gion, a test can be made of Hoyle’s hypothesis* that 
temperature lapse rate is of greater significance than 
ia now believed. If, for example, the field strength 
during one-way transmission reaches the value cal- 
culated for a flat earth while the modified refractive 
index lapse rate is still positive, then something must 
be wrong with either the modified index concept or 
the method of calculating refractive index from mete- 
orological data. 
When the modified index lapse rate is relatively 
constant with altitude, a fairly simple transformation 
makes the atmosphere nonrefracting and the effective 
earth radius greater or smaller than the actual radius, 
and ray tracing should then be valid in the interfer- 
encé field. If no rays reach the receiver, the diffracted 
field must supply all the energy received. 
The propagation path extended from Datelan to 
Gila Bend, Arizona, a distance of 47 miles over desert 
terrain. A 3,200-mc transmitter was located on a tower 
at a height of 53 ft above ground at Datelan, with the 
receiver 35 ft above the ground in the control tower 
at the Gila Bend airfield. There is a gentle rise of 
ground from Datelan to Gila Bend with a total rise 
in elevation of 402 ft, or about 8.5 ft per mile. The 
intervening terrain is remarkably uniform, without 
trees or large irregularities, and there are no build- 
ings except in the immediate vicinity of the trans- 
mitter and receiver locations. 
Figure 1 shows the diurnal variation of surface air 
temperature at Ajo, Gila Bend, and Phoenix for 
December 16 and 17, 1944. These typical data show 
the uniformity of conditions over that region and 
the effect of radiation cooling on the air mass near 
the surface. 
Figure 2 shows the variation of the soil tempera- 
ture with time at Datelan for the same period, as well 
as temperature changes at 25, 50, 100, and 500 ft 
above the earth. 
The general topography around Gila Bend in con- 
Junction with the diurnal variation in the prevailing 
surface wind vector shows an interesting condition. 
Hourly wind vector observations during November 
‘and December 1944 showed that by 1900 the prevail- 
‘ing wind was downslope toward the lower elevations. 
This flow of cold air into the area of the link may be 
tesponsible for the overall cooling of the air up to 
several hundred feet during the night. At present it 
is not clear how much of this effect should be attrib- 
uted to radiation and eddy diffusion of heat toward 
the earth, although on nights with wind speeds from 
calm to a gentle variable breeze it is difficult to at- 
tribute the entire transport of heat to the latter 
processes. 
Some pertinent data are tabulated in Table 1 show- 
ing the time at which the signal was first detected and 
completely lost and the general atmospheric and 
ground conditions nearest these times. On the after- 
noon of December 14, the sky was overcast, and the 
signal was detected about an hour earlier than on the 
other evenings. 
Figure 3 shows the field strength data for a typical 
day plotted in decibels below free space. The maxi- 
mum and minimum for half-hour intervals are shown 
so that the fading range is apparent. The meteorologi- 
cal data for the period are given in the form of modi- 
fied refractive index curves relative to a fictitious earth 
tadius of 4a/3, the time of the sounding being given 
on each curve. 
The diurnal variation in field strength is quite pro- 
nounced and regular. The maximum range of fields 
Measured was around 46 db, the maximum field gen- 
erally occurring at times when the inversion layer was 
thickest. There is no significant correlation between 
strong fields and the amount that M decreases at some 
elevation above the antennas. In fact, at times such 
as 0900 on December 17 the field strength is quite 
high and yet M shows little indication of trapping. 
In most cases strong fields occur at times when the 
4a/3 modification of the index of refraction gradient 
is near zero or varying slowly with altitude. 
APPENDIX 
The general results of this experiment may be sum- 
marized in the following way. Over the desert loca- 
tion a ground-based temperature inversion was found 
each pight due to radiation cooling of the under- 
lying surface. This temperature inversion produced a 
strong index of refraction gradient in the first few 
hundred feet above the earth. 
The 10-cm nonoptical link showed a marked diur- 
nal variation in field strength in close correlation 
with the building up and intensification of the tem- 
perature inversion. The strongest fields generally ac- 
companied modified index gradients approaching zero 
in the first few hundred feet above the earth’s surface. 
The trapping criterion most widely accepted here- 
tofore specifies that, at some elevation above the 
transmitter and receiver antennas, M should be less 
than at the antennas. The data herein reported seem 
to indicate that this criterion is neither necessary nor 
sufficient to insure strong ‘fields below the optical 
horizon. The strongest fields observed at 10 cm ap- 
proached the flat earth value, assuming a reflection 
value of unity for the earth. 
ANTIGUA, WEST INDIES 
PROPAGATION IN S AND X BANDS 
IN LOW-LEVEL OCEAN DUCTS 
General Description" 
ad [be EXISTENCE of low-lying ducts over the seas of 
the world, particularly in the trade wind belt, has 
been known for the past 2 years. Measurements made 
by the British and by Washington State College and 
the Naval Research Laboratory have consistently in- 
dicated the presence of ducts ranging in thickness 
from 20 to 50 ft in regions where the trade wind 
followed a long over-water trajectory. These ducts 
are known to vary in intensity and thickness with 
wind velocity during the trade wind season. It was 
considered advisable to investigate the possibility 
that such ducts would permit greatly extended ranges 
‘on surface craft and very low-flying aircraft by prop- 
erly sited radar installations. 
Diseussion by representatives of the Chief of Naval 
Operations, NDRC, and the Naval Research Labora- 
tory resulted in organization of a project to make an 
experimental investigation of meteorological and prop- 
agational conditions in an area of the Caribbean 
theater where such ducts are persistent, with a view 
to determining their operational usefulness. It was 
decided that a one-way ship-to-shore transmission 
path over water would provide the most direct data 
for analysis, and such a system was set up, using 
transmitting and receiving equipment provided by 
the Radiation Laboratory. The transmitters were in- 
stalled in a patrol craft assigned for the project, there 
being no larger vessel available, with transmitting an- 
tenna heights of 16 and 46 ft. 
The site chosen for the receivers at the land-based 
end of the link was at Judge Bay on the island of 
Antigua in the Leeward Island group of the British 
West Indies. Antennas were installed on a tower 50 
ft from the water’s edge, at heights of 14, 24, 54, and 
94 ft, for both S- and X-band receivers, 
Antennas for both S- and X-band transmitters were 
installed on the patrol craft at heights of 16 and 46 
ft. These consisted of parabolic reflectors arranged 
to permit transmission forward or astern, su that 
transmissions could be made on both the outward 
and inward legs of the runs. The S-band transmitter 
peak power output was 42 kw, and its antenna pro- 
vided a measured gain of 27 db. Output on X band 
‘was 31 kw, the antenna providing a measured gain 
of 29 db. Later in the experiment an S-band antenna 
was installed at a height above the water of 8 ft. 
Tests were made with this antenna on two suns. Ad- 
ey Ut. R. W. Bauchmam J. 8. Naval Research Labora- 
ry. 
487 
equate switching arrangements to permit tests with 
the different antennas were provided, and power out- 
puts were measured by means of directional couplers 
and thermistor bridges. 
Meteorological measurements from the ship con- 
sisted of detailed temperature and relative humidity 
readings taken on a rigging running from a boom 
extending out over the water amidship to the yard- 
arm about 46 ft above the water. Low-level sounding 
equipment of Washington State College design was 
used for all meteorological measurements. Balloon 
ascents to heights of 600 ft from the stern of the ship 
were also made when conditions permitted. Hourly 
observations of sea temperature, wind, and sling psy- 
chrometer readings from the bridge were made. It was 
impossible to obtain satisfactory soundings on the 
rigging or by use of balloons and kites when running 
away from the tower into the wind because of the 
large amount of water taken over the bow and the 
resulting salt spray. Shipboard observations during 
outward runs were therefore confined to the hourly 
wind velocity, sea temperature, and sling psychrom- 
eter readings. On return runs with the wind, bal- 
loon and rigging soundings were made. It was neces- 
sary to estimate the height above the surface for 
readings taken below 10 ft because of the severe 
pitching and rolling motion of this type of ship, and 
therefore very few such readings were made. 
At the receiving end of the radio path, the antennas 
for S band were 48-in. parabolic dishes with a gain 
of about 30 db. The X-band antennas were 48-in. 
dishes cut to 2 ft in the horizontal dimension to 
broaden the horizontal acceptance angle. This was 
done to eliminate the effects of minor deviations of 
the ship from a radial course. These antennas had a 
measured gain of 35 db. Midway in the experiment 
an X-band antenna was mounted at the base of the 
tower at a height of 6 ft, since results up to that 
time indicated the lowest available antenna height 
on X band gave the strongest signals. All antennas 
were mounted on swivels to allow alignment on any 
course over a 40-degree arc and were connected by wave 
guide and stub-supported coaxial cable. 
Two S-band and two X-band receivers feeding 
Esterline-Angus recording milliammeters were kind- 
ly furnished by the Radiation Laboratory. The S- 
band receivers had a minimum sensitivity of 110 db 
below 1 w, while the X-band receivers had a mini- 
mum sensitivity of 105 db below 1 w. It was necessary 
to use automatic frequency control on the X-band 
receivers, but manual tuning was employed on the 
S-band receivers because of the greater frequency 
stability of the S-band magnetron. The receivers were 
calibrated with standard test sets before every run 
and checked upon ‘the completion of each test. Indi- 
vidual calibration curves were then used in plotting 
the results of each run. Since only two receivers on 
each band were available, an r-f switching arrange- 
ment similar to that used on the ship was employed. 
Two-way voice communication between the ship 
and shore station was maintained at all times for 
coordination of operations. The facilities of an Army 
radio direction-finding station on the island were 
available to obtain bearings on the ship. 
Meteorological measurements were made at the 
shore station during operations by means of kite 
flights and a guy rigging running from the water’s 
edge to 10 ft above the top of the tower. Detailed 
soundings in the first 100 ft were then taken by slid- 
ing the measuring instruments up and down. the 
rigging. Since the duct conditions important in this 
investigation were always below 100 ft, only occa- 
sional kite soundings (two to three a day) were made 
to check the higher levels. Most of the data accumu- 
lated were taken on the tower rigging where detailed 
soundings could be made. Wind speeds at the surface 
and 100 ft levels were recorded hourly. Hygrother- 
mographs were placed at the antenna levels and con- 
tinuous records taken to determine the diurnal varia- 
tion of temperature and relative humidity, if any. 
