Medical Officers of the Army of India.

55

ordinary. It has already been mentioned that there may be heavy rain in
June, or even in both May and June, as in 1887, and yet no marked increase
of fever will occur until July. This shows that the increase does not depend on
either the variations of temperature, which will be most marked at the beginning of
the rains, or on chills produced by getting wet (a very strong point in favour of
my theory), and it has already been shown that the absence of fever in June
is due to the water being too low down to cause it during that month. Now,
in the year 1892, not only is the rainfall deficient, but it is also very evenly
distributed, not quite 12 inches having fallen during any one month, so that
it is evident that in this year the water-level must have been lower and more
constant than in any other year of the series, and hence the nearly complete
absence of fever. Moreover, it will be seen that there was only 331 inches
of rain between October 1891 and May 1892, so that the water-level must
have been unusually low at the beginning of the rains in this year (at least
forty feet), which, again, is an important factor, as it would lessen the influence
of the small amount of rain that actually did fall, in the rainy season, in raising
the water-level. Thus, it is evident that everything combined to make the
conditions least favourable for the incidence of fever during that year. On
the other hand, there was a low rainfall in 1890, but it was more unevenly
distributed, and there was more rain in March and May, so that, although
the fever in the rains (for that in the early months of the year is due to an
entirely different cause) was less than the average, yet it was greater than in
1892. Thus, all the peculiarities of the distribution of the fever during the
rainy seasons are easily explained on the above-mentioned supposition, but
that which occurs in the cold weather months does not come under the same
laws, and remains to be considered.

       That the amount of rain in the early months of the year does not in-
fluence the fever rate during the same months, is evident from a comparison
of the small amount of fever during the first five months of 1895, in spite
of unusually heavy rain during the same period, with the high rate of fever
in the same months of 1889, accompanied by a low rainfall. For a long time
I was at a loss to account for the considerable variation in the amount of
fever in the early months of the different years, although the fact that it did
not coincide with the rainfall was not surprising, seeing that the ground-water
level during the cold weather months of 1895-96, the years in which I took
measurements, reached a depth of 30 feet from the surface by the end of
November, and remained at below that level until the onset of the next rainy
season, and the very slight variations which it undergoes at that depth could
not be expected to influence the fever rate. In fact I believe that the malarial
organisms will be found in largest numbers in the damp earth at or just above
the water level, and hence, during the cold weather, will be too far down to be
able to gain access to the air in large enough number to cause many fresh