1. The bacterial content of the air of the Underground Railways, when the average of all results of the bacteriological investigations is taken, does not numerically compare unfavourably with the outside air of London.

2. The ratio of the number of organisms growing at room temperature appears to be about 14 for railway air to 10 outside air. For those growing at body temperature the ratio is considerably higher, namely 2 to 1 respectively. The mean per litre, for room temperature organisms, is about 9 in railway air, 6·3 in the outside air; for body temperature organisms 4·6 for railway air, 2·2 for outside air.

3. The bacterial content of platform air, except on the City and South London Railway, would appear to be higher than that of carriage air; the total mean for platform air being 52 and for carriage air 42·8 organisms per 5 litres, or a ratio of 16·4 and 13·5 respectively to 10 of the open air. The higher proportion in platform air is generally speaking to be accounted for by the greater amount of draught and dust disturbance.

4. The ratios of the total bacterial content of railway carriage air and carriage and platform air on the six lines to open air are estimated in the following proportions:

Spontaneous agglutination is a not uncommon occurrence, especially in the case of organisms of the enteric and diphtheria group, where it causes some difficulty in serological diagnosis. Many investigations have been made to clear up the principles of this phenomenon. Recently Arkwright (1921) has thrown much light on this problem. He was able to isolate from old pure cultures of dysentery, typhoid, paratyphoid and enteritidis (Gaertner) bacilli, two variants which he named the “R” (rough) and “S” (smooth). The “R” variant agglutinates spontaneously in physiological saline and in broth, while the “S” variant forms good emulsions in both. Specific immune sera prepared for these two variants show only slight cross-agglutination. Arkwright regards the two forms as potentially inherent in most of the individuals in the young culture, which may become segregated in different individuals in old cultures. “Rough” varieties have been observed also by Zoeller (1922) in B. dysenteriae Shiga and by Schütze (1921) in B. paratyphosus B.

“Health is more affected by atmospheric conditions than by any other influence. From the moment the new born child utters its first cry, until life finally ceases, he is directly and immediately dependent on the atmosphere he breathes for continued existence. Records of sickness and of mortality clearly show that the nearer atmospheric conditions are to those of the open country, the better is the health of the population, and that the more these conditions are modified owing to urbanisation and industrialisation, the more pronounced is the ill-effect upon general health” (1).

“Virulence” is an ill-defined term and its use frequently gives rise to disputes, which remain unresolved because the disputants have no common basis other than that the term applies to organisms which cause disease. Thus it is imperative that the writer clearly defines the meaning of the term virulence as used in this paper, that the issue may not be confused by the reader approaching the subject from a different point of view.

1. There is some evidence that an injection of 0·00001 c.c. of horse serum causes an appreciable degree of active immunity to horse serum in rabbits.

2. Rabbits actively immune to horse serum receiving a series of injections of antitoxic horse serum at intervals of three to ten weeks eliminated the antitoxin at rapidly increasing rates.

3. An intravenous injection of antitoxic horse serum given six weeks after the second of a series of 0·5 c.c. in a rabbit already actively immune to horse serum was eliminated so rapidly owing to excess precipitin that less than 1/1000th remained 24 hours after injection.

4. Another rabbit which, six days after a previous injection, was given

serum intravenously, lost over 99 per cent. of the antitoxin within 15 minutes of the injection.

5. The maximum of circulating antitoxin detected after the second subcutaneous injection into a sensitised rabbit was only 1/15,000th of the amount injected.

6. Antitoxic horse serum injected intravenously into rabbits at weekly intervals was rapidly eliminated and less than 2 per cent. could be detected 24 hours after each injection.

7. After 12 daily intravenous injections of antitoxic horse serum into a rabbit, the antitoxic content was only one-half of that after the first injection; later injections were not eliminated so quickly and precipitin formation gradually ceased.

8. The rate of elimination of antitoxic horse serum in a rabbit may be greatly delayed by daily injections of normal horse serum.

In recent studies on experimental epidemiology, and especially in a series of reports by Webster (1922, 1923 a, b, c, d, e, f and g), attempts have been made to standardise certain of the factors concerned in the spread of bacterial infection among mice. One such factor is dosage, to which great importance has been attached by Amo⋅s (1922 a and b) and also, though to a somewhat less extent, by Webster, whose later work leads him to emphasise the importance of host-susceptibility in determining the form of the epidemic wave. Webster would appear to regard variations in dosage as decisive, in the sense that an effective dose must be applied; but once the critical limit of dosage is exceeded he would give first place to host-susceptibility in determining the subsequent course of events.

In the preceding report (Topley and Ayrton, 1923 a) we have described a technique by which measurements may be obtained of the excretion of B. enteritidis (aertrycke)1 in the faeces of mice.