306 
rise at somewhat below 30 km, and at 50 km is still 
rising. Above 42 km, however, the rate of rise falls off. 
The temperatures are very much higher than those 
given by Cox and by rocket soundings at White Sands. 
Penndorf [15] does not assume radiative equilibrium 
and considers only ozone heating and cooling. He con- 
siders the effect of water vapor in the ozonosphere to 
be negligible. Above 3000 A the solar radiation is taken 
from Abbot’s table; from 3000 A to 2000 A the radiation 
is assumed to be that of a black body at 5910K. Rates 
of heating and cooling in the ozonosphere are computed 
for different temperature distributions. Penndorf’s re- 
sults on daytime heating and nighttime cooling corre- 
spond to rather high equilibrium temperatures during 
the day. with little change during the night. According 
to Penndorf the maximum rate of heating and most of 
the ultraviolet absorption both occur near 50 km. 
Gowan’s and Penndorf’s calculations agree quali- 
tatively with sound propagation and rocket results in 
showing a temperature rise in the upper ozonosphere. 
The computed ozone heating, however, 1s much greater 
than Cox’s measurements and rocket findings would 
indicate. 
At present, calculations of ozone heating involve so 
many uncertainties that conclusions based upon them 
must be regarded with considerable caution. It appears 
highly desirable to perform a number of rocket experi- 
ments to measure the total energy absorbed at various 
atmospheric levels from the solar and terrestrial radia- 
tions. Such data would provide a firmer foundation for 
temperature calculations than is now available. 
Free Oscillations in the Atmosphere 
Observed amplitudes of solar tides produced in the 
earth’s atmosphere are about one hundred times as 
great as one normally would expect. Kelvin suggested 
that this might be a resonance phenomenon, due to a 
free atmospheric oscillation: period of 12 hr. Evidence 
also is available to show the existence of a second free 
period in the atmosphere. The speed of large explosion 
waves, such as those generated by the Krakatau erup- 
tion and the Great Siberian Meteor, corresponds to a 
period of 10.5 hr. 
Tt is to be expected that the natural periods of the 
atmosphere are directly associated with its tempera- 
tures. Pekeris has, in fact, shown that an atmosphere 
with the temperature distribution set forth in curve 7 
of Fig. 8 possesses both the 10.5- and the 12-hr periods 
[14]. His results are further substantiated and clarified 
by later work of Weekes and Wilkes [18]. The 10.5-hr 
period is a property of the lower atmosphere with the 
temperature distribution revealed by balloon measure- 
ments. The 12-hr period, on the other hand, is asso- 
ciated with the upper atmosphere, and requires the 
indicated temperature drop above the hot ozonosphere. 
Pekeris’ conclusions, based upon analysis of a number 
of special cases, were that the atmosphere must become 
cold again at or above 80 km. 
Meteors and Atmospheric Temperatures 
A meteor striking the earth’s atmosphere is heated 
by impact with the molecules of air. Depending upon 
THE UPPER ATMOSPHERE 
the speed, size, and material of the meteor, the heating 
causes Incandescence along some portion of the meteor 
path. The altitudes at which incandescence begins and 
ends depend also upon the density of the atmosphere. 
In a pioneering work on meteor studies, Lindemann 
and Dobson [12] developed a theory which enabled 
them to calculate atmospheric densities at the heights 
of appearance and disappearance of a meteor from its 
speed, size, and composition. The theory: applies to an 
isothermal atmosphere, whereas it now appears that 
the region in which meteors are observed visually is 
definitely not isothermal. Nevertheless, the theory can 
provide a curve of densities from which an average 
temperature can be estimated. The densities obtained 
by Lindemann and Dobson were not consistent with 
an average temperature of 220K in the neighborhood of 
stratospheric values, but were consistent with a much 
higher temperature, about 300K. 
More recently F. L. Whipple [20] presented a theory 
relating the mass, shape, composition, speed, decelera- 
tion, and luminosity of a meteor to atmospheric 
densities. From a study of available photographic 
meteor data he obtained a curve of log-density versus 
altitude, with which a number of temperature dis- 
tributions can reasonably be associated. In his paper 
Whipple states: 
The best solution appears to be one in which the height-log 
p curve corresponds to a flat temperature maximum of about 
375K near the 60-km level, a rapid drop to 250K near 80 
km, and a constant or slowly rising temperature at greater 
heights to about 110 km. 
Rocket Measurements in the Upper Atmosphere 
Since the spring of 1946 rockets have been employed 
at White Sands, New Mexico, for measuring pressures 
and temperatures in the upper atmosphere [1, 2, 10]. 
Initially the German V-2 was the only large rocket 
available, but at present the V-2 and the Navy’s Aero- 
bee and Viking are all being used. So far, pressure 
measurements extend only to 130 km, although some of 
the rockets have exceeded 100 miles in altitude. Instead 
of being measured directly, atmospheric temperatures 
are computed from the pressure data. 
On each rocket there is an area, usually just ahead of 
the tail fins, where ambient pressures exist during 
flight. Gages are mounte in this area to measure 
atmospheric pressures directly. As a check on the ac- 
curacy of readings from gages so mounted, it is cus- 
tomary to compare measurements at low altitudes with 
balloon data acquired just before and just after the 
rocket flight. In the past there has been excellent agree- 
ment between rocket and balloon measurements. 
Gages are also mounted at various points on the 
nose of the rocket. Measurements from such gages 
must be converted to ambient pressures by some method 
such as application of the Taylor-Maccoll theory. Stag- 
nation pressures, recorded with gages at the nose tip, 
provide essentially a measure of density. 
Temperatures are computed from the pressure versus 
altitude curve using the relation 
ee ALG 
p dh RT’? 
