HOWIE

Micron 40 (2009) 285–301
Contents lists available at ScienceDirect
Micron
journal homepage: www.elsevier.com/locate/micron
Review
Optical properties of amyloid stained by Congo red: History and mechanisms

Alexander J. Howie a,*, Douglas B. Brewer b
a
Department of Pathology, University College London, London, UK
b
Department of Pathology, University of Birmingham, Birmingham, UK
A R T I C L E I N F O A B S T R A C T
Article history: Amyloid stained by Congo red has striking optical properties that generally have been poorly described
Received 1 September 2008 and inadequately explained, although they can be understood from principles of physical optics.
Accepted 2 October 2008 Molecules of Congo red are orientated on amyloid fibrils, and so the dye becomes dichroic and
birefringent. The birefringence varies with wavelength in accordance with a fundamental property of all
Keywords: light-transmitting materials called anomalous dispersion of the refractive index around an absorption
Amyloid peak. The combination of this and absorption of light, with modification by any additional birefringence
Anomalous colours
in the optical system, explains the various colours that can be seen in Congo red-stained amyloid between
Birefringence
crossed polariser and analyser, and also when the polariser and analyser are progressively uncrossed.
Congo red
Dichroism These are called anomalous colours.
Polarising microscopy ß 2008 Elsevier Ltd. All rights reserved.
Contents
1. Introduction: amyloid and Congo red . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

2. Optical properties of Congo red-stained amyloid and orientated Congo red: the background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
3. Dichroism of Congo red . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
4. Birefringence and retardance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
5. Types of birefringence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
6. The colours seen in Congo red-stained amyloid and orientated Congo red between crossed polariser and analyser . . . . . . . . . . . . . . . . . . 291
7. Historical survey of colours described in Congo red-stained amyloid and orientated Congo red between crossed polariser and analyser . . . . . 292
8. Normal dispersion of the refractive index, and destructive interference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
9. Anomalous dispersion of the refractive index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
10. Sign of birefringence, and compensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
11. Contribution of absorption to the colours of Congo red-stained amyloid and orientated Congo red between crossed polariser and analyser . . . . 297
12. Measurements of retardance of Congo red-stained amyloid and orientated Congo red, with suggested explanations of the colours seen . . . . . 298
13. Anomalous colours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
14. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
1. Introduction: amyloid and Congo red history.) Because many misquotations, misunderstandings and
misattributions of historical matters related to amyloid have been
In 1953 Missmahl and Hartwig wrote: ‘Die Entdeckung der passed from paper to paper, extracts from the most relevant papers
Doppelbrechung in der Amyloidsubstanz hat eine längere und sich are copied from the original, with translations of German and French
mehrfach wiederholende Geschichte’. (The discovery of the extracts by DBB, to allow readers to check the meaning for
birefringence of amyloid has a long and many times repeated themselves. The spelling of the German in quotations and titles of
papers in the list of references is copied exactly from the original and
so may not follow precisely the current reformed German spelling.
* Corresponding author at: Department of Cellular Pathology, Royal Free
Hospital, London NW3 2QG, UK. Tel.: +44 207 794 0500x35641;
The quotation from Missmahl and Hartwig (1953) can be used
fax: +44 207 435 3289. as an example of confusion about optical properties of amyloid,
E-mail address: [email protected] (A.J. Howie). because most of the practical and theoretical study has not been on
0968-4328/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.micron.2008.10.002
286 A.J. Howie, D.B. Brewer / Micron 40 (2009) 285–301
birefringence of amyloid alone, but amyloid stained by Congo red. treatment with sulphuric acid. This gave cellulose a property that
Indeed, the more important factor in the properties is Congo red. unmodified cellulose did not have, which was reaction with iodine
Congo red is a synthetic compound given a variety of formal to give blue (Puchtler and Sweat, 1966).
chemical names, but each molecule can be considered to consist of Application of the term amyloid to human disease was initially
two molecules of naphthionic acid (1-naphthylamine-4-sulfonic inconsistent and confused (Puchtler and Sweat, 1966; Schwartz,
acid or 4-aminonaphthalene-1-sulfonic acid) coupled to one 1970). Virchow, who is usually said to have introduced the term
molecule of benzidine (diaminophenyl or 4-(4-aminophenyl)ani- into pathology, wrote a few papers which are difficult to
line) by azo groups, –N=N–. This is now given the Colour Index understand, such as Virchow (1854). He used amyloid first as a
number 22120 and name direct red 28 (Horobin and Kiernan, morphological description of various concentric structures resem-
2002). The chemical was supposedly first made in 1883 by a bling starch granules, including corpora amylacea in the brain,
method patented in 1884 by a German chemist, Paul Böttiger, who previously named by Purkinje. Later he used amyloid to mean an
had been looking for a pH indicator (Steensma, 2001). Böttiger may abnormal material detected by chemical tests, which he thought
not have been the first to make Congo red (Armstrong, 1935). resembled cellulose rather than starch. This use of the term was
Congo red was found to be a direct dye for cotton, which meant eventually to lead to the modern meaning of amyloid as
that there was no need for a mordant, or a chemical required to fix extracellular, insoluble fibrils derived from various misfolded
the colour in materials. Cotton is cellulose from plant cell walls. proteins.
The linear Congo red molecules align themselves along the linear
molecules of cellulose, with hydrogen bonds between the two 2. Optical properties of Congo red-stained amyloid and
substances (Wälchli, 1945). Congo red was the first direct dye, and orientated Congo red: the background
had great commercial success. Congo Red was originally a
trademark, and other early trade names for this included Congo, Amyloid stained by Congo red has striking optical properties
various types of Cotton Red and Direct Red, Cosmos Red, and that have mostly been badly described and inadequately
Chokusetsu Red (Rowe, 1924). Chokusetsu means direct in explained. These properties are largely due to the fact that the
Japanese. dye molecules have an orientated arrangement on amyloid fibrils,
The dye is too easily washed out of fabrics to be useful these as they do on cellulose molecules, with hydrogen bonding between
days for most colouring purposes, and there are other drawbacks amino and other groups on the dye and hydroxyl groups on
such as the toxicity of benzidine used in its manufacture, but amyloid (Puchtler et al., 1962). Solutions of Congo red, in which the
Bennhold (1922) in a short technical paper described the specific molecules are randomly orientated, generally do not show these
staining of amyloid with Congo red, and this is now one of its major properties.
uses. Congo red molecules can also be orientated in ways that have
Bennhold was not the first to use Congo red as a stain for been used in optical investigations. These include crystallisation
microscopy. Griesbach had done this in 1886 in a study of many and use of flowing solutions, which are technically difficult
azo dyes, but he did not mention amyloid (Griesbach, 1886). (Wälchli, 1945); polishing or stroking of dried drops of a solution,
Bennhold is credited as the first to report its value in detection which do not give particularly well orientated preparations
of amyloid. He discovered this in a study of intravenous (Missmahl, 1957; Taylor et al., 1974); and smearing of drying
injections of Congo red, which others had used to determine drops with a single stroke in one direction on a glass slide, which is
blood volume (Bennhold, 1923). Bennhold was a physician in the simplest way to obtain clean, well-orientated material that can
Hamburg at that time, and investigated the distribution and be used repeatedly in investigations of optical properties (Zocher,
elimination of Congo red in 21 normal people and 82 patients 1925; Missmahl, 1957). In these preparations the asymmetric
with various conditions. molecules become orientated with their long axis in the long axis
He found that in patients with what would now be called the of needle-shaped crystals, in the direction of flow, and in the
nephrotic syndrome there was more rapid clearance of Congo red direction of polishing or smearing.
from the blood than in controls and most other conditions, with When stained by Congo red, amyloid and other materials, such
appearance of the dye in the urine in nephrotic patients, which as cellulose paper and cotton, appear various shades of red in
suggests that Congo red was bound to albumin or other plasma unpolarised light, and so do solutions of Congo red and
proteins. In 10 patients said to have amyloidosis complicating preparations of orientated Congo red. This is because the dye,
pulmonary tuberculosis, most had even more rapid clearance, as usually used, has its maximal absorption of wavelengths in the
including when no Congo red appeared in the urine. Necropsy on a blue/green part of the visible spectrum. When white light is
patient with amyloidosis who died 20 h after injection gave the altered by most absorption of these wavelengths, the appearance
explanation, because the liver and spleen appeared red macro- is red.
scopically, and unstained frozen sections of liver, spleen and Congo red has to be in alkaline or weakly acid solutions for
kidney had red areas on microscopy, which showed that Congo red red to appear, when it is the disodium salt. In strongly acid
specifically bound to amyloid. conditions, as the free acid, the maximal absorption moves to
Intravenous injection of Congo red was used in clinical practice longer wavelengths in the yellow and orange, giving a violet or
for about 40 years in the diagnosis of amyloid. For example, this blue colour. This explains how Congo red can be used as a pH
was still in a guide called Today’s Tests (1968). Biopsy of an indicator, changing colour in the pH range 3–5 (Horobin and
affected organ with staining of sections by Congo red is now the Kiernan, 2002). The possibility of change of colour is another
usual way to make a diagnosis of amyloid. reason why Congo red is rarely used to dye clothes now. The
Amyloid, meaning starch-like, was first used by Vogel and peak absorption may also move to longer wavelengths when
Schleiden (1839) as a name for a material in cell walls of the seeds there is increased binding to a substrate, which is called the
of some plants: ‘Das Amyloid kommt gebildet in der Natur vor’. bathochromic shift (Amelin and Tret’yakov, 2003). The effects of
(Amyloid occurs already formed in nature.) This had a property of pH and binding may explain some of the reported differences in
starch, namely reaction with iodine to give a blue colour, but did optical properties of Congo red-stained amyloid and orientated
not have all the properties of starch. The name amyloid was Congo red, for instance in the wavelengths at which a property is
applied later in several different senses, including to cellulose after at its maximum.
A.J. Howie, D.B. Brewer / Micron 40 (2009) 285–301 287
3. Dichroism of Congo red literature. Rotation of either the polarising filter or the specimen
stage changes the appearance of orientated Congo red from red to
Congo red-stained amyloid and Congo red orientated in other virtually colourless, but the background stays uniformly bright,
ways show dichroism. This term is sometimes misused, but in because the unaided human eye is not able to detect that light is
physical optics, dichroism means that depending on the plane of polarised.
linearly polarised light illuminating a material, the specimen Neubert (1925), and later Wälchli (1945), showed that smears
shows different amounts of absorption of light, which may be seen of Congo red absorbed light polarised parallel to the direction of
as different colours or different intensity of a colour (Born and smearing more strongly than light polarised at right angles,
Wolf, 1999). confirming that orientated Congo red is dichroic. The change in
Congo red molecules only absorb light of the appropriate intensity of red is easy to see in smears of Congo red (Figs. 1–3), but
wavelength that is polarised parallel to their orientation, and do may be difficult to detect in sections of Congo red-stained amyloid,
not absorb light polarised perpendicularly. This can be seen on a in which amyloid fibrils are often haphazardly arranged (Howie
microscope fitted with only one polarising filter, which is either a et al., 2008).
polariser, which is between the light source and the specimen, or Some crystals show different colours as a manifestation of
an analyser, which is between the specimen and the observer. Such dichroism, but this is not usually the case with orientated Congo
filters were often of the type called Nicol prisms or Nicols in the old red, which just changes intensity of the red colour. Congo red
Figs 1–3. Three images of the same field of a smear of Congo red, with the specimen rotated to different angles relative to the marked plane of polarisation of either a polariser
alone (p) or an analyser alone (a). In Fig. 1, streaks, such as the one arrowed in each image, are almost colourless when perpendicular to the plane of polarisation. There are
intervening areas that are dark red, and do not change colour on rotation. These are where the dye dried haphazardly because it was not streaked by imperfections in the slide
used to make the smear. In Fig. 2, streaks at 458 to the plane of polarisation have a redder appearance than in Fig. 1. In Fig. 3, streaks are deepest red when parallel to the plane
of polarisation. This illustrates dichroism in the orientated streaks that change intensity of colour, with intervening non-orientated, non-dichroic areas. Reproduced from Lab.
Invest. 2008; 88: 232–242 with permission.
the two axes, and is reduced the more the dichroic object is turned
from an orientation of 458 to the polariser.
This mechanism was postulated to be part of the explanation of
the brightness of Congo red-stained amyloid between crossed
polariser and analyser (Goldstein, 1969; Taylor et al., 1974). The
contribution of this mechanism was shown to be negligible at the
thicknesses of sections of Congo red-stained amyloid in routine
use, in which the amount of absorption by Congo red was too small
to have a significant rotatory effect (Howie et al., 2008).
4. Birefringence and retardance
Another term like dichroism that is frequently mentioned in the

literature on amyloid and Congo red, but rarely defined, is
birefringence. The definition of this term depends on the refractive
index, which is the ratio of the velocity of light in a vacuum or in air,
which are virtually the same, to the velocity in a light-transmitting
medium (Brewer, 1957; Bennett, 1967; Hallimond, 1970). The
velocity of transmission and the refractive index are reciprocal, and
Fig. 4. Optical density of streaks of Congo red measured at three alignments
between the streaks and the polariser plane, without an analyser: 08 or parallel,
so a refractive index higher than another means that the velocity of
uppermost line; 458, middle line; and 908 or perpendicular, lowest line. This transmission is lower. The refractive index varies with tempera-
confirms that orientated Congo red is dichroic. Reproduced from Lab. Invest. 2008; ture, but this is not relevant in everyday microscopy.
88: 232–242 with permission. Some media such as various types of glass are isotropic and
have only one refractive index, irrespective of the plane of
crystallised in certain rigorous conditions as the potassium salt is polarised light illuminating the material. Others are anisotropic,
reported to show red and blue as dichroic colours (Mitchell, 1950). and have more than one refractive index. A birefringent material
Spectrophotometric examination of smears of Congo red transmits light polarised in one plane at a faster velocity than that
confirms the explanation of the change in intensity of red polarised perpendicularly, and has two extremes of refractive
depending on the plane of polarisation of light (Howie et al., index. Birefringence may be loosely defined to mean that a
2008). The absorption of light and the red colour of a smear aligned material has two refractive indices, although strictly speaking the
at 458 to the plane of polarisation are between those at 08, when the material has two extremes of refractive index, with a range
smear is orientated parallel to the plane of polarisation, absorption between the two limits. The orientations of the material with the
of blue/green is maximum, transmittance of blue/green is highest and lowest refractive indices are called the slow axis and
minimum, and red is strongest, and those at 908, when the smear the fast axis. The amount of birefringence is the difference between
is perpendicular, absorption of blue/green is minimum, transmit- the two limiting indices, and as the refractive index has no units,
tance of blue/green is maximum, and red is weakest (Fig. 4). the birefringence is a number with no units.
The transmittance of light at 458 is halfway between the In microscopy, birefringence can be detected by examination of
minimum at 08 and the maximum at 908, even though the a specimen between crossed polariser and analyser. This is often
absorption curve at 458 does not lie halfway between the other called examination in polarised light, which is ambiguous, because
two. This is because absorption is measured not on a linear scale this suggests use of only a polariser or only an analyser. Sometimes
but on a logarithmic scale, as optical density, which is the this arrangement is called use of cross-polarised light, which is
logarithm of the ratio of the intensity of light illuminating the inaccurate, because this implies that somehow the plane of the
specimen to the intensity of light transmitted by the specimen. polarised light has been rotated to be transverse to its initial
Dichroism may make a material appear bright against a dark orientation. The light is linearly polarised in the usual way, and the
background in the optical field if there is a marked difference in analyser, not the light, is crossed. Another phrase that has been
absorption between two axes of the material, which means that used is double or doubly polarised light, although the light is only
it is strongly dichroic (Goldstein, 1969; Hallimond, 1970). polarised once, and cannot pass the crossed analyser.
Investigation of this property requires examination on a Between crossed polariser and analyser, birefringent materials
microscope fitted with both a polariser and an analyser. When appear bright against the dark background. Examples are
the planes of polarisation of polariser and analyser are arranged unstained collagen bundles, starch granules, and oxalate crystals.
at right angles and they are said to be crossed, the background is Few materials appear uniformly bright at every position as the
maximally dark, because the linearly polarised light has no specimen is rotated. Some birefringent structures such as starch
component in a perpendicular plane and no light can pass the granules show dark bands with a Maltese cross appearance that
analyser. seem to rotate as the specimen is rotated. Ring-shaped birefringent
If the strongly absorbing axis of the material is arranged at 458 structures show four points that are dark and again appear to
to the plane of the polariser, light entering the material can be rotate. Linear birefringent structures become dark at some points
considered resolved into two vectors at right angles to each other, when the specimen is rotated.
one in the absorbing axis and the other in the relatively non- As mentioned in Section 3, strong dichroism can also make an
absorbing axis. The intensity of light is reduced in the vector in the object appear bright between crossed polariser and analyser, but
absorbing axis. When the vectors re-enter air, they recombine into this mechanism is negligible in Congo red-stained amyloid and will
one wave, and the resultant linearly polarised wave may have its not be considered further in relation to Congo red.
plane rotated towards the plane of the non-absorbing axis. This The explanation of the brightness of a birefringent object is
means that the light now has a projection in the plane of the related to the retardance. This is the birefringence multiplied by
analyser, and some can be transmitted to the observer. The effect of the specimen thickness. The dependence of retardance on speci-
this mechanism depends on the difference in absorption between men thickness explains how birefringent effects are stronger as the
thickness of microscopic sections increases. Retardance equals the

difference in the distance travelled in air, after leaving the object,
between a light wave polarised in the fast axis, and a light wave
polarised perpendicularly, which is delayed relative to the other
wave by travelling through the slow axis. To show this effect, the
waves must be coherent, which means they are generated
simultaneously by the same source, are of the same frequency,
and are capable of interference with each other.
Although the frequency of a light wave is a practically
unalterable property, and wavelength can vary depending on
the velocity, wavelength is more easily understood than frequency
in relation to colour in microscopy and is in common practice, and
will be used in this text, with the meaning of wavelength of a
particular wave in air. Retardance can be expressed as a linear
distance or as a phase difference, which means the distance divided
by the wavelength of the transmitted light. Measurement of
retardance can be done in various ways, and birefringence can be
calculated if the specimen thickness is known. Because the
thickness is not usually known accurately, retardance is the
commonly reported surrogate measure of birefringence.
When a birefringent object is aligned with its slow and fast axes
at 458 to the plane of polarised light, a light wave entering the
object from air can be considered to be resolved into two coherent Fig. 5. Diagram of production of elliptically polarised light. A birefringent object
with its slow and fast axes at 458 between crossed polariser and analyser interacts
vectors, one in the slow axis and the other in the fast axis. The with linearly polarised light to produce elliptically polarised light, some of which is
vector in the fast axis emerges from the object into air before the in the plane of the analyser, and can be transmitted, to make the object appear
other, and when the delayed vector emerges, they recombine into bright against a dark background. Reproduced from Bulletin of the Royal College of
one wave. Before entering the object, the wave vibrates in only one Pathologists October 2008; 144: 263–266 with permission.
plane perpendicular to the direction of travel. If viewed end-on by
an observer facing the oncoming light, the wave vector traces a plane of polarisation between crossed polariser and analyser. If
straight line in the plane of the polariser. After leaving the object, the object is not moved but the polariser and analyser are
the tip of the combined vector may no longer trace a straight line in uncrossed, the birefringent effects progressively decline. The
the plane of the polariser, depending on the retardance. reasons are that first for any given retardance the size of the ellipse
If the retardance is nil, as it is for a non-birefringent object, or if reduces, and second the background becomes lighter as more light
the retardance is a whole wavelength or a multiple of this, the is directly transmitted through the polariser and analyser. This
combined vector still traces a straight line as linearly polarised quenches the brightness of the object, which is most obvious
light in the plane of the polariser, and so no light can pass the between crossed polariser and analyser because the background
crossed analyser. When examined in light of relevant wavelengths, is dark.
the object appears as dark as the background. If the polariser and analyser are kept crossed but the object is
For almost any other retardance, the tip of the combined vector rotated, the effects reduce again as the size of the ellipse reduces, to
now traces an ellipse if viewed end-on from the direction of travel. disappear when the fast and slow axes are parallel and
This is elliptically polarised light (Fig. 5). The ellipse gives some perpendicular to the plane of polarisation, since the light wave
light in the plane of the analyser which can be transmitted, and so cannot be resolved into vectors in both axes.
the object appears bright against the dark background. The tip This explains how birefringent objects with a linear structure
of the combined vector of elliptically polarised light rotates become as dark as the background at every 908 during rotation, and
either clockwise or anticlockwise if viewed end-on, depending on how ring-shaped birefringent structures have four dark points, one
the relative orientations of the polariser and the object, and on the in each quadrant, at positions corresponding to the planes of the
retardance, which affects the position in time and space of the polariser and analyser (Fig. 6). These dark areas appear to rotate as
resultant of the vectors in the fast and slow axes. the specimen is rotated, but actually they remain fixed in the
If the retardance is a quarter of a wavelength, or three quarters, planes of the polariser and analyser. Similarly this explains the
or one wave and a quarter, and so on, the tip of the vector traces a appearance of a dark Maltese cross in otherwise bright objects such
circle when seen end-on. This is circularly polarised light, which as starch granules, and how the cross seems to rotate if the slide is
can be considered a special type of elliptically polarised light, in rotated, although the cross remains fixed relative to the polariser
which the principal axes of the ellipse are equal. and analyser. These birefringent objects have a circular arrange-
The dimension of the ellipse in the plane of the analyser, at right ment of one axis and radial of the other.
angles to the plane of the polariser, increases with retardance until
the maximum is reached when there is a phase difference of half a 5. Types of birefringence
wavelength, or a wavelength and a half, and so on. At these
retardances, the combined vector now vibrates as linearly Birefringence is of four types (Bennett, 1967).
polarised light in the plane of the analyser. This can be considered
another special type of elliptically polarised light, in which one axis (1) Intrinsic birefringence is that due to asymmetric arrangement
of the ellipse has no length. Maximum transmission of light and of molecules or molecular components that affect the velocity
brightness of the object are when the retardance is half a of linearly polarised light. This is a different property from
wavelength, or a wavelength and a half, and so on. absorption of light, which reduces the intensity of transmitted
The brightness of an object due to birefringent effects is light. Intrinsic birefringence is the most important type in
reduced when its slow and fast axes are no longer at 458 to the Congo red molecules.
Linearly polarised light can be considered a mixture of two

coherent beams of circularly polarised light of equal amplitude and
opposite direction of rotation, left- and right-handed, with phases
simultaneously equidistant from the plane of polarisation. If there
is circular birefringence, the plane of linearly polarised light is
rotated, and the material is said to show optical rotation or optical
activity. This property arises because molecules or combinations of
them have a handedness of structure that cannot be superimposed
on their mirror image, irrespective of their orientation. Optical
rotation is independent of the plane of polarisation of light
illuminating the circularly birefringent material, and the amount of
rotation is proportional to the concentration of the material and
the distance the light beam travels through the material, which is
the optical path length.
This is a different meaning of the word rotation from
mechanical rotation of a polarising filter or the specimen stage,
and from the apparent direction of rotation of elliptically or
circularly polarised light (Section 4). Optical rotation is also a
different mechanism from rotation of the plane of linearly
polarised light by a strongly dichroic material, which is dependent
on the plane of polarisation relative to the dichroic object, as
Fig. 6. Section of kidney stained by Congo red between crossed polariser and
described in Section 3.
analyser. Amyloid in tubular basement membranes appears green against the black Optical rotation varies with wavelength and is maximal around
background, except at the four positions parallel to the planes of polariser and an absorption peak, where it changes markedly, with a change
analyser. This is the appearance of Congo red-stained amyloid under ideal from rotation of the plane of linearly polarised light in one
conditions. Reproduced from Lab. Invest. 2008; 88: 232–242 with permission.
direction to rotation in the other direction, which is called
anomalous optical rotatory dispersion. A circularly birefringent
(2) Form birefringence is that due to orientated, asymmetric material usually also shows circular dichroism, which means
particles larger than molecules, or to micellar aggregates of different absorption of left- and right-handed circularly polarised
chains of molecules, which have at least one dimension smaller light. This is also maximal around an absorption peak and changes
than the wavelength of light, and whose refractive index is linearly polarised light to elliptically polarised.
different from that of the surrounding medium. Form Combined changes in optical rotation and circular dichroism
birefringence varies with the difference in refractive index around an absorption peak are the Cotton effect, named after Aimé
between the particles or micelles and the medium, and is August Cotton (1869–1951), a French physicist who described the
abolished when the refractive indices are the same. effect in 1895 (Cotton, 1895a,b). This effect on linearly polarised
light may mean that some light is now in the plane of the analyser,
This type of birefringence is shown by measuring the which can make a circularly birefringent structure appear bright
retardance of a material such as amyloid in an unstained section between crossed polariser and analyser.
which is successively mounted in media of different refractive Congo red in solution had no optical activity, but addition of a
indices. The birefringence is highest in media of high or low suspension of amyloid fibrils produced anomalous optical rotatory
refractive indices. The curve of retardance against the refractive dispersion centred on 520 nm in the green part of the spectrum,
index of the medium is U-shaped. When the refractive index of the and circular dichroism maximal between 500 and 550 nm. These
medium is the same as that of the material, which is at the lowest findings meant that Congo red-stained amyloid showed a Cotton
point of the U, the retardance due to form birefringence is nil. In effect, which was postulated to explain optical properties of Congo
this circumstance if retardance is still present, it is due to intrinsic red-stained amyloid (Benditt et al., 1970).
birefringence of the material. If this had a significant part in the optical properties, Congo red-
Unstained amyloid fibrils were shown to have both intrinsic stained amyloid at any position relative to the plane of linear
birefringence and form birefringence by Romhányi (1949), polarisation would allow some light to pass a crossed analyser, and
giving a more detailed account of studies first published in should never appear as dark as the background, whatever the plane
1943. Both types of birefringence, especially intrinsic birefrin- of polarisation. The observation that in Congo red-stained amyloid
gence, are weak in unstained amyloid, and barely produce there are four positions corresponding to the planes of the polariser
enough retardance to give appreciable transmission of light by a and analyser at which the material appears as dark as the
crossed analyser. In amyloid stained by Congo red, any background (Fig. 6) shows that in practice there is a negligible
retardance produced by birefringence of amyloid is insignificant contribution of circular birefringence to the optical properties
compared with the greater intrinsic birefringence of orientated (Howie et al., 2008).
Congo red. This is not surprising because first, even if concentrations of
Congo red and amyloid were similar to those in sections, the study
(3) Strain birefringence is produced by mechanical stress in a of optical rotation used an optical path length of 10 cm rather than
material, including when a solution is made to flow. Stressed a length of about 10 mm comparable to the thickness of sections
glass shows strain birefringence, and this is relevant in (Benditt et al., 1970), and second, optical activity and circular
polarising microscopy. dichroism generally have a 1000-fold weaker effect than linear
(4) Circular birefringence, as opposed to linear birefringence which birefringence in materials that show both circular and linear
is of the three other types, means that a material has different birefringence (Moxon and Renshaw, 1990).
refractive indices for left- and right-handed circularly polarised The difference between these types of birefringence was said
light (Barron, 2004). to be much less than this in Congo red-stained amyloid, but the
measurements to support this claim are not easily found in the

paper (Taylor et al., 1974): ‘Since the bound Congo red induces
optical activity in the complex, an ellipticity due to circular
dichroism is added to the ellipticity due to [linear] birefringence.
The dispersion of [linear] birefringence produces a larger
ellipticity compared to circular dichroism by a factor of about
20.’
In the following sections, birefringence only refers to linear
birefringence.
6. The colours seen in Congo red-stained amyloid and

orientated Congo red between crossed polariser and analyser
By far the main interest in optical properties of amyloid has

been in the striking colours that can be seen transmitted by a
crossed analyser when amyloid is stained by Congo red. These
colours are almost always said to be caused by birefringence, Fig. 7. Section of kidney stained by Congo red between crossed polariser and
although how birefringence can produce the colours is hardly ever analyser. Amyloid in an arterial wall appears yellow/green and blue/green, or
attempted to be explained. yellow and green. This is a typical appearance in everyday practice. Reproduced
from Bulletin of the Royal College of Pathologists October 2008; 144: 263–266 with
Even the description of the colours is almost always inaccurate.
permission.
Various phrases have been used, such as ‘green birefringent colour’
or ‘green birefringence’ or ‘green dichroism’ or ‘green-red
dichroism’. Probably the most popular phrase is ‘apple-green The colours change in a different way when the polariser or
birefringence’. Although this seems to have appeared in the early analyser is rotated from the accurately crossed position. If pure
1970s, the originator is difficult to discover. The author of the first green is the initial colour, rotation one way gives progressive
published mention of ‘apple-green birefringence’ that we have changes through yellow, orange and bright red to dull red, while
traced (Francis, 1973) said that he had not thought of the term, but rotation the other way gives light green or blue/green, bright white
had heard or read it used by someone else, now unidentifiable and then a neutral, colourless appearance. A streak of Congo red
(Francis, R.J., personal communication, 2008). shows only one effect at a time, while a section of Congo red-
Neither dichroism nor birefringence can properly be said to be stained amyloid shows both effects at the same time, but in
‘apple green’ or any other colour. Both are optical properties, different parts of the field. The colours are modified if combina-
arising from absorption and the refractive index respectively, that tions of colours rather than just green are seen between crossed
relate to differences in interaction between a material and light polariser and analyser (Figs. 9 and 10) (Howie et al., 2008).
polarised in different planes, and can explain colours, but cannot These asymmetric effects are known to pathologists and are a
have an intrinsic colour themselves (Sections 3 and 4). Even if the help in the diagnosis of amyloid. Possibly uncrossing of the
terms are used informally to describe colours, dichroism is polariser and analyser is the explanation of the colours described
misused because it is not detected between crossed polariser by Schwartz (1970).
and analyser, as described in Section 3, and there are various Even if the brightness of the colours is not strong, detection of
problems with the colour ascribed to birefringence. colours against a dark background is much easier than either
Green may appear the only colour when Congo red-stained detection of dichroic effects against a light background or detection
amyloid is examined between crossed polariser and analyser of faint staining by Congo red in unpolarised light. This is why
(Fig. 6). This does not change colour when the slide is rotated, pathologists routinely examine Congo red-stained sections
although there are always four positions at each 908 during between crossed polariser and analyser if the sections are
rotation at which there are bands as dark as the background in any suspected to contain amyloid or if exclusion of amyloid is
particular structure in a section. necessary.
In practice, to see only green is unusual, and much more
commonly, there are two or more colours (Fig. 7). These are often
yellow/green and blue/green, which if they are mentioned at all are
usually just called yellow and green. The green seen with yellow in
any particular specimen has a different shade from the green that
could theoretically be seen on its own, which itself may vary
between specimens. Which shade of green should be considered
the genuine ‘apple green’ is purely arbitrary. Occasionally more
definite yellow and blue are seen. When more than one colour is
seen, they exchange positions usually when the specimen is
rotated by 908 (Fig. 8). Sometimes the colours vary across the
microscopic field.
Even when published illustrations are said to show ‘apple-green
birefringence’, few actually show just one colour, and the fact that
more than one colour is evident in most figures is hardly ever
mentioned. One author who opposed this habit was Schwartz
(1970), although the colours he described are unusual: ‘(Green)
was never the only color of birefringent material evoked by Fig. 8. The section in Fig. 7 rotated 908, with the image presented in the same
polarized light: glaring white and yellowish or brownish-golden orientation. The colours have exchanged positions. Reproduced from Bulletin of the
strips and spots were very often present too’. Royal College of Pathologists October 2008; 144: 263–266 with permission.
Congo, ont également la propriété de mettre en relief la

birefringence de l’amyloide’. (Note added during [proof] correc-
tion: other dyes, notably Congo red, have equally the property of
enhancing the birefringence of amyloid.)
In the short paper written in collaboration with Florkin, the
description of the effect of iodine on amyloid is repeated, and the
effect of Congo red in increasing the birefringence is confirmed
(Divry and Florkin, 1927). They also described similar results with
amyloid deposits in a so-called sago spleen and in experimental
amyloidosis in the mouse, which suggested that these staining
reactions are a general property of amyloid.
They described only an increase in the birefringence of Congo
red-stained amyloid, but not a change in colour. Missmahl (1957)
claimed that Divry (1927) ‘bemerkte als erster die Zunahme der
Stärke der Doppelbrechung am Amyloid nach Kongorotfärbung,
außerdem fiel ihm das Auftreten einer grüner Polarisationsfarbe
Fig. 9. The section in Fig. 8 with polariser and analyser slightly uncrossed. The auf’. (Divry first noticed the increase in the strength of the
colours are now yellow/orange and a different blue/green. Reproduced from birefringence of Congo red-stained amyloid. He was also struck by
Bulletin of the Royal College of Pathologists October 2008; 144: 263–266 with the appearance of a green polarisation colour.)
permission.
Similarly Diezel and Pfleiderer (1959) wrote that ‘Divry and
Florkin (1927) beobachteten als erste, daß die durch Kongorot
7. Historical survey of colours described in Congo red-stained angefärbten amyloiden Strukturen bei Betrachtung im polarisier-
amyloid and orientated Congo red between crossed polariser ten Licht, unter Verstärkung der Doppeltbrechung, grün aufleuch-
and analyser ten’. (Divry and Florkin were the first to observe that amyloid
structures stained with Congo red, when observed in polarised
Bennhold (1922, 1923) simply described the specific staining of light, as well as the increase in birefringence, lit up green.)
amyloid with Congo red and made no observations with polarised Both these attributions are incorrect. Wolman and Bubis
light. Divry (1927) and Divry and Florkin (1927) noted that the (1965), Benditt et al. (1970) and others repeat similar incorrect
staining of amyloid with Congo red increased the birefringence. attributions.
These observations were first made in an investigation by Divry There is not surprisingly some confusion in the literature during
into the nature of cerebral lipids (Divry, 1927). The amyloid he the Second World War. During this period Romhányi, working in
studied was in so-called senile plaques in the brain. He described the University Department of Pathology in Budapest, in a brief
the staining of the plaques with Lugol’s iodine, producing a abstract of a communication given to the Hungarian Pathology
mahogany-brown colour, and giving an increased birefringence Society on 2–3 October 1942, described birefringence in Congo
compared with the faint birefringence of unstained plaques. As a red-stained amyloid which was up to five times the birefringence
result of this the iodine-stained plaques stood out when examined of unstained amyloid in water (Romhányi, 1943). He suggested, as
between crossed polariser and analyser as tiny luminous points had Divry and Florkin (1927), that this resulted from binding of
against a dark background, and they could be readily counted. Congo red to orientated micelles of amyloid.
In the main part of this paper he merely described Congo red as In this and a later paper (Romhányi, 1949), in which he again
staining the plaques of amyloid specifically, and did not mention reported the increase in birefringence produced by Congo red, he
any effect on the birefringence (Divry, 1927). The effect on the made no comment on the colour of Congo red-stained amyloid
birefringence was only mentioned in a footnote added during between crossed polariser and analyser, nor did he refer to the
correction of the proofs, on the page of legends to the figures: ‘Note work of Divry (1927). It is not possible to say whether his
en cours de correction: d’autres colorants, et notamment le rouge observations were made independently of those of Divry.
Ladewig (1945) in a short letter to the journal Nature seems to
have been the first to report colours in Congo red-stained amyloid
examined between crossed polariser and analyser. Curiously, he
does not make a single reference to any previously published work,
but as he was working in Istanbul during wartime such literature
may not have been available to him. He described the colours
changing twice from yellow to green during rotation of the slide
through 3608.
This is important because Ladewig made an accurate, indepen-
dent observation, unbiased by the thought that there should only be
a green colour. This observation is also important as it demonstrates
that the same area of a section or smear can appear green or yellow
according to the set-up of the microscope, and the colours do not
depend on the thickness of the specimen (Figs. 7 and 8).
In fact there had been an earlier description by Neubert (1925)
of a colour seen in orientated Congo red between crossed polariser
and analyser that was different from the colour in unpolarised
white light. He examined the birefringence of a variety of textile
Fig. 10. The section in Fig. 9 with further uncrossing of polariser and analyser in the
same direction. The colours are now red and pale blue/green. Reproduced from
fibres, celloidin, cellulose, nitrated cellulose, silk acetate, ramie
Bulletin of the Royal College of Pathologists October 2008; 144: 263–266 with fibre and nitrated ramie fibre, before and after staining with Congo
permission. red, and he also examined smears of Congo red.
Colours would not be apparent in preparations examined in 1953 (Missmahl and Hartwig), several years before this was
light of a single wavelength, called monochromatic light, even demonstrated by electron microscopy, is but one indication of its
using a range of different wavelengths, as Neubert did for many of inherent possibilities’.
his observations. Materials would only have appeared bright or Romhányi (1971) seems to have been a little put out by
dark in the colour of the light used. He recognised that his Missmahl’s attempt to discard his previous error, and he wrote,
preparations between crossed polariser and analyser in white light again in English: ‘It is not quite clear why Heller et al. (1964)
could show a different colour from Congo red in unpolarised light, referring to Missmahl and Hartwig (1953) claimed that they
but he was not particularly concerned with colours. He did not predicted the micellar ultrastructure of amyloid on polarisation
report a range of these, and just said that smears of Congo red optical grounds’.
appeared yellow.
Wälchli (1945) was interested in determining the mechanism 8. Normal dispersion of the refractive index, and destructive
of the interaction between Congo red and cellulose fibres, but he interference
did not describe seeing any unexpected colours during studies of
birefringence, even though he was aware of the work of Neubert For colours to appear, wavelengths of light must be transmitted
(1925). It may be that he only examined the dyed fibres in unequally. Birefringence, alone or combined with other factors,
monochromatic light of different wavelengths. must affect wavelengths differently through the spectrum to give
To return to amyloid, the study by Missmahl and Hartwig colours transmitted by a crossed analyser.
(1953) showed that Congo red-stained amyloid appeared green, Materials that transmit light but do not absorb it, such as many
and only green, between crossed polariser and analyser. Although types of glass, have a slightly different refractive index for different
they included a reference to the paper of Ladewig (1945), they did wavelengths (Hartshorne and Stuart, 1970). There is a gradual
not mention his precise description of how the colours seen in reduction in refractive index as the wavelength increases. This is
Congo red-stained amyloid changed from green to yellow and back called normal dispersion of the refractive index. Because the
to green with rotation of the specimen. refractive index for violet is higher than that of blue, and blue of
They also disagreed with the suggestion made by Divry (1927), green, and so on, white light is split into the spectrum when a beam
Romhányi (1943) and Ladewig (1945) that these findings indicated passes through a triangular prism, and red is refracted least.
that amyloid has an ordered micellar structure. Instead, they A birefringent material that does not absorb light, such as
interpreted their findings to be due to pre-existing connective unstained amyloid, shows normal dispersion of the refractive
tissue fibres included in the amyloid deposits and masked by them. index in both the fast and slow axes. The birefringence of unstained
Thus Missmahl and Hartwig (1953) wrote: ‘Daher möchten wir amyloid, which is the difference between the refractive indices of
auf Grund dieser Befunde über die Doppelbrechung an Amyloida- the two axes, is the same at every wavelength (Taylor et al., 1974).
blagerungen im Gewebe im Gegensatz zu Romhány und Ladewig For any particular specimen, all wavelengths have the same
aus der Anisotropie keine Rückschlüsse mehr auf die Feinstruktur retardance. Wavelengths are transmitted equally by unstained
dieses Eiweißes ziehen, sondern annehmen, daß die Doppelbre- amyloid. Between crossed polariser and analyser, this appears
chung auf die in den Geweben präexistenten und durch white, which is another way to say colourless.
Amyloidabscheidung maskierten Fasern zurückzuführen ist’. This does not mean that every birefringent, non-absorbing
(Because of this, in contrast to Romhány and Ladewig, we do material appears white. This is because even though the retardance
not wish to draw any conclusions as to the fine structure of this is the same at every wavelength, the phase difference is not the
protein from the findings on the birefringence of amyloid same. For instance, a retardance of 400 nm is a whole wavelength
deposition in the tissues, but we accept that the birefringence of light in the violet band, but is rather over half of 700 nm, the
arises from pre-existing fibres in the tissues that are masked by the wavelength of light in the red band. At this retardance, violet is not
amyloid deposition.) transmitted because its phase difference means it is linearly
In a later study, Missmahl (1957), who has been mentioned a polarised in the plane of the polariser and cannot pass the crossed
few times and was a physician in Tübingen under the direction of analyser. Other wavelengths are elliptically polarised and are
Bennhold, made important observations on the birefringence of passed by the analyser to different extents. The appearance is
amyloid fibrils and non-amyloid fibres before and after staining white lacking violet, which is yellow.
with Congo red. He also investigated smeared preparations of This effect is called destructive interference. There are hardly
Congo red. These showed dichroism, and also, between crossed any noticeable effects of interference until the retardance is over
polariser and analyser, ‘demselben Grün wie die amyloidhaltige about 300 nm. As the retardance progressively increases above
Faser’ (the same green [colour] as the amyloid-containing fibres). this, various wavelengths and combinations of them in turn cannot
This is the usual finding in smears of Congo red (Howie et al., 2008), pass the analyser, and the resultant interference colours form a
which makes the report by Neubert (1925) of a yellow colour series known as Newton’s scale (Bennett, 1967; Hallimond, 1970;
difficult to understand. Hartshorne and Stuart, 1970).
Diezel and Pfleiderer (1959), who were in the Institute of The relatively large values of retardance required to produce
Pathology at the University of Heidelberg, reported green and destructive interference are uncommon in biological material. An
yellow colours in Congo red-stained amyloid between crossed example of something that can give such retardance is surgical
polariser and analyser, as Ladewig (1945) had done. They wrote suture material in sections of pathological specimens. The
that Romhányi (1956) described the colour as green. In fact, he did retardance of unstained amyloid is so small that interference
not. He was mainly concerned to refute the mistaken claim by does not occur, and no colour is seen.
Missmahl and Hartwig (1953) that the birefringence was not due A commonly ignored fact in the literature on Congo red-stained
to amyloid, but was due to fine connective tissue fibres. amyloid is that green is produced by absorption from white light of
Missmahl persisted in this highly individual and incorrect wavelengths at both ends of the spectrum, meaning violet and red
interpretation of the structure of amyloid until 1964, when he light. Absorption of either colour alone does not give green, but
sought to disown it (Heller et al., 1964). In this paper, written in gives yellow and blue respectively. Any green colour can be
English, he and his co-authors wrote in reference to polarising matched by a suitable mixture of yellow and blue. To give green by
microscopy: ‘The prediction of the fibrillar structure of amyloid in destructive interference, the lowest retardance of white light must
be between 700 and 800 nm, when violet is retarded by

approximately two whole wavelengths and red by approximately
one whole wavelength. Some higher values of retardance also give
green (Bennett, 1967; Hartshorne and Stuart, 1970).
Destructive interference was suggested by Wolman and Bubis
(1965) as the explanation of the ‘green polarization color’ of
smeared crystals of Congo red, polished dried smears of Congo red,
and sections of Congo red-stained amyloid between crossed
polariser and analyser. ‘When a near-perfect crystal of Congo red of
a given thickness is examined under crossed polars when situated
at 458 to the polarizer axis, the polarized light which enters the
object is split into two rays, a red one which runs parallel to the
length of the crystal and a white one perpendicular to it. The red
component of the white ray is obviously of the same amplitude
(and therefore intensity) as the red ray. On being brought together
by the analyzer the two rays can produce phenomena of
interference. Whenever the retardation of the slow ray is equal
to half the wavelength of red light, the red ray will cause complete
extinction of the red component of the white ray. . .. Extinction of
red produces the appearance of the complementary color, green.’
There are several errors in this. The main ones are that a
retardance of half a wavelength actually gives maximal transmis-
sion by a crossed analyser, as described in Section 4, and that
removal of red from white light gives blue, not green. In addition,
Fig. 11. Diagram of anomalous dispersion of the refractive index around an
there was no attempt to measure retardance to support or refute absorption peak. In a light-absorbing axis, the refractive index changes markedly at
the hypothesis. wavelengths around a peak, indicated by the vertical dotted line. On the shortwave
A retardance of half the wavelength of red light, as required by side, the index falls rapidly. On the longwave side, the index is high, and falls rapidly
the erroneous suggestion of Wolman and Bubis (1965), is about at first. Away from the peak, or in a non-absorbing axis, there is a gradual decrease
in refractive index with increasing wavelength, which is normal dispersion.
310–350 nm, which would appear light yellow. No published Reproduced from Bulletin of the Royal College of Pathologists October 2008; 144:
retardance of orientated Congo red has ever approached these 263–266 with permission.
values, which are about the lowest necessary for interference to be
apparent. An understanding of the explanation of normal and anomalous
Retardance depends on the thickness of a specimen, and so dispersion requires a profound knowledge of physics: ‘For an
varies between specimens and papers. The highest value of adequate treatment of dispersion, it would be necessary to go
retardance for smears of Congo red was 130 nm measured by deeply into the atomic theory of matter’ (Born and Wolf, 1999).
Neubert (1925), 54.5 nm measured by Missmahl (1957), and about Fortunately this is not essential for an understanding of the
18 nm measured by Taylor et al. (1974). For Congo red-stained properties of Congo red-stained amyloid.
sections of amyloid, the highest value was 18.1 nm measured by All light-transmitting materials show anomalous dispersion as
Missmahl (1957), 21 nm measured by Diezel and Pfleiderer (1959), well as normal dispersion, although for those that do not absorb
and about 20 nm measured by Taylor et al. (1974). The highest visible light, such as various types of glass, the region of anomalous
values in the study of Howie et al. (2008) were 95.4 nm on smears dispersion is in ultraviolet or infrared wavelengths.
and 86.6 nm on sections, and sections generally had lower Orientated Congo red is birefringent, and because Congo red
retardances than smears. absorbs light, its birefringence varies with wavelength in the
Destructive interference has no contribution in practice to the visible spectrum. Light is only absorbed when polarised parallel to
colours of Congo red-stained amyloid. the long axis of Congo red molecules. This axis is the one that in the
intrinsic birefringence of Congo red has more effect on the velocity
9. Anomalous dispersion of the refractive index of light than the non-absorbing axis.
The refractive index of the non-absorbing axis follows normal
The explanation of green and other colours seen between dispersion, while the index of the absorbing axis shows anomalous
accurately crossed polariser and analyser depends on a property of dispersion around the absorption peak, which is in the blue/green
any light-absorbing material which is called anomalous dispersion part of the spectrum. The birefringence is the difference between
of the refractive index around an absorption peak (Zocher, 1925; the two indices. The difference in absolute values is largest in
Born and Wolf, 1999). the region of the absorption peak (Fig. 12). The observed curve of
There is a dramatic change in refractive index around the retardance is smoother than that expected from the theory of
wavelengths that represent the peaks of absorption of a material anomalous dispersion (Fig. 11). This is the usual finding, because
(Fig. 11). The refractive index, which in accordance with normal first the narrow waveband filters used in measurement of
dispersion reduces gradually as wavelength increases, declines retardance do not give completely monochromatic light, and
sharply on the shortwave side of an absorption peak to a minimum. second there are damping effects within atoms (Born and Wolf,
The index then jumps to a maximum on the longwave side of the 1999).
peak, declining at first sharply after that, but then more gradually, Because retardances of Congo red-stained amyloid and
following normal dispersion again. Normal dispersion curves orientated Congo red are in practice all below half of any
theoretically approach a limiting value of refractive index as they wavelength (Section 8), the largest birefringence should give the
gradually decline with increasing wavelength, but on the longwave largest elliptical polarisation (Section 4), and the colour trans-
side of a region of anomalous dispersion, the limiting value is mitted by a crossed analyser should depend on the net effect of
higher than on the shortwave side. birefringence at different wavelengths. In theory for Congo red the
example, dichroic preparations with two absorption maxima.

They reported quantitative observations on about 160 dyes
including Congo red, although its properties were not specifically
described. In about three-quarters of these dyes, the smears
showed a simple dichroism, meaning that there was only one
absorption peak, and the curve of birefringence confirmed their
theoretical predictions. Surprisingly, in this great mass of data they
did not mention that they observed any unexpected colours with
the changing conditions between crossed polariser and analyser.
The direction of the change in sign of birefringence around an
absorption peak can be used to determine the orientation of dye
molecules. In collagen stained by toluidine blue, the birefringence
was positive on the shortwave side of the absorption peak of the
dye, and became negative on the longwave side (Brewer, 1957).
This indicated that toluidine blue molecules were orientated with
their light-absorbing intramolecular bonds at right angles to the
long axis of collagen molecules. The opposite change in Congo red-
stained amyloid confirms that Congo red molecules absorb light
polarised parallel to the long axis of amyloid fibrils (Howie et al.,
Fig. 12. Measured retardance of streaks of Congo red (red dots), and of orientated 2008).
Congo red-stained amyloid on sections (green dots), aligned at 458 to the plane of
polarisation between crossed polariser and analyser. Absolute values are maximum
To understand the consequences of this change in sign of
around 500 nm, the absorption peak of Congo red, with a change from negative birefringence, an account of compensation is required (Bennett,
retardance at shorter wavelengths to positive at longer. Reproduced from Lab. 1967; Hallimond, 1970; Born and Wolf, 1999). A compensator is an
Invest. 2008; 88: 232–242 with permission. optical device that compensates or neutralises a phase difference
produced by birefringence, which means that the phase difference
analyser should give most transmission of wavelengths around the is changed to nil or to a whole wavelength. This converts
absorption peak (Howie et al., 2008). elliptically polarised light to linearly polarised. There are various
This mechanism is due to the intrinsic birefringence of types of compensator, such as a quartz wedge, used by Neubert
orientated Congo red molecules and could be taken as the (1925). Another type, an elliptical compensator, is a birefringent
explanation of the green colour of orientated Congo red under plate of known retardance, which is positioned in the optical path
these conditions, although in fact it is only part of the explanation. between the specimen and the analyser, and is rotated to give
polarised light of variable ellipticity. The direction of rotation of the
10. Sign of birefringence, and compensation elliptically polarised light can be changed by altering the direction
in which the compensator is turned relative to the polariser.
Another consequence of anomalous dispersion that helps to This compensator converts elliptically polarised light of a
explain green and other colours is that not only is the birefringence specific wavelength to linearly polarised when the ellipse
largest around the absorption peak, but there is also a change in produced by the compensator is opposite in direction of rotation
sign of the birefringence. Birefringence is conventionally called and at least equal in dimensions to the ellipse produced by the
positive when the slow axis, which means the one with the higher birefringent object. The resultant linearly polarised light is in the
refractive index, is orientated parallel to a distinctive feature such plane of the polariser or is extremely close to this plane, and so
as long axis of a fibre or direction of smearing, rather than light of the compensated wavelength cannot be perceived to be
perpendicular to the feature, when the birefringence is negative. passed by a crossed analyser.
For Congo red, at wavelengths above the absorption peak, the An elliptical compensator can be used to measure the
higher refractive index is in the long axis of the molecules, and the retardance of materials with relatively slight birefringence, such
birefringence is positive. At wavelengths below the peak, the long as Congo red-stained amyloid (Howie et al., 2008). This investiga-
axis has a lower refractive index than the relatively constant index tion requires the use of monochromatic light, in case not every
of the non-absorbing axis at right angles, and the birefringence is wavelength is compensated at any setting of the compensator.
negative (Figs. 11 and 12) (Howie et al., 2008). The compensator may be rotated until the material becomes
A change in sign of birefringence around an absorption peak is maximally dark, and from the retardance of the compensator and
not only found in orientated Congo red. This is a general property of its angle of rotation, the retardance of the material can be
all orientated dichroic dyes, but is not widely known in calculated. As the compensator is rotated, the background
microscopy. The change in sign at different wavelengths means becomes lighter, because the linearly polarised light in the
that the slow axis of transmission of any particular birefringent background is converted to elliptically polarised, and is trans-
material is not necessarily always in the same orientation relative mitted by the analyser. This makes judgment of the darkest
to the arrangement of the material. appearance of the material difficult. In practice, an easier and more
Zocher (1925) used the theory of normal and anomalous accurate method is to rotate the compensator until the brightness
dispersion of the refractive index to show that birefringence should of the material and the background match, which is the half
change in sign around an absorption peak, and confirmed this by compensation method of Bear and Schmitt (1936) (Brewer, 1957;
making measurements of the birefringence of smeared prepara- Bennett, 1967).
tions of 50 dyes, not individually listed. These apparently included The elliptical compensator also easily determines the sign of
Congo red, because he mentioned Congo red in passing in a later birefringence from the direction it is rotated to give compensation,
section on dichroism of the 50 dyes. but this may not be so straightforward with other types of
In a second paper Zocher and Jacoby (1927) considered in more compensator.
detail the theoretical derivation of the curve of birefringence For retardances on opposite sides of the absorption peak,
against wavelength in more complicated cases such as, for negative and positive birefringence give opposite directions of
rotation of elliptically polarised light. Just as the unaided human

eye is not able to detect whether light is polarised, effects produced
by different signs of birefringence cannot be distinguished without
optical aids. Because the human eye cannot detect the direction of
rotation of elliptically polarised light, colours passed by opposite
signs of birefringence at the same point are blended to give a single
colour.
In perfect conditions, Congo red-stained amyloid examined
between crossed polariser and analyser appears green, which is the
colour blended from blue as a result of effects of negative
birefringence at wavelengths shorter than that of the absorption
peak, and yellow, from positive birefringence at longer wave-
lengths (Fig. 6) (Howie et al., 2008).
This confirms the finding of Missmahl and Hartwig (1953) that
a pure green colour may be seen in Congo red-stained amyloid
between crossed polariser and analyser.
In reality, conditions are usually not perfect for polarising Fig. 14. The section in Fig. 8 with an elliptical compensator in the light path, rotated
microscopy. There is additional birefringence in the optical system to give even more yellow and blue. This shows that the green in Fig. 13 is a blend of
on many routine microscopes, especially strain birefringence in yellow from effects of positive birefringence and blue from effects of negative
birefringence. Reproduced from Bulletin of the Royal College of Pathologists
components such as objective lenses, glass slides, and coverslips
October 2008; 144: 263–266 with permission.
(Section 5). This birefringence acts as a compensator and converts
elliptically polarised light produced by the birefringence of
orientated Congo red to linearly polarised, which cannot be Figs. 7 and 8 can be removed by appropriate rotation of an elliptical
passed by the analyser. compensator to give uniform green.
Any particular compensating agent can act at any point only on This figure also shows that the pure green that can theoretically
elliptically polarised light with one direction of rotation, and so at be seen in any particular specimen of Congo red-stained amyloid
that point can only affect light produced as a result of birefringence between crossed polariser and analyser is a different shade from
of one sign, for example, negative. That direction of rotation may the green seen in the specimen when there is also yellow, which is
be produced at other points in the optical field by positive a bluer shade of green (Figs. 7 and 8). In addition, a comparison of
birefringence, in this example, if the fast and slow axes of Congo Figs. 6 and 13 indicates that even the pure green can appear
red at these other points are at an orientation of 908 relative to the different, depending on the specimen, conditions of staining and
axes at the first point. In the whole field, both negative and positive illumination, and conditions of photography, which emphasises
birefringent effects may be compensated, but at different points. further that ‘apple green’ has been used to describe several
This means that if there is additional birefringence in the optical different colours, as mentioned in Section 6.
system, blue resulting from effects of negative birefringence may Fig. 14 shows that rotation of an elliptical compensator can also
be partly or completely compensated, changing the net colour seen mimic and exaggerate the effect of strain birefringence, giving
by the observer from green in ideal conditions to yellow/green or yellow and blue, which exchange positions when the compensator
yellow. Similarly, compensation of yellow produced by effects of is turned the opposite way. This confirms that the pure green is
positive birefringence gives blue/green or blue (Figs. 7 and 8). actually produced by balanced transmission of blue, representing
An elliptical compensator can be used to restore elliptical wavelengths shorter than the absorption peak of Congo red, and
polarisation after strain birefringence has had a compensating yellow, the net colour of wavelengths longer than the peak (Fig. 12)
effect. Fig. 13 shows that effects of strain birefringence seen in (Howie et al., 2008).
The yellow and green colours reported by Ladewig (1945), seen
in different parts of a section of Congo red-stained amyloid, which
exchanged positions when the microscopic field was rotated by
908, are explained by unintentional strain birefringence in the
optical system (Figs. 7 and 8). Strain birefringence may not be
uniformly distributed across a microscopic field, in which case the
amount of compensation and the colours seen may also not be
uniform (Howie et al., 2008).
Missmahl (1968), reported in English during discussion at a
meeting, insisted that only green should be seen, but he realised
the importance of strain birefringence, even though he did not
explain how it had an effect: ‘If you see green and yellow in one
picture, or green and red, then your microscope is wrong. . .. The
optical system must be absolutely strain free.’
This statement is correct if a microscopist considers that it is
essential to see only green to make a diagnosis of amyloid. In
everyday practice there is no need to use a perfect polarising
microscope, or to insert an elliptical compensator to neutralise
Fig. 13. The section in Fig. 8 with an elliptical compensator in the light path, rotated strain birefringence, because the optical properties of Congo red-
to compensate strain birefringence and give a uniform green. This green has a
stained amyloid are sufficiently well recognised to be character-
different appearance from that in Fig. 6, which is a different specimen,
photographed at a different time and under different conditions. Reproduced istic even when there is strain birefringence in the optical system.
from Bulletin of the Royal College of Pathologists October 2008; 144: 263–266 with Incidentally, although strain birefringence explains green and
permission. yellow as implied by Missmahl (1968), green and red are much
more plausibly explained by uncrossing of polariser and analyser, die das Objekt in Diagonalstellung zeigt, ist keine reine Inter-
as described in Section 6 (Fig. 10). ferenzfarbe, sondern eine Mischung dieser mit der Absorptions-
farbe’. (The colour that the object shows in the diagonal position is
11. Contribution of absorption to the colours of Congo red- not a pure interference colour but a mixture of this with the
stained amyloid and orientated Congo red between crossed absorption colour.) In Section 8, interference was shown to have no
polariser and analyser contribution.
Missmahl (1962) explicitly dismissed any contribution from
Anomalous dispersion of the refractive index around the absorption: ‘Da an dünnen Schichten, wie sie in 5–10 mm dicken
absorption peak of Congo red, with compensation from strain histologischen Schnitten vorliegen, die Absorption des Farbstoffes
birefringence if present (Section 10), explains some optical zwischen gekreuzten Polars nicht in Erscheinung tritt, muss
properties of Congo red-stained amyloid, but was mentioned in dementsprechend die Doppelbrechung grün sein’. (In thin films as
Section 9 to be not the only contributor to the colours. As light of in 5–10 mm thick histological sections the absorption of the dye
the appropriate wavelengths and in the appropriate plane of between crossed polars does not become evident. Accordingly the
polarisation passes through the absorbing axis, some is inevitably birefringence must be green.)
absorbed. The optical density spectrum of orientated Congo red Both absorption and birefringent effects are maximal around
between crossed polariser and analyser is not just that expected the absorption peak of Congo red, but have opposite outcomes.
from birefringence alone, but is the net result of absorption and Absorption reduces the transmission of light while birefringent
birefringent effects (Fig. 15) (Howie et al., 2008). This is analogous effects give transmission. Absorption does not cancel birefringent
to insertion of a coloured screen in the optical path of a non- effects completely to give no transmission which would appear
absorbing birefringent material (Goldstein, 1969). black, nor do the two processes balance each other to give equal
This means that the green appearance of Congo red-stained transmission of all wavelengths which would appear colourless,
amyloid between crossed polariser and analyser in ideal conditions because their effects are not mirror images of each other (Fig. 15).
(Fig. 6) is a different shade from the green that would be seen if This is because birefringent effects are symmetric around the
only birefringent effects contributed to it. Similarly, the blue and peak, and persist longer towards the red end of the spectrum than
yellow that are respective results of effects of negative birefrin- absorption, which is asymmetric and declines rapidly on the
gence at wavelengths below the absorption peak and positive longwave side of the peak. There is not a symmetrically rapid
birefringence above the peak (Fig. 14), and are blended to give decline in absorption on the shortwave side because Congo red has
green (Fig. 13), are different shades of blue and yellow from those another absorption peak in the ultraviolet, and the two curves of
that would be seen if absorption had no effect. What the absorption overlap at the violet end of the visible spectrum (Fig. 4)
theoretically unabsorbed shades of green, blue and yellow would (Howie et al., 2008).
be can only be guessed, given the difficulties of prediction of colour Because in ideal conditions green is the transmitted colour of
perception from wavelength distributions, especially in low levels Congo red-stained amyloid between crossed polariser and
of illumination (Davson, 1990; Born and Wolf, 1999). analyser, there seems an assumption in some papers, not
There seems to have been no measurement of the optical supported by measurement, that there must be maximum
density curve of orientated Congo red between crossed polariser transmission of green wavelengths, which are approximately
and analyser before that of Howie et al. (2008). Many papers 510–570 nm, or 490–570 nm if blue/green wavelengths are
appear to assume that birefringence, usually by an unstated included. The maximum transmission is actually of yellow and
mechanism, is solely responsible for the colours transmitted by the orange wavelengths, about 575–600 nm, as a result of the net
analyser. Neubert (1925) realised that absorption must occur, effects of absorption and birefringence (Fig. 15). The appearance is
although he thought that the colour of a smear between crossed not yellow or orange because there is least transmission of violet at
polariser and analyser was due to this and interference: ‘Die Farbe, about 420 nm and red at 700 nm, and relatively more absorption at
the two ends of the spectrum is perceived as green, irrespective of
the position in the spectrum of the most transmission and least
absorption (Howie et al., 2008).
The physiology of colour vision is complicated but is a crucial
part of the perception of green. When light levels are reduced, as in
a microscopic field when polariser and analyser are crossed,
wavelengths in the green, compared with longer wavelengths such
as those in the yellow and orange, appear brighter than they do in
usual levels of illumination (Davson, 1990). This increases the
perception of green in these conditions. Also of physiological
relevance, there has been no formal study of the influence of
defective colour vision on the perception of the colours produced
by Congo red, but so-called colour blindness occurs in pathologists
just as in the general population (Rigby et al., 1991). In theory this
must affect the colours seen and reported. Perhaps this is why
Neubert (1925) reported only yellow in smears of Congo red
between crossed polariser and analyser.
Even in ideal conditions, green is only seen in sections or on
smears between crossed polariser and analyser if the amount of
Fig. 15. Optical density spectrum of orientated Congo red aligned at 458 to the plane Congo red is by definition enough to produce perceptible effects,
of polarisation between crossed polariser and analyser (red line), corresponding to but does not exceed a certain thickness. As thickness increases,
the observed green colour in Figs. 6 and 13, with curves of the two factors
contributing to it. The green line is the contribution of birefringence expected from
both absorption and birefringent effects increase, but absorption
retardance. The blue line is the contribution of absorption of streaks at 458 to a becomes increasingly dominant (Goldstein, 1969). The net effect is
polariser. Reproduced from Lab. Invest. 2008; 88: 232–242 with permission. that shorter wavelengths such as blue and green are increasingly
absorbed and transmitted less, longer wavelengths such as yellow 12. Measurements of retardance of Congo red-stained amyloid
and red are still not absorbed but have increased transmission, and and orientated Congo red, with suggested explanations of the
the colour changes progressively from green through yellow to colours seen
orange and then to red with increasing thickness.
This was shown by Howie et al. (2008) on smears of Congo red, Congo red-stained amyloid and orientated Congo red show a
and by Diezel and Pfleiderer (1959) using sections of liver change in the sign of the retardance, and so of birefringence,
containing amyloid at nominal thicknesses from 10 to 100 mm. around the absorption peak, in accordance with anomalous
The minimum and maximum thicknesses of Congo red required to dispersion of the refractive index in the absorbing axis and normal
give green are not known, because observations such as those of dispersion in the non-absorbing axis (Figs. 11 and 12) (Howie et al.,
Diezel and Pfleiderer (1959) on sections depend upon variables 2008). With absorption, and strain birefringence if present, this
such as the amount of amyloid at any point and the conditions of explains the colours seen between crossed polariser and analyser
staining, even if the stated thickness of a section is accurate. (Sections 9–11).
Wolman and Bubis (1965) suggested that green is only seen in a There have been a few other published measurements of the
narrow range of thicknesses, between 2 and 20 mm. In the retardance of Congo red, but surprisingly none has shown a change
thicknesses of sections used in everyday practice, the problem is in the sign of birefringence. All have reported only positive
more likely to be insufficient amyloid to give perceptible birefringence. This may be because a change in sign was not
birefringent effects than an excess of amyloid, giving colours recognised by the method of measurement, and the authors were
other than green in ideal conditions. Colours produced by the unaware of the theory of anomalous dispersion, which is not
mechanism of excessive absorption can be distinguished from mentioned by any of them. Without knowledge of this theory, any
those resulting from compensation by strain birefringence because explanation of the mechanism that produces the colours is
strain birefringence gives colours that change when the slide is incomplete and unsatisfactory.
rotated, while colours due to absorption in thick specimens stay Neubert (1925) was the first to show that the birefringence
the same in every quadrant during rotation (Howie et al., 2008). of orientated Congo red varies with wavelength. In his study
The contribution of absorption to the optical density curve of of textile fibres, he measured the retardance over a range of
Congo red orientated at 458 between crossed polariser and wavelengths. In all cases he found that the birefringence of
analyser is that of Congo red orientated at 458 to a polariser or unstained fibres did not vary significantly with wavelength, but
an analyser alone (Figs. 4 and 15). When the polariser or analyser is staining with Congo red increased the birefringence, and the
rotated from the crossed position, the absorption either increases, increase was different at different wavelengths. At some
if the plane of the polariser or analyser approaches the plane of the wavelengths the dyed fibres appeared isotropic, meaning non-
absorbing axis of the orientated Congo red, or decreases, if the birefringent, whereas at other wavelengths the birefringence was
polariser or analyser is turned the opposite way towards the non- increased.
absorbing axis, as expected from dichroism (Section 3). At the same From the fact that the dyed fibres were dichroic, he deduced
time, irrespective of the direction of rotation of the polariser or that the dye molecules were orientated and that the variation of
analyser, the background lightens, and birefringent effects decline, birefringence with wavelength was in some way a consequence of
as described in Section 4. this. He found that smears of Congo red were similarly dichroic and
The net result of the varying combination of decreasing that their birefringence varied with wavelength in the same way as
birefringent effects and increasing dichroic effects is that the the birefringence of Congo red-stained fibres.
colour changes from green in perfect conditions either to yellow, He measured the retardance of a smear of Congo red at five
orange, bright red and then dull red as the polariser or analyser is different wavelengths. He did not draw a graph of these results but
turned towards the absorbing axis of Congo red, or to light green or presented them in a two line table, with an unusual arrangement of
blue/green, bright white, and then colourless as the filter is turned longest wavelengths to the left and shortest to the right, probably
the other way (Howie et al., 2008). The two extreme appearances to reflect the order of increasing frequency of radiation. In the more
are those of pure dichroism. usual order, his values of retardance were nil at a wavelength of
The Cotton effect was proposed as the explanation of this 430 nm in the violet/blue, 30 nm at 480 nm in the blue, 110 nm at
change of colour by Benditt et al. (1970): ‘Slight uncrossing of the 530 nm in the green, 130 nm at 580 nm in the yellow, and 90 nm at
polarizer in one direction produces a yellow to yellow-orange color 630 nm in the red.
in the stained material, and uncrossing a similar amount in the He gave this description of the colour of smears, with an
opposite direction produces a blue-green color. . .. The rotation of explanation: ‘Auffallend ist, daß die Schichten verschiedener Dicke
the yellow-orange and blue-green light in opposite directions is im weißen Licht zwischen gekreuzten Nikols alle ein leuchtendes
obviously consistent with a Cotton effect.’ This has already been Gelb zeigen. Mit hilfe des Quarzkeiles kann man nun leicht
mentioned in Section 5 to be negligible. Isotropie des aufgestrichenen Kongorots für die Strahlen von den
If there is strain birefringence in the optical system, as there Wellenlangen l = 430–450 nm (also für blaues Licht) feststellen.
usually is, which introduces compensation and gives colours other Auf Grund der Kenntnis der Isotropie des Objekts für diese Farben
than green when Congo red is between crossed polariser and ist dann ohne weiteres verständlich warum die Schichten alle
analyser, rotation of either polariser or analyser gives various other zwischen gekreuzten Nikols gelb aufleuchten.’ (It is striking that
colours, from the combination of the usual birefringent and films of Congo red of different thickness in white light between
dichroic effects interacting with the effects of strain birefringence. crossed Nicols all show a bright yellow [colour]. One can easily
Such colours can be replicated under controlled conditions in establish, with help of a quartz wedge, that the smear is isotropic
smears of Congo red by use of an elliptical compensator rotated to for light of wavelengths of 430–450 nm (that is for blue light). On
different extents and in opposite directions, with gradual the basis of the knowledge of the isotropy of the films for these
uncrossing of the polariser and analyser (Howie et al., 2008). colours it can be readily understood why the films light up yellow
This variety of colours can be easily seen in practice but has between crossed Nicols.)
hardly ever been reported (Figs. 9 and 10). Indeed, some published By ‘readily understood’ he meant that failure to transmit blue
figures said to illustrate ‘apple-green birefringence’ actually show light would change white to yellow. He did not give a reason why
effects of uncrossing of the polariser and analyser. birefringence varied with wavelength.
Wälchli (1945) in a detailed study confirmed and extended the Diezel and Pfleiderer’s explanation of the change of retardance
work of Neubert (1925). He examined the optical properties of with wavelength is confused. They were misled by their attempt to
molecules of Congo red and other substances in solution, including explain the green and yellow colours they describe, which they
in flowing solutions. He found that the birefringence of Congo red suggested arose because the section was thicker in the yellow
was positive with respect to the direction of flow and so from this areas than the green areas. The only evidence they presented for
observation he deduced that the birefringence was positive with this was a claim that the birefringence in the yellow areas was on
respect to the long axes of the molecules or micelles. Whether he average twice as great as in the green areas. They gave no details of
used white light or monochromatic light for this observation is how these measurements were made.
unclear. He calculated the birefringence of 100 crystals of Congo In making such measurements the long axis of amyloid fibrils
red at only one wavelength, 656 nm, which is red light. must be arranged at 458 to the plane of polarisation, because in this
He repeated the observation of Neubert (1925) on smears of position the retardance is at its maximum, and in settings away
Congo red. He found again that the smears were positively from this position in either direction the retardance becomes less
birefringent with respect to the direction of smearing. He copied (Section 4). In a histological section containing amyloid, it may be
Neubert’s results on retardance, reversing the order of the difficult to find fibrils of sufficient length lying in the plane of the
wavelengths. He assumed arbitrarily that Neubert’s smear was section to enable such an orientation to be made. Further it would
0.93 mm thick and on that basis converted Neubert’s five values of be surprising if during rotation of the slide necessary to orientate it
retardance to birefringence, and produced a graph of birefringence appropriately the green or yellow areas did not change colour, as
against wavelength, with a peak at 580 nm, which is yellow light. Ladewig (1945) showed. The most likely explanation of the green
This appears to be the first graph published of the relation between and yellow seen by Diezel and Pfleiderer is not different
birefringence and wavelength in Congo red. thicknesses, but effects of strain birefringence (Figs. 7 and 8).
Wälchli also measured the effect of Congo red on the Missmahl (1962, 1968) twice published a graph of retardance of
birefringence of ramie fibres and cellophane film. Like Neubert, Congo red-stained amyloid against wavelength with 21 evenly
he found that the birefringence was increased and the increase was spaced points. This had a sharply convex appearance, rising from a
not the same at all wavelengths. It was least in blue light of retardance of 8 nm at a wavelength of 440 nm, to about 18 nm at
wavelength 430–450 nm and greatest in green light of wavelength 530 nm, and falling to about 8 nm at 640 nm, with a slight
550–560 nm. He offered no explanation of this. concavity on the down slope. This graph has a close resemblance to
Missmahl (1957) found that staining with Congo red made no that of Diezel and Pfleiderer (1959). Missmahl did not give a
significant difference to the retardance of non-amyloid fibres and reference or describe how he had made the measurements. In an
that the retardance was the same in green light, wavelength earlier work he had only made measurements of retardance at two
520 nm, and red light, wavelength 680 nm, the only two wavelengths (Missmahl, 1957). His explanation of the optical
wavelengths he used. With amyloid fibrils he found that the properties of Congo red-stained amyloid was complicated and
retardance was increased by staining with Congo red, and the confused (Missmahl, 1957, 1962).
increase was greater in green light, with a change of retardance Taylor et al. (1974) used a method called phase modulation
from 8.7 to 18.0 nm, than in red light, with a change from 8.7 to microspectrophotometry to make various measurements on
9.7 nm. He measured the retardance of a smear of Congo red and sections of amyloid, either unstained or stained with Congo red,
found a marked difference in the retardance measured in green and preparations apparently made by buffing dried Congo red on a
light compared with red light, 54.5 nm compared with 15.2 nm. slide with a glass rod. The retardance was only measured at
Missmahl was aware that such differences in retardance with wavelengths between 420 nm in the violet and 590 nm in the
wavelength must have an important influence on the appearance yellow. For apparently one section of Congo red-stained amyloid
of Congo red-stained amyloid in white light between crossed and one smear of Congo red, the graphs showed a small positive
polariser and analyser. Partly because he thought that under such retardance below about 490 nm in the blue, either no retardance or
circumstances Congo red-stained amyloid appeared only green, his a tiny positive retardance at about 500 nm in the blue/green, and a
explanation was severely incomplete. He wrote that green was peak of positive retardance at about either 560 nm in the green or
seen because the retardance was greater in green light than in red 580 nm in the yellow.
light, which was accordingly less transmitted. This is the same These curves of retardance against wavelength are completely
error as the explanation of Wolman and Bubis (1965). Both violet different in shape from those published by Wälchli (1945), Diezel
and red must be relatively more absorbed and relatively less and Pfleiderer (1959), and Missmahl (1962, 1968), all of whose
transmitted to give green. retardances were positive, and Howie et al. (2008), who had both
Diezel and Pfleiderer (1959) extended the observations of positive and negative retardances (Fig. 12). Possible explanations
Neubert (1925). They measured the retardance of Congo red- of this discrepancy are that the sudden change in retardance
stained amyloid at 14 wavelengths throughout the visible around an absorption peak was detected but the retardances at the
spectrum. They published a graph in which the retardance, shortest wavelengths were not recognised to be negative, or that
beginning at approximately 8.5 nm at a wavelength of 440 nm in somehow their method confounded effects of birefringence and
the blue, rose steeply to a maximum of about 21 nm at a absorption.
wavelength of 550 nm in the green, and then descended with a Taylor et al. (1974) reported ‘classical green birefringence’ in
slight concavity to a minimum of about 6 nm at a wavelength of Congo red-stained amyloid between crossed polariser and
680 nm in the red. analyser. They gave a highly complicated and unsatisfactory
For some reason they described this graph in the legend to the explanation of this, based on three processes: the Cotton effect,
figure as being of birefringence (‘Doppeltbrechung’) plotted which was mentioned in Section 5 to be insignificant in practice;
against wavelength. They defined birefringence correctly as rotation of the plane of linearly polarised light by different
retardance or path difference (‘Gangunterschied’) divided by absorption in two axes, which was mentioned in Section 3 to be
section thickness. They actually plotted retardance against also insignificant in practice; and birefringence, although they did
wavelength, as is evident from the values of retardance they give not make it clear how, due to this mechanism, ‘a drastic increase in
in a table. They also labelled the vertical axis with the Greek capital retardation occurs in the green portion of the spectrum with a peak
letter for gamma, which is the accepted symbol for retardance. value between 560 and 570 nm’.
13. Anomalous colours anomalous polarization color’. Taylor et al. (1974) used the phrases
‘anomalous polarization colors’ and ‘green-blue anomalous color’.
The features of Congo red-stained amyloid that can be seen in
various conditions of polarised light can be explained using
14. Conclusion
principles of physical optics. A final problem is to decide a
description that can be used in everyday practice to report the
Congo red-stained amyloid between crossed polariser and
finding of the characteristic colours. For various reasons, many of
analyser can be said to show an anomalous colour, such as green, if
the expressions used are inadequate or inaccurate, such as ‘green
that is the only colour, or anomalous colours, such as yellow/green
dichroism’ or ‘apple-green birefringence in polarised light’ or
and blue/green, commonly called yellow and green. This combina-
‘apple-green birefringence in cross-polarised light’.
tion, or yellow and blue, is more often seen in practice than pure
Colours transmitted by an analyser when a material is
green. The various other colours apart from red seen when the
examined between crossed polariser and analyser are called
polariser or analyser is progressively rotated from the crossed
anomalous, meaning they are different from the colour seen in
position are also anomalous colours, and are explained by the
unpolarised white light, but they are not explained by Newton’s
combination of absorption and changes in birefringence due to
scale of interference colours (Hartshorne and Stuart, 1970).
anomalous dispersion of the refractive index, usually with
Anomalous colours are well known in crystallography but seem
additional effects of strain birefringence in the optical system.
neglected in biological microscopy. One of us described anomalous
colours in collagen stained with eosin, toluidine blue, and other
dyes (Brewer, 1957). References
An analysis of anomalous colours, not in Congo red, was given
by Perutz and Mitchison (1950), who described such colours in red Amelin, V.G., Tret’yakov, A.V., 2003. Adsorption-bonded azo reagents in chemical
tests based on the principles of precipitation paper chromatography. J. Anal.
cells in sickle cell anaemia and in crystals of reduced normal Chem. 58, 740–747.
haemoglobin. These showed a deep blue colour between crossed Armstrong, H.E., 1935. Chemical industry and Carl Duisberg. Nature 135, 1021–
polariser and analyser, which became completely extinguished at 1025.
Barron, L.D., 2004. Molecular Light Scattering and Optical Activity, 2nd ed. Cam-
every 908 of rotation of the specimen. With varying amounts of bridge University Press, Cambridge, pp. 1–10.
compensation in one direction the blue colour changed to violet Bear, R.S., Schmitt, F.O., 1936. The measurement of small retardations with the
and finally to green. When the compensator was rotated in the polarizing microscope. J. Opt. Soc. Am. 26, 363–364.
Benditt, E.P., Eriksen, N., Berglund, C., 1970. Congo red dichroism with dispersed
opposite direction the colour changed to pale pink and finally to
amyloid fibrils, an extrinsic Cotton effect. Proc. Natl. Acad. Sci. U.S.A. 66, 1044–
white. They noted that the colours were completely different from 1051.
Newton’s scale of interference colours seen in more strongly Bennett, H.S., 1967. The microscopical investigation of biological materials with
birefringent material. polarized light. In: Jones, R.M. (Ed.), McClung’s Handbook of Microscopical
Technique. 3rd ed. Hafner, New York, pp. 591–677.
The explanation they proposed for these findings was identical Bennhold, H., 1922. Eine spezifische Amyloidfärbung mit Kongorot (a specific
to that of Zocher (1925), although he did not report anomalous staining of amyloid with Congo red). Münch. Med. Woch. 69, 1537–1538.
colours (Section 10). Their ideas were apparently reached Bennhold, H., 1923. Über die Ausscheidung intravenös einverleibten Kongorotes bei
den verschiedensten Erkrangungen insbesondere bei Amyloidosis (on the
independently of Zocher, because they did not refer to his work. elimination of intravenously absorbed Congo red in various diseases, in parti-
They suggested that the colours resulted from anomalous cular in amyloidosis). Deutsches Arch. Klin. Med. 142, 32–46.
dispersion of the refractive index in the absorbing axis compared Born, M., Wolf, E., 1999. Principles of Optics. Electromagnetic Theory of Propagation.
Interference and Diffraction of Light, 7th ed. Cambridge University Press,
with normal dispersion in the non-absorbing axis. Cambridge, pp. 12, 49, 95–103, 820–823, 843.
The retardance of one haemoglobin crystal varied with Brewer, D.B., 1957. Differences in the fine structure of collagen and reticulin as
wavelength in a complex manner. Perutz and Mitchison plotted revealed by the polarising microscope. J. Path. Bact. 74, 371–385.
Cotton, A., 1895a. Absorption inégale des rayons circulaires droit et gauche dans
birefringence but accepted that their values were approximate, certains corps actifs (unequal absorption of right and left circular rays in some
because the crystal thickness was unknown. There were two peaks active bodies). Comp. Rend. Acad. Sci. 120, 989–991.
of birefringence, a large one at about 460 nm in the blue and a Cotton, A., 1895b. Dispersion rotatoire anomale des corps absorbants (anomalous
rotatory dispersion of absorbing bodies). Comp. Rend. Acad. Sci. 120, 1044–
smaller broader one at about 600 nm in the yellow/orange. The
1046.
variation was related to the absorption curve of reduced Davson, H., 1990. Physiology of the Eye, 5th ed. Macmillan, London, pp. 395–402.
haemoglobin in solution. The two peaks were displaced a little Diezel, P.B., Pfleiderer, A., 1959. Histochemische und polarisationsoptische Unter-
to the longwave side of the two peaks of the absorption curve. suchungen am Amyloid (histochemical and polarising optical investigations of
amyloid). Virchows Arch. Pathol. Anat. 332, 552–567.
All birefringence was positive, although Perutz and Mitchison Divry, P., 1927. Etude histochimique des plaques seniles (histochemical study of
had correctly expected a change of sign on theoretical grounds. senile plaques). J. Belg. Neurol. Psychiatry 9, 643–657.
They mistakenly wrote that ‘the refractive index of the strongly Divry, P., Florkin, M., 1927. Sur les proprietes optiques de l’amyloid (on the optical
properties of amyloid). Comp. Rend. Soc. Biol. 97, 1808–1810.
absorbed ray should become negative on the shortwave side of Francis, R.J., 1973. Amyloid: its nature and histological demonstration. In: Cook, H.C.
absorption bands’. They meant retardance, because a negative (Ed.), Histopathology: Selected Topics. Baillière Tindall, London, pp. 1–28.
refractive index is impossible. Goldstein, D.J., 1969. Detection of dichroism with the microscope. J. Microsc. 89,
19–36.
Perutz and Mitchison were working with a material with more Griesbach, H., 1886. Weitere Untersuchungen über Azofarbstoffe behufs Tinction
complicated optical properties than Congo red, which has only one menschlicher und thierischer Gewebe (further investigations of azo dyestuffs as
peak of absorption in the visible spectrum (Fig. 4). They ignored the a stain of human and animal tissues). Z. Wiss. Mikrosk. 3, 358–385.
Hallimond, A.F., 1970. The Polarizing Microscope, 3rd ed. Vickers, York, pp. 87–88,
effect of absorption which would interact with birefringence to
129–133, 258–259.
influence the colours, and did not measure the optical density Hartshorne, N.H., Stuart, A., 1970. Crystals and the Polarising Microscope, 4th ed.
curve of haemoglobin between crossed polariser and analyser. Edward Arnold, London, pp. 159–161, 275–282, 426–427.
Heller, H., Missmahl, H.P., Sohar, E., Gafni, J., 1964. Amyloidosis: its differentiation
Their explanation of the colours was accordingly incomplete.
into peri-reticulin and peri-collagen types. J. Path. Bact. 88, 15–34.
The word anomalous has been used in the literature on Congo Horobin, R.W., Kiernan, J.A., 2002. Conn’s Biological Stains, 10th ed. BIOS Scientific,
red, but until the work of Howie et al. (2008) this was not defined Oxford, pp. 132–134.
or explained. Missmahl and Hartwig (1953) attributed the green Howie, A.J., Brewer, D.B., Howell, D., Jones, A.P., 2008. Physical basis of colors seen in
Congo red-stained amyloid in polarized light. Lab. Invest. 88, 232–242.
appearance to ‘eine anomale Absorptionsfarbe’ (an anomalous Ladewig, P., 1945. Double-refringence of the amyloid-Congo-red-complex in his-
absorption colour). Missmahl (1968) described ‘an intensive green tological sections. Nature 156, 81–82.
Missmahl, H.P., 1957. Polarisationsoptischer Beitrag zur Kongorotfärbung des Romhányi, G., 1949. Über die submikroskopische Struktur des Amyloids (on the
Amyloid (a polarising optical contribution to the Congo red staining of amy- submicroscopic structure of amyloid). Schweiz. Z. Pathol. 12, 253–262.
loid). Z. Wiss. Mikrosk. 63, 133–139. Romhányi, G., 1956. Zur Frage der submikroskopischen Struktur des Amyloids (on
Missmahl, H.P., 1962. Die Beeinflussung des polarisierten Lichtes durch die Amy- the question of the submicroscopic structure of amyloid). Z. Pathol. Anat. 95,
loidsubstanz (the influence of amyloid on polarised light). Ann. Histochim. 130–138.
(Suppl. 2), 225–234. Romhányi, G., 1971. Selective differentiation between amyloid and connective
Missmahl, H.P., 1968. Reticulin and collagen as important factors for the localization tissue structures based on the collagen specific topo-optical staining reaction
of amyloid. The use of polarization microscopy as a tool in the detection of the with Congo red. Virchows Arch. A Pathol. Anat. 354, 209–222.
composition of amyloid. In: Mandema, E., Ruinen, L., Scholten, J.H., Cohen, A.S. Rowe, F.M. (Ed.), 1924. Colour Index. 1st ed. Society of Dyers and Colourists,
(Eds.), Amyloidosis: Proceedings of the Symposium on Amyloidosis, University Bradford, p. 93.
of Groningen, The Netherlands, September 24–28, 1967. Excerpta Medica Schwartz, P., 1970. Amyloidosis: Cause and Manifestation of Senile Deterioration.
Foundation, Amsterdam, pp. 22–33. Charles C. Thomas, Springfield, Illinois, pp. 10–14, 26.
Missmahl, H.P., Hartwig, M., 1953. Polarisationsoptische Untersuchungen an der Steensma, D.P., 2001. ‘Congo’ red: out of Africa? Arch. Pathol. Lab. Med. 125,
Amyloidsubstanz (polarising optical investigations of amyloid). Virchows Arch. 250–252.
324, 489–508. Taylor, D.L., Allen, R.D., Benditt, E.P., 1974. Determination of the polarization optical
Mitchell, P., 1950. Crystallization of Congo red. Nature 165, 772–773. properties of the amyloid-Congo red complex by phase modulation microspec-
Moxon, J.R.L., Renshaw, A.R., 1990. The simultaneous measurement of optical activity trophotometry. J. Histochem. Cytochem. 22, 1105–1112.
and circular dichroism in birefringent linearly dichroic crystal sections. 1. Intro- Today’s Tests, 1968. Tests for amyloidosis. Br. Med. J. 4 (5630), 564–565.
duction and description of the method. J. Phys. Condens. Mat. 2, 6807–6836. Virchow, R., 1854. Zur Cellulose-Frage (on the cellulose question). Virchows Arch.
Neubert, H., 1925. Über Doppelbrechung und Dichroismus gefärbter Gele (on the Pathol. Anat. 6, 416–426.
birefringence and dichroism of coloured gels). Kolloidchem. Beiheft Supp. 20, Vogel, T., Schleiden, M.J., 1839. Ueber das Amyloid, eine neue Pflanzensubstanz (on
244–272. amyloid, a new plant material). Ann. Phys. Chem. 56, 327–329.
Perutz, M.F., Mitchison, J.M., 1950. State of haemoglobin in sickle-cell anaemia. Wälchli, O., 1945. Beitrag zur Kenntnis des Aufbaues der Zellulose (a contribution to
Nature 166, 677–679. the knowledge of the structure of cellulose). Schweiz. Arch. Wiss. Technik. 11,
Puchtler, H., Sweat, F., 1966. A review of early concepts of amyloid in context with 129–138, 181–189, 216–222, 241–246.
contemporary chemical literature from 1839 to 1859. J. Histochem. Cytochem. Wolman, M., Bubis, J.J., 1965. The cause of the green polarization color of amyloid
14, 123–134. stained with Congo red. Histochemistry 4, 351–356.
Puchtler, H., Sweat, F., Levine, M., 1962. On the binding of Congo red by amyloid. J. Zocher, H., 1925. Über die optische Anisotropie selektiv absorbierende Stoffe und
Histochem. Cytochem. 10, 355–364. über mechanische Erzeugund von Anisotropie (on the optical anisotropy of
Rigby, H.S., Warren, B.F., Diamond, J., Carter, C., Bradfield, J.W., 1991. Colour selectively absorbing materials and on the mechanical production of aniso-
perception in pathologists: the Farnsworth-Munsell 100-hue test. J. Clin. Pathol. tropy). Naturwissenschaften 50, 1015–1021.
44, 745–748. Zocher, H., Jacoby, F.C., 1927. Über die optische Anisotropie selektiv absorbierender
Romhányi, G., 1943. Über die submikroskopische Struktur des Amyloids (on the Farbstoff (on the optical anisotropy of selectively absorbing dyes). Kolloidchem.
submicroscopic structure of amyloid). Z. Path. Anat. 80, 411. Beiheft 24, 365–417.

HOWIE

Uploaded by

Copyright:

Available Formats

HOWIE

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

HOWIE

Uploaded by

Copyright:

Available Formats

Micron 40 (2009) 285–301

Contents lists available at ScienceDirect

Optical properties of amyloid stained by Congo red: History and mechanisms

1. Introduction: amyloid and Congo red . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

4. Birefringence and retardance

Another term like dichroism that is frequently mentioned in the

thickness of microscopic sections increases. Retardance equals the

Linearly polarised light can be considered a mixture of two

measurements to support this claim are not easily found in the

6. The colours seen in Congo red-stained amyloid and

By far the main interest in optical properties of amyloid has

Congo, ont également la propriété de mettre en relief la

be between 700 and 800 nm, when violet is retarded by

example, dichroic preparations with two absorption maxima.

rotation of elliptically polarised light. Just as the unaided human

You might also like