in the village if no imported cases had been introduced. This
would probably be looked upon by the villagers as a serious
epidemic due to some climatic cause.
On the other hand, suppose that there have been 40
Anophelines per person in the village. Then the static
malaria there would be zero. But if 500 of the villagers are
now replaced by 500 infected persons, the 40 Anophelines,
though insufficient to maintain static malaria, would be suffi-
cient to spread the disease to some extent among the villagers ;
as will be seen by comparing examples 3 and 6 of section 28
— that is, the examples with 40 and with o Anophelines.
In four months the former will have 500, 450, 409, 376, 348
202 MALARIA IN THE COMMUNITY [SECT.
cases, some of them among the healthy villagers ; and the
latter will have only 500, 400, 320, 256, 205 cases, all among
the imported persons. Thus in the former there will be 50,
89, 120, 143 cases among the villagers, who will, of course,
suffer from an epidemic. But in the end the epidemic will
die out.
In this village then, 40 Anophelines per person will suffice
to spread the disease to some extent from the imported cases
to the healthy villagers, though the various small epidemics
which may be caused in this way will tend ultimately to die
out. But suppose that owing to the blocking of some stream
the Anopheline factor is increased to 60. The imported cases
will now not only spread the disease to the healthy villagers,
but the epidemic, instead of dying out, will continue to increase
until the static limit of one-third of the population is reached.
The village, which previously suffered only from sporadic out-
breaks due to imported cases, will now become permanently
infected. Yet this great change will be due merely to the
small increase of Anophelines from 40 to 60 pei' person — an
increase which few observers would be able to detect.
It is unnecessary to labour this matter further — I have
already laboured it perhaps too much. The point to be
understood is that a small Anopheline variation — too small
to be easily observed — may yet make a great difference in
the local endemicity. Below a certain figure the Anophelines
will be, according to the laws of chance, too few to produce
enough new infections to keep pace with the natural recoveries.
Above that figure they will be sufficient for the purpose, and
the malaria ratio will rise to a definite static limit.
So far, this theorem has been already given in section 28 ;
but we now perceive another point of importance. Though
the Anophelines are too few to maintain a static endemicity,
yet if any at all are present they may suffice (provided that
the other factors b, s, i, r do not possess prohibitive values) to
cause small temporary spreadings of the disease from imported
3o] ISOLATED ABNORMAL AREAS 203
cases among the local population. Only in the case where
there are practically no Anophelines will the imported cases
have no effect at all. The reader will easily grasp this law
by comparing the examples in section 28.
The reader should also remember another point. Our
formulae are based upon the calculation of average chances.
Thus, if the average Anopheline factor is over forty, and a
single case of malaria is imported, we argue that the chances
are that the disease will spread in the locality. We do not
say that the disease must or will spread to a certainty. The
laws of luck come in here. The distribution of Anophelines
cannot be absolutely uniform in any locality. The imported
case may come to reside at a spot where there are few
Anophelines, so that the chances are that the disease will not
spread at all. Or he may reside where the Anophelines are
a little more numerous, so that only a small temporary epidemic
may be caused. Or, if he happens to live where there are
many Anophelines, a permanent rise in the malaria rate may
be produced. Yet in all these cases the average Anopheline
factor for the whole area under consideration may be anything.
Thus, again, the imported case may have no gametids in his
blood ; or he may arrive in winter and recover before the
mosquito season ; or he may use mosquito - nets. This will
be a case of good luck ; but by bad luck none of these events
may occur, and an epidemic may be started or increased.
We must now consider some special cases of importation.
(17). Isolated supervialanous or infravialariotis areas. —
Suppose the case of a country in which the static malaria is
generally low, but in which an isolated area becomes, from
any cause, more highly malarious : what will be the effect of
that patch of high endemicity upon the health of the surround-
ing country of low endemicity ? Evidently, unless the inhabi-
tants of the supermalarious patch are strictly confined to it,
some of them will tend to wander away from it, especially if,
as often happens, they are actually driven out of it by its
204 MALARIA IN THE COMMUNITY [Sect.
unhealthiness. If there are any Anophelines at all in the
surrounding country, these emigrants, when infected, may cause
local spreading, slight, severe, transient or permanent, according
to the local factor. In any case they will tend to increase the
surrounding endemicity. To calculate the increase from the
variation formula, we must add, month by month, the imported
cases to the number of cases remaining over from the previous
month. Thus, though the static malaria ratio of the surrounding
country might be naturally low, the continued importation of cases
from the focus of high endemicity might raise it very considerably.
It may be said that this process cannot continue for ever,
because the focus would soon become depopulated. This
certainly happens in some very unhealthy spots ; but in others
the population of the focus is maintained by constant immigra-
tion, so that it continues to be a perennial source of danger
to the neighbours. And its influence may be felt for consider-
able distances.
Now, consider the converse case — that in which an isolated
patch of low endemicity exists in the midst of a larger area
of high endemicity. The effect should be just the opposite.
There should be obviously less malaria in the whole country
than if the inframalarious patch had not existed at all ; and
the larger the healthy patch, the larger its influence will be.
In other words, we assume that inequalities of malaria
distribution have results similar to those of inequalities of heat
distribution or mosquito distribution. The superheated objects
or supermalarious areas tend to add heat or malaria to the
surrounding objects or areas. But I am not clear that this
resemblance always holds. Thus the people, and perhaps the
Anophelines, in the supermalarious patch may, after a time,
become so immune that they will no longer add very much
to the surrounding ratios. Conversely, the non-immunes in
an inframalarious patch may raise the ratio there owing to
" extraneous infection " (19).
(18). Effect of mass - migratio7is. — Suppose that a large
30] EXTRANEOUS INFECTION 205
number of persons, with a given malaria ratio, are suddenly
and simultaneously moved into a country possessing a different
malaria ratio : what will be the effect ? Such events frequently
happen, as when a regiment of soldiers coming from England
or from a very malarious country is moved into a moderately
malarious country ; or when numbers of coolies are collected
in a locality which has a different malaria ratio from that in
which they were collected.
The effect will probably be that there will be an epidemic
in both cases. An inframalarious regiment of non-immunes
from England is sure to suffer to a degree determined by the
local factors. Its own men will be the principal ones to suffer,
but the numerous fresh cases among them will tend to react
on the local population. In the case of the introduction of a
supermalarious regiment, the surrounding local population will
suffer most, and will in turn react on the men of the regiment.
In both cases the malaria ratio will ultimately tend to settle
down to the static ratio of the locality. Examples of both
of these cases are of frequent occurrence. Large bodies of
labourers collected for various kinds of engineering works are
liable to cause outbreaks, either among themselves or among
the surrounding population, according to the respective endemic
ratios and the local factors. This will be especially the case
if the aggregation of human beings causes an actual increase
in the local Anopheline factor (9). In such a case the aggrega-
tion may have all the bad effects of a patch of high endemicity
described in the last subsection.
(19). Extraneous infection. — By this I mean the infection of
people during occasional visits outside their usual habitation —
a very important case. There are many villages the malaria
ratio of which under natural conditions would be low, but in
which it is maintained at an unnaturally high level by the
fact that the inhabitants work in more malarious areas at
certain times of the year. Thus in the village of Mazi (575
inhabitants), situated several hundred feet above the Plain of
2o6 MALARIA IN THE COMMUNITY [Sect.
Kopais in Greece, I found malaria infection in twenty out of
forty children examined [1906]; but there were scarcely any
possible breeding - pools among the broken hills round the
village. The apparent anomaly was easily explained by the
fact that every year nearly the whole population descends
to the plain for the harvesting in August (the most malarious
month), and camps there for days or weeks at a time. As
the most malarious months are also generally those of the
greatest agricultural activity, extraneous infection is probably
one of the principal equalisers of malarial distribution. In
fact, the result is the same if the factors of endemicity are
temporarily raised in a given locality, or if the people
temporarily migrate to another locality where the factors of
endemicity are higher. The local endemicity depends upon
the local factors, only if the local population remains stationary.
Of course if there are any local Anophelines at all, persons
infected extraneously will tend to spread the disease to some
extent among those who do not leave the locality. Thus
adults infected while harvesting at a distance may infect
their children when they return home.
(20). Slow changes zk endemicity. — Seasonal, annual and
other variations such as we have considered are constantly
occurring in every malarious country ; but, besides these,
certain slow but persistent changes, sometimes affecting very
large areas, are often observed, or at least reported. It is
stated that a whole country is gradually becoming more
and more, or less and less, malarious. As undeniable examples
I may quote the entry and increase of malaria in Mauritius,
and its disappearance from Britain. How explain the
phenomena ? They may be due, not to one, but possibly
to any or all the factors concerned in the malaria equation.
Let us first consider each of the possible factors in turn.
{a) The Anopheline factor, that is, the number of carrying
insects per person, may slowly and persistently vary. As
shown in (15) a scarcely-appreciable variation in this factor
3o] SLOW CHANGES 207
may produce a striking change in the endemicity. Thus with
60 Anophelines per person the static malaria ratio should
stand at about one-third ; but it should stand at double that
with the Anophelines at 120. Yet few persons would be
able to observe this difference in the number of the insects ;
and as no census has been taken in the past, the scientific
student would possess no standard for comparing the present
with it. Yet many things may occur or concur to influence
this factor to a considerable degree.
Thus it is well known that long but inexplicable undu-
lations occur in the weather-cnrv&s. These often affect only
certain tracts of country, and have marked results upon the
local crops and prosperity. They may thus quite possibly
make a considerable percentage of difference in the Anopheline
output, besides tending to influence the recovery factor. It
would be difficult to state precisely what the exact effect
of the weather might be. Excessive rainfall might either
increase or decrease the Anopheline output, especially of
certain species, according to local breeding conditions. Thus
we should observe the change in the malaria without being
able to draw a very clear correlation curve between it and
the weather.
Again, according to Walter [1908], afforestation has a
marked effect on the number of rainy days. The great
exhalation of moisture from vegetation tends to increase the
afternoon thunder showers which occur so frequently over
tropical lands (but not seas) at certain seasons. It has con-
siderable influence on crops, and doubtless helps to keep
breeding-pools filled. Irrigated cultivation is evidently likely
to have a large similar effect, especially with extensive
irrigation works and with certain corps. High -class roads
and railways tend to obstruct natural surface drainage and
also to provide numerous breeding-pools in the " borrow-pits "
along their route (section 57) ; and I have seen the undoubted
effect of them in Sierra Leone, Lagos, Panama and Mauritius.
2o8 MALARIA IN THE COMMUNITY [Sect.
But whether the increase of total breeding surface produced
by them is sufficiently large to modify the average Anopheline
factor over a great area is another question. Probably
irrigated cultivation has a larger effect. General increase of
population may, by section 29 (7), increase the mosquito out-
put out of proportion to the increase of the human population,
by supplying more food to the female Anophelines ; and
diminution of prosperity may have the same result, by (12).
Thus many things may modify the Anopheline factor ; and
as we possess no exact method of measuring mosquitos it
will always be difficult to decide how or why the factor has
been changed, or if it has been changed at all.
{b) Changes in the carrying and biting factors may occur
as the result of a gradual predominance of certain species of
Anophelines. As we now know, the fauna and flora of a
country frequently change to a considerable extent. Those
originally found in ocean islands, like Mauritius and the
Seychelles, are apt to be almost swamped by imported
species. Quite possibly, therefore, more " virulent " species
of Anophelines may gradually increase very greatly in certain
areas, at least for a time under certain cycles of weather or
with certain kinds of cultivation.
{c) Human iuimigratioti into certain areas, especially in
connection with large planting, railway, mining or irrigation
works, may tend to create /^a of malaria (17); to enhance
the local rate by the importation of cases or even of non-
immunes; to increase the convection factors s and b, by destitu-
tion or bad housing ; or to raise the Anopheline output by
additional food.
id) The habits of the people may change in consequence
of prosperity, famines, pestilences or increase of civilisation.
Thus they may use better houses, or houses of two storeys,
or mosquito - nets, or glass windows. Or a larger proportion
of medical men may practise among them ; or quinine may
3o] EXAMPLES 209
be more accessible ; or they may abandon cultivation for trade
or take to cities — and so on.
It will always be a very difficult matter to select the most
probable factor. The final result will generally be due to
the balance of all the forces, some pulling this way and some
that. S. R. Christophers and C. A. Bentley have recently
[1909] considered that a slow increase in certain Indian areas
is due not so much to the Anopheline factor as to (c) — that
is, immigration in connection with large economic develop-
ments. Their paper is thoughtful, but, unfortunately, they do
not deal with their subject quantitatively. They give us no
estimate of the degree of the reputed epidemics, of the pro-
portion of immigrants and of area occupied by them to total
population and area, of their methods of enumerating the
Anophelines, and, indeed, of comparing the effects of any of
the possible factors. It is evident that immigration and allied
factors may have some results, but it is necessary to calculate
/iozv much compared with those of other influences. Without
some attempt at exact quantitative estimates, the scientific
study of the diffusion of malaria is as impossible as that of
the diffusion of heat. Many similar medical papers on the
subject have the same deficiency.
(21). Examples: Mauritius, Isniailia, Clairfond, Britain. —
In section i reasons were given for the opinion that malaria
might have been comparatively scarce, or entirely absent, in
Greece and Italy before the historic period, and that after
about 550 B.C. it probably increased greatly in amount. This
theorem can scarcely be proved ; but I will now describe briefly
the similar events which have occurred within the memory of
living persons in Mauritius and Reunion.
The island of Mauritius, situated about 550 miles (885
kilometres) east of Madagascar, and lying just within the
southern tropic, covers an area of 705 square miles (1,825
square kilometres), rises to a height of about 600 metres in
the centre, from which it slopes down more or less gradually
o
2IO MALARIA IN THE COMMUNITY [Sect.
to the sea. The climate is the usual tropical marine climate,
warm, equable and humid, and the vegetation is rich. Dis-
covered in 1 507, it was peopled by Dutch and then by French
settlers, who were always introducing slaves from Africa and
Madagascar. In 18 10 it was taken by the British. The
slaves were set free in 1834, and from next year onward
an immense importation of coolie labour from India com-
menced and continued. Thus more than 20,000 coolies were
introduced in each of the years 1843, 1854, 1858 and 1865,
and in 1859 no less than 48,377 were imported. Quite
certainly large numbers of these immigrants must have been
infected with malaria, yet, equally certainly, no endemic
malaria occurred in the island all this time. I examined
the point with the utmost care for the purposes of my Report
to the Government [1908], looked through endless returns
and studied the voluminous evidence placed before a Com-
mission of Enquiry. A. Davidson [1892] and many other
writers have reached the same conclusion. In Saint-Pierre's
classical novel Paul and Virginia (1789) no mention is made
of malaria in the island. Dr Lorans, the late lamented Medical
Director, who always lived there, informed me that there
was none when he was a boy. At that time the planters
and whites lived in houses scattered all over the country,
and they and the British troops and officials lived entirely
free from local infection. Relapses of malaria and other fevers
such as relapsing fever, were, of course, noted among the
immigrants, but the malaria did not spread, though a few
medical men suspected the existence of occasional sporadic
endemic cases since (?) 1857. In 1865, however, a small
outbreak commenced in a marshy district near the capital,
Port Louis. Next year the epidemic reached the capital, and
in 1 866- 1 867 attained alarming proportions. It was like an
epidemic of plague or cholera. One quarter of the inhabitants
of Port Louis died from all causes in 1867, and 6,224 out of
a population of 87,000 in one month (April). The whole
3o] EPIDEMICS IN MAURITIUS 211
death-rate of the island rose from 32/1000 in 1866 to 120/1000
in 1867, the total population being 360,000. The disease
spread in two years all over the lower parts of the island, and
has remained, and probably increased, ever since (section 32).
In the sister island of Reunion, 125 miles distant, precisely
similar events occurred at the same time. There also there
had been no previous endemic malaria in spite of the intro-
duction of slaves from Madagascar and of Indians from
India.
How explain this extraordinary and terrible event, which
shattered existing notions about malaria? The present carrier
is P. costalis, an African, not an Indian mosquito. If it had
existed in the islands in large numbers the epidemic would
have commenced long previously — certainly with the beginning
of the copious Indian immigration, thirty years before its actual
commencement. If it had existed in small numbers, below
the endemic limit, it ought still to have spread the disease
to some extent, according to the reasoning in (16); but most
authorities denied that this occurred at all before 1865. More-
over, on both these suppositions, the chances are that the
epidemic would not have commenced simultaneously in both
islands, especially because the Indian immigration into
Mauritius was much greater than that into Reunion. I can
only conclude, therefore, that P. costalis was introduced from
Africa or Madagascar simultaneously into both islands,
probably by the same vessel, a year or two before the
epidemic ; that it spread rapidly and soon multiplied over
the endemic limit ; and that from this point the success and
ultimate permanence of the malarial invasion was assured.
Of course the inhabitants would not be able to distinguish
the new-comer among the swarms of innocuous mosquitos
already present, and there appears to be no reason why,
after being once introduced, it should not have reached the
endemic limit of numbers in a year or two. This therefore
appears to be wholly a case of the Anopheline factor. The
212 MALARIA IN THE COMMUNITY [Sect.
human factors had been present for at least thirty years
without causing the disease.
A similar case is that oi Isviailia (section 53). Situated on
the Suez Canal in the desert, it remained free from malaria until
the fresh-water canal was completed in 1877, when malaria
appeared almost immediately. Since then the disease con-
tinued to increase with fluctuations, until it was banished in
1902 by proper mosquito reduction. Here again the outbreak
was probably due almost entirely to the Anopheline factor.
A more difficult case is that of an outbreak in the
neighbourhood of Clairfojid Marsh in Mauritius (section 30
(i) map), already frequently mentioned. This marsh existed
on a plain 1,400 feet (427 metres) above sea-level, and the
epidemic is described in my Report [1908] by Dr de Chazal
who had been practising in the neighbourhood since 1890,
and by myself, and also by Major Fowler [1908]. The history
is as follows : —
Before 1903 Dr de Chazal, who lived not far from the marsh,
had observed a few cases of malaria, but only among Indians.
In that year, however, cases began to occur among well-to-do
people living in the neighbourhood, and the epidemic now
developed rapidly. The cases at the neighbouring dispensary
were as follows : —
1904 1905 1906 1907
346 843 1,147 1,487
and similar rises in the cases of malaria at seven other
dispensaries, situated not far away, also occurred. In 1906
the British soldiers in barracks (see subsection i) began to
be affected. In 1908 Major Fowler and I witnessed an
epidemic of seventy-one cases in January among these troops,
and on making a house-to-house visitation, found a spleen
rate of 241/339=71%, and an average spleen of 4-12 among
the Indian children living round the marsh. The disease was
being carried by P. costalis, which our " moustiquiers " found
3o] MALARIA IN BRITAIN 213
with considerable difficulty, and not by M. mauritianus which
abounded round the marsh. The marsh was now drained by
the Government ; quinine was simultaneously distributed in
the houses by ourselves and a dispenser; and the epidemic
ceased.
How explain it? The marsh had been there as long as
any one could remember, and P. costalis, though it prefers sea-
level, had probably been breeding in it in small numbers all
the time. Indians as well as the richer classes had been living
near at hand, and imported cases had certainly been common.
Why, then, did not the epidemic commence before? Up to
ig02 the population round the marsh had been comparatively
scarce, but in that year a great increase of Indians took place
owing to the advent of a regiment of Sepoys, and many of
these Indians built huts close to the marsh waters. Here, then,
might be a case of creation of a focus of high endemicity ; but
still I find it difficult to explain the focus without supposing an
increase in P. costalis, pari passu with that of the human popula-
tion. The case is most probably explained by an increase due
to more abundant food in the new huts built close to the marsh,
according to sections 29 (7) and 30 (9 and 12) ; and possibly by