58 
NATURE 
[SEPTEMBER 18, 1913 
view of the theodolite either by a trick of eyesight 
or by the accident of clouds, long before any such 
distance has been reached. Consequently, the in- 
vestigation is limited to clear weather, and for 
the most part to the lower layers of the atmo- 
sphere. The lower half of the troposphere, say, 
up to five kilometres from the surface, is the 
region specially under observation, but when the 
sky happens to be clear the investigation can be 
extended to much greater heights. Mr. Cave 
gives twelve examples of series of observations 
beyond 11 kilometres, and one up to 18 kilo- 
metres. He is therefore able to devote a chapter 
to the winds of the upper layer, the stratosphere, 
and he supports the general conclusion that the 
wind falls off rapidly as the boundary of the 
stratosphere is approached and passed, though we 
must wait to learn whether this result is charac- 
teristic of the stratosphere or merely character- 
istic of the weather when the stratosphere comes 
under observation. 
Useful chapters will be found devoted to 
methods of observing and their accuracy, and to 
the rate of ascent of balloons, with an examination 
of the effect thereupon of the orographical fea- 
tures of the neighbourhood. The author then 
takes up the meteorological applications of his 
results. This section takes the form, for the most 
part, of a study of the relation of the strength and 
direction of currents aloft to the distribution of 
pressure and temperature at the surface, and leads 
up to an important diagram on p. 75 showing the 
upper winds in relation to a hypothetical distribu- 
tion of high and low pressure at the surface. The 
diagram represents increased velocity aloft in the 
westerly currents of a “low,” and in the south- 
westerly and north-westerly currents of a “high,” 
but a diminished velocity in the easterly wind of 
a “high.” To judge by the text, an unchanging 
current might have been represented in the more 
central area where there is little pressure 
gradient, and certainly a reversal of the north- 
easterly current on the south-eastern side of a 
high-pressure area. But the most striking feature 
of the diagram is a strong north-westerly upper 
current increasing with height (across the surface 
isobars and the south-westerly surface winds) 
from the central region of a “low” to the eastern 
region of the neighbouring “high.” This note- 
worthy current which must be closely associated 
with the dynamical structure of the atmosphere is 
rightly selected as one of the types of structure to 
which attention is specially called. It is in line 
with observations of cirrus cloud in front of a low- 
pressure area. 
1 The falling off of wind with height in the stratosphere can be showa to 
be a logical consequence of the higher temperature of the region of lower 
pressure. 
NO. 2290, VOL, 92 | 
formula, to which that used by Mr. Cave approxi- 
The relations to the sequence of weather on 
many occasions are set Out in detail, but the 
results are not easily generalised except in the_ 
special case of the reversal of the current over 
a north-easterly wind, which is shown in many | 
instances to be the precursor of rain and thunder- 
storms. : 
Some attention is given to the relation of the 
direction and strength of the wind in the upper 
air to the pressure gradient at the surface, 
assuming that the gradient wind is tangential to 
the isobars. As regards direction, there are useful 
diagrams showing the relation of the gradient 
directions to the surface winds for the three types 
of structure, viz. : (1) “solid current,” (2) increase 
of velocity aloft, and (3) decrease of velocity aloft ; 
the last shows the decreasing winds to be limited 
to cases of surface wind between north and south- 
east. As regards strength, gradient velocities are 
calculated from the usual formula : 
V =y/(2up sina), 
and the increase of velocity with height for some 
situations is attributed to an increase of gradient 
deduced from the distribution of temperatures in 
the lowest layers indicated in the published 
weather-charts. 
The increase of gradient is calculated from the 
surface temperature by a rough and ready formula 
which, considering the local influences upon tem- 
perature and other circumstances, is sufficiently 
accurate for Mr. Cave’s immediate purpose; but 
it may be useful to give here a more accurate 
mates. Neglecting the effect of humidity, which 
is certainly small and usually unknown, the 
increase of pressure difference in millibars for h 
metres of height measured from any level is given 
by the formula : 
sn-spmorsa (2090), 
where p is the pressure at one place, p+Ap that 
at another on the same level, @ and A@ are the 
temperature and temperature difference, and Ap) 
is the pressure difference at h metres above the 
given level. 3 
Near the surface p/@ is approximately equal 
to 3, so that for the first k kilometres from the 
surface the formula would become approximately : 
Ap. — Aps= r00k( 48 = *f). 
The approximate formula used by Mr. Cave is 
practically identical with this, except that the 
term Ap/p is omitted. The omitted term may 
be sufficiently small to be neglected for the surface 
layers when Ap is not large, because p is 
