We may summarise our results, including those recorded in our two recent reports (Topley and Ayrton, 1924 b and c) under the following headings.

(1) The antigenic structure of B. aertrycke (Mutton).

(a) B. aertrycke (Mutton) undergoes natural dissociation into two sharply distinguished varieties, one containing that antigenic constituent which is characteristic of this bacterial species or variety, the other that antigenic constituent which B. aertrycke (Mutton) shares with the bacterial group to which it belongs. We have obtained no evidence that both the type and group antigen may be fully developed in one and the same bacillus. In these respects our results confirm entirely those recorded by Andrewes (1922).

(b) Both the type and the group varieties of this organism, when grown for more than 16 hours at 22° C., or for that length of time at 37° C., tend to undergo an antigenic alteration such that they become agglutinable with an entirely different type of antibody, to which they did not previously respond. The antigen developed as a result of this alteration would seem to be the same for the type and the group varieties. Since plating from such altered cultures gives colonies of the usual type or group variety, it would seem probable that the altered bacilli are dead, dying or senescent organisms.

(c) Although type or group cultures readily give rise to individuals of the alternative antigenic variety in ordinary laboratory media, yet this change is neither so rapid nor so unpredictable in its occurrence that it is impossible to determine the nature of any given colony, or to obtain a culture of the serological variety required.

(d) There is some evidence that type variants arise in group cultures more readily than do group variants in type cultures.

(2) The relations between the presence of type or group antigen and other biological attributes of B. aertrycke.

(a) Roughness and smoothness vary independently of the presence of group or type antigen. Rough forms may react as group or type.

(b) There is no difference between the virulence of the type and group varieties of B. aertrycke. While the virulence of all smooth strains is high, the virulence of all rough strains is low.

(c) There is some evidence that group strains of B. aertrycke give rise more readily to faecal excretion than do type strains, but the difference is not great. Rough strains appear, in this as in other respects, to be ill fitted for a parasitic existence, and do not give rise to persistent faecal excretion.

(d) There is a well-marked tendency for group strains, which have gained access to the tissues, to give place to the type variety. The longer the sojourn in the tissues, the more complete is this replacement. Type strains, under similar conditions, do not tend to be replaced by group strains.

There is a similar, but less marked, tendency for group strains to be replaced by type strains in the intestinal canal.

(3) The excretion of B. aertrycke in the faeces, after administration by the mouth, and its relation to other phenomena of infection.

(a) The excretion of B. aertrycke in the faeces, after administration per os, may be continuous or intermittent, or may not occur in a sufficient degree to be detected by the technique employed.

(b) The faecal excretion, in those mice which succumb to infection, is no more regular in its course than in those mice which survive. It is a frequent occurrence for mice to die with typical lesions of enteric infection, without ever having excreted B. aertrycke in detectable amounts.

(4) The condition of mice which have survived the oral administration of B. aertrycke.

(a) A high proportion of mice, which have survived for 42 days after the oral administration of B. aertrycke, yield cultures of this organism from their spleens when examined post-mortem. In many series of experiments the proportion of survivors showing such positive spleen cultures has exceeded 50 per cent.

(b) The strains of B. aertrycke isolated from the spleens of such survivors appear to be possessed of the normal degree of virulence.

(c) The presence of agglutinins in the blood serum can be demonstrated in only a very small proportion of such surviving mice.

(5) The relation of dosage to the phenomena studied.

(a) With a single dose, there is a definite relation between the number of viable B. aertrycke administered and the frequency of the phenomena indicating infection. The larger the dose the higher is the mortality rate, the more frequent and persistent is the faecal excretion, and the higher is the percentage of the survivors which harbour B. aertrycke in their tissues.

As the dose decreases there is, at first, a rapid fall in the frequency of these various phenomena, but when the dose falls below a certain limit further decreases produce relatively little effect, over the range of doses studied.

(b) Repeated administration of small doses of B. aertrycke results in a high frequency of persistent faecal excretion.

In conclusion we should wish to express our thanks to Miss E. R. Lewis, for her assistance during a large part of the investigations recorded above.