Haematoxylin and Counterstains
Anthony E Woods

Introduction
Haematoxylin is the most widely used natural dye in histotechnology. It will stain tissue components such as myelin, elastic and collagenic fibres, muscle striations, mitochondria and so on, but its most common application is as a nuclear dye in the standard H&E, the primary staining method for tissue section analysis.

Haematoxylin is obtained from the logwood, Haematoxylon campechianum, a tree of the order Leguminosae (Genus Eucaesalpinieae) and so named because of the reddish colour of its heartwood (from the Greek haemato, blood, and xylo, wood) and young leaves.1 There are several varieties of Haematoxylon but H. campechianum gives the best coloured wood. The tree, which can grow up to 15 metres, is indigenous to Central America (particularly the Campeche and Yucatan regions of Mexico) but is grown commercially in the West Indies. The heartwood is very hard and heavy and may vary from dark orange to purplish red.2 The crude product is obtained from the chipped heartwood by hot water or steam, purified by ether extraction, dried and recrystallized from water.3-6 Alternatively, the aqueous extract is precipitated from solution with urea.7 The crude logwood product also contains tannins, resins, quercetin and a small amount of volatile oil.2,8 Haematoxylin can be produced synthetically9 but the natural product is still the agent in common use.
Extracts of logwood were originally taken to Europe (by the Spanish) after the invasion of Mexico by Cortes in 1520, although it was used as a dye in the Americas before this time. The Spanish also recognized that the colour of the dye could be altered by mixing with other agents including iron alum (black), potash alum (blue) and salts of tin (red).10 The dye was being used in England by 158011 and is mentioned by Robert Hooke12 as a colouring agent for fluids. Waldeyer13 is frequently credited with having introduced haematoxylin as a biological stain,14-17 using it unsuccessfully to stain neuronal axons. However, mention of haematoxylin as a specimen dye is made in Quekett's 1848 reference on microtechnique and it is possible that Reichel used it as early as 175818 In these first attempts, however, haematoxylin was used only as a direct dye and the results were poor.
Indeed, the potential for haematoxylin in histology may never have been realized had not Böhmer,19 inspired by the textile industry, combined a metallic mordant with haematoxylin to stain tissues adequately. Although the formulation that Böhmer proposed is no longer in use, coupling haematoxylin with a mordant is still the major form in which the dye is used. After Böhmer's discovery the histological use of haematoxylin evolved rapidly, eventually overtaking carmine as the stain of choice for standard preparations. Numerous formulations using various metallic mordants were devised, many of which-although modified-are still in use. The more enduring examples include the methods of Delafield,14 Ehrlich,20 Mayer21 and Harris,22 which were mordanted with aluminium; those of Weigert,23 Weil,24 Verhoeff25 and M. Heidenhain,26 which use iron as the mordant; and that of Mallory,27 which mordants with tungsten.

Structure and characteristics
Some of the physicochemical properties of haematoxylin are given in Table 1. In the pure form haematoxylin is a colourless or slightly beige powder which gives a colourless solution in water.28 Commercially available haematoxylin is rarely this pure,29-30 although some manufacturers do specify a minimum dye content, which may be as high as 95%.

The structural formula for haematoxylin (Fig.1) was originally described in the early 1900s.31-32 Later studies have largely confirmed this view,5,9,33 and also indicate that the molecule may not be planar.34-35 Pure aqueous haematoxylin cannot stain. The active staining agent is not the principal substance but rather an oxidation product, haematein.

Formation of haematein
Haematein can be produced naturally through exposure to air and sunlight or UV light (a process referred to as 'ripening') although the reaction may take three months or longer. Examples of naturally ripened preparations include Delafield's and Ehrlich's.
The use of a chemical oxidant significantly quickens the process but does carry a risk of developing ineffective reaction products, and may also produce a solution with a shorter shelf life than the naturally ripened dye. Various agents can be used as oxidants,33,36 including potassium permanganate, iodine, sodium iodate, sodium periodate, potassium periodate, hydrogen peroxide or mercuric oxide. Of the more common formulations, Mayer's and Ehrlich's utilize sodium iodate, whereas Harris' preparation relies upon both vigorous boiling and addition of mercuric oxide to generate haematein. Although chemical oxidation can produce a working dye virtually immediately, it must be performed carefully to avoid over-oxidizing the haematoxylin. Several substances are produced as a consequence of excessive oxidation, although the most significant is a quinone-carboxylic acid (Fig.2), referred to as oxyhaematein.28,33 Using the exact amount of oxidant necessary will generally prevent over-oxidation, but it should also be noted that oxidation does not end immediately haematein is formed but continues with exposure of the solution to air. In addition, although some argue to the contrary,29 haematoxylin powder may oxidize during prolonged storage,30,33,37 and the use of such material as though it were fully reduced will unavoidably give rise to over-oxidation products.
One means of avoiding this problem has been to use half the normally required amount of chemical oxidant.17 The 'half-oxidized' solution produced has a greatly extended shelf life since unoxidized haematoxylin forms a reservoir which, through natural ripening, gradually replaces haematein that is over-oxidized or exhausted by staining.
The rate of oxidation is also influenced by the nature of the solvent. Haematein forms rapidly in alkaline solutions but slowly in an acidic environment. Adding alcohol and glycerol will further slow oxidation;38 alcohol also acts as a preservative.39

ASSESSING HAEMATEIN CONTENT
The activity of the final staining solution can be assessed by several means, including refractometry,40 thin-layer chromatography30 and spectroscopy29,41 although using the solution to stain a tissue section is probably the most immediately indicative test. Spectrophotometric analysis of a freshly pre-pared, aqueous solution of pure haematoxylin shows a maximum absorp-tion at about 290 nm. As the solution ages, two peaks, one at 445 nm and the other at 560 nm, begin to appear. With further oxidation there is a reduction in the height of these two bands, with that at 560 nm often diminishing more markedly. Over time these peaks continue to decrease in intensity and may disappear altogether, while a new broad band develops at about 390 nm.
The actual spectra obtained from spectrophotometric analysis depend upon the pH and composition of the solutions tested. Haematein becomes increasingly unstable below pH 5 and undergoes irreversible over-oxidation above pH 10. Within this range the absorbance at 445 nm steadily decreases with increasing pH (alkaline solutions develop a bluish tint), while the band at 560 nm gradually rises.42 The isosbestic point is at 486 nm.35 The absorption bands at 445 nm and 560 nm probably represent the neutral and monoanion forms of haematein, respectively, as these exist in equilibrium at low concentration. The appearance of the band at 390 nm indicates the development of over-oxidation products.35 The addition of aluminium (at a constant pH), commonly used as a mordant for haematein, has the effect of increasing the absorbance and hence darkening the solutions.43 In this case haematein is indicated by a band at 430 nm, while a 560 nm peak represents the absorption maximum for the alum haematein complexes.35,43 Palmer & Lillie28 noticed that the ratio between the absorbance peaks at 560 and 430 nm rises to over 3.0 during oxidation and remains between 3 and 4 throughout the effective life of the stain. Solution decay was signalled by a fall in the A560/A430 ratio to as little as 0.5. The use of haematein itself as the primary stain to avoid many of the problems associated with its generation from haematoxylin44 has not been successful largely because of the poor solubility of haematein in water and ethanol. The purity of commercially available haematein also varies significantly,29,39,45 precluding its use on a regular basis.

Haematoxylin as a stain
As described earlier, haematoxylin cannot stain but must be oxidized to haematein (usually at an acid pH) which acts as the dye. (Despite this, the staining solution formed is still traditionally referred to as haematoxylin.) However, even at this stage except for a few applications, direct staining is unsuccessful and it is necessary to include a mordant for haematoxylin to stain tissues effectively. The combination of mordant and dye is known as a 'lake' and in the case of haematoxylin-mordant such lakes are often positively charged, behaving as cationic dyes at low pH. Various metal salts have been used as mordants with haematoxylin, but only those containing aluminium, iron or tungsten are still in common use.

Aluminium haematoxylin
Complexes between aluminium and haematoxylin are called haemalums or alum haematoxylins. The aluminium is usually provided by adding ammonium or potassium aluminium sulphate, but sodium alum can also be used. Aluminium salts are not oxidants and therefore in preparing alum haematoxylin it is necessary to include an oxidizer (or use a naturally ripened solution of haematoxylin). The molecular configurations of the dye-mordant (D-M) structures which occur have not been completely resolved. There is general acceptance that 1:1 (D:M) cationic chelates46 do form in the dye solution, which also contains free haematein and aluminium ions. Additionally, there is evidence that a 1:2 (D:M) cationic species41,43 develops in the presence of excess metal ions rather than the 2:1 (D:M) possibly anionic34 chelate traditionally advanced.46-47
Recently, Bettinger & Zimmerman41 have suggested that several cationic dye complexes existing at low pH contribute to the reddish colour typical of acidified alum haematoxylin solutions. With mild alkaline treatment these complexes are converted into a neutral chelate which gives a blue colour. These structures can be continually regenerated by reversing the acid-alkaline condition, with the solution colour successively changing between red and blue. When used to stain tissues, cationic D-M complexes are attracted to negatively charged sites, displaying a particular affinity for polyphosphates.48 In the case of alum haematoxylin complexes, the major tissue binding site is thought to be the phosphoric acid residue in nucleic acid (nuclear DNA/cytoplasmic and nucleolar RNA) with the linkage, at least initially, being electrostatic and occurring through the aluminium ion (Fig.3).41,43,49-50 Fig 3 demonstrates the probable binding of alum haematoxylin (in acid solution) via aluminium ion to phosphoric aacid residue in nucleic acids (after Bettinger and Zimmerman41 and Baker49) The pH of the staining solution also influences this reaction and glacial acetic acid can be added (following the practice of Ehrlich) to enhance nuclear staining.17,38-39 This occurs because the pH of such acidified solutions (around pH 2.0-3.0) is higher than the isoelectric point of nucleic acids (pH 15-2.0) which will thus express a negative charge` and more readily attract and bind the cationic dye lake.

The bond formed between alum haematoxylin and tissues is unusual in that it cannot be broken by ethanol, as normally occurs with cationic dyes. This may be a function of differential solubility in alcohol,47 but it is also thought that the stability of the bond is due to a change from an ionic to covalent linkage.41,49,52 Under the appropriate conditions (high haematein concentration, pH not too low), an alum haematoxylin solution will stain cationic proteins such as the histones associated with nucleic acid.34,43,53-54 In this case it is likely that anionic forms of free haematein and alum haematoxylin are responsible for the reaction.

PROGRESSIVE AND REGRESSIVE STAINING
Alum haematoxylin solutions can be used as either progressive or regressive stains. Solutions such as Mayer's, Harris' and Gill's, in which the mordant to dye ratio is between 8 and 16 to 1 (referred to as the critical quotient ratio)` or greater (and provided that the pH is around 2.5) will in principle tend not to overstain and can be used progressively (although the strong concentration of haematein used in Harris' haematoxylin does lead to overstaining). Maintaining a high mordant content has the further advantage of favouring nuclear-only staining.54 Progressive reactions provide a mechanism for highly selective staining but often require extended reaction times to maximize dye-tissue binding and produce sections of high contrast.
On the other hand, as regressive techniques are designed to overstain, solutions can be prepared so that only brief staining is required. For example, if the mordant quotient is kept below the critical level (by using relatively more dye) the solution will not only stain nuclei but also a range of tissue components including cationic proteins and some mucosubstances. Preparations such as Ehrlich's, which fall into this category, therefore require differentiation to achieve acceptable nuclear staining. Removal of excess dye needs to be carefully controlled, frequently microscopically, but even with this step the regressive approach often takes relatively less time and is generally favoured for this reason.
Differentiation can be achieved in one of two ways. Placing the section in a solution containing aluminium ions in great excess will have the effect of drawing dye away from the tissue-bound aluminium and redistributing it into the differentiating solution.49 This process of mordant differentiation will eventually remove all dye from the tissue. Consequently, decolourization needs to be controlled (microscopically) and at the appropriate level slides must be washed in water to halt the process.
More commonly, sections are differentiated using a weak acid solution (normally prepared in 70% ethanol to improve control over the differentiation). The acid competes for anionic sites breaking the tissue-mordant bond.43 Although microscopic control is again necessary, acid differentiation is generally quicker, particularly if large numbers of slides are being stained at the same time.
In both types of differentiation stain is removed from all tissue components at more or less the same rate. Structures which bound more dye initially will retain more and remain coloured for longer. Differentiation is discontinued when only the structures of interest (usually nuclei) are still clearly visible. Over-differentiation can be corrected by washing out the differentiator and returning the sections to the haematoxylin solution. The use of acid as the differentiator also favours the reddish form of haematein, both soluble and bound. Although the dye complexes are stable in this form it has become standard practice to convert haematein to the blue complex, as this contrasts more effectively with various anionic counterstains. This process of 'blueing' requires that sections be washed in a solution of pH>5.0.43 In most cases ordinary tap water will suffice, but may take a few minutes. Virtually instantaneous conversion can be obtained with a weak ammonia solution, a dilute solution of lithium carbonate or Scott's tap water substitute.55
Alum haematoxylin solutions have become the standard, universal means of staining cell nuclei for microscopic examination. Practically every section of normal and diseased tissue will be examined and presumptively identified using an alum haematoxylin to colour nuclei. The major disadvantage of alum haematoxylin as a stain is its susceptibility to acids, which limits the range of counterstains that can be used. Alum haematoxylin staining is also influenced by other factors, including the concentration and age of staining solutions as well as the fixation and processing to which the tissue was subject. Nevertheless, staining with alum haematoxylin is an immensely versatile procedure and, with experience, results that are consistent in intensity and effect can be achieved.

Haematoxylin And Eosin Staining Of Paraffin Sections
SECTION PREPARATION
Most standard tissue fixatives are suitable. Cut sections at 3-4 µm.
REAGENTS REQUIRED
1 Ehrlich's haematoxylin (after Ehrlich20)
haematoxylin (Cl 75290) 6 g
absolute alcohol 300 ml
distilled water 300 ml
glycerol 300 ml
glacial acetic acid 30 ml
ammonium or potassium aluminium sulphate 30 g
sodium iodate 0.9 g
Dissolve the haematoxylin in the alcohol before adding the other ingredients in the order given. Then mix the solution overnight. The addition of sodium iodate artificially ripens the haematoxylin so that it may be used immediately. Alternatively, sodium iodate can be deleted and the mixture ripened by exposure to warmth and sunlight for approximately two months. The naturally ripened form has a longer shelf life.

Harris Alum Haematoxylin (after Harris22)
haematoxylin (Cl 75290) 2.5 g
absolute alcohol 25 ml
distilled water 500 ml
ammonium [*] or potassium alum 50 gm
mercuric oxide (yellow) 1.25 g
Dissolve the haematoxylin in the absolute alcohol. Dissolve the alum in water, using heat if necessary. Mix the two solutions, rapidly bring to the boil and carefully add the mercuric oxide a little at a time. Cool rapidly by immersing the flask into iced water. The addition of 20 ml of glacial acetic acid is optional but gives sharper nuclear staining; this must be added just before use and the stain filtered. The solution is ready for staining as soon as it is cool.

Or

Gill's haematoxylin17
Distilled water 730 ml
Ethylene glycol 250 ml
Haematoxylin (CI 75290) 2 gm
Sodium iodate 0.2 gm
aluminium sulphate 17.6 g
glacial acetic acid 20 ml
Combine the reagents in the order given and mix for 1 hour at room temperature. The stain can be used immediately.

Or

Carazzi's haematoxylin (after Carazzi56)
haematoxylin (Cl 75290) 0.5 g
potassium iodate 0.01 g
potassium alum 25 g
glycerol 100 ml
distilled water 400 ml
Combine haematoxylin and glycerol Dissolve the potassium iodate in a little of the water and prepare the alum using the remainder. Mix the haematoxylin and alum solutions and then carefully add the potassium iodate.

2. Differentiator
1% hydrochloric acid in 70% alcohol

3. Alcoholic eosin solution
95% alcohol 3.91
eosin Y (CI 45380) 5 g
phloxine B (CI 45410) 0.5 g
glacial acetic acid 20 ml

4. Ammonia water
0.04% aqueous ammonia

METHOD
1 . Dewax and re-hydrate sections.
2. Place in haematoxylin solution for 5-10 minutes.
3. Wash sections in running water.
4. Differentiate sections in 1 % acid-alcohol and then wash well in water. Repeat if more stain needs to be removed. This step requires microscopic control to ensure that only nuclei are stained.
5. Rinse ('blue') in ammonia water (or similar) for 1 minute.
6. Rinse sections briefly in distilled water.
7. Counterstain in eosin for 2-5 minutes.
8. Wash well in water.
9. Dehydrate, clear and mount in neutral mounting medium.

RESULTS
Nuclei-blue
Cytoplasm, red blood cells and connective tissue-shades of pink

Iron haematoxylin
Solutions of haematoxylin which incorporate iron as the mordant are effective not only as nuclear stains but will also demonstrate many other structures, including muscle striations, various cellular inclusions and organelles (particularly mitochondria), myelin, keratin, some mucosubstances and elastic fibres. No single iron haematoxylin method will display all of these com-ponents and most (regressive) procedures require selective differentiation to identify specific structures. Since the dye lake is normally dark blue to black, iron haematoxylin techniques applied appropriately can provide staining of exceptional contrast which is particularly suited to photomicrography. Ferric ions in the form of ferric chloride or iron alum (ferric ammonium sulphate) are normally used as the mordant, although staining is also effective with ferrous salts.57 The interaction between the ferric ions and haematoxylin is not fully understood. It is clear, however, that ferric ions will act as oxidant as well as mordant, and consequently iron haematoxylin solutions, once prepared, rapidly deteriorate as the oxyhaernatein content increases. The effective staining life of these solutions is thus very short. It is also necessary to keep iron haematoxylin solutions strongly acidic both to enhance nuclear staining (see previous discussion) and to prevent precipitation of the dye lake.38,58
The strategies for preparing and applying iron haematoxylin solutions to avoid or minimize these difficulties have changed little since their original development and can be divided into two broad categories, sequence and combined.

SEQUENCE TECHNIQUES
With this approach sections are treated initially in the ferric salt solution, washed and then transferred to the haematoxylin bath. Partially ripened haematoxylin can be used, but it is not essential as tissue-bound ferric ions will oxidize haematoxylin and generate haematein. Separating iron and haematoxylin solutions ensures that the haematoxylin has a long shelf life: however, these two-stage procedures are somewhat lengthy. The bond between iron haematoxylin and tissues is apparently more stable than that formed with alum haematoxylin, particularly in terms of resistance to acid attack.52 As with alum haematoxylin, nucleic acids appear to be a major binding site for ferric ions, with phosphoric acid groups being the likely receptor.59 Proteins, including nucleoproteins, are also thought to be target sites binding to either lysine and arginine residues` or other points carrying suitably spaced carboxyl and hydroxyl groups.34,59 The most notable sequence iron haematoxylin method is that of Martin Heidenhain,26 I which is a modification of the afterchroming procedure (in which the stain is applied before the mordant) pioneered by his father, Rudolf Heidenhain.60 Sections are jet black when removed from the staining solution and successive brief washes in the differentiator (a ferric salt solution) are necessary to demonstrate individual tissue features (mordant differentiation). Colour is lost because metallic ions in solution draw haematein away from the bound form. Over-differentiation can be corrected simply by restaining with haematoxylin, since the bound iron is still retained by the tissue.34
In practice, differentiating Heidenhain's haematoxylin can be difficult because the rate of colour loss tends to accelerate with time. Also, even with careful control the final staining is often quite patchy (Fig.4). The basis for this mottling is not clear, but it may relate to permeability effects,61 the level of dye bound initially, or the number of bonds between the dye and the tissues.59

Heidenhain's Iron Haematoxylin (after Heidenhain26)
SECTION PREPARATION
Adequate results can be obtained with tissues fixed in formalin, Susa's fixative, Bouin's fluid and usually with chrome fixatives such as Zenker's, but these may require longer times. Cut sections at 3-4 µm.

REAGENTS REQUIRED
1. Iron alum solution
ferric ammonium sulphate (violet crystals) 5 g
distilled water 100 ml

2. Haematoxylin solution
haematoxylin (CI 75290) 0.5 g
absolute alcohol 10 ml
distilled water 90 ml
This solution must be allowed to ripen for 4-5 weeks.

METHOD
1. Dewax and re-hydrate sections.
2. Transfer to the iron alum solution for 30-35 minutes at 56oC or at room temperature for 12-24 hours.
3. Rinse rapidly in water.
4. Transfer to the haematoxylin solution and leave for 30-35 minutes at 56oC or at room temperature for 12-24 hours.
5. Rinse rapidly in water.
6. Differentiate in 5% iron alum.
7. Wash the slide in running water for 5 minutes.
8. Counterstain if required.
9. Dehydrate, clear and mount.

RESULTS
The following tissue elements may be demonstrated in the order given, depending upon the degree of differentiation employed:
mitochondria, cross-striations of muscle fibres, cytoplasm, nuclear membrane, yolk, chromosomes, chromatin, nucleoli and centrioles - black

ALUM HAEMATOXYLIN - CELESTIN BLUE62
This is a sequential procedure which uses an alum haematoxylin but renders it acid resistant. Sections are treated first with a solution of celestin blue B (an oxazine dye) and iron alum (ferric ammonium sulphate). After rinsing in water, sections are stained in alum haematoxylin (Mayer's haemalum is effective) which combines with the tissue-bound celestin blue-ferric alum to produce a celestin blue-ferric alum-haemalum complex that is highly stable in the presence of acid.58 This method is used frequently to substitute for iron haematoxylin as a nuclear stain in trichrome procedures (which use acidic counterstains) since differentiation is not required and the ingredients have a relatively longer shelf life.

Mayer's Haemalum - Celestin Blue (after Lendrum & WFarlane62)
SECTION PREPARATION
Formalin fixation is suitable. Cut sections at 3-4 pm.
REAGENTS REQUIRED
1. Celestin blue solution
iron alum 2.5 g
distilled water 50 ml
celestin blue (CI 51050) 0.25 g
glycerin 7ml
Dissolve iron alum in distilled water overnight at room temperature and then add celestin blue. Boil for 3 minutes, filter when cool and add glycerin.

2. Alum haematoxylin (Mayer's is given)
haematoxylin (Cl 75290) 1.0 g
distilled water 1.0 l
potassium or ammonium alum 50,0 g
citric acid 1.0 g
chloral hydrate 50,0 g
sodium iodate 0.2 g
The haematoxylin, potassium or sodium alum and sodium iodate are dissolved in distilled water. The chloral hydrate and citric acid are added and the mixture boiled for 5 minutes, cooled and filtered. The stain is ready for immediate use.

3. Differentiator
0.25% hydrochloric acid in 70% alcohol

METHOD
1. Dewax and re-hydrate sections.
2. Stain nuclei in celestin blue solution for 10 minutes.
3. Rinse in water.
4. Stain with alum haematoxylin for 10 minutes.
5. Rinse in water.
6. Apply counterstain as required.

RESULTS
Nuclei- black

COMBINED (SINGLE SOLUTION) TECHNIQUES
Combined techniques are those in which mordant and haematoxylin are prepared as separate solutions and mixed immediately before use. It is important that correct quantities be used in preparation, as these working solutions are inherently unstable and their effective life is quite brief. Common examples of combined iron haematoxylin methods include those of Weigert and Weil. Weigert's iron haematoxylin is often used as a nuclear stain when applying weakly acidic counterstains (such as van Gieson's stain63). Nevertheless, the iron haematoxylin - tissue link which forms is not absolutely stable and excess bound dye can be removed with a suitable acid differentiator. Weigert's haematoxylin can also be used progressively as a nuclear stain, provided that sufficient acid or ferric ions are added to ensure selectivity.

Weigert's Haematoxylin and van Gieson's Stain
SECTION PREPARATION
Formalin fixation is suitable. Cut sections at 3-4 µm.
REAGENTS REQUIRED
1. Weigert's iron haematoxylin
Solution A:
haematoxylin (CI 75290) 1 g
absolute alcohol 100 ml

Solution B:
30% aqueous ferric chloride 4ml
distilled water 100 ml
hydrochloric acid 1 ml
Mix equal parts of solutions A and B immediately before use.

2. van Gieson's stain
saturated aqueous picric acid 100 ml
1% acid fuchsin (CI 42685) 10 ml
The above quantities may be prepared and kept as a stock solution, but it is better to use a freshly prepared solution containing 5 ml of saturated aqueous picric acid and 0.75 ml of 1 % acid fuchsin, which gives more precise and sharp staining.

3. Differentiator
1 % hydrochloric acid in 70% alcohol

METHOD
1 .Dewax and re-hydrate sections.
2. Stain with Weigert's haematoxylin for 40 minutes.
3. Wash in water.
4. Differentiate in 1% acid-alcohol.
5. Rinse ('blue') in ammonia water or similar and then rinse in distilled water.
6. Cover the section with van Gieson's stain, and leave for 45 seconds to one minute. Flick off the stain.
7. Air-dry.
8. Clear and mount.

RESULTS
Nuclei - blue black
Collagen - bright red
Cytoplasm, red blood cells - yellow

Weil's stain will demonstrate myelin with the likely binding site being phospholipid.53 It is a regressive technique which uses two differentiation steps. The first, mordant differentiation, removes the bulk of the stain. After washing in water sections are placed in a borax-ferricyanide solution which oxidizes bound dye to a colourless end product (oxidant differentiation). This second stage must be carefully monitored to ensure that the myelin is left with adequate colour.

VERHOEFFS HAEMATOXYLIN
Verhoeff's technique is a single-solution iron haematoxylin to which iodine has been added. It was developed primarily as a stain for elastic fibres (Fig.5) but will also demonstrate nuclei and myelin14 and can be adapted for electron microscopy.65-67
The technique is regressive (using ferric chloride as the differentiator) and, unless carried out carefully, stain can be removed from the smaller fibres. Sections can be repeatedly stained and differentiated if necessary, which means that results tend to be variable and highly subjective. The role of iodine is unclear. It has been thought to act in converting haematoxylin to haematein, but since the iodine appears to prolong the staining life of the solution it is more likely to have an anti-oxidant effect.64 The iodine may also function as a trapping agent for the dye lake,68 possibly through complex formation with'the metallic component. The staining mechanism is probably by way of hydrogen bonding between hydroxyl groups in the dye and carbonyl groups in the tissue, although specific reactive sites have not been identified.

Verhoeff's Method25
SECTION PREPARATION
Most fixatives are suitable. Cut sections at 3-4 µm.

REAGENTS REQUIRED
1. Verhoeff's haematoxylin
5% alcoholic haematoxylin (5 g haematoxylin, CI 75290, in 100 ml absolute alcohol 30 ml
10% ferric chloride 12 ml
Verhoeff's iodine 12 ml
(Prepared by mixing (in the order given) 2 g of iodine and 4 9 of potassium iodide in 100 ml of distilled water.)

2. Lugol's iodine
Prepared by mixing 1 g of iodine and 2 g of potassium iodide in 100 ml of distilled water.

3. 5% sodium thiosulphate

4. Differentiator
2% ferric chloride

METHOD
1. Dewax and re-hydrate sections.
2. Cover the section with Lugol's iodine for 5 minutes.
3. Wash in water.
4. Treat the section in sodium thiosulphate for 2 minutes.
5. Wash in water.
6. Stain in Verhoeff's haematoxylin solution for 15-30 minutes (until the sections are jet black).
7. Wash in water.
8. Differentiate in ferric chloride (if too much stain is removed sections may be replaced in haematoxylin solution).
9. Wash in water and then in 95% alcohol to remove the iodine staining.
10. Wash in water for 1 minute.
11. Counterstain if required (van Gieson's stain for 45 seconds to 1 minute is recommended).
12. Air-dry (if van Gieson's stain is used).
13. Clear and mount.

RESULTS
Elastic fibres and nuclei - black to blue-black
Cytoplasm and muscle - yellow
Collagen - red

Tungsten haematoxylin
Solutions of haematoxylin mordanted with tungsten are quite unique in that the single solution gives two staining colours (reddish-brown and blue). The most widely used preparation is that developed by Mallory.27,69
Although originally intended for neuroglial fibres, a range of other structures can also be stained by Mallory's method. These include nuclei, centrioles, mitochondria, fibrin, red blood cells (RBC), cardiac and skeletal muscle striations, myelin and some microorganisms, all of which stain blue, while collagen, reticular and elastic fibres, cartilage and bone matrix appear reddish-brown.39,70-71 The most useful application for the procedure in histopathological diagnosis is in detecting striational changes in muscle disease and neuroglial (astrocytic) changes in lesions of the CNS.72 A major advantage of the technique over iron haematoxylin in staining muscle striations is that, being progressive, it requires no differentiation and thus determination of the staining end-point is not entirely arbitrary. The tungsten is usually in the form of phosphotungstic acid (hence the common abbreviation PTAH-phosphotungstic acid haematoxylin). This is not an oxidant and staining solutions only become active through natural ripening of the haematoxylin. Alternatively, haematein or chemically oxidized haematoxylin, which provide solutions of immediate but brief activity, can be used. The lake formed between haematoxylin and tungsten is thought to contain anionic 1:2 M:D complexes,73 but the basis for the two-coloured staining characteristic is not known. A metachromatic effect in which orthochromatic red colour (absorption maximum at 560 nm) gives rise to the blue (absorption maximum at 610 nm) on reaction with tissues has been suggested.74 Conversely, a polychromatic mechanism is proposed based on chromatographic evidence which indicates separate blue, red75 and yellow76 components in the staining solution. The mechanism of PTAH staining in tissues is also unclear, but various target sites have been implicated.74,76-77 The differential permeability of tissue structures to the dye complexes is also probably involved, as the molecular size of the red fraction appears to be larger than the blue.77
Fixation in Zenker's fluid followed by treatment of sections with potassium permanganate is usually recommended for optimal results.78 The potassium permanganate may augment dye retention in the tissues,68 but the use of Zenker's fixative is not essential.53 Staining may be enhanced by mordanting sections in iodine solutions, iron alum or saturated mercuric chloride.71

Mallory's Phosphotungstic Acid Haematoxylin (PTAH)27,69
SECTION PREPARATION
Formalin fixation is suitable. Cut sections at 3-4 µm.

REAGENTS REQUIRED
1. Haematoxylin solution
haematoxylin/haematein (Cl 75290) 1 g
phosphotungstic acid 20 g
distilled water 1 l
Dissolve the haematein and the phosphotungstic acid separately in distilled water, using gentle heat. When cool, combine the solutions and make up to 1l. If haematoxylin is used it can be ripened immediately by adding 0. 177 g of potassium permanganate or by exposing the solution to light and warmth for 5-6 weeks.

2. 2.5% potassium dichromate in 5% acetic acid

3. Lugol's iodine

4. 5% sodium thlosulphate

5. 0.25% potassium permanganate

6. 2% oxalic acid

METHOD
1 Dewax and re-hydrate sections.
2. Sensitize overnight in acetified potassium dichromate solution.
3. Wash in running water for 5 minutes.
4. Treat with Lugol's iodine for 5 minutes.
5. Wash in water.
6. Treat with sodium thiosulphate solution for 1 minute.
7. Wash in water.
8. Treat with potassium permanganate for 10 minutes.
9. Wash in water.
10. Treat with oxalic acid for 1 minute.
11. Wash in water.
12. Stain in PTAH solution at 56oC for 1-3 hours.
13. Wash in water and replace in PTAH if understained.
14. Dehydrate, clear and mount.

RESULTS (Fig.6)
Nuclei, centrioles, neuroglia, fibrin and cross-striations of muscle fibres - blue
Collagen, reticulin and bone, ground substance - yellow to red

Other mordants
There are many other metallic salts which form coloured complexes with haematoxylin.79 Several of the more notable forms, described below, are applied only infrequently, having been confined to the role of special staining solutions.

CHROMIUM
Chrome alum haematoxylin preparations stain lipoproteins,70 myelin, phospholipids and cytoplasmic granules in B cells of the anterior pituitary and pancreatic islet.53,68 The chromium-haematoxylin complex, being cationic, will also bind to nuclei at low pH, probably through phosphate groups. This association is highly stable and its dogged resistance to differentiation discourages more frequent use.

MOLYBDENUM
Complexes of molybdenum and haematoxylin are used principally to stain collagen,77,80 although original techniques were applied to neural tissue.81 By increasing the concentration and adding dioxane to the active solution reticulum, basement membranes and skeletal muscle striations are also coloured, suggesting that dye penetrability is a factor in the staining mechanism.68

COPPER
Copper haematoxylin will stain fatty acids,70 acidophils in the pituitary,82 myelin sheaths and mitochondria.68

LEAD
Applied as a sequence method, lead-haematoxylin has a particular affinity for axis cylinders, but up to 6 weeks in the mordant may be necessary for an effective result. A single-solution lead haematoxylin has been developed83 and this stains various cells of the diffuse neuroendocrine system, binding to carboxyl groups in proteins as well as nuclei, nucleoli, keratohyalin, calcium deposits, nerve fibres, smooth and striated muscle.

Haematoxylin without a mordant (direct staining)
Under appropriate conditions, free haematein is weakly anionic and will stain cationic tissue components-particularly collagen, but also elastin, erythro-cytes and contractile elements in smooth muscle-yellow to orange-brown.43 Non-mordanted, dilute solutions of haematoxylin (haematein) also colour nuclei through binding to basic nuclear proteins.84
Aqueous solutions of haematoxylin may be used to identify lead, copper or iron (haemosiderin) deposits in tissues.80 The reaction is probably due to dye-metal chelate formation and is reportedly more sensitive in the case of iron than the ferrocyanide reaction of Perls.39 However, the procedure is not particularly specific, as calcium and other metals will also form a blue-black haematoxylin lake85 giving a false positive result.
In the Clara86 haematoxylin procedure, non-mordanted haematoxylin reacts with arginine residues and stains keratin, keratohyalin, basic nucleoproteins and eosinophil granules.87 Iron pigments and enterochromaffin cell granules (which probably contain iron) also react through chelate formation.

Substitutes for haematoxylin
A number of dyes have been recommended as substitutes for haematoxylin, particularly during the early to mid-1970s when haematoxylin was in short supply. Although some closely mimic the features of haematoxylin, none will serve as a complete replacement. The more notable examples include the following.

Celestin blue B, iron alum88 This provides an effective alternative to alum haematoxylin in the routine H&E method. Staining is effective after 5-30 minutes.

Gallein89-90
Combined with ferrous sulphate and a small amount of ferric chloride, gallein will stain muscle striations, RBC, myelinated nerve fibres and some elastins. Can be used to substitute for Weigert's iron haematoxylin with van Cieson's stain, but combines poorly with eosin.

Gallocyanin - chrome alum10,88,91
This is a cationic dye which gives strong nuclear staining at low pH but requires 24-48 hours for an effective result, virtually eliminating it as a practical alternative.

Mordant blue 392
This is prepared with iron alum and used as a replacement for haematoxylin in routine H&E procedures. It has reduced value in neural tissue, as it also stains myelin.

Thionin bromide93
This is recommended for cytological preparations. Advantages over haematoxylin are that it is quick and simple to use, nuclei and cytoplasm are stained in one step and the reagent has a long shelf life.

Fluorone black94
This is a useful substitute for haematoxylin in iron haematoxylin - van Gieson procedures (but is inferior to iron gallein).

Other uses for haematoxylin
Haematoxylin will act as an acid-base indicator, a property which is probably due to haematein present in the solution. A 0.5% solution of haematoxylin in ethanol will change from red to yellow between pH 0 and 1.0, and from yellow to violet over the range pH 5.0-6.0.95
The role of haematoxylin in the textile industry has decreased significantly in recent years, but it is still used for dyeing wool fibres and some synthetics (Colour Index96). Haematoxylin is also mixed with iron salts to colour inks blue or black.5,37
Haematoxylin has been reported to have medicinal value as a mild astringent' in some intestinal disorders.97 It may also be effective in sup-pressing acute inflammation,98-99 by inhibiting histidine decarboxylase and thus modulating histamine activity on capillary permeability.100-101 Administration of haematoxylin (and haematein) may also lessen chemical or physically induced capillary damage,102 retard adrenalin breakdown103 and minimize the effects of some toxins.104 There is a report that haematoxylin may bind to DNA, leading to cell degeneration and necrosis in some tumours.105

Counterstains
In general terms, counterstains are used to gain information supplementary to that given by the primary stain. Frequently, this only amounts to demonstrating the overall morphology or histology of the tissue concerned; however, in some cases, the judicious selection of a counterstain can result in significant additional information.

Counterstains may consist of single dyes or dye mixtures and---depending upon the primary target-may colour the cell nucleus, cytoplasm or specific tissue structures (such as collagen). To be effective, a counterstain should be of a contrasting, subtle colour which does not intrude on the major stain. Some commonly used counterstains are given in Table 2.

Eosin
The ubiquitous H&E combination was proposed106 shortly after the discovery of eosin in 1871,107 although aniline blue was the first counterstain to haematoxylin.108 Originally, eosin was used alone to colour tissues,109 but its role now is almost exclusively in double and multiple staining procedures.

The most frequently used form is eosin Y (Cl 45380) which gives yellowish-pink shades (from the Greek eos, dawn) and can be prepared as either an alcoholic (2% solubility)'or aqueous (40% solubility) solution. Eosin B (Cl 45400), erythrosin B (Cl 45430) and phloxine B (Cl 45410), along with various similar xanthene dyes, can be used as suitable alternatives.39 Eosin Y (M.W. 691.9) is a tetrabrominated derivative of fluorescein with maximum absorption in water between 515 and 518 nm.110 Commercial preparations may also contain fluorescein and tribrornofluorescein in sufficient quantities to influence staining colour111 as the dye becomes paler with less bromine.110 The molecule carries one negative charge between pH 3 and 5 and two negative charges above pH 5.112

Proteins are generally cationic below pH 6 and will thus bind eosin,51 probably through the bromine groups.112 The reaction is influenced by fixation113 with tissues prepared in Zenker's fluid, in particular, staining strongly. The selectivity and strength of eosin staining can also be enhanced by adding a small amount of glacial acetic to the dye solution.7
As a counterstain for haematoxylin, aqueous eosin solutions range between 0.5% and 2% in strength (1% is most common). Alcoholic formulations generally contain less eosin. Correctly applied, eosin should stain various tissue structures shades of pink-especially collagen, cell cytoplasm and erythrocytes. The addition of a very small amount of phloxine can further improve the result (Fig.7).

References
1 Willis JC. A dictionary of the flowering plants and ferns. 6th ed. Cambridge: Cambridge University Press, 1951.
2 Wallis TE. Textbook of pharmacognosy. 5th ed. London: J&A Churchill, 1967. 3 McClung CE. Haematoxylin. Science 1923;58:515.
4 Corm HJ. The haematoxylin problem. Stain Technol 1927;2:1-3.
5 Duff DH. Logwood colouring matters. Dyer 1950;103:271-3.
6 Lalor GC, Martin SL. Studies on haematoxylin and haematein. The colouring principles of logwood. I-Absorption spectra of pure compounds in various solvents and a spectrophotometric method of analysis for haematoxylin and haematein. j Soc Dyers Colourists 1959;5:513-7.
7 Stevens A. The haematoxylins. In: Bancroft JD, Stevens A, eds. Theory and practice of histological techniques. 3rd ed. Edinburgh: Churchill Livingstone, 1990.
8 Chevreul PM. De la seconde partie du Memoire sur le Bois de Campeche et sur la nature de son principe colorant. Ann Chim 1810;82:126-47.
9 Morsingh F, Robinson R. The synthesis of diazilin and haematoxylin. Tetrahedron 1970;26:281-9.
10 Lillie RD. The hematoxylin shortage and the availability of synthetic substitutes. Am J Med Technol 1974;40:455-61.
11 Leggett WK Ancient and medieval dyes. Brooklyn: Chemical Publishing, 1944.
12 Hooke R. Micrographia. London: Royal Society, 1665. Facsimile reproduction, New York: Dover, 1961.
13 Waldeyer R. Untersuchungen über der Ursprung und den Verlauf des Axencylinders bei Wirbellosen und Wirbelthieren sowie über dessen Endverhalten in der quergestrieften Muskelfaser. Henle u. Pfeifer. Z Ration Med 1863;20:193-256.
14 Conn HJ. The history of staining. Logwood dyes. Stain Technol 1929:4;37-48.
15 Lynch MJ, Raphael SS, Mellor LD, Spare PD, Inwood MJ14. Medical laboratory technology and clinical pathology. 2nd ed. Philadelphia: WB Saunders, 1969.
16 Culling CFA. Handbook of histopathological and histochemical techniques. 3rd ed. London: Butterworth, 1974.
17 Gill GW, Frost JK, Miller KA. A new formula for a half-oxidised hematoxylin solution that neither overstains nor requires differentiation. Acta Cytol 1974;18:300-11.
18 Bracegirdle B. A history of microtechnique. New York: Cornell University Press, 1978.
19 Bbhmer F. Zur pathologischen Anatomie der Meningitis cerebromedularis epidemia. Aerztl Intelligenzb (Munchen) 1865;12:539-50.
20 Ehrlich P. Hamatoxylinlösung. Z Wiss Micr 1886;3:150.
21 Mayer P. Ueber das Färben mit Hämatoxylin. Mitt Zool Stat Neapel 1891;10:170-86.
22 Harris HR On the rapid conversion of haematoxylin into haematein in staining reactions. J Appl Microsc 1900;3:777-80.
23 Weigert K. Eine Kleine Verbesserung der Hämatoxylin-van Gieson-Methode. Z Wiss Mikr 1904;2L1-5.
24 Weil A. A rapid method for staining myelin sheaths. Arch Neurol Psychiat 1928;20:392-3.
25 Verhoeff FH. Some new staining methods of wide applicability. Including a rapid differential stain for elastic tissue. JAMA 1908;50:876-7.
26 Heidenhain M. Neue Untersuchungen über die Centralk6rper und ihre Beziehungen zum Kern-und Zellen-protoplasma. Arch Mikr Anat 1894A3:423-758.
27 Mallory FB. On certain improvements in histological technique. J Exp Med 1897;2:529-33.
28 Palmer RW, Lillie RD. Spectroscopic and staining studies of the ripening and overripening of aluminium hematoxylins. Histochemie 1965;5:44-54.
29 Imes NK Jr, Sanders DC, Crane CR, Clark G. Assaying actual hematein content of commercial hematoxylins and hemateins. Stain Technol 1969;44:167-72.
30 Marshall PN, Horobin RW. A simple assay procedure for mixtures of hematoxylin and hematein. Stain Technol 1974;49:137~42.
31 Engels P, Perkin WH Jr, Robinson R. Brazilin, haematoxylin, and their derivatives. Part IX. On brazilein, haematein and their derivatives. J Chem Soc 1908;93:1115-62.
32 Perkin WH jr, Robinson R. Brazilin and haematoxylin. Part VIII. Synthesis of brazilinic acid, the lactones of dihydrobrazilinic and dihydrohaematoxylinic acids, anhydrobrazilinic acid &c. The constitution of brazilin, haematoxylin and their derivatives. J Chem Soc 1908;93:489-517.
33 Marshall PN, Horobin RW. The oxidation products of haematoxylin and their role in biological staining. Histochem J 1972;4:493-503.
34 Puchtler H, Meloan SN, Waldrop FS. Application of current chemical concepts to metal-hematein and brazilein stains. Histochem 1986;85:353-64.
35 Bettinger CL, Zimmermann HW. New investigations on hematoxylin, hematein, and hematein-aluminium complexes. I. Spectroscopic and physio-chemical properties of hematoxylin and hematein. Histochem 1991;95:279-88.
36 Clark G. Comparison of various oxidants for alum hematoxylin. Stain Technol 1974;49:225-7.
37 Green Fj. The Sigma-Aldrich handbook of dyes, stains and indicators. Milwaukee, Wisconsin: Aldrich Chemical Company, 1990.
38 Cole EC. Studies on hematoxylin stains. Stain Technol 1943;18:125-42. 39 Lillie RD, Fullmer HM. Histopathologic technic and practical histochernistry. 4th ed. New York: McGraw-Hill, 1976.
40 Arshid FM, Desai JN, Duff DJ, Giles CH, Jain SK Macneal IR. A study of certain natural dyes. I-The absorption of brazilwood and logwood colouring matters by fibres. J Soc Dyers Colourists 1954;70:392-401.
41 Bettinger CL, Zimmermann HW. New investigations on hematoxylin, hematein, and hematein-aluminium complexes II. Hematein-aluminium complexes and hemalum staining. Histochem 1991;96:215-28.
42 Lalor CC, Martin SL. Studies on haematoxylin and haematein, the colouring principles of logwood. II-Behaviour in aqueous solutions at varying pH and the pK values. J Soc Dyers _Colourists 1959;75:517-21.
43 Baker JR. Experiments on the action of mordants. 2. Aluminium-haematein. Q J Microsc Sci 1962;103:493-517.
44 Kornhauser SI. Hematein-its advantages for general laboratory use. Stain Technol 1930;5:13-15.
45 McClure TD, Clark G. A comparison of assay values with actual staining obtained from commercial hemateins. Stain Technol 1970;45:261-3.
46 Arshid FM, Connelly RF, Desai JN, Fulton RG, Giles CH, Kefalas JC. A study of certain natural dyes. II-The structure of the metallic lakes of the brazilwood and logwood colouring matters. J Soc Dyers Colourists 1954;70:402-12.
47 Marshall PN, Horobin RW. The mechanism of action of 'mordant' dyes-a study using preformed metal complexes. Histochemie 1973;35:361-71.
48 Marshall PN, Horobin RW. Measurements of the affinities of basic and 'mordant' dyes for various tissue substrates. Histochenue 1973;36:303-12.
49 Baker JR. Experiments on the action of mordants. 1 'Single-bath' mordant dyeing. Q J Microsc Sci 1960;101:255--72.
50 Scott JE, Willett IH. Binding of cationic dyes to nucleic acids and other biological polyanions. Nature 1966;209:985-7.
51 Elias JM, Chemistry of H&E staining. Am J Med Technol 1974;40:513-14.
52 Baker JR. Cytological technique. 5th ed. London: Science Paperbacks,1966.
53 Hrapchak RB. Selective staining with hematoxylin, applications and theory: a review. Am J Med Technol 1976;42:371-9.
54 Prento P, Schulte E. Staining involving metal complex dyes. In: Lyon H, ed. Theory and strategy in histochemistry. Berlin: Springer-Verlag, 1991.
55 Scott SG. On successive double staining for histological purposes. J Pathol Bact 1912;16:390-8.
56 Carazzi D. Eine neue Hämatoxylinlbsung. Z Wiss Mikr 1911;28:273-4.
57 Lillie RD, Pizzolata P. Mechanisms of iron II and iron III sequence hematoxylin stains. J Histochem Cytochem 1972;20:116-29.
58 Slidders W. A stable iron-haematoxylin solution for staining the chromatin of cell nuclei. J Microsc 1968;90:61-5.
59 Puchtler H, Sweat F. On the mechanism of iron-hematein stains. Histochemie 1964;4:197-208.
60 Heidenhain R. Eine neue Verwendung des Hämatoxylin. Arch Mikr Anat 1885;24:468-70.
61 Horobin RW. Understanding histochemistry: selection, evaluation and design of biological stain. New York: John Wiley, 1988.
62 Lendrum AC, Mc17arlane D. A controllable modification of Mallory's trichromic staining method. J Pathol Bact 1940;50:381-4.
63 Van Gieson I. Laboratory notes of technical methods for the nervous system. New York Med J 1889;50:57-60.
64 Puchtler H, Waldrop FS. On the mechanism of Verhoeffs elastica stain: a convenient stain for myelin sheaths. Histochem 1979;62:233-47.
65 Brissie RM, Spicer SS, Hill Bj, Thompson NT. Ultrastructural staining of thin sections with iron hematoxylin. J Histochem Cytochem 1974;22:895-907.
66 Brissie RM, Spicer SS, Thompson NT. The variable fine structure of elastin visualised with Verhoeffs iron hematoxylin. Anat Rec 1975;181:83-94.
67 Spicer SS, Brissie RH, Thompson NT. Variability of dermal elastin visualized ultrastructurally with iron hematoxylin. Am J Pathol 1975;79:481-98.
68 Thompson SW. Selected histochemical and histopathological methods. Springfield: Charles C Thomas, 1966.
69 Mallory FB. A contribution to staining methods. J Exp Med 1900;5:15-20.
70 Gurr E. Staining animal tissues: practical and theoretical. London: Leonard Hill, 1962.
71 Drury RAB, Wallington EA. Carleton's histological technique. 4th ed. New York: Oxford University Press, 1967.
72 McKeever PE, Balentine J1). Nervous system. In: Spicer SS, ed. Histochemistry in pathological diagnosis. New York: Marcel Dekker, 1987.
73 Terner JY, Gurland J, Gaer F. Phosphotungstic acid- hematoxylin; spectrophotometry of the lake in solution and in stained tissue. Stain Technol 1964;39:141-53.
74 Terner JY. Phosphotungstic acid-hematoxylin. Reactivity in vitro. J Histochem Cytochem 1966;14:345-51.
75 Longley JB, Burtner HJ. Preliminary observations on the mechanism of the Mallory phosphotungstic acid hematoxylin method. J Histochem Cytochem 1962A0:664.
76 Puchtler H, Waldrop FS, Meloan SN. On the mechanism of Mallory's phosphotungstic acid - haematoxylin stain. J Microsc 1980;119:383-90.
77 Bulmer D. Observations on histological methods involving the use of phosphotungstic and phosphomolybdic acids, with particular reference to staining with phosphotungstic acid /haematoxylin. Q J Microsc Sci 1962;103:311-23.
78 Bancroft JD, Stevens A. Histopathological stains and their diagnostic uses. Edinburgh: Churchill Livingstone, 1975.
79 Pizzolato P, Lillie RD. Metal salts-hematoxylin staining of skin keratohyalin granules. J Histochern Cytochem 1967;15:104-110.
80 Mallory FB. Pathological technique. Philadelphia: WE; Saunders, 1938.
81 Mallory FB. Phospho-molybdic acid haematoxylin. Anat Anz 1891;6:375-6.
82 Rasmussen AT. Copper hematoxylin, a stain for the acidophils of the human hypophysis. Proc Soc Exp Biol Med 1936;34:760-2.
83 Solcia E, Capella C, Vassallo C. Lead-haematoxylin as a stain for endocrine cells: significance of staining and comparison with other selective methods. Histochenlie 1969;20:116-26.
84 Fullmer HM, Lillie RD. Dilute unmordanted hematoxylin as a stain for basic nuclear protein. J Histochem Cytochem 1962;10:502-3.
85 Pearse AGE. Histochemistry: theoretical and applied. 3rd ed. Edinburgh: Churchill Livingstone, 1972.
86 Clara M. Untersuchungen ilber die spezifische Fdrbung der Kömchen in der basalgekörnten Zellen des Darmepithels durch Beizenfarbstoffe. Zuglerch III. Bietrag zur Theorie der Hämartoxylinfiibungen. Z Zellforsch 1935;22:318-52.
87 Lillie RD, Pizzolato P, Donaldson PT. The CLARA hematoxylin reaction. Acta Histochem Bd. 1974;49:S204-19.
88 Lillie RD, Pizzolato P, Welsh RA, Holmquist ND, Donaldson PT, Berger C. A consideration of substitutes for alum hematoxylin in routine histologic and cytologic diagnostic procedures. Am J Clin Pathol 1973;60:817-19.
89 Lillie RD, Pizzolato P, Donaldson PT. Iron gallein in van Gieson technics, replacing iron hematoxylin. Stain Technol 1973;45:348-9.
90 Lillie RD, Pizzolato P, Donaldson PT. Hematoxylin substitutes: gallein as a biological stain. Stain Technol 1974;49:339-45.
91 Horobin RW, Murgatroyd LB. The composition and properties of gallocyanin - chrome alum stains. Histochem J 1968;1:36-54.
92 Llewellyn BD. Mordant blue 3: a readily available substitute for hematoxylin in the routine hematoxylin and eosin stain. Stain Technol 1974;49:347-9.
93 Schulte EKW, Wittekind DH. Standardized thionin-eosin Y: a quick stain for cytology. Stain Technol 1989;64:255-6.
94 Lillie RD, Pizzolato P, Donaldson PT. Hematoxylin substitutes: fluorone black and methyl fluorone black (9-phenyl- and 9-methyl 1-2,3,7-trihydroxy-6-fluorone) as metachrome iron alum mordant dyes. Stain Technol 1975;50:127-31.
95 Long C, ed. Biochemists' handbook. London: E&FN Spon, 1961.
96 Colour Index. Society of Dyers and Colourists, Bradford (UK) and the American Association of Textile Chemists and Colorists, Research Triangle Park, North Carolina (USA). 3rd ed. (with supplements). 1971-82.
97 Reynolds EF. Martindale: the extra pharmacopoeia. 28th ed. London: Pharmaceutical Press, 1982.
98 Hikino H, Taguchi T, Fujimura H, Hiramatsu Y. Antiinflammatory principles of Caesalpinia sappan wood and of Haematoxylon campechianum wood. Planta Med 1970;31:214-20.
99 Gabor M. Anti-inflammatory substances of plant origin. In: Vane JR, Feurina SH, eds. Anti-inflammatory drugs. Berlin: Springer-Verlag, 1979.
100 Gabor M and Szorady I. Die beeinflussung der experimentellen senfölentziindungen n-dt hAmatoxylin. Acta Physiol Acad Sci Hung 1952;3:405-7.
101 Gabor M, Szorady I, Dirner Z. The inhibiting effect of the members of the haematoxylin group on the action of histidine-decarboxylase. Acta Physiol Acad Sci Hung 1952;3:595-600.
102 Gabor M. Pathophysiology and pharmacology of capillary resistance. Budapest: Akademiai Kiado, 1974.
103 Gabor M, Horvath M, Kiss L, Dirner Z. Über die den adrenalineffekt prolongierende wirkung der glieder der hämatoxylingruppe an überlebenden organen und in vivo. Acta Physiol Acad Sci Hung 1952;3:585-90.
104 Cabor M, Horvath B. Untersuchungen über die wirkung von << P-vitamin-artigen >> pfianzlichen farbstoffen auf das isolierte und in situ arbeitende froschherz. Arch Int Pharmacody 1953;95:68-77.
105 Rogers TD. Light and electron microscope studies on the chemotherapeutic effect of a combination of dimethyl sulfoxide and hematoxylin on a transplantable lymphosarcoma. Diss Abstr Int 1970;31:1734-B.
106 Wissowsky A. Ueber das Eosin als reagens auf Hämoglobin und die Bildung von Blutgefässen und Blutkörperchen bei Sängethicrund Hühnerembryonen. Arch Mikr Anat 1876;3:479-96.
107 Conn HJ. The history of staining: aniline dyes in histology. Stain Technol 1930;53-12.
108 Poole WH. A double staining with haematoxylin and aniline. Q J Microsc Soc 1875;23:375-7.
109 Fischer E. Eosin als Tinctionsmittel für mikroskopische Präparate Arch Mikr Anat 1876;12:349-52.
110 Lillie RD. HJ Conn's biological stains. 9th ed. Baltimore, MD: Williams and Wilkins, 1977.
111 Marshall PN, Bentley SA, Lewis SM. A procedure for assaying commercial samples of eosin. Stain Technol 1975;50:107-13.
112 Birkedal-Hansen H. Eosin staining of gelatine. Histochemie 1973;36:73-87.
113 Hairston MS. The effect of pH on staining by eosins. Stain Technol 1955;30:299-304.