The results of this investigation may be summarised as follows:

1. The qx or cage-age mortality curve of a herd increases rapidly to a maximum and thereafter descends to an approximately constant level which is much above the level of the qx for normal mice of ages within the range of real, physiological, age of animals living in the herds.

2. At the latest ages under observation the principal factor of mortality is still the specific factor, so that the advantage produced by selective mortality and active immunisation is brought to a standstill far above the zero line. Exposure in a herd under the conditions of these experiments will not produce an ultimate population fully resistant to the specific factor of infection.

3. It is probable that a large proportion, perhaps a majority, of the members of a herd become infected early in herd life, and gradually increase their degree of immunity because variations of environmental conditions, so far as these are measured by the general herd mortality, become less and less influential on the ultimate mortality of the exposed to risk the later the point in time chosen for measurement.

4. Study of the form of the age-mortality curve is still proceeding; at present we cannot offer an adequate mathematical description of it which takes due account of the biological factors requiring attention.

5. Quite provisionally, we attribute the difference in time of the maxima of the qx curves of Pasteurella and aertrycke epidemics to a difference of average interval between infection and death.