View report page

45
flea prevalences recorded in 1920 and 1921 than were actually observed. In illustration of the complicated
character of the issues arising a word may be said on a parallel instance, which is presented on study of
the relation between rainfall and scarlet fever incidence. Referring to the diagram on page 41
the years 1890 and 1891 had been wet years with, it will be seen, scailet fever below the average;
1892-3 were dry years, with high prevalence of scarlet fever; then 1894 was wet, and although the
rainfall was not excessive it came (see diagram) at the end of an exhaustive epidemic, with obvious effect.
Then followed eight successive dry years 1895-1902 : The influence of deficiency of rain upon the scarlet
fever in the first half of this period is clearly apparent; but in the latter half scarlet fever, notwithstanding
the dryness, tended to fall below the average. The excessively wet year 1903 then followed, with marked
effect upon the scarlet fever record even in the succeeding dry year. The four years thereafter were dry
with the exception of 1906, which was sufficiently wet to restrain to some extent the then oncoming
epidemic. In 1908 the rainfall was only just below the average, but in 1909 and, more especially, 1910
the rainfall was in excess, a fact clearly reflected in the scarlet fever curve. Then four years supervened
of rainfall rather below the average, with epidemic scarlet fever, though the deficiency of rain in 1911
and 1914 was almost negligible. The four succeeding years 1915-1918 were all very wet, closing down
perhaps prematurely the epidemic of 1913-1914. There then follow the years 1919-20 with rainfall
below the average, and finally 1921 and its record drought and excessive scarlet fever. Although the
effect of the rainfall can thus be broadly discerned from year to year, and is at times very clearly defined,
the coefficient of correlation of rainfall and scarlet fever prevalence over the period of thirty-one years
(-.57) is not very significant, the explanation offered being that the effect produced by an excess of
rain depends upon the susceptibility of the population at the particular time.
The relation between flea prevalence and the scarlet fever curve may be similar to that for rainfall.
A high prevalence of fleas would not necessarily connote an excessive amount of scarlet fever or vice
versa ; other determining factors, for example, the proportion of susceptible units in the population at the
time, must also be taken into account. The size of a fire does not depend upon the number of sparks
which light it. In this connection reference may be made to the contention of Professor Tendeloo of
Leyden, that we must have in mind the "constellation" point of view in considering cause and effect
relationships in medicine. The "Lancet" (1922, p. 410) explains this astronomical metaphor as being
intended to indicate that there are "certain prominent factors and others less conspicuous, whilst
beyond lie possibilities of still undiscovered influences." Under such conditions close correlation, say
of scarlet fever prevalence with rainfall or with fleas, is not to be expected ; and yet both the latter
may exert influence upon the epidemic waves.
Certain special factors operating in the present case may be mentioned. In the first place it
was suggested in last year's report (p. 54) that some proportion of the cases of scarlet fever notified
among school children late in the year really belonged to the influenzal group of diseases, and that
influenza was thus exerting a disturbing influence; a special report by Dr. Armstrong was prepared on
the "catarrhal sore throat or glandular fever" prevalent in the schools at the close of 1920 (loc. cit. p. 89).
At the close of 1921, moreover, it was again suspected that some of the cases of scarlet fever notified
were really cases of influenza. The surmise that confusion between scarlet fever and influenza has been
occurring is supported by the fact, to which Mr. Spear has called my attention, that a sharp rise in the
case-mortality of "scarlet fever" has occurred in each of the recent influenzal prevalences ; this rise
was quite a considerable one in the prevalence of February-March 1919 (see diagram). It is interesting
to observe that the occurrence of skin eruptions was commented upon in the Times of 16th April, 1921,
and 15th January, 1922. Moreover in a discussion in Manchester on scarlet fever and scarlatiniform
rashes (Brit. Med. Journal, 18th February, 1922, p. 27) Dr. McClure remarked that "though he
believed that the great mass of cases notified were truly scarlet fever yet there was a number in
which differential diagnosis between mild scarlet fever and mild influenza was not clear."
And then there is the second point, already glanced at in last year's report (p. 86), viz., the recent
behaviour of the flea curve considered in its relation to prevalence of scarlet fever. The percentages
of flea-bitten school children at the times of seasonal maxima of scarlet fever in 1919 and 1920 were about
17.5 per cent. at the time of the autumnal maximum of 1919, and 22 per cent, at the corresponding
maximum of 1920. In 1921, a slight rise in the flea curve occurred a few weeks before the maximum of
scarlet fever was attained (as had been the case in each of the preceding 12 years for which records are
available), but the percentage of flea-bitten school children did not show an increase upon that of 1920,
but reverted practically to the 1919 standard, although the scarlet fever of 1921 actually exceeded that
of 1920. As was remarked last year, this phenomenon was perhaps only to be expected, for the question
of the extent of accumulation of susceptible children here necessarily assumes considerable importance.
It must be realised, in fact, that (as the diagram on page 46 shows) in 1920 the seasonal rise was from a
plateau of some 300 cases a week in the early part of the year to a peak of some 1,000 cases a week, while
the corresponding rise in 1921 was from a plateau nearly twice as high (five to six hundred cases a week) to
a slightly higher peak still (1,200 a week). Moreover, and this is the main point, the level of scarlet fever
prevalence in London from 1916 onwards until early in 1919 had been but rarely as high even as 200
cases a week. Thus it may be calculated that, at the commencement of the present major prevalence
of scarlet fever in 1920, there were for this reason alone at least fifteen thousand more " susceptible
children" in London than at the commencement of the preceding major prevalence of scarlet fever in
1913. Hence, in the main, the sustained prevalence in the early half of 1921, with a rise to a still
higher level in the autumn.
These occurrences in fact, illustrate Dr. Ransome's contention of some 40 years ago, that " a
certain density of the population at susceptible ages is necessary before a disease can spread with the
vigour of an epidemic." This law was made use of 16 years ago to illustrate the form of the London
measles curve (Milroy Lectures 1906); it may be applied to epidemic throat diseases also, for here too