That out breaks of plague are associated with a more or less simultaneous mortality amongst rats is an observation of great antiquity. After the discovery of the bacillus by Yersin and Kitasato this concomitant disease of rats was identified as plague, and during the last twelve years the association of rat plague and human plague has been shown to be almost invariable. The relationship has been so striking that many observers who have studied the question on the spot, e.g. Yersin, Ogata, Simond, Thompson, Koch, Gaffky and many others, have arrived at the opinion that from an epidemiological point of view one must regard plague as essentially a rat-disease in which human beings may participate.

In the following experiments, which are a repetition of those of Gauthier and Raybaud (1902, 1903), it is shown that, in the presence of the common Indian rat flea (Pulex cheopis, Rothschild), plague may spread from a plague-infected rat to a healthy rat confined in close proximity, but in such a way as to prevent contact with the body or excreta of the sick rat.

The following conclusions appear to us to be justified as a result of the experiments cited above:—

(1) Close contact of plague-infected animals with healthy animals, if fleas are excluded, does not give rise to an epizootic among the latter. As the go-downs were never cleaned out, close contact includes contact with faeces and urine of infected animals, and contact with, and eating of food contaminated with, faeces and urine of infected animals, as well as with pus from open plague ulcers. This conclusion is justified from Experiments I and II and especially from Experiment VI, in which close contact with infected animals was maintained for nearly a month. In Experiment I the animals were inoculated cutaneously, and in many cases had developed ulcers at the site of inoculation.

(2) Close contact of young, even when suckled by plague-infected mothers, did not give the disease to the former.

(3) If fleas are present, then the epizootic, once started, spreads from animal to animal, the rate of progress being in direct proportion to the number of fleas present. Thus, in the experiments during the months of December, January and February, the epizootic was very rapid in those go-downs, namely, Nos. 1 and 2, Experiments IV and V, in which the flea population was abundant and was kept up by a natural supply from the roof: it was much slower in Experiment VII, go-down No. 5, in which the flea supply was kept up artificially; and, finally, it was slowest of all in go-down No. 6, Experiment VI, in which there was no definite natural supply of fleas, and from which the fleas were daily removed for a period of six days, after which removal only a comparatively small number could be caught.

(4) An epizootic of plague may start without direct contact of healthy animal and infected animal. Thus, in the case of Experiment V in go-down No. 1 the healthy guinea-pigs were not put in until the last inoculated guinea-pig had died and been removed.

(5) We have in Part II shown by direct experiment that the rat flea can convey plague from rat to rat. Further experiments (Experiments VI, VII and XI) of a similar nature with the fleas removed from infected go-downs are now recorded.

(6) Infection can take place without any contact with contaminated soil. Thus in go-downs Nos. 1 and 2 (Experiments VIII, IX), guinea-pigs placed in wire cages two inches above the ground developed plague; also the monkey in go-down No. 2 (Experiment XIII) was never in contact with the ground.

(7) Aerial infection is excluded. Thus, guinea-pigs suspended in a cage two feet above the ground did not contract the disease, while in the same go-down those animals allowed to run about and those placed two inches above the floor became infected. Further, the monkey surrounded by tangle-foot was exposed as much to aerial infection as the control animal which contracted the disease.

Animals protected from fleas by means of a sufficiently broad layer of “tangle-foot” and placed in plague-infected houses do not contract plague, but the control animals, not so protected, on several occasions (24 per cent.) developed the disease. Out of 247 fleas caught on the “tangle-foot,” 60 p.c. were human, 34 p.c. were rat and 6 p.c. were cat fleas. Plague-like bacilli were demonstrated in the stomach contents of one out of 85 human fleas dissected, and of 23 out of 77 rat fleas.

Ceratophyllus fasciatus1 is the flea usually found on Mus decumanus in Great Britain, and this is also the case, apparently, throughout Northern and Central Europe. The common house mouse, Mus musculus, occasionally harbours this parasite also, though its usual flea is Ctenopsylla musculi.

The key to the anatomy of the mouthparts of the flea and to the understanding of the mechanism of blood sucking is the recognition of the pricking epipharynx and of the hypopharynx with its salivary pumping apparatus. In general it may be stated that the epipharynx makes a way through the skin for the mandibles, that the mandibles enlarge and lacerate the hole thus bored and convey into it the salivary secretion which is pumped, by means of the salivary pump, down the salivary canal contained in them, and that the aspiratory pharynx aspirates blood from the now congested wound along the blood channel formed by the approximation of the epipharynx and the two mandibles. The whole complex of parts is adapted to form two channels, an effluent along which saliva is pumped down and an affluent along which blood is pumped up.

1. Twenty-six passages from rat to rat have been effected without recourse to artificial media. The time taken was 89 days.

2. Our results furnish no evidence that the virulence of the bacillus used for the passages undergoes any alteration.

3. The experiment gives evidence of the varying susceptibility to plague of the Bombay rat.

In the preceding paper we have put forward the result of some experiments in which B. pestis was passed through a series of 26 rats by means of the subcutaneous method and without the intervention of any culture on artificial media. From this series of experiments it appeared that the virulence of the organism was not affected by passage through the ordinary rats of Bombay.

We have previously referred to the fact that a certain number of Bombay rats were immune both to subcutaneous and to cutaneous inoculations of virulent plague bacilli. In paper V, p. 505 above, we were able to record that 59% of rats in Bombay could withstand the rubbing on scarified surfaces of emulsions of spleens of rats which had died of acute plague. The fact that a large number of rats in India, where plague had occurred, were more or less immune has been noted by Simond and Walton.

From these two series of experiments we can conclude that a cowdung floor, grossly contaminated with a virulent culture of B. pestis, can retain its infectivity for 48 hours but not longer, and that a chunam floor, treated in the same way, does not remain infective even for 24 hours.

The blood of plague-infected rats may contain an enormous number of plague bacilli, even as many as 100,000,000 per c.c. having been found before death. On the other hand rats occasionally die from plague with little or no septicaemia. An insect sucking the blood of most rats shortly before death would imbibe a large number of bacilli.

While the blood of a rat may have as many as 100,000,000 organisms in a c.c. the urine may have none at all, or at least less than 10 per c.c. Plague bacilli were discovered in the urine in 29 per cent. of the cases. When the urine does contain plague bacilli they are always present in much fewer numbers than in the blood.

The faeces of rats dead of plague and the blood of which contained abundant bacteria are not highly infective and would appear to play little part in the spreading of the epizootic.

An investigation of the blood of 28 patients suffering from plague has been made, involving the examination of 74 specimens. The B. pestis was not found in the blood of five patients whose illness ended in recovery; nor was it found in seven of the fatal cases. The salient facts ascertained from an analysis of the remaining 16 fatal septicaemic cases may be recapitulated thus:

(1) Microscopical examination of the blood cannot be regarded as a trustworthy criterion of the degree septicaemia.

(2) A severe septicaemia may be present at a comparatively early stage of the disease and for a considerable number of hours before death.

(3) The septicaemia may be of an irregular or fluctuating type.

(1) It is a matter of considerable interest that, with the exception of rat No. 1832, the rats were caught alive during a period when no rats with acute plague were found and when no human cases were occurring in the locality.

(2) It cannot be regarded merely as a coincidence that the lesions were in every instance situated within the abdomen. With our present knowledge it would be rash to speculate as to the method of infection in these cases, but it may be conceded that the most reasonable interpretation of the lesions involves the assumption of an intestinal infection.

(3) Examination after the animals were killed furnished no indication that their health suffered in any way; the organs were apparently normal and there was no emaciation.

(4) Although the number of cases is admittedly small and although it is difficult to estimate virulence with any degree of accuracy, yet it is noteworthy that of the animals inoculated directly from, and with subcultures of, these abscesses only one survived.