430 
upper layers of the equatorial current appeared to 
have been raised unchanged through the height 6h. 
Provided that the polar air was not more than 2 or 
3 km. in depth, anticyclones formed in this way would 
be ‘‘ warm ’”’ anticyclones, and would possess the 
features associated with such. But there are almost 
certainly cases where the encroaching polar air 
extends right up to the base of the stratosphere, and 
these appear to have all the characteristics of the 
cold, rapidly-moving anticyclone. This cold air, 
passing as it does into latitudes warmer than those 
where it acquired the main features of its existing 
temperature distribution, is heated from the bottom 
upwards, and becomes sufficiently unstable to pro- 
vide within itself moderate rain and much cloud, 
but probably not persistent heavy rain. (It seems 
likely also that anticyclones do reach us in which 
there is either no polar surface air or only a negligible 
amount. Their formation was probably a much more 
gradual though similar process, and took place in more 
southerly latitudes.) 
Mr. Dines has referred to the difficulty of maintain- 
ing the polar air im situ. The patch of polar air with 
which we are dealing may be described as a roughly 
circular one of 1000 or more km. in diameter; in the 
case of a ‘‘warm’”’ anticyclone we may limit its 
depth at the deepest part to 2 or 3 km. ; in the case 
of a ‘‘ cold ’’’ anticyclone the depth in the centre may 
include the whole thickness of the troposphere. It 
appears to be maintained im situ, so fay as it 1s main- 
tained, by the currents which produced it.. But 
actually the motion of most “ cold’’ anticyclones— 
i.e. those of the deep polar air—does strongly resemble 
that of the flat drop of mercury on the laboratory 
table. 
This problem was dealt with hydrodynamically by 
Exner in 1918 (Sitzungsber. Akad. Wiss., Wien Ila, 
127, 1918, pp. 795-847). He assumed as the initial 
conditions the existence of a mass of cold dense air 
(at rest or in motion) covering a small portion of the 
earth’s surface and surrounded on all sides and above 
by warmer, less dense air. Particular points made 
by him include—(1) that the rotation of the earth 
renders possible the maintenance (at a slight inclina- 
tion to the horizon) of a definite fixed bounding 
surface between the cold and the warm air; (2) that 
if a long ridge of cold air divides into two ridges 
flowing apart like cold waves, then the square of the 
velocity of separation of these waves is proportional 
to the depth of the cold air and to the difference of 
density between the cold and the warm; (3) also 
that in such a case friction with the earth’s surface 
results in a shallow cold film being left over the whole 
area traversed by the waves and in the consequent 
gradual reduction in the height of the waves. 
There is another consideration which supports the 
view that an anticyclone is of complex structure, and 
that is the frequency with which the air above an 
“inversion ’’ of temperature can be shown to be of 
different origin from that below. It has usually been 
said that the surface layers were being cooled by 
radiation, also that there was outflow of air in these 
layers, and that the upper air, descending and settling, 
was being warmed adiabatically. When, however, 
an attempt is made to apply numerical data, cases 
arise where the change of temperature at a given 
point in space appears to have taken place much more 
rapidly than can be provided for by the most favour- 
able time scale of the assumed operating causes. 
But in particular it is difficult to see why these causes 
should lead rapidly to the formation of comparatively 
sharp discontinuities of temperature of the order of 
10° F., and also how they can lead to other than a 
very unstable vertical distribution of temperature. 
It seems much simpler, being provided with air of 
NO. 2787, VOL. 111 | 

NATURE 
[ MarcH 31, 1923 
about the appropriate temperatures to northward 
and southward respectively, to explain the formation 
of anticyclones and their temperature distribution by 
means of the horizontal motion and interaction of 
these ‘“‘ polar ’’ and “‘ equatorial ’’ currents. 
* A. H. R. Gotpie. 
Wimbledon, S.W.19, March 8. 
The Phantom Island of Mentone. 
On a fine dark night, looking towards the point of 
Mentone from the sea-front about the middle of the 
West Bay, the appearance is presented of a dark 
island rising out of the sea in the gap which separates 
the lights of Mentone from those of Bordighera, some 
ten miles distant. This ‘“ phantom island ” appears 
to be about 200 feet high, and from its darkness one 
would imagine it to be thickly covered with vegetation, 
its sides rising steeply out of the water. It is directly 
opposite, and quite near the sea-front of Mentone, 
from which it is separated by a very narrow channel 
of water. It appears, in fact, to be quite close to 
Mentone. 
The explanation of this curious optical illusion is 
comparatively simple. The lights of Mentone and 
those of Bordighera present the appearance of being 
ranged round a curved bay, and they throw their 
reflections on the water, but they are separated by 
the East Bay, which is not seen, and by a dark, 
unilluminated portion of the coast. The correspond- 
ing part of the sea is devoid of reflections, and the 
impression is produced of a dark obstacle breaking the - 
continuity of the line of lights and of their reflections 
inthe water. This effect has been seen by independent 
observers on several occasions. 
G. H. Bryan. 
University College of North Wales, 
March 6. 

Ball Hardness and Scleroscope Hardness. 
In the ball hardness test Meyer found that L=ad". 
By combining this relation with Brinell’s formula H = 
L/A, it can be shown that the hardness number when 
the ball is immersed up to its diameter is 2p". 
This value has been called the “‘ ultimate hardness ”” 
(H,), and is independent of the initial condition of the 
metal with regard to cold work. 
Several attempts have been made to obtain a 
relation between standard Brinell and scleroscope 
numbers. The results have been more or less unsatis- 
factory. If, however, values of H, be plotted against 
the scleroscope numbers of metals in the annealed 
condition, the points lie on a smooth curve which is 
independent of the ball diameter. The following 
results have been obtained by the writer using balls 
of r mm. and ro mm. diameter: 










Sample, a. n. Hu. ii 
Tin 5°53 | 27185 54 35 ro mm, 
Zinc’. 7 eee ert 25 II 
Steel A 74 2:288 | or 27 
ie 185 2°292 | 231 51 
rat. > ab2 2°292 | 327 64 
3 JNg4z 2293 | 428 73 
Armco 94 2164 | 60 au Imm, 
Steel 2N 5. -|surz 2185 | 71 23 
Pedy . | 150 2:247 | 96 27 
S90 . | 264 2:298 | 168 41 
Manganese ‘ 
Steel 453 2-303 | 288 50 





