Attempts to demonstrate reversed passive anaphylaxis in the guinea-pig with crystalline egg albumin as sensitizing antigen have been uniformly negative.

When purified anti-pneumococcal antibody globulin was used as sensitizing antigen, reversed anaphylactic shock could be elicited in guinea-pigs by the intravenous injection of precipitins for the antibody globulin.

The mild reactions which could be elicited when the total globulins from the serum of normal rabbits were used as sensitizing antigen are probably dependent on the presence of small amounts of y globulin.

Reversed passive anaphylaxis, like direct anaphylaxis, is dependent on a cellular mechanism, and the success of experiments in which rabbit antibody globulin was used as sensitizing antigen depends on the acceptability of the antibody to the cells of the guinea-pig's tissues.

Antigenic differences between antibody globulins and total normal globulins from rabbit serum are noted.

Measurements of the rate of ventilation were made on 312 occasions in thirty-one rooms in (a) a block of older flats, (b) a block of modern flats, (c) three villa residences, and (d) a modern university building.

The rate of air change was determined by liberating a quantity of coal gas into the room and measuring the rate of its subsequent disappearance by means of a katharometer.

The effects on natural ventilation of air gratings and flues were studied in unheated rooms in two blocks of flats. The effective orifices of the gratings were 10·5 sq. in. in the modern flats and 6 sq. in. in the older ones. With the air gratings or flues (if any) closed the rates of air change were 0·55 and 0·71 per hour respectively. With the air gratings open the air change was increased to 0·83 per hour in the modern flats and to 1·04 per hour in the older ones.

The flues in the modern flats in connexion with gas fires were 4 in. diameter pipes (cross-sectional area, 12·6 sq. in.); in the older flats the ordinary fire-places were connected to standard flues and the area of the register openings was 53 sq. in. When the flues were sealed, the air changes per hour in the modern and older flats were 0·84 and 0·72 respectively and when the flues were opened these rates of ventilation were increased to 1·17 and 2·06 air changes per hour. In nine rooms in three other houses the average ventilation rates with the flues closed and open were respectively 0·82 and 1·76 air changes per hour.

The ordinary chimney flue, even without a fire at its base, was an efficient ventilator, but the flue of smaller dimensions was far less effective.

With windows open almost any required amount of ventilation could be obtained, and the observations showed that even when windows are opened only slightly the ventilation is distinctly increased. Thus, in flueless rooms of the modern flats, with doors and windows closed, but with the air grating open, the ventilation amounted to 0·83 air changes per hour, but with one casement only open to a width of 3/8 in. it was increased to 1·79 air changes.

The effects of various factors on the rate of air change are discussed and illustrated. Of these, outside wind speed appears to have most influence. The evidence regarding the effect on ventilation of the difference in temperature between the room air and the outside air is very conflicting. In the present observations, when outside wind speed is taken into account, no definite effect of temperature difference on air change is seen. But the experiments were made in unheated rooms and the temperature difference rarely exceeded 10° F., and in the majority of cases was much less.

The various recommended standards of air supply are discussed. Although certain bye-laws require a fresh-air supply of 1000 cu. ft. per person per hour, it is probable that where there is no overcrowding, as in the normally occupied dwelling-room, a ventilation rate of not less than 600 cu. ft. per person per hour should be sufficient to prevent body odours from becoming apparent.

The results of this inquiry show that in closed, flueless rooms, the air supply is likely to be much below 600 cu. ft. per hour per occupant.

On account of the low ventilation rates observed in flueless rooms it is desirable that the question of the use of flueless gas heaters in such rooms be reconsidered.

The complex group of organisms with cultural and biochemical properties closely similar to the classical Cl. welchii are at present differentiated into four types according to the quality and number of the toxins they elaborate in vitro.

The present study is concerned with the specific affinities of the bacterial antigens of representative strains from these four types. The experiments confirm earlier reports concerning a wide diversity in the specificity of the somatic antigens of the classical type A strains. The bacterial antigens of the types B, C and D, however, have been shown to possess interesting similarities and affinities:

(1) The O antigen of type A (classical Cl. welchii) strains is strictly strain specific.

(2) The type B (L.D. bacillus) strains isolated in Great Britain share an identical O antigen but a strain from South Africa had a strictly specific O component.

(3) The O antigen of the type C (B. paludis) strains isolated in Great Britain is identical.

(4) The type D (B. ovitoxicus) strains exhibit wide diversity in the specificity of their O component. Thirteen strains were placed in not less than seven groups.

(5) Cross O reactions between various types is negligible or entirely absent.

(6) In addition to the heat-stable O antigen certain strains possess a heat-labile somatic component. This antigen occurs only among those types (B and D) that produce ɛ toxin but it is probably not present in all strains.

(7) Colony variants reminiscent of the typical rough forms of Gram-negative organisms are frequently encountered among strains of each type. Certain of these forms appear as stable variants and they are characterized serologically by the loss of O antigen specific for the strain.

(8) The rough forms still produce toxins characteristic of the type to which the parent culture belongs.

A form of ultramicroscope apparatus is described which enables the evaporation of mist droplets to be followed for long periods. A rough method of estimating the size distribution in aerosols is also given.

The rates of evaporation of particles of various phenol solutions have been studied, the solutions chosen being of possible interest as aerial germicides. It is shown that the typical evaporation curve for a binary mixture consists of three parts, which in actual cases need not all be present. Sometimes the more detailed history of an evaporation can be deduced from the results.

The approximate effective size distributions in mists of hexyl-resorcinol—propylene-glycol solutions have been measured and compared with the evaporation rates.

It is shown that the efficiency of aerial germicides is partly a function of their volatility, and three rough classes are distinguished. But biological evidence is adduced which suggests that the solubility of the germicides in water may be of equal importance, and that the liquid in which the germicide is dissolved is not without effect.

A centrifuge is described which enables the maximum particle size in a mist to be limited to a desired value.

Methods of estimating quantitatively germicidal aerosols are discussed, and a method described which has been found to give reasonably accurate results in practice. Tentative experiments have been made to determine whether the materials studied are likely to cause corrosion of metals, etc., after pro-longed use.

The collision theory of aerial disinfection is discussed, as is also the applicability of Langmuir's treatment of the evaporation of fine particles in mixtures.

The development of the experimental technique involved in the biological investigations appertaining to air disinfection has been discussed, together with the numerous fallacies arising and the means by which these have been overcome.

The construction of a special pair of experimental chambers is described, and details given, both as regards the bacteria and germicides, of the exact experimental conditions under which the most consistent results can be obtained.

The requisite characteristics of germicidal mixtures which could be utilized for air disinfection are considered, and examples of suitable germicides and solvents are given. The method of recording and evaluating the results of a varied assortment of experiments is explained to facilitate inter se comparison.

Examples are given of the lethal effectiveness in the air of several germicides in broth emulsions, both with and without the presence of serum, on stock cultures of selected types of bacteria. The sensitivity to the germicides of the bacteria making up the flora of the average normal saliva is compared with that of broth emulsions of bacteria.

10% hexyl-resorcinol dissolved in propylene-glycol+0·05 % sulphonated lorol (“S2”) was the most effective all round germicidal mixture of those tested. Various other germicides were found equal in germicidal activity in the air, but for reasons stated have been deemed unsuited for use in the presence of man.

The relationship between germicidal efficiency in the air and mist particle size and persistence has been studied by means of centrifuged mists. Particles of 0·5–1·0μ radius are shown to be the most useful in dealing with bacterial particles containing organic matter. Laboratory cultures, emulsified in broth, have been sterilized in mist dispersions by droplets of bactericides at least as small as 0·25μ radius.

Tests on the penetrating abilities of bactericidal mists and bacteria through certain cloths have been performed.

Experiments to demonstrate the surface sterilizing properties of aerosols have been performed, and examples of differences in effect on horizontal and vertical surfaces are noted.

The degree of tolerance of man and animals (mice) for a number of phenolic substances and some organic solvents has been compared with the degree of tolerance of bacteria for the same substances.

The tests on man have been confined to the detection by the special senses of the substances suspended in the air as a fine mist.

The tests on animals were designed to show acute and chronic effects, the results being recorded in terms of clinical symptoms and pathological changes in the organs. A record was also kept of the findings as regards protozoal parasites of the host, and compared with the normal incidence among ordinary laboratory stock.

Tests on bacteria were instituted for the purpose of gaining information on the host-parasite tolerance ratio for our germicides when in the air and when in the test-tube.

Before the foundation in 1893 of the Journal of Pathology and Bacteriology by the late Professor Sir G. Sims Woodhead, there was no English journal devoted to pathology, bacteriology and kindred subjects. In 1901 Professor G. H. F. Nuttall founded the Journal of Hygiene, the second English journal of this type, and edited it till his death in December 1937. Owing to the increasing number of papers on parasitological subjects, which he received for publication in this journal, he founded Parasitology in 1908 and was its chief editor till 1933. He was very proud of both journals, and for some years made himself personally responsible for the heavy expenses of production. Within a few years of its foundation each began to pay for itself, and afterwards both were taken over by the Cambridge University Press.

1. A Clostridium, non-pathogenic to laboratory animals and not producing toxin, but in some ways resembling Cl. oedematiens, was isolated from the hip joint in a case of “bone-taint” in beef.

2. It was demonstrated in the synovial fluid, the periosteum, the Haversian canals, the bone marrow of the femur, and in associated muscular and connective tissue and blood vessels.

3. Examination of forty hip joints from normal carcases indicated that thirty-eight were sterile, one heavily infected with a Streptococcus, possibly of the faecalis type, and one with a rod, probably a member of the Proteus group.

1. The types of strains of tubercle bacilli isolated from fifty cases of non-pulmonary tuberculosis in children, 0–14 years of age, in Eire have been determined. In cervical gland disease ten (71·4%) out of fourteen strains were bovine, in bone and joint disease four (18·2%) out of twenty-two were bovine and in meningitis three (25%) out of twelve were bovine. Only two cases of primary abdominal tuberculosis were examined. Both yielded human strains.

2. The difficulties in assessing the true significance of the results recorded are briefly discussed.

3. The technique used and some bacteriological observations made are described.

1. Sulphapyridine is an efficient prophylactic or therapeutic agent against intradermal infection with the strain of Vibrion septique selected for test, whereas, in intramuscular infection the drug is unreliable and saves only a relatively small proportion of animals.

2. The strain of Cl. welchii (type A) selected for test is not insusceptible to the action of sulphapyridine. A significant degree of protection against intradermal infection may be obtained however, only when a prophylactic dose of the drug is given. The drug is apparently without action against intramuscular infection with this strain.

3. Under the particular experimental conditions sulphanilamide is much less effective than sulphapyridine against infection with either V. septique or Cl. welchii.

4. There is no evidence that sulphapyridine given per os neutralizes the toxins of V. septique or Cl. welchii injected intravenously.

5. It is possible to control infection with V. septique by antitoxin or antibacterial serum given at a time when sulphapyridine is of little use. Cl. welchii antitoxin has a marked therapeutic effect in infection with this strain of the organism but sulphapyridine given only after infection is apparently without action on the course of the disease.

6. In intradermal infection with V. septique the combined action of sulpha-pyridine and antitoxin or of sulphapyridine and antibacterial serum effects a saving in life much greater than would be expected if a mere summation effect was in question. A similar effect was observed in intramuscular infection provided the administration of the drug was sufficiently prolonged. No such synergic effect is produced by the combined action of antitoxic and antibacterial serum.

7. In Cl. welchii infection the combined action of antitoxin and sulpha-pyridine produces a noticeable synergic effect but the evidence on this point is less clearly defined.

8. The pathogenesis of infection with spore-bearing anaerobes is discussed in relation to prophylaxis or therapy with chemotherapeutic agents, antitoxin, antibacterial serum and combinations of such.

9. The statistical aspect of summation effects is briefly considered pending a forthcoming mathematical analysis, by Dr K. Mather, of the figures obtained in experiments on synergic action.

1. The vertical distribution of bacteria has been determined in Windermere (north and south basins), Thirlmere, and Esthwaite Water; these lakes are oligotrophic bodies of water in the English Lake District. Studies on the horizontal distribution of bacteria have been confined to the north basin of Windermere.

2. Bacterial counts at the mouth of the main river flowing into Windermere were about 200 times as great as counts in samples of water from the lake at a distance of 200 m. offshore from the point of entry of the river. In general there was a gradual decrease in numbers of bacteria with increasing distance from this 200 m. point but such decreases were relatively insignificant.

3. During the winter period of circulation of water in the lakes numbers of bacteria were approximately the same at all depths on any one occasion. When the water in the lakes was stratified, however, the numbers of bacteria decreased with increasing depth in the epilimnion, but below the thermocline the numbers were approximately the same at all depths, except in the layer of water immediately above the mud where the bacterial population was somewhat greater.

4. Progressive depletion of dissolved oxygen with depth in the hypolimnion during the summer months did not stimulate the growth of anaerobic bacteria, nor was the degree of depletion related to numbers of bacteria.

5. Periodic fluctuations in numbers of bacteria which took place at all depths were found to be directly correlated with the amount of rain which had fallen in the drainage area during the week before the samples were taken. The correlation is most significant during the winter months, possibly on account of the high rainfall during those months. During the summer the relation was most marked in the upper layers of Windermere north basin, but in the lower layers to which rain had no access there was no correlation. In the south basin of Windermere, where the bacterial population has become more stable, numbers were not affected by the relatively small summer rainfall. There are indications that short early summer fluctuations in numbers of bacteria unrelated to rainfall may be related to maxima in algal growth.

6. Changes in the temperature and concentration of inorganic substances in the lake waters have not shown any relation to fluctuations in the bacterial population, nor has any noticeable bacteriostatic effect of sunlight been apparent.

The question of testing for synergic action of treatments of Vibrion septique infection in mice is considered. Henderson and Gorer's data are used as the basis of the analysis.

Their six experiments are tested for heterogeneity and one experiment is found to be discrepant. This is then removed and only the remaining five used in the later working.

A x2 formula is derived for testing the significance of the evidence for synergic action. Sulphapyridine shows such interaction in effect with both antitoxin and antibacterial serum. These last two show no interaction.

The most efficient design of such experiments is then considered, and it is shown that the death rate in the single treatment tests should be kept at least as high as 90 %, and that whereas the numbers of mice assigned to the single treatment tests should be equal, that assigned to the joint treatment should be 1·5 times as great.

A graphical method for testing the significance of the results is described. These graphs also allow of an opinion being formed as to the number of mice which would be necessary to detect an interaction of given magnitude.

1. In order to obtain evidence as to what extent the urban excess in cancer mortality is real, an analysis has been made, in respect of both location of tumour and geographical region, of the cancer mortality statistics for Scotland, 1931–7 inclusive.

2. Of the various locations, the greatest urban excess is shown by cancer of the respiratory tract; this is considered to be largely, though probably not entirely, due to better diagnosis and more accurate certification of deaths in the cities. Next come the buccal cavity and uterus; in both these locations certification is likely to be relatively accurate, so that a real excess would seem here to be present. The other accessible locations, the skin and the female breast, do not show significant urban excesses, but the numbers are small, and the existence of a real urban excess is not excluded. The results are in general agreement with those of the English experience 1911–20.

3. Further analysis of the data presented by Russell for Scotland, 1923–8, reveals an excess in the industrial counties of deaths for cancer of the buccal cavity, and perhaps a smaller one for cancer of the female breast. These data are, therefore, not inconsistent with our findings.

4. As compared with the period 1921–30, the figures for 1931–7 demonstrate, for ages 25–65, a marked improvement in cancer of the buccal cavity and uterus. A recorded fall in cancer of the digestive organs may, in part, be due to more accurate certification in respect of the primary location. The failure of mortality from breast cancer to fall may be due to the same cause. The large increase in lung cancer is almost certainly due, in part, to better diagnosis. For all sites together, the cancer mortality between ages 25 and 65 remained constant for men, and fell by 8% in the case of women. The standardized rate for all ages and both sexes combined increased by 1·3%, a rise which is 2·2 times its standard error.

5. The urban excess cannot be entirely explained as due to incomplete certification in the rural areas. The implied effect of social and industrial environment in stimulating tumour growth is in harmony with the known facts regarding occupational and social cancer, and emphasizes the importance which the control of adverse environmental factors may have in the reduction of cancer incidence.

1. 656 samples of sera taken at random from natives in Southern Rhodesia were examined for the presence of typhoid Vi agglutinins.

2. Forty-nine or 7·47% were shown to contain Vi agglutinins in titres of 1 in 5 to 1 in 12·5.

3. Samples of faeces and urine from twenty-six of the positive Vi reactors were cultured for Bact. typhosum, which was isolated from only one individual who was shown to be a chronic urinary carrier.

4. The significance of these findings is discussed.

The study of the bacteria of natural waters is dependent on a simple and effective sampling technique. Any routine sampling procedure designed to obtain a vertical series of samples at one station must not only exclude any possibility of contamination of the sampling vessel and its contents while it is lowered to, and raised from, the depth at which the sample is taken, but must also be capable of straightforward operation on a boat or ship under rough conditions. The sampling vessels should be of a standard, interchangeable and robust pattern, and their preparation and sterilization should be a matter of simple routine. It should be possible to use them repeatedly and they should be of sufficient capacity to permit physical and chemical as well as a bacteriological examination of the sample.

1. The types of tubercle bacilli have been determined in the sputum of 515 cases of pulmonary tuberculosis occurring in the middle and south of Scotland.

2. Of the 515 cases 484 were human (476 eugonic and eight dysgonic) and thirty-one were bovine infections.

3. With the exception of the strains from one case (case 28) all the bovine strains, seventy in number, were typical culturally and fully virulent for rabbits.

4. The attenuated strains, two in number, from case 28 were slightly less virulent than typical bovine strains for rabbits and (one strain) for guinea-pigs.

5. The percentage of bovine infections found in this series, including the Cumberland case, during the years 1931–9 was 6·0, but excluding that case it was 5·8.

6. The percentage of bovine infections found by Munro during about the same period and covering the same regions was 5·0%.

7. In Munro's series strains of bovine tubercle bacilli were obtained from fifty-eight out of 1165 persons (5·0%). Five of his cases yielded attenuated bovine strains and in one of these the pulmonary tuberculosis was preceded by tuberculosis of the thoracic spine.

8. In my series the attenuated tubercle bacilli came from a case (case 28) of pulmonary tuberculosis which was preceded nearly 20 years previously by tuberculosis of the lower dorsal spine.

Dr Munro and others have made post-mortem examinations on cases of phthisis pulmonalis due to bovine bacilli, but I wish to defer reference to these until we can review them altogether.

In this series there are seven instances of cervical gland enlargement and one instance (case 28) of spinal tuberculosis occurring previous to the development of phthisis pulmonalis. These, I think, are examples of alimentary infection with the bovine tubercle bacillus. Thus, with the three autopsies previously mentioned, there are eleven cases, or about one-third, which are almost certainly alimentary in origin. As for the rest of the cases, 20 in number, no glandular enlargements in neck or abdomen were detected but the majority, if not all, were probably alimentary in origin, since all the persons drank a lot of raw milk and only five came into direct contact with cattle in their employment.

Experiments on the bactericidal action of ultra-violet light have been made to determine the shape of the survival curve and the dependence upon radiation intensity of the rate of death. Bact. coli, Bact. prodigiosum and spores of B. mesentericus were irradiated with approximately monochromatic light of wave-length 2537 A. The survival curves obtained were exponential and the rate of death was accurately proportional to the intensity over an intensity range of 500:1.

By comparing these results with data previously obtained of the action of X-rays on the same organisms it was established that one ionization produced by X-rays is as effective as some hundreds of ultra-violet quanta. This is interpreted to mean that the quantum yield in whatever chemical change leads to the loss of viability in the irradiated bacteria is, for 2537 A., between 0·01 and 0·001.

The poliomyelitis epidemic of 1937 in Denmark was investigated by means of information obtained about each individual case.

It occurred in the usual season, culminating in September. The districts involved were limited largely to Zealand, including Copenhagen. It is of particular interest that the disease occurred also in rural districts that were the sites of similar epidemics in 1933 and 1934.

It is noted as a peculiar fact that poliomyelitis seems to be unable to invade the whole of Denmark within one season.

The 1937 epidemic comprised 1200 notified cases, of whom 599 developed paralysis and 113 died. Everything suggests that the number of non-paralytic cases was far greater than the number of cases notified. Influenza and angina were notified with striking frequency in the districts where poliomyelitis prevailed.

The morbidity, frequency of paralysis and mortality were greater among males than among females.

A change in the age distribution during the past 35 years is pointed out. In 1937, 30% of the patients were over 15 years, and about 14% over 20 years; of those who died 45% were over 15 years, 20% over 20 years.

Convalescent serum was not employed in Copenhagen, but in the provinces 489 patients were treated with, and 400 without, serum. The principles of the treatment were the same as in 1934. Serum was given early in the disease, as a rule intravenously.

1. A fluid enrichment medium containing sulphite and bismuth based on Read's modification of one of Wilson & Blair's media has been devised and the claims of Read (1939) and Seal (1939) as to the value of such a medium has been confirmed.

2. A saccharose mannitol sulphite bismuth alcohol phenol red agar medium is described and has been found to allow rich growth of the true cholera vibrio and to inhibit the growth not only of B. coli and B. lactis aërogenes but of many vibrios liable to be mistaken for the true vibrio cholerae.

3. Thirty-one strains of true cholera vibrios, obtained from the National Collection, grew rapidly and profusely in the media.

4. Of twenty-five cholera-like and para-cholera strains six grew well, whilst the growth of nineteen was scanty or nil.

5. Of eleven El Tor vibrios five grew in both fluid and solid bismuth media, five grew only on the solid medium and in one no growth occurred in either medium. The El Tor colonies were much smaller than those of the epidemic cholera strains.

1. During a period of 1 year the incidence of infection with haemolytic streptococci was determined among children on admission to, and during residence in, the children's wards of University College Hospital, London.

2. The increased incidence among those in residence, 13·1% of whom acquired haemolytic streptococci during their stay, was due principally to outbreaks of infection by these organisms. By the slide agglutination method the majority of strains could be assigned to one or other of Griffith's thirty types and the presumed path of their spread studied.

3. The source of the infection in most of the outbreaks was not ascertain-able from the data collected, but in general it did not appear to be in the patients themselves.

4. With a few exceptions, isolation by cubicles was an effective measure in preventing the spread of haemolytic streptococci.

5. In general the haemolytic streptococci seemed to spread from patient to patient, though the exact mode of spread was not ascertained.

6. The clinical manifestations of infection by any one type of haemolytic streptococcus was varied, e.g. wound suppuration, sore throat, pyrexia, otitis media. Many of the infections were “latent”.

7. In addition to the haemolytic streptococcal outbreaks, there were two outbreaks of diphtheria (fifteen patients), two of Bact. sonnei dysentery (ten patients and one nurse), one of catarrhal jaundice (one patient and three nurses), six cases of intercurrent gastro-enteritis of infancy and one case of measles.

8. In all, at least 104 (or 18·9%) of the 551 patients of these wards developed intercurrent infections, manifest or latent, during their stay in hospital.

Ten generations of mice (with one exception) were inoculated with a partially purified toxic fraction isolated from Bact. typhi-murium, the next generation being bred each time from the survivors to the test. By this selective process a stock was produced with a substantially increased power of resistance to the toxin. For instance, in the fourth and fifth generations (combined), at which point of time the dose inoculated was stabilized at 4 mg., 36% of the mice with resistant ancestors survived inoculation compared with 12% of normal mice derived from the same original stock; in the ninth and tenth generations (combined) 64% of the resistant stock survived and only 14% of the controls. Also, those mice which failed to survive inoculation took slightly longer to die, on the average, in the resistant stock than in the normal stock.

The mice of each generation were not inoculated until about the age of 12 weeks, but to test the possibility of the transference of a passive immunity from mother to young the seventh generation was bred from without previous inoculation. The eighth generation was composed therefore of descendants of animals selected for resistance by inoculation but its own parents had not been tested. Its survival rate, 59%, was substantially greater than that of the normal stock, 17%, proving that the differences observed in previous generations were not the results of a transference of passive immunity.

Some mice of the ninth and tenth generations were inoculated with the living organism Bact. typhi-murium in place of the toxic fraction derived from it. Although these generations were relatively highly resistant to inoculation with the toxin itself they were less resistant than normal mice to the living organism.