380 
tenuous envelope of gas called the corona. Recent evi- 
dence shows that the corona is extremely hot, having a 
temperature of about 1,000,000K compared with the 
photospheric temperature of about 6000K. The cause 
of this high temperature is still a mystery. 
All of the solar phenomena described above vary 
with time in intensity or character. The variations of all 
the features seem to be in phase. Only the sunspots 
have been observed for a long enough period of time to 
include a large number of cycles of variation. Accord- 
ingly, sunspot variation is discussed here as an indicator 
of solar variation. 
The relative sunspot number is an index of the number 
of spots and groups of spots visible on the solar surface. 
Reasonably reliable records of sunspot numbers extend 
back to 1749, although the earlier years of this record 
are much less reliable than the later ones. These records 
show that the sunspot number has alternately increased 
and decreased with a period that has averaged about 
eleven years. The length of this period, however, has 
varied from seven to seventeen years and the sunspot 
number has also varied from maximum to maximum. 
The beginning of a solar cycle is characterized by the 
appearance of a few spots near 30°N and § heliographic 
latitudes. As the cycle progresses, the number of spots 
increases and the latitudes of most frequent occurrence 
move toward the solar equator. The largest number of 
spots occurs when the spot zones are near 16°N or 8S. 
After sunspot maximum, the spot zones proceed equa- 
torward and at sunspot minimum there are only a few 
spots near the equator and a few spots in higher 
latitudes marking the beginning of a new cycle. 
Sunspots possess intense magnetic fields. The po- 
larities of these fields vary characteristically, with a 
period twice that of the ordinary sunspot cycle. The 
two large spots in a group, which are called the leader 
and follower spots, have opposite polarities. The leader 
spots in the Northern and Southern Hemispheres also 
have opposite polarities. Moreover, spots in the same 
relative position in their groups and in the same hemi- 
sphere have opposite polarities from one sunspot cycle 
to the next. 
Along with sunspot variations, other solar phenomena 
show characteristic changes. Faculae, flocculi, flares, 
and prominences are more numerous and intense at 
sunspot maximum, although they are by no means 
absent at sunspot minimum. The corona is approxi- 
mately circular at the time of sunspot maximum, but is 
flattened at the poles near sunspot minimum. The 
intensity of the coronal emission lines is also greater at 
sunspot maximum than at minimum, according to 
recent observations. 
The energy output of the sun is difficult to measure 
at the earth’s surface, because of absorption and scatter- 
ing of sunlight by the earth’s atmosphere. Nevertheless, 
the Smithsonian Institution has for the past forty years 
undertaken the routine measurement of the solar con- 
stant. This is the amount of solar energy incident at the 
outer edge of the earth’s atmosphere per unit area and 
time reduced to mean solar distance. These measure- 
ments, of course. do not include solar energy to the 
COSMICAL METEOROLOGY 
violet of about 2900 A, which is completely absorbed in 
the earth’s upper atmosphere. This energy is estimated 
and included in the solar constant. 
The value of the solar constant is near 1.94 cal em~ 
min-!. C. G. Abbot, who has supervised most of the 
measurements, claims that the solar constant varies 
from time to time by a few per cent of the mean value. 
This claim has been disputed by many other scientists, 
who feel that the uncertainties involved in the measure- 
ments are at least as great as the suspected range of 
variation. In any case, Abbot’s work gives an upper 
limit to the amount of solar variability during the past 
forty years in the part of the spectrum available for 
routine measurement. The measuring program of the 
Smithsonian Institution is described fully in the Annals 
of the Smithsonian Institution and in numerous papers 
of the Smithsonian Miscellaneous Collections. 
Abbot has claimed that the variations in the solar 
constant primarily reflect large variations in the violet 
and ultraviolet part of the spectrum. This claim seems 
to be verified by some measurements of Pettit [16]. 
Pettit measured the ratio of solar radiation at 3200 A 
to that at 5000 A. Over a period of years from 1924 to 
1931, this measured ratio varied by about 40 per cent 
of its mean value, presumably reflecting variability in 
the ultraviolet. Unfortunately, observing conditions 
were not ideal and atmospheric factors were suspected 
to be the cause of at least part of this variation. On the 
other hand, none of the specific tests applied by Pettit 
showed instrumental or atmospheric effects of this mag- 
nitude. To some extent, at least, the ratio varied m a 
manner similar to that of the sunspot number. 
Farther into the ultraviolet, below about 2900 A, 
solar energy is absorbed in the upper atmosphere and 
cannot be observed at the earth’s surface. In the ultra- 
violet, however, there is abundant indirect evidence of 
solar variability because of effects on the upper atmos- 
phere. The frequency of occurrence of magnetic storms 
and of aurorae, as well as the ionization of the upper 
atmosphere, clearly varies over the solar cycle. The 
correlation between upper-atmospheric phenomena and 
sunspot number becomes progressively poorer as the 
parameters involved are averaged over shorter and 
shorter periods of time. Thus, there is little or no corre- 
lation between daily sunspot numbers and daily values 
of the various indices of upper-atmospheric conditions. 
This would make it appear that all spots do not affect 
the earth directly but rather are correlated with the 
occurrence of the solar phenomena that do affect the 
atmosphere. Effects in the upper atmosphere indicate 
clearly solar variations only in the short wave lengths 
of hydrogen and helium emission (\ < 1216 A) and vari- 
ations in particle emission of the sun. For the wave- 
length interval 1216-3200 A, there is apparently no 
evidence to indicate whether the sun’s energy is or is 
not variable. 
The present state of knowledge of solar variability, 
then, reveals little or no variability in the infrared and 
visible spectrum, perhaps some moderate variability in 
the violet and near ultraviolet, and almost certainly 
rather large variability in the far ultraviolet. The solar 
