EXTENDED-RANGE WEATHER FORECASTING 
dients as well as the temperature difference between 
the eastern and western parts of the circulation system 
are decisive for the mean pressure changes that are 
observed over the Azores and Iceland during periods of 
approximately a month’s duration. Thus, from these 
pressure gradients and temperature contrasts, at least 
for the spring, the change of the pressure difference 
between the Azores and Iceland can be computed from 
one month to the next with an error of approximately 
six-tenths of the standard deviation. These correlations 
are presented in Table II, in which the correlation 
coefficients that exceed the upper limit of chance are 
italicized. 
Cosmic Influences on Grosswetter. Unimportance of 
Moon and Planets. No one has ever proved in a single 
case to date that durmg or after a given arrangement 
of the moon or the planets any weather phenomenon 
occurred more frequently or less frequently than would 
have been expected from chance. On the contrary, proof 
has been established from extensive data and by means 
of statistical criteria [14; 17, pp. 93-96] that certain 
819 
tropics the relationship is generally quite poor. There 
is a better-than-chance correlation only between the 
fluctuations of the level of Lake Victoria in East Africa 
and sunspots, but even this correlation is not persistent. 
The correlation coefficients between the annual means 
of relative sunspot numbers and the annual means of 
the water level of Lake Victoria (mean of maximum 
and minimum readings at Kisumu according to World 
Weather Records) are: 
+ 0.87 for 1899-1924, 
+ 0.07 for 1925-1943, 
+ 0.58 for 1899-1943. 
The utilization of these weak correlations for prog- 
nostic purposes does not seem feasible. However, sev- 
eral investigations [9, 18, 19] have revealed that over 
the whole earth, particularly im both temperate zones, 
a double, sometimes a triple, fluctuation of nearly all 
meteorological elements occurs within a sunspot cycle. 
The world-wide distribution of this fluctuation is shown 
by the curves of Fig. 2. The bottom curve represents 
Tasie II. CoRRELATION COEFFICIENTS OF ZONAL PRESSURE AND TEMPERATURE DIFFERENCES with MontHiy CHANGES OF 
NortH Arrantic MrripionaL PRESSURE DiIrrERENCE* (1874-1923) 
Station pairs Jan. Feb. 
Mar. 
Apr. May June July Sept. Oct. Nov. Dec. 
Rome—Ponta Delgada 
Indianapolis—Ponta 
+0.36)+-0. 48|+-0.37|+-0.63|+0.59|+-0. 46|--0.35|-++0.32)-+-0.16]+-0. 48|+-0.21)+-0.27 
+0.42|+-0.62|-++0.66|+-0.56|+-0.41|+-0.44\+-0. 48|--0. 29|+-0. 23/-+-0. 46|+-0.41|-+0.34 
—0.39|—0.55|—0.55|—0.58|—0.26|—0.59|—0.58|—0.19|—0.30|—0. 43)—0.35)—0.16 
—0.32|—0.45|—0.47|—0. 47|—0.06|—0.36|—0. 23/—0.31|—0.50|—0.39|—0. 44|—0.36 
Fes, | Det 
Haparanda—Stykkisholm 
Jacobshavn—Stykkisholm 
Temperature | Tromsé—Western Green- 
difference land 
—0.16|—0.42|—0. 44)—0.41 
0.21/—0.69)/—0.16 
0.43|—0.63)—0.51|—0.39|—0.34 
* The pressure and temperature differences for the station pairs at the left for the indicated month are correlated with changes 
in pressure difference between Ponta Delgada and Iceland from the indicated month to the next; correlation coefficients that 
exceed the scope of chance are italicized. 
coincidences claimed by “‘astro-meteorologists” lie well 
within the range of chance. 
The influence of the moon upon the atmosphere is 
restricted to producing atmospheric tides, which repre- 
sent diurnal pressure variations that amount to only 
44000 of the changes associated with weather develop- 
ments. Likewise, there is no notable influence of the 
planets on terrestrial weather. The hypothesis by See, 
according to which the planets are supposed to influ- 
ence the sun’s activity by their power of attraction 
upon a hypothetical stream of matter directed toward 
the sun, is no longer tenable. According to Waldmeier 
[55], each sunspot cycle, from one sunspot minimum 
to the next, represents a separate phenomenon that 
can be caused only by processes in the sun’s interior 
(eruption hypothesis). 
Relationships between Grosswetter and Sunspot Cycles. 
K6ppen was the first to show the existence of an average 
period of eleven years in the air temperature. However, 
according to more recent investigations [3; 5; 17, pp. 
97-105], no significant parallelism or antiparallelism 
of any one meteorological element to the variation of 
sunspots exists in the temperate zone, and even in the 
the dependence of the departure of the mean annual 
temperatures in Apia, Samoa, and Colombo, Ceylon, 
on the year’s position in the sunspot cycle. This curve 
shows that a single oscillation prevails in the tropics, 
but that even here a double oscillation is superposed. 
The temperature maximum does not coincide with the 
sunspot minimum, as would be the case with pure anti- 
parallelism, but occurs two years before. Summarizing 
all phenomena, we find that about 2-214 yr before the 
sunspot extremes, the general (zonal) atmospheric cir- 
culation is increased. The subtropical high-pressure 
belts are stronger and displaced poleward in midsum- 
mer, at least in the North Atlantic and European re- 
gion. Moreover, the pressure difference between winter 
and summer over the Asiatic continent is diminished, 
and central Europe has dry midsummers as a rule. 
In years of sunspot extremes (7.e., minima as well 
as maxima) and shortly before and afterward, the 
following phenomena are observed: The planetary cir- 
culation is weakened, the mean pressure in the high- 
pressure belts is lowered, the difference between winter 
and summer pressure in Asia is increased, and the 
