It is best to begin with antibodies which conform to their original definition as new serological properties attributable to an antigenic stimulus.

Their origin is still disputed. They are often regarded as cellular secretions due to adsorption or ingestion of antigen. For my part, I consider that they are formed in the plasma by filtration through capillary endothelium which has adsorbed antigen.

The precise specificity which characterises many antibodies is difficult to explain. I have suggested that it may be due to an interplay of reactions between the dextro- and laevo-rotatory forms of particular combining affinities.

But antibodies are not always the exact counterparts of their antigens. I have discussed some of the possible reasons why animal idiosyncrasies may be responsible for irregular results. A defective yield may also be due to the antigen, which may be present in a masked condition.

It is often taken for granted that the characters of the antibodies found in the serum of an actively immunised animal are identical with those actually present in the animal's circulation. I have given reasons for thinking that this assumption is frequently erroneous.

Coming now to the more elastic conceptions of “antibodies,” which relate to properties not attributable to an antigenic stimulus, the part played by alexin in the circulation of the normal animal has first to be considered. It is not a distinctive chemical entity, nor is it simply a chemico-physical condition of the plasma; it is a complex of chemical and physical factors which defy analysis. The dual mechanism represented in vitro between alexin and “immune body” does not afford a true picture of alexin's activities in vivo.

The humoral factors in natural resistance are too complex to be regarded as a dual mechanism. The difference between susceptibility and resistance often depends on that precise sequence of events in the plasma's activities which is peculiar to the species or even to the individual.

Acquired immunity is a reconstruction of the normal mechanism, with the emergence of a definitely new property. It is not simply the introduction of a new antibody but involves also utilisation, with readjustment, of the complex biological conditions normally present. As a result of the antigenic stimulus: (1) antibodies may be formed which are the precise mirrors of their antigens; (2) by a similar but less precise mechanism, antibodies may arise which are not the exact counterpart of any particular antigen; (3) the antigenic stimulus may lead to complex humoral changes which are not identifiable as antibodies.

Current opinions on immunological principles, as expressed by well known authorities, exhibit divergencies which cannot be satisfactorily reconciled with each other. There is, however, general recognition that the present status of knowledge is highly deficient. Though the time is not yet ripe for anything like a finally satisfactory reconstruction of these principles, efforts in this direction are needed, with readjustment of the old ideas by the aid of hypotheses about what is going on in the living body.

The characters by which the “R” variant of a given bacterial type is distinguished from the “S” variant are to-day widely known. Most of them are morphological and in most bacterial types it is not difficult to say whether a single colony is smooth or rough; but every bacteriologist is familiar with those borderline colonies, which are apparently smooth, but which later prove to be serologically rough. It sometimes happens that by special stimulation the reversion from an “R” culture into the “S” form succeeds, but that, notwithstanding the smooth appearance of the isolated colonies, the organisms persist in showing some characters which belong to the “R” variant, such as instability in saline (Bruce White, 1926).

In our investigation of our subject we have perused the literature, analysed the data of other workers, conducted a number of animal experiments, considered the bearings clinical and pathological of a case of presumed toluene poisoning, and conducted an investigation into the blood state and symptoms of co-workers with this case and yet are unable to give a categorical answer to the question. What are the relative toxicities of toluene and benzene? We would almost feel inclined to echo the words of another worker in the experimental field with regard to benzene, Neumann (1915), who concludes: “The fact that my results so often do not correspond either with one another or with those of Selling, so often indeed do not correspond to expectation, justifies the conclusion that, although benzene is to be regarded as a powerful leucotoxin, it manifests great individual differences in its mode of action and it is very difficult to lay down exactly what action is to be expected.”

We should be inclined to put the case somewhat differently however and put forward the following view as our tentative finding.

The actions of benzene and toluene are very similar. That of benzene is the more powerful because it is a more volatile substance and can therefore reach higher concentration than toluene. This greater volatility at the same time is a reason why it is eliminated more quickly and its effects are less lasting than those produced by toluene. The action of both these substances is on young body cells, probably whatever they are and wherever they may be found, but is most easily discoverable because of its action on the young bone marrow cell and the consequent effect on the cells of the peripheral blood. It is not a specific action but operates, according to circumstances, on the stem cells of the myeloid leucocyte or the erythrocyte. This does not mean that the substances have little or no effect on the lymphocyte, for the lymphoid tissue in the body is so very abundant that it would only be in extreme cases that the effect on that cell would be manifest peripherally. Like many poisons, indeed like many of the drugs which are used in Medicine, the first effects of these substances are stimulating. Leucocytosis, erythrocytosis and possibly thrombocytosis are early manifestations or the manifestation of slighter action, with histological evidence of hyperplasia. This is succeeded by the truly toxic or destructive action and the production of leucopenia, anaemia, thrombopenia and the effects due to action on other organs. The ultimate result in clinical terms is an agranulocytic anaemia accompanied by purpura haemorrhagica and, as the pathological manifestation an aplastic bone marrow. The fatal termination is commonly an infection to which there is no resistance possible in an individual deprived of defence cells and possibly also of defensive substances in the body fluids. Powerful solvents of fat as both these chemical substances are, they produce an action through this means also, on nerve cells, capillary endothelium and parenchyma of organs with consequences represented in nervous symptoms and in haemorrhages. This then is our reading of the results of our investigation and explanation of the variability which may be found, a variability which is regarded mainly as one of phase, although we do not exclude idiosyncrasy altogether as explanation.

1. Experience of 4½ years of intravenously administered scarlatinal antitoxin, in 1204 cases, confirms the tentative conclusions reached in 1928, viz. that the acute stage of the disease can generally be arrested in 12–24 hours, complications almost wholly prevented, and the period of morbidity reduced to a little over a fortnight.

2. These effects were obtained in hospital wards where over 50 per cent, of the cases, generally about 75 per cent, and including all the worst, were so treated, and it is possible that the reduction of mass infection obtained in the wards by this means is an essential condition for obtaining such results. Under such conditions relapses and cross-infections were very rare.

3. Some evidence was obtained that intravenous antitoxin administered in the acute stage of scarlet fever largely prevents the onset of septic complications, including the late septic type itself, scarlatina anginosa; and that it may be of striking benefit in the treatment of the septic type, even at a late stage.

4. The importance of proved clinical potency and freedom from dangerous reactions of each individual batch of serum to be employed, is stressed.

5. Indications and contra-indications are stated.

6. Certain precautions and details of technique are recommended.

1. An agar medium is described which enables a direct estimation of the numbers of coliform bacilli in a water or milk sample to be made.

2. An enrichment medium for B. lactis aerogenes is one containing in every 100 c.c. of peptone water 0·5 g. sodium citrate, 0·25 g. soluble starch and 1 c.c. of a 1 per cent, watery solution of neutral red.

3. Enrichment media for B. coli are: (a) lactose peptone water containing in every 100 c.c., 1 g. sodium sulphite anhydrous and 2 c.c. of 1 per cent, rosolic acid in absolute alcohol; (b) lactose peptone water containing in every 100 c.c., 2 c.c. of a 40 per cent, solution of hexamine in water.

4. Importance is attached to the capacity to ferment starch possessed by B. lactis aerogenes.

5. The application of these media to the examination of water and milk is indicated.

Hecht (1923) states that the haemolytic activity of serum complement disappears on standing, but that the activity can be restored by the addition of heated fresh serum. It has also been shown (Gordon and Wormall, 1928; Gordon, 1930;. Gordon and Carter, 1932) that the bactericidal action of serum involves the activity of a complement similar to haemolytic complement and that inactivation of the latter by dilute ammonia and congo red causes, at the same time, inactivation of the bactericidal complement. The experiments now to be described show that the similarity of the two complements is exhibited once again in the effect of long standing on serum.

In 1890 Klein described two types of colony of the diphtheria bacillus on gelatin; one smooth and circular, the other radiating and resembling a daisy. Cowan (1927) observed two types of colony, but one of the types was thought to be due to roughening. Parker (1928) observed three types of colony, some of which were considered ordinary rough variants.

1. The annual number of notified cases of erysipelas is approximately 17,000, and assuming complete notification of the disease the general case fatality is approximately 6 per cent.

2. The death-rate, deaths in terms of the population, varies according to age, being highest at the beginning and end of life and at a minimum between the age of 5 and 10 years. The mortality of males is identical with that of females up to age 25, but is afterwards in excess.

3. The disease has in recent years a well-marked seasonal incidence—a winter and spring excess with a summer defect. In this respect it resembles scarlet and puerperal fevers and, although its seasonal incidence has changed in the course of time, the alteration has not been nearly so pronounced as that for scarlet fever.

4. The incidence is highly correlated with overcrowded conditions, the correlation coefficient being in some periods as high as + 0·83 in the divisions of London and + 0·70 in the sanitary districts of Glasgow.

5. The morbidity from erysipelas is fairly well correlated in time with that from scarlet fever and erysipelas, but in London, in the urban and in the rural districts of England and Wales, the spacial correlation is very small. On the other hand, in Glasgow the spacial correlation between the incidence of erysipelas and that of scarlet fever is highly negative, −0·718 ±0·109, whereas between erysipelas and puerperal fever the association is positive, + 0·689±0·109 during the period 1903–8, but these values were much reduced when the partial coefficients were calculated.