308 
beeause of inaccuracies introduced in differentiating 
the pressure curves to obtain temperatures. Havens 
estimates that the pressure measurements permit the 
determination of an average temperature for a layer 20 
km thick correct to within +10K. The error for a 10- 
km layer would be on the order of 20K. The curves 
of actual temperature variation with altitude must, 
therefore, be regarded as only indicative. 
The double hump on curve (5), Fig. 8, appears to be 
real. Its presence was suspected in the reduction of 
data on other flights, and strongly suggests two distinet 
heating processes. The lower hump is due to ozone 
absorption in the ultraviolet above 2000 A. The upper 
hump may be due to oxygen absorption in the ultra- 
violet below 2000 A. 
125 
100 
ALTITUDE (KM) 
25 
300 
TEMPERATURE (°K) 
Fig. 8—Temperatures calculated from pressures measured 
150 200 350 
on rocket flights above White Sands Proving Ground, 
New Mexico. The points indicated by x with their probable 
errors are temperatures calculated from the ratio of stagna- 
tion pressure at the nose of the rocket to ambient pressure. 
Sea-level composition is assumed throughout. The curves in 
the drawing are: (1) 7 March 1947, 1123 MST; (2) 22 January 
1948, 1318 MST; (3) 5 August 1948, 1887 MST; (4) 28 January 
1949, 1020 MST; and (5) 3 May 1949, 0914 MST (taken from 
{10]). Also shown are: (6) Balloon flight shown in Fig. 1; and 
(7) Temperature distribution assumed by Pekeris for calcula- 
tion of atmospheric oscillations (taken from [14] Pekeris: Proc. 
roy. Soc.,(A) 158: 658 (1937), by permission of The Royal 
Society). 
The increase above 80 km indicated by the dotted 
extension of curve (1), Fig. 8, was drawn so as to have 
the average value of about 260K between 80 and 120 
km. This average value, indicated at the 100-km level, 
is the sole calculated pomt on which the dotted portion 
of the curve is based. The heating of this atmospheric 
level is possibly due to absorption of solar X-radiation 
such as that recently discovered in V-2 flights by T. R. 
Burnight. 
Temperatures and Pressures at Extreme Altitudes 
Rocket measurements [1, 2, 10] and Cox’s sound 
propagation studies [5] indicate a continually increasing 
THE UPPER ATMOSPHERE 
temperature above the minimum near 80 km. The 
former measurements extend to about 130 km, and the 
latter to about 172 km, although Cox states that he 
places little faith m the highest point on his curve. 
According to Whipple a positive temperature gradient 
above 80 km is consistent with data from meteor 
studies. From the energy distribution in the nitrogen 
bands of the auroral spectrum Rosseland and Steens- 
holt [17] report a temperature of 347K (at 110 km 
approximately), correcting a much lower value obtained 
earlier by Vegard. 
It appears reasonable to assume a general tempera- 
ture increase to great heights, with a corresponding 
lessening in the rate of decrease in pressure. For one 
thing, if the temperature were essentially constant 
above 100 km, the atmosphere at 400 km would be 
practically nonexistent, as a simple calculation will 
show, whereas observations of aurorae indicate appre- 
ciable densities at these heights. Secondly, helium ap- 
pears to be escaping from the earth’s atmosphere at a 
rate which requires temperatures on the order of 1000K 
between 600 and 800 km [18, pp. 21-24]. Thirdly, 
Babcock’s measurements of the width of the night-sky 
line 5577 A indicate temperatures on the order of 1200K 
at several hundred km [13, p. 509]. 
In comparison with information available on the at- 
mosphere below 100 km, data on temperatures and 
pressures at higher altitudes are meagre indeed. To be 
sure there is an appreciable amount of spectroscopic 
data; and ionospheric studies also indicate high tem- 
peratures in the ionospheric layers. Nevertheless, at 
present, conditions in the outer reaches of the earth’s 
atmosphere are largely a matter of speculation. 
Conclusion 
Even a casual survey of the literature shows wide 
differences in the temperatures and pressures ascribed 
to the upper atmosphere. This is apparent in the limited 
selection of material presented above. Some of the 
differences can be traced to diurnal and seasonal effects 
and to geographical differences. Also atmospheric con- 
ditions are closely associated with conditions im the sun, 
which fluctuate continually. On the other hand, most 
of the methods of determining atmospheric tempera- 
tures and pressures cannot be proved free of systematic 
errors. The rather involved ozone calculations of Gowan 
and Penndorf, for example, and Whipple’s meteor 
theory, contain many uncertainties which cannot be 
resolved without additional information. The most 
direct measurements are those made by balloons and 
rockets. 
It would be worth while to compare critically the 
various methods of determining pressures and tempera- 
tures in the upper atmosphere, at least to pomt up 
systematic differences in the results. At present, accom- 
plishment of such a comparison is hindered by lack of 
any single standard. To provide such a standard, data 
should be available from a single credible method cover- 
ing different times and seasons and a wide range of 
geographical and solar conditions. 
Because of their on-the-spot character, rocket meas- 
