1. The potency of PPD tuberculin made from AN 5, a bovine strain, was high when WR medium was used and low when BAI medium was used. PPD tuberculin made from strains of the human type grown on BAI medium was intermediate in potency between the two bovine PPDs.

2. The standard deviation of the difference between responses to duplicate injections of human type PPD at 2·0 mg./ml. concentration was 2·59 in the Islington and 1·84 in the Weybridge experiment on tuberculous animals, and 1·11 at Islington and 1·13 at Weybridge on tubercle-free animals. Differences greater than 1·5 mm. were relatively infrequent on tubercle-free animals.

3. The Elliott ‘401’ electronic computer at Rothamsted was used to calculate discriminant functions and to estimate the discriminating power of different ‘tests’, each test involving an injection of a mammalian and an injection of avian tuberculin.

4. There were some differences between the mammalian preparations in their power to discriminate between tuberculous and tubercle-free animals, but these differences were closely allied to differences in potency. It seems probable that if concentrations of equivalent potency were used, there would be little if any difference in their discriminating power.

5. The contribution of avian tuberculin to the discrimination test was demonstrated; the greatest contribution came from a concentration of 0·25 mg./ml. used in the Islington experiment. The Weybridge results were quite consistent with the assumption that a maximum contribution is made by a concentration less than 0·5 mg./ml.

6. Methods for deducing rules for the interpretation of comparative tuberculin tests, using a mammalian and an avian tuberculin are illustrated and results given for the type of population sampled in these experiments.

7. The principle on which this rule might be modified where the population is believed to be free of infection, is outlined and discussed.

We should like to acknowledge the advice and constructive criticisms during the planning of the experiments of Dr A. W. Stableforth and Dr H. H. Green of Weybridge, and also the care with which Miss C. N. Hebert carried out the preliminary arrangement of the large quantity of data. Thanks are also due to the staff of the Islington abattoir for their willing co-operation.

1. ‘Pure fimbrial antisera’ were prepared for Sh. flexneri strains of O-serotypes 1a, 2b, 3 and 4a and 5, by injecting rabbits with a living fimbriate-phase culture and absorbing the crude immune serum with a non-fimbriate-phase culture of the same strain to remove antibodies for the non-fimbrial (somatic) antigens. These sera gave at high titre a rapid, loosely floccular agglutination of all fimbriate-phase Flexneri cultures, caused adhesion of their fimbriae visible by electron-microscopy and inhibited their haemagglutinating activity. ‘Non-fimbrial antisera’, prepared by injection of non-fimbriate-phase cultures, were devoid of these activities; they gave somatic-type agglutination with fimbriate bacilli at lower titres than with homologous non-fimbriate bacilli, the fimbriae tending to mask the O-antigens and confer relative O-inagglutinability.

2. Heating at 90°C. detached the fimbriae from the bacilli, so that these lost their haemagglutinating activity, fimbrial serum agglutinability and fimbrial agglutinin-binding power. When heated for 21/2 hr. at 100°C., the detached fimbriae retained agglutinin-binding power, but lost their immunogenicity.

3. Cross-agglutination and absorption tests showed that the antigenic composition of the fimbriae was identical in all flexneri strains, irrespective of O-serotype. The flexneri pure fimbrial antisera agglutinated at low titre sixty out of sixty-six fimbriate Bact. coli strains. Absorption tests with a flexneri and three coli pure fimbrial antisera showed that the fimbriae of Sh. flexneri contained a major ‘flexneri-specific antigen’, found in only one coli strain, one or more minor ‘flexneri-coli antigens’ shared by a few coli strains. The coli fimbriae also contained a major, ‘coli type-specific antigen’ shared in groups of several related strains. The flexneri sera did not react with fimbriate strains of Bact. cloacae, Salmonella and Proteus.

4. Fimbrial agglutinins for Sh. flexneri were found in eighty of eighty-one normal human sera at titres from 30 to 1920, and in some pre-immunization rabbit sera at titres of 30 to 60.

A simple method, applicable in routine work, for simultaneous air sampling of total bacteria, on serum agar, and Str. salivarius and Bact. coli on specific media, is described.

Simultaneous air sampling including, in addition to detection of total bacteria, Str. salivarius and Bact. coli as indicators of buccal and faecal air-borne bacteria, respectively, reveals the kind and source of air-borne pollution much better than the simple total bacterial count.

The air in various medical and public institutions was examined; the results are presented and discussed. The air pollution of the premises examined was relatively high, especially in the operating theatres. The air of infants' wards was found to be heavily polluted with Bact. coli as a result of poor nursing technique and this might have had bearing on the epidemic of infant diarrhoea.

Rinderpest neutralizing antibody was found to be transferred from the immune dam to the calf via the colostrum. No antibodies were detected in the sera of calves before suckling. The colostrum of immune cows contained rinderpest antibodies to a higher titre than that of the serum; 30–48 hr. after the ingestion of such colostra, newborn calves possessed high antibody levels in their sera, levels greater than those of their dams' sera but less than those of the colostrum ingested.

After the neonatal period the serum titres of calves declined linearly. The mean half-life of maternally derived rinderpest antibody in calves was 36.7 days and the extinction point 10.9 months.

Two young calves, from susceptible dams, which each ingested daily for 5 weeks 1 gallon of milk containing rinderpest antibodies failed to show evidence of their absorption from the intestinal tract.

Calves which were the progeny of rinderpest-susceptible dams were inoculated with lapinized rinderpest virus when 1 day to 2 months old. Their serological response, as measured by neutralizing antibody titre 21 days later, did not differ from that of adult susceptible cattle.

In calves born of rinderpest-immune cows, the serological response to caprinized rinderpest virus depended upon the level of maternally derived antibody. There were two critical levels, one below which all calves produced antibodies (serum titre of 0.7 or less) and another above which no antibody was formed (serum titre of 2.2 or more). Thus all animals aged 8 months or more at the time of inoculation were actively immunized but no calf aged 3 months or less reacted in this manner.

The response of calves with maternally derived antibody titres between 0.9 and 2.0 varied. There was a significant inverse relationship between the preinoculation titre and the amount of antibody formed during the following 3 weeks. A similar inverse relationship was shown between the pre-inoculation titre and the rate of production of antibody and its titre 1 year following inoculation.

Calves which possessed colostral antibody, and which were actively immunized by caprinized virus inoculation, did not necessarily show the usual clinical reaction. When such animals failed to become actively immunized there was no sensitization of the antibody-forming mechanism, a fact demonstrated by the lack of an anamnestic response to subsequent exposure to rinderpest virus antigen.

I owe thanks to many people: to Dr S. E. Piercy, Deputy Director, E.A.V.R.O., for advice and encouragement; to Mr G. R. Scott, Senior Virologist, E.A.V.R.O., for advice and encouragement and, with Mr G. J. Knight, for help in the statistical analysis of the data; to Dr M. H. French lately head of Division and Mr G. Lampkin, Joint Animal Industry Division of E.A.A.F.R.O. and E.A.V.R.O. for facilities for the inoculation of calves from rinderpest-immune dams with rinderpest vaccine; Mr J. M. Nightingale, Sasamua Estate, South Kinangop, and Mr J. Armstrong, Naivasha, for facilities for immunizing calves from rinderpest-susceptible dams; Mr C. S. Rampton, Mr N. Kerani and Mr F. Mwithiga for technical assistance.

This and the previous paper (see p. 427) are published by permission of the Director, East African Veterinary Research Organization.

1. The rat population of a main sewer was studied by means of census baits.

2. In a series of poison treatments extending over 4–5 months all rats visiting the access points were eliminated.

3. At the end of this period the population was 40–65% of what it was at the beginning.

This work would not have been possible without the generous assistance once more of Mr A. R. J. Vickery, Assistant District Engineer of the Chief Engineer's Department of the London County Council and members of his staff, and without the ready co-operation of the Borough Engineer, the Medical Officer of Health and the Chief Health Inspector of the Metropolitan Borough of Bethnal Green, through which area the sewer ran.

Suppose that λ is the average density per unit volume in a suspension of infective particles such as virus particles. To estimate λ the usual method is to make up a series of inocula of various dilutions containing expected numbers of particles …λi−1, λi, λi+1,… which are known multiples of λ. Each of these is then tested by inoculation in a host such as an egg. We consider only the case where the dilution series is twofold (λi=λ2i, say) and the same number of eggs, N, is tested at each dilution. Then if we are sure that an egg is infected if and only if the inoculum contains at least one infective particle, the probability that the egg remains sterile is Pi=exp {− λ2i}. If each particle is not certainly infective but has a probability p of infecting the egg, the probability of sterility of an egg chosen at random is exp {− λp2i} provided that p does not vary from egg to egg. It is then only possible to estimate λp from the results.

The air of operating theatres ventilated by powerful extractor fans was sampled during operations with special media for Cl. welchii and Staph. aureus, and also for total organisms. Counts made with the slit-sampler showed the presence of Cl. welchii in all samples in a range slightly higher than that of Staph. aureus (mostly between five and twenty-five colonies per 100 cu.ft.). There was no evidence of a build-up of either Cl. welchii or Staph. aureus during the course of operations, nor were there any consistent peaks of contamination during operations.

Settle plates exposed on numerous days in the operating theatres showed similar counts of Cl. welchii and Staph. aureus—the latter slightly more abundant in this series. Settle-plate counts from many parts of the hospital and from the road outside provided evidence that airborne Cl. welchii comes into the hospital from outside, while Staph. aureus is contributed to the air inside the hospital. This view was supported by study with the slit-sampler and with settle plates in a dressing station ventilated with filtered air, where Cl. welchii was rarely found and Staph. aureus momentarily abundant during dressings; staphylococci but not Cl. welchii were dispersed by the operators in a mock operation. The numbers of airborne Cl. welchii inside and outside the hospital were higher in dry than in rainy weather.

Of the patients in the Burns Unit 35% had at some stage Cl. welchii in their burns; contamination was commoner in the more extensive burns. In contrast to Staph. aureus, Cl. welchii was usually present in small numbers and for a short period.

Cl. welchii was found in the burns of twenty out of thirty (66%) patients whose rectal swab taken on admission showed the presence of Cl. welchii; a significantly smaller proportion of patients whose rectal swabs were free from Cl. welchii picked up the organism while in hospital (eight out of thirty-two, 25%; x2 = 9.4, P < 0.01). This observation supports the view that burns and possibly other wounds acquire Cl. welchii by self-infection at least as often as from the environment. While it is obvious that plenum ventilation of operating theatres with filtered air is desirable, it is uncertain to what extent this will reduce the small hazard of post-operative gas gangrene.

We wish to thank Miss G. Hogg and Mr L. Hurst, F.I.M.L.T. for technical assistance, the surgeons and nursing staff of the Birmingham Accident Hospital for their co-operation, Mr N. R. Gill for photography, and Professors A. A. Miles and C. L. Oakley for kindly sending us strains of Cl. welchii.

The growth of Salmonella typhimurium in mice was studied by performing viable counts on pooled viscera and, in the case of fatally infected mice, by determining the relation between size of dose and mean death time. The results suggested that this system conformed to a general model which postulates (a) that the organisms causing the response (death in the present experiments) increase in vivo at a rate which is constant for all doses, and (b) that the response is certain to occur when their number reaches a constant figure. It follows that logarithm of dose (for doses > 1 ED50) should be linearly related to mean latent period (the interval between inoculation and response). A search of the literature showed that this relation was observed in many reported infective systems.

The interpretation of the observations on partially resistant mice (LD50 = 102.5−106 organisms) was not straightforward for the hypothesis of independent action (Meynell & Stocker, 1957) states that only a random fraction of the inoculum usually initiates each response. Multiplication of this fraction is considered to cause a reduction in host resistance so that other organisms surviving in vivo began to increase in numbers and to contribute to the fatal infection. The observations showed that this could well occur as (i) the counts on all mice, whether fatally infected or not, increased exponentially for 2–3 days after inoculation; (ii) the growth curves of fatally infected mice showed exponential increase of the whole inoculum as postulated; and (iii) the length of the death time was found to be determined by the growth of the whole inoculum and not by the predicted random fraction. The viable count per fatally infected mouse at the time of death was approximately constant at 108.75 organisms. There was no evidence that the course of the infection was influenced by organisms which may have been killed in vivo as the requirements of the model were satisfied by the observed behaviour of the viable organisms alone. The rate of increase of organisms in a fatally infected animal was found to be less, the greater the LD50 of the host; which suggested that the degree of host resistance was determined by successive random events occurring after inoculation.

The latent period/log dose curve provides a new test of the hypothesis of independent action, since this predicts that most responses to doses equal to or less than 1 ED50 are each initiated by only one inoculated organism. Hence, the latent period should become approximately constant at such doses. This was found to be so in the present experiments.

The authors are grateful to the Medical Research Council for meeting part of the expenses of this work by a grant to G. G. M. They also wish to acknowledge the help of Dr W. R. Bryan, Dr B. Mandel, Dr D. B. W. Reid and Dr G. T. Stewart who provided unpublished data.

1. Ultrasonic vibration has been shown to be a highly effective means of releasing vaccinia virus from infected HeLa cells.

2. The titre of vaccinia virus obtained from a suspension of infected HeLa cells following ultrasonic vibration is closely similar to that obtained by rapid shaking of the suspension with glass beads. Since these methods are dissimilar in their disruptive action and the titres obtained are higher than those obtained by other methods, they are probably close to the maximum possible yield of virus from these cells.

3. Ultrasonic vibration and shaking with glass beads give higher yields of vaccinia virus from infected HeLa cells than grinding with sand, grinding in a glass TenBroeck tissue disintegrator and three cycles of freezing and thawing.

I am indebted to Prof. H. B. Maitland whose interest stimulated this work. I also wish to thank Dr J. C. N. Westwood for much helpful criticism and Mr R. F. W. Southey for his technical assistance.

A comparison has been made between the bacteriophage and serological methods for identifying types among coagulase-positive staphylococci. In only a few cases was one particular serotype clearly related to a particular phage-type, and in several cases single ‘types’ recognized by one method contained several different ‘types’ when tested by the second method. Nevertheless, when sets of strains isolated in the investigation of epidemics or of the nose and skin carrier state of particular individuals were tested by the two methods, consistent results were obtained: an individual combination of phage-type and serotype appeared to be stable.

The principal advantages of phage-typing are the facts that it is able to recognize more types than the serological method, and that the distinctions between the types are based on a greater number of different reactions. The advantages of the serological method are the smaller number of untypable strains and its greater technical simplicity.

Thanks are due to Miss Joan Rippon, Ph.D., for assistance in the analysis of the phage-typing results.

An investigation has been made of methods of estimating the antitoxin content of small samples of human sera by haemagglutination procedures.

The direct agglutination test appears to be satisfactory when dealing with groups of sera, but was unreliable for testing individual sera.

The discrepancies observed in the direct agglutination test were greatly reduced by adopting an agglutination inhibition test at the LA/100 level of test.

The difference in agglutinating capacity of both avid and non-avid sera per in vivo unit of antitoxin did not appear to be related to avidity.

Field trials were conducted in the San Francisco Bay region to test the effectiveness of 5 and 10% DDT powder in rodent bait-boxes against fleas on native wild rodents and in the rodent nests. The principal species concerned were known flea vectors of sylvatic plague, Malaraeus telchinum and Hystrichopsylla linsdalei, and the meadow vole, Microtus californicus, an important plague reservoir. Preliminary trials with the bait-boxes showed that Microtus and other small native rodents would enter them quite readily and feed upon rolled oats. About 90 g of either 5 or 10% DDT powder, placed on the floor boards of the bait-boxes, was picked up by the rodents and resulted in effective immediate control of all flea species. Residual control of all flea species on Microtus (with the possible exception of Hystrichopsylla linsdalei) was obtained over a period of several months. Analyses of Microtus nests showed considerable amounts of DDT which had been transported to them by the rodents. All flea species in nests were effectively controlled by 5% DDT for at least 132 days after the DDT was removed from the area.

The field work was performed by Alva R. Kinney and Robert L. Martin. The fleas were identified by Harold E. Stark. The DDT analyses of rodent nests were done by the personnel of the Technical Development Laboratories, CDC, Savannah, Georgia, under the direction of J. A. Jensen.

The disappearance of the resistant mutant from an artificial mixture of normal antibiotic-sensitive organisms and a resistant mutant after varying numbers of subcultures is known as the Welsch phenomenon. Welsch's observations have been confirmed and experiments devised to show that the resistant strain disappears because it grows more slowly than the normal sensitive strain which eventually, during the successive subcultures, overgrows it.

1. Sera taken from seventy-nine individuals who had been examined for poliomyelitis antibodies shortly before a Type 1 epidemic, were examined for the development of antibodies to polio-virus 6 months after the epidemic.

2. Ten of eleven (91%) susceptibles aged 5–7 years, 44% of those aged 8–9 years, 25% of those aged 10–14 and only one of fifteen (7%) 15 years or more old had developed neutralizing antibodies to Type 1 polio-virus.

3. In the school which the younger subjects attended, one boy developed paralytic poliomyelitis while the school was in session, but the amount of minor illness in the epidemic period was not significantly different from the 3-year average.

4. The comparison between these findings and the notification rate for the province suggested that the chance of an infection with polio-virus during this epidemic, leading to a clinically recognizable attack of the disease, might have been as much as eighteen times higher in adolescents and adults than in the 5–9 age group.

5. The epidemiological significance of these results is discussed.

We have pleasure in acknowledging our indebtedness to Dr J. H. S. Gear for the report on the sera collected before the epidemic, to Dr W. Murphy, Medical Officer of Health, for data about notifications, to Dr A. M. Douglas for assisting by collecting some of the bloods, to the Principal and staff of John McGlashan College for their readiness to assist us at all times and to all those subjects who allowed us to take a further specimen of their blood for this investigation.

Ringtail in infant rats is shown to depend on humidity over a wide temperature range. It is not due to infection or diet, but probably to chronic cooling by evaporation, convection and conduction.

Rats should either be bred in cages with solid floors, or else be given environmental conditions resembling those in the natural nest, which keep the young at a low saturation deficit during the early post-natal period.

I am glad to express my thanks to M. Potter, Senior Animal Technician, for his invaluable aid in keeping the records of the experimental programmes, and to G. Dimmock for his help in building and maintaining apparatus. I am also grateful to the University of London for the loan of a Cambridge portable potentiometer.

1. In a children's home with an average population of about 460 children aged from 0 to 15 years, 473 attacks of acute sore throat, 354 attacks of otitis and 314

attacks of other acute febrile respiratory tract illness were recorded in the 30 months from November 1950 to April 1953.

2. The attack rate for all forms of illness and at all ages was higher among children in a reception group, who were mostly recently arrived in the home, than it was in the permanent residents.

3. Of all ifinesses with the principal signs or symptoms referable to the throat, the proportion that yielded Streptococcus pyogenes rose from about 34 % in the 1–2-year- old children to about 74% in the children aged 10 years or more. Streptococcal sore throats made up about 8% and about 38 % of all febrile respiratory tract illnesses in the same two age groups.

4. The attack rate for streptococcal sore throat varied from month to month between 0 and 6 5 % of the children at risk; only one period of substantially increased prevalence was observed and there was no indication of any seasonal trend. Non-streptococcal sore throat varied between FO and 20 % per month. It was not possible to recognize any epidemics of streptococcal otitis; non-strepto coccal otitis had a definite winter prevalence associated with the rise in the total of non-streptococcal respiratory disease.

5. There was a slight indication that tonsillectomized children had rather milder attacks of sore throat than non-tonsifiectomized under the routine sulphamezathine treatment adopted. About 86 % of all children stifi harboured the causative streptococcus in their throat on discharge from hospital. The tonsifiectomized children lost their throat, but not their nose, streptococci in convalescence more rapidly that the non-tonsillectomized. Among throat carriers, the tonsifiectomized children carried streptococci in their nose more often than the non-tonsilectomjzed at all stages in their illness and convalescence.

6. Bacteriological examination of excised tonsils did not often reveal the presence of streptococci that had been undetected in throat swabs. Twenty-one of twenty- two children who had had a streptococcal respiratory tract ifiness treated with suiphamezathine within the 3 months preceding their tonsifiectomy were found to harbour the same streptococcus in their tonsifiar tissue; only four of eighteen children whose illness had been treated with penicillin still harboured the streptococcus.

Streptococcal infections in children. I 61

1. Of 318 Polish refugees now living in this country 30% gave a history of typhus fever; nearly half of these had complement-fixing antibodies to a titre of 1–5 or more to epidemic typhus. A further 12% without such a history similarly had these antibodies making a total of 22% of 318 refugees. None of the 174 British born controls had antibodies up to this titre.

2. Brill-Zinsser disease may occur from time to time among immigrants or visitors from typhus endemic areas to this country, but it is unlikely to be a serious problem.

3. The clinical diagnosis of Brill-Zinsser disease may not be easy because the characteristic rash is not always to be seen. Any patient of Eastern European origin or who has a history of typhus complaining of fever and severe headache should be examined serologically by the rickettsial agglutination or the complement-fixation test for specific epidemic typhus antibodies. The Weil-Felix reaction may be positive, but a negative result will not exclude a diagnosis of Brill-Zinsser disease.

We acknowledge with gratitude the help, interest and advice of Dr E. S. Murray, of the School of Public Health, Harvard University and Dr Janet Niven and also the technical assistance of Mr J. P. Brooks, FIMLT. We have to thank Messrs Lederle for their kindness in supplying the rickettsial antigens and in addition those who helped us to obtain the specimens and the volunteers who supplied them.

Children in a residential home whose tonsils had been removed had lower attack rates for streptococcal sore throat than tonsillectomized children. Tonsillectomy did not have a consistent effect on respiratory tract illness.

Tonsillectomy has a similar effect on the outcome of individual exposures to streptococcal infection; the difference was more marked for illness than for simple colonization.

The apparent benefit from tonsillectomy could not be attributed to the ageing of the child, and was considered real. Its magnitude did not seem, however, great enough to justify a policy of extensive tonsillectomy, particularly in view of the possible risks to the child.

A child who had been ill with infection due to one streptococcal type rarely became ill when re-exposed to the same type. In addition there was some evidence in children with tonsils of an immunity, perhaps short lived, which was clearly not type-specific.

Following primary vaccination no antibody was detected up to the tenth day. After this time neutralizing antibody and antihaemagglutinin were present in the majority of individuals, while, with the technique used, complement-fixing antibody was found in less than half those examined. Neutralizing antibody may be found in the blood more that 20 years after primary vaccination; by this time very few show antihaemagglutinin and then only in low titre. Complement-fixing antibody was not found more than 6 months after primary vaccination.

In revaccinated individuals the antibody response tended to be higher than after primary vaccination and, when it occurred, often appeared within a week. This suggests that prompt revaccination of smallpox contacts may offer even greater protection than does primary vaccination.

The neutralization test is a better measure of immunological response than are either of the other two tests, for these may give a negative result even though there may be a well-marked rise in neutralizing antibody. The ‘early’ type of reaction to revaccination was associated with an antibody response in only about half the subjects examined. In the remainder, the skin reaction, although identical in appearance, must be regarded as allergic in nature and not accompanied by increase in immunity.