May 28, 1914] 
NATURE 321 
tts advantage is that it is the time of the morning 
weather chart, and hence the results can be plotted on 
the chart with the knowledge that all the observations 
shown are simultaneous. But the objections are two- 
fold. In many cases figures obtained at great expense 
have to be rejected, because they are obviously falsified 
by solar radiation. This must happen if the balloon 
does not burst, and the sun is high, whereas if the sun 
is near to or below the horizon it is of no consequence 
if the balloon does or does not burst. This gives 
cause for doubt also about some of the printed figures 
since it is not too clear in all cases as to what may be 
accepted and what must be rejected. Secondly, it is 
impossible to get the annual or the daily variation 
from observations at a fixed hour. The daily variation 
is, of course, hopeless, and not knowing the law of 
the daily variation, it is uncertain whether the same 
correction for the hour should be applied both summer 
and winter. The result is that the annual variation 
above 12 km., as shown_by the English ascents, many 
of which were made at sunset before the international 
time of 7 a.m. was fixed, differs by 3° C. from the 
Continental value, but it is very improbable that there 
is any real difference. A plan has now been adopted 
by which the string carrying the instrument uncoils 
after the balloon is started, and since last winter a 
very much longer string has been employed in the 
English ascents. This avoids the difficulty of starting 
with a long string in rough weather, and it will be 
interesting to see what effect the plan will have on the 
records. The change of length is from 44 to 132 ft. 
W. H. Dives. 
May 20. 
‘ 
Transmission of Electric Waves Round the Bend of 
the Earth. 
IN a paper on the transmission of electri: waves 
round the earth’s surface, read by Prof. H. M. Mac- 
donald before the Royal Society on February 12, some 
conclusions are recorded which cast new light on the 
problem of long-distance wireless telegraphy. Prof. 
Macdonald’s point of view is that of simple diffraction, 
and the paper is the latest one of a notable series of 
attempts by a number of eminent mathematicians. 
in the present paper the author reduces his formulz 
fo figures, and thus makes comparison wiih experi- 
ment easy. The most extensive quantitative experi- 
ments yet made over great ranges are those of L. W. 
Austin in 1910 (Bulletin Bureau of Standards, vol. vii., 
No. 3), and those of J. L. Hogan in 1913 (Electrician, 
August 8, 1913). From the former experiments Austin 
and L. Cohen deduced a formula which has been 
corroborated by Hogan’s results. This formula may 
be written :— 
Z=ce-ax/VA/ (Ax), 
where 7 is the current in amperes in the receiving 
antenna at the distance x kilometres for the sending 
station, Ais the wave-length of the radiation in kilo- 
metres, a@ has the value o-o015, and c, like a, is a 
quantity which does not depend on A or x. This 
formula was deduced from daylight experiments ex- 
tending over larger ranges of A and x than those 
used in the table below. 
By aid of this formula Prof. Macdonaid’s calcula- 
tions can be quickly compared with the results of 
experiment. In the following table the first column 
contains the number of miles between sender and 
receiver, and the remaining columns contain the ratios 
of the effect at various distances to that at 419 miles. 
R, is the ratio, calculated on the diffraction theory. 
between the electric fields, R,, is the ratio found by 
measurement between the currents in the same re- 
NO. 2326, VoL. 93] 
| 
| rather, is contradicted by 
ow = 
_ for rather long ranges. 
'borne out by the experience of wireless telegraph 
ceiving antenna when moved to the successive dis- 
tances in turn. 
Miles} A=320m. A625 m. | A=1220m. | A=2560 m. | A=5c00 m. 
| 
(ia, me = 
Wh Petonets ay | Ra | Re | Rm | Ra | Rm | Ra | Rm 
| | ty } : a 
| | | | 
419 I Tee ts Aare: ier 1 I I I | Dy ei vz I 
536 | 07304 | 0481 | 0.392 | 07554 |0°464 | 0°605 |0°537 | 0660 | 0°585 | 07693 
675 0°128 | 0°286 | 0184 | 0°349 | 0 256 0°418 | 0°315 | 0°467 
814 | 0°9764| 0218 | 0128 | 0'282 | 0°178 | 0°336 
1970 | 0°0392| 0°148 | 0°0637| 0°184 
1257 | - | 00321) 0°134 
From the table it seems fair to draw the conclusion 
that diffraction accounts for a large proportion of 
the observed effects up to distances of, perhaps, 2000 
miles. The proportion is much larger than has 
hitherto been demonstrated, and compels the admission 
of diffraction into the list of phenomena contributing 
to the practical success ot wireless telegraphy. 
But the Austin-Cohen formula expresses a remark- 
able experimental fact which is not explained, but, - 
the diffraction theory; 
namely, that for each distance there is a best wave- 
length. The formula indicates that this’ optimum 
wave-length is given by 4A\=a*x*, and, consequently, 
that under the best condition 
b= Ab te)> -X —° S047 (XK 214006 10", 
These equations are, broadly, 
engineers. ‘The equation for A shows that as the 
wave-length is increased the effect at a given place 
first increases and then decreases. Here the theory 
of diffraction appears to fail, for the diffraction effect 
at any fixed point should increase steadily with in- 
crease of wave-length. On the other hand, the hypo- 
thesis of the refraction of electric waves in the atmo- 
sphere when it is ionised by sunlight seems more 
| promising. For while radiation of; very short wave- 
| length is lost into space by the rays suffering too 
little bending, radiation of great’ wave-length, by 
being too strongly refracted, is lost in the ground 
b ~eer, the oscillator and the receiver; and thus an 
opti sum wave-length is easily conceivable, 
W. ECccLEs. 
University of London, University College, May 18. 
Some Phenomena of Clay Suspensions. 
THE interesting letters in Nature on the cellular 
structure of emulsions induced me to test the be- 
haviour of clay which I have been accumulating for 
some time for evaporation experiments. The clay is 
>tained by the usual sedimentation method, and is 
tnat fraction of the soil which does not settle in 
“venty-four hours in a beaker containing dilute 
«a nonia to a depth of 85 cm. The suspension is 
{ mn ~aporated to dryness in vacuo. 
if some of this dried clay be well shaken up with 
strong ammonia solution and poured. into a_ Petri 
dish, the usual network, mentioned by Mr. Wager 
in Nature for May 7, gradually develops. Occa- 
sionally a different pattern appears, only the angles of 
the network are formed, and the surface thus has a 
pitted appearance. 
The pattern persists for a few minutes and then 
orally becomes blurred. The two cases are shown 
in F..s. 1 and 2 respectively. In neither case are 
groups of cells formed. Although the structure 1s 
« ite sharp to the eye, the lack of contrast makes 
the photographic difficulties considerable. [T am in- 
- 
