C L I N I C A L

A N D

E X P E R I M E N T A L

Representation of aberrations Atchison

OPTOMETRY
INVITED REVIEW

Recent advances in representation of

monochromatic aberrations of human eyes
Clin Exp Optom 2004; 87: 3: 138148
David A Atchison MSc(Optom) PhD
GradCertEd FAAO FOSA
School of Optometry, Queensland
University of Technology, Queensland,
Australia
Submitted: 31 January 2004
Revised: 29 April 2004
Accepted for publication: 1 May 2004

The field of aberrations of the human eye is moving rapidly, being driven by the desire
to monitor and optimise vision following refractive surgery. In this paper, I discuss the
different ways of representing aberrations of the human eye, the terminology used, how
wave aberrations are used to determine refractions, the influence of pupil size on aberrations, how to compare right and left eye aberrations, how aberrations can be manipulated into different forms, how to make corrections for changes in wavelength, the appropriate ocular axis, and corneal and lenticular components of the aberrations.

Key words: aberrations, refractive correction, wavefront, Zernike polynomials

Prior to the past decade, the field of aberrations of the human eye was almost solely
the preserve of visual optical scientists,
such as myself. The reader may ask why
ophthalmic practitioners should now be
concerned about aberrations when they
have managed for years without them.
There is a number of reasons. In normal
eyes, the conventional sphero-cylindrical
errors dominate but when these are corrected, higher-order aberrations limit our
visual capabilities. In abnormal eyes with
conditions such as keratoconus, higherorder aberrations can be more important
than the conventional refractive errors.
Some treatments for refractive errors, such
as corneal refractive surgery, intra-ocular
lens implants and orthokeratology, can
unintentionally introduce considerable
amounts of higher-order aberrations.
For convenience, aberrations of optical
systems are usually divided into two

groups, the monochromatic and the chromatic aberrations. As the names suggest,
the first group refers to aberrations that
are present when only one wavelength of
light is considered, while the second group
refers to the additional aberrations that
occur when the chromatic dispersion effects of optical media are considered. This
review is about the first group, although
the reader should keep in mind that the
monochromatic aberrations at one wavelength are different from those at another
wavelength. This is particularly important
when dealing with instruments that measure refraction and/or aberrations at infrared wavelengths, and correction is required for the visible spectrum.
Aberrations are often considered to be
optical defects additional to refractive errors but with the use of Zernike aberration terminology and with instruments
that measure both, the distinction between
Clinical and Experimental Optometry 87.3 May 2004

138

aberrations and refractive errors is becoming less useful. It is becoming more common to refer to the refractive errors as
lower-order aberrations and to other effects, such as spherical aberration and
coma as higher-order aberrations. I would
like to emphasise to readers that, for normal eyes at least, the refractive errors are
the most important optical defects and
higher-order aberrations are of consequence only once these refractive errors
are corrected.
This paper addresses the representation
of monochromatic aberrations. Another
paper will consider methods used to measure aberrations, the magnitudes of aberrations in the general population and how
these are affected by various factors (for
example, age, refractive error, accommodation and refractive surgery) and how
aberrations affect spatial visual performance.

Representation of aberrations Atchison

Figure 1. Wave, transverse and longitudinal aberrations of an optical system in (A) image space and (B) object space. (B) includes the OSA
co-ordinate system in the pupil. For clarity, the aberrations have been exaggerated and the object plane in (B) has been placed close to the eye
rather than at infinity.

TAYLOR ABERRATIONS AND

ZERNIKE ABERRATIONS

Aberrations of any optical system can be

represented in three ways (Figure 1a):

1. Wave aberration
We can image a spherical wavefront moving outwards from an originating point
object. The aberrations of an optical system, such as the eye, prevent the wavefront
from remaining spherical as it passes
through the system. This aberrated
wavefront can be compared with an ideal
spherical wavefront, whose centre of curvature on the image side of the system is
at the ideal image position. For the ideal
wavefront the image point is located in a
position so that it produces an image that
is a perfect copy of any object (allowing
for magnification differences and diffraction effects). The aberrated wavefront may
be compared with the ideal wavefront at
any point in the optical system but a convenient position for doing this is at the exit
pupil of the system. This is the image of
the limiting aperture (stop) as seen from
the image side of the system. At any point
on the exit pupil, the departure of the
aberrated wavefront from the ideal

wavefront, multiplied by the refractive index in the vitreous, gives the wave aberration associated with that point. If the
wavefront is ahead of the ideal wavefront,
the wave aberration is positive. Wave aberrations are small quantities and are usually expressed in micrometres or wavelengths. At a wavelength of 500 nm, one
micrometre (m) is equivalent to two
wavelengths.

2. Transverse aberration
Rays describe the path of a particular point
of the wavefront through the optical system. The transverse aberration of a single
ray is its departure from its ideal position
at the image surface. Usually it has both
horizontal (') and vertical components
('). Transverse aberrations are often
expressed as distances in millimetres or as
angles in minutes of arc.

3. Longitudinal aberration
This is the departure of the intersection,
where it occurs, of a ray with a reference
axis away from its ideal intersection. It may
have different values in the horizontal and
vertical planes. It may be measured as a
length (typically millimetres) or as a power
(dioptres).
In the case of the eye, we do not have
Clinical and Experimental Optometry 87.3 May 2004

139

access to the image side of the optical system and so it is sometimes convenient to
start with the object at the image position
(on the retina) and then to measure the
aberrations of the eye on the object side
of the system (Figure 1b). Instead of being referred to the exit pupil, they are referred to the entrance pupil, which is the
image of the stop (the iris) as seen from
the object side of the optical system. Again,
the wave aberration is positive where the
wavefront is ahead of the ideal wavefront.
As the ideal condition for the eye is emmetropia, the ideal wavefront leaving the
eye will be a plane, although for reasons
of space I have shown the ideal wavefront
as converging to a plane near the eye in
Figure 1b.
The wave aberrations of an optical system such as the eye can be represented by
a wave aberration polynomial function
W(X, Y) = W1 + W2X + W3Y + W4X 2 + W5XY
+ W6Y 2 + W7X 3 + W8X 2Y + W9XY 2 + W10Y 3
+ W11X 4 + W12X 3Y + W13X 2Y 2 + W14XY 3
+ W15Y 4 + higher order terms
[1]
where X and Y are horizontal and vertical
co-ordinates in the entrance pupil. This
polynomial function is sometimes referred
to as a Taylor series. The terms are ar-

Representation of aberrations Atchison

ranged in orders according to the sum of

their powers in X and Y. As an example,
W14X Y 3 is a fourth-order term. The coefficients of some of the terms can be described as follows:
W1: piston coefficient, usually set to zero
so that the aberration at the centre of
the wavefront is zero
W2 and W3: tilt (prismatic or distortion)
coefficients
W4, W5 and W6: defocus and astigmatism
coefficients
W7, W8, W9 and W10: coma-like coefficients
W11, W 12, W 13, W 14 and W 15: spherical
aberration-like coefficients
The values of these coefficients depend
on the position of the object in the field.
If the optical system is rotationally symmetrical and the object point is along one
meridian, say the X-axis, some of the terms
drop out and other terms can be equated,
to leave us with
W(X,Y ) = W1 + W2X + W4(X 2 + Y 2) + W6X 2
+ W7(X 2 + Y 2)X + W11(X 2 + Y 2)2 + higher
order terms
[2]
Here, the third term is defocus, the fourth
term is astigmatism, the fifth term is primary coma and the sixth term is primary
spherical aberration.
As eyes are not rotationally symmetrical
about an appropriate reference axis, ocular aberrations are not described well by
equations such as equations (2) but require the use of an equation such as equation (1). The lack of rotational symmetry
can be attributed to the departure of the
line of sight from the best fit to the optical axis (because of component tilts and
displacements), no rotational symmetry of
refracting surfaces, and possibly no rotational symmetry of the lens refractive
index.
A problem with the Taylor series is that
the terms are not independent of one another. This means that as additional terms
are added to the series, the lower-order
terms are altered. The Taylor series is not
the only mathematical series capable of
representing the wave aberrations of an
optical system. One series in which the
terms are independent and which has

gained considerable recent popularity is

the Zernike polynomial function series.
Zernike aberrations have several useful
features as discussed below. Different numbering and order schemes exist but here I
use the OSA system that is gaining widespread acceptance in the vision community1 and has been adopted by the American National Standards Institute. 2
Formally, it has the form
k

W(,) =

cnmZ nm(,)

[3]

n = 0 m = -n
n - |m| = even

m
n
m
n

|m|
n
|m|
n

{NN RR

()cos(m), for m 0
()sin(|m|), for m < 0
[4]

where Rn|m| is a radial polynomial that is a

function of , Nnm is a normalisation term
and |m| is the absolute value of m. The Rn|m|
term is given by
Rn|m|()
(n - |m| )/2

s=0

Nnm = 2(n + 1)

(-1)s(n-s)!
n-2s
s![0.5(n + |m|)-s]![0.5(n - |m| - s]!
[5]

where the index n is the highest power of

the radial polynomial and the index m
describes the meridional frequency of the
sinusoidal component of the Zernike polynomial. The normalisation term N nm is
given by
Nnm = n + 1

[6a]

Clinical and Experimental Optometry 87.3 May 2004

140

[6b]

for m 0. The significance of Nnm will be

explained shortly.
The description of wave aberrations
above in terms of Zernike polynomials uses
a double indexing scheme in n and m. Although this is the preferred scheme, there
is a single index scheme that can be represented as the wave aberration series

with W(,) being a polar representation

of the wave aberration, Z nm a particular
Zernike polynomial and c nm the coefficient
of the Zernike polynomial. is the relative distance from the centre of the pupil
and thus ranges from zero to 1. is a meridian measured in radians and ranging
from zero to 2. It is measured from the
positive X-axis (to an observers right when
looking at a persons eye) with an anticlockwise angle (Figure 1b). This corresponds to the angular system used in ophthalmic optics. k is the maximum order of
the polynomial series. The Zernike polynomial function Z nm is defined as
Z nm(,) =

for m = 0 and

W(,) = c j Z j

[7]

j=0

where Z j is the j th Zernike polynomial function and c j is its coefficient. j is related to

n and m by
j = [n(n + 2) + m]/2

[8]

The number j, order n and frequency m

of the Zernike polynomials are given in
Table 1 up to the sixth order and more
extensive lists are available in other references.3
Figure 2 shows the Zernike polynomial
function pyramid in which order n
changes vertically and frequency m
changes horizontally. Plots are given as a
function of position in the pupil for terms
up to the fourth-order. The Zernike polynomials vary across the pupil with values
within the range -Nnm and +Nnm. I will now
explain some of these polynomial functions.
The order of the first row is n = 0. The
single function has a value of 1 and is
called piston. It does not have much significance, except that its coefficient is usually manipulated to make the wave aberration zero in the centre of the pupil.
The order of the second row is n = 1,
that is, these are the first-order functions.
These are tilts or prisms, one in the (vertical) y - direction and one in the (horizontal) x - direction. They are rotated functions of one another. This rotational
nature continues down the pyramid, so
that in any row a function with a negative
value of the index m is a rotated form of
the function with the same but positive
number for m.
The order of the third row is n = 2, that

Representation of aberrations Atchison

Number Radial
j
order n

Frequency
m

Polar representation

Monomial representation

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

0
-1
1
-2
0
2
-3
-1
1
3
-4
-2
0
2
4
-5
-3
-1
1
3
5
-6
-4
-2
0
2
4
6

1
2sin()
2cos()
62sin(2)
3(22-1)
62cos(2)
83sin(3)
8(33-2)sin()
8(33-2)cos()
83cos(3)
104sin(4)
10(44-32)sin(2)
5(64-62+1)
10(44-32)cos(2)
104cos(4)
125sin(5)
12(55-43)sin(3)
12(105-123+3)sin()
12(105-123+3)cos()
12(55-43)cos(3)
125cos(5)
146sin(6)
14(66-54)sin(4)
14(156-204+62)sin(2)
7(206-304+122-1)
14 (156-204+62)cos(2)
14(66-54)cos(4)
146cos(6)

1
2y
2x
26xy
3(2x2+2y2-1)
6(x2-y2)
8(3x2y-y3)
8(3x2y+3y3-2y)
8(3x3+3xy2-2x)
8(x3-3xy2)
10(4x3y-4xy3)
10(8x3y+8xy3-6xy)
5(6x4+12x2y2+6y4-6x2-6y2+1)
10(4x4+4x2y2-3x2-4x2y2-4y4+3y2)
10(x4-6x2y2+y4)
12(5x4y-10x2y3+y5)
12(15x4y+10x2y3-12x2y-5y5+4y3)
12(10x4y+20x2y3+10y5-12x2y-12y3+3y)
12 10x5+20x3y2+10xy4-12x3-12xy2+3x)
12(5x5-4x3-10x3y2-15xy4+12xy2)
12(x5-10x3y2+5xy4)
14(6x5y-20x3y3+6xy5)
14(24x5y-20x3y-24xy5+20xy3)
14(30x5y+60x3y3+30xy5-40x3y-40xy3+12xy)
7(20x6+60x4y2+60x2y4+20y6-30x4-60x2y2-30y4+12x2+12y2-1)
14(15x6+15x4y2-20x4+6x2-15x2y4-15y6+20y4-6y2)
14(6x6-5x4-30x4y2-30x2y4+30x2y2+6y6-5y4)
14(x6-15x4y2+15x2y4-y6)

0
1
1
2
2
2
3
3
3
3
4
4
4
4
4
5
5
5
5
5
5
6
6
6
6
6
6
6

Table 1. Zernike polynomials, up to the sixth-order, in polar and monomial representation

is, these are second-order polynomial functions. These are important for determining refraction. The central term is
rotationally symmetric and is called
defocus. Its equivalent as a longitudinal
aberration is mean spherical equivalent
(also called mean sphere). On either side
of it are the astigmatisms. The polynomial
function for which m = -2 is called oblique
astigmatism and has a maximum value of
+2 along the 45 degree meridian and a
minimum value of -2 along the 135 degree
meridian. The other polynomial function,
for which m = +2 is the same but rotated
by -45 degrees to have a maximum value
along the horizontal meridian of +2 and a
minimum value along the vertical merid-

ian of -2. Knowing the pupil size, the

defocus and astigmatism coefficients can
be converted into a conventional refraction (see below).
The order of the fourth row is n = 3.
These are the first of what are referred to
as the higher-order polynomial functions.
The two functions in the middle are
comas, one vertical and one horizontal.
The order of the fifth-row is n = 4. The
middle one of these is the rotationally symmetric polynomial function known as
spherical aberration. If its coefficient is
positive, we can consider the periphery of
the pupil to be myopic relative to the centre of the pupil. Spherical aberration will
be mentioned below.
Clinical and Experimental Optometry 87.3 May 2004

141

I have shown only the first few orders of

the Zernike pyramid in Figure 2. As you
move down the pyramid and away from
the centre, the shapes of the terms get
more complicated. As you go down the
pyramid, the number of positive and negative regions continues and as you move
away from the mid-line, the number of
repetitions of the pattern increases as you
go around the pupil.
As mentioned above, the Zernike polynomials are independent of or orthogonal
to one another. This means that when you
multiply one polynomial by another at all
points in the pupil and sum the results,
the answer is zero. Formally, for two
Zernike polynomial functions Z mn and Z m'n',

Representation of aberrations Atchison

) in object space, shown in Figure 1,

which are related to the wave aberration
by partial derivatives: 4

Figure 2. Three-dimensional representation of Zernike function pyramid up to the fourthorder

provided either n n' or m m', this is

expressed as
= 2 = 1

Z nmZ n'm' dd = 0

=0 =0

[9]

Each Zernike polynomial function can

be given a monomial representation, obtained by converting it from a polar to a
Cartesian (x, y) co-ordinate system, where
(x, y) are normalised values of the actual
pupil co-ordinates and
x = cos(), y = sin()

[10]

Both polar and monomial representations

of Zernike polynomial functions are
given in Table 1. Most Zernike polynomial
functions contain various orders of the
Taylor series, for example, the monomial
representation of Zernike polynomial
function Z 5-1 (secondary vertical coma) is
10x 4y + 20x 2y 3 + 10y 5 - 12x 2y - 12y 3 + 3y,
which includes first-, third- and fifth-order

Taylor terms. Each Zernike polynomial

function can be referred to as balanced.
This means that the highest order Taylor
term appearing in it is balanced by lowerorder Taylor terms to minimise the variance of the Zernike polynomial function
across the pupil.
When representing aberrations of an
optical system, a decision has to be made
on the level of aberrations required. For a
Taylor series, as higher terms are added,
fitting techniques will change the lowerorder Taylor terms. Because of the
orthogonal nature of the Zernike polynomial series, at least theoretically adding
additional polynomial functions will not
change the coefficients of the lower-order
functions.
The fitting of Zernike polynomials assumes fitting a continuous function over
a circle but in practice, an optical system
has to be sampled discretely. Usually, aberrations are obtained as transverse aberrations (', ') in image space or (,
Clinical and Experimental Optometry 87.3 May 2004

142

' or =

-l d W(X,Y )

dX

' or =

-l d W(X,Y )

dY

[11]

where l is the radius of the reference

wavefront and is the refractive index in
the object or image space (depending on
the space in which the aberrations are
being measured). Alternatively, transverse
aberrations can be derived in angular
terms, in which the expression -l/ is deleted from the right hand sides of both
equations (11). A least squares fitting procedure is used to determine the coefficients. Whereas Zernike polynomials are
orthogonal, their derivatives are not.
Hence in practice, a fitting of additional
Zernike polynomials may lead to some
change in underlying terms. These
changes are minimised when the number
of sampled points is much larger than the
number of terms to be fitted, and there is
regular sampling across the pupil. If required, there are procedures available to
make the derived functions orthogonal,
for example the Gram-Schmidt procedure.5
A normalisation term Nnm appears in each
Zernike polynomial function (equation
[4], equation [6], Table 1). This makes the
variance of each Zernike polynomial function as 1. This means that the square of
the coefficient of the Zernike polynomial
function represents the contribution of
that polynomial function to the total variance of the wave aberration. Summing the
squares of the coefficients for Zernike
polynomial functions of a particular order
gives the contribution of that order to the
variance. Here, variance has the same
meaning as it does in statistics. When determining the total variance, it is customary to ignore both the piston coefficient,
the value of which is arbitrary, and the tilt
coefficients as although these contribute
to image position they do not contribute
to its quality (at least in monochromatic
light). Many instruments use the rootmean-square (RMS) of the wave aberration

Representation of aberrations Atchison

(the square root of the variance) as a measure of image quality. In mathematical

terms
RMS =

(Cnm )2

n >1, all m

[12]

As the coefficients cnm are relative to the

size of the pupil, they must change if pupil
size changes. The change is not simple
because of the different Taylor terms appearing in a Zernike polynomial function.
There are different ways in which Zernike
coefficients, determined at one pupil size,
can be determined at a second pupil size.
One way is to make measurements at the
second pupil size, or in the case of a smaller
pupil size isolate the measurements corresponding to it. The latter procedure is
used in at least some commercial instruments. Another way is to convert Zernike
coefficients to Taylor coefficients and then
use these to recalculate the Zernike coefficients at the second pupil size.6 There
are other methods7,8 and in my experience
these give similar results. Although it is
valid to interpolate from a larger to a
smaller pupil size, the reverse procedure
is not advisable, nor possible with the
aberration-measuring instruments with
which I am familiar. Comparing aberrations of different eyes, or the aberrations
of a single eye at different times, is valid
only if this is done at the same pupil size.
Table 2 and Figure 3 show some of the
points made in the previous paragraphs
from the practitioners viewpoint. Table 2
shows the first 28 Zernike polynomial functions and their coefficients for a patients
eye, with the latter determined for fourand six-millimetre diameter pupils. The
names for several Zernike polynomials
have been included. The considerable differences between the magnitudes of the
coefficients at the two pupil sizes are evident, with change in sign for some of the
terms. The piston and tilt terms are usually ignored in the calculation of RMS as
described above. That leaves us with the
second- and higher-order aberrations. The
second-order aberrations are correctable
with lenses, so as shown in the Table, sometimes an instrument might give both the
total RMS (from second-order and higherorder coefficients) and the higher-order

Zernike
Name
polynomial

Coefficient
4 mm pupil ( m)

Coefficient
6 mm pupil ( m)

Z 00
Z 1-1
Z 11
Z 2-2
Z 20
Z 22

0.427
0.707
0.859
-0.025
0.309
-0.276

1.733
1.214
1.404
-0.034
1.356
-0.456

0.118
0.025
0.015
0.039
0.007
-0.009

0.228
0.098
0.046
0.141
0.006
-0.017

0.045

0.357

0.022
0.004
-0.006
-0.024
-0.002
-0.007
0.007
-0.001
0.010
0.009
0.001
-0.002
-0.001
-0.013
0.011
0.439
0.143
-0.20/-0.68 x 3
+0.21/-0.89 x 179
+0.28/-0.92 x 178

0.053
0.009
-0.007
-0.025
-0.015
0.008
-0.007
-0.027
0.015
-0.019
0.003
0.069
-0.011
0.014
-0.010
1.507
0.472
-0.79/-0.50 x 2
+0.38/-0.72 x 179
-0.05/-0.84 x 178

Z 3-3
Z 3-1
Z 31
Z 33
Z 4-4
Z -24
Z 40
Z 42
Z 44
Z 5-5
Z 5-3
Z 5-1
Z 51
Z 53
Z 55
Z 6-6
Z 6-4
Z 6-2
Z 60
Z 62
Z 6-4
Z 6-6

Piston
Vertical tilt
Horizontal tilt
Oblique astigmatism
Defocus
With-/against-the-rule
astigmatism
Oblique trefoil
Vertical coma
Horizontal coma
Horizontal trefoil
Oblique quatrefoil
Oblique secondary
astigmatism
Spherical aberration
With/against the rule
secondary astigmatism
Quatrefoil

Secondary vertical coma

Secondary horizontal coma

Secondary spherical aberration

Total RMS
Higher order RMS
Refraction 2nd order
Refraction, 2nd and 4th orders
Refraction, 2nd, 4th and 6th orders

Table 2. Zernike aberration coefficients of the eye, the wave aberration maps of which are
shown in Figure 3, and derived quantities

RMS (from third-order and above coefficients) as shown in the fifth-last and fourthlast rows, respectively, of the Table.
Figure 3 shows a three-dimensional map
of the patients aberrations, in which the
height represents the magnitude of aberration as a function of position in the
pupil. A projection (contour map) of the
aberrations is shown as part of this figure.
In Figure 3a, the tilt has been removed
(or set to zero) and the piston manipuClinical and Experimental Optometry 87.3 May 2004

143

lated so that the wave aberration is zero at

the centre of the pupil. In Figure 3b, the
second-order terms have also been set to
zero, to show just the higher-order aberrations. Note the change of scale between
Figures 3a and 3b.
SPHERICAL ABERRATION
One aberration that has been of considerable interest to investigators for several

Representation of aberrations Atchison

Figure 3. Three-dimensional wave aberration maps with a 6 mm pupil of a patient following

refractive surgery. The left map includes aberrations of the second- and higher-orders and
the right map shows only the higher-order aberrations. The same colour scale in 0.05 m
steps has been used for both maps but the physical scales are different.

decades is spherical aberration (Figure 2).

Although early investigators knew that
ocular aberrations were not usually
rotationally symmetrical within the pupil,
many have considered the aberration
measured to be conventional primary
spherical aberration, a rotationally symmetrical aberration given by the term
W11(X 2 + Y 2)2 in equation (2). Care needs
to be taken when interpreting results in
early studies, particularly where measurements have been made along a single
meridian. Smirnov 9 and Howland and
Howland10,11 re-emphasised the asymmetric nature of the eyes aberrations from the
measurements with two-dimensional psychophysical raytracing and the crossedcylinder aberroscope, respectively. The
spherical aberration term continues to
intrigue: as represented by the fourthorder Zernike aberration term c 04 5(64 62 + 1), its coefficient is positive for most
people in the unaccommodated state12-14
and its change following refractive surgery
is a useful monitor of the quality of the
optics of the eye produced by that surgery.
With (X,Y ) being the co-ordinates in the
pupil, the coefficient W11 in the expression
W11(X 2 + Y 2)2 in equation (2) gives the
magnitude of primary spherical aberration
for one millimetre pupil semi-diameter.
Where the aberration is high such as for
dilated pupils, there may be need for a

secondary spherical aberration term that,

following the numbering system in equations (1) and (2), would be W22(X 2 + Y 2)3.
Atchison, Cox and Scott3 provided equations showing conversions between Taylor
and Zernike polynomial coefficients, although without using normalising terms.
In the case of the fourth-order Zernike
aberration coefficient c 04 and using the corresponding normalisation term, the conversion from Taylor coefficients up to the
eighth order is given by

( W16 + W48 + W16 )R

W W
W W
+(
+
+
+
R
64
64 64 64)
W
W W W W
+(
+
+
+
+
R
64 448 448 448 64 ) [13]

5c 04 =

11

13

15

22

24

26

28

37

39

41

43

45

where W11, W13, W15, W22, W24, W26, W28, W37,

W39, W41, W43 and W45 are coefficients of X 4,
X 2 Y 2, Y 4, X 6, X 4 Y 2, X 2Y 4 , Y 6 , X 8 , X 6Y 2,
X 4Y 4, X2Y 6 and Y 8, respectively.

HOW MANY ZERNIKE ABERRATION

TERMS ARE NEEDED TO FIT THE
WAVE ABERRATION?
Above, I wrote that because of the
orthogonal nature of the Zernike series
Clinical and Experimental Optometry 87.3 May 2004

144

and at least theoretically, adding additional terms will not change the lower
terms. The Zernike fit to a wavefront aberration should contain a sufficient
number of terms to model the wavefront
aberration as closely as possible. At some
point, the error of the fit associated with a
coefficient becomes as significant as the
coefficient itself, so that the contribution
of that Zernike polynomial to the model
is no better than adding random noise. It
is important to have an appropriate test
for determining the optimal number of
terms to fit15,16 but these are not yet available in commercial instruments. Using
inappropriate terms can contribute to
giving poor estimates of refraction (see
below). Another thing that must be considered is that taking multiple measurements of the eye may give considerable
variation relative to the mean of some coefficients, which may not be significantly
different from zero.17 As a rough guide,
from my experience for eyes that seem to
have normal optics, there is probably no
need to go above the sixth-order for pupils up to at least six-millimetre diameter.
There may be occasions where higherorder aberrations are needed, for example, in radial keratotomy with eight cuts
at regular intervals, there may be significant higher-order coefficients such as c -88
and c 88.18
USING ZERNIKE ABERRATIONS TO
DETERMINE REFRACTION
INFLUENCE OF PUPIL SIZE
One possible criterion for refraction is to
provide a correcting lens that minimises
the RMS of the wave aberration. This is
achieved by neutralising the second-order
Zernike aberrations. This will be confusing to the practitioner because in the
presence of considerable higher-order
terms, particularly in the fourth-order,
refraction will be dependent on the
pupil size. As an example, comparing
refractions on this basis for pupils of twoand six-millimetre diameter can give
discrepancies of up to one dioptre in
mean spherical equivalent,6 but subjective
refractions show little change across a
wide range of pupil sizes.19-22

Representation of aberrations Atchison

Another criterion is to neutralise the

Taylor defocus and astigmatism terms appearing in equation (1). This is a paraxial
refraction, in which we are determining
the refraction based on rays traced into/
out of the eye for a small central part of
the pupil, that is, we are ignoring the influence of rays passing through the pupil
periphery. The Taylor defocus and astigmatism terms appear in even-order
Zernike polynomial functions and the
equations using Zernike coefficients to
convert these into a correction are:6
M = - (43c 02 - 125c 40 + 247c 06 - 409c 08
+ 6011c 100 - )/R 2
[14]
J180 = -(26c 22 - 610c 24 + 1214c 62 - 2018c 82
+ 3022c 102 - )/R 2
[15]
J45 = -(26c -22 - 610c -24 + 1214c -26 - 2018c -28
+ 3022c 10-2 - )/R 2
[16]
Here M is the mean spherical equivalent,
J180 represents a (Jackson) cross-cylinder
that has powers of equal magnitude but
opposite sign along the 180 and 90 degree
meridians and J45 represent a cross-cylinder that has powers of equal magnitude
but opposite sign along the 45 and 135
degree meridians, c 02, c 40 et cetera are
Zernike coefficients, and R is the pupil
semi-diameter for which the aberrations
apply. The negative signs in front of the
left brackets are to convert from an error
into a correction.
The M, J180, J45 system is an excellent way
to represent the refraction, being readily
amenable to statistical manipulation. Conversion to conventional sphero-cylinder
form S/C x with a negative cylinder is
given by23
C = -2(J1802 + J452)

[17]

S = M - C/2

[18]

= [tan-1(J45/J180)]/2

[19]

If J180 is 0, there will be an indeterminate

result. In this case, if J45 < 0, = 135 degrees
and if J45 0, = 45 degrees. To keep
within the clinical conventional range of
0180 degrees, the following equations

must be applied:
If J180 < 0
= + 90

[19a]

If J180 0 and J45 0

= + 180

[19b]

For completeness, the reverse of equations (17) to (19) are:24

M = S + C/2

[20]

J180 = -(C/2) cos(2)

[21]

J45 = -(C/2) sin(2)

[22]

The two sets of equations are useful for

manipulations when refractions are to be
combined for example, averaging obliquely crossed-cylinders.
There has been recent investigation into
how Zernike aberrations should be used
to determine objective refractions that will
best match subjective refractions. In a
study of 200 eyes, Thibos and associates13
found that an approximation to the paraxial criterion using the second- and fourthorder terms in equations (14) to (16) gave
a superior match to subjective refractions
compared with the minimum RMS criterion. The differences between objective
and subjective refractions for six millimetre pupils were -0.3 0.3 D (more myopic
for mean objective refraction) and 0.0
0.3 D for minimum RMS and paraxial
refractions of the objective data, respectively. The poorer match for the minimum
RMS criterion is probably because it places
too much emphasis on the outer regions
of the pupil.
An example of the use of different orders of Zernike polynomials to determine
an objective refraction is shown in the last
three rows of Table 2 with four- and sixmillimetre diameter pupils. This patient
has very high spherical aberration following LASIK surgery (c 04 coefficient +0.358
m for a six-millimetre pupil, compared
with a normal population mean of about
0.10 0.10 m14,25-27) and this has a considerable effect on refraction estimates.
When the second-order alone is used, the
Clinical and Experimental Optometry 87.3 May 2004

145

difference in refraction is 0.51 D but

reduces to approximately 0.25 D when
either fourth- or both fourth- and sixthorders are used as well as the secondorder.
Recently, a number of aberration-based
image quality metrics, such as the volume
under the three-dimensional modulation
transfer function, has been investigated
and many have been found to provide
better estimations of subjective refraction
than does the paraxial objective refraction
criterion.28-32 It is likely that commercial
instruments will use such metrics to
estimate refraction.
LEFT AND RIGHT EYE DATA
When right and left eyes are compared,
or both right and left eyes data are used
in pooled data sets, account should be
taken of the expected nasal-temporal
asymmetry in both eyes. This can be done
by treating right eye and left eye data separately, or by altering the signs of some of
the Zernike coefficients for one eye. 12
ANSI Z80.28 20042 recommends altering signs of left eye coefficients, for which
the Zernike polynomial functions have
negative, even m indices (such as Z -22 , Z -44
and Z -24 ) or positive, odd m indices (such
as Z 13 and Z 33).
THE SPHERICAL EQUIVALENT TO
WAVE ABERRATIONS
Most ophthalmic practitioners are more
familiar with longitudinal aberrations than
with wave aberrations. To help practitioners, Thibos33 proposed that an aberration
coefficient be converted to a spherical
equivalent SE as
SE = -43c nm/R 2

[23]

where R is the pupil semi-diameter. This

equation uses the first term on the right
hand side of the refraction equation (14)
and assumes that coefficients of the same
magnitude, but of different Zernike aberration polynomials, have the same effect
on vision. A spherical equivalent based on
a range of wave aberration coefficients can
be determined where

Representation of aberrations Atchison

and c 33 of the trefoil polynomial functions

Z 3-3 andZ 33 are +0.228 m and + 0.141 m,
respectively, as given in the last column of
Table 2, the combined Zernike polynomial
function Z33 is given by
Z33(,,) = 83 cos[3( - 19.4o)]

Figure 4. Two-dimensional wave aberration maps with a 6 mm pupil of a patient following

refractive surgery. The left map shows the aberration with the polynomial function Z -33, the
middle map shows the aberration with the polynomial function Z33 and the right map shows
the combination of the two aberrations to give the new Zernike polynomial function Z 33 with
a coefficient c33 .

SE = -43 (c nm)2/R 2

[24]

Here the spherical equivalent could be

calculated for second- and higher-order
aberration orders or perhaps just the
higher-orders of aberration. This formula
provides a way of comparing image quality of eyes with different pupil sizes.
MANIPULATIONS OF ZERNIKE
ABERRATIONS AND OTHER
ABERRATION SYSTEMS
The Zernike polynomial system of describing aberrations is increasing quickly in
acceptance but it has its limitations. It is not
always the best way to describe a measured
wavefront error if there are local irregularities of significance present, because the
number of terms needed becomes large if
a good fit to the data is desired. At least
one manufacturer shows aberration plots
taking local irregularities into account.
Porter and colleagues12 and Thibos and
associates13 investigated the existence of a
more efficient set of functions than
Zernike polynomials, accounting for as
much of the variance of population data
with the minimum number of functions.
Their principal component functions
were linear combinations of Zernike aberration polynomials but they provided
only a slightly more efficient description
of population data.
Campbell34 suggested a way to reduce
the number of Zernike polynomial func-

tions by a factor of approximately two, by

combining polynomial functions in a way
similar to that given for the astigmatic components (equations [17] and [19]). This
method is also described in the ANSI
standard.2 The polynomial functions to be
combined are those of the same radial
order, n, and having meridional indices m
of the same magnitude but with opposite
signs. This gives new polynomial functions:
Znm(, , )
= Nnm Rn|m|()[cos(m( nm)]

[25]

where here m takes only positive values

and Rn|m|() and Nnm are already defined by
equations (5) and (6), respectively. Coefficients and angular parameters are given
by
cnm = (c -m
) 2 + (c mn )2
n

[26]

and
tan-1

nm =

-m
n
m
n

( cc )
[27]

|m|

The wave aberration across the pupil is

now given by
W(, ) =

cnmZnm (,,nm)

[28]

all n and m

instead of equation (3). Campbell 34

showed contour maps of the derived
Zernike polynomial functions. As an example of this system, if the coefficients c 3-3
Clinical and Experimental Optometry 87.3 May 2004

146

and is accompanied by the coefficient

c33 = +0.268 m (Figure 4).
Although Campbells system reduces the
number of Zernike polynomial functions,
it introduces an angle so there is no reduction in the amount of information needed.
Other polynomial systems have been
investigated including Bhatia-Wolf polynomial functions, which, as well as being
orthogonal in relative x and y pupil coordinates like Zernike polynomial functions, are also orthogonal in . Iskander
and co-workers35 found the Bhatia-Wolf
polynomials to model corneal elevation
data better than Zernike polynomials but
more polynomials are needed for the
former for a particular radial order n. This
system of polynomials may be useful in fitting irregular wave aberration patterns but
at the expense of appreciation of the
importance of individual terms.
CHANGE OF WAVELENGTH
Most instruments use infrared wavelengths
and a correction in the defocus coefficient
can be made for visible wavelengths using
the Indiana Chromatic eye model,36 an
adaptation of Emsleys reduced eye,37 that
models well the longitudinal chromatic
aberration of human eyes. The change in
mean spherical equivalent correction M
for this eye, from the infrared to the visible, is given by the hyperbolic equation
M = 633.46[1/( IR - 214.102) - 1/(v
- 214.102)]
[29]
where IR is the infrared wavelength of an
instrument and v is a visible wavelength,
both in nanometres. As an example, for
the wavelengths 840 nm and 550 nm, M
is -0.87 D. From this equation and a truncated form of equation (14), the corresponding change in defocus wave aberration coefficient c 02 is

Representation of aberrations Atchison

c 20 = -M.R 2/(43)

[30]

refraction corrections using modified

forms of equations (14) to (16)

where R is the pupil semi-diameter. Probably, equation (29) gives slight overestimates because it does not take into
account the increased penetration of
infrared radiation compared with that of
visible radiation. There are also subtle
effects depending on the length of the eye
and its refractive error.38 One manufacturer gives a complex equation that allows
for increased penetration into the eye of
0.125 mm and a changing eye length depending on refractive error.39 It gives an
approximate mean spherical correction
from 840 nm to 550 nm of -0.7 D, with
very small changes for wave aberration
coefficients other than c 02.

where is the off-axis angle.

OFF-AXIS ABERRATIONS

REFERENCE AXIS

Off-axis refraction has been measured for

several years using various techniques.
Often the findings were given as refraction components in the horizontal (H)
and vertical meridians (V ). The relationship between H and V with the conventional sphere S, cylinder C and axis is

A working group of the Optical Society of

America recommended that the reference
axis for on-axis vision should be the line
of sight.40 This is defined in ANSI Z80.28
20042 as The line from the point of interest in object space to the centre of the
entrance pupil of the eye and continuing
from the centre of the exit pupil to the
retinal point of fixation (generally the
foveola) and more simply by Atchison and
Smith38 as the line joining the fixation
point and the centre of the entrance
pupil.
One problem with the line of sight is
that the pupil centre moves slightly with
changes in pupil size caused by accommodation or illumination changes, and this
is an issue when comparing aberrations at
a range of pupil sizes.41,42

H = S + C cos2() V = S + C sin2()

[31]

Sometimes we refer to tangential (Ta)

and sagittal (Sa) corrections, where the Ta
correction is along the meridian of the
visual field investigated and the Sa correction is perpendicular to this. For refraction in the horizontal visual field, the tangential meridian corresponds to the
horizontal meridian and the sagittal meridian corresponds to the vertical meridian. Ta and Sa are related to the M, J180, J45
system by
M = (Ta + Sa)/2 J180 = (Ta - Sa)/2

[32]

Unfortunately this approach does not give

a value for J45.
Assuming that the off-axis pupil is elliptical, Atchison and Scott16 developed a
method to determine wave aberration coefficients from Hartmann-Shack images.
For off-axis vision along the horizontal
meridian of the visual field, the wave aberration coefficients can be converted into

M = -[(23C 02 - 65C 40 + 127C 60 - 209C 80 +

3011C 100 - )(1 + cos2)
+ (6C 22 - 310C 24 + 614C 26 - 1018C82
+ 1522C102 )sin2]/(R 2cos2) [33]
J45 = -(26C2-2 - 610C4-2 + 1214C6-2 - 2018C 8-2 +
3022C 10-2 - )/(R 2cos)
[34]
J180 = -[(23C 20 - 65C 40 + 127C 60 - 209C 80 +
3011C 100 - )sin2
+ (6C 22 - 310C 42 + 614C 26 - 1018C82 +
1522C102 - )(1 + cos2)]/(R 2cos2)
[35]

CORNEA AND LENTICULAR

COMPONENTS OF ABERRATIONS
Several studies have investigated the corneal and lenticular contributions to aberrations of the whole eye. The lenticular
contribution has been estimated by laser
raytracing along one meridian of in vitro
lenses,43,44 measuring the aberration of the
eye after neutralising the cornea45 and subtracting the corneal aberrations from
those of the whole eye.46 Most recent studies have used the latter approach. RayClinical and Experimental Optometry 87.3 May 2004

147

tracing is performed through the anterior

cornea following topographic measurement. The raytrace is usually done from
infinity into the eye, even when the total
aberrations are determined by an out-ofeye method, such as the Hartmann-Shack
sensor technique, and this may produce
small errors in the calculation of internal
aberrations. Typical refractive indices used
for the cornea are 1.3375 and 1.376. Use
of the former value provides an estimate
of the aberrations of the total cornea but
ignores possible aberrations of the posterior cornea. If a refractive index of 1.376
is used, we are estimating the aberrations
of the anterior cornea. The posterior cornea contributes to the internal aberrations, although its contribution is likely to
be small because of the small refractive
index difference between the cornea and
the aqueous.
For a valid determination of the corneal
and lenticular contributions to the total
aberrations, the anterior corneal topographical data should be re-referenced
from the keratometric axis measured with
corneal topographers to the line of sight.
This is possible if the displacement between the corneal sighting centre and the
vertex normal is known, as is often the case
for videokeratometers. Alternatively, the
correction can be made if the angle
lambda between the line of sight and the
pupillary axis has been measured.23,47 In
many studies measuring corneal aberrations, no attempt has been made to correct for the difference between the
keratometric axis and the line of sight, and
this will introduce errors.
In investigations of the corneal aberrational component, there is considerable interest in the shape of the anterior cornea
and in particular how this changes following corneal refractive surgery. This surface
is sometimes approximated as a conicoid
c[X 2 + Y 2 + (1+Q)Z 2] - 2Z = 0

[36]

where the Z-axis is the optical axis, the

X-axis and Y-axis are in the equatorial
plane of the lens, c is the vertex curvature
and Q is the asphericity. If Q < 0 the surface flattens away from the vertex and is
referred to as being prolate, if Q = 0 the

Representation of aberrations Atchison

surface is spherical and if Q > 0 the surface steepens away from the surface and
is referred to as being oblate. A number
of authors use p to represent the
asphericity, where

15.

16.

p=1+Q

[37]

ACKNOWLEDGEMENT

17.

I thank Dion Scott for producing Figures

1 and 2.
18.
REFERENCES
1. Thibos LN, Applegate RA, Schwiegerling
JT, Webb R and VSIA Standards Taskforce
Members. Standards for reporting the optical aberrations of eyes. J Refract Surg 2002:
18: S652-S660.
2. American National Standards. American
National Standard for OphthalmicsMethods for reporting optical aberrations of eyes
ANSI Z80.28-2004.
3. Atchison DA, Cox MH, Scott DH. Mathematical treatment of ocular aberrations:
a users guide. Washington DC, Optical
Society of America, TOPS 2000; 35: 110-130.
4. Welford WT. Aberrations of optical systems.
Bristol: Adam Hilger, 1986; 98.
5. Wang JY, Silva DE. Wave-front interpretation with Zernike polynomials. Appl Optics
1980; 19: 1510-1518.
6. Atchison DA, Scott DH, Charman WN.
Hartmann-Shack technique and refraction
across the horizontal visual field. J Opt Soc
Am A 2003; 20: 965-973.
7. Schwiegerling J. Scaling Zernike expansion
coefficients to different pupil sizes. J Opt
Soc Am A 2002; 19: 1937-1946.
8. Campbell CE. Matrix method to find a new
set of Zernike coefficients from an original
set when the aperture radius is changed. J
Opt Soc Am A 2003; 20: 209-221.
9. Smirnov HS. Measurement of wave aberration in the human eye. Biophysics 1961; 6:
687-703.
10. Howland B, Howland HC. Subjective measurement of high-order aberration of the
eye. Science 1976; 193: 580-582.
11. Howland HC, Howland B. A subjective
method for the measurement of monochromatic aberrations of the eye. J Opt Soc
Am 1977; 67: 1508-1518.
12. Porter J, Guirao A, Cox IG, Williams DR.
Monochromatic aberrations of the human
eye in a large population. J Opt Soc Am A
2001; 18: 1793-1803.
13. Thibos L, Hong X, Bradley A, Cheng X.
Statistical variation of aberration structure
and image quality in a normal population.
J Opt Soc Am A 2002; 19: 2329-2348.
14. Wang L, Koch DD. Ocular higher-order
aberrations in individuals screened for re-

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

fractive surgery. J Cataract Refract Surg 2003;

29: 1896-1903.
Iskander DR, Collins MJ, Davis BA. Optimal modeling of corneal surfaces with
Zernike polynomials. IEEE Trans Biomed
Imaging 2001; 48: 85-97.
Atchison DA, Scott DH. Monochromatic aberrations of human eyes in the horizontal
visual field. J Opt Soc Am A 2002; 19: 2180-2184.
Cheng X, Himebaugh NL, Kolbaum PS,
Thibos LN, Bradley A. Test-retest reliability
of clinical Shack-Hartmann measurements.
Invest Ophthalmol Vis Sci 2004; 45: 351-360.
Schwiegerling J, Greivenkamp JE, Miller JM,
Snyder RW, Palmer ML. Optical modeling
of radial keratotomy incision patterns. Am J
Ophthalmol 1996; 122: 808-817.
Koomen M, Tousey R, Scolnik R. The
spherical aberration of the eye. J Opt Soc
Am 1949; 39: 370-376.
Charman WN, Jennings JA, Whitefoot H.
The refraction of the eye in relation to
spherical aberration and pupil size. Br J
Physiol Opt 1978; 32: 78-93.
Atchison DA, Smith G, Efron N. The effect
of pupil size on visual acuity in uncorrected
and corrected myopia. Am J Optom Physiol
Opt 1979; 56: 315-323.
Mrochen MC, Bueeler M, Seiler T. Influence
of higher order optical aberrations on refraction. Invest Ophthalmol Vis Sci 2002; 43: 82S.
Salmon TO, Thibos LN. Videokeratoscopeline-of-sight misalignment and its
effect on measurements of corneal and internal ocular aberrations. J Opt Soc Am A
2002; 20: 657-669.
Thibos LN, Wheeler W, Horner D. Power
vectors: an application of Fourier analysis to
the description and statistical analysis of refractive error. Optom Vis Sci 1997; 74: 367-375.
Thibos LN, Bradley A, Hong X. A statistical model of the aberration structure of
normal, well-corrected eyes. Ophthalmic
Physiol Opt 2002; 22: 427-433.
Cheng X, Bradley A, Hong X, Thibos LN.
Relationship between refractive error and
monochromatic aberrations of the eye.
Optom Vis Sci 2003; 80: 43-49.
Wang Y, Zhao K, Jin Y, Niu Y, Zuo T. Changes
of higher order aberration with various
pupil sizes in the myopic eye. J Refract Surg
2003; 19: S270-S274.
Guirao A, Williams DR. A method to predict refractive errors from wave aberration
data. Optom Vis Sci 2002; 80: 36-42.
Cheng X, Thibos LN, Bradley A. Estimating
visual quality from wavefront aberration measurements. J Refract Surg 2003; 19: S579-S584.
Cheng X, Bradley A, Thibos LN. Predicting subjective judgment of best focus with
objective image quality metrics. J Vision
2004; 4: 310-321.
Marsack JD, Thibos LN, Applegate RA.
Metrics of optical quality derived from wave

Clinical and Experimental Optometry 87.3 May 2004

148

32.

33.
34.

35.

36.

37.
38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

aberrations predict visual performance. J

Vision 2004; 4: 322-328.
Thibos LN, Hong X, Bradley A, Applegate
RA. Accuracy and precision of objective
refraction from wavefront aberrations. J
Vision 2004; 4: 329-351.
Thibos LN. Wavefront data reporting and terminology. J Refract Surg 2001; 17: S578-S583.
Campbell CE. A new method for describing the aberrations of the eye using Zernike
polynomials. Optom Vis Sci 2003; 80: 79-83.
Iskander DR, Morelande MR, Collins MJ,
Davis BA. Modeling of corneal surfaces with
radial polynomials. IEEE Trans Biomed
Imaging 2002; 49: 320-328.
Thibos LN, Ye M, Zhang XX, Bradley A.
The chromatic eye: a new reduced-eye
model of ocular chromatic aberration in
humans. Appl Optics 1992; 31: 3594-4000.
Emsley HH. Visual Optics, Vol 1, 5th ed.
London: Butterworths; 1952.
Atchison DA, Smith G. Optics of the Human Eye. Oxford United Kingdom:
Butterworth-Heinemann, 2000.
Ma L, Atchison DA, Charman WN. Off-axis
aberrations following corneal refractive
surgery. J Cataract Refract Surg. In press.
Applegate RA, Thibos LN, Bradley A,
Marcos S, Roorda A, Salmon TO, Atchison
DA. Reference axis selection: Subcommittee report of the OSA working group to
establish standards for the measurement
and reporting of the optical aberration of
the eye. J Refract Surg 2000; 16: S656-S658.
Walsh G. The effect of mydriasis on
pupillary centration of the human eye. Ophthalmic Physiol Opt 1988; 8: 178-182.
Wilson MA, Campbell MCW, Simonet P.
Change of pupil centration with change of
illumination and pupil size. Optom Vis Sci
1992; 69: 129-136.
Sivak J, Kreuzer RO. Spherical aberration
contribution of the crystalline lens. Vision
Res 1983; 23: 59-70.
Glasser A, Campbell M. Presbyopia and the
optical changes in the human crystalline
lens with age. Vision Res 1998; 38: 208-229.
Millodot M, Sivak J. Contribution of the
cornea and lens to the spherical aberration
of the eye. Vision Res 1979; 19: 685-687.
El Hage SG, Berny F. Contribution of the crystalline lens to the spherical aberration of the
whole eye. J Opt Soc Am 1973; 63: 205-211.
Atchison DA. Anterior corneal contribution to peripheral aberrations of human
eyes. J Opt Soc Am A 2004; 21: 355-359.

Authors address:
David Atchison
School of Optometry
Queensland University of Technology
Victoria Park Road
Kelvin Grove QLD 4059
AUSTRALIA