292 RADIO WAVE PROPAGATION EXPERIMENTS 
Attention is here called to the already well-known 
fact that in the optical spectrum region dissymmetry 
appears in the angular distribution of the scattered 
radiation. That is to say, the larger the parameter p 
or the nearer the drop diameter to the wavelength, the 
greater the power scattered in the direction of the 
propagation in comparison with that scattered back- 
ward or at 180° to the direction of propagation. 
The back-scattering cross section or radar cross 
section of water drops is given in the form of a series 
in ascending powers of the parameter p. Table 14 
contains the results of numerical computations of 
these radar cross sections for water drops in the diam- 
eter range 0.05 to 0.55 cm and wavelength range 3 
to 100 cm. 
The radar cross section allows the determination 
of a radar attenuation constant. The radar absorp- 
tion coefficient, or the double of the attenuation con- 
stant, is the fraction of the incident power scattered 
backward by a layer of unit thickness of the echoing 
medium. Table 15 contains the numerical values of 
this radar absorption coefficient in different rains of 
known drop size distribution, and Figure 17 is its 
graphical representation. Table 16 is a somewhat 
modified form of Table 15, in so far as it gives in 
decibels the fraction of the incident power scattered 
backward by a 1-km layer of different rains. The 
theoretically predicted back scattering seems to be 
in fair agreement with the rather few experimental 
data on the power received in radar observations of 
rains or rain clouds. 
In conclusion it may be stated that, in view of the 
scarcity of meteorological data and the irregularities 
inherent in meteorological phenomena, the theory 
provides a satisfactory picture of the propagation of 
microwaves through a variety of precipitation forms 
present in the atmosphere. 
K-BAND ABSORPTION — 
EXPERIMENTAL? 
Our knowledge of the attenuation of K-band radia- 
tion in the normal atmosphere is based upon the 
theory outlined by Van Vleck and upon a number of 
experiments, some of which were undertaken to ob- 
tain data needed in the theory, others of which were 
attempts to measure directly absorption by the at- 
mosphere. 
The width of the rotational lines of water vapor in 
the infrared has recently been measured in work at 
the University of Michigan. The width of the oxygen 
lines responsible for the strong absorption at 0.5 cm 
and the rather small effect at K band are inferred 
from experiments at the Radiation Laboratory. The 
absorption in oxygen was measured directly at sev- 
>By E. M. Purcell, Radiation Laboratory, MIT. 
eral wavelengths in the neighborhood of 0.5 to 0.6 
cm. The gas was contained in a wave guide about 6 m 
long. This guide could be evacuated and then filled 
with gas to any desired pressure between zero and 
roughly 1,000 mm Hg. The radiation was obtained 
as the second harmonic generated in a crystal rectifier 
fed by a K-band oscillator. The source was amplitude 
modulated at audio frequency, and the signal was 
detected by a second crystal at the far end of the 
wave-guide path. The attenuation in the gas was de- 
termined by comparing the signal received with the 
guide evacuated to that received with gas present in 
the guide. The absorption of pure oxygen, at various 
pressures, as well as that of controlled mixtures of 
oxygen and other gases, was measured. The results 
confirm the predictions of the theory in a very con- 
vincing manner and suggest a value of the line width 
lying between 0.05 and 0.02 cm. 
Direct measurements of atmospheric absorption 
at K band have been made by a group at the Radia- 
tion Laboratory using a K-band radar set in an air- 
plane. For this purpose, the set was provided with 
fixed attenuators which could be switched in or out 
of the system. Both r-f and i-f attenuators carefully 
calibrated were used. The experiment consisted in 
flying a straight level course away from a known 
target and determining the maximum range to which 
the target could be seen with and without attenuation 
in the system. The maximum ranges involved were of 
the order of 30 miles. From the results a value for 
the attenuation in the atmosphere can be calculated, 
assuming free space propagation, and this value in 
turn correlated with the meteorological data. The 
latter were obtained from radio-sonde flights at MIT. 
After making allowance for the rather small oxygen 
effect, the results are best represented by a figure of 
0.02 db per nautical mile for 1 g/m®* of water vapor. 
Tn several of the flights the target was an accurately 
made 4-ft corner reflector. This provides an independ- 
ent upper limit to the attenuation, since all system 
parameters (antenna gain, S/N, etc.) were known, and 
one can calculate how far the corner should have been 
seen with any supposed amount of atmospheric atten- 
uation. The upper limit estimated in this way is about 
0.04 db per nautical mile for 1 g/m* of water vapor. 
An entirely different method for measuring attenu- 
ation in the atmosphere has been developed at Radia- 
tion Laboratory. It is possible to measure the apparent 
radiation temperature of any matched r-f load, includ- 
ing an antenna, with great precision (—1C). In the 
case of an antenna, the temperature measured is the 
temperature of whatever the antenna is looking at, 
that is, the temperature of whatever would absorb the 
energy emitted from the antenna if the antenna were 
transmitting. When the antenna is pointed at the sky, 
the temperature measured is some mixture of the tem- 
perature of outer space and the temperature of the air, 
