1. Plasma heated to 56° C. for 4 min. is not clotted by staphylococcal coagulase or thrombin, but when heated in the presence of glycine remains sensitive to the action of these clotting agents.

2. With glycine, plasma no longer clots after heating at 62° C.

A passive protection test with brucella antibodies, based on the spleen counts technique, has been described. This test has been found more sensitive than that based on lethal challenge.

Normal human sera have been shown to possess marked protecting power.

Protective activity of human sera was considerably increased following vaccination with living streptomycin-dependent brucellae, or after natural disease.

The bearing of these findings on several brucellosis problems has been briefly discussed.

The author wishes to express his thanks to Prof. A. L. Olitzki for constant advice and encouragement; to Dr H. V. Muhsam, Department of Statistics, the Hebrew University, for generous help in the statistical analysis; and to Miss R. Avraam for technical assistance.

Dilution series of viruses often do not fit the exponential curve which they should fit if the usual assumptions of the theory of dilution series were true. Suppose the density of the virus suspension is λ, the average number of virus particles per unit volume of the inoculum (we suppose that unit volume is the amount actually used in inoculating the egg or other test material). Let C be the proportion of infective virus so that Cλ is the effective density of virus particles, and X is a factor depending on the dilution used such that CλX is the expected number of infective particles in the inoculum.

1. The diphtheria antitoxin titres of several thousand samples of cord blood have been analysed with relation to the histories of the mothers.

2. Among mothers without histories of diphtheria or of immunization those born in 1933 or later showed a higher degree of natural immunity.

3. The distribution of antitoxin titres among mothers who had suffered from diphtheria was similar to the average distribution among the unimmunized.

4. The differences in proportions with any antitoxin in different groups of unimmunized mothers were almost entirely made up of those with titres above 0·04 and below 1·0 unit per ml.

5. Among the artificially immunized mothers there were many more with low titres, i.e. below 0·04, and a few more with high titres, i.e. over 1·0, than among the unimmunized.

6. Mothers immunized between the ages of 9 and 14 with the peak at 11 or 12 years showed the highest immunity rate.

7. The greatest number of high titres occurred among recovered cases subsequently immunized.

Groups of guinea-pigs were passively immunized against diphtheria toxin with homologous antitoxic serum so that their sera contained, at the start of the experiment, 1·0, 0·1, 0·01 or 0·001 unit/ml, respectively. They were then actively immunized with one, or two spaced, injections of 0·1 Lf of A.D.F. Two control groups were included, one passively immunized only and the other actively immunized only.

Passively produced serum titres of 0·01 and 0·001 unit/ml. did not interfere with active immunization in any demonstrable way.

A titre of 0·1 unit/ml. did interfere with active immunization, markedly 4 weeks after the primary, slightly 2 weeks after the secondary, and markedly 14 weeks after the secondary, stimulus.

A titre of 1·0 unit/ml. interfered with active immunization, markedly 4 weeks after the primary, and 2 and 14 weeks after the secondary, stimulus. This titre, however, did not completely annul the effect of the primary stimulus. The highest observed serum titre was obtained at the 32nd, instead of at the 4th week, as in the guinea-pigs actively immunized only.

In large measure the results confirm those of Barr and her colleagues who found that, in human babies, an initial ‘passive’ titre of 0·04 unit/ml. serum did not interfere with active immunization, whereas a titre of 0·1 unit/ml, led to unsatisfactory immunization.

C. botulinum types A and B toxins required larger amounts of antitoxin for neutralization when fractions of the test dose of toxin were used than would be expected.

The degree of this dilution effect, i.e. the dilution ratio, varied with different sera. The dilution ratios with type A sera could be improved by allowing a greater time for combination to occur, or by mixing toxin and antitoxin in high concentration and diluting these mixtures.

The dilution ratios of type A sera from horses rose as immunization proceeded, the greatest rise occurring during the 3-month rest period following three small doses of toxoid.

A type A serum could be separated into fractions of different dilution ratio by precipitation with increasing concentrations of ammonium sulphate.

Type A antitoxins prepared by immunization of groups of guinea-pigs with two doses of toxoid, had dilution ratios far lower than those found in hyperimmune horse sera.

The dilution ratios of sera were lower when tested against a toxic toxin than against a more toxoided one.

A number of type A sera were tested at one level of test against a number of test toxins. There was no evidence of heterogeneity of the toxins and antitoxins tested.

The serum ratios were found to be correlated with the dilution ratios except where the serum ratios exceeded 0·64. Flocculation time was slow and varied from 6 to 48 hr. with different sera tested against the same toxoid.

I am greatly indebted to Miss M. Barr for her constant advice and encouragement in the execution of this work and in the preparation of this paper. I am grateful to Dr M. Sterne for the supply of C. botulinum type A and B toxins and for the benefit of his experience of animal botulism.

Mice and rats which are normally resistant to histamine become more susceptible to its lethal action after an injection of H. pertussis. This so-called sensitization to histamine is not an anaphylactic phenomenon. It is due to the action of a component of H. pertussis, the histamine-sensitizing factor (HSF), which in some unknown way overcomes the physiological mechanism in rats and mice that makes them more resistant than other species to histamine. Guinea-pigs which appear not to possess this mechanism and are about 200-fold more susceptible, weight for weight, than rats and mice, were not made more sensitive by H. pertussis.

After a stated dose of vaccine, sensitization was detectable in 48 hr., reached a maximum in 3–4 days, remained at this level for about 2 weeks and gradually disappeared. The effects of dosage, route of injection, and weight and sex of mice have been examined.

The HSF was found in strains of H. pertussis; it was not found in H. parapertussis, H. bronchisepticus or H. influenzae. It was only slightly affected by heating at 70° C. for 1 hr. but was destroyed at 80° C. in ½ hr. It was destroyed when bacteria were disintegrated by shaking with glass beads, or by grinding after being freeze-dried. It was found in the supernatant fluid of a partially autolysed vaccine.

HSF was antigenic. Antisera were prepared in rabbits. Anti-HSF combined with HSF and neutralized its histamine-sensitizing activity. Bacteria treated with antiserum in vitro absorbed anti-HSF and did not thereafter sensitize mice.

Antiserum protected mice passively against the sensitizing action of vaccine, presumably by combining with HSF. After sensitization had developed the sensitive state was not affected by antiserum.

Although HSF is an antigen there was no indication that histamine-sensitization was due to its antigenicity.

The HSF was differentiated from heat-labile and heat-stable toxin, haemagglutinin, capsular material and agglutinogen.

The preparation V 17, which is a small fraction of the disintegrated bacteria, adsorbed on red cell stromata (Pillemer et al. 1954) had a high histamine-sensitizing value. Compared with whole bacterial vaccine it caused little production of agglutinin in mice; in rabbits it caused a slower and smaller production of agglutinin and a faster and greater production of anti-HSF. For this reason immune rabbit sera may have a high agglutinin titre and a low anti-HSF value or vice versa. Anti-HSF rabbit serum protected mice against sensitization by V 17.

Vaccines could be graded according to their histamine-sensitizing activity. This did not always correspond to their grading by agglutinin production in mice. The relation of HSF to the immunizing antigen of H. pertussis and the use of histamine-sensitization to indicate the immunizing potency of pertussis vaccines are discussed.

We wish to thank Glaxo Laboratories Ltd. for the supply of reference vaccine, and the Whooping Cough Immunization Committee of the Medical Research Council for vaccines and sera.

1. Strains of Proteus derived from animals produced septicaemia in mice when given in large doses intravenously, but smaller doses caused necrotic lesions in the kidneys. The organisms failed to produce kidney lesions when administered intra-peritoneally or subcutaneously but did cause septicaemia.

2. Killed suspensions of Proteus failed to induce kidney lesions.

3. Strain differences in pathogenicity were found, some strains producing naked-eye lesions in the kidney, others giving microscopic changes only in this organ; of those strains which caused no changes in the kidney tissue, some persisted for several days and others were eliminated rapidly.

4. The mechanism of this localization in the kidney is discussed.

The writer wishes to thank Dr A. Wilson Taylor and Prof. R. Lovell for their advice and helpful suggestions and Mr R. Hood for the photographs.

1. Gastro-enteritis caused by certain varieties of Bact. coli is a highly infectious disease among babies, and it has been suspected that the infectivity is associated, partly at least, with the very large numbers of pathogens excreted in the faeces. Very large numbers were found in the faeces, but large numbers of pathogens were also found in dysentery, salmonella food-poisoning and in some cases of paratyphoid fever.

In cases of gastro-enteritis the dilution of faeces containing 5 to 50 pathogenic Bact. coli per ml. was usually 10−8; in Sonne dysentery and salmonella food-poisoning the corresponding dilution was 10−6. It is doubtful if even this difference can account for the differences in infectivity.

2. When normal Bact. coli outnumber pathogenic Bact. coli, shigellae or salmonellae by as little as 10: 1 it is difficult to identify the pathogens unless a selective culture medium can be used.

3. Ordinary nutrient broth is generally able to act as an enrichment medium for the isolation of pathogens from faeces.

Previous workers (Stuart-Harris, Pownall, Scothorne & Franks, 1953; May, 1953)have reported studies on the bacterial flora of the sputum in cases of chronic bronchitis. Both pneumococci and Haemophilus influenzae have been reported frequently in the sputum but the source of these organisms, which might be either the patient's own upper respiratory tract or the environment, has not been traced.A long-term study of individuals with and without chronic bronchitis and of their families was therefore begun in an attempt to discern the role of exogenous infection in this disease. Observations were limited to the pneumococcus because serological identification of individual strains was readily possible in this species.

1. Production figures are given for a small mouse colony.

2. High productivity depends upon careful selection of breeding stock.

3. This selection has been based on an analysis of individual records. Both the keeping and the analysis of records are easy and simple.

4. In addition to providing a basis for selection, these records make possible a forecast of future production.

I particularly wish to thank Miss D. Lawrence, the animal technician whose skill in looking after this colony has contributed so much to its successful performance.

1. In bacteriological investigations of stretches of water an estimate of the number of a certain type of bacterium in a unit of volume of the water is required so that the pattern of the pollution of the area can be discerned. Usually it is known or guessed that the density of bacteria,δ say, lies in a range δL up to δH. Several samples of water are taken and each is tested for a positive reaction which would indicate the presence of at least one bacterium of the type being searched for. When it is assumed that bacteria are spread in a Poisson manner through the water in each sample an estimate of their density can be made from the number of positively reacting samples.If the range δL to δH is large it is unlikely that one volume of water for each sample will give a very efficient estimate of the density;e.g. if this volume is in an average position with respect to (δL to δH), say, 2/(δL +δH), and the true density is near an end of the range, all the reactions are likely to have the same sign and the sample gives no discrimination. Even worse estimation for some densities will occur if a very small or very large volume is used. The setting of reasonably sized confidence intervals for the density requires a series of volumes designed to locate the density with about equal accuracy no matter where it is in (δL to δH). Hence the usual set of volumes is the dilution series discussed by several writers, for example recently by Cochran (1950).