26 SECTIONAL ADDRESSES 
ten cases; and with dry Januaries also there is a similar two-to-one 
chance of a prolongation of the character. It is this persistence, especially 
when it is preceded by abnormal features in other regions, that seems 
now to hold out most promise of reliability in forecasting. In agricultural 
countries in which a failure of the rains involves a national calamity, the 
desirability of making preparations in advance has long ago led to efforts 
at prediction ; and the demand has been so great that the supply has 
been forthcoming before its quality would bear the most cursory 
examination. The causes of unusual weather seem hopelessly obscure 
to the layman ; and hence primitive ideas, surviving in the depth of our 
natures from countless ages of magical practices, still come to the 
surface in connection with it. In India I have been officially asked 
what is the need of an expensive and difficult scientific inquiry into the 
causes of drought when Hindu astrology will indicate what is coming ; 
and many a country that claims to be dominated by Western science fails 
to recognise that events in weather obey the ordinary laws of physics and 
chemistry. The almost universal idea that weather must repeat itself 
after a certain number of years finds its origin, I believe, ultimately in 
the ancient belief in the control of our affairs by the heavenly bodies 
with their definite cycles—a belief which clearly shows itself in the 
supposed influence of the moon on the weather. Be that as it may, the 
faith in periods is so deep-seated that even in scientific discussions the 
ordinary tests for validity are very often ignored : more than once I have 
seen in journals of repute the artless remark of an author that if he were 
to limit his results to those which would satisfy the criteria of reality 
he would obtain few results of interest ! 
Another regrettable feature of current practice, even in important 
memoirs, is that of classing together processes with true periods and 
those sometimes called ‘quasi-periodic,’ of which the period varies. 
If our ideas are to be applied with success in the present enterprise 
their currency must be stabilised, and no good can come of attempting 
to pass off a vague surge of a few years as a three-year period. 
After these preliminary remarks I propose to make a rapid sketch 
of the relationships that have been found between seasonal features in 
different parts of the world, then to describe the efforts that have actually 
been made to issue long-range forecasts, and finally to consider the 
directions from which improvements can be hoped for. 
In the collection of World Weather Records, of which the publication 
was made possible by American generosity six years ago, there are about 
a thousand series of monthly data of pressure, temperature and rainfall ; 
and these form but a scanty network. If quarterly values were com- 
puted and correlation coefficients between each pair for contemporary 
seasons, as well as for seasons one quarter before and after, we should 
have about four million coefficients. Co-ordination and generalisation 
are imperatively called for, and the development of the subject lies in the 
discovery of regions over which the variations are linked together. 
After preliminary efforts by Buchan, Hoffmeyer, Blanford, de Bort, 
Hann, Meinardus and Pettersson, the far-reaching possibilities were 
first visualised by Hildebrandsson, who plotted pressure curves for ten 
