The final choice of a suitable dilution of a given toxin will be made, partly from the difficult comparison with other known toxins, but mainly from the results of clinical observations in patients in the early stage of scarlet fever and in convalescents.

It is difficult to titrate scarlet fever toxin with accuracy. It will probably be found that by careful choice of subject and by several readings of the reactions at different intervals after the injection, one can differentiate between toxins of values x and 2x.

Recently conflicting statements have been published in various countries on the prevalence and increased mortality from cancer, i.e. malignant disease in general. Statistics show a definite upward trend in the annual death rate from cancer, but opinions differ as to whether the increased mortality represents the actual increase in the incidence of the disease or is explained by such factors as the increased average age at death, improved methods of diagnosis and more careful records of the cause of death.

A method has been devised whereby the potency of any given carcinogenic agent can be compared with that of a standard agent. This is called the standard carcinogenic potency of the agent (P.). The relative potency of the agent (R.P.), calculated from P., permits one to estimate the effect on the agent of different chemical treatments, dilutions, etc., the percentage increase or decrease in activity of the agent being at once shown. A unit of carcino-genicity (U.C.) has been established, and a concentration of 500 carcinogenic units per cubic centimetre is considered to induce the maximum relative response on animals when the agent is applied twice a week. These standards have been devised as a natural sequence to the accumulation of a vast amount of experimental data the correlation of which necessitated some expedient more rapid and covering a wider field than the graphs formerly utilized.

This paper is written from the practical point of view of a laboratory worker who is daily in direct contact with the cases submitted to him for diagnosis. Its purpose is to prove the inefficiency of the Widal reaction as a diagnostic test, and to urge its replacement by a more satisfactory method. Such a method is to be found in the application of the discoveries and theories on agglutination of Weil and Felix (1920), Felix (1924 and 1929), and Felix and Olitzki (1928 and 1929).

1. During the summer months 1928–9 an epizootic occurred in De Aar and neighbouring districts among small veld rodents, mainly affecting Namaqua gerbilles (Desmodillus auricularis).

2. Many of the animals dying showed a haemorrhagic septicaemia, the causative organism of which appeared to be a Pasteurella rather smaller in size than Pasteurella (Bacillus) pestis.

3. In a considerable proportion of the animals found dead there was no obvious septicaemia.

4. A description is given of the Pasteurella, which is believed to be a hitherto undescribed species, and for which the name Pasteurella desmodilli is suggested.

5. This organism was found to be highly pathogenic for rabbits, less so for guinea-pigs and mice, and non-pathogenic for rats and fowls.

6. It was found by Rhodes to be highly pathogenic for Lobengula gerbilles (Tatera lobengulae) from Cape Town, but for animals of this species from Johannesburg it proved to be much less virulent, and even when it killed them there was frequently no septicaemia.

7. The above observation seems to indicate the presence of a considerable degree of immunity among the Johannesburg lobengulas, as compared with those from Cape Town.

8. It is suggested that a similar condition of partial immunity among the Namaqua gerbilles (for which there is other evidence adduced) might account for the finding of many of them dead without obvious septicaemia.

9. The existence of this disease introduces a complication in the diagnosis of plague among veld rodents.

The culture representing the new type to be described was obtained by one of us (S. H. W.) under rather unusual circumstances. Three fatalities occurred at intervals of 4 and 6 months among the children of one family living in Ashington, Northumberland, a coal-mining town of 30,000 population; in each case death was due to acute gastro-enteritis and supervened within 24 hours of their falling ill. On the death of the third child a bacteriological examination was made post-mortem and S. aertrycke found in abundance in the organs. The suggestion was made at the inquest that a carrier must exist in the stricken family and specimens of faeces from the surviving members were examined by S. H. W. From the mother's specimen a Salmonella was isolated agglutinating with the sera not only of S. paratyphi B, but also of S. enteritidis (Gaertner). This Salmonella, which we propose to call S. newcastle, was obviously different from the strain of S. aertrycke obtained from the fatal case, and the suggestion that the mother had been a carrier infecting her family should almost certainly be dropped; she had had no suspicious illness at any time. The probability is rather that conditions in the house and habits of the inmates were responsible for the succession of cases, though no evidence could be obtained as to possible infection of food by rodents, or consumption of food likely to be specifically infected; the sanitation was of poor quality (privy midden).

1. The addition to the Nairobi prison diet of either a mineral supplement (closely reproducing in composition the ash of cows' milk) or of a small dose of cod-liver oil, or a combination of these two, produced in adult African prisoners a distinct improvement in body weight.

2. Similarly an addition of olive oil and iodine produced a like effect but less in degree. The significance of this will require further investigation.

3. Evidence is adduced which suggests that one of the most important constituents of this soluble mineral mixture is calcium. Thus:

(a) The prison diet would appear on the basis of previous work to be deficient in the absolute amount of calcium present, nor does it seem rich in vitamins A and D. (In this connection it seems apposite to mention that cases of “night blindness” have frequently been reported.)

(b) The blood calcium of the 42 experimental subjects was initially at a low normal level.

(c) In the groups receiving the mineral mixture and/or cod-liver oil (which is a recognised calcium absorption promoting agent) the level of the blood calcium was invariably raised.

(d) As indicated above, the body weights in these groups increased. There appears to be, further, a certain amount of correlation between the increase in body weight and that of the blood calcium level.

(e) Balance experiments show that given favourable conditions (i.e. the addition of the mineral mixture alone or along with cod-liver oil) calcium will be absorbed in amount considerably in excess of that absorbed from the basal diet alone.

4. In the course of the work some evidence has emerged that certain other factors, e.g. intestinal parasites and minimum temperature, may respectively influence the blood calcium level and the body weight increase. Should these prove in future work to be real influences they would not vitiate our present conclusions.

1. Of five experimental subjects—growing East African native boys—four proved to have definitely negative calcium balances on the diet in use at the Nairobi Prison, while the fifth was merely in calcium equilibrium.

2. The diet in question was found to be so low in calcium (0·32 grm. Ca per day) that the occurrence of negative balances is not altogether surprising.

3. These negative calcium balances became positive on the addition of calcium as milk or in a mineral mixture with or without cod-liver oil.

4. Cod-liver oil alone was not effective in bringing about any appreciable reduction of a negative balance.

5. Our findings showed that 0·43 grm. Ca daily was the maintenance requirement for the subjects under experiment. This figure is practically the same as that found by Sherman in a summary of work on white races.

6. On giving additional calcium in a mineral mixture it was found that after the maintenance requirement had been satisfied, as much as 50 per cent. of any excess might be retained. The calcium of milk, on the other hand, was found to be utilised more efficiently; the corresponding retention figure being over 70 per cent.

7. The influence of the various supplements on growth is briefly discussed.

In the two preceding communications we have discussed the calcium balances and weight increase of natives, both adult and growing, on certain institution diets, with and without additions.

Various considerations pointed to a relative or absolute deficiency of calcium in the above diets. The experimental additions made, therefore, were (1) a mineral mixture reproducing closely the ash of cows' milk; (2) cod-liver oil; (3) cod-liver oil plus the mineral mixture; (4) olive oil plus KI; (5) milk.

Pioneer studies seemed to resolve the organisms of the Salmonella group into four types or groups sharply marked off from one another by their agglutinative reactions: S. typhosus, S. paratyphosus A, S. enteritidis Gärtner and a “paratyphoid B group.” Later work has shown the heterogeneity of the old paratyphoid B group, which approximates to what we may now term the series of diphasic types; and further that S. typhosus and S. paratyphosus A, far from standing in serological isolation, display striking and peculiar relationships to certain members of this series: the flagellar antigen of the typhoid bacillus closely resembles that of the specific phase of the diphasic Stanley type, less closely that of the Bombay type; both in flagellar and somatic antigen S. paratyphosus A coincides with the specific phase of the diphasic Sendai type (typical paratyphus A bacillus of Aoki and Sakai, 1926).

In 1922 Iwaschenzoff described from Russia an epidemic of relapsing fever attended by high mortality. From many of the sufferers organisms of the Salmonella group were isolated. These organisms, termed S. paratyphus N1, were apparently responsible for the severity of the cases. In 1921–2 the mortality rate in uncomplicated cases of relapsing fever was 5·5 per cent.; whereas it was 50 per cent, in cases showing secondary infection by Paratyphus N bacilli.