These studies, carried out during the period 1940–2, were undertaken to test the relative merits of hypochlorous acid gas and sprayed solutions of sodium hypochlorite, when these systems are used as aerial disinfectants, and to ascertain the factors determining their efficiencies in this capacity. The ulterior motive was the use of such systems as a counter measure against the danger of air-borne infection. The more practical aspects of their large-scale application, in shelters and other compartments, have been the province of others of our colleagues and will be reported upon separately.

1. In the period 1923–39, more than one-fifth of all deaths in the age group 0-2 years in the Administrative County of London are recorded as having been caused by bronchitis or broncho-pneumonia. It is very unlikely that the returns have been exaggerated by more than 10–15% through the inclusion of deaths from secondary broncho-pneumonia following measles, whooping cough or influenza. In the 8 years in which measles epidemics occurred, the peak of the measles mortality was from 2 to 17 weeks after, and never before, that of the mortality from broncho-pneumonia.

2. The winter rise in mortality from broncho-pneumonia is not associated with any comparable increase in the carrier rates for pneumococci and haemolytic streptococci.

3. The broncho-pneumonia mortality of infants in the Metropolitan Boroughs is strongly correlated with sub-standard housing incidence, social index and percentage below the poverty line. The relatively small value found for the coefficient of regression between broncho-pneumonia mortality and sub-standard housing may be due to the less rigorous winter temperatures occurring in the more overcrowded central areas.

4. Low external temperature appears to be the most important predisposing cause of fatal broncho-pneumonia in infants; their weekly death-rate from this disease in mid-winter is about seven times that in the late summer. This great seasonal fluctuation in mortality is also found for persons over 55 years in the Administrative County, and for persons of all ages in both the urban and rural districts of the Home Counties.

5. The weekly mortalities of infants from broncho-pneumonia in the Administrative County are most strongly correlated with low temperatures 2 weeks earlier. The two variables are so associated that the mortality rises sharply when the mean temperature of the coldest day of the week falls below 40° F. There is no indication that low temperatures in the spring are associated with higher death-rates from broncho-pneumonia than the same temperatures in the autumn, so that the data provide no support for the supposition that susceptibility increases during the winter in consequence of the progressive depletion in the body of protective nutrient substances.

6. Possible measures to reduce the mortality of infants from broncho-pneumonia are discussed.

A total of 1040 individuals has been examined for the typhoid carrier state by the O, H and Vi agglutination tests. Some observations have been made upon the technique of the Vi test.

Estimation of the O and H agglutinins proved unsuitable as a preliminary test owing to the high proportion of positive reactors. The Vi test, in contrast, gave a positive result in only a small percentage of those tested and brought the cultural work subsequently required within practical limits. All positive reactors were examined by cultural methods, and in no case were the organisms of enteric fever recovered.

The relative merits of the various tests available for the detection of carriers are discussed and certain tentative conclusions are expressed.

The Group I paracolon bacilli formed biochemically a stable group. They fermented glucose, maltose, mannite, and dulcite, but not saccharose. Most of them could be distinguished from Salmonella organisms by the late fermentation of lactose or the production of indole. Many resembled the Newcastle bacillus in being late-dulcite fermenters. Serologically 75% of the strains were identical, each containing the A antigen. Some formed a separate group containing the A antigen, and Bact. alkalescens antigen and the C antigen of some Group III strains. The A strains are probably identical with those of Dudgeon & Pulvertaft (1927), but their strains were described as being motile. All Group I strains had minor agglutinins in common with the dysentery bacilli, particularly when freshly isolated. Most of Dudgeon's strains were isolated from cases of urinary infection, but in the present investigation almost all were isolated from the discharges of infants with diarrhoea and enteritis. This organism may be regarded as pathogenic or facultatively so, and deserves a name. I suggest it would be appropriate to attach to it the name of Dudgeon, who was the first to call attention to it.

Biochemically Group III bacilli resembled Group I strains but differed in fermenting saccharose. Some were motile. The metabolic and biochemical properties examined were stable. Serologically many (40%) possessed the A antigen, but also possessed a type-specific antigen, E. Others (37%) possessed both C and D antigens, usually C but sometimes D was dominant. These strains are related to the dysentery bacilli by the possession of common, minor antigens. A group of serologically heterogeneous strains remained (23%), generally unrelated to the dysentery bacilli.

Most Group II bacilli were stable biochemically, but some were not. Many strains of Group IV were unstable. Serologically, Groups II and IV were mostly heterogeneous, although 24% were included in the A, B, C, D, E pattern. Together they formed but a minority of all the paracolon bacilli.

Further serological investigation is required to discover new antigens and to link up the paracolon bacilli isolated from various sources.

This work was carried out under a whole-time grant from the Medical Research Council of Ireland to whom I am indebted.

I wish to record my very best thanks to my director, Prof. R. A. Q. O'Meara, for his continuous interest and invaluable advice at all times and without whose aid this research would not have been possible. My gratitude and thanks are due to my colleague, Dr J. St L. O'Dea, for his unfailing co-operation, particularly in obtaining all the specimens so quickly, in assisting at the post-mortems, and for many other services too numerous to mention.

Thanks are also due to the authorities, doctors and nurses of St Ultan's Hospital, Temple Street Hospital, St Kevin's Hospital, Clonskeagh Fever Hospital, Cork Street Hospital, and the Child Welfare Clinics in Dublin, particularly to Dr F. N. Elcock and Dr T. Murphy for their support and co-operation; to the ‘Diarrhoea and Enteritis’ Committee of the Department of Local Government and Public Health for their supervision of the research; and to the staff of the Department of Bacteriology, Trinity College, Dublin, where the research was carried out, for the help and facilities given.

I would like to thank all those general practitioners in Dublin who co-operated by the early notification of cases and in other ways. Finally, I wish to record my appreciation to the Statistics Branch, Department of Industry and Commerce, for their interest and kindness in furnishing a valuable report.

In a discussion of a series of tests connected with the fastness of colours on textile fibres, Villavecchia (1918) mentions the use of neutral ammonium acetate or common salt solution for measuring the colour resistance to perspiration. Thus in the case of coloured cotton, this is immersed for 10 min. alongside an equal quantity of white cotton yarn in a 0·1% aqueous ammonium acetate at 80°C., and the extent of colouring of the yarn as well as the degree of stripping of the coloured sample noted after drying without rinsing. An odd number of degrees of fastness is arbitrarily assigned to the coloured specimen according to the result, e.g. V degrees if neither the original tint nor the whiteness of the yarn is changed. Similarly with linen, hemp and ramie. For coloured wool both methods are used, viz. with sodium chloride and with ammonium acetate. In the former case, the alteration in colour is noted after simply dipping the wool in salt solution and allowing to air-dry out at room temperature. In the latter case, an equal quantity of white zephyr wool in addition to the white cotton is present. V degrees in all cases are stated to be conventionally adopted between industrial associations and dyestuff manufacturers for materials which are fast enough to reveal no change whatever in the testing bath. For coloured silk there are no precise data, and usually it is immersed in distilled water for several days to ascertain if any colour is lost.

Since Green & Mellanby (1928, 1930) reported that vitamin A protected the body against infection, this problem has been studied by various authors (Verder, 1928; Lassen, 1930, 1931; Seidmon & Arnold, 1931–2; Boynton & Bradford, 1931; McClung & Winters, 1932; Stryker & Janota, 1941). In general, the results showed that A avita-minotic animals were more susceptible to spontaneous or experimental infections than controls kept on a normal diet. However, in so far as we know, no experimental work has been done on the inverse problem—the effect of infection on vitamin A metabolism.

Green & Mellanby (1928, 1930) were the first to direct attention to a specific relationship between vitamin A deficiency and infection. They studied the occurrence of spontaneous infections in vitamin A-deficient rats and concluded that animals fed on a diet deficient in vitamin A, but complete in all other respects, develop multiple infective lesions and die. Since then vitamin A has become known as the ‘anti-infective’ vitamin.