(1) These experiments prove that the fumigator is safer to use than the dumping fixture, because high concentrations of the poison gas are avoided, so that dangerous pockets are not developed.

(2) The poison gas is more evenly distributed over the different parts of the section of the ship under treatment.

(3) High concentrations, however, can be developed^ at will in any part which requires special attention.

(4) The maximum concentrations of the poison gas were obtained in every part of the section in a shorter time, when the fumigator was used, than when the dumping fixture was employed.

(5) The artificial ventilation caused by the fan, which is capable of blowing into the section 500 to 1000 cubic feet of fresh air per minute, according to the speed of the motor, materially assists the rapid clearing of the poison gas from the ship, after it has accomplished its work. It is thus possible to complete the fumigation and ventilation of a ship in 3 hours using the cyanide fumigator, while more than 4 hours are required for the dumping fixture.

(6) Even more important is the fact that smaller quantities of cyanide can be used with greater efficiency in the fumigator, than larger quantities, with less efficiency in the dumping fixture.

(7) The general conclusion is arrived at, that, from the point of view of efficiency, safety and economy, the fumigator is superior to the dumping fixture. Creel has shown that, in respect to the time required for completing the fumigation, and in respect to the thoroughness of the process, both in killing rats and insects, the dumping fixture is superior to the Clayton gas machine. It follows, therefore, that the cyanide fumigator is very much superior in these respects to the Clayton gas machine.

The chemical changes which occur in the dissolved solids of sewage on discharge into sea-water have been investigated by several workers (see references 1–11). Some observation on the difference of these changes to those occurring in polluted fresh water have been made by Letts and Richards (6,11). In regard to waters whose saline composition is not such as to preclude their admissibility for potable purposes on this account, Liversidge(12) has made a brief communication on the variation in the free and albuminoid ammonia.

In a recent series of papers, Flexner (1922), Lynch (1922), Amoss (1922a and b) and Webster (1922a and b) record the results obtained at the Rockefeller Institute in an experimental investigation of epidemics among laboratory animals. These experiments bear a close similarity to those which have been recorded in previous papers of the present series (Topley, 1919, 1921 a and b, 1922 a and b, and Topley, Weir and Wilson, 1921). Their publication affords an opportunity of comparing the results obtained in two independent enquiries, which have been carried out along slightly different lines, and of describing in more detail certain points in the general technique of our own experiments, which have an important bearing on the significance of the results obtained.

Susceptible mice have been placed in contact with infective individuals of the same species, separated from them after a definite period, and placed in contact with a further group of normal mice. This process has then been repeated again and again, the number of groups exposed varying, in different experiments, from three to twelve. In this way the spread of infection has been limited to that occurring between any one group and the group immediately following it.

Under these circumstances infection has spread far less readily than was the cases in other experiments, in which the mice were retained as a single population, and in which infection could spread in any direction among all the individuals at risk.

In our previous studies on the spread of bacterial infection among mice, the resistance of the host has not been especially investigated; although the importance of this factor has been repeatedly referred to in discussing the experimental results obtained.

From consideration of the above evidence the following conclusions may be drawn:

(1) The prevalence of diphtheria-like bacilli in the fauces of scarlet fever patients on admission to hospital does not definitely follow that of diphtheria outside either in seasonal or yearly prevalence.

(2) The majority of such patients show no clinical evidence of the added infection. It is believed that they are instances of scarlet fever developing in diphtheria carriers, and the presence of the organism does not affect the prognosis. However, cases which present such clinical evidence as a mem branous angina are to be considered as instances’ of true double infection and in them the prognosis is bad.

(3) Bacteriological examination of a series of cases belonging to the former group reveals the presence of true virulent diphtheria bacilli in about 36 per cent., and these organisms may persist in a virulent state for a considerable period.

(4) Routine bacteriological examination to find these cases and their sub sequent isolation materially reduces the incidence of post-scarlatinal diphtheria.

In connexion with our inverstigations upon the methods of casuation of food Poisoning outbreaks we considered it advisable to examine a further series of rats.

Transbaikalia or Zabaikalia is a vast stretch of land lying to the north-west of Manchuria, bounded on the east by River Argun from China and on the west by Lake irkutsk and Irkutsk Mountains from the Irkutsk region. Its most important cities are Chita (capital), Stretinsk, Nerchinsk, Kiakhta and Borzia. The region around borzia for a hundred miles is composed of undulating hillocks inhabited by a species of large (cat-size) marmots or tarabagans (Arctomys bobac) whose skin is much sought after for the manufacture of imitation sable. The Russians and Buriats find the flesh of these animals whole-some, and often bunt them for the meat and the fat which they use for greasing shoes. It is among the trappers and eaters of tarabagans that the early case of bubonic plague are usually reported.

(1) The existence of sporadic cases of natural plague among tarabagans has again been established.

(2) The tarabagan is easily susceptible to pneumonic plague produced by the inhalation of B. pestis in spray form, and may contract, besides the acute infection, chronic type of the lung affection.

(3) Our inhalation experiments have proved contacts to be liable to pulmonary plague infection in both acute and chronic forms. The existence of carriers among tarabagans seems also probable.

(4) Tarabagans can be infected with plague by feeding upon artificially infected food, contracting an alimentary type of the disease. The B. pestis is easily cultivated from the intestinal tract of animals suffering from plague.

(5) Our evidence shows that subacute or chronic plague may exist among tarabagans in Mongolia and Siberia, and thus form a connecting link in the epidemiology of plague in those regions.