In 1927 Hilda M. Woods and I contributed to this Journal a study of some aspects of a syndrome called “Status Thymicus,” “Status Thymico-Lymphaticus” or simply “Status Lymphaticus.” We were moved by three considerations: (1) that Prof. H. M. Turnbull had placed at our disposal a long series of weighings of the Thymus, the alleged principal villain of the piece; (2) that every year a not inconsiderable number of deaths are put to the account of this syndrome; (3) that the first volume of Prof. J. A. Hammar's monograph on the Thymus had recently been published. Since our publication, the annual quota of victims shows no signs of diminution (Table I) and the second volume of Prof. Hammar's work has appeared.

1. Farr's 63 healthy districts are, as a whole, representative of the stationary districts of England and Wales, i.e. those districts in which growth or decline of industry has, on the whole, been absent and in which mortality rates are consequently free from the influence of industrialisation, though not necessarily unaffected by urbanisation.

2. These healthy districts maintained, until 1901–10, the advantage they held over England and Wales as regards their relative death rates in 1851–60.

3. From 1851–60 to 1901–10 the death rate of the healthy districts remained roughly constant at 76 per cent, of the death rate of England and Wales. Both improved their total death rate to the extent of 30 per cent, of what it was in 1851–60, the rates for this decennium being—England and Wales 21·17, healthy districts 16·13 per 1000.

4. The population of the healthy districts has been unfavourably constituted for a low death rate throughout the period.

5. That the periods during which the most improvement was made in lowering the death rate—not only in the healthy districts, but in England and Wales as a whole—were 1881—90 and 1901–25.

6. By 1921–5 the position of the healthy districts had become a little less favourable, their death rate having risen from 76 to 86 per cent, of that of England and Wales. It is quite possible, however, that this is due, not so much to a falling-off in the rate of improvement in the healthy districts, as to an exceptional increased improvement in backward, very unhealthy districts.

7. For the following diseases, the healthy districts show improvement at faster rates, of varying degrees, than England and Wales as a whole, during the period 1851–1925. Measles, scarlet fever, whooping cough, diphtheria, pulmonary tuberculosis, and respiratory diseases.

On the other hand, the data relating to diarrhoea seems to indicate less improvement, though inherent deficiencies in the data make this a matter open to doubt.

The death rate from cancer, which has increased considerably in the whole country during the last forty years, appears to have increased at a slightly faster rate in the healthy districts.

But, as discussed in the report, the untrustworthiness of the data, relating to comparisons of individual disease death rates over long periods of time, make it essential to regard the points enumerated in conclusion 7 rather in the manner of interesting possibilities than as proven facts.

8. When the 141 administrative areas, which in 1920–5 corresponded to Dr Farr's original healthy districts, were classified as to whether the majority of their occupied persons worked in other districts or in their own district, the standardised death rates found were as follows:

I. Districts in which 50 per cent, of occupied persons work in other districts. Death rate = 10·55 per 1000.

II. Districts in which more than 50 per cent, of occupied persons work within the district. Death rate = 9·90 per 1000.

When the second of these classes is sub-divided according to whether 50–75 per cent, or over 75 per cent, of occupied persons work within the district, the death rates are:

(a) 50–75 per cent, of occupied persons working within the district. Death rate = 9·69 per 1000.

(b) 75–100 per cent, of occupied persons working within the district. Death rate = 10·14 per 1000.

When the same class is sub-divided according to whether the majority of workers are engaged in non-agricultural or agricultural pursuits, the death rates in each sub-class are:

(a) Mainly non-agricultural. Death rate = 9·85 per 1000.

(b) Mainly agricultural. Death rate = 9·98 per 1000.

As diverse opinions are held about the significance of systemic influences in cancer, the subject needs some reconsideration. What are “systemic influences”? In the literature on cancer this preliminary question is usually ignored, presumably because it is thought that the answer is self-evident. With this view I do not agree. I think one must begin by forming a general conception about the nature of systemic influences. What are they like in the normal body? What is their position in bacteriological immunity, about which knowledge is more advanced than in cancer? Ideas derived from a discussion of these two questions ought to provide a useful base for approaching the problem as it concerns malignancy.

In the normal body the systemic influences which the plasma exerts upon the tissues form a complex system presenting three aspects, the chemical, the chemico-physical, and the vitalistic. For example, sometimes it may be said that the plasma's activity is due to a special chemical substance, such as a hormone; sometimes the predominant factor is due to the balance of its colloidal constituents; and not infrequently its action can only be attributed to those properties of living matter which cannot be reproduced in the chemical or the physical laboratory. These three aspects of systemic influences are not independent factors but have to be correlated; and the essential difficulty of the subject is to assign to each of them its appropriate significance.

In natural immunity and resistance towards bacteria, these normal systemic influences are in possession of the field and it is upon their activities that the fate of the bacterial intruders largely depends. Where the immunity is non-specific, as in the inability of saprophytes to grow in living tissues, the defensive factor bears a prominently vitalistic aspect. The mechanism of bacterial destruction seems largely to depend on the circumstance that the bacteria find themselves in a living animal environment where they cannot remain in the resting stage; they must endeavour to grow but they perish in the attempt because the medium is unsuitable.

What is the nature of “alexin” as a natural defensive mechanism? The idea that it is a special chemical substance secreted by some cells of the body must be abandoned. In vitro, it is a property due to the chemico-physical lability and colloidal complexity of fresh serum, in virtue of which the serum promotes interactions which would not take place in a more stable medium. In vivo, the plasma possesses similar chemico-physical properties in a more complex and more effective form, supplemented by its vitalistic capacities as living material. For these properties of the circulating plasma the term “alexin” is not appropriate.

As regards specific manifestations of natural immunity, how is one to explain the selective action of normal systemic influences on bacteria which are pathogenic for some species of animals but not for others? Selection naturally suggests special chemical attributes of the plasma; but species immunity has not been identified with the presence of distinctive chemical substances and it is not likely that it ever will be. One has to fall back on the chemico-physical attributes of the plasma which constitute its general “make up,” as characteristic of a particular species. And these attributes must be regarded not as a system in stable equilibrium but as a dynamic system involving an ordered sequence of reactions.

The most important feature of true natural immunity is that, when the bacteria have been disposed of, the condition of the plasma remains as it was before their intrusion. Its activities have not been due to antibodies, in the accepted serological sense, and the destroyed bacteria have not behaved as antigens.

In most of the literature on acquired immunity the chemical conception stands out very conspicuously. Bacterial protein behaves as an antigen and stimulates certain cells of the host to secrete an antibody; that is regarded as the basis of immunity. After allowing for the operation of chemico-physical laws, the predominant feature remains that an immunological reaction is essentially the interplay between two chemical entities, an antibody (agglutinin, lysin, tropin, etc.), and its corresponding antigen. This conception is considered preferable to the much less concrete ideas of interactions between systemic influences and living bacterial protoplasm.

Whilst appreciating the value of precise chemical data, I consider that this view of acquired immunity is one-sided and inadequate. Systemic influences (other than serological antibodies) cannot be left out of account in the conception of interactions between living bacteria and the living animal body. One needs a scheme which will help to correlate natural with acquired systemic influences, to bridge the gap between specific and non-specific factors, and to modify the conception of an antibody as a special chemical entity, specially secreted by certain cells in response to the stimulus of a foreign protein. Within such a scheme, as I have endeavoured to show, an explanation may be found for what may be called the routine production of antibodies by antigens.

Coming now to cancer, one must first insist on the commonsense view that the transplantation of grafts is a special and relatively unimportant line of experiment which, whatever interests it may possess in other respects, does not help to explain either established autogenous cancer or the genesis of cancer. In these grafting experiments certain systemic influences emerge which cannot be explained as due to the production of antibodies by antigens. This is to be expected, on the analogy of similar manifestations of antibacterial systemic influences. But neither natural nor acquired systemic resistance to the taking of a graft involves anything which may be regarded as a new kind of systemic influence peculiar to cancer.

In the case of established autogenous cancer there does seem to be a new kind of systemic influence which is directly attributable to the disease. This influence, as is found by animal experiment and by observation on human malignancy, inhibits, or tends to inhibit, the creation of a second and independent malignant growth in the same animal body. Apparently products of the existing cancer pass into the circulation and cause other tissues to lose their susceptibility to influences which might ultimately have produced a malignant variant. The mechanism of this inhibitory action is obscure and is probably more complex than the chemical influence of a particular cancerous product upon normal cells. Whatever may be the right explanation, the observed facts indicate that it is something new, which is created by the cancerous condition; they afford no proof whatever that, before the cancer existed, there were in the circulation special systemic influences which were favourable or unfavourable to the genesis of cancer.

The idea that there are systemic influences concerned with the genesis of cancer has assumed many forms and is often expressed ambiguously. Does it mean that normal cells have a “natural tendency” to malignancy and will actually become malignant if freed from systemic control? I do not accept this “natural tendency”; unrestrained growth does not suffice to explain the origin of cancer. What is meant by “systemic control”? My view is that such control regulates normal cells and that cancer cells are independent of it; I do not agree that there is a special kind of antimalignant systemic control which may destroy the fully fledged cancer cell. What is the nature of “susceptibility” to the change into the cancerous condition? I regard it as essentially a cellular property, not as a humoral or systemic influence, though I admit that irritant material which gains access to the circulation may increase the susceptibility of particular cells. What is meant by “resistance” (either local or systemic) to cancer? Owing to the recuperative powers of the animal body, local disturbances of metabolism are often corrected and there is a return to the normal condition; some of these disturbances, if left uncorrected, might have led to cancer and the fact that they have been corrected may, if one likes, be called resistance to the genesis of cancer. It is also known that true cancerous foci or metastases may remain quiescent for a considerable time. But I do not agree that such quiescence has been shown to be attributable to a specific kind of antimalignant “resistance” (either local or systemic).

Whilst there is no satisfactory evidence, either direct or indirect, of a systemic influence which causes cancer, systemic influences are so complex and obscure that this possibility cannot be definitely excluded. But there does not seem to be any cogent reason for dissenting from the view that the production of the malignant variant is due to its local environment.

In a previous communication the authors drew attention to the occurrence of complement-fixation by normal serum along with various non-antigenic substances, and recorded the results of a study of this phenomenon with particular reference to the serum principle or principles concerned in the reaction (Mackie and Finkelstein, 1928). Though no conclusive interpretation of the results could be offered, a close analogy was established between the reacting agent of the serum and a natural antibody, e.g. a natural haemolysin, and the observations appeared to be of significance in relation to natural immunity phenomena.

The maintenance of a constant supply of healthy guinea-pigs is an important part of the work of laboratories which are engaged in the production of diphtheria and tetanus antitoxin. The subacute infective processes to which malnutrition predisposes enhance the effect of the test dose of toxin, with the result that irregular deaths among the animals under test render difficult the titration of the antitoxin. For this reason it is desirable that breeding stocks should be kept in suitable animal houses under the best hygienic conditions possible. This policy has been followed in the Serum Department of the Institute for many years; there have been comparatively few introductions of stock from outside sources, and within recent years special attention has been devoted to diet.

Comparatively little work has been published upon the bacteriological investigation of the urine in the healthy woman either normally or during pregnancy and the puerperium, and the findings in such work as has been published are contradictory. In 1912, Leith Murray, Stenhouse Williams and A. J. Wallace, working together upon the coliform organisms in the female urinary tract, came to the conclusion that typical Bacillus coli is found in a considerable percentage of all female urines taken under conditions precluding all sources of contamination, and that ordinarily they have no apparent pathological significance. John Hewitt, in 1923, published a report upon 34 cases of urinary infection during pregnancy and the puerperium. He agreed with Leith Murray and other workers who had quoted figures to show that healthy pregnant women may have B. coli in the urine in the absence of pus cells.

Young normal rats are put on vitamin A-free diet and, after development of pronounced xerophthalmic symptoms, inoculated by mouth and by subcutaneous injection with Breslau bacilli (paratyphoid). The course of infection and bacteriological autopsy findings show a marked decrease in the resistance of these animals to this infection as compared with the findings in rats on adequate diet.

The experiments do not show any change in the mechanism of infection in pronounced vitamin A deficiency.

1. An outbreak is described of fatal epidemic enteritis due to B. dysenteriae Sonne which occurred in the autumn, 1928, in the infant wards of a hospital in Glasgow.

2. The clinical histories of two fatal cases are given.

3. The bacteriological and immunological examination of two strains of B. dysenteriae Sonne is described.

4. The results are given of a series of animal inoculations undertaken to study the virulence of B. dysenteriae Sonne and the nature of the lesions which it produces. To these are added the results of certain feeding experiments.

1. The average number of guinea-pigs born in albino litters is 3·000 as against 3·097 in the case of the mixed group, but the difference between the mean values is not statistically significant. Hence, it will be seen that there is little or no relationship between the colour of guinea-pigs and the number of their progeny.

2. It may be said that there is a relationship between fertility and the period of the year. There is a tendency for fewer litters to be born during the quarter, January to March, and likewise for the fertility to be lowest during this period, as the mean number of births per litter is 2·65 for albino guineapigs and 2·67 for the mixed class, both values being significantly below the mean for the whole period.

3. The mean weight of the albino guinea-pigs at birth is 81·2 ± 0·36 grm., the corresponding value for the cream, cream and white class is 82·6 ± 0·43, but the difference is of no statistical importance. Hence we conclude that the weight of a guinea-pig at birth is not affected by its colour. Once again attention is centred on the January–March quarter as the most unfavourable period, since there is a tendency for guinea-pigs of either colour born in these months to be below the normal weight.

4. When allowance was made for the effects of selection on our data, there was no material difference between the rates of growth for the two types of guinea-pigs and, furthermore, the period of the year at which littering occurred exercised no apparent influence.

5. The rate of mortality during the first thirteen days of life amongst albino guinea-pigs is 5·58 per 1000 per day, and amongst cream, cream and white guinea-pigs 4·73 per 1000 per day, but the difference probably represents nothing fundamental because, when the mortality is studied according to the size of the litter, the rates are sometimes in the reverse direction.

6. Finally, there is, in the present data, nothing to suggest that albino guinea-pigs are as regards fertility, growth and mortality, significantly different from cream, cream and white guinea-pigs.

The results are collected in Table I, and it may be concluded that:

1. Under ordinary conditions in the south of England B. tuberculosis may remain alive and virulent in cow's faeces exposed on pasture land for at least 5 months during winter, for 2 months during spring and for 4 months during autumn. In summer no living organisms were demonstrated after 2 months.

2. Under special conditions, e.g. protection from direct sunlight, the survival period may be 4 months during summer. In autumn faeces protected from earthworms, etc., yielded bacilli after 6 months.

Three cases of continued fever are described in which the Salmonella type “Dublin” was isolated from the blood. The cultural and serological behaviour of the three strains so obtained are compared with those of 34 strains of typical S. enteritidis (Gärtner) (including strains from commercial rat virus) and with those of 12 other strains of S. dublin (including six from calf-dysentery in Denmark). The suggestion is made that the “Dublin” type is of bovine habitat and that cows' milk is the vehicle of human infection.

1. African pouched rats (Cricetomys gambianus) are very suitable animals for plague experiments on account of their tolerance of captivity and their extreme susceptibility. After inoculation with small doses of plague culture of ordinary virulence they always die.

2. Broth-grown vaccine was much more efficient as a prophylactic than agar-grown vaccine. When two doses were given a survival rate of 56 per cent, was obtained with the former and 25 per cent, with the latter.

3. A vaccine composed of carbolised spleen pulp from animals, which had died of acute plague, was rather more efficient than broth vaccine, giving a survival rate of 75 per cent. But it is not likely to be of practical value, as preparation is difficult, there are undesirable local effects and its superiority over broth vaccine is not very great. Its efficiency is attributed to the fact that the organisms which it contains are a true “body strain.” The method of preparation is similar to that of Hindle's yellow fever prophylactic.

4. A number of experiments are described in which virulence was abolished or reduced by passage through immune or partially immune rats. This occurred only when the passage animal showed a considerable degree of immunity, as indicated by the length of the survival period after the test infection. Cultures made from the abscess at the site of infection were less often attenuated than those from the liver of an immunised animal, but this was observed only in the case of large well-defined abscesses. It is suggested that a firm abscess wall protects organisms inside it against the action of antibodies appearing in the general circulation.

A culture attenuated by passage through an immunised rat was restored to normal virulence by successive passages through normal rats, but the process was slower than in the case of cultures attenuated by long cultivation, five passages being required in the former case and only one in the latter.

5. Cultures used in the later work, unlike those used in the earlier work, lost virulence rapidly when kept at room temperature. Experiments and observations are described which suggest that old strains, even when subjected to rat passage, are less stable as regards virulence than those recently isolated from human epidemic cases, and that the high temperature of the Gold Coast has an adverse effect upon virulence.

6. Attempts to isolate stable variants of B. pestis were not successful. Colonies vary much in form. Three types of colony are mentioned, but all intermediate grades were observed. Certain old avirulent strains had a greater tendency to produce large irregular colonies than recent strains, but the relation was so loose that the form of colony gave little indication of the degree of virulence. In old avirulent strains the bacilli tend to become longer and filaments of great length may be present. Salt stability tends to become less as a strain grows old and avirulent, but the change is not very marked and is apt to be masked by individual fluctuations.

Agglutinins and precipitins. Cultures of S. aureus isolated from infections in man were used for the inoculations. It is well known that rabbits are very susceptible to this organism, and previous inoculation with vaccines, attenuated cultures, or with carefully graduated doses of the live coccus does not confer adequate protective immunity, although the animal's serum contains immune bodies, such as agglutinins and precipitins, to a high titre.

In 1929, for the third time in the present century, cerebrospinal fever occurred in Glasgow in epidemic form. The first epidemic, which was by far the most severe, was in 1906–8, the total number of cases recorded being 1363. Descriptions of the events of these years were given by Dr A. K. Chalmers (1906) in the Annual Reports of the Medical Officer of Health of Glasgow for 1906 and 1907, and also in a paper read before the Epidemiological Society. In this outbreak the fatality rate of cases treated in hospital was given as 74·8 per cent. by Currie and Macgregor (1908). The total number of cases admitted to hospital was 910. The incidence of the disease was highest during the months of January to May of 1907, and there was a slight recrudescence in the same months of the following year. The characteristics of the epidemic form of cerebro-spinal fever were present in so far as there occurred several instances of multiple infection in the same house, and a number of cases showed the exanthem peculiar to the disease.

1. Analysis of a series of 135 strains of B. aertrycke indicates that at least four well-defined types may be recognised on the basis of morphology, colonial formation, virulence to mice, and to a less extent agglutination by salt, serum, and acids. These types comprise: (i) type A—the normal smooth virulent form; (ii) type B—a smooth form of low virulence; (iii) type C—a roughish, so-called ichthyotic, form of low virulence; (iv) type D—the really rough avirulent form. Between these well-defined types certain transitional forms may sometimes be recognised, the virulence of which is frequently of an order intermediate between that of the A type on the one hand, and the B, C, or D types on the other.

2. Evidence is brought to show that if a virulent strain is sub-cultured daily, under conditions known to lead to a fall in virulence of the whole culture, a gradual replacement occurs of the virulent type A bacilli by the avirulent B, C or D types. So long as an adequate proportion, which is very small, of type A bacilli persists in the culture the virulence of the whole culture remains fairly high; but when all type A bacilli have been replaced by bacilli of the B, C or D types the virulence of the whole culture falls considerably.

Good results were obtained by Ferry, Pryer and Fisher (1925) in the treatment of cases of scarlatina with concentrated antitoxin, obtained from horses treated with toxin of scarlatina derived from streptococci. This antitoxin was supplied commercially by Parke, Davis and Co. It has been used in the present investigation.

1. A given strain of Bact. aertrycke was sub-cultured daily, except on Sundays, in 5–7 c.c. of casein broth, pH 7·4–7·6, under different partial pressures of oxygen. Most experiments were continued till about ninety sub cultures had been made. Virulence determinations were made from time to time by intraperitoneal inoculation of twenty mice with about 100 living organisms, the control and the experimental cultures generally being tested simultaneously. The results showed that under anaerobic conditions there was no appreciable change in virulence. Under the usual aerobic conditions, and in partial pressures of oxygen varying from 1 to 21 per cent., the virulence declined, the extent and rapidity of the fall increasing with the pressure of oxygen supplied. In high partial pressures of oxygen, varying from 40 to 100 per cent., there was a slight increase in virulence, the increase being greatest in cultures incubated in pure oxygen.

2. Some of the literature dealing with incitants to microbic dissociation is discussed, particularly that dealing with variations in oxygen pressure.

3. In explanation of the experimental results obtained, a tentative hypothesis is put forward with reference to the dissociation of Bact. aertrycke. It is suggested that under aerobic conditions, and under low partial pressures of oxygen generally, substances are formed in the medium, as the result of growth, which favour dissociation, and lead to the appearance of weakly virulent or completely avirulent variants. Under anaerobic conditions these substances either are not formed at all, or appear in amounts too small to be of significance. In cultures incubated in high partial pressures of oxygen, these substances are formed, probably in large quantity, but owing to the high alkalinity of the medium and the abundant oxygen present, they undergo rapid destruction. This explanation is based on the fact that many protein degradation products, particularly those with a phenolic grouping, prove markedly unstable in an alkaline medium in the presence of free oxygen.

As an alternative hypothesis it is suggested that the maintenance of the organisms in the virulent state depends on the intracellular pH. Under anaerobic conditions it is supposed that the metabolism is such as to lead to no marked change in the H-ion concentration within the cell. In the presence of air and of low pressures of oxygen generally, owing to the breakdown of the protein in the medium and the production of ammonia, the intracellular reaction becomes more alkaline; whereas in high pressures of oxygen the increased amount of CO2 produced is sufficient to compensate for the production of ammonia, and the intracellular pH remains more or less at its original value. This hypothesis rests on the observation of Stephenson and Whetham (1924) that in high pressures of oxygen Bact. coli may produce considerably larger quantities of CO2 from lactic acid than under ordinary aerobic conditions; and on the demonstration by Jacobs (1920 b) that, owing to the ease with which CO2 is able to penetrate the living cell, the intracellular pH may differ markedly from that of the medium to which the cell is exposed.

4. It is of interest to note that Bact. aertrycke, when incubated under partial pressures of 40 to 75 per cent. of oxygen, rapidly passed into the nonmotile O form.

1. A study has been made of natural agglutination as exemplified by the reactions of the serum of nine animal species with a variety of bacteria.

2. End-titres are recorded, and the fact is noted that sera of different animal species show an order of agglutinating activity which is almost constant for all organisms used. Ox, pig and horse sera give consistently strong reactions, while specimens from rabbit, guinea-pig and rat react weakly or not at all. Sheep, human and cat sera occupy an intermediate position. Variations are noted, however, with different individual specimens of serum from the same species.

3. Organisms of the series tested can also be grouped in order according to their apparent susceptibility to agglutination by normal sera.

4. The serum of young animals is found to be deficient in the agglutinating principle.

5. The agglutinating effect shows a thermolability intermediate between that of complement and the immune agglutinins. Complete inactivation occurs as a rule after exposure to 60° C.–65° C. for half-an-hour. For certain strains the serum principle is inactivated at much lower temperatures.

6. Lability curves show marked irregularity. In certain cases a zone of relative inactivation is produced at a temperature of 55° C.

7. The natural agglutinating substance is found to be present in greater degree in the carbonic acid insoluble fraction of serum than in the carbonic acid soluble fraction. In this respect it differs from the immune agglutinins, which are chiefly located in the carbonic acid soluble moiety.

8. The agglutinating principle for each organism can be absorbed completely by the homologous strain, when a variable lowering of the end-titre for other unrelated organisms results. A similar lowering of activity for these organisms may be produced by treating the serum with non-specific physical absorbents. Charcoal and Kieselguhr were used to demonstrate this.

9. By the technique of double absorption it can be shown that agglutination depends on non-specific and specific factors and it is concluded that normal serum agglutinates bacteria in virtue of a twofold mechanism:

(a) A non-specific effect reacting in varying degree with all organisms and removable by treatment with a finely divided absorbent.

(b) A series of specific effects reacting as true “natural antibodies.” These specific antibody-like principles exist for a wide variety of organisms. Absorption of any one organism removes the homologous effect leaving the remainder quantitatively unimpaired.

10. The question of bacterial variation and receptor analysis in relation to the natural agglutinins is being studied and will be reported on at a later date.

1. The virulence of a given strain of B. aertrycke was tested on 54 occasions during the 3-year period, October 1926–October 1929. Each test was made by inoculating intra-peritoneally about 100 viable organisms into 20 parti-coloured mice, usually of 17–23 grm. in weight, drawn at random from the general laboratory stock. The average number of specific deaths caused by this procedure was 12·71 ± 3·39, but very marked deviations were noticeable, varying from a minimum of 5 to a maximum of 20. These deviations occurred without any observable regularity, and appeared to have no definite seasonal relationship. Evidence is brought to suggest that they were chiefly determined not by variations in the dosage or in the virulence of the strain, but by fluctuations in the resistance of the mice.

2. The meaning of the term “resistance” is discussed, and it is suggested that the term “fluctuating immunity” should be used to denote those variations in resistance which occur from time to time, and which are due to alterations in the physiological behaviour of the animal dependent on changing environmental conditions. This fluctuating immunity should be distinguished from innate, inherited, orgenetic immunity on the one hand and from acquired immunity on the other.

3. It is pointed out that the existence of fluctuations from time to time in the susceptibility of animals to infection renders fruitless any attempt to obtain an absolute measure of virulence. The most that can be done is to compare the relative virulence of two or more strains injected simultaneously.

1. Three hundred and two samples of sera from normal people, and 847 samples from cases suspected of enteric, or from contacts, have been examined for heat labile agglutinins against B. typhosus and allied organisms.

2. Sera from the “not normals” show a higher percentage of agglutinins against nearly all the organisms tested than sera from the “normals.”

3. In the contact group of the “not normals” the incidence of agglutinins is practically the same as in the “normal” group.

4. Sera from inoculated persons show a higher percentage of agglutinins against those organisms included in the T.A.B. vaccine used during the late war, than the sera from non-inoculated persons.

5. There is a higher percentage of agglutinins for B. typhosus, B. paratyphosus A and B. paratyphosus B among males than among females, especially in the age groups 20–30 and 30–40.

6. The proportion of normal males whose sera agglutinate B. typhosus, B. paratyphosus A and B. paratyphosus B and the group suspension is significantly lower than corresponding proportion noted in an investigation recorded by Rosher and Fielden in 1922. The proportion of agglutinating sera for these strains among normal females is substantially the same in our findings as in theirs.

7. The associations between ability to agglutinate different pairs of organisms are shown to be statistically significant in those cases which might arise from T.A.B. inoculation.

8. For these reasons it seems likely that the present distribution among the adult male population of agglutinins for B. typhosus, B. paratyphosus A, B. paratyphosus B and for Salmonella strains allied to B. paratyphosus B when in the group phase, was largely determined, in the period 1925–6, by the results of T.A.B. inoculation carried out during the war. It is probable that this effect is slowly disappearing.

9. Statistically significant associations were found between certain other pairs of organisms, which could not be explained as resulting from inoculation with T.A.B. vaccine, or as a result of any common antigenic stimulus. These associations were found mainly in the “not normal” group; and their meaning, at the moment, remains obscure. It may be noted that the number of sera agglutinating these particular pairs of suspensions was, in most cases, small.

10. Our results, so far as they yield any information on this point, do not suggest that infection with any one of the organisms studied by us increases the probability of the patient's serum agglutinating some other bacterium, with which the infecting organism is not antigenically related.

11. High titres were found chiefly in sera from the “suspected of enteric” group. The normal group usually gave low titres.