The analogy between the modifications undergone by the red blood corpuscles and other cells inside the macrophages and the changes that take place in the intestinal cells of Planarians and Actinians, suggests that the resorption of formed elements must undoubtedly be regarded as a true intracellular digestion. It would, however, be a very important matter to be able to support this conclusion by even more convincing proofs. The study of the artificial digestion that is observed in vitro in the case of the macerated mesenterial filaments of Actinians has furnished a very valuable argument in favour of the enzymatic nature of intracellular digestion. Animal exudations are not well adapted for this special line of study. We can only obtain them as the result of the injection of different substances, solid or fluid, which are greedily absorbed by phagocytes. If we collect the exudations at a moment when the number of these cells is still considerable we must withdraw along with them many digestive substances which interfere with our observation. We may therefore with advantage turn our attention to masses of phagocytes collected in organs. As it is mainly the macrophages which effect the resorption of cells, it is evident that we must choose the centres where they are formed in order to investigate the digestive ferments. Let us take, then, the lymphatic glands of the mesentery, the glandular portion of the omentum and the spleen, the three pre-eminently macrophagic organs, and let us see if, with an extract of them, prepared with physiological salt solution (0·75% of sodium chloride), any digestive effect is to be obtained.

Macerate the three organs mentioned of a guinea-pig and mix the extracts thus obtained with red blood corpuscles of the goose, corpuscles that have already given us information in connection with the phenomena of resorption in the living organism. In almost all the guinea-pigs a solution of the red blood corpuscles of the goose by the extract of the glandular portion of the omentum may be observed. The mesenteric glands likewise give an extract which in most cases has a solvent action. The extract from the spleen is only active in a limited number of cases. In all these examples the extracts from macrophagic organs bring about the solution of the haemoglobin, but leave intact the membrane and nucleus of the corpuscles. In this respect there exists, then, a certain difference between this and the digestion of red corpuscles in the macrophages of exudations, where the membrane and even the nucleus are in the end completely dissolved. This difference may be explained by the fact that in the preparation of the extract in physiological salt solution, one part only of the soluble digestive ferment may be set at liberty.

The solvent action of extracts of macrophagic organs must in fact be attributed to the presence of a soluble ferment in the cells of which these organs are made up. As the diastases are distinguished, in general, by their great sensitiveness to heat, we tried the action of our extracts after a preliminary heating, when it was found that a temperature of 56° C., applied for three quarters of an hour, completely abolished the solvent action of the extracts upon the red blood corpuscles of the goose. The soluble ferment of macrophagic organs, to which we propose to give the name of macrocytase^[110] or macrophage ferment, is in many respects analogous to the actino-diastase of Mesnil, described in the preceding chapter.

With a view to obtain more complete information on the cytases I suggested to Tarassewitch that he should make a detailed study of them; this he has carried out in my laboratory. He has demonstrated that the macrophagic organs of other mammals than the guinea-pig, especially those of the rabbit and dog, exert the same solvent action on the red blood corpuscles. He has also established the fact that this action applies not only to the red corpuscles of the goose but extends also to those of several other birds and mammals. Tarassewitch succeeded in confirming the injurious action of heat on macrocytase. Extracts of macrophagic organs which contain much debris in suspension, when heated for an hour at 55°·5 C. in certain cases lose their solvent property for red blood corpuscles; sometimes this temperature brings about merely a weakening of the macrocytase. In order to destroy it surely and completely, the suspensions must be heated at 58°·5–62° C. for an hour. If, however, instead of heating the entire suspension, we first pass it through filter paper, the clear fluid filtrate is deprived of its diastatic action even after it has been heated at 55°·5 C. for three quarters of an hour.

Of all the other organs of which extracts have been kept in prolonged contact with the red blood corpuscles of birds, the pancreas alone has shown a very well-marked digestive action. Extracts of the salivary glands exerted a feeble solvent action on a certain quantity of the red corpuscles. The other organs, such as the liver, kidneys, brain, spinal cord, ovary, testicles, suprarenal capsules and placenta, exercised no such action. Even bone marrow, in agreement with my results published some years ago, showed itself quite inactive.

The blood serum of guinea-pigs which I employed in my researches, as well as that of the animals examined by Tarassewitch, has not shown itself capable of dissolving the red blood corpuscles of the goose, although the macrophagic organs dissolve them easily. It has long been known, however, that the serum of the blood of many animals will destroy the red corpuscles of a different species. This demonstration was afforded during the period when attempts were being made to transfuse the defibrinated blood of mammals, especially of the sheep, into man. This practice had to be abandoned, in consequence of the difficulties resulting from the solution of the human red corpuscles. Later, Daremberg^[111] and Buchner^[112] set themselves to study this haemolytic action of serums systematically. They found that it was due to a particular substance to which Buchner gave the name of alexine or protective substance. Of indeterminate chemical composition, this substance is allied to albuminoid substances. It is destroyed when heated to 55°–56° C. and only acts in the presence of certain salts. When these salts are removed from the serum by dialysis, it loses its haemolytic power; but as soon as the salts are replaced in proper proportion this power reappears. Later, Buchner^[113] compared the action of alexine to that of soluble ferments and referred it to the category of the digestive diastases. According to him the same alexine is capable of dissolving the red blood corpuscles of several species of Vertebrates. Bordet^[114], in a series of researches made in the Pasteur Institute, confirmed this view. He came to the conclusion that the alexines of the various species of animals differ from one another. Thus, the alexine of the blood serum of the rabbit is not the same as that found in the serum of the guinea-pig or dog. Nevertheless each of these alexines is capable of exerting a solvent action on the red blood corpuscles of several species.

Ehrlich and Morgenroth^[115], in a series of memoirs on the solution of red blood corpuscles, have combated the idea that there is only a single alexine in one and the same serum. Moreover, they state that alexine always requires for its action the aid of another substance, and that matters are much more complicated than at first sight appears. They maintain that in each normal serum a number of different substances are found, each one of which only attacks a single species of red blood corpuscle. They point out that the solution of the red corpuscles by the normal serum takes place through the combined action of two different substances and cite several cases where a normal serum, after being heated to 55° C. and so deprived of its haemolytic power, again becomes capable of dissolving the red corpuscles when some normal serum from another species, which of itself is destitute of the solvent property, is added to it. Let us quote an example from Ehrlich and Morgenroth. The normal serum of the goat readily dissolves the red blood corpuscles of the rabbit and guinea-pig, but if heated for half-an-hour at 55° C., it loses this power. On the other hand, the normal serum of many horses shows itself powerless to dissolve the red corpuscles of these rodents. Here, then, are two serums, equally incapable of effecting the solution of the red corpuscles of the rabbit and guinea-pig. Yet, when they are mixed together and to them a few drops of blood from one of the rodents cited is added, haemolysis takes place readily. The heated goat’s serum then, has, retained in it something that resists a temperature of 55° C., a substance which, by itself, leaves the red blood corpuscles intact; but which, when combined with a second substance present in the horse’s serum, causes their solution. Ehrlich gives to the first substance, that is to say that found in the heated goat’s serum, the name of intermediary body (“Zwischenkörper”). The second substance, present in the unheated horse’s serum, is designated by him the complement. In order that a normal serum may dissolve the red corpuscles, it is not sufficient that it should possess a single substance, the alexine of Buchner. It must, to exert this action, contain two distinct substances which are very often found together in the same normal serum. Unheated goat’s serum was only capable of dissolving the red blood corpuscles of the rabbit because a particular complement and intermediary substance were both present. Deprived of its complement at 55° C., the serum is solvent only when we add to it another substance that is contained in the normal serum of a different species (horse). Continuing their researches in this direction, Ehrlich and Morgenroth have come to the conclusion that the normal serum of a single species may contain several intermediary substances, each one acting on a single species of red blood corpuscles. Further, that normal serum must contain several or even many different complements.

Ehrlich and Morgenroth carried on researches on the intermediary substances in normal serums and found several in addition to that already mentioned. The serum of the normal dog readily dissolves the red blood corpuscles of the guinea-pig. When heated to 57° C. it loses this property; but with the addition of normal guinea-pig’s serum the property is regained. In the serum of the normal dog there exists, then, besides the complement, at least one intermediary substance. The same result can be obtained with several combinations of serums of normal mammals, heated or unaltered^[116]. Yet it often happens, as Ehrlich and Morgenroth themselves point out, that the demonstration of the presence of the intermediary substance in normal serums is accompanied with marked difficulties. Bordet, also, who has studied this question very thoroughly, often failed completely in his attempts to make normal serums, that were incapable of producing haemolysis, active by the addition of heated serums of other species of animals. Thus he observed that normal fowl’s serum readily dissolves the red corpuscles of the rabbit. When heated to 55°–56° C. this serum loses its haemolytic power which cannot be restored by the addition of any normal serum. He thinks therefore that, in this example, haemolysis is produced solely by the alexine, without the assistance of any intermediary substance in the serum of the normal fowl. P. Müller^[117], whilst confirming Bordet’s experimental results, considers that, in this case also, there is the intervention of an intermediary substance. When he mixed heated fowl’s serum with a small quantity of unaltered fowl’s serum the solution of the red corpuscles of the rabbit is not brought about. When, however, instead of adding a little unheated normal fowl’s serum, he added the same quantity of serum from a fowl previously treated with physiological salt solution, the red corpuscles of the rabbit were dissolved without any difficulty. Müller explains this difference as due to the fact that the serum of the treated fowl contains more complementary substance than does that of the normal fowl.

We see, then, from this example that the analysis of the phenomena taking place in the solution of the red corpuscles by normal serums is beset with very great difficulties. For this reason it is much more profitable to make researches in this direction, using more active serums, where the demonstration of the two substances can be made simply and exactly. This desideratum has been supplied by J. Bordet, when preparateur in our laboratory; he described an easy method of increasing the haemolytic power of serums.

As stated above, guinea-pigs that have received an intraperitoneal injection of goose’s blood digest the corpuscles, although the peritoneal fluid exerts no haemolytic action. In vitro, the extract of their macrophagic organs certainly dissolves the red corpuscles, whilst the blood serum usually fails to do so. Now, if a second or a third injection of goose’s blood be made into the peritoneal cavities of the same guinea-pigs, partial solution of the corpuscles takes place in the peritoneal plasma and the serum of the blood acquires new properties: it becomes capable of clumping the red corpuscles, that is to say of agglutinating them; afterwards it dissolves them in vitro.

J. Bordet^[118] has shown that the injection of the blood of one species of Vertebrate (mammal or bird) into the peritoneal cavity or under the skin of an animal of a different species, always produces in the blood serum of the latter the haemolysing substance. This haemolysing substance is specific or nearly so, that is to say it dissolves the red corpuscles of the species which has furnished the injected blood and also, but more feebly, the red corpuscles of allied species. Consequently, with guinea-pig’s serum, treated with goose’s blood, we obtain the greatest solvent action on the red corpuscles of the goose, though there is a certain haemolysis of the red corpuscles of some other birds. This rule, thoroughly established by Bordet, has been the starting-point for a large number of researches on haemolysis and amongst others of those which bear on the intermediary substance of normal bloods.

Bordet demonstrated very definitely a fact of fundamental importance—that in the blood serums of animals treated with blood from a different species, there exist two distinct substances which only dissolve the red blood corpuscles when they are combined. Here the duality of the haemolytic agent cannot be doubted, as it may in certain examples of normal serums. Each time that we deprive the serum of a treated animal of its solvent action by heating at 55°–56° C., this property can be restored to it with certainty by the addition of a little normal serum which, by itself, is incapable of bringing about haemolysis. The heated serum of these injected animals loses the power of dissolving the corresponding red corpuscles, but it retains its other acquired property—the agglutination of the corpuscles. The red corpuscles, aggregated into voluminous masses quite visible to the naked eye, remain intact indefinitely, if left in the prepared and heated serum. But as soon as we add to them a trace of normal blood (taken from one of a number of species of Vertebrates), the solution of the red corpuscles is not long in taking place. Under these conditions an action of two substances is set up; one of these substances is found in the heated serum of the injected animal, and the other in unheated normal serum. The first of these substances which not only resists a temperature of 55°–56° C., but stands, without alteration, heating to 60°–65° C., corresponds to the intermediary substance of Ehrlich. By Bordet it has been termed “substance sensibilisatrice^[119].” The second substance, a common one, found in normal serums and destroyed at 55°–56° C., is the alexine of Buchner and of Bordet, or the complement of Ehrlich.

The ease with which one can demonstrate the co-operation of two substances in the haemolysis by the serums of animals treated with the blood of a different species, is due to the fact, that during the course of this treatment the animal organism produces a quantity of an intermediary or sensibilising substance. In fresh animals which have not been treated, it is often very difficult to demonstrate the presence of this substance. Bordet has established the fact that the serum of animals which have been injected several times with the blood of a different species, contains almost the same amount of alexine as does untreated serum. On the other hand, the sensibilising substance makes its appearance in large quantity as the result of these injections. Von Dungern^[120] has confirmed this result and contributes the interesting fact that the sensibilising substance is found even in great excess in the serum of treated animals. When he adds to this serum blood that has not been heated, he produces a haemolysis that is more than thirty times as active as when the serum of the prepared animal alone is used. From the quantitative point of view, then, there is no relation between the amount of the two substances in the serum of prepared animals.

It may be suggested that the sensibilising or intermediary substance is the same as that which produces the agglutination of the red corpuscles. But careful researches have thoroughly demonstrated the difference between the two substances that have this character in common, both resist heating to 55°–60° C. and even beyond this point.

Having established this co-operation of two substances in haemolysis the intimate mechanism of their action was next studied. Here I must give pride of place to the discovery by Ehrlich and Morgenroth that the intermediary (or sensibilising) substance links itself to its corresponding red corpuscles. A serum, capable of dissolving the red corpuscles of a different species, is heated to 56° C. which causes it to lose this solvent property. When a certain number of these corpuscles are added to it, such corpuscles remain intact although they are agglutinated. It is sufficient, after some hours of contact, to centrifugalise the mixture in order to separate a limpid serum from the mass of red corpuscles, the former being now entirely deprived of its intermediary substance, that is to say it has become incapable of dissolving the red corpuscles even with the addition of a large quantity of the “complement” (normal serum, unheated). On the other hand, the red corpuscles, having fixed (linked) all the intermediary substance, dissolve very rapidly when placed in contact with normal serum which contains the necessary quantity of the complement (or alexine). This fundamental experiment has been confirmed and repeated by many observers and has now become classic. The idea that the intermediary (or sensibilising) substance links itself to the red corpuscle, without dissolving it, is generally accepted and may be regarded as permanently settled. We should do well, then, instead of designating by all sorts of synonyms the substance in serums which resists the action of a temperature of 55°–65° C., to apply to it, once for all, the name of fixative substance or simply that of fixative. This name is short, expresses the essential character of the substance and gives rise to no misunderstanding, as do the other names proposed up to the present (amongst them that of philocytase employed by myself in some of my earlier publications).

Another of Ehrlich and Morgenroth’s experiments has furnished the proof that the complement does not fix itself to the red corpuscles only. A normal serum, unheated, which, by itself, is quite as incapable of dissolving the red corpuscles as the fixative alone, is mixed with some defibrinated blood. After the centrifugalisation of this mixture, it is easy to demonstrate that the supernatant fluid has lost none of its complement (alexine), whilst the red corpuscles have fixed none.

If, instead of an inactive serum, we take a serum which is capable of dissolving the red corpuscles and which consequently contains the two haemolysing substances, and if we place it in contact with the corresponding red corpuscles, at a temperature between 0° and 3° C., the solution will not take place (Ehrlich and Morgenroth). Under these conditions the fixative certainly attaches itself to the red corpuscles, but the alexine remains in solution, unused. It is only necessary, however, to heat the mixture up to 30° C. to bring about rapid haemolysis.

From their very ingenious experiments, as a whole, Ehrlich and Morgenroth conclude that the fixative possesses two different affinities: one for the red corpuscle and another for the complement. Of these two affinities the stronger is that which links it to the red corpuscle, for this is manifested at a very low temperature. In order that the fixative may combine with the complement a much higher temperature is requisite. Ehrlich comes to the conclusion that the molecule of the fixative possesses two haptophore groups, or groups capable of chemical combination. The first of these links it to a corresponding molecule of the red corpuscle to which he gives the name of receptor; the second combines the fixative with the molecule of the complement and in this way introduces the latter into the red corpuscle. These investigators give a diagram which greatly facilitates the understanding of their hypothesis (Fig. 19). They seek to prove that the combinations of the fixative with the red blood corpuscle and with the complement follow the law of definite multiples and that these phenomena must, in consequence, be looked upon as being of a purely chemical character.

The hypothesis advanced by J. Bordet does not accord very well with the theory we have just set forth. He could never convince himself that the fixative combines with the complement. He was of opinion rather that the fixative, retained by the corpuscle, exercises upon it a mordant action which enables it to absorb the alexine. The alexine is supposed to attach itself to the sensibilised red blood corpuscle as a dye attaches itself to a mordanted element. Bordet rests his interpretation mainly on the fact that the absorption of alexine by the sensibilised corpuscles does not follow the elementary laws of chemical combination, especially those of definite multiples.

Nolf^[121] has sought to define more accurately the part played by these two substances in the solution of the red blood corpuscles. He agrees with Bordet, that in this phenomenon the fixative plays the same part that the mordants do in dyeing. Linked to the red corpuscle the fixative renders it more greedy for alexine, exactly as the mordant facilitates the fixation of the dye on the fibre of the textile fabric. Under these conditions the alexine (complement), finding itself in large quantity inside the red corpuscle, exercises upon it its hydrating action, thus bringing about the diffusion of the haemoglobin and often even the solution of the corpuscular stroma.

Nolf compares the solvent action of alexine upon the red corpuscle to that of certain mineral salts, such as ammonium chloride. He passes in review the various properties of alexines and finds them very similar to the solvent action of certain salts. Even the peculiarity of alexine, of remaining inactive at a temperature of 0°–3° C., is shared by ammonium chloride which, alone of all the salts studied by Nolf, exercises no solvent action under these conditions. But Nolf found it impossible to push these analogies further, and especially to sensibilise, by the fixative, the red corpuscles to the action of quantities, which were of themselves inactive, either of ammonium chloride or of any other salt.

London^[122] hoped by fresh experiments to solve the problem of the mode of action of the two substances which act in haemolysis. He pronounced in favour of the theory that they entered into chemical combination with the red corpuscles. But the facts accumulated up to the present do not enable us to make a positive statement as to the exact nature of the reaction which is set up during the solution of the red blood corpuscles; this is not astonishing in view of the fact that it is impossible to isolate the haemolysing substances in a pure state.

It may, however, be admitted that the action of alexine (complement) comes under the category of phenomena that are produced by soluble ferments. Buchner^[123] maintains that there is an analogy between this substance and the diastases (or enzymes); Bordet^[124], from the appearance of his first publications on haemolysis, has expressed himself in favour of this view. Ehrlich and Morgenroth^[125], in their two first memoirs, very distinctly put forward the same idea. “We shall not deceive ourselves”—they say—“if we attribute to the addiment (syn. complement, or alexine) the character of a digestive ferment.” In one of their last memoirs^[126] they no longer express themselves in so decided a fashion. Nevertheless we are still quite justified in maintaining this proposition. The substance which dissolves the red blood corpuscles of Mammals or a portion only of those of Birds, undoubtedly presents very great analogies to the digestive ferments. As has been mentioned repeatedly, it is very sensitive to the action of heat and is completely destroyed by heating for one hour at 55° C. In this respect it closely resembles the macrocytase of macrophagic organs which also dissolves red corpuscles. As it is the macrophages which ingest and digest the red blood corpuscles in the organism, it is evident that alexine is nothing but the macrocytase which has escaped from the phagocytes during the preparation of the serums.

We know that the leucocytes contain quite a series of soluble ferments of which some are set at liberty after the blood has been drawn from the vessels. It is thus that plasmase, or fibrin-ferment, is set free from the leucocytes to combine with fibrinogen to produce the clot. This is not the only soluble ferment of leucocytic origin. It has been known for some time that in addition to this coagulating ferment the leucocytes contain ferments which are especially digestive or decoagulating. Thus Rossbach^[127] has demonstrated the presence of amylase in the leucocytes of different organs, especially the tonsils. Arthus has confirmed this discovery and Zabolotny^[128] has completed it by his observations on the phenomena which appear in the peritoneal cavity of animals into which wheat flour or starch were injected. He observed that the small granules are quickly ingested by isolated leucocytes, whilst the large granules are surrounded by quite a layer of phagocytes. He agrees with several other writers, that the amylase found in defibrinated blood has its origin in leucocytes.

Leber^[129], in the course of his researches on inflammation, made the observation that the pus of a hypopyon that was absolutely aseptic digests coagulated fibrin at a temperature of 25° C. and liquefies gelatine very readily. Achalme^[130] has confirmed this and has added several other interesting data. He investigated the soluble ferments of pus and directed his attention amongst others to experimental pus, set up by the injection of spirit of turpentine. In addition to amylase and a ferment which liquefies gelatine, Achalme has discovered in pus, saponase (lipase), casease, and a ferment closely allied to trypsin. This last readily digests fibrin and also attacks coagulated white of egg; in the products of this digestion Achalme found peptone but could not always obtain leucin and tyrosin. He never succeeded in demonstrating the presence of sucrase, inulase, emulsin or lactase in pus. On the other hand he found large quantities of oxydase, thus confirming the discovery of Portier^[131] who was the first to demonstrate that these ferments met with in the blood are, in the living animal, found inside leucocytes. By a large number of experiments, carried out on most diverse representatives of the animal kingdom, Portier was able to establish the important fact that the oxydases which are found in many organs or in the fluid of blood withdrawn from the organism really originate in leucocytes as they deteriorate and break up. In this respect, then, they resemble fibrin-ferment very closely.

To complete the list, already considerable, of leucocytic ferments, I must further cite the anticoagulating soluble ferment whose existence in Mammals has been so well demonstrated by Delezenne.

All this evidence encourages us, then, to support the thesis that alexine is one of the numerous intraleucocytic soluble ferments and that it only passes into the fluids as the result of rupture or of damage to the phagocytes. Nolf (l.c.) has recently pronounced against this view; we must therefore examine his arguments closely. In the first place he takes his stand on the analogies between the solution of the red blood corpuscles by the serums and by certain salts. It must not be forgotten, in connection with his theory, that haemolysis is but one example, out of many, of the action of alexines. Of all the formed elements the red corpuscles are the most delicate; they are readily broken up by all sorts of agents (moderate heat, water, salts, etc.). Further, there are numerous other cells (white corpuscles, spermatozoa, and inferior organisms) which resist the action of salts much better, which, nevertheless, are very injuriously affected by the action of the alexines.

Nolf lays special stress on the experiments in which, after keeping red blood corpuscles in prolonged contact with active serums, he has looked in vain for the peptone reaction. He prepared his mixtures in sealed tubes or flasks, and kept them in an incubator at 37° C. for 24–48 hours, or even for weeks. Under these conditions the haemoglobin is transformed into metahaemoglobin, but peptones never appear. Nolf concludes therefrom “with confidence, that the alexines do not exert the slightest peptonising effect on the albuminoids of the corpuscle” (l.c. p. 672).

To this conclusion it must be objected that peptone is not the only product of the digestion of albuminoids by soluble ferments. Under certain conditions the disintegration is carried much further, in others it is arrested at an earlier stage. Thus human urine which contains pepsin, never gives the peptone reaction with fibrin; the digestion of the latter only goes on up to the stage of protalbumose. When, however, the urinary pepsin is fixed on flakes of heated fibrin which are submitted to digestion in acidulated water the digestion proceeds further and gives as final products deuteroalbumose and peptone^[132]. Now, under the conditions in Nolf’s experiments the digestion would be very quickly stopped, because, at the temperature of 37° C., alexine very soon loses its strength. Investigators who have experimented with haemolytic serums know well that, even when kept at a low temperature, alexine may lose its activity within 24 hours.

It has been mentioned above that Nolf sought in vain for a parallel between haemolysis by salts and that by serums, in what relates to the action of the fixative. He was unable to find anything comparable to this action amongst salts, although digestion by soluble ferments offers undoubted analogies. I need only recall further the discovery of enterokynase, the soluble ferment of the digestive juice of the dog, which actively stimulates the action of pancreatic ferments, and especially that of trypsin. The recent researches of Delezenne (communicated to the International Congress of Physiology held at Turin in September 1901) support this conclusion in a very important fashion. As already pointed out in Chapter III the enterokynase of the intestinal juice exerts an action comparable with that of the fixatives of haemolytic serums. Alone, it does not act as a solvent ferment, but when it attaches itself to the fibrin it aids the action of the trypsin in a marked degree. In pancreatic digestion enterokynase plays the part of the fixatives in the solution of red corpuscles.

The analogy between the resorption of formed elements and intestinal digestion extends even beyond this. When we inject, into the peritoneal cavity or under the skin of various animals, blood from a different species, the blood serum of the former becomes haemolytic for the red corpuscles of the latter. The solution of these red corpuscles is effected by the alexine of the serum, whose activity is rendered very great owing to the presence of a quantity of specific fixative. This same fixative appears also in the fluids of animals to whom, instead of injecting blood, we simply give it by the mouth. This fact has been established by Metalnikoff^[133].

Another fact in favour of the close relationship between the fixatives and enterokynase consists in the presence of both in the lymphatic (lymphopoietic) organs. The fixatives which aid the solution of red corpuscles are found specially in the mesenteric glands. Enterokynase, as demonstrated by Delezenne, is found not only in the intestinal juice, but also in Peyer’s patches, the solitary glands, the mesenteric glands, and the leucocytes of exudations and of the blood.

Supported by these various facts we are quite justified in regarding the haemolysing substance of serum as containing two soluble ferments, of which one, alexine, corresponds to trypsin, the other, the fixative, resembling enterokynase. The alexine, whose nature is gradually disclosing itself with more precision, should bear the name of cytase or cell-ferment. The cytase of the macrophagic organs, or macrocytase, comes under this category. According to the researches of Tarassewitch it also acts more actively when there is added to it some of the fixative found in the serum (heated to 56° C.) of prepared animals.

We have said that in the living animal the macrocytase is localised in the phagocytes of the organs and of the blood. Thus, when goose’s blood is injected into the peritoneal cavity of the guinea-pig the red blood corpuscles are digested within the macrophage and not in the fluid of the peritoneal exudation. When, however, the same kind of blood is injected a second or a third time, it is found that a certain number of the red corpuscles become permeable and lose their haemoglobin, which they give up to the fluid of the exudation, and only the membrane and the nucleus remain. These are at once ingested by the macrophages which under these conditions manifest a real excess of activity. Instead of sending out small processes, as they do after the first injection of blood, these phagocytes move about like true Amoebae, sending out broad pseudopodia, and ingest not only the remains of the red corpuscles but also those still intact^[134] (Fig. 20). Under these conditions macrocytase must undoubtedly be found in the peritoneal plasma. It is, however, easily demonstrable that this ferment was not preformed in the fluid but has escaped from the leucocytes that have undergone phagolysis. After the rapid injection of alien blood the phagocytes of the peritoneal lymph gather into clumps, become immobile, and for a time lose their phagocytic power. It is only after the lapse of a longer or shorter period that the leucocytes recover from the phagolysis, arrive in great numbers in the peritoneal cavity and display their phagocytic energy.

If the damage to the phagocytes—the phagolysis—is the actual cause of the setting free of the intraleucocytic ferment, we have only to prevent this phagolysis in order to inhibit the solution of red blood corpuscles in the fluid of the exudation. For this purpose it is sufficient to prepare guinea-pigs (which have already received several injections of goose’s blood) by means of an injection of fresh broth, of physiological salt solution, or of carbonic acid into the peritoneal cavity on the eve of the decisive experiment. Such injection at once provokes phagolysis, which is then followed by an abundant exudation of leucocytes. When, next day, a dose of red blood corpuscles of the goose (deprived of serum by centrifugalising) is introduced into the peritoneal cavity thus prepared phagolysis is no longer produced, or very feebly, and is of very short duration. Under these conditions the solution of the red corpuscles by the peritoneal fluid is reduced to a minimum, and in its place an extremely rapid and considerable ingestion of red corpuscles by the macrophages may be observed. In order that the experiment may be completely successful it is advisable to use goose’s blood heated to 37° C. or thereabouts for the injection.

Even when the red corpuscles of the goose are introduced, not into the peritoneal cavity but into the subcutaneous tissue of guinea-pigs that have received several injections of goose’s blood, we can easily prevent the extracellular solution of the red corpuscles which takes place, as already indicated, in the normal guinea-pig. As in this case the goose’s serum which is mixed with the corpuscles contributes to the haemolysis, it must be suppressed by centrifugalising the defibrinated goose’s blood and by washing the corpuscles with normal saline solution.

Collectively, the facts I have just described clearly indicate that the phagocytes must be regarded as the source of the haemolytic ferment. The macrocytase remains in the body of these cells so long as they are in a normal condition; but immediately they are injured, in consequence of the sudden introduction of foreign substances into the peritoneal cavity, a portion of the macrocytase escapes and acts on the red corpuscles as if it had been employed in vitro.

As the conclusion I have just formulated is of fundamental importance in the study of resorption and immunity it is necessary to support it by as many arguments as possible. For this reason, therefore, I feel obliged to draw the attention of the reader to another example of the resorption of formed elements.

We have already spoken of the resorption of spermatozoa in the peritoneal cavity, and of the part played by the macrophages in this phenomenon. As a result of this resorption, just as after that of red blood corpuscles, the organism acquires new properties of the same character. Landsteiner^[135] and the writer^[136] have shown that the blood serum and the peritoneal fluid of animals that have been injected with the spermatic fluid of bull, rabbit, or man, become spermotoxic, that is to say, they render the corresponding spermatozoa motionless and kill them. These fluids, however, never acquire the power of dissolving, even partially, these elements. The disappearance and final solution of the spermatozoa is only effected within phagocytes, and almost exclusively in the macrophages.

Moxter^[137] has demonstrated that the spermotoxin which appears in the serum of prepared animals consists of two substances, corresponding to those present in the haemolytic serums. These are the macrocytase (alexine, complement) and the fixative (intermediary or sensibilising substance). For him they are identical with those which dissolve the red corpuscles. Without dwelling on the subject we may say that the macrocytase which dissolves the red corpuscles and that which arrests the motion of the spermatozoa are really identical in the same species of animal, as is accepted and developed by Bordet. On the other hand, it is impossible to accept Moxter’s theory of the identity of the two fixatives. They must be regarded as different; this we have attempted to prove in one of our memoirs^[138] and is in accordance with the law of the specificity of fixatives in general.

The question which interests us more especially at this moment is where are these two constituent substances of the spermotoxin to be found and how do they behave in the living organism? This question has been very thoroughly studied by Metalnikoff^[139] in my laboratory. His experiments have been closely followed by me, and in presenting their principal results I can bear witness to their correctness.

The spermotoxin obtained by Metalnikoff is distinguished from the haemotoxins we have discussed up to the present in that they develop, not as a result of the injection of cell elements from a different species, but as a result of the introduction into the organism of spermatozoa from the same species, the guinea-pig. We have here, then, to deal with what has been termed autospermotoxin.

The serum of the normal guinea-pig acts but feebly on the spermatozoa of this species, which, under its influence, remain motile for several hours. When, however, guinea-pigs have received one or several injections of the spermatozoa of their own species, their serum and peritoneal lymph become distinctly toxic and render the spermatozoa motionless in a few minutes. In male guinea-pigs so prepared the serum acquires this toxic property not only for the spermatozoa of other male guinea-pigs, but likewise for those of the individual itself which furnishes the serum. This latter, then, becomes distinctly autospermotoxic.