Strains of Staphylococcus pyogenes able to grow in 0·1 unit/ml. of penicillin produced penicillinase and were regarded as resistant. The majority of the remaining strains had the same sensitivity as the standard sensitive staphylococcus.

Five hundred and three strains were isolated from pre-clinical medical students from 1950 to 1953, and 24% were penicillin-resistant. Two per cent were streptomycin-resistant and none were resistant to chloromycetin, aureomycin or terra-mycin. Two hundred and forty-six of the strains were isolated from persistent and intermittent carriers, but only 14% were penicillin-resistant.

Another 183 strains were isolated from persons unattached to hospital, and 14% were resistant to penicillin. Only one strain was streptomycin-resistant; one was chloromycetin-resistant and none were resistant to the other antibiotics.

Among persistent and intermittent carriers, more than 7 times as many penicillin-resistant strains were isolated from those who had had previous penicillin therapy as from those who had not.

It was concluded that the administration of antibiotic increases the number of resistant strains in the community.

Suppose that λ is the average density of virus particles per unit volume. If x is a dilution of this and unit volume is applied to an egg (or plate in other problems) the probability that the egg remains sterile is

provided that if a particle is present, it will infect the egg. To make a dilution assay we choose dilutions x1, …, xm (m levels) and apply these to n1, …, nm eggs. If these result in r1, …, rm sterile eggs we can estimate λ by maximum likelihood. The theory has been given by Barkworth & Irwin (1938), and full references to work on this problem will be found in Finney (1952). If we plot the quantities r1/n1, …, rm/nm against x1, …, xm we get a set of point whose fit to the curve (1) can be tested by a χ2 test. In a number of situations, however, it is found that (1) does not give a good fit. The estimation of λ is then completely invalid. In the present paper we consider why this happens, what types of curve may be fitted to the data and what they imply, and we also give a simple rapid test for such data fitting an exponential curve.

An account is given of experiments designed to study the spread of myxomatosis in populations of rabbits living under natural conditions on a number of sites in the Eastern Riverine Plain of south-eastern Australia. In five different trials the disease spread from inoculated rabbits, but failed to gain momentum and died down within a few weeks of its introduction.

In December 1950, when the disease was persisting at low incidence on one test site and seemed to have died out in the others, an epizootic broke out in the neighbourhood which spread, in a few months, over the greater part of south-eastern Australia. The only factor, apart from climatic ones, which could account for this sudden change in the activity of the disease was the development of large populations of two rabbit-feeding mosquitoes, Anopheles annulipes and Culex annulirostris. A close correlation was demonstrated, on the flats bordering the Murray River, between the distribution of these insects and myxomatosis activity.

The author is indebted to Mr F. N. Ratcliffe, officer-in-charge, Mr J. le G. Brereton, Mr J. Calaby, and Dr R. Mykytowycz, all of the Wildlife Survey Section, Commonwealth Scientific and Industrial Research Organization, for assistance given during the course of the work and to Mr Ratcliffe, and Prof. F. Fenner of the Australian National University, for guidance in the presentation of the results.

Field observations on myxomatosis in wild rabbit populations of the Riverine Plain of south-eastern Australia, extending between August 1951 and March 1953, were reported. General observations on vector abundance and disease incidence were made over a large area. Detailed analysis of the population counts, age structure and immune status of the rabbit population, and the occurrence of insect vectors, were made at one site, Lake Urana.

The case-mortality rate in the first epizootic was in the neighbourhood of 99·8%, whereas in the second epizootic it was about 90 %, and many more chronically sick rabbits were observed in the later outbreaks. The virulence of eight strains of virus recovered from mosquitoes at Lake Urana in December 1952 and four strains recovered at Corowa at the same time was found to be somewhat reduced when they were compared with the original standard laboratory strain of virus. It is thought that this attenuation is the major reason for the observed change in the case-mortality rates.

The mosquitoes, Anopheles annulipes and Culex annulirostris were by far the most important insect vectors of myxomatosis in this region.

Thanks are due to Messrs F. N. Ratcliffe, B. V. Fennessy, and H. J. Frith of the Wildlife Survey Section of C.S.I.R.O. for helpful discussion and assistance during the field work and in the preparation of the paper. The technical assistance in the field of Messrs C. S. Hale and K. L. S. Harley is gratefully acknowledged, as is the generous help afforded by the landholders of Lake Urana, especially Mr J. Muldoon, and by Mr W. Quirk, rabbit inspector.

1. A quantitative comparison has been made of the observed mortalities from respiratory tuberculosis with those expected from a simple scheme of generation mortality.

2. This scheme gives a good account of the observed rates until about 1930, but is unsatisfactory thereafter.

3. It is concluded that these discrepancies cannot be due to chance, and that the generation effect has become obscured by other factors.

4. In particular the shift of maximum mortality into later life is not directly explained by the hypothesis, nor is such a shift a necessary consequence of the generation hypothesis of mortality.

I am greatly indebted to Dr J. C. McDonald for advice and constructive criticism during the writing of this paper; my thanks are also due to Dr G. S. Wilson for his comments.

1. The production of weaned young from pairs or groups of three, four and five mice, each containing one male, has been compared.

2. For long-term large-scale production, small breeding groups of one male with two females are proposed.

I should like to thank Dr A. S. Parkes, F.R.S., for much helpful criticism, and Dr Walter Perry for the statistical tests.

Staphylococcal α- and β-toxins inhibited the spontaneous activity of isolated rabbit intestine, the former leading to an increase in tone and the latter to a decrease. The γ-toxin and δ-haemolysin had no effect in concentrations likely to be met in extracts of cultures, nor did δ-lysin enhance the action of β-toxin. Culture extracts in which the haemolytic toxins had been neutralized by antiserum sometimes produced an increase in tone or amplitude or both, which was moderate in degree but rapid in onset. This stimulation was more common with strains isolated from food-poisoning outbreaks than from infective lesions. It was occasionally observed with coagulase-negative strains and was therefore considered to be unrelated to enterotoxin. No effect on intestine was observed which could be considered specific for food-poisoning strains.

We are indebted to Dr R. E. O. Williams of the Central Public Health Laboratory, Colindale, for providing food-poisoning strains of staphylococci, to Dr H. W. Smith for strain CN, and the Wellcome Research Laboratories for gifts of sera.

Factors affecting the formation and destruction of sulphide in the Thames Estuary have been discussed. Measurements of oxidation-reduction potentials show that the effect of oxygen in the water on conditions in deposits of mud is limited to the surface layer. The rate of formation of sulphide increases with temperature, doubling approximately for each 10° C. rise. Reduction of sulphate to sulphide will occur at temperatures as low as 5° C., but even at 25° C. is inhibited by traces of dissolved oxygen. Mud in suspension in the estuary can produce as much sulphide as compact layers of similar mud which might be more anaerobic.

Oxidation of sulphide in the Thames Estuary is shown to be of a purely chemical nature; the rate of oxidation, when oxygen is present, is increased by the presence of suspended matter and iron. In the Thames Estuary, however, the rate is limited by the rate at which oxygen enters the water.

Much of the experimental work described in this paper was carried out by Miss A. Laird, Mr A. E. Sutton and Mr R. B. Peel of this Laboratory. The paper is published by permission of the Department of Scientific and Industrial Research.

Salmonella typhi or S. paratyphi B were found in very large numbers in the faeces of enteric carriers. Of twenty-four carriers, four were found negative by the fullest examination. Of the twenty found positive almost all harboured many millions of bacilli per gram of faeces.

A minute inoculum of one drop (1/50 ml.) of a 1:1000 dilution of faeces on a culture plate only rarely failed to reveal all the positives without the use of an ‘enrichment’ medium and the result of such a procedure was a culture plate with virtually a pure culture of the pathogen.

I am grateful to the Medical Superintendents of the Mental Hospitals at Cardiff, Bridgend and Denbigh, and the Medical Officer of Health for Brecon for submitting specimens for examination.

1. When toxic filtrates derived from coagulase-positive staphylococci including twelve food-poisoning strains, thirty pyogenic strains of human and twelve of animal origin were applied to isolated rabbit gut in vitro, two characteristic responses were obtained, associated with the α- and β-lytic titres respectively, and abolished by the corresponding antisera. They were independent of the origin of the strains.

2. A non-specific response is also described which is thought to be due to the peptone in the medium.

3. No effect that could be attributed to enterotoxin, such as was reported by Anderson (1953), was observed and it is concluded that isolated rabbit gut is unlikely to prove a useful indicator of its presence.

We are indebted to Drs Mary Barber, J. M. Talbot, F. C. O. Valentine and H. Williams Smith for strains; to Dr R. E. O. Williams for strains and bacteriophage typing; to Prof. C. E. Dolman and Messrs Burroughs Wellcome for strains and antisera; to Mr F. Crisp of the Division of Experimental Biology, National Institute for Medical Research, Medical Research Council, for rabbit gut; and Dr H. O. Schild for continued interest and practical help.

In earlier reports on the physique of Oxford undergraduates attending the pilot student health service from 1947 to 1950 emphasis was laid on body shape. Parnell (1952) pointed out the tendency for centrally placed Sheldonian somato-types, that is those with more balanced mid-range body proportions, to occur more frequently in the Oxford sample than in American universities, a tendency incidentally which advances further according to the level of university performance attained. Tanner (1952) emphasized the tendency for Oxford students to show less mesomorphy, that is muscle and bone development, than American students but he reports a comment from Sheldon himself that the distribution at Harvard resembled the Oxford distribution more closely. This approach, by the analysis of body shape, to the problem of physical characteristics accompanying academic selection is of particular interest, but in calculating the somatotype, height is deliberately ignored by being placed as the denominator of all the body proportions employed. Height, however, is an index of body size, as opposed to body shape, and has an importance of its own. This report concerns observations on Oxford undergraduates in which more attention is paid to body size, and particular attention also to variation of weight during university residence.

A number of workers have shown that antibodies to the influenza A viruses which are demonstrable in the sera of normal adults undergo periodic fluctuations in titre related to epidemic occurrence of the disease. The rise in antibodies accompanying infection and demonstrable by neutralization, agglutination-inhibition and complement-fixation tests is used as a method of diagnosis which is admitted to be of greater sensitivity than actual recovery of virus from the throat. Studies of the population during an epidemic have also revealed the existence of subclinical infection with rise in antibodies comparable to that occurring in those suffering clinical illnesses. But surveys of sera from large samples of the population show no general upward shift in antibody levels unless an actual epidemic occurs (Martin, 1940). The studies made by Francis, Magill, Rickard & Beck (1937), Hoyle & Fairbrother (1937), Rickard, Lennette & Horsfall (1940) and Martin (1940) also indicate the relative impermanence of the enhanced antibody levels consequent upon an epidemic. A relatively rapid decrease in antibody is shown both by the neutralization and complement-fixation tests during the 3–6 months after an epidemic and then a slower fall occurs until the next epidemic again causes a rise in titres. There have, however, been relatively few studies on large samples of populations situated in different geographical areas before and during an outbreak.

Bacteriological surveys of beach pollution were made at a seaside resort before and after the installation of a new sewer outfall.

The presumptive coliform count on samples of sea water was found to be a satisfactory index of contamination, and showed good agreement with sanitary survey and with the results of float-test observations. The faecal coli count was in general constantly related to the results of the presumptive test, but there was some indication that coliform bacilli of the I.A.C. groups persisted or multiplied in the sand around the outfall.

The degree of contamination of the sea water at a given sampling point was shown to depend not only on its distance from the outfall but on various meteorological and tidal factors.

For bacteriological surveys of beach pollution, it is suggested that a large series of tests by a relatively simple technique is preferable to a more limited investigation by more complex methods.

The possibility of laying down bacteriological standards for bathing beaches is discussed. An upper permitted limit of the order of 1000 coliform organisms per 100 ml. would discriminate between areas on the beach investigated which were unsatisfactory by sanitary survey before and after the installation of the new sewerage scheme, and those parts of the beach which were apparently not subject to serious pollution with sewage.

I am greatly indebted to Lt.-Col. J. M. Linton Bogle, consulting engineer responsible for the sewerage scheme described, for his interest in and practical help with this investigation; to members of the staff of the local authority concerned for help with sampling, and to Dr A. F. Ollerenshaw and Mr A. Partridge, who were in local charge of the temporary laboratory where the presumptive tests were carried out, in 1950 and 1948 respectively.

1. Straight-chain α-amino-acids are all equally effective in causing the lysis of Bact. coli.

2. Various modifications of this structure always result in a decrease of this lytic activity.

3. The optical isomers of α-amino-acids are equally effective in causing lysis.

4. There is a correlation between the lytic and the bactericidal powers of the amino-acids and their derivatives.

1. Non-avid tetanus antitoxin appears to have a higher titre in relation to an avid standard if titrations are made against test toxins with a high toxoid content. The least avid serum tested appeared to be 2·87 times as potent when titrations were made against a toxin grown for 35 days than when a 2-day growth was used. Using the same test doses avid sera did not show this discrepancy.

2. A mixture of short-growth toxin and formol toxoid reacted similarly to the long-growth toxin.

3. Some non-avid tetanus antitoxic sera showed evidence of difference in goodness of fit with two test toxins; one serum was 2·49 times the titre of another when tested against one toxin and only 1·37 times when tested against another toxin.

4. Diphtheria toxoid has less affinity than toxin for antitoxin; this difference is greater with formol toxoids of poor immunizing efficiency; the difference is further accentuated if non-avid anti-toxin is used.

5. The Danysz effect may be very small if non-avid antitoxin is used.

6. In the presence of excess antitoxin, toxin or formol toxoid can combine with 3, 4 or more times its equivalent as judged by in vivo titrations for free antitoxin.

7. The serum ratio, i.e. in vivo titre/in vitro titre of high or low ratio antitoxin in multiple combinations is nearer unity than that of the original antitoxin.

1. The production of soluble antigen in the chorioallantoic membranes at 6 hr. can be used to measure the infectivity of influenza virus preparations.

2. Intracellular virus is much more infective when tested by soluble antigen production at 6 hr. than when tested by conventional methods.

3. It is suggested that intracellular virus differs from extracellular because more virus particles per cell are required to initiate infection.

The author is greatly indebted to Dr L. Hoyle for his advice and encouragement throughout this work. Mr G. Field gave valuable help with the mathematics.

Desert rodents often have been used successfully as laboratory animals in parasitology (Abdel Azim & Cowper, 1950; Adler & Feldman-Muhsam, 1952; Delanoe, 1929; Kuntz & Malakatis, 1954; Nicolle, Anderson & Colas-Belcour, 1927; Sergent & Poncet, 1951 a, b) and virology (Durand & Laigret, 1932; Gear & Davis, 1942; Noury, 1948; Snyder, Zarafonetis & Liu, 1945; Van Rooyen & Danskin, 1944; Zarafonetis, 1945). However, with the exception of investigations into the role of some of these animals as mammalian reservoirs of plague (Ristorcelli, 1938; Tikhomirova, 1934; Wassilieff, 1933), no reports have been found of their use for experimental bacterial infections.

In certain types of titrations the experimental results are obtained as percentages, say of mortality, at different dilutions in geometric progression of the test material. The result of a titration is then expressed as the 50% end-point which is estimated by interpolation as follows.

A technique is presented for the titration of the virus of foot-and-mouth disease in culture in surviving epithelial tissue from the tongues of cattle. The cultures are incubated in cups on Perspex plates, and the detection of virus multiplication is by a complement-fixation test made on the culture in each cup.

On the basis of comparative titrations in culture and in cattle, the method has been found to be as sensitive for the detection of virus as the titration by intradermal inoculation of the tongue of cattle. The method can also be applied in the detection of antibody in neutralization tests.

We wish to record our thanks to Messrs E. Scoates and P. Mitchell for their technical assistance, and to Messrs H. M. Smith, R. H. Compton and R. L. Jackson for their part in the design and fabrication of various bottle rotators, plate shakers, Perspex lids and the inoculating box.

1. The growth in synthetic cream of certain pathogenic bacteria may be inhibited by two methods:

(a) By limiting the ingredients to cooking fat or other fatty material, excluding margarine or butter, emulsifying agent such as methyl cellulose, sugar and salt. The cream must not then be left in contact with confectionery from which nutritive material may be absorbed. This last condition renders the method impracticable as a means of control.

(b) By the addition to synthetic cream, including those containing milk and egg, of a bacteriostatic agent such as hydrogen peroxide.

2. Contaminants may grow in emulsified fats in the absence of protein. These are probably lipolytic organisms which are able to utilize the products split from the fat. Organisms such as coagulase-positive staphylococci, Salmonella paratyphi B, and Bacterium coli, tend to die out within 24 hr. or after a few days in creams lacking protein.

3. The concentration of H2O2 necessary to inhibit the growth of test organisms is dependent on the ingredients of the cream. In the presence of butter, milk and egg yolk at least three times the concentration of H2O2 is necessary as when emulsified fat without added protein is used. The suggestion is made that concentrations of H2O2 from 0·005 to 0·02% should be added to all commercially produced synthetic cream, the actual concentration used depending on the constituents of the cream.

4. To prevent the growth of contaminating organisms in commercially produced synthetic cream, H2O2 should be added immediately after pasteurization and cooling when the bacterial count is still low. Even comparatively high concentrations of H2O2 will not inhibit or control the growth of bacteria already present in large numbers.

5. The inhibiting power of 0·012% of H2O2 remained effective in a fat-emulsion cream and also in a laboratory cream containing margarine, egg and milk for 3 weeks at + 4 and 20° C.; at higher temperatures of storage the stability of the oxidizing agent is doubtful.

6. Since synthetic cream is exposed to so many different sources of contamination, particularly in the bakery, and since there is no method of controlling the growth of pathogenic organisms that have gained access to it other than by the addition of a bacteriostatic agent, it is strongly recommended that the present regulations on preservatives should be altered to allow the addition of H2O2 in suitable concentration to synthetic cream.

We are grateful to all those who have helped in the supplies of material, and in particular we wish to thank the Public Health Department, St Pancras Town Hall; Mr E. Capstick, United Dairies Ltd.; Mr J. Valentine Backes, President, Bakery Allied Traders' Association; and also Messrs Bakcos Catering Supplies Ltd., A. Bellamy and Co. Ltd., Farma Cream Products Ltd., Krema Ltd., Malga Products Ltd., Ramsay, Braddon and Co. Ltd., Quality Foods Ltd., Sunnyside Products Ltd., Unicream Ltd., and Vitacream Ltd.

We are indebted to Mr L. Bailey for his technical assistance.