiculously as Britannia, sits
in on the test in a position of splendid isolation across the
and introduces the letters C and K to the recommended
room. All may not be as it seems, however. As Strube was a range from which letters should be selected (BSI 4274-1
fanatical anti-Nazi and favourite of Churchill it is possible 2003). In the USA Sloan, who pioneered much of the work
that the artist’s intention is to illustrate the priority of the on standardisation, selected a sans serif style (Sloan 1951,
more easily seen letters (courtesy of Mr Neil Handley MA Sloan et al. 1952). Her preference was for 10 letters in a 5
AMA, Curator British Optical Association Museum, College × 5 grid style (S O C D K V R H N Z) (Sloan 1959). Her
of Optometrists, London).
legibility ratings ranged from 0.90 (O) to 1.10 (Z).
© The Daily Express Regrettably, many currently available charts, though using
The original sketch for this cartoon, in which Chamberlain sans serif letters, do not consider legibility ratings.
is described as an ‘oculist’, is in the collection of the BOA Legibility is also affected by ‘contour interaction’, a term
Museum at the College of Optometrists. coined by Flom to describe the effect of neighbouring
contours on the discriminability of detail (Flom et al.
1963). Flom distinguishes between ‘contour interaction’,

57
AJ Jackson & IL Bailey

which affects resolution of detail, and ‘crowding’, which

Increments on
affects assignment of attention. Both affect performance in
visual acuity scales
recognising optotypes on a visual acuity chart. Bailey
studied the effects of spacing between letters and between
rows for different optotypes. Using British Standard
letters, he found that a twofold increase in letter spacing
changed the visual acuity score by 0.03 log units (Bailey
1988). Spacing was noted to have more of an effect when
assessing acuity with Landolt rings than when using
Sloan-styled letters (Raasch & Bailey 1984). Bailey
recommends letter and row spacings should equal letter Geometric
width. Reciprocal
Arithmetic
Row progression ‘Conventional’
6/6
Superimposed on the debate surrounding letter style has 6/60
been an equally vociferous debate about chart design and
row progression ratios. Both geometrical and arithmetical
progressions, using a variety of progression rates, have
been utilised over time. Snellen chose for his original
chart seven rows of letters which subtended 5min of arc at
distances of 6, 9, 12, 15, 21, 30 and 60m. The 1911 edition
of his chart changed the size progression to 6, 8, 12, 18,
24, 36 and 60-m letters. The six smaller rows approximate
a geometric or logarithmic progression with a multiplier of
Figure 5 This figure shows the distribution of increments
1.4 (1.4 = √2 = 100.15 = 0.15 log unit). Using this factor, letter on the visual acuity charts and their dependence on the
size doubles every second row (6, 8.5, 12, 17, 24, 34). mathematical system by which the size increments are
determined.
Figure 5 illustrates the consequences of using three The diamonds illustrate increments that follow a geometric
different size progression sequences. One has a geometric or logarithmic size progression. There is a constant ratio
(logarithmic) progression with a constant multiplier ratio between the sizes in adjacent rows. It is now widely
accepted that this system is the most appropriate for visual
of 100.1 (6/6, 6/7.5, 6/9.5, 6/12 ...), the second uses a simple
acuity scaling. It is the basis for the log MAR and visual
arithmetic progression in height (6/6, 6/12, 6/18, 6/24 ...) acuity rating scales. Here, there are 10 steps between 6/6
and the third uses equal steps of the reciprocal of size with and 6/60.
0.1 increments on the decimal scale. The sequence on the The open circles show the increments on a scale that
decimal scale is 1.0, 0.9, 0.8, 0.7 ... with the corresponding follows the reciprocal of visual angle. This is the basis of
Snellen sizes being (6/6, 6/6.7, 6/7.5, 6/8.6, 6/10 ...): the the decimal visual acuity scale. The increments are
figure is scaled so that the increments are equal on the relatively compacted at the small size and sparsely spaced
at the larger size. Here there are nine steps between 6/6 and
geometric scale. Three steps correspond to a twofold 6/60. The sequence is 1.0, 0.9, 0.8, 0.7, etc (6/6, 6/6.66,
change in size over the entire range. With the arithmetic 6/7.5, 6/8.57, 6/10, 6/12, 6/15, 6/30, 6/60).
scaling sequence, three steps cover a four-times change in The open squares show the scale for an arithmetic
size at the small end, and a 1.43 times change in size at progression. This scale has equal steps in terms of
the large end. For the reciprocal progression, three steps minimum angle of resolution (1, 2, 3, 4, 5, etc in minutes of
at the small end represent a 1.43 change in size, whereas arc) or in terms of the Snellen fraction denominator (6/6,
6/12, 6/18, 6/24, 6/30, 6/36, 6/42, 6/48, 6/54, 6/60). The
at the larger end of the scale, three steps cover a four increments are quite sparse at the smaller sizes and
times change in size. Also shown on this figure is the compacted at the larger size. Here, there are nine steps
‘conventional’ size series commonly used in the UK (6/4, between 6/6 and 6/60.
6/5, 6/6, 6/7.5, 6/9, 6/12, 6/18, 6/24, 6/36 and 6/60). This The filled triangles illustrate the ‘conventional’ scale most
series uses optotypes tightly spaced at the smaller sizes widely used in the UK. The increments are most densely
and more sparsely spaced at the larger sizes. packed at the smaller sizes and more sparsely spread at
the larger size. There are nine steps between 6/6 and 6/60,
the sequence being (6/6, 6/7.5, 6/9, 6/12, 6/18, 6/24, 6/36,
Westheimer (1979) took advantage of the eye’s natural 6/60).
decline of visual acuity away from the fovea to show that
the ‘probability of seeing’ function, a graph showing

58
 Visual Acuity

percentage correct recognition as a function of size, was patients, with ample accommodative reserves, attention
essentially the same for all levels of acuity, provided that may need to be paid to the potential for defocus induced
visual acuity was plotted on a logarithmic scale. There is a by proximal accommodation if the chart is at a close
broad acceptance that logarithmic (constant-ratio) scaling distance.
should be used for visual acuity measurement, and the
most popular scaling ratio is now 0.1 log units (100.1, Acuity designation
1.2589 or approximately √2). This progression, which was
3

proposed by Green (1905) and endorsed by the American Whereas visual acuity is an expression of the angular size
Committee on Optics and Visual Physiology (Ogle 1953; of the smallest detail just resolvable by the eye, there are
Spaeth et al. 1955), has had more recent advocates in several alternative ways of specifying the value clinically.
Keeler (1956), Sloan (1951) and Bailey & Lovie (1976).
The Snellen fraction expresses the angular size of the
On this scale, 10 steps represent a 10-fold change, three a optotype as a fraction, specifying the test distance d as the
twofold change, and one step represents a change of about numerator and the distance at which the just-resolvable
5/4. This sequence, based on a 6m reference, is 6, 7.5, 9.5, letters should be positioned in order to subtend an angle
12, 15, 19, 24, 30, 38, 48 and 60. For 20ft, the sequence is of 5min of arc as the denominator D (visual acuity = d / D)
20, 25, 32, 40, 50, 63, 80, 100, 125, 160, 200. (Figure 6). A benefit of expressing visual acuity in this way
is that the clinician records the testing distance, a factor
Test distance that can be important in paediatric eye care and low-
vision work. At a viewing distance of 6m, a patient with a
The original Snellen chart was designed for use at a testing best corrected visual acuity of 6/24 will just be able to read
distance of 20 Parisian feet (6.5m). (One Parisian inch is the 24m letter, which subtends 5min of arc at a distance
equivalent to 1.0658 English inches.) By 1875, new of 24m. In the USA, test distances are expressed in feet,
editions of the Snellen chart were calibrated in metric whereas in most European and Commonwealth countries,
units and were produced for use at 6m. Charts for use at a metric notation is used (6/6 = 20/20, 6/60 = 20/200, etc).
5m testing distances were, however, soon to become On the European continent, following the notation used
available. Both Landolt and Monoyer advocated the use of by Monoyer (1875), the Snellen fraction is usually
the 5m test distance (Landolt 1899, Monoyer 1875). The reduced to a decimal notation. An acuity of 6/6 thus
XI International Ophthalmologic Congress in 1901 ratified becomes 1.0, whereas a visual acuity of 6/60 becomes 0.1
a recommended working distance of 5m (cited in Bennett in the decimal notation. With the decimal notation,
1965). Historically, the UK has stayed with a 6m standard information on test distance is lost. Further, and more
test distance, whereas in the USA the 20ft standard has importantly, visual acuity scores in decimal notation can
been adopted. Given that clinical rooms are rarely 6m become confused with acuity scores expressed as log MAR.
long, test distances are often achieved using indirect
charts viewed in a mirror. The US Committee on Vision of Those familiar with the task of checking European and
the National Academy of Science/National Research
Council recommended that 4m be the standard test
distance (NAS-NRC 1980). Current British standards state Eye with 6/6 Angle = 5 min of arc = 1/12 degree

that the testing distance should not be less than 4m (BSI

4274-1 2003).
E
6m
E
12m
E18m

In low-vision work, the concept of a variable working

distance is widely used and charts may be presented at Eye with 6/18
various distances, sometimes as close as 50cm from the
eye. At close distances, attention may need to be paid to
the accommodation demand induced by reducing the
E
6m 12m
E18m

working distance. The dioptric depth of field of the eye

provides a tolerance to small errors of focus. It is generally
assumed that the refractive correction should be no more Figure 6 The Snellen fraction (e.g. 6/6, 6/12, 6/18, etc.), of
than +0.25D from the optical correction that would be which the first number (numerator) indicates the test
distance in metres and the second number (denominator)
optimal for the viewing distance that is being used. For indicates the height of the letter. The height of the letter is
low-vision patients with poor visual acuity, wider dioptric indicated by the distance at which a letter of that height will
tolerances are often allowed. When testing younger subtend a specific angle, namely, 5min of arc = 1/12 degree.

59
AJ Jackson & IL Bailey

North American reference articles for comparable studies value of 0.02 log MAR units can be assigned to each letter
will be familiar with the problems which can arise. read correctly. Other forms of acuity expression have also
Alternative notations have been proposed by various been advocated. Bailey (1988) suggested the use of a
authorities over the years. Blaskoviscs (1924) proposed visual acuity rating (VAR), derived from log MAR acuity
the ‘oxyoptre’, which was the reciprocal of the visual angle values (VAR = 100 – 50 log MAR). Using this system 6/6
(in degrees) of the smallest recognisable target. Javal equates to VAR = 100, 6/60 becomes VAR = 50 and 6/600
(1900) proposed that ‘letter height’ should be used, becomes VAR = 0. There is one VAR point per letter on a
whereas Swaine (1925), who coined the term ‘visual Bailey–Lovie or Early Treatment of Diabetic Retinopathy
badness’, suggested that the Snellen fraction should be Study (ETDRS) chart. Another designation system, visual
inverted. efficiency (VE), rarely used in the UK, was proposed by
Snell and Sterling in 1926 in an attempt to quantify visual
Minimum angle of resolution loss for medicolegal purposes (Snell & Sterling 1926).
They proposed the efficiency formula VE = 0.2
MAR is used to express, usually in minutes of arc, the (MAR–1)/9. An individual with ‘normal’ (20/20) vision
angular size of the critical detail within a just-resolvable would be allocated a VE rating of 100%, while an individual
optotype. It is usually assumed that the critical detail is with a best corrected visual acuity of 20/200 would have a
one-fifth of the letter height. Thus, an individual with a VE rating of 20%. This method, which has been reviewed
best corrected visual acuity of 6/6 can resolve critical by Hofstetter (1950), was adopted by the American
detail, represented by a letter limb width of 1min of arc. If Medical Association in 1955 (Spaeth et al. 1955).
the best corrected acuity were 6/60, the minimal limb Somewhat similar ratings for field loss and motility
width resolvable would be 10min of arc. The MAR is, in restrictions were subsequently developed and a
fact, the reciprocal of the decimal acuity value. The log permanent impairment index was introduced in 1984
MAR designation was introduced in 1976 with the (American Medical Association 1984).
Bailey–Lovie chart (Bailey & Lovie 1976). It expresses the
visual acuity as the common logarithm of the MAR. On In many circles, within the UK at least, it has become
this scale, the 6/6 acuity (MAR = 1) becomes 0 on a log commonplace to record visual acuities of less than 6/60 as
MAR chart (log10 (1.0) = 0). A 6/60 acuity (MAR 10) CF, HM, PL or NPL. CF stands for count fingers and
becomes log10 (10.0) = 1.0 (Table 2). Using charts with indicates that the patient can recognise the number of
five letters per row and a 0.1 log unit progression of size, a fingers held up by the clinician. This is an unnecessarily

Table 2. Comparison of different visual acuity designations.

Snellen Snellen Decimal Minimum angle Visual acuity Visual Conventional Keeler Snellen Snellen
fraction fraction of resolution Log MAR rating efficiency Snellen Chart fraction fraction
(metres) (feet)
notation (MAR) (VAR) % A chart (metres (4m)) (feet)

In the conventional Snellen and Keeler A series columns, brackets around the acuity values indicate that the size is
only a moderately close equivalent to the values in the other seven columns. For the Keeler A series, exact Snellen
fraction equivalents are shown in parentheses.

60
 Visual Acuity

crude way of estimating visual acuity. Finger widths will stroke widths that were only one-seventh of the letter
vary from one clinician to another, separations between height. An important feature was that size progression was
fingers will differ from one presentation to the next and logarithmic. Successive sets of letters, moving up the
the luminance and complexity of the background will chart, were larger, by a factor of 1.25, than their
affect the visibility of the fingers, so the visual task is not predecessors. On the original chart, there were eight A1
well-controlled. Working distances are often not recorded (6/6) letters and only one A20 (6/416) letter. In the USA,
by those who choose to express visual status in terms of Sloan developed a visual acuity chart using her family of
counting fingers. For patients whose visual acuity is less 10 5 × 5 sans serif letters and a logarithmic size
than 6/60, appropriate measures of visual acuity may be progression with a multiplier ratio = 1.2589 = 10√10 =
made by holding a visual acuity chart at a close distance. 100.1 (Sloan 1959). An important advance in the
If a 60m letter cannot be read at a distance of, say, 50cm measurement of visual acuity came with the development
or even closer, then attempts at obtaining a meaningful and introduction of the Bailey–Lovie design principles
score of visual resolution may be abandoned. which were applied to new distance and near vision charts
in the 1970s (Bailey & Lovie 1976). The design principles
Hand movements (HM) is not a measurement of visual cause the visual acuity task to be standardised at each size
acuity, but rather a more basic and coarse classification of level. Other features that combined with the logarithmic
visual function. It should be reserved for those cases size progression to standardise the task are: (1) the same
where the patient is unable to see any chart optotypes at number of optotypes in each row; (2) spacings between
any distance. With this level of vision, the observer can optotypes and between rows are proportional to the size of
only detect shadows or large objects moving across the the optotypes; and (3) the mix of optotypes has
visual field. Perception of light (PL) indicates an ability of approximately the same level of difficulty. This means size
the eye to detect a light source which may be a ceiling is the only significant chart variable from one size level to
light, or an ophthalmoscope bulb held close to the eye. the next. Their design uses a geometrical progression of
Such a classification of vision should be reserved for eyes size and they chose a factor of 1.2589 (= 100.1) as their
incapable of detecting hand movements. It can be useful multiplier ratio. Bailey and Lovie introduced the term ‘log
to ask if the direction of the light source can be identified, MAR’ as a measure of visual acuity used in conjunction
as an ability to judge the direction of major light sources with a visual acuity chart. On their log MAR scale, there
can be functionally valuable when it comes to orientation are 10 steps between 6/60 and 6/6, with the size
awareness. The NPL (no perception of light) classification progression being 60, 48, 38, 30, 24, 19, 15, 12, 9.5, 7.5,
of vision refers to total blindness. 6. On the charts, the log MAR values are in increments of
0.1 log units: 6/60 = 1.0 log MAR, 6/48 = 0.9 log MAR, 6/38
Clinicians are often required to provide visual acuity = 0.8 log MAR, and so forth (Table 2). Every third line on
measurements in order to determine an individual’s the chart sees a doubling, for larger letters, or a halving,
eligibility for benefits, for privileges or for occupational for smaller letters, of letter size, so doubling or halving
assessment purposes. It can be useful to have information causes a change in the log MAR value by 0.3. Each line
on hand on the World Health Organization classification of differs in size from the rows immediately above or below
central visual status (World Health Organization 1979) by a factor of 1.2589 (0.1 log units). On currently available
and local definitions that are relevant to occupational charts, there is a 20-fold range of size with 14 rows of
standards (Association of Optometrists 2002), driving letters which, assuming a 6m viewing distance, provide a
(Taylor 1995) and visual impairment registration (Bunce size range from log MAR 0.8 (6/38 equivalent) to log MAR
et al. 1998, Evans 1995). –0.5 (6/1.9 equivalent). On the Bailey–Lovie charts, there
are five letters per row, with the letters selected from the
group of 10 recommended in the 1968 British Standards
Modern Test Chart Characteristics (BSI 4274-1 1968). The spacing between adjacent letters
is equal to one letter width, and the separation between
Logarithmic progressions rows is equal to the width of the letters in the larger row.
The combination of letters on each line is selected to
The first of the modern logarithmically based test charts ensure about equal average difficulty from row to row.
was introduced to UK-based practitioners in 1956 by
Charles Keeler. The Keeler charts were referred to as A Visual acuity measurements recorded using Bailey–Lovie
series charts and were marketed as particularly useful for charts can be determined using either row-by-row scoring
those involved in low-vision practice (Keeler 1956). The or letter-by-letter scoring. Common clinical practice is to
chart layout was generally similar to that of a conventional record the acuity as the smallest line on which the patient
Snellen chart. At the larger sizes, the sans serif letters had is able to identify either 50% of 80% of letters correctly.

61
AJ Jackson & IL Bailey

This methodology is, however, too coarse to identify small Bailey–Lovie chart, a difference of five letters can usually
or moderate changes in acuity. A small change in the be taken as being significant. That is, there is a 95%
number of letters read, even by one letter, can easily cause confidence that two visual acuity scores will be no more
the acuity score to change by one row. A change in visual than four letters different from each other when there has
acuity score cannot be taken to be real or significant been no change in vision. Thus a change of five letters can
unless the change is by two rows or more (Bailey et al. be taken as being a real or significant change. Confidence
1991). The row-by-row scores can be refined by applying limits may be broader for some patient groups in whom
the plus or minus notations, commonly used when acuity measurements are less consistent (Bailey et al.
expressing Snellen fraction measurements of acuity (6/6 ± 1991).
2). Giving credit for every letter read improves sensitivity
substantially (Figure 7). The Bailey–Lovie design The term ‘log MAR chart’ is often used to describe charts
principles allow each letter to be given the same value in that follow the Bailey–Lovie design principles (Figure 8).
log MAR or VAR units. Every extra letter read changes the The most widely used log MAR chart is the ETDRS chart
log MAR score by –0.02 and the VAR score by 1. This designed for the Early Treatment of Diabetic Retinopathy
facilitates scoring visual acuity letter-by-letter. With the Study by Ferris et al. (1982). These charts, of which there
five-letters-per-row chart, each letter can be assigned a are three, use a combination of the 10 Sloan 5 × 5 letters
value of 0.02 log MAR units (VAR = 1), with the result that rather than the 5 × 4 British standard letters and they
each additional letter read on a subsequent line changes were designed for use at 4m following the
the log MAR score by –0.02 and the VAR by +1. The recommendations of the Committee on Vision of the
patient who reads all letters on every row down to the 0.20 National Academy of Sciences National Research Council
log MAR line (VAR = 90 = 20/32) and one letter on the 0.1 in the USA (NAS-NRC 1980). The use of standardised
log MAR line would thus be allocated a log MAR score of illumination is encouraged and both direct and retro
0.18 (VAR = 91). Another patient who reads all letters on illumination cabinets have been designed for use with
every row down to the 0.20 log MAR row, four out of five ETDRS charts (Ferris & Sperduto 1982).
on the 0.10 row and three out of five on the 0.00 row
scores, using individual letter scoring, a log MAR score of Strong and Woo produced the Waterloo chart, which uses
0.06 (0.20 – 0.08 – 0.06 = 0.06). In VAR units the a logarithmic progression, with letters in columns
equivalent score is 97 (90 + 4 + 3 = 97). becoming progressively smaller from left to right. They
When giving credit for every letter read using a also added contour bars at the end of each row and
Frequency

By row
By letter

Test / retest discrepancy (letters)

one row = five letters = 0.1 log unit

Figure 7 Distribution of test–retest discrepancies expected Figure 8 A selection of currently available log MAR charts:
when scoring visual acuity, recorded for normally sighted a high–contrast Bailey–Lovie chart, a low–contrast Bailey-
subjects, on a Bailey–Lovie log MAR chart using Lovie letter chart, an ETDRS chart, Glasgow Acuity cards
letter–by–letter and row–by–row scoring criteria. The 95% and Kay Crowded Symbol Cards.
confidence limit for change using letter–by–letter scoring is
± 5 letters. For row by row scoring it is ± 2 rows.

62
 Visual Acuity

column, thus ensuring that contour interaction affected all Test Chart Luminance
letters equally (Strong & Woo 1985). Log MAR principles
have also been applied to the development of distance The sensitivity of the visual system is such that it is
visual acuity charts utilising Chinese characters (Woo & capable of responding to illuminated targets over a very
Lo 1980). Similar principles have been applied to both wide range of intensities. The normal healthy eye can
optotype and picture symbol acuity charts developed for operate over a very wide range (7–8 log units) of
the assessment of vision in young children (Hyvarinen et luminance, encompassing both scotopic and photophobic
al. 1980, Jayatunga et al. 1995). McGraw and Winn vision. Visual acuity does vary with changes in chart
developed the Glasgow acuity card log MAR-based system, luminance. The British Standard recommendations are
which incorporates contour interaction bars, for assessing that internally illuminated charts should have a minimum
vision in preschool children (McGraw & Winn 1993). background luminance of 120 candela per square metre
Recent studies indicate that visual acuity results recorded (cd/m2), with newly installed charts having a minimum
using Kay picture, crowded symbol and crowded letter log background luminance of 150 cd/m2 in order to allow for
MAR-based tests, on children with amblyopia, are ageing of the system (BSI 4274-1 1968). Externally
comparable (Jones et al. 2003). illuminated charts should have a minimum illuminance of
480lux, although new installations should measure
600lux. Recommendations in the most recent British
Other Test Chart Design Features Standards remain unchanged, although guidance on
background illumination has been withdrawn (BSI 4274-1
Historically, acuity charts were produced as printed 2003). The recommended lighting levels, for the
panels designed for use with direct illumination. Most externally illuminated (807–1345lux) and internally
were printed on card and, as a result, were susceptible to illuminated (343cd/m2) charts advocated by Ferris, are
discoloration and ageing. Charts printed on plastic panels higher than those recommended in British Standards
were more durable and could, in addition, be wiped down (Ferris & Sperduto 1982). Sheedy and colleagues have
with a damp cloth, thus rendering them grime- and dirt- reported that, when chart luminance is within the
free. Given the lack of availability of 6m testing facilities, moderate photopic range (40–600cd/m2), doubling the
many visual acuity charts are printed in reverse format so chart luminance only alters the visual acuity score by a
they can be viewed via a mirror in rooms 3–4m in length. little less than 0.02 log units (Sheedy et al. 1984). They
In an attempt to introduce standardisation of testing recommended that chart luminances be in the range
conditions, internally illuminated acuity testing 80–320cd/m2. Within any given clinical environment, it
equipment was produced, the most commonly used in has been recommended that chart luminance should be
British hospitals being the rotating Snellen drum. Even kept constant with a tolerance of ±15% (Bailey 1998) and
with these units, charts are subject to ageing and care that variance across a chart should not exceed 20% (BSI
needs to be taken in the selection of background bulb 4274-1 2003).
intensity. Log MAR charts are now also available as
internally illuminated units with interchangeable panels
to allow variation in the letter sequencing. Testing Protocols
An alternative to the panel chart is the projector chart. Many of the important principles and practices associated
Typically their display area is relatively small in angular with acuity testing have been outlined by Johnston
size and this limits the length of the rows and the number (1991). Irrespective of the type of chart used, or the test
of print sizes that can be presented at the same time. The distance selected, it is important to record both
angular size of the projected letters will be independent of monocular and binocular acuities. Good clinical practice
viewing distance provided the projection screen is is to commence the assessment of visual functions by
equidistant from both the projector and the patient. In recording habitual visual acuities with the patient
short rooms, mirrors may be used to increase the optical wearing his or her own distance spectacles or contact
path length from projector to screen and from screen to lenses. Adapting the practice of recording data for the
patient. The magnification of the projector may be varied, right eye (Rt/OD) first, whilst ensuring total occlusion of
allowing precise calibration of the angular size of the the left eye, minimises the risk of erroneously recording
letters for a given eye-to-patient distance. laterally transposed data on the record card. Thereafter
left-eye data (Lt/OS) should be recorded using, if possible,
a separate acuity chart. This can be readily achieved using
a Snellen rotating drum system, hand-held charts or
alternatively by utilising different charts from the

63
AJ Jackson & IL Bailey

Bailey–Lovie or ETDRS sets. Generally, binocular acuities pathology, is about 6/4.8. Even in those over the age of 75
(Binoc/OU) will be equal to or marginally better than best years, provided the eyes are free of disease, the average
monocular acuity. Sometimes, such as in the case of visual acuity is marginally better than 6/6 (Elliott et al.
congenital nystagmus, the difference may become as great 1995).
as 0.3 log MAR units. It is often useful to record unaided
acuities (Rt/Lt and Binoc), or ‘visions’, as they are often For clinicians to monitor changes in ocular health and
called in the UK, in order to ascertain the extent to which visual status, visual acuity measurements should be made
the refractive correction improves visual status. In cases with appropriate test charts, viewed under controlled
where an occupational report is requested, it is not conditions, and credit should be given to every additional
uncommon for employers to request information on both letter read correctly. The significance of this statement is
habitual and unaided visual acuities as well as best borne out by the publication of two British Medical
corrected acuities. Best corrected visual acuities, Journal papers on the subject in the mid-1990s (McGraw
recorded monocularly and thereafter binocularly, should et al. 1995, Pandit 1994). McGraw et al. concluded that
be recorded after full refraction. Pinhole acuities (PH) are doctors must be extremely cautious when assigning
measures of visual acuity through a 1–1.5mm pinhole. clinical importance to changes in acuity of two lines or
Pinhole acuities should be determined when the best less because of the inherent variability of the Snellen
corrected visual acuity is less than expected or where chart, whereas Pandit stressed the practical importance of
there is reason to suspect that visual impairment is standardising working distances and illumination
predominantly the result of medial irregularity as, for conditions when assessing visual acuity using Snellen
example, in keratoconus or cortical cataract. charts in general practice.

To ensure that a given patient’s threshold lies within the

range of the chart, all letters should be read at the largest The Impact of Glare
size and no letters should be legible at the smallest size. If
a patient correctly reads half or more of the letters in one Light scatter from lenticular or corneal media haze
row, he or she should be encouraged to attempt to read produces veiling luminance over the retina. This reduces
any letters in the following row. In clinical research the contrast of the retinal image which produces glare
protocols, it is common practice to have all patients read disability. Substantial visual disability and glare
from the top of the chart, even for patients with excellent discomfort may be encountered in certain lighting
visual acuity. This practice can be laborious when using conditions, such as when facing oncoming headlights
log MAR charts where, for normally sighted subjects, the when driving at night, or when driving towards the setting
threshold size will commonly be at the 12th or 13th row sun. In the consulting room, lighting conditions are rarely
on the chart. In the interest of time, patients may be sufficient to induce significant glare disability or
encouraged to start reading three rows above the expected discomfort. Sometimes, vision may, however, be reduced
threshold. Using log MAR charts, care has to be taken to when reading material printed on glossy paper,
encourage patients with macular defects, and those with illuminated by an inappropriately positioned luminaire. A
neurological impairment, as the visual task can be difficult clinical simulation of distance vision under glare can be
and exhausting. Isolating letters, or sets of letters can be achieved by using the brightness acuity tester. This device
useful and flip charts such as the Sonksen–Silver charts allows acuity to be assessed, viewing through a 12mm
may be useful in such circumstances (Sonksen & Silver diameter peephole in a 60mm hemispherical light-
1988). For more expeditious testing, Rosser et al. designed diffusing bowl which can be set at three different levels of
charts with an abbreviated size range, different letter glare intensity. Neumann et al. (1988) found the
designs and fewer letters per row (Rosser et al. 2001). An brightness acuity tester to be an accurate predictor of
alternative approach advocated by Camparini and outdoor high-contrast acuity in cataractous patients.
colleagues is to estimate the threshold region by having Glare has an even more dramatic impact on visibility
the patient read only the first letter in each row of ETDRS when the test targets are of low contrast (Elliott &
charts until mistakes are made (Camparini et al. 2001). In Bullimore 1993).
both cases, quicker testing is achieved with modest
reductions in reliability and validity.
Contrast and Acuity Charts
There is a popular belief that 20/20 vision (6/6 or log MAR
= 0.0) represents ‘normal’ vision. Elliott and co-workers Although a review of contrast sensitivity is outside the
have shown that the average best corrected visual acuity scope of this paper, it should be acknowledged that
in those under the age of 50, who show no signs of ocular fundamental to vision is the eye’s ability to differentiate

64
 Visual Acuity

luminance differences within an image formed on the Electronic Display Technology and Clinical
retina. The difference in luminance (∆L) that can just be Measurement of Visual Acuity
detected is proportional to the luminance of the
background on which the increment is displayed. This is Computer-controlled displays offer some important
expressed as Weber’s law (∆L/L = constant), where L is the advantages for clinical measurement of visual acuity and
overall background or prevailing luminance (Tunnacliffe related functions. For visual acuity charts, optotype
1994). The threshold contrast ratio (or Fechner constant) sequences may be varied, so memorisation becomes less
is about 0.02–0.03 over the full range of photophobic of a problem, especially when it is necessary to take
luminance (1–1000 cd/m2). Under scotopic conditions, the repeated measurements. Patients’ responses can be
Fechner constant rises exponentially to 0.7 and the eye recorded online and presentation protocols may be
becomes less good at detecting differences in luminance modified to enhance efficiency of testing, reliability of
levels within a target. When quantifying the contrast results and storage of data. Systematic variations may
inherent in a conventional black-on-white chart, it is easily be made in stimulus variables such as luminance,
common practice to express contrast (C) as C = ∆L/L or contrast, spacing arrangements, optotype sizes, exposure
C = (Lmax – Lmin)/Lmax. The British Standard recommend- time, and so forth. Until recently it had been difficult to
ations are that conventional acuity charts should have a obtain high luminances on electronic screens. Pixel
contrast factor of 0.9 (90%), which is close to the structure imposes some limitations in that, for reasonable
theoretical maximum with printed charts (BSI 4274-1 shape fidelity, letters need to be about 10 pixels or more
1968). Sloan suggested that the lowest acceptable contrast in height, and the need for more pixels may be more
ratio for the assessment of conventional high-contrast pronounced for Landolt ring and tumbling E targets with
acuity was 0.84 (84%) (Sloan 1951). their more regular structures. The smallest letters on the
chart should be just beyond resolution for normally
In recent years, increased understanding about the impact sighted subjects.
of ocular and visual pathway pathology on contrast
sensitivity function has resulted in the development of Pixellation still imposes important limitations. If the
low-contrast acuity charts. Regan letter charts, which are smallest letters on the chart are 10 pixels high, a row of
visual acuity charts with eight letters on each of five lines, five 5 × 5 letters would need to be 110 pixels long, allowing
are available at 96%, 7% and 4% contrast levels. They are for one-letter spacing between letters, and similar spacing
designed for use at 3m and are scored on a letter-by-letter at either end of the row. It would be 88 pixels for 5 × 4
basis (Regan 1988). Also available is the reduced-contrast optotypes. For a chart covering a 20-fold size range (6/60
Bailey–Lovie visual acuity chart, which is identical to the to 6/3), the length of the largest row would thus need to be
high-contrast chart in design, except that the letters are 2200 or 1760 pixels. With the separation between rows
printed at an 18% contrast level (Bailey 1982). The Regan being equal to the height of the letters, the height of the
and Bailey–Lovie low-contrast visual acuity charts are chart would need to be at least 187 times the height of the
used to determine the smallest letters that can be read at smallest letters for a 20-fold range of sizes. Thus, for a 6/60
two or more different contrast levels. Less well-known are to 6/3 size range, a display screen would need to have in
the low luminance charts (Smith-Kettlewell Institute Low the order of 2000 pixels in both directions. The pixellation
Luminance or SKILL charts) designed by Haegerstrom- constraints are not so restrictive in the vertical dimension
Portnoy et al. (1997). These charts utilise black letters on because vertical scrolling could be used to limit the
a dark-grey background and they have been shown to be display to a size range of immediate relevance.
sensitive at detecting retinal and optic nerve disease.
For letters to be 10 pixels high and to represent a 6/3
A significantly different test is the Pelli–Robson low- visual acuity angle the pixels need to subtend 15 seconds
contrast letter chart which measures contrast threshold, of arc at the eye. This means the pixel size would need to
not visual acuity. This chart consists of Sloan letters of be 0.22mm for a 3-m viewing distance, or 0.44mm for 6m.
equal size (49mm) but of progressively decreasing The screen width or height would thus need to be 45 or
contrast. Letters are grouped in 16 sets of three, each set 90cm. Current display screen technology is very close to
decreasing in contrast by 0.15 log units. Those at the top being able to accommodate the high resolution and large
left-hand corner of the chart have 89% contrast whilst area required for a log MAR acuity chart with a size range
those at the bottom right have 0.5% contrast (Pelli et al. extending from 6/60 to 6/3. A particularly interesting
1988). This chart provides a contrast threshold development is the BV-1000 automated subjective
measurement at a single letter size. refractive system, which has recently come to the market
in the UK. The system utilises single Landolt Cs oriented

65
AJ Jackson & IL Bailey

randomly in one of four directions. The optotype sizes access to controlling target characteristics, luminances
follow a logarithmic sequence and presenting visual acuity and layouts will enable clinicians to make more thoughtful
levels range from 6/60 (log MAR =1.0) to 6/3.8 (log MAR = systematic evaluations of the visual capabilities of their
–0.2). The single optotype presentation means the results patients.
are not directly comparable to results from optotypes in
log MAR chart format (Dave 2003).
References
American Medical Association (1984) Evaluation of permanent visual
Acknowledgements impairment. In: AMA Council on Scientific Affairs (eds) Guides to the
Evaluation of Permanent Impairment. AMA Physicians Desk
Reference for Ophthalmology, 2nd edn, pp. 141–51. Chicago
The authors wish to offer particular thanks to Miss E
Association of Optometrists (2002) AOP Members’ Handbook 2002.
Elliman, Mrs J Martin and Ms S O’Connor, whose Section C Vision Standards. London: Association of Optometrists
assistance in preparing the manuscript was invaluable. Bailey IL (1982) Simplifying contrast sensitivity testing. Am J Optom
Physiol Opt 59 (suppl.),12
Bailey IL (1988) Measurement of visual acuity – towards
Conclusion standardisation. In: Vision Science Symposium: A Tribute to Gordon
G Heath, pp. 217–30. Bloomington: Indiana University

Visual acuity, measured with letter charts, is likely to Bailey IL, Bullimore MA, Raasch TW, Taylor HR (1991) Clinical
grading and the effects of scaling. Invest Ophthalmol Vis Sci 32,
remain the principal test of central visual function used by 422–32
clinicians. Advances in medical and optical technology Bailey IL (1998) Visual acuity. In: Benjamin WJ, Borish IM, Lampert
will inevitably lead to eye care practitioners being more R (eds) Borish’s Clinical Refraction, ch. 10. Philadelphia: WB
interested in accurately measuring small changes in Saunders
vision. Practitioners will need better measurements to Bailey IL, Lovie JE (1976) New design principles for visual acuity
letter charts. Am J Optom Physiol Opt 53 (11), 740–5
assess the outcome of sophisticated refractive surgery or
Bennett AG (1965) Ophthalmic test types. Br J Physiol Opt 22,
the benefits achievable through using new designs of 238–71
ophthalmic appliance. Decisions about treatments, access
Blaskoviscs L (1924) The new unit of visual acuity and its practical
to funds to pay for treatment, or eligibility for certain use. Arch Ophthalmol 53, 476–85
benefits and privileges, will increasingly depend more on Borish IM (1970) Clinical Refraction, Visual acuity, 3rd edn, pp.
clinical data, and tightly specified criteria, and less on a 345–422. Chicago, IL: Professional Press
clinician’s declaration of opinion. BSI 4274–1 (1968) Visual Acuity Test Types Pt 1: Specification for
Test Charts for Clinically Determining Distance Visual Acuity.
London: British Standard Institute
The clinical community is slowly moving towards
improving visual acuity measurement. There is an BSI 4274–1 (2003) Visual Acuity Types Pt 1: Test Charts for Clinical
Determination of Distance Visual Acuity: Specifications. London:
increased awareness that chart design affects visual acuity British Standards Institute
scores. There is also a growing acceptance of the need to Bunce C, Evans J, Fraser S, Wormald R (1998) BD8 certification of
standardize significant chart design features such as the visually impaired people. Br J Ophthalmol 82 (1), 72–76
choice of optotype, the number of letters at each size, the Camparini M, Cassinari P, Ferrigno L, Macaluso C (2001) ETDRS –
layout and spacing arrangements of the charts, and the fast: implementing psychophysical adaptive methods to standardise
visual acuity measurement with ETDRS charts. Invest Ophthalmol
sequence of the size progression. Scoring visual acuity in Vis Sci 42 (6), 1226–31
fine steps, giving additional credit for each additional
Coates WR (1935) Visual acuity test letters. Trans Inst Ophthalm
letter read, should become the rule rather than the Opticians III, 1–38
exception as clinicians become more interested in getting Cowan A (1928) Test cards for the determination of visual acuity: a
more precise and reliable measures of vision. review. Arch Ophthalmol NY 57, 283–92
Dave T (2003) Topcon BV 1000 automated subjective refraction
As with the assessment of visual fields, progress in visual system. Optom Today April 4th 1–6
acuity measurement, by real-world clinicians, is likely to Dennett WS (1885) Test type. Trans Am Ophthalmol Soc 4, 133–9
be gradual until such times as computer-based technology Duke-Elder SW (1937) The subjective examination of the eye. In:
becomes adequate and affordable. It will not be long. Duke-Elder SW (ed.) Textbook of Ophthalmology, vol. II, ch. 29.
London: Henry Kimpton
When computer-based acuity measurement is made easy,
Elliott DB, Bullimore MA (1993) Assessing the reliability,
and incorporated into examination routines, clinicians discriminative ability and validity of disability glare tests. Invest
will embrace visual acuity tests with randomly generated Ophthalmol Vis Sci 34 (1), 108–19
letter sequences, efficient automated procedures to Elliott DB, Yang KCH, Whitaker D (1995) Visual acuity changes
establish threshold, and automatic registering of precise throughout adulthood in normal healthy eyes: seeing beyond 6/6.
visual acuity scores with measures of reliability. Ready Optom Vis Sci 72 (3), 186–91

66
 Visual Acuity

Evans J (1995) Causes of Blindness and Partial Sight in England and Raasch T, Bailey I (1984) Choice of optotype and spacing affect visual
Wales 1990–1991. Studies on Medical Population Subjects No. 57. acuity scores. Invest Ophthalmol Vis Sci 25 (suppl. 86), 145
London: HMSO
Regan D (1988) Low contrast letter charts and sinewave grating tests
Ferris FL, Sperduto RD (1982) Standardised illumination for visual in ophthalmological and neurological disorders. Clin Vision Sci 2,
acuity testing in clinical research. Am J Ophthalmol 94, 97–8 235–50
Ferris FL, Kassoff A, Bresnick GH, Bailey IL (1982) New visual acuity Rosser DA, Laidlaw DAH, Murdoch IE (2001) The development of a
charts for clinical research. Am J Ophthalmol 94, 91–6 ‘reduced log MAR’ visual acuity chart for use in routine clinical
practice. Br J Ophthalmol 85, 432–6
Flom MC, Weymouth FW, Kahneman K (1963) Visual resolution and
contour interaction. J Opt Soc Am 53 (9), 1026–32 Sheedy JE, Bailey IL, Raasch TW (1984) Visual acuity and chart
luminance. Am J Optom Physiol Opt 61 (9), 595–600
Green J (1868) On a new series of test letters for determining
acuteness of vision. Trans Am Ophthalmol Soc 1 (3), 68–71 Sloan LL (1951). Measurement of visual acuity. AMA Arch
Ophthalmol 45, 704–25
Green J (1872) Some Improvements in Test Letters. London: IVth
International Ophthalmology Congress Sloan LL (1959) New test charts for the measurement of visual acuity
at far and near distances. Am J Ophthalmol 48 (6), 807–13
Green J (1905) Notes on the clinical determination of the acuteness
of vision, including the construction and gradation of opto-types and Sloan LL, Rowland WM, Altman A (1952) Comparison of three types
on systems of notation. Trans Am Ophthalmol Soc 10 (3), 644–654 of test target for the measurement of visual acuity. Q Rev Ophthalmol
8, 4–16
Haegerstrom-Portnoy G, Brabyn J, Schneck M, Jampolsky A (1997)
The SKILL card: an acuity test of reduced luminance and contrast. Snell AC, Sterling S (1926) An experimental investigation to
Invest Ophthalmol Vis Sci 38 (1), 207–18 determine the percentage relation between macular acuity of vision
and macular perception. In: Menasha G (ed.) Contributions to
Hartridge H, Owen HB (1922) Test types. Br J Ophthalmol 6, 543–549
Ophthalmic Science, pp. 52–62. Banta
Hofstetter HW (1950) The AMA method of appraisal of visual
Snellen H (1862) Letterproeven tot Bepaling der Gezigtsscherpte.
efficiency. Am J Optom Arch Am Acad Optom 27 (2), 55–63
Utrecht: PW Vander Weijer
Hyvarinen L, Nasanen R, Laurinen P (1980) New visual acuity test for
Snellen H (1864) Test Types for the Determination of the Acuteness of
pre-school children. Acta Ophthalmol 58, 507–11
Vision. Official War Office edition (2nd edn). Utrecht: PW Vander
Javal E (1900) Réforme de la notation de l’acuité visuelle. Ann Oculist Weijer
124, 155–9
Sonksen PM, Silver J (1988) The Sonksen Silver Acuity System. Test
Jayatunga R, Sonksen PM, Bhibe A, Wade A (1995) Measures of acuity System and Instruction Manual. Windsor: Keeler
in primary school children and their ability to detect minor errors in
Spaeth EB, Fralick FB, Hughes WF, Scheie HG (1955) American
vision. Dev Med Child Neurol 37, 515
Medical Association council on industrial health: special report,
Jenkins FA, White HE (1976) Fundamentals of Optics, 4th edn, p. estimation of loss of visual efficiency. Arch Ophthalmol 54, 462–8
329. Kogakusha: McGraw–Hill
Stiles WS, Crawford BH (1933) The luminous efficiency of rays
Johnston AW (1991) Making sense of the M, N and log MAR systems entering the eye pupil at different points. Proc R Soc (Lond) Biol 112,
of specifying visual acuity. Probl Optom (Lippincott) 3 (3), 394–407 428–50
Jones D, Westall C, Averbeck K, Abdolell M (2003) Visual acuity Strong G, Woo GC (1985) A distance visual acuity chart incorporating
assessment: a comparison of two tests for measuring children’s vision. some new design features. Arch Ophthalmol 103, 44–6
Ophthalmol Physiol Opt 23 (6), 541–6
Swaine W (1925) The relation of visual acuity and accommodation to
Keeler CH (1956) Visual aids for the partially sighted. Trans ametropia. Trans Opt Soc 9–27
Ophthalmol Soc UK 76, 605–14
Taylor JF (ed.) (1995) Medical Aspects of Fitness to Drive. London:
Landolt E (1899) Nouveaux opto-types pour la détermination de Medical Commission on Accident Preventions
l’acuité visuelle. Arch Ophthalmol Paris 19, 465–471
Tunnacliffe A (1994) Visual acuity updated: 4. Br J Optom Dispens 2
McGraw PV, Winn B (1993) Glasgow acuity cards: a new test for the (12), 455–8
measurement of letter acuity in children. Ophthalmol Physiol Opt 13
Walker JPS (1942) Test type. Br J Ophthalmol 26, 556–559
(4), 400–3
Weber EH (1834) De Pulsu, Resorptione, Auditu et Tactu
McGraw P, Winn B, Whitaker D (1995) Reliability of the Snellen chart.
Annotationes Anatomicae et Physiologicae. Leipzig: CF Koehler
BMJ 310, 1481–2
Westheimer G (1964) Pupil size and visual resolution. Vis Res 4, 39–45
Monoyer M (1875) Echelle typographique décimale pour mésurer
l’acuité de la vue. CR Hebd Séance Acad Sci Paris 80, 1137–8 Westheimer G (1979) Scaling of visual acuity measurement. Arch
Ophthalmol 97, 327–30
NAS–NRC (National Academy of Science: National Research Council
Vision Committee on Vision) (1980) Recommended standard Woo G, Lo P (1980) A Chinese word acuity chart with new design
procedures for the clinical measurement and specification of visual principles. Singapore Med J 21, 689–92
acuity (report 39). Adv Ophthalmol 41, 103–48
Woodruff EW (1947) Visual acuity and the selection of test letters. In:
Neumann AC, McCarty GR, Locke J, Cobb B (1988) Glare disability Some Recent Advances in Ophthalmic Optics, pp. 59–70. London:
devices for cataract: a consumer’s guide. J Cataract Refract Surg 14, Hatton Press
212–16
World Health Organization (1979) Guidelines for Programmes for the
Ogle KN (1953) On the problem of an international nomenclature for Prevention of Blindness. Geneva: World Health Organization
designating visual acuity. Am J Ophthalmol 36, 909–21
Pandit JC (1994). Testing acuity of vision in general practice: reaching
recommended standards. BMJ 309, 1408
Pelli DG, Robson JG, Wilkins AJ (1988) The design of a new letter
chart for measuring contrast sensitivity. Clin Vision Sci 2 (3),
187–200

67
 AJ Jackson & IL Bailey

Multiple Choice Questions

This paper is reference C5333b. Four College credits are available. Please use the inserted answer sheet. Copies can be obtained
from Optometry in Practice Administration, PO Box 6, Skelmersdale, Lancashire WN8 9FW. There is only one correct answer for
each question.

1. The diffraction pattern of the image of a point source 7. Regarding optotypes for visual acuity measurement
of light (a) British Standard letters are all 5x4 letters with serifs.
(a) Has a Gaussian profile. (b) Sloan letters are all 5x5 letters without serifs.
(b) Consists of a series of concentric circles. (c) Keeler chart letters all have stroke widths that are
(c) Has a central disk surround by rings. 20% of the letter height.
(d) Is independent of wavelength.
(e) Is independent of pupil size. (d) Landolt’s Rings were developed before Snellen’s
Letters.
2. Monochromatic light from a point object, focused by (e) Snellen’s original letter set were all 5x5 letters.
the optics of the eye, forms a patch of light on the
retina. The light distribution across the patch is 8. The term “Oxyoptre” as suggested by Blaskovics is a
referred to as the measure of
(a) Ramsden disk. (a) Retinal image blur in diopters.
(b) Petzval surface. (b) Optical defocus.
(c) Interval of Sturm. (c) The reciprocal of the visual angle in degrees.
(d) Contrast sensitivity function (d) Spatial frequency.
(e) Point spread function. (e) Longitudinal chromatic aberration in diopters.

3. In a system free of aberrations or defocus, visual 9. Regarding the designation of visual acuity
resolution (a) The Snellen fraction is a measure of letter height.
(a) Is poorer with smaller pupils. (b) The MAR is the angle that the letter (or optotype)
(b) Is better with smaller pupils. height subtends at the eye.
(c) Is independent of the wavelength of light. (c) If a chart follows the Bailey-Lovie format, each letter
(d) Is better when the point spread function is larger. has a value of 0.02 logMAR units.
(e) Is better if the areas of the foveal cones are larger. (d) LogMAR notation can only be used if there is the
same number of letters on each row.
4. Under photopic conditions, the differential contrast (e) In decimal notation, 2.0 is equivalent to 6/12.
threshold remains a constant ratio of approximately
(a) 80% 10. Which of the following is NOT true about charts that
(b) 20% follow the Bailey Lovie principles
(c) 15% (a) The progression of size should follow a constant
(d) 1% ratio.
(e) 0.1% (b) There should be the same number of letters (or
optotypes) per row.
5. For a 3-bar target, spurious resolution (c) The letters (or optotypes) chosen for each row
(a) Causes vertical lines to appear horizontal. should be balanced for difficulty.
(b) Causes noticeable colored fringes. (d) Between-row spacing should be proportional to letter
(c) Causes a doubling of the spatial frequency. height.
(d) Causes the 3 lines to appear to be 2. (e) Between-letter spacing should be proportionally
(e) Causes the 3 lines to appear to be 4. greater when letter sizes are smaller.

6. Hyperacuities 11. Comparing the merits of row-by-row and letter-by-

(a) Occur only for sinusoidal grating targets. letter scoring methods
(b) Occur only under binocular viewing and bi-foveal (a) If using letter-by-letter scoring, you are more likely
fixation. to get exactly the same score for test and retest.
(c) Occur because of vertical orthophoria. (b) If using row-by-row scoring, you are less likely to
(d) Occur because of pooling information from the two have 1 row difference between test and retest.
eyes. (c) If using row-by-row scoring, you are less likely to
(e) Occur because of pooling of spatial information have 2 rows difference between test and retest.
across receptors. (d) If using letter-by-letter scoring, you ignore errors
made at sizes larger than the smallest row on which
more than 50% of the letters were read correctly.
(e) If using letter-by-letter scoring, you are substantially
more able to identify changes in visual acuity.

Questions for this paper are continued on inside back cover

68

View publication stats