In a paper which has quite recently appeared Diabella has investigated these relations very thoroughly, and his results partly correct, and partly confirm those of Hammerschlag. Diabella found from his comparative estimations that differences of 10% hæmoglobin (Fleischl) correspond in general to differences of 4.46 per thousand in the specific gravity (Hammerschlag's method). Nevertheless with the same amount of hæmoglobin, differences up to 13.5 per thousand are to be observed; and these departures are greater the richer the blood in hæmoglobin. Regular differences exist between men and women; the latter have, with the same amount of hæmoglobin, a specific gravity lower by 2 to 2.5.

Should the parallelism between the number of red blood corpuscles and the amount of hæmoglobin be considerably disturbed, the influence of the stroma of the red discs on the specific gravity of the blood will then be recognisable. Diabella calculates, that with the same amount of hæmoglobin in two blood testings, the stroma may effect differences of 3-5 per thousand in the specific gravity.

Hence the estimation of the specific gravity is often sufficient for the determination of the relative amount of hæmoglobin of a blood. It is only in cases of nephritis and in circulatory disturbances, and in leukæmia, that the relations between specific gravity and quantity of hæmoglobin are too much masked by other influences.

The physiological variations which the specific gravity undergoes under the influence of the taking in and excretion of fluid do not exceed 0.003 (Schmaltz). From what has been said, it follows that all variations must correspond with similarly occurring variations in the factors that underlie the amount of hæmoglobin and the number of corpuscles.

More recent authors, in particular Hammerschlag, v. Jaksch, v. Limbeck, Biernacki, Dunin, E. Grawitz, A. Loewy, have avoided an omission of many earlier investigators; for besides the estimation of the specific gravity of the total blood, they have carried out that of one at least of its constituents, either of the corpuscles or of the serum. The red blood corpuscles have consistently shewn themselves as almost exclusively concerned with variations in the specific gravity of the total blood; partly by variations in number, or changes in their distribution; partly by their chemical instability; loss of water and absorption of water, and variations in the amount of iron.

The plasma of the blood on the contrary—and there is no essential difference between plasma and serum (Hammerschlag)—is much more constant. Even in severe pathological conditions, in which the total blood has become much lighter, the serum preserves its physiological constitution, or undergoes but relatively slight variations in consistence. Considerable diminutions in the specific gravity of the serum are much less frequently observed in primary blood diseases, than in chronic kidney diseases, and disturbances of the circulation. E. Grawitz has lately recorded that in certain anæmias, especially posthæmorrhagic and those following inanition, the specific gravity of the serum undergoes perceptible diminutions[3].

There are still therefore many contradictions in these results, and it is evidently necessary in a scientific investigation always to give the specific gravity of the serum and of the corpuscles, in addition to that of the total blood.

A method closely related to the estimation of the specific gravity is the direct estimation of the dried substance of the total blood, "Hygræmometry"; the clinical introduction of which we owe to Stintzing and Gumprecht. This method is really supplementary to those so far mentioned, and like them can be carried out with the small amounts of blood obtainable at the bedside without difficulty. Small quantities of blood are received in weighed glass vessels: which are then weighed, dried at 65°-70° C. for 24 hours and then weighed again. The figures so obtained for the dried substance have a certain independent importance; for they do not run quite parallel with those of the specific gravity, amount of hæmoglobin or number of corpuscles. The normal values are, for men 21.26%, for women 19.8%.

A further procedure for obtaining indirect evidence of the amount of hæmoglobin is the determination of the volume of blood corpuscles in 100 parts of total blood. For this estimation a method is desirable, which allows of the separation of the corpuscles from plasma in blood, that is as far as possible unaltered. The older methods do not fulfil this requirement; since they recommend either defibrination of the blood (quite impossible with the quantities of blood which are generally clinically available); or keeping it fluid by the addition of sodium oxalate or other substances which prevent coagulation. The separation of the two constituents can be effected by simply allowing the blood to settle, or with the centrifugal machine, specially constructed for the blood by Blix-Hedin and Gärtner ("Hæmatocrit").

For these methods various diluting fluids are used, such as physiological saline solution, 2.5% of potassium bichromate and many others. According to H. Kœppe they are not indifferent as far as the volume of the red blood corpuscles is concerned; and a solution which does not affect the cells must be previously ascertained for each specimen of blood. For this reason attention may be called to the proceeding of M. Herz, in which the clotting of the blood in the pipette is prevented by rendering the walls absolutely smooth by the application of cod-liver oil. Kœppe has slightly varied this method; he fills his handily constructed pipette, very carefully cleaned, with cedar wood oil, and sucks up the blood, as it comes from the fingerprick into the filled pipette. The blood displaces the oil, and as it only comes into contact with perfectly smooth surfaces, it remains fluid. By means of a centrifugal machine, of which he has constructed a very convenient variation, the oil as the lighter body is completely removed from the blood; and the plasma is also separated from the corpuscles. Three sharply defined layers are then visible, the layer of oil above, the plasma layer, and the layer of the red blood corpuscles. In as much as the apparatus is calibrated, the relation between the volumes of the plasma and corpuscles can be read off. No microscopical alterations in the corpuscles are to be observed.

Though this procedure seems very difficult of execution, it is nevertheless the only one, which has really advanced clinical pathology. The results of Kœppe—not as yet very numerous—give the total volume of the red corpuscles as 51.1-54.8%, an average of 52.6%.

M. and L. Bleibtreu have endeavoured indirectly to ascertain the relation of the volume of the corpuscles to that of the plasma. Mixtures of blood with physiological saline solution in various proportions are made, in each the amount of nitrogen in the fluid which is left after the corpuscles have settled is estimated. With the aid of quantities so obtained they calculate mathematically the volume of the serum and corpuscles respectively. Apart from the fact that a dilution with salt solution is also here involved, this method is too complicated and requires amounts of blood too large for clinical purposes. Th. Pfeiffer has tried to introduce it clinically in suitable cases, but has not so far succeeded in obtaining definite results. That, however, the relations between the relative volume of the red corpuscles and quantity of hæmoglobin are by no means constant, is well shewn by conditions (for example the acute anæmias) in which an "acute swelling" of the individual red discs occurs (M. Herz), but without a corresponding increase in hæmoglobin. The same conclusion results from recent observations of v. Limbeck, that in catarrhal jaundice a considerable increase of volume of the red blood corpuscles comes to pass under the influence of the salts of the bile acids.

As we have several times insisted, the quantity of hæmoglobin affords the most important measure of the severity of an anæmic condition. Those methods which neither directly nor indirectly give an indication of the amount of hæmoglobin are only so far of interest that they possibly afford an elucidation of the special pathogenesis of blood diseases in particular cases. To these belong the estimation of the alkalinity of the blood, which in spite of extended observations has not yet obtained importance in the pathology of the blood.

A value to which perhaps attention will be more directed than it has up to the present time by clinicians is the rate of coagulation of the blood, for which comparative results may be obtained by Wright's handy apparatus, the "Coagulometer." In certain conditions, particularly in acute exanthemata, and in the various forms of the hæmorrhagic diathesis, the clotting time is distinctly increased, or indeed clotting may remain in abeyance. Occasionally a distinct acceleration in the clotting, compared with the normal, may be observed. Wright has further ascertained in his excellent researches, that the clotting time can be influenced by drugs: calcium chloride, carbonic acid raise, citric acid, alcohol and increased respiration diminish the clotting power of the blood.

Recently Hayem has repeatedly called attention to a condition, that is probably closely connected with the coagulability of the blood. Although coagulation has set in, the separation of the serum from the clot occurs only very slightly or not at all. Hayem asserts, that he has found such blood in Purpura hæmorrhagica, Anæmia perniciosa protopathica, malarial cachexia: and some infectious diseases.

For such observations large amounts of blood are needed, which are clinically not frequently available. Certain precautions must be observed, as has been ascertained in the preparation of diphtheria serum, so that the yield of serum may be the largest possible. Amongst these that the blood should be received in longish vessels, which must be especially carefully cleaned, and free from all traces of fat. If the blood-clot does not spontaneously retract it must be freed from the side of the glass with a flat instrument like a paper-knife without injuring it. If no clot occurs in the cold, a result may perhaps follow at blood temperature.

In spite however of all artifices and all care, it is here and there, under pathological conditions, impossible to obtain even a trace of serum from considerable amounts of blood. In a horse for example which was immunised against diphtheria, and had before yielded an unusually large quantity of serum, Ehrlich was able to obtain from 22 kg. of blood scarcely 100 cc. serum, when the animal was bled on account of a tetanus infection.

Perhaps a larger rôle is to be allotted in the diseases of the blood to these conditions. Hayem already turns the incomplete production of serum to account, for distinguishing protopathic pernicious anæmia from other severe anæmic conditions. A bad prognosis too may be made when for example in cachetic states this phenomenon is to be observed.

A few methods still remain to be mentioned which test the resistance of the red blood corpuscles to external injuries of various kinds.

Landois, Hamburger and v. Limbeck ascertain for instance the degree of concentration of a salt solution, in which the red corpuscles are preserved ("isotonic concentration," Hamburger) and those which cause an exit of the hæmoglobin from the stroma. The erythrocytes are the more resistant, the weaker the concentration which leaves them still uninjured.

Laker tests the red blood corpuscles as regards their resistance to the electric discharge from a Leyden jar, and measures it by the number of discharges up to which the blood in question remains uninjured.

Clinical observation has not yet gained much by these methods. So much only is certain, that in certain diseases: anæmia, hæmoglobinuria, and after many intoxications, the resistance, as measured by the methods above indicated, is considerably lowered.

THE MORPHOLOGY OF THE BLOOD.

A. METHODS OF INVESTIGATION.

A glance at the history of the microscopy of the blood shews that it falls into two periods. In the first, which is especially distinguished by the work of Virchow and Max Schultze, a quantity of positive knowledge was quickly won, and the different forms of anæmia were recognised. But close upon this followed a standstill, which lasted for some decades, the cause of which lay in the circumstance that observers confined themselves to the examination of fresh blood. What in fact was to be seen with the aid of this simple method, these distinguished observers had quickly exhausted. That these methods were inadequate is best shewn by the history of leucocytosis, which after the precedent of Virchow was in general referred to an increased production on the part of the lymphatic glands; and further by the imperfect distinction between leucocytosis and incipient leukæmia, which was drawn almost exclusively from purely numerical estimations. It was only after Ehrlich had introduced the new methods of investigation by means of stained dry preparations, that the histology of the blood received the impulse for its second period.

We owe to them the exact distinction between the several kinds of white blood corpuscles, a rational definition of leukæmia, polynuclear leucocytosis, and the knowledge of the appearances of degeneration and regeneration of the red blood corpuscles, and of their degeneration in hæmoglobinæmic conditions. The same process, then, has gone on in the microscopy of the blood that we see in other branches of normal and pathological histology: by advances in method, advances in knowledge full of importance result. It is therefore little comprehensible, that an author quite recently should recommend a reversion to the old methods, and emphatically announce that he has managed to make a diagnosis in all cases, with the examination of fresh blood. At the present time, after the most important points have been cleared up by new methods, in the large majority of cases, this is not an astonishing achievement. For any difficult case (for instance the early recognition of malignant lymphoma, certain rare forms of anæmia, etc.) as the experienced know, the dry stained preparation is indispensable. The object of examining the blood, is certainly not to make a rapid diagnosis, but to investigate exactly the individual details of the blood picture. To-day, we can only take the standpoint, that everything that is to be seen in fresh specimens—apart from the quite unimportant rouleaux formation, and the amœboid movements—can be seen equally well, and indeed much better in a stained preparation; and that there are several important details which are only made visible in the latter, and never in wet preparations.

As regards the purely technical side of the question, the examination of stained dry specimens is far more convenient than that of fresh. For it leaves us quite independent of time and place, we can keep the dried blood with few precautions for months at a time, before proceeding to further microscopic treatment; and the examination of the preparation may last as long as required, and can be repeated at any time. On the contrary, the examination of the wet preparation is only possible at the bedside, and must be conducted within so short a time, on account of the changeability of the blood, clotting, destruction of white corpuscles and so forth, that a searching investigation cannot be undertaken. In addition the preparation and staining of dried blood specimens is amongst the simplest and most convenient of the methods of clinical histology. In the interest of its wider dissemination, it will be justifiable to describe it more in detail.

We must also mention here the use of the dry preparation in the estimation of the important relation between the number of the red and of the white corpuscles; and also of the relative numbers of different kinds of white blood corpuscles.

For this purpose, faultless specimens, specially regularly spread, are indispensable. Quadratic ocular diaphragms (Ehrlich-Zeiss) are requisite, which form a series, so that the sides of the squares are as 1:2:3 ... :10, the fields therefore as 1:4:9 ... :100. The eye-piece made by Leitz after Ehrlich's directions is more convenient, in which, by a handy device, definite square fractions of the field can be obtained. The enumeration is made as follows. The white blood corpuscles are first counted in any desired field with the diaphragm no. 10, that is with the area of 100. Without changing the field, the diaphragm 1, which only leaves free a hundredth part of this area, is now put in, and the red corpuscles are counted. The field is then changed at random, and the red corpuscles counted in a portion of the area which represents the hundredth of that of the white. About 100 such counts are to be made in a specimen. The average of the red corpuscles is then multiplied by 100 and so placed in proportion with the sum of the white. If the white corpuscles are very numerous, so that counting each one in a large field is inconvenient, smaller sections of the eye-piece 81, 64, 49, etc. may be taken.

The important estimation of the percentage relation of the various forms of leucocytes is effected by the simple "typing" of several hundred cells, a count which for the practised observer is completed in less than a quarter of an hour.

α Preparation of the dry specimen.

To obtain unexceptionable preparations cover-slips of particular quality are necessary. They should not be thicker than 0.08 to 0.10 mm., the glass must not be brittle or faulty, and must in this thickness easily allow of considerable bending, without breaking. Every unevenness of the slip renders it useless for our purposes. The glasses must previously be particularly carefully cleaned, and all fat removed. It is generally sufficient to allow the slips to remain in ether for about half-an-hour, not covering one another. Each one still wet with ether is then wiped with soft, not coarse, linen rag or with tissue-paper. The slips now are put into alcohol for a few minutes, are dried in the same manner as from the ether, and are kept ready for use in a dust-tight watch-glass. Bearing in mind, that these cover-slips are not cut out from a flat piece but from the surface of a sphere, it is evident that only with glasses thus prepared, can it be expected that a capillary space should be formed between two of them, in which the blood spreads easily. For with the smallest unevenness or brittleness of the glass it is an impossibility for the one to fit every bend of the other. And it is only then that the slips can be drawn away one from another, without using a force which breaks them.

To avoid fresh soiling of the cover-slips, and above all the contact of the blood with the moisture coming from the finger, the cover-glass is held with forceps[4] to receive the blood. We recommend for the under cover-glass a clamp forceps a, with broad, smooth blades; the ends may be covered with leather or blotting-paper for a distance of about 1/2 in. For the other cover-slip a very light spring forceps b, with smooth blades, sharp at the tips, is used, with which a cover-glass can be easily picked up from a flat surface. The lower slip is now fixed by one edge in the clamp forceps, and held ready in the left hand. The right hand applies the upper glass with the forceps b to the drop of blood as it exudes from the puncture, and takes it up, without touching the finger itself. The forceps b is then quickly brought to a and the slip with the little drop of blood allowed to fall lightly on the other. In glasses of the right quality the drop distributes itself spontaneously in a completely regular capillary layer. With two fingers of the right hand on the edge of the upper glass, it is now carefully pulled from the lower, which remains fixed in the clamp, without pressing or lifting. Frequently only one, the lower, shews a regular layer, but occasionally both are available for examination. During the desiccation in the air, generally complete in 10-30 seconds, the preparations must naturally be protected from any dampness (for example the breath of the patient).

The extent of surface which is covered depends on the size of the drop, the smaller the latter, the smaller the surface over which it has to be spread. Large drops are quite useless, for with them, the one cover-glass swims on the other, instead of adhering to it.

Although a written description of these manipulations makes the method seem rather intricate, yet but little practice is required to obtain an easy and sure mastery over it. We have felt compelled to describe the method minutely, since preparations so often come under our notice which, although made by scientific men, who pursue hæmatological investigations, are only to be described as technically completely inadequate.

The specimens so obtained, after they are completely dried in the air, should be kept between layers of filter-paper in well closed vessels till further treatment. In important cases, preparations of which it is desirable to keep for some considerable time, some of the specimens should be kept from atmospheric influences by covering them with a layer of paraffin. The paraffin must be removed by toluol before proceeding further. The preparations must naturally be kept in the dark.

β. Fixation of the dry specimen.

All methods of staining available for the blood require the fixing of the proteids of the blood. A general formula cannot be given, since the intensity of the fixation must be regulated in accordance with the kind of stain that is chosen. Relatively slight degrees of hardening suffice for staining in simple watery solutions, for example, in the triacid fluid, and can be attained by a short, and not too intense action of several reagents. For other methods, in which solutions that are strongly acid or alkaline are employed, it is however necessary to fix the structure much more strongly. But here, too, an excess as well as an insufficiency must be guarded against. It is easy with the few staining fluids that are in use to ascertain the optimum for each.

The following means of fixation are employed.

1. Dry Heat.

A simple plate of copper on a stand is used, under one end of which burns a Bunsen flame. After some time a certain constancy in the temperature of the plate is reached, the part nearest to the flame is hottest, that farther away is cooler. By dropping water, toluol, xylol, etc. on to it, one can fairly easily ascertain that point of the plate which has reached the boiling temperature of the particular fluid.

Far more convenient is Victor Meyer's apparatus, used by chemists. This consists of a copper boiler, modified for our purpose, with a roof of thin copper-plate, perforated for the opening of the vapour tube. Small quantities of toluol are allowed to boil for a few minutes in the boiler, and the copper-plate soon reaches the temperature of 107°-110°.

For the ordinary staining reagents (in watery fluids) it is enough to place the air-dried preparation at about 110° C. for one half to two minutes. For differential staining mixtures, for instance the eosin-aurantia-nigrosin mixture, a time of two hours is necessary, or higher temperatures must be employed.

2. Chemical means.

a. To obtain a good triacid stain, the preparations may be hardened, according to Nikiforoff, in a mixture of absolute alcohol and ether of equal parts, for two hours. The beauty of specimens fixed by heat is however not quite fully reached by this method.

b. Absolute alcohol fixes dried specimens in five minutes sufficiently to stain them subsequently with Chenzinsky's fluid, or hæmatoxylin-eosin solution. It is an advantage in many cases, especially when rapid investigation is required, to boil the dried preparation in a test-tube in absolute alcohol for one minute.

c. Formalin in 1% alcoholic solution was first used by Benario for fixing blood preparations. The fixation is complete in one minute, and the granulations can be demonstrated. Benario recommends this method of fixing, especially for the hæmatoxylin-eosin staining.

These methods are described as the most suitable for blood-investigation in general. For special purposes, for instance, the demonstration of mitoses, blood platelets, etc., other hardening reagents may be used with advantage: Sublimate, osmic acid, Flemming's fluid, and so forth.

γ. Staining of the dry specimen.

Staining methods may be classified according to the purpose to which they are adapted.

We use first those which are suitable for a simple general view. For this it is sufficient to use such solutions as stain hæmoglobin and nuclei simultaneously. (Hæmatoxylin-eosin, hæmatoxylin-orange).

Occasionally a stain is desirable which only brings out, but in a characteristic manner, a special kind of cell, e.g. the eosinophils, mast cells, or bacteria. Single staining is attained on the principle of maximal decoloration. (Cp. E. Westphal.)

Finally, we have panoptic staining; that is, by methods which bring out, as characteristically as possible, the greatest number of elements. Although we must use high magnifications with these stains, we are compensated by a knowledge of the blood condition that cannot be reached in any other way. A double stain is generally insufficient, and at least three different dyes are used.

Successive staining was formerly used for this purpose. But everyone who has used this method knows how difficult it is to get constant results, however careful one may be in the concentration and time of action of the stain.

Simultaneous staining offers undoubted and important advantages. As there is much obscurity with regard to the principle on which it rests we may here shortly explain the theory of simultaneous staining.

We will begin with the simplest example: the use of picro-carmine, a mixture of neutral ammonium carmine and ammonium picrate. In a tissue rich in protoplasm, carmine alone stains diffusely, though the nuclei are clearly brought out. But if we add an equally concentrated solution of ammonium picrate, the staining gains extraordinarily in distinctness, in as much as now certain parts are pure yellow, others pure red. The best known example is the staining of muscle with picro-carmine, by which the muscle substance appears pure yellow, the nuclei pure red. If, however, instead of ammonium picrate we add another nitro dye which contains more nitro-groups than picric acid, for example the ammonium salt of hexa-nitro-diphenylamine, the carmine stain is completely abolished, all parts stain in the pure aurantia colour. The explanation of this phenomenon is obvious. Myosin has a greater affinity for ammonium picrate than for the carmine salt, and therefore in a mixture of the two combines with the yellow dye. Owing to this combination it is not now in a condition to chemically fix even carmine. Further, the nuclei have a great affinity for the carmine, and therefore stain pure red in this process. If, however, nitro dyes be added to the carmine solution, which have an affinity for all tissues, and also for the nuclei, the sphere of action of the carmine becomes continually smaller, and finally by the addition of the most powerful nitro body, the hexa-nitro compound, is completely abolished. Connective tissue and bone substance, however, behave differently with the picro-carmine mixture, in as much as here the diffuse stain depends exclusively on the concentration of the carmine, and is quite uninfluenced by the addition of a chemical antidote. This staining can only be limited by dilution, but not by the addition of opposed dyes. We must look upon the latter kind of tissue stain not as a chemical combination, but as a mechanical attraction of the stain on the part of the tissue. We may also say: chemical stains are to be recognised by the fact that they react to chemical antidotes; mechanical stains to physical influences; of course always assuming, that purely neutral solutions are employed, and that all additions, which alter the chemical relation of the tissues such as alkalis and acids, or which raise or limit the affinity of the dye for the tissues, are avoided. A further consequence of this view is, that all successive double staining may be serviceably replaced by simultaneous multiple staining, if the chemical nature of the staining process is settled. In contradistinction, in all double stains, which can only be effected by successive staining, mechanical factors are concerned.

In the staining of the dry blood specimen, purely chemical staining processes are concerned, and therefore the polychromatic combination stain is possible in all cases.

The following combinations are possible for the blood:

1. Combined staining with acid dyes. The best known example is the eosin-aurantia-nigrosin mixture, in which the hæmoglobin takes on an orange, the nuclei a black, and the acidophil granulations a red hue.

2. Mixtures of basic dyes. It is possible straight away to make mixtures consisting of two basic dyes. As specially suitable we must mention fuchsin, methyl green, methyl violet, methylene blue. On the other hand, mixtures of three bases are fairly difficult to prepare, and the quantitative relations of the constituents must be exactly observed. For such mixtures, fuchsin, bismarck brown, chrome green, may be used.

3. Neutral mixtures. These have played an important part in general histology, from the time that they were first introduced by Ehrlich into the histology of the blood up to the present day; and deserve before all others a full consideration.

Neutral staining rests on the fact, that nearly all basic dyes (i.e. salts of the dye bases, for instance, rosanilin acetate) form combinations with acid dyes (i.e. salts of the dye acids, for instance, ammonium picrate) which are to be regarded as neutral dyes, such as rosanilin picrate. Their employment offers considerable difficulties as they are very imperfectly soluble in water. A practical application of them was first possible after Ehrlich had ascertained that certain series of the neutral dyes are easily soluble in excess of the acid dye, and so the preparation of solutions of the required strength, readily kept, was made possible. Among the basic dyes which are suitable for this purpose are those particularly which contain the ammonium group, especially methyl green, methylene blue, amethyst violet[5] (tetraethylsafraninchloride), and to a certain extent pyronin and rhodamin also. In contradistinction to these, the members of the triphenylmethan series, such as fuchsin, methyl violet, bismarck brown, phosphin, indazine, are in general less suited for the purpose, with the exception of methyl green already mentioned. The acid dyes specially suited for the production of soluble neutral stains are the easily soluble salts of the polysulpho-acids. The salts of the carbonyl acids and other acid phenol dyes are but little suitable: and least of all, the nitro dyes. Specially to be mentioned among the acid dye series are those which can be used for the preparation of the neutral mixtures: orange g., acid fuchsin, narcëin (an easy soluble yellow dye, the sodium salt of sulphanilic acid—hydrazo-β-naphtholsulphonic acid).

If a solution of methyl green be allowed to fall drop by drop into a solution of an acid dye, for instance orange g., a coarse precipitate first results, which dissolves completely on the further addition of the orange. No more orange should be added than is necessary for complete solution. This is the type of a simple neutral staining fluid. Chemically the above-mentioned example may be thus explained; in this mixture all three basic groups of the methyl green are united with the acid dye, so that we have produced a triacid compound of methyl green.

Simple neutral mixtures, which have one constituent in common, may be combined together straight away. This is very important for triple staining, which can only be attained by mixing together two simple neutral mixtures, each consisting of two components. A chemical decomposition need not be feared. We thus get mixtures containing three and more colours. Theoretically there are two possibilities for such combinations:

1. Staining mixtures of 1 acid and 2 basic dyes,

2. Staining mixtures of 2 acids and 1 base, in particular the mixture to be described later in detail of

and the corresponding combinations with methylene blue, and amethyst violet may be mentioned.

The importance of these neutral staining solutions lies in the fact that they pick out definite substances, which would not be demonstrated by the individual components, and which we therefore call neutrophil.

Elements which have an affinity for basic dyes, such as nuclear substances, stain in these neutral mixtures purely in the colour of the basic dye; acidophil elements in that of one of the two acid dyes; whilst those portions of tissue which from their constitution have an equal affinity for acid and basic dyes, attract the neutral compound, as such, and therefore stain in the mixed colour.

The eosine-methylene blue mixtures are exceptional in so far, that it is possible with them, for a short time at least, to preserve active solutions, in which with an excess of basic methylene blue, enough eosin is dissolved for both to come into play. A drawback however of such mixtures is, that in them precipitates are very easily produced, which render the preparation quite useless. This danger is particularly great in freshly prepared solutions. In solutions, such as Chenzinsky's, which can be kept active for a longer time, it is less. Hence fresh solutions stain far more intensely and more variously than older ones, and are therefore used in special cases (see page 46). If the stain is successful the appearances are very instructive. Nuclei are blue, hæmoglobin red, neutrophil granulation violet, acidophil pure red, mast cell granulation deep blue, forming one of the most beautiful microscopic pictures.

For practical purposes, besides the iodine and iodine-eosine solution described below (see page 46) the following are especially used:

1. Hæmatoxylin solution with eosin or orange g.

The fluid must stand for some weeks. The preparations, fixed in absolute alcohol, or by short heating, stain in from half-an-hour to two hours. The hæmoglobin and eosinophil granules are red, the nuclei stain in the colour of hæmatoxylin. The solution must be very carefully washed off.

2. In the practical application of the triacid fluid, particular care must be taken, as M. Heidenhain first shewed, that the dyes are chemically pure[6]. Formerly granules, apparently basophil, were frequently observed in the white blood corpuscles, particularly in the region of the nucleus. They were not recognised, even by practised observers (e.g. Neusser) as artificial, but were regarded as preformed, and were described as perinuclear forms. Since the employment of pure dyes these appearances, whose meaning for a long time puzzled us, are but seldom seen.

Saturated watery solutions of the three dyes are first prepared, and cleared by standing for some considerable time. The following mixture is now made:

These fluids are measured in the above-mentioned order, with the same measuring glass; and from the addition of methyl green onwards the fluid is thoroughly shaken. The solution can be used at once, and keeps indefinitely. The staining of the blood specimen in triacid requires only a little fixation, cp. page 35. The stain is completed in five minutes at most.

The nuclei are greenish, the red blood corpuscles orange, the acidophil granulation copper red, the neutrophil violet. The mast cells stand out by "negative staining" as peculiar bright, almost white cells, with nuclei of a pale green colour.

The triacid stain is very convenient. It is much to be recommended for good general preparations; it is indispensable in all cases where the study of the neutrophil granulations is concerned.

3. Basic double staining. Saturated, watery methyl-green solution is mixed with alcoholic fuchsin.

The stain, which only requires a small fixation, is completed in a few minutes, and colours the nuclei green, the red blood corpuscles red, the protoplasm of the leucocytes fuchsin colour. It is therefore specially suited for demonstration preparations of lymphatic leukæmia.

4. Eosin-methylene blue mixtures, for example Chenzinsky's fluid:

This fluid is fairly stable, but must always be filtered before use. It only requires a fixation of the specimen for five minutes in absolute alcohol. The staining takes 6-24 hours (in air-tight watch-glasses) at blood temperature. The nuclei and the mast cell granulations stain deep blue, malaria plasmodia light sky blue, red corpuscles and eosinophil granules a fine red.

This solution is particularly suited for the study of the nuclei, the baso and eosinophil granulations, and it is used by preference for anæmic blood, and also for lymphatic leukæmia.

5. 10 c.c. of a 1 per cent. watery eosin solution, with 8 c.c. methylal, and 10 c.c. of a saturated watery solution of methylene blue are mixed, and used at once, see page 41. Time of staining 1, at most 2 minutes. The staining is characteristic only in preparations very carefully fixed by heat. The mast cell granulations are stained pure blue, the eosinophil red, the neutrophil in mixed colour.

6. Jenner's stain consists of a solution in methyl alcohol of the precipitate formed by adding eosine to methylene blue.

Precipitate allowed to stand 24 hours, and then dried at 55°. It is then made up to 1/2% in methyl alcohol (Merck). The stain may be obtained from R. Kanthack, 18, Berners Street, London, ready for use. It is exceedingly sensitive to acids and alkalis. Fixation is effected by heat. Time of staining 1-4 minutes.

Before we pass to the histology of the blood, two important methods may be described, for which the dried blood preparation is employed directly, without previous fixation: 1. the recognition of glycogen in the blood; 2. the microscopic test of the distribution of the alkali of the blood.

1. Recognition of glycogen in blood.

This may be effected in two ways. The original procedure consisted in putting the preparation into a drop of thick cleared iodine-indiarubber solution under the microscope, as had been already recommended by Ehrlich for the recognition of glycogen.

The following method is still better. The preparation is placed in a closed vessel containing iodine crystals. Within a few minutes it takes on a dark brown colour, and is then mounted in a saturated lævulose solution, whose index of refraction is very high. To preserve these specimens they must be surrounded with some kind of cover-glass cement.

By the use of better methods the red blood corpuscles which have taken on the iodine stain stand out, without having undergone any morphological change. The white blood corpuscles are only slightly stained. All parts containing glycogen on the contrary, whether the glycogen be in the white blood corpuscles, or extracellular, are characterised by a beautiful mahogany brown colour. The second modification of this method is specially to be recommended on account of the strong clearing action of the lævulose syrup. In using the iodine-indiarubber solution a small quantity of glycogen in the cells may escape observation owing to the opaqueness of the indiarubber, and occasionally too by the separate staining of the same. The second more delicate method is for this reason recommended, in the investigation of cases of diabetes and other diseases[7].

2. The microscopic test of the distribution of alkali in the blood.

These methods rest on a procedure of Mylius for the estimation of the amount of alkali in glass. Iodine-eosine is a red compound easily soluble in water, which is not soluble in ether, chloroform, or toluol. But the free coloured acid, which is precipitated by acidifying solutions of the salt, is very sparingly soluble in water. It is, on the contrary, very easily soluble in organic solvents, so that by shaking, it completely passes over into an etherial solution, which becomes yellow. If this solution be allowed to fall on glass, on which deposits of alkali have been formed by decomposition, they stand out in a fine red colour as the result of the production of the deeply coloured salt.

In its application to the blood, of course the vessels used for staining as well as the cover-glasses must be cleaned from all adhering traces of alkali by means of acids. The dry specimen is thrown directly after its preparation into a glass vessel containing a chloroform or chloroform-toluol solution of free iodine-eosine. In a short time it becomes dark red. It is then quickly transferred to another vessel containing pure chloroform, which is once more changed, and the preparation still wet from the chloroform is then mounted in canada balsam. In such preparations the morphological elements have preserved their shape completely. The plasma shews a distinct red colour, whilst the red corpuscles have taken up no colour. The protoplasm of the white corpuscles is red, the nuclei appear as spaces, because unstained (negative nuclear staining). The disintegrated corpuscles and the fibrin which is produced, shew an intense red stain. These stains are peculiarly instructive, and shew many details which are not visible in other methods. The study of these preparations is really of the highest value, since they allow the products of manipulation of the dry preparation and every error of production to stand out in the most reliable manner, and so render possible a kind of automatic control. The scientific value of this method lies in the fact that it throws light on the distribution of the alkali in the individual elements of the blood. It appears that free alkali reacting on iodine-eosine is not present in the nuclei; these must therefore have a neutral or an acid reaction. On the contrary the protoplasm of the leucocytes is always alkaline, and the largest amount of alkali is held by the protoplasm of the lymphocytes. We call particular attention, in this connection, to the strong alkalinity of the blood platelets.