492 i 
is about 0.8 at the start, about 0.4 in the middle, and 
over 0.9 at the end. It is suggested that reflection at 
small glancing angles from a sharp inversion top pro- 
vides a possible explanation and that the first part 
of the pattern in Figure 9 corresponds to an increase 
in height (the first deep minimum occurring when 
the inversion top is just above the transmitter) and 
the latter part to a decrease in height at a different 
rate. The rates of climb and fall turn out to be about 
200 ft per hour and 300 ft per hour respectively. No 
local soundings are available to check this hypothesis, 
But the calculated rates of change of height are quite 
possible. 
The general meteorological data for the night (illus- 
trated in Figure 9) are as follows: cloudless, follow- 
ing a fine day ; temperature inversion of about 6 F in 
(approximately) the first 500 ft at 0600 GMT ; ground 
mist about dawn. It is clearly desirable to obtain ade- 
quate soundings at periods when this type of effect is 
observed, especially on account of the widely held view 
{hat the index changes which occur at heights of the 
order involved (about 500 ft above ground level) are 
inadequate to account for reflection coefficients of the 
size implied by the pattern observed here. 
Difficulties of Existing Theory 
In this section a few general characteristics of the 
radio observations which appear to be at variance with 
previous theoretical conclusions will be summarized 
1. The most obvious point as far as the Irish Sea 
data are concerned is the failure of the soundings to 
provide an adequate guide to the signal variations. The 
fault may lie in the limited nature of the soundings or 
in the method of interpreting them, but it is clear that 
the problem is by no means as simple as was supposed 
when the soundings were started. 
2. The minimum levels obtained (where they are 
high enough to be measured) usually agree tolerably 
well with the expected values. For the Irish Sea paths, 
as well as in other measurements, the maximum level 
rarely goes above that calculated for flat-earth condi- 
tions; this level is practically the same as that for 
free space conditions for all the Irish Sea centimeter 
links except for paths BF on S band (only) where 
there is a difference of about 10 db. If complete guid- 
ing were a common phenomenon over paths of the 
lengths actually used, it would appear that levels above 
free space should oceur much oftener and more con- 
tinuously than is observed. It appears to be a useful 
working assumption that the level obtained under 
favorable conditions over nonoptical paths is nearly 
that calculated for rays with the same curvature as 
the earth. 
3. The fact that the Irish Sea results show that (for 
centimeter waves) the advantage lies only rarely with 
the smaller heights and shorter wavelengths suggests 
that the importance of complete guiding as a criterion 
for siting stations may have been overemphasized. 
4. The good correlation obtained in a number of 
eases between land-sea temperature difference and 
signal level suggests that the importance of tempera- 
ture may be greater than is indicated by existing 
theory. 
5. It appears very difficult to account for the max- 
imum levels reached on the basis of existing theory. 
6. The interference patterns of the type discussed 
still await an adequate explanation. 
PROPAGATION WORK 
AT THE NATIONAL PHYSICAL 
LABORATORY? 
Analysis and Study of Centimeter and 
Meter Wave Propagation over Sea 
(Irish Sea Experiment) 
This project utilizes the results of radio transmis- 
sions being conducted between two sending stations 
>By W. Ross, British Central Scientific Office. 
APPENDIX. 
in South Wales and receiving stations in North Wales 
and Scotland, jointly by the Admiralty, Ministry of 
Supply and Air Ministry. 
The contribution of the National Physical Labo- 
ratory to the installations being used for this investiga- 
tion’ has been chiefly connected with the monitoring 
equipment used at both sending and receiving stations 
to insure that the radiation from the former and the 
sensitivity of the latter are maintained constant, so 
that any variations on the field strength records are 
known to be due to transmission effects in the atmos- 
phere: The instruments required for the S band are in 
an advanced state of production, while for the X band 
the necessary field strength meter for the transmitters 
has been developed, but some development work is 
still required on the standard radiator for the receiver 
calibration. In accordance with a recent agreement as 
to the limitation of the scope of the investigation, all 
work on instruments for other wavelength bands has 
been put in abeyance. 
Study of Centimeter Wave 
Propagation over Land (Whitwell 
Hatch to Wembley) 
A transmitter operating on a wavelength in the S 
band has been installed at the Admiralty Signal Es- 
tablishment, Whitwell Hatch, and a continuous record- 
ing is being made of the field intensity of the radiation 
received at the Research Laboratories of the General 
Electric Company, Wembley, over a land path of 38 
miles. Except for some houses and trees within about 
5 miles of the receiver, the path is a clear optical one; 
originally the transmitter was also partially obscured 
by some trees, removal of the tops of which produced 
a rise in received field of 10 db. Field strength re- 
cording has been in progress over this link since 
March 1943, and during the intervening 18 months 
there has been a seasonal variation, with the average 
daily value in November and December at least 10 db 
below the value in July and August. Two reports*®’7 
have already been issued on the results obtained from 
an analysis of the records, and a third will be pre- 
pared shortly. 
Among the main conclusions so far reached are the 
following. Cloudy weather, either by day or night, 
tends to produce a signal steady to within about 2 db, 
while on days of clear and variable skies, the signal 
exhibits slow (period 5 to 10 minutes) fluctuations. of 
the order of 3 or 4 db, with sharper and more rapid 
fluctuations superposed. These rapid fluctuations are 
accentuated by the presence of strong wind. On nights 
of clear or variable skies with temperature inversions 
near the ground, the field intensity is from 5 to 10 
db above the daytime level and is usually accompanied 
by variations in the absence of wind. On clear radia- 
tion nights, when the wind is too strong to permit the 
development of a temperature inversion, the peak of 
signal intensity does not occur. Fog affects the field 
strength differently according to its depth. A shallow 
autumn fog causes a sharp decrease in signal strength, 
while the widespread and more established type of fog 
experienced in winter may sometimes cause marked 
interference type of fading with unusually high peak 
values and at other times may have no apparent effect 
on the field strength. 
The sending and receiving stations on this link are 
now being equipped with field strength monitoring 
arrangements to improve the overall accuracy of the 
radio recording, and a daily statement of the meteoro- 
logical conditions over the path is being supplied to 
supplement the ground station records already avail- 
able. It is contemplated that this link should remain 
in operation for a further period of 6 to 12 months. 
FADING IN A LINE-OF-SIGHT 
EXPERIMENT IN ENGLAND: 
This experiment was carried out between Aberporth, 
South Wales, and the summit of Mt. Snowdon (3,600 
°By F. Hoyle, Ultra Short Wave Panel, Ministry of Supply, 
England. 
ft). The transmitters were mounted at a height of 
120 ft above sea level at Aberporth, the receivers being 
on Mt. Snowdon. The length of path was about 60 
miles as compared with 85 miles optical range. 
Two separate radio circuits were used, one on a 
wavelength of 9 cm and the other on 10 cm. For a 
standard atmosphere the phase difference between the 
ray reflected from the sea and the direct ray was about 
4.5 radians on 10 cm and about 5 radians on 9 em, 
while under “flat earth” conditions the corresponding: 
phase differences were about 12 radians and 13 radians 
respectively. As the atmospheric conditions varied, an 
interference pattern was obtained arising from varia- 
tions in the phase difference. The chief characteristic 
of the interference pattern was that in a plot of radio 
signal strength against time the peaks are broad and 
flat and the minima are sharp and deep. The provision 
of the two circuits insured that the variations due to 
alterations in the phase difference could easily be 
distinguished from yariations due to other causes 
(described below). The reason for this is that the 
interference minima on one circuit tend to occur at 
times when the signal strength is high on the other 
circuit. 
The length of time required for the radio signal 
strength to complete a cycle of the interference pat- 
tern was usually about 2 hours. The receivers and 
transmitters were very carefully calibrated in order 
to show that the received field strength at interference 
maximum was equal to twice the free space field. This 
result was established to within an accuracy of +2 db. 
For about 20 per cent of the time (June 1, 1944 to 
September 30, 1944) the normal interference pattern 
was replaced by an entirely different form of signal 
fading. The period of the fading was about 5 minutes, 
the field strength at maximum was usually between 
10 and 12 db above the free space signal, and the 
peaks of signal were sharp and the minima broad. The 
latter characteristic is entirely different from interfer- 
ence fading between two rays which must lead to 
broad peaks and narrow minima; it is more akin to 
receiver noise or the form of the signal echo received 
from “window” on radar sets. Thus it would seem 
plausible to suppose that the signal was the result of 
a large number of contributions with uncorrelated 
phases. 
The type of fading described in the previous para- 
graph is especially interesting in view of the very high 
signal maxima. It was shown that the occurrence of 
these variations was not associated with the reflection 
at the surface of the sea. It would appear therefore 
that atmospheric conditions can exercise a very im- 
portant effect on the propagation of radio waves over 
a completely optical path. 
TEMPERATURE EFFECTS ON 
NONSTANDARD RANGES‘ 
Experimental work carried out in the Irish Sea has 
shown the following three characteristics, all of which 
are in disagreement with existing theory. 
1. It is well known that the present theory requires 
the contribution of the temperature gradient to be in 
general small compared with the contribution of the 
water vapor gradient. In fact if we write 
dp de aT 
where » = refractive index, 
e = partial pressure of water vapor, 
T = temperature in degrees absolute, 
h = height coordinate, 
a,b are positive constants, 
then except in rare cases the present theory requires 
the (a) (de/dh) term to be large compared with the 
(b) (d2'/dh) term. Thus, since the radio propagation 
aBy F. Hoyle, Ultra Short Wave Panel, Ministry of Sup- 
ply. England. 
