It has long been recognized that there is a relationship between environment and health and examples of this association are easy to cite—overcrowding tends to spread infection. But our knowledge of the exact part played by occupational as distinct from economic and social environment is as yet imperfect.' The liability of the worker to ill-health, apart from any inborn constitutional weakness, is determined broadly by two factors. The first is the influence of the type of work directly upon the worker, an influence inherent in the actual occupation and the conditions under which the work is done. This might be regarded as the direct occupational risk. The second is the economic and social environment conditioned by the occupation followed: this might be termed the indirect occupational risk. This distinction was recognized by the Eegistrar-General who stated that both factors are essentially components of the total occupational risk since ' If a man is obliged by his place of work and his rate of pay to live in an insanitary area, the extra risk of death involved is, in a wider but very real sense, part of the occupational risk'. The purpose of the present paper is to try to disentangle the parts played by each in determining the occupational mortality from a specific disease like tuberculosis. The difficulty which previously presented itself was the absence of a standard which could be regarded as a measurement of the indirect occupational risks. The Registrar-General, in his last Occupational Mortality Supple-ment, has provided an approximate index, by publishing data on the mortality of males following particular occupations and also that of their wives. His viewpoint was.'. for no trade could longer be regarded as directly prejudicial to health if it were found to entail as much excess risk for the wife as for the husband. In such a case excess mortality would evidently be in the main attributable to the social conditions implied.' Greenwood likewise stated that ‘the wives share the social and geographical advantages of their husbands’ occupation but, with relatively ummportant exceptions, are not exposed to their specific occupational risks’. The purpose of the present paper is to try to disentangle the parts played by each in determining the occupational mortality from a specific disease like tuberculosis. The difficulty which previously presented itself was the absence of a standard which could be regarded as a measurement of the indirect occupational risks. The Registrar-General, in his last Occupational Mortality Supplement, has provided an approximate index, by publishing data on the mortality of males following particular occupations and also that of their wives. His viewpoint was. ‘..for no trade could longer be regarded as directly prejudicial to health if it were found to entail as much excess risk for the wife as for the husband. In such a case excess mortality would evidently be in the main attributable to the social conditions implied.’ Greenwood likewise stated that ‘the wives share the social and geographical advantages of their husbands’ occupation but, with relatively ummportant exceptions, are not exposed to their specific occupational risks'.

In a previous paper (Twort, Baker, Finn & Powell, 1940) we were able to show that the bactericidal properties of phenolic -aerosols were correlated with their physical characters, though our experiments were not sufficiently numerous or varied to enable us to estimate separately the effect of the different factors operative. Here we propose to describe similar tests which we have carried out on the germicidal smokes used by Twort & Baker (1940, 1941), namely, those produced by smouldering cardboard impregnated with potassium nitrate, and by dropping incense and Peru balsam on to hot metal surfaces. There were also used mixtures obtained by distilling Peru balsam under reduced pressure: fraction I, b. p. 109–110° C; fraction II, b. p. 120–138° C; fraction III, bp 148–155° C, at 3 mm. Hg, and the residue from this operation.

Human serum was dried in this department at the outbreak of war with a view to its possible use as a stable substitute for whole blood for transfusion purposes. In the absence of any definite knowledge of dosage, the arbitrary unit of 200 c. c. was chosen, because this was approximately the amount of serum which could be obtained from a pint of clotted blood, and also because this amount of serum could conveniently be frozen before drying in 12 oz. medical-flat bottles. The technique of freezing was to place the bottles of serum on racks in a cold-room maintained at –20° C, the racks being slightly inclined from the horizontal so that the serum froze in a wedge, thus giving a reasonably large surface from which evaporation could occur during the drying process. Large amounts of human serum have been dried by sublimation from the frozen state on apparatus based on the experimental model described by Greaves & Adair (1939), after prefreezing by this simple technique.

In 1930 the author (Graham-Smith, 1930) showed that of spores of B. anthracis inoculated on 7 July 1907, on to dry, sterile pieces of canvas, measuring about 0·5 × 0·5 in., which were kept in a Petri dish in a cupboard with a glass door in the laboratory and therefore exposed to diffuse daylight at room temperature in a dry atmosphere, large numbers were capable of germinating in ordinary laboratory media after 10 years and progressively smaller numbers after 17 and 20 years. Very few germinated after 22 years and none after 22½ years.

Papillary variation includes all heritable changes of character which take place in the papillae of bacterial colonies and the corresponding variations in liquid media.

The following is a short summary of what is known about these variations. Papillary variation was first observed in B. coli-mutabile, and other coliform bacteria show the same variations. This race forms white colonies on lactose neutral-red plates. After 1–5 days’ incubation minute papillae, which become red and increase in size, appear on their surfaces. Subcultures from thesepapillae give colonies of two kinds, white and red. The white colonies continue to throw red, the reds breed true (Neisser, 1906; Massim, 1907). Similar variation to different sugars occurs in other related bacteria: B. dysentenae Sonne varies to lactose, B. dysentenae Mexner to maltose, B. paracolon to saccharose, B. coh-mutabile, B. acidi-lactici and B. typhosus to dulcite (Twort, 1907; Reiner Miiller, 1909; Burn, 1910; Stewart, 1926, 1927). Vacation takes place only on the sugar concerned, but, once it has taken place, the character remains fixed, even after long growth on other media. Variation is not determined by length of exposure to the sugar. It occurs at (or after) the end of the logarithmic phase. If this phase is kept going by daily subculture (in the appropriate sugar) the bacteria of the chain carried on from culture to culture do not vary, however long they are exposed (Stewart, 1927; Deere, Dulaney & Michelson, 1939). On the other hand, if the logarithmic phase is curtailed by heavy inoculation of a culture, variation occurs early (Stewart, 1927). If the bacterium is grown on a plate without sugar until the colonies are well formed and the sugar is then added, with care not to disturb the colonies, variants appear after very short exposure (Stewart, 1927, 1928). A variation 33–2 may be completed in, two distinct steps; thus a race of B. typhosus formed white colonies on dulcite with pink papillae, subcultures from which gave white and pink colonies; the pink colonies formed dark red papillae from which on subculture came pink and dark red colonies, the latter bred true.

The experiments now described were carried out in Lagos and Khartoum, and although the hoped-for end-result, the production of a sheep strain of neurovaccinia, was not realized, certain points of interest have arisen and are recorded in the hope that they may be of help to other workers concerned with vaccinia vaccine production.

The early work on the chemistry of the Voges-Proskauer reaction is too well known to need consideration here. Several workers tried to hasten and strengthen the reaction by the addition of various oxidizing agents, but with little success. O'Meara (1931) made the first real advance when he investigated anew the chemistry of the reaction and found the accepted views incorrect in several respects. As a result he added creatine to the culture, used a much stronger alkaline solution than had hitherto been customary, and so achieved a marked improvement in the rapidity of development and intensity of the colour. He recommended the use of 24 hr. cultures.

Instances are recorded in the literature of mammalian -tubercle bacilli having lost their virulence completely during long cultivation on artificial media. In a paper published in 1925 I gave an account of the results of investigating two such strains which had been deposited in the National Collection of Type Cultures by Dr Nathan Raw. These strains were stated by Dr Raw to be lineal descendants of strains which had been given to him in 1906, one (human) by Prof. R. Koch, who had isolated it from the sputum of a case of human pulmonary tuberculosis, the other (bovine) by Prof. A. Calmette, who had obtained it from a mesenteric gland of a cow.

The following account of a milk-spread epidemic of scarlet fever in the burgh of Elgin may be of interest to those who are at times faced with similar problems.

HISTORY

The number of cases of scarlet fever notified in the burgh each year from 1920 to 1938 is given in Table 1. This table also shows that the periods of

Table 1. Scarlet fever notifications, 1920–38

maximal incidence, centring on 1921, 1926 and 1934, are related roughly to the periods of maximal incidence of all cases in the area, which area includes the counties of Moray and Nairn and the burghs situated within the counties. A prolonged epidemic occurred in the whole area from 1933 to 1935, the age distribution of cases at different stages of that epidemic (Douglas, 1935) being given in Table 2. It is regretted that present circumstances do not permit of a more detailed analysis of the age and sex distributions of this epidemic. There was no evidence that the disease was spread by milk.

In the summer of 1940 an outbreak of paratyphoid fever occurred in Kettering and in the districts surrounding it. The total number of cases was 222. Most of them were mild, and the period of vague symptoms before a doctor was called in was in some as long as three weeks to a month. A few cases were more severe, and one or two at the beginning showed toxic symptoms which led to a diagnosis of meningitis. There was one death. The centre of the epidemic was in Kettering, but the surrounding districts were also affected. The population of this area in peace time is about 133,000: it has been somewhat increased by the influx of evacuees. The age and sex distributions are shown in Table 1.

The incidence was greater in women and children than in adult males. Fig. 1 shows in graphic form the number of notifications week by week. In a few instances the agglutination test of the patient's serum was the only examination made but in the majority Bact, paratyphosum B was isolated either from the blood, faeces or urine. The methods used in isolating the infecting organisms will be described in a separate paper. Many of the cases which were excreting small numbers of paratyphoid bacilli for only a short period were detected only by the use of selective ‘enrichment.’ media, and the experience gained in the epidemic showed that a negative result based only on the use of a comparatively non-selective medium such as McConkey agar is of little value and may be dangerously misleading. Most of the specimens were examined at Leicester Emergency Public Health Laboratory, but a few, especially at the beginning, were sent to private laboratories.

Since the original definition of the coli-aerogenes group of organisms by MacConkey (1905,1909), research has been continually directed in an endeavour to find a more rapid and definite distinction between its respective members.

In general, for the purposes of water analysis, the conform organisms can be divided into two main groups: faecal coli, of which Bact. colitype I can be considered the typical member, and the non-faecal organisms comprising the intermediate-aerogenes-cloacae (I.A.C.) organisms. All available evidence tends to show that Bact. colitype I is by far the most frequent organism in the human and animal intestine. Levine (1921) has summarized the figures of various workers and obtained a percentage ratio of 93·4:6·6 as the ratio of coli:aerogenes in faeces, while Bardsley (1938) has shown that in only a very few cases does the number of I.A.C. organisms exceed the faecal coli in faeces.

In a large institution whose population can be considered as forming a semiclosed community there frequently exist conditions favourable to the spread of infectious diseases. The organization of such an institution to avoid epidemics is an accepted difficulty of administration. The occurrence of an epidemic in a semi-closed community offers opportunity for thorough investigation, and is, therefore, of more than usual interest.