6 NATURE 
[Marcu 5, 1914 
to —57-:1° C. is shown, the latter agreeing with the 
value which is usually found in Europe. In former 
balloon ascents made at Batavia, temperature records 
have been obtained only twice for heights above 20 km. 
In one of those cases (August 6, 1913), an increase 
similar to that of December 4 was recorded, viz., 
—82-6° C., at 17 km., and —63-7° at 22 km.; in the 
other case, however (October 2, 1912), the temperature 
showed a much smaller increase (—80° C. to 
—75° C.), the balloon reaching 23 km. 
Regarding the trustworthiness of the records, the 
scale-values of the thermo- and barograph may be 
entirely depended upon, the instruments having been 
subjected to thorough verification before the respective 
ascents. 
The only seemingly prejudicial circumstance is that, 
though the balloon was liberated before sunrise, the 
sun rose above the horizon as the balloon entered the 
stratosphere, so that the insolation may have caused 
an apparent rise of temperature, notwithstanding 
the sun was low in the sky. Comparison of the 
temperature records for ascent and descent, however, 
prove that this has not been the case. The balloon 
descended at a quicker rate than it ascended; accord- 
ingly ventilation was more efficient in the downward 
movement, and consequently any heating effect of the 
sun-rays smaller. Thus we should expect higher tem- 
peratures during the ascent, even if we take into 
account that the sun had risen higher in the mean- 
time. On the contrary, however, on December 4 the 
temperatures during ascent were lower than, and on 
August 6 nearly equal to, those recorded during 
descent. The higher temperatures in the descent 
on December 4 are easily explained as an 
effect of sluggishness of the thermograph, especially 
as the sign of the difference between ascent and descent 
changes with the temperature gradient, when going 
from the troposphere to the stratosphere, quite in 
accordance with any effect of sluggishness. The 
values of temperature and ventilation are given here 
for heights from 15 km. upwards :— 
Height ilenmmeratire Hae ete: Ventilation 
inkm. Ascent Descent ge ture aigher — (m.p. sec.) 
in ascent Ascent Descent 
“OR "Gs "GC. . 
15 -76'4 —79°6 —78'0 202 oes 1'6 
155 eNO ue 4 Ok | STkS 50 
16 - $3°1 —89'1 - 86°1 6°0 o>4 1°6 
16°5 — 87°! -90°9 —89'0 Se) 
17 —89°5 -—897 —89°6 o'2 03 0'9 
75s! 8018. — 85:8. ye — 87-3 a0 
18 Org | comran)  $Becges | aso o'4 0°5 
19 On gt) 7420. '— Sore 64 02 06 
20 —74°5 68:0) 9 — 77 —5°6 o'2 06 
2I —69°7 -66°7. —68:2 —3°'0 o'2 O7 
22 —67°8 — 64'2 = 00109 136 oO'2 o'7 
23 —640 -—62°'0 — 63°0 20 o-2 O'7 
24 — 60'0 — 60°0 — 60°0 foe) O'2 On 
25 TOO be OA = 5054 oo o'2 o°2 
26 = 52 S572 =] 0'O o'2 o'2 
As to the temperature gradients, it may be re- 
marked that in the stratosphere they show a succession 
of low and very high (>1-0) values. 
I believe the remarkable behaviour of temperature 
in the tropical stratosphere, revealed by these upper- 
air soundings, will furnish a key to the explanation 
of the two salient features of the stratosphere, viz., 
the lowering of the temperature at its base and its 
rise in height, when proceeding from the poles to the 
equator. W. vAN BEMMELEN. 
Batavia, January 23. 
NO. 2314, VOL. 93] 
The Vertical Temperature Distribution in the 
Atmosphere. 
In the observations described by Dr. van Bemmelen 
in the foregoing letter, the vertical temperature distri- 
bution in the tropics is so typically represented that it 
seems worth while to consider a little more closely the 
essential difference between the curve obtained from 
them and curves obtained in temperate latitudes, and 
to discuss its probable cause. This difference, as Dr. 
van Bemmelen has pointed out already, chiefly relates 
to the greater height of the stratosphere and the large 
and rapid increase of temperature in it. 
As to the vertical temperature distribution in higher 
latitudes, theory has already been able to give account 
of its principal properties. These theories (Hum- 
phreys, Gold, Emden‘), however, deal only with those 
latitudes and not with the particular features of the 
tropical atmosphere. I will briefly formulate the 
results of these researches so far as they will be used 
in the further discussion. 
(1) When convective temperature equilibrium is sup- 
posed to exist and the decrease of water vapour with 
height is taken into account, Gold finds that above 
the isobaric surface of a quarter atmosphere, radia- 
tion has a heating effect, below a cooling influence. 
In Europe this surface has a height of 9500 metres. 
With a slightly different conception as to the distribu- 
tion of water vapour Emden calculates nearly the same 
height for the limit between heating and cooling effect, 
viz., 8950 metres. 
(2) According to Emden, his equations used for the 
lower regions do not hold good in the upper part of 
the troposphcve in consequence of the very small quan- 
tity of water vapour. Taking this small amount of 
vapour into account, he derives a minimum radiation 
temperature of —59° C. Supposing the condition 
in (1) to be gradually changing into those mentioned 
in (2), his theory fully agrees with the facts observed 
over the tropics. 
(3) The equations of Emden show at greater eleva- 
tions a gradual increase of temperature in the strato- 
sphere, in agreement with the results of observation. 
The very low temperatures, observed in the upper 
part of the tropical troposphere, which are about 
30° C. below those observed at the same height in 
temperate regions, must be ascribed to another effect. 
Besides the radiation and the resulting vertical con- 
vection currents, which explain the principal features 
of the gradients in higher latitudes, in the low-pressure 
belt of the tropics the rising air currents of the general 
atmospheric circulation cooperate. They disturb the 
temperature distribution as determined by radiation, 
and shift the troposphere to greater heights. 
Dr. van Bemmelen, who found at Batavia the upper 
limit of the anti-trade winds at about the same level 
as the base of the stratosphere, thereby proved that 
the convection currents reach as high as the upper 
limit of the troposphere. 
Also from a theoretical point of view, it is evident 
that the vertical convection currents do not reach 
higher than the limit of strato- and tropo-sphere, as 
this would be inconsistent with the stability of the 
temperature gradients of the stratosphere; on the 
other hand, it is improbable that their height remains 
much below this limit, for without convection the high 
radiation temperatures would rapidly take possession 
of their dominion. 
At higher levels the conditions in the stratosphere 
will rapidly approach those of the temperate latitudes 
at the same height. Therefore the marked increase 
of temperature, as shown by the observations is 
exactly what might be expected. 
1 R. Emden: Uber Strahlungsgleichgewicht und atmospharische Strah- 
lung. Ein Beitrag zur Theorie der oberen Inversion. Sitz. Ber. d. math. 
| phys. Klasse d. K. B. Akad. d. Wiss. zu Miinchen, 1913, Heft 1. 
