Ninety-seven groups of five-membrane pock counts from 115 titrations of vaccinia virus preparations were examined in an attempt to identify the statistical distribution followed by such counts, and to obtain an estimate of the overall coefficient of variation of the technique.

It was found that the distribution of counts was neither Poissonian nor normal, and that the count variances ‘wandered’ so extensively that no valid estimate of the overall coefficient of variation was possible. Nor was it possible to assess the intrinsic reproducibility of the test, although it is empirically a useful one.

The growth curve of rabbit skin-adapted vaccinia virus in the chick chorioallantoic membrane incubated in Hanks' solution showed a drop in titre of virus for about 10 hr. followed by growth. At least 25% of virus, sometimes more, remained infective. A similar fall in titre was observed in heated membranes in which the virus did not grow and this occurred also when membranes, either normal or heated, were infected and disintegrated before incubation.

The growth curve of virus in minced chick-embryo was similar to that in chorioallantoic membrane.

Virus in cell suspensions prepared from chick embryo and incubated in a nutrient medium showed only a small loss of infectivity before growth in some experiments and rarely dropped below 65–70 % of the original titre in others.

These results throw considerable doubt on the view that loss of infectivity preceding growth of vaccinia virus should be interpreted as an essential part of a growth cycle.

Data concerning numerous severe epidemics of poliomyelitis were surveyed for information useful in estimating the size of the immune component of the population by age groups of 1 year. The New York epidemic of 1916 and the Mauritius epidemic of 1945 were chosen as the most suitable for this purpose. It is shown that there was a regularly progressive decline in attack rates for successively older age groups from 2 to 8 years in New York. The attack rate in 8-year-olds was less than 10 % as large as that in 2-year-olds. It is noted that a difference of this magnitude could be accounted for by immunizing infections amounting to 30 % per year for 6 years.

A similar analysis of age specific attack rates during the Mauritius epidemic shows progressive declines of 28, 60, 50 and 53 % for the successive age groups 5–9. An average annual infection rate of 45 % over a 4 year period could account for the ten-fold difference in infection rates between 5-year-olds and 9-year-olds.

An immunizing infection rate of 30 % a year would lead to a pattern of immunity in which the seven youngest age groups had a total susceptible component of 30 % at the beginning of a ‘ poliomyelitis season’. Twenty per cent would remain susceptible at the end of the season when spread of virus terminated.

An annual immunizing infection rate of 45 % would bring about a situation in which 21 % of the five youngest age groups were susceptible at the start of a period of viral prevalence and 12 % at the end. Alternatively, one could consider that on Mauritius there may have been a continuous prevalence of virus in a population in which approximately 15 % of the five youngest age groups were susceptible at any time.

It is suggested that the essentially similar attack rates among children 3–5 years old on Mauritius may have reflected a 3-year period during which homotypic virus was not prevalent, in contrast to its great prevalence during the years prior to that time.

A more detailed discussion and analysis of additional data concerning age-specific attack rates in poliomyelitis will be found in the thesis submitted by Dr Sample to the University of Washington School of Medicine entitled, ‘Some observations on statistical and theoretical epidemiology of infectious diseases, principally poliomyelitis’, in 1955. This is obtainable by Inter-library Loan from the Health Sciences Library, University of Washington, Seattle.

We wish to express our sincere thanks to Dr Blair M. Bennett and Dr William E. Reynolds of the University of Washington for their advice on the statistical and epidemiologic aspects of this study.

Experiments were performed which show that virulent strains of Salmonella paratyphi A contain an additional antigen which has all the properties of a Vi antigen.

The anti-Vi immune serum combined in suitable proportions with an anti-0 immune serum exerted maximal bactericidal effect on virulent strains.

An outbreak of encephalomyclitis of unknown actiology occurred in the Royal Free Hospital Teaching Group in 1955. The disease which affected about 9% of the population was probably a viral infection spread by personal contact. People of all ages were attacked but those under 30 were most susceptible. Institutional life was an important factor in determining the case incidence which was highest among nurses, orderlies and resident domestic staff. The clinical picture was one of encephalomyelopathy, with lympho-reticular reaction and relapsing myalgia.

A study has been made of various factors involved in performing neutralization tests with vaccinia virus on chorio-allantoic membranes.

The common practice of using a single serum concentration has been shown to result in a test which is qualitative only. The linear relationship between the logarithm of the pock count of the surviving virus and the logarithm of the serum dilution can be used to determine the dilution of serum which will neutralize 50% of a virus suspension. By this method it is possible to distinguish between sera whose potencies vary by about fourfold.

I gratefully acknowledge the advice and criticism given by Dr J. C. N. Westwood and Dr I. A. Macpherson, as well as the help given by Mr S. Peto with the statistical treatment of results. My thanks are also due to Mr H. B. Maber and Corporal M. Fitzpatrick, R.A.M.C., for their technical assistance.

Treatment of serum with the insoluble polysaccharide residue of yeast cells (zymosan) has been shown (a) to decrease the bactericidal power of serum for certain bacteria; (b) to remove from some sera a factor essential to the interaction of zymosan with the third component of complement. The adsorbed protein (properdin) was eluted from zymosan and found to contain α2, β and some γ globulins. It was shown to restore partially the normal characteristics to treated serum (RP).

Properdin levels have been estimated in the sera of eighteen burned patients and in twelve normal subjects. There was evidence of a reduced level in the patients' sera. The limitations of the assay method are discussed.

Bactericidal activity of normal serum containing properdin was tested against nineteen Gram-negative bacilli isolated from burns. Ps. pyocyanea, Proteus and Klebsiella, which are commonly found in burns, were with one exception resistant or slightly sensitive to the bactericidal action of this serum. Bact. coli, which is uncommon in burns, was usually highly sensitive.

From our preliminary studies reported here it would seem that the properdin system cannot be expected to kill the flora in established infections of burns, and properdin is therefore unlikely to prove useful as a therapeutic agent. However, it probably has an important role as one of the defence mechanisms of uninfected tissues.

We wish to thank Dr Derrick Rowley and Dr A. C. Wardlaw for helpful discussions, and Dr W. Weiner for his advice and co-operation. We are grateful to Miss Stella Baar, F.R.I.C., for N and P analyses.

A study has been made of the properties of soluble antigen in the brains of infant mice infected intracerebrally with the Flury strain of rabies virus.

Soluble antigen is produced at the same time as infective virus, and reaches a high concentration in a period of 2–3 days.

It can be partially purified by precipitation at pH 4·3. It is partially resistant to the action of trypsin, RNAse and DNAse. It is relatively stable at pH 6–10.

Experimental results suggest that the soluble antigen remains antigenically active after heating at 56° C. and treatment with 0·5% phenol or 0·35% formal-dehyde, but that such heating markedly reduces the ability to stimulate formation of neutralizing antibody.

Rabbits and mice appear to differ in the production of neutralizing antibody following immunization against soluble antigen in which residual live virus was inactivated by heat, phenol or formaldehyde.

It is suggested that this difference may depend on the different susceptibility to traces of incompletely inactivated virus remaining in the immunizing antigens.

The authors are grateful to Miss T. Madsen for her assistance in some aspects of this work. Dr N. Sapeika kindly made available facilities for the in vitro anaphylaxis experiments.

Financial assistance was received from the Nkana-Kitwe and Chingola Poliomyelitis Research Funds.

In seeking a more satisfactory test than the conventional mouse-protection tests for the assessment of typhoid-paratyphoid vaccines in relation to their O and Vi antigens, it was decided to undertake a series of experiments with vaccines containing the somatic antigens of Salmonella paratyphi C and to use this organism to challenge mice thus immunized. Salm. paratyphi C contains the same Vi antigen as Salm. typhi and moreover it is capable of giving rise to a true infection in mice in contrast to the acute intoxication which follows a fatal dose of Salm. typhi.

Large doses, of the order of 100 × 106, of this paratyphoid organism injected intraperitoneally into mice resulted in acute intoxication with death in 24–72 hr., similar in every way to the picture with typhoid bacilli, but when the dose was reduced 100-fold or more an infective illness was set up which after an interval of 5–10 days or more, during which the organisms multiply freely in the tissues, terminated fatally. Typhoid bacilli do not produce a true infection in mice.

Living vaccines were prepared of Salm. typhi Ty6S (an almost pure Vi strain) and Salm. cholerae-suis (contains the same O-antigen complex as Salm.paratyphi C). These were used separately, and in combination, to immunize different batches of mice by the subcutaneous route against a subsequent intraperitoneal challenge with a Vi-positive culture of Salm. paratyphi C at both high- and low-dose levels. The different challenge doses were to assess the relative values of the 0 and Vi components in protecting against intoxication (high dose) and infection (low dose), respectively.

Against intoxication pure Vi vaccines were almost as effective as Vi + O vaccines, whereas pure O vaccines gave little or no protection.

Against fatal infection the combined Vi + O vaccine offered good protection and the pure O vaccine was but little inferior, whereas the pure Vi vaccine appeared to be without significant effect.

Quite small doses of Salm. paratyphi C administered by the intra-gastric route can give rise to a general infection in mice, but in order to obtain consistent results much larger doses were found necessary (about 300 x 106 organisms). Mice immunized with the same three vaccines and subsequently challenged by the intragastric route with this large dose showed significant protection throughout. The combined, O + Vi, vaccine did not appear to have any advantage over the pure O vaccine, but both were markedly superior to the pure Vi vaccine. Further experiments on the same lines are however necessary to see if the results obtained in this single experiment are reproducible.

The relationship of the above findings to the evaluation of agents employed in man for the prophylaxis of enteric fever and the need for further experimental work of a similar nature are discussed.

We wish to thank Dr F. Kauffmann for strains of Salm. paratyphi C, Lt-Col. T. E. Field, whose results we have freely quoted, and Dr S. Peto for advice on statistical evaluation of our results.

The Mahoney, Type I, poliomyelitis virus undergoes two ‘cycles’ of multiplication in the CNS of suckling mice—the first within 48 hr. after intracerebral inoculation, and the second approximately 10 days later. The former may not be associated with paralysis whilst the latter usually is.

By to-and-fro passage in suckling mouse brain and monkey kidney-tissue culture there is a gradual increase in the amount of virus obtained during the early multiplication cycle in suckling mouse brains and associated with this is the occurrence of early paralysis in 80% or more of these mice. The virus capable of causing early paralysis in infant mice is gradually lost by direct mouse-to-mouse passage.

The occurrence of two variants is postulated, one of these apparently nonparalytic and multiplying selectively in suckling brains.

The author would like to thank the late Prof. M. van den Ende for his stimulating discussions and advice, Dr P. K. Olitsky of New York for his constant interest, and Mr T. C. Norcott for his able technical assistance.

1. The A.S.O. titre of cord blood is about 1½ times that in the mother's blood taken 4 weeks before delivery.

2. The distribution of A.S.O. levels in cord bloods from multiparae with children at school is not significantly different from that in multiparae whose children are under school age, or in primiparae.

3. Among these groups, only primiparae show a significantly increased A.S.O. level in winter.

4. The distribution of A.S.O. levels among the five social classes defined by the Registrar-General's occupation code showed a significant trend towards higher levels in the lower social grades.

We wish to thank Mr D. M. W. Maxwell, F.R.C.S., F.R.C.O.G. and Miss D. Slade, S.R.N., S.C.M. at Taplow, and Professor J. C. McC. Browne, F.R.C.S., F.R.C.O.G., Dr W. Hayes, D.Sc., F.R.C.P.I. and Miss M. Tweedle, S.R.N., S.C.M. at Hammersmith Hospital for their kind help.

A detailed investigation has been made of the pock-counting method for titrating vaccinia virus on embryonated eggs with a view to eliminating the wide scatter of results hitherto inseparable from the technique, and of assessing the precision of which it is capable. A detailed description is given of the technique evolved.

Reasons are given for supposing that the distribution of counts should follow the Poisson series, and considerable progress is claimed towards the achievement of such a distribution.

The duration of bacteraemia in guinea-pigs following the subcutaneous inoculation of brucellae can be used as a measure of the virulence of the strains. Strains of low virulence (e.g. Brucella abortus S19) are almost always cleared from the blood stream within a week, whereas virulent strains produce a bacteraemia that regularly persists for 4 weeks and may last very much longer.

All of four strains of Br. melitensis isolated from cow's milk in Great Britain behaved in the same way as the typical and virulent strains of Br. abortus and Br. melitensis tested, regularly producing prolonged bacteraemia. These results do not support the suggestion that the strains of Br. melitensis isolated in this country are of low virulence.

The influenza virus serum flocculation previously reported (Belyavin, 1955) opened up a number of lines of investigation. One of obvious importance was extension of the reaction to other viruses belonging to both related and unrelated groups. Indeed, other workers in this laboratory have already achieved the flocculation of poliomyelitis viruses by specific antisera (Smith, Sheffield, Lee & Churcher, 1956) and flocculation of both mumps and Newcastle disease viruses is now reported in this communication. The ease with which the viruses of the mumps-influenza group can be flocculated by homologous rabbit antisera suggested that the technique may be applicable as a method of antigenic analysis. If so, it would have the advantage of being much simpler than the standard haemagglutination inhibition and complement-fixation tests. The exploration of this possibility forms the basis of this paper. A large-scale antigenic survey involving numerous virus strains has not been attempted, greater emphasis being placed on the examination of techniques and their applicability to the end in view. The investigation has also revealed new phenomena peculiar to the direct virus flocculation reaction.

1. When GDVII virus was injected into the brains of mice, subsequent intracerebral injections of mixed pertussis and diphtheria prophylactic significantly shortened the incubation period. Subsequent intramuscular injections of mixed prophylactic had no influence on the course of the infection.

2. Intramuscular inoculation of either GDVII virus or EMC virus results in a marked localization of paralysis in the injected limb. With GDVII virus, the localization probably arises from the use of the sciatic nerve route for entry into the c.n.s., and is unaffected by injection of substances into the opposite limb. With EMC virus, localization can be modified by injections into the opposite limb.

3. When GDVII or EMC virus was injected by a peripheral route other than intramuscular, intramuscular injections of vaccines or normal mouse brain suspension induced localization of paralysis in the injected limb.

4. Intramuscular inoculation of vaccines increased mortality in mice injected into the same limb muscle with GDVII virus. Within the limits tested the later the vaccines were injected relative to the virus, the greater was the effect.

5. When the vaccines used were more than a few months old, this facilitating effect was replaced by a protective effect. Mice injected intramuscularly with aged vaccines were significantly more resistant than the controls to GDVII virus injected into the same muscle.

6. Intracerebral inoculation of vaccines, and to a lesser extent other substances, increased mortality in mice which had received intramuscular inoculations of GDVII virus, by allowing virus to pass directly from the blood into the brain. This resulted in a predominance of cerebral symptoms when the vaccines were injected immediately after the virus, but not when the interval was lengthened to 48 hr.

7. The mode of action of modifying factors in neurotropic virus infections is discussed in the light of these results.

This work was done in the Department of Zoology, Oxford University, during the tenure of a grant from the Medical Research Council. I also wish to thank the Nuffield Foundation for their support of the virus research carried on in the Department.

I am most grateful to Dr F. K. Sanders for his advice and criticism, as well as for provision of technical facilities, and to Dr Donald Michie for assistance with the statistical treatment of the data, and for much other help.

The behaviour of sheep cells sensitized with tetanus toxins and toxoids in the presence of antitoxin has been investigated.

Sensitized sheep cells were agglutinated in the presence of tetanus antitoxin. The agglutinin titres of such direct agglutinin systems were roughly proportional to the antitoxin content of the sera with some marked exceptions. The discrepancies were associated with the presence of non-specific agglutinins (antibodies other than specific antitoxin) and with the immunization history of the horses providing the sera.

When sensitized cells were used as indicators of the presence of free antitoxin in toxin or toxoid-antitoxin mixtures, the correlation with in vivo tests was quite good at the L+ level of test, but not so reliable at the 0·01 L+ level of test.

Sensitized cells preserved with formalin were found to be more reliable and the results with such materials were more reproducible than those obtained with freshly prepared sensitized cells.

Cells sensitized with different concentrations of antigen gave different titres by direct agglutination with simple dilution of antitoxin; this difference was very much less in haemagglutination inhibition tests.

The LA dose (the volume of toxin or toxoid which when mixed with one unit of antitoxin gave a + + ± end-point) varied according to the sensitivity of the suspension used and the purity of the components of the reacting mixtures.

Prozones (zones of agglutination inhibition in the region of antibody excess) were found to be associated with the immunization history of the horses from which the sera were obtained. In general the more thorough the immunization the longer the prozones found. The length of prozones was proportionately increased with decrease in the concentration of sensitizing antigen. Possible reasons for the behaviour of certain antitoxins in these tests have been discussed.

I am indebted to Prof. M. Raynaud of the Pasteur Institute, Garches, for the supply of tetanus toxoids of high purity.

I am very grateful to Mr Charles Newman for considerable technical assistance.

1. Influenza virus can be labelled with 35S by cultivation in fertile eggs into which radioactive methionine has been introduced.

2. When virus labelled with 35S is fractionated by ether treatment the nucleoprotein soluble antigen fraction contains 15 % of the 35S, the haemagglutinin also contains 15%, while the denatured envelope protein contains 70% of the 35S.

3. When sulphur-labelled virus is introduced as a primary inoculum in fertile eggs and extracts of the chorioallantoic membrane are made 1–20 % of 35S is recovered as amino-acid, while the remainder is present as non-infective material of particle size similar to that of the original virus.

4. It is suggested that on entry into the cell the virus nucleoprotein is hydrolysed with the release of amino-acid and free nucleic acid, while the virus envelope protein and haemagglutinin remain on the cell surface.

The authors are indebted to the Medical Research Council for a grant in aid of expenses, and to Messrs. Eli Lilly and Co. for the generous provision of the Spinco centrifuge used in this work. The radioactive methionine was obtained from the Radiochemical centre at Amersham.

1. The electrostatic precipitation of single cell bacterial aerosols has been studied. Two types of electrostatic precipitators of the wire-in-tube construction have been developed. In one type the collector is a tube with nutrient agar, in the other a vertical water-film.

2. Collection efficiencies of 90 % and more could be obtained at sampling rates of 25 l./min. and tensions varying from 5 to 10 kV.

3. The survival of organisms in electrostatic precipitators has been measured in comparison with other samplers, viz. the slit sampler, membrane filters and impingers. The equivalence of the electrostatic precipitators and other samplers in sampling spore aerosols could be demonstrated. Aerosols containing vegetative organisms were sampled equally well in the agar tube precipitator as in the slit sampler. In the precipitator with water-film collector there is a loss of c. 25 % of viable cells.

4. The practical advantage of the electrostatic precipitators is their very low air resistance.

The authors feel greatly indebted to Prof. Dr J. A. Cohen and Dr F. Wensinck for their helpful criticism of the manuscript, to Mr W. de Vries who designed and constructed the generator for extremely high tension; to Mr R. van Strik, Philips-Roxane (Weesp, Netherlands), for statistical advice and computations and to Miss H. Kranenburg for technical assistance.

Quantitative studies have shown that the haemolytic activity of canine complement was in all essentials similar to guinea-pig complement. Canine serum which had been heated at 56° C. markedly enhanced the haemolytic activity of canine complement. The complement-enhancing power of heated canine serum was found to be increased with the concentration of serum up to 20%. Further concentration up to 50 % serum dilution, or the use of undiluted serum, resulted in a reduction of this effect.

Serum aliquots were either pro-complementary or anti-complementary to cn, depending on whether they were heated at 56 or 62° C.

Factors that possibly affect the complement-enhancement were investigated and evidence is presented which indicates that the electric charge on the protein particle is probably responsible for either the pro-complementary or anti-complementary serum activity; the mechanism of this, however, still remains obscure.

The present work shows also that canine complement can be used instead of guinea-pig complement for testing canine sera with virus antigen by the complement-fixation reaction, an advantage of this being that the use of canine complement permits the testing of canine sera which are anti-complementary to guinea-pig complement.

The data obtained are discussed in relation to canine complement and its specific activity.

The authors are indebted to Professor W. I. B. Beveridge for his constructive advice, to Dr R. R. A. Coombs for his helpful comments on the complement activity, and to Dr Christine E. Rice of the Animal Diseases Research Institute, Canada, who has generously supplied the zymosan.

Miss W. Barr and Miss M. Baker gave fine technical assistance throughout all the experiments. Miss M. Lambourne assisted in preparing the manuscript and graphs.