There is a summary "pdf" slide show on
web site in addition to the write-up which I have "clipped
and pasted" from that site,
below (I've been afraid that I'd lose the site if the URL changed).
Using the ASCP book on morphology, we are said to be using the "ASCP
system of morphologic classification". We stain smears with
the Pap stain and use rough estimations of percentage "normal",
without detailing the various abnormal forms. From that
The goal of estimating correctly a man’s
fertility potential has long been of great interest to researchers.
The term ‘male infertility ’ does not constitute
a defined clinical syndrome but rather a collection of different
conditions exhibiting a variety of etiologies and varying prognoses
(1). It is therefore unreasonable to hope for tests that will indicate
with absolute certainty that a man will be fertile. On the other
hand, when the results of these tests are subnormal, they should
not be used as an absolute diagnosis of sterility.
Semen analysis is a keystone in the clinical workup
of the infertile man patient. Since first published in 1980, the
WHO Manual for Andrology Laboratories (35,36) has gained worldwide
acceptance as a source of standard methodology for human semen
analysis. In the latest edition (1992) the guidelines for morphology
evaluation have been revised significantly.
Sperm morphology is assessed routinely as part
of standard laboratory analysis in the diagnosis of human male
infertility. This practice has its origins in the work of Mac Leod & Gold
(1951) which showed that sperm morphology was significantly different
in fertile compared to infertile man.
Despite this standardization, human semen evaluation
continues to be influenced by subjectiveness of the investigator
and a lack of objective measurements for sperm morphology continues
to be a problem.
There is an ongoing debate on which criteria should
be applied to define normal spermatozoa and which classification
of abnormal forms is most appropriate.
Studies on sperm morphology should concentrate
to obtain measurements and biological data of spermatozoa which
are functionally active. Only then the definition of normal can
be achieved and clinically useful criteria can be adopted.
However, the definition of a normal spermatozoon
as described by WHO in 1992 is different from that used by other
authors (17,21,22,31). The evaluation of the morphology of human
spermatozoa varies widely between and sometimes even within laboratories.
While most investigators agree on the appearance of a normal spermatozoon,
standardized analysis is difficult because of the use of different
staining techniques which are not always suitable for optimal examination
from head to tail (Figure 1). For example, the techniques for preparing
morphology specimens have been expanded from three to five.
The difficulty in classifying human sperm morphology
is mainly caused by the large variety of abnormal forms found in
the semen of infertile men. Only certain types of abnormalities
can be analyzed objectively (11).
The definition of ‘morphological normal’ is
still discussed, as well as the clinical relevant limits for the
rate of pathologic forms.
What is a normal spermatozoon?
The WHO’s definition of a normal spermatozoon
is not based on any biological data. As a consequence, the implementation
of new standards has resulted in some controversy. However, sperm
morphology was found to correlate more closely with fertilization
rates than sperm count and motility (2).
Because it is not possible to determine the fertilizing
potential of individual human spermatozoa, physiological endpoints
other than fertilization must be studied to obtain insight into
the mechanisms by which sperm morphology influences the fertilization
process, e.g. examination of morphologic characteristics of spermatozoa
recovered from cervical mucus and/or those binding to the zona
The strict criteria of sperm morphology (e.g.
Kruger’s criteria) use the examination of spermatozoa that
had penetrated cervical mucus for the definition of normal spermatozoon.
Even by using strict criteria it is difficult to determine if results
between studies are really comparable.
In order to avoid subjectivity, over the past
20 years numerous studies describe image analysis techniques in
the assessment of sperm morphology. These techniques allow objective
characterization of different sperm forms. Automated methods may
help, but there remains a lack of biological data to support the
use of computer-aided semen analysis in clinical settings.
Recent evidence suggests that sperm morphology
assessment by relatively simple and inexpensive methods can provide
prognostic information similar to that obtained from some of the
more elaborate sperm function tests.
Criteria of the World Health Organization
The WHO manual in 1987 describes a normal spermatozoa: ‘An
oval head shape with a regular outline and acrosomal cap covering
more than one-third of the head surface. The head: length :
3-5 µm, width : 2-3 µm; lenght/width ratio :
1,5-2. The midpiece: 7-8 µm, long, straight and regular in
outline, slender, less than 1/3 of width of the head. The tail:
at least 45 µm in lenght, slender, uncoiled and regular in
Spermatozoa were classified into normal, having
head defects (amorphous, small, large, pyriform, tapering), midpiece
defects (including cytoplasmic droplets) and tail defects (Table
1). At least 200 spermatozoa need to be examined in an attempt
to reduce technical variation.
Sperm morphology is determined using brightfield
illumination at x1000 magnification, after preparing air-dried
Papanicolaou-stained smears. All slides were read blind by an experienced
highly trained technician who produced consistent and reliable
In keeping with the move to a more strict definition
of sperm morphology, WHO has redefined what they consider to be
a normal spermatozoon and subsequently set an ‘empirical
reference value of 30% normal forms and above as normal’ (36).
A normal spermatozoon has an oval head shape with
regular outline and a well-defined acrosomal region covering 40-70
% of head; vacuoles occupy less than 20% of the head area. The
head: length : 4-5,5 µm, width : 2,5-3,5 µm
and lenght/width ratio : 1,5-1,75; no cytoplasmic droplets
more than 1/3 of the size of a normal sperm head. No dimensions
and no description of a normal midpiece are mentioned. Defects
are described, e.g. insertion of the tail in more than 90% of the
head’s longitudinal axis is abnormal. No dimensions and no
description of a normal tail are mentioned, only defects are described.
Using this classification scheme, all borderline
forms are considered abnormal.
large, small, tapering, amorphous, pyriform, vacuolated or double
heads or any combination of these.
Neck or midpiece defectsare:
bent or abnormal thin.
short, multiple hairpin, broken, irregular width or coiled tails,
tails with terminal droplets or any combination of these.
A minimum of 200 spermatozoa is counted and a
stage micrometer is used to aid interpretation.
The teratozoospermia index is a measure of the
average number of defects per spermatozoa, which could be used
to improve the correlation between sperm morphology and fertility.
Basic semen evaluation is performed after liquefaction
of the specimen. Two morphology slides are prepared for each patient
and are stained by the quick-stain technique ( Diff-Quick solution
1 and 2). Special care is taken to clean the slides thoroughly
with 70% ethyl alcohol before use and not more than 5 ul of semen
is used to prepare the slides as thin as possible. The slides are
reported on the same day. The morphology is evaluated by two independent
The Strict Criteria
Another important evaluation aiming to clarify
male fertility is the strict criteria for sperm morphology (Kruger’s
or Tygerberg’s criteria1986) (17).
Strict criteria of sperm morphology established
by Kruger et al. define normal spermatozoa as having an oval configuration
with a smooth contour ( Figure 2). The head in lenght is 5-6 µm,
the diameter( width) is 2,5-3,5 µm and the width/lenght ratio
is 1/2-3/5. The acrosome is well-defined, comprising 40-70% of
the distal part of the head. No abnormalities of the neck, midpiece
or tail and no cytoplasmic droplets of more than half of the sperm
head are accepted. Borderline forms are considered abnormal.
is divided into two categories:
slightly amorphous, with a head diameter of
2,0 - 2,5 um, with slight abnormalities in the head’s
shape but with normal acrosome.
severely amorphous, with no acrosome at all
and those with an acrosome smaller than 30% or larger than
70% of the sperm head; completely abnormal shapes also are
put into this category.
Neck defects are
also classified in two categories:
slightly amorphous, referred to those sperms
with debris around the neck or a thickened neck but with a
normal shaped head.
severely amorphous, referred to those sperms
with a bended neck or midpiece of more than 30%, or a severely
amorphous head shape, as described.
All other abnormal sperm forms- round, small,
large, tapered, double head, double or coiled tail, cytoplasmic
droplets-are classified following the WHO classification.
Normal and borderline forms grouped together are
called ‘the morphology index ’.
Patients with a morphology index less than 30% will have a severe
reduction in fertilization as compared with patients having an
index greater than 30% (17). In Kruger’s practice, the normal
forms considered alone are called the ‘percentage of ideal
forms’ (PIF). A PIF greater as 4% is considered favorable
and less than 4% unfavorable. At least 200 cells per slide are
to be evaluated. A micrometer in the eyepiece of the microscope
is used for routine measurements.
In order to improve the strict criteria, in 1987
Menkveld defined a normal spermatozoon, based on the
appearance of spermatozoa found in good periovulatory cervical
mucus of the upper part of the endocervical canal. These spermatozoa
are usually in an apparently homogenous population( Figure 3).
The morphological classification used by Menkveld is based on a
modification of the methods of Mac Leod and Gold (1951) and Eliasson
The head must have a smooth oval configuration
with a well-defined acrosome comprising approx. 40-70% of it. The
normal head has a lenght of 3-5 µm, width of 2-3 µm
and a width/lenght ratio of 3/5- 2/3. The range of variation
within the normal population is shown in Figure 4. The most important
difference between this classification and other methods is that
Menkveld regards borderline-normal head forms and/or spermatozoa
with nearly oval heads with no gross abnormalities as abnormal.
Neck, midpiece or tail defects are considered abnormal. The midpiece
lenght is 1,5 times the head lenght, the width is less than 1 µm,
slender and axially attached. The tail measures 45 µm in
length, appears uniform, uncoiled and is slightly thinner than
the midpiece. Cytoplasmic droplets (remnants) which comprise less
than ½ the size of the sperm head are accepted. These are
Spermatozoa are classified into seven groups (Figure
5): normal (whole sperm), large, small, elongated (tapering), duplicated
and amorphous heads, all with or without the presence of a cytoplasmic
droplet and/or tail, neck and/or midpiece defects. The seventh
group consist of spermatozoa with a normal head but with a tail
and/or a neck and/or a midpiece defect and/or the presence of a
At least 100 but preferably 200 spermatozoa are
evaluated. Inexperienced workers should use a built-in micrometer
when they begin with morphology evaluations. The normal dimensions
for spermatozoa stained with the Diff-Quik method (used by Kruger)
are larger than those based on the Papanicolaou method.
In order to improve the evaluation of sperm morphology
, Davis and Gravance (1994) have emphasized the sensitivity of
sperm classification methods when only two morphometric variables
are used (lenght and width of the head, for example). Based on
linear models as an appropriate mean of describing the size relationship
between phenotypic characters, they show that small changes can
significantly alter the percentage of normal sperm within a specimen.(9).
A new expression of sperm morphology parameters
is the sperm deformity index (SDI), described in 1996 (2). This
is a method by where the whole spermatozoon is assessed by the
strict criteria and classified more than once if more than one
deformity exists. Both normal and abnormal sperms are considered
and the average number of deformities per sperm is determined to
give a value to this index. This index reflects the balance between
the prevalence of sperms with multiple structural deformities and
the proportion of sperms with normal morphology in a semen sample.
Computer-assisted methods of sperm morphology
Over the past 20 years, in order to avoid subjectivity,
numerous studies that incorporate image analysis techniques in
the assessment of sperm morphology have appeared. These techniques
allow objective characterization of different sperm forms.
The method of Moruzzi (1988) for quantification
and classification of human sperm morphology by computer-assisted
image analysis is semi-automated (25). Sperm heads are imaged through
a microscope (NA=1.3) sampled at 0,125 um intervals and measured
on an image analyzing system. Measurements included stain content,
lenght, width, perimeter, area, and arithmetically derived combination.
Additionally, each sperm image is optically sectioned at right
angles to its major axis, to give a measure of lengthwise heterogeneity
of shape. According to this method, the percentage of normal sperm
heads can be accurately predicted using just four sperm head measurements.
This classification procedure distinguishes normal from abnormal
sperm with 95% accuracy and correctly assigned 86% of the sperm
to one of 10 shape classes.
The system described by Perez-Sanchez (1994) is
based on the evaluation with a video digitizer board, a brightfield
microscope with a 100x immersion objective, two monitors and the
image analysis software (31)(Figure 6). Cells are displayed live
on the video monitor and each sperm head image is processed automatically
using a specific image analysis program for image enhancement and
thresholding. Analysis of the sperm midpiece and tail is not included
in the program. The system detects the boundary of the sperm head
and the outline is displayed as white overlays superimposed on
the video image. The set of morphometric parameters used by Perez
constitutes a set of characteristics which is valid for characterization
of the majority of morphological types of spermatozoa.
The association between semen quality and male
infertility has been known for more than 40 years.
Having reviewed the literature, it seems clear
that strict morphology has a clinical relevance, being an excellent
biomarker of sperm fertilizing capacity, in vivo and in vitro,
independent of motility and concentration (27).
Sperm morphology evaluation is considered to be
a highly subjective procedure because, unlike the haematopoietic
cells for example, the difficulty in classifying human sperm morphology
is caused by the large variety of abnormal forms found in the semen
of infertile and fertile men. Only certain types of abnormality
can be quantitated objectively (11).
Normal sperm morphology needs to consider two
points. The first one is the proportion of spermatozoa with normal
morphology in semen and the second is the definition and the characterization
of the normal spermatozoa.
According to WHO criteria, a normal ejaculate
must have at least 30% normal sperm.(36). For the stricter criteria,
fertile men have > 14% normal forms in their semen and men with < 4%
of normal forms are subfertile. According to Kruger’s criteria,
IVF outcome was suboptimal when normal sperm morphology was less
than 14% and worst if it was under 4%.(17). The sperm deformity
index is a more reliable predictor of the outcome of fertilization
in vitro than the proportion of normal sperm morphology.(2).
WHO recommends that each laboratory recruits fertile
men (a reference population) in order to investigate and determine
the real cut-off values for normality in that laboratory (28).
These men are very difficult to recruit, therefore only a few laboratories
actually perform this analysis (3).
During the past 15 years there has been an increase
in total motile sperm count, secondary to an increase in semen
volume, and a decline in normal morphology. Both are independent
of the age and the duration of abstinence in fertile men (5).
There is a debate on this subject and some of
the arguments discussed by different authors are the analyze techniques
used by laboratories and the influence of environmental factors
on sperm morphology.
A number of authors, under in vivo conditions
of conception, have demonstrated that the pregnancy rate was significantly
higher in cases of better sperm morphology (13). But, particularly
under the usual in vivo conditions of conception, pregnancy rates
are influenced by a multiplicity of different parameters.
It seems that sperm morphology evaluated for strict
criteria has definitive advantages over the other (liberal) criteria
evaluation methods in the prediction of in vivo and especially
in vitro fertilization rates (22).
Other publications’ unavoidable conclusion
is that sperm morphology according to WHO classification can be
a vigorous predictor of IVF outcome (24). Therefore, further investigations
of semen morphology and male fertility according to WHO may be
There are papers which emphasize the importance
of inclusion of abnormalities in the acrosomal region during sperm
morphology assessments and suggest that previous emphasis placed
on sperm head dimensions in the classification of normality is
justified only in cases of extreme deviation of the normal (12).
Neither abnormal sperm concentration, nor abnormal
sperm morphology are significantly associated with recurrent spontaneous
Kruger’s opinion is that the existing classification
of abnormal and normal shaped sperms are in need of revision by
those involved in the field. Scanning electron microscope is worth
further evaluation as a tool in the accurate scoring of normal
The advantage of using strict criteria in morphology
evaluation is the fact that the measure is reproducible between
patients and between different technicians performing the test.
It also allows the clinician to classify the patient into one of
two specific groups ( < 14% and >14% normal morphology),
giving a reliable criteria to plan the approach for future IVF
For clinicians responsible interpreting results
to patients, it is important to emphasize that 0% strict criteria
does not necessarily implicate that no normal sperm is present:
it only signifies that no sperm meeting a set of criteria was observed
in the sample (24).
Nevertheless, even in cases of severe teratozoospermia
fertilization may be possible; most promising is the development
of intracytoplasmic sperm injection ( ICSI ) as the treatment of
first choice in these cases.(27).
Another conclusion is that only one standard method
should be recommended for the preparation of morphology slides
in order to ensure inter-laboratory comparability of results and
to enhance the value of sperm morphology analysis for predicting
fertility (23). At present there are five techniques of sperm preparation.
For practical purposes, the Papanicolaou method seems the most
suited, due to its widespread use.
Sperm morphology must only be considered as an
indicator of fertilization potential, not as an absolute indicator
of sterility. The aim should be to find the minimum sperm parameters
associated with specific methods of seminal treatment and with
culture techniques which can lead to a single, common strategy
for male infertility problems.
The Human Fertilization and Embryology Authority
(1991) and WHO (1993) strongly suggest that for the treatment of
infertility both partners must be evaluated in order to exclude
- Aitken, R.J., Baker, H.W.G. and Irvine, D.S.,(1995).The diagnosis
of male infertility by semen quality : On the nature of
semen quality and infertility. Hum. Reprod. 10, 248-250.
- Aziz, N., Buchan, I., Taylor C., Kingsland C.R. and Lewis-Jones
I.,(1996). The sperm deformity index : a reliable predictor
of the outcome of oocyte fertilization in vitro. Fertil. Steril.
- Barrat, C.L.R., Naeeni, M., Clements, S. and Cooke, I.D.,(1995).
Clinical value of sperm morphology for in-vivo fertility :comparison
between World Health Organization criteria of 1987 and 1992.
Hum. Reprod. 10, 587-593.
- Barrat, C.L.R.,(1995). On the accuracy and clinical value of
semen laboratory tests. Hum. Reprod. 10, 250-252.
- Benshushan, A., Shoshani, O., Paltiel, O., Schenker, J.G. and
Lewin, A.,(1997). Is there really a decrease in sperm parameters
among healthy young men ? A survey of sperm donations during
15 years . J. Assist. Reprod. Genet. 14, 347-353.
- Comhaire, F., Schoonjans, F., Vermeulen, L. and De Clercq,
N.,(1994). Methodological aspects of sperm morphology evaluation :
comparison between strict and liberal criteria. Fertil Steril.
- Check, J.H.,(1995). Value of sperm morphology. Fertil. Steril.
- Coetzee, K., Kruge, T.F. and Lombard, C.J.,(1998). Predictive
value of normal sperm morphology : a structured literature
review. Hum. Reprod. Update, 4, 73-82.
- Davis, R.O. and Gravance, C.G.,(1994). Consistency of sperm
morphology classification methods. J. Androl. 15, 83-91.
- de Agostini, A.I., Schorderet-Slatkine, S., Fournier, J., Luthi,
R. and Campana, A.,(1998). Criteria of evaluation for human sperm
head morphology. 1998, Manuscript in preparation (available upon
- Eggert-Kruse, W., Schwarz, H., Rohr, G., Demirakca, T., Tilgen,
W. and Runnebaum, B.,(1996). Sperm morphology assessment using
strict criteria and male fertility under in-vivo conditions of
conception. Hum. Reprod. 11,139-146.
- Garrett, C., Liu, D.Y. and Baker, H.W.G.,(1997). Selectivity
of the human sperm-zona pellucida binding process to sperm head
morphometry. Fertil. and Steril. 67, 362-371.
- Haidl, G. and Schill, W.B.,(1993). Sperm morphology in fertile
men. Arch. Androl. 31, 153-157.
- Hill, J.A., Abbott, A.F. and Politch, J.A.,(1994). Sperm morphology
and recurrent abortion. 61, 776-778.
- Hofmann, G.E., Santili, B.A., Kindig, S., Scott, R.T. Jr. and
Johnson, C.A.,(1996). Intraobserver, interobserver variation
of sperm critical morphology : comparison of examiner and
computer-assisted analysis. Fertil. Steril. 65, 1021-1025.
- Karabinus, D.S. and Gelety, T.J.,(1997). The impact of sperm
morphology evaluated by strict criteria on intrauterine insemination
success. Fertil. and Steril. 67, 536-541.
- Kruger, T.F., Acosta, A.A., Simmons, K.F., Swanson, R.J., Matta,
J.F. and Oehninger, S.,(1988). Predictive value of abnormal sperm
morphology in in vitro fertilization. Fertil. Steril. 49, 112-117.
- Kruger, T.F., du Toit, T.C., Franken, D.R., Acosta, A.A., Oehninger,
S.C., Menkveld, R. and Lombard, C.J.,(1993). A new computerized
method of reading sperm morphology (strict criteria) is as efficient
as technician reading. Fertil. Steril. 59, 202-209.
- Kruger, T.F., du Toit, T.C., Franken, D.R., Menkveld, R. and
Lombard, C.J.,(1995). Sperm morphology : assessing the agreement
between the manual method (strict criteria) and the sperm morphology
analyzer IVOS. Fertil. Steril. 63, 134-141.
- Lindheim, S.R., Barad, D.H., Zinger, M., Witt, B., Amin, H.,
Cohen, B., Fisch, H. and Barg, P.,(1996). Abnormal sperm morphology
is highly predictive of pregnancy outcome during controlled ovarian
hyperstimulation and intrauterine insemination. J. Ass. Reprod.
Genet. 13, 569-572.
- Menkveld, R., Stander, F.S., Kotze, T.J., Kruger, T.F. and
van Zyl, J.A.,(1990). The evaluation of morphological characteristics
of human spermatozoa according to stricter criteria. Hum. Reprod.
- Menkveld, R. and Kruger, T.F.,(1995). Advantages of strict
(Tygerberg) criteria for evaluation of sperm morphology. Int.
J. Androl. 18, 36-42.
- Meschede, D., Keck, C., Zander, M., Cooper, T.G., Yeung, C.H.
and Nieschlag, E.,(1993). Influence of three different preparation
techniques on the results of human sperm morphology analysis.
Int. J. Androl. 16, 362-369.
- Morgentaler, A., Fung, M.Y., Harris, D.H., Powers, R.D. and
Alper,M.M.,(1995). Sperm morphology and in vitro fertilization
outcome : a direct comparison of World Health Organization
and strict criteria methodologies. Fertil. Steril. 64, 1177-1182.
- Moruzzi, J.F., Wyrobek, A.J., Mayall, B.H. and Gledhill, B.L.,(1988).
Quantification and classification of human sperm morphology by
computer-assisted image analysis. Fertil. Steril. 50, 142-152.
- Oehninger, S. and Kruger, T.,(1995). The diagnosis of male
infertility by semen quality. Clinical significance of sperm
morphology assessment. Hum. Reprod. 10, 1037-1038.
- Ombelet, W., Menkveld, R., Kruger, T.F. and Steeno, O.,(1995).
Sperm morphology assessment : historical review in relation
to fertility. Hum. Reprod. Update 1, 543-557.
- Ombelet, W., Wouters, E., Boels, L., Cox, A., Janssen, M.,
Spiessens, C., Vereecken, A., Bosmans, E. and Steeno, O.,(1997).
Sperm morphology assessment : diagnostic potential and comparative
analysis of strict or WHO criteria in a fertile and a subfertile
population. Int. J. Androl. 20, 367-372.
- Ombelet, W., Vandeput, H., Van de Putte, G., Cox, A., Janssen,
M., Jacobs, P., Bosmans, E., Steeno, O. and Kruger, T.,(1997).
Intrauterine insemination after ovarian stimulation with clomiphene
citrate : predictive potential of inseminating motile count
and sperm morphology. Hum.Reprod. 12, 1458-1463.
- Ombelet, W., Bosmans, E., Janssen, M., Cox, A., Maes, M., Punjabi,
U., Blaton, V., Gunst, J., Haidl, G., Wouters, E., Spiessens,C.,
Bornman, M.S., Pienaar, E., Menkveld, R. and Lombard, C.J.,(1998).
Multicenter study on reproducibility of sperm morphology assessments.
Arch. Androl. 41, 103-114.
- Perez-Sanchez, F., de Monserrat, J.J. and Soler, C.,(1994).
Morphometric analysis of human sperm morphology. Int. J. Androl.
- Seibel, M.M. and Zilberstein, M.,(1995). The diagnosis of male
infertility by semen quality. The shape of sperm morphology.
Hum. Reprod. 10, 247.
- Seracchioli, R., Porcu, E. and Flamigni, C.,(1995). The diagnosis
of male infertility by semen quality. Sperm morphology is not
the only criterion of male infertility. Hum. Reprod. 10, 1039-1041.
- Vawda, A.I., Gunby, J. and Younglai, E.V., (1996). Semen parameters
as predictors of in-vitro fertilization : the importance
of strict criteria sperm morphology. Hum. Reprod. 11, 1445-1450.
- WHO.,(1987). WHO laboratory manual for the examination of human
semen and semen-cervical mucus interaction. Cambridge University
Press, Cambridge. 1-67 pp.
- WHO.,(1992). WHO laboratory manual for the examination of human
semen and sperm-cervical mucus interaction. Cambridge University
Press, Cambridge. 1-107 pp.
- Infertility in Eastern Europe, a survey of the current status
in 9 countries. Lüdicke, F., Brandrup-Lukanow, A.,
Horga, M. and Campana, A. Presentation at WHO/HRP/FIGO Symposium
on Reproductive Health in Eastern Europe, FIGO Congress, August
1997, Copenhagen. WHO (in press).
8 August 2002; latest update 12 May 2003)