PLANNING MALARIA CONTROL 
287 
for the years 1926-1939, inclusive. It is 
apparent that either graph would serve to 
show the relative annual intensity of ma¬ 
laria. Incidentally the graphs show the 
effect of sustained efforts at improving the 
reporting of malaria cases and deaths. 
Prior to 1932, the reported case-fatality 
ratio was ten or less, whereas, after 1931 it 
exceeded ten. During the last three years 
shown, it has been from 25 to 30. Simi¬ 
larly, Fig. 2 is an array of 10-year, monthly 
malaria case- and death-rates for the same 
state. Considering the known defects in¬ 
herent in the data, the two graphs exhibit a 
remarkable degree of parallelism in trend. 
This gives a useful picture of the seasonal 
occurrence of each attribute exemplifying 
the well-known lag in the rise and decline 
of deaths as compared to cases. It also 
shows what is either a reportorial or actual 
difference in the case-fatality ratios in the 
winter versus the other months. 
Thus, employed conservatively and with 
a proper understanding of their limitations, 
either morbidity or mortality statistics may 
serve to point out certain useful time-ma¬ 
laria relationships, provided the area or 
population involved is sufficiently large. 
They may also be used to advantage in de¬ 
fining certain space-malaria relationships, 
again provided that the time span is great 
enough. Some extra caution, however, is 
required in interpreting the distributional 
maps thus developed, of which Fig. 3 is an 
example. 
While geographic distributions of mor¬ 
tality are probably more dependable than 
those of morbidity for the reasons pointed 
out above, it must be remembered that they 
show little more than the intensity-distribu¬ 
tion of estivo-autumnal fever. This may 
or may not be associated with tertian ma¬ 
laria. Where tertian predominates to the 
exclusion of other types of malaria through¬ 
out the season, morbidity should be plotted 
instead of—or as well as—mortality, assum¬ 
ing that cases have been reported with some 
regularity for several years. 
Rate maps and unit maps have different 
distributional significance. The first show 
variations of attributes expressed as pro¬ 
portions of the population. Thus in areas 
including large cities many of whose inhabi¬ 
tants are at minimal risk of malaria, rate 
maps may be misleading because they indi¬ 
cate low rates but do not reveal the effect 
of their application to large populations. A 
good example of this is shown in Fig. 3. 
Chatham County, in the most easterly pro¬ 
jection of the state, has a low malaria death- 
rate, but this is due mainly to the fact that 
95,271 of 116,803 persons in Chatham 
County live in Savannah where exposure 
to malaria is virtually negligible. Never- 
