338 
The half-thickness of the Fj-region ranges roughly be- 
tween about 10 and 30 km, depending largely upon 
time of day, season, and solar activity. Very occasionally 
during disturbance a well-formed Fj-region appears at 
night. This latter phenomenon is not understood. The 
recombination coefficient, between electrons and posi- 
tive ions is of the order of 1 X 10~ ce sec for the 
F\-region. 
The F,-Region 
The F.-region of the ionosphere has been extensively 
studied because its variations present peculiarities not 
immediately evident in the lower regions. When the 
F,-region is mentioned it is customary to think of the 
general F-region at night and the F.-region in the day- 
time as being of the same character. During intervals 
when the F>- and F-regions are not separately identi- 
fiable the region unresolved is thought of as the general 
F-region and as having F,-region characteristics pre- 
dominantly. 
An outstanding oddity of the F>-region (or the gen- 
eral F-region) is that maximum electron density occurs 
everywhere in the Northern Hemisphere in midwinter 
when the solar altitude at noon is much lower than in 
midsummer. From this behavior it is evident that 
F.-region maximum electron concentration does not 
follow solar altitude changes in the same manner as do 
the H- and F\-regions. Berkner and Wells, and later 
Seaton and Berkner, demonstrated the existence of a 
nonseasonal component in the F»,-region electron con- 
centration which is world-wide in appearance and may 
be caused by the annual change in distance between 
earth and sun. Maximum free-electron density in the 
F.-region is quite sensitive to solar activity, varying 
some 300 per cent during the sunspot cycle, but this 
nonseasonal variation persists throughout. 
Frequently, depending upon location, season, and 
solar activity, the maximum concentration of electrons 
does not occur near noon. While there is always an after- 
noon maximum, the time of its occurrence varies from 
about an hour to several hours after midday and there 
frequently is a pre-noon maximum. The effect of such 
variations is often to create a minimum at noon. This 
noon decrease in F>-region electron population was orig- 
inally thought to be caused by heating and consequent 
expansion of the region. Later it was suggested that such 
a proposal was untenable when Northern and Southern 
Hemisphere data were examined over the same time 
interval. In more recent years a different approach to 
the description of the ionosphere has arisen in which 
electron concentration is replaced by total electron 
content in the region. This concept is particularly im- 
portant for the F.-region. When F>-region behavior is 
examined by this newer measure, much if not all of the 
peculiar bimodal behavior is removed, leaving a single 
maximum occurring after noon. It has been demon- 
strated that the F,-region may grow thinner over an 
interval of time, leaving the electron concentration un- 
changed but reducing the total electron content. In 
other words, maximum concentration is not necessarily 
a correct indication of the behavior of this region, and 
THE UPPER ATMOSPHERE 
the use of concentration alone may lead to false con- 
cepts especially when rates of electron production, re- 
combination processes, and other factors are sought. 
Early unexplained asymmetry of the F,-region has 
to some extent been clarified through consideration of 
data in terms of geomagnetic coordinates. It is in this 
layer that strongest evidence exists pointing towards 
ionizing agents other than ultraviolet light from the 
sun. All characteristics of the F>-region display wide 
variations when compared with those of the lower layers. 
One of the important methods for study of the iono- 
sphere is examination of electron-concentration varia- 
tions during eclipse of the sun. The E- and Fj-regions 
show behaviors in good agreement with theory, but 
eclipse effects in the F2-region have been difficult to 
detect. One reason for difficulty in isolating such effects 
in the F-region is of course the small recombination 
coefficient between electrons and positive ions, which 
is of the order of 1 X 101° to 1 X 10-" ce sec at 
these great heights. Another reason for difficulty is the 
immense variability which is always evident in the 
electron density of the F,-region. Some investigators 
have pointed out that if an important part of the F»- 
region ionization is caused by corpuscular bombard- 
ment, eclipse of the ultraviolet light would not coin- 
cide with eclipse of the particle stream. Corpuscular 
eclipse has been sought without success. There is some 
evidence, although not wholly satisfactory, of ultra- 
violet light eclipse in the F,-region. The semithickness 
of this region ranges between about 30 and 100 km. 
Free-electron densities in the F-region range between 
about 2 X 104 electrons per cc for winter night and 
3 X 10° electrons per cc for daytime during maximum 
solar activity. The free-electron population in the F.- 
region is probably the result of photo-ionization by 
ultraviolet light of the nitrogen molecules. [onization 
of atomic nitrogen and atomic oxygen may also con- 
tribute during the middle of the day. 
The G-Region 
The highest ionospheric region for which any evi- 
dence exists is the G-region (see Fig. 3). This region may 
be present in reality. Occasionally data of the type 
shown in Fig. 2 develop an additional retardation be- 
yond the F.-region penetration frequency. In some rec- 
ords where additional reflections appear, the character 
of the data suggests Rayleigh scattering from patches 
of ionization of the order of a wave length or less in 
diameter. On other occasions, however, the data are 
so suggestive of another higher region that credence 
must be given this possibility. There is certainly no 
a priori reason why such a G-region should not be 
present. It may always be present but, because of rarity 
of the atmosphere in the vicinity of 400 km, may only 
develop free-electron concentrations greater than those 
in the F,-region during unusual conditions. So little 
is known of the G-region that it can only be said of the 
free-electron concentration there that it is almost always 
less than that of the F.-region. Height of maximum con- 
centration is probably of the order of 400 to 500 km. 
