21

of the Indus where this river first debouches from the hills in the north of the
Mianwali District. A tract occupying part of Ludhiana and extending up
between the Hoshiapur and Jullunder Districts, will be observed to lie across,
but not to extend along, the Sutlej valley.

      But though the fringes of the epidemic show a relation to rivers, etc., the main
epidemic areas show comparatively little relation to any physical features and affect
without exception the thanas over a very wide extent of country.

      The conditions in such areas are altogether extraordinary. In the northern
epidemic area throughout a tract covering a great part of Gujrat, half of Gujran-
wala and a part of the Lahore District, there is not a thana in which the death rate
for the month of October did not rise over seven times the normal rate, whilst the
death rate for the thana of Gujrat itself was eleven times greater than the normal ;
and if we can credit the figures there are villages in the Gujrat thana showing in
some cases a mortality as much as thirty times the normal.

      In the special focus at Bhera there are scarcely any villages showing a death
rate below 200 per mille, whilst the majority ranged between 200 and 400, and one
shows the extraordinary figure of 641 per mille.

      If it were not for Bhera town itself, which shows an equally high mortality
rate, we should hesitate to credit these extraordinary values. Yet if there be
errors of registration, one would think they must err on the side of under-
estimating rather than overestimating the mortality.

      In the southern area scarcely a thana south of Delhi had a death rate under
seven times the normal, whilst again there are areas in the centre of the epidemic
area which have been even more severely affected, the rates varying from ten to
seventeen times the normal.

      In the thana of Nuh for example, where there normally are about 125 deaths
in the month, there occurred in October 1908 no less than 1,404 deaths. In the
thana of Hodal, which normally has about 100 deaths, there occurred in October
1,083 deaths. In the thana of Ferozepore, with a normal of 130 deaths per
month, there occurred 2,346 deaths.

      In the case of the southern area the nucleus is almost coincident with the
Gurgaon District. The population of the Gurgaon District is nearly three-quar-
ters of a million. In October the recorded death rate per mille was 267. This
being the average of all villages in the district, it is more than probable that
many of them must have had rates much higher. As a rule the maximum
intensity reached by the mortality in 1908 was about 400 per mille. Such effects
are not seen elsewhere than within the circumscribed zone of an epidemic
nucleus.

      In the case of towns if we refer to the figures we shall see that those situated
in areas of special intensity suffered most severely in the epidemic of 1908.