145 
NATURE 
[APRIL It, 1912 
associated with lowest sun-spot conditions, while the 
frequency of tropical cyclones and abundance of rain- 
fall, which should prima facie show the same rela- 
tionship, suggested that there was increased move- 
ment and evaporation about the maximum epoch of 
the sun-spot period. As the results obtained from the 
different groups of facts could not be reconciled, the 
discussion of the subject was for some years in abey- 
ance. Broader views are, however, now being taken; 
and it is realised that many sun-spots may mean 
increased rainfall in one part of the world and de- 
creased in another, or the like inversion of any other 
meteorological element. Also, the development of 
means of obtaining more complete records of changes 
upon the sun’s surface, in addition to those mani- 
fested by sun-spots, has encouraged further inquiry 
into the subject of solar influence. 
Full knowledge can only be secured when the new 
methods have been used for many years, but, so far 
as the discussion has proceeded, it indicates that there 
is no real inconsistency in the earlier conclusions, 
and that studies of the sun offer the most promising 
prospects of success in long-range weather prediction. 
Meteorological analysis of observations made at a 
solar physics observatory has become even more 
important than consideration of the results from an 
astronomical point of view. As was remarked in a 
report of the U.S. Weather Bureau a few years 
ago :—‘‘Advances in -the period and accuracy of 
weather forecasts depend upon the exact study and 
understanding of atmospheric pressure over large 
areas, and a determination of the influences, prob- 
ably solar, that are responsible for ordinary and 
extraordinary distributions of atmospheric pressure 
upon the earth’s surface.” 
We may take it for granted that the weather of 
any region is determined mainly by the barometric 
pressure and the interchange of areas of high and 
low barometer. The most important variations to 
consider are, therefore, those of atmospheric pressure, 
and the conclusions arrived at may then be used for 
comparison with variations of solar phenomena in- 
dependently determined. The unit of inquiry should, 
however, be the world, and not one particular region ; 
and there should be no assumption, as was formerly 
common, that solar changes indicated by sun-spots or 
other phenomena would affect the whole of our globe 
simultaneously and in the same direction at any par- 
ticular epoch. Investigations carried on in recent 
years in this spirit have led to results which are both 
stimulating and valuable. For convenience we will 
refer first to the meteorological conclusions, then to 
solar changes, and finally to the relationships found 
between sun and earth. 
Periodic plus and minus ‘‘pulses”’ of rainfall in 
India were described by Sir Norman and Dr. W. J. S. 
Lockyer in a paper read before the Royal Society in 
November, 1900, and were shown to be related (some- 
times inversely) to similar variations at Mauritius, 
Cordoba (South America), the Cape of Good Hope, 
and other places. Attention was then devoted to an ex- 
amination of the variations of pressure over the Indian 
and other areas, and a period of about 3:5 vears 
(always referred to later as the 3-8-year periodicity) 
was found in the mean variation of pressure over the 
whole of India and at individual stations, and also in 
other large areas. An inverse variation was found 
in the pressures at Cordoba, and, referring to it, the 
authors remarked :—‘‘The cause, therefore, which 
raises the mean value for the low-pressure months 
over the Indian area would appear to lower the mean 
value of high-pressure months at Cordoba simul- 
taneously. In fact, we have a see-saw”’ (Proc. Roy. 
Soc., Ixx., June 19, 1002). The area affected by this 
barometric see-saw was extended in a later paper 
NO. 2215, VOL. 89] 
” 
(Proc. Roy. Soc., 1xxi., December 4, 1902) to Ceylon, 
Java, Mauritius, and Australia, and further results 
were described in 1904. 
H. F. Blanford and Hildebrandsson had previously 
found similar evidence of reciprocal barometric varia- 
tions in widely separated regions, but the extension 
of the investigation to about a hundred stations in 
various parts of the earth led to the important result 
that there exists a world-wide barometric see-saw 
between two nearly antipodal parts of the earth, one 
region about India and its neighbourhood showing 
exactly opposite effects as regards atmospheric 
pressure in any year to those felt in a region which 
includes South America and the southern parts of the 
United States. 
The pressure variations in the British Isles year by 
year do not go up or down with either side of this 
see-saw, but appear to be a mixture of both types. 
During some years the British area is enveloped in 
the pressure system that extends over the large region 
of which India is about the centre, while for another 
series it is dominated by the conditions of atmospheric 
pressure experienced in the region of which South 
America is the middle portion. On this account the 
cycle of 3-8 years distinctly exhibited in the meteor- 
ological records of two great regions of the world 
appears in British meteorology at a period of about 
three years, in the course of which variations are 
sometimes very noticeable. An examination of the 
records of annual rainfall at Greenwich for sixty years 
shows this three-year cycle very clearly. Taking a 
series of years, it is found that two wet years are 
followed by a dry year; but after eight years a re- 
versal takes place, two dry years being followed by a 
wet one. This peculiar result is due to the com- 
bination, in the British Isles, of the Indian and South 
American pressure systems, which have a definite 
period of change—one going up while the other goes 
down—in a period of about 3:8 years. 
Passing now to solar conditions, we have, first of 
all, the well-known sun-spot period of about eleven 
years. This is the mean length of the period; and it 
must be remembered that epochs of maximum 
activity, as indicated by frequency and magnitude of 
sun-spots, do not follow those of minimum activity at 
constant intervals, but vary from about three to five 
years. Moreover, when a critical examination is 
made of solar phenomena, it is found that successive 
cycles differ from one another in certain respects. 
Dr. Lockyer discovered that underlying the ordinary 
sun-spot period there is another of greater length, 
namely, about thirty-five years. In other words, the 
sun has to pass through about three cycles of activity 
before it reaches the same state as it was before. 
Sun-spots must not, of course, be taken as the sole 
criteria of the sun’s condition; and it was pointed out 
by the Lockyers in 1900 that ‘there seems. little 
doubt that in the future the measure of the change 
in the amount of solar energy will be determined by 
the amount and locus of the prominence area ”’ (Proc. 
Roy. Soc., Ixvii.). The solar latitudes in which sun- 
spots most frequently appear vary with the epoch of 
the sun-spot period, and the eruptive prominences 
have their maxima in the same latitude as the spots. 
A detailed examination of the records connected with 
solar spots and prominences revealed subsidiary 
maxima and minima of about 3:8 years (tbid., Ixx., 
x03), so that three waves of solar activity have to be 
considered having approximate periods of four, 
eleven, and thirty-five years. The condition of the sun 
at any time represents the algebraic sum of these 
coefficients of activity. 
The correlation of the factors of terrestrial weather 
now becomes possible. In the first place, the old view 
that the sun’s influence upon weather must be general 
“ 
