——————y—————— 
PRESENT STATE OF OUR KNOWLEDGE OF THE UPPER ATMOSPHERE. 103 
larities in occasional ascents indicates that there is in places limited con- 
vection, and the considerable inversion of temperature frequently found at 
the dividing surface suggests that there may be oblique convection similar 
to that for anticyclones in the lower atmosphere. Any fall of temperature 
arising from such convection would tend to disappear owing to the effect 
of radiation. In general, however, interchange of air in the upper region 
would be mainly by advection and the two regions might be appropriately 
named advective and convective regions,’ expressing the characteristic 
difference between them. 
The height of the dividing surface will be denoted by H,, and the 
temperature at this height by T.. 
Although H, varies with the latitude, the observations available are 
insufficient to enable an accurate expression for the relation to be obtained. 
Teisserenc de Bort? found from simultaneous ascents at Trappes and at 
Kiruna on the Arctic circle that the value of H, was practically the same 
for the two places, but the value of T, was slightly lower for Kiruna. 
Towards the equator, on the other hand, the value of H, is considerably 
greater than for temperate latitudes. Rotch and Teisserenc de Bort failed 
to reach it over the Atlantic with balloons reaching 15 km. It has been 
reached, according to Assmann,*® by a German expedition which sent up 
balloons from a steamer on Lake Victoria Nyanza in 1908. Two of these 
reached heights of 17 and 19°8 km., and both entered an isothermal region, 
the temperature in which was lower than that found in temperate latitudes. 
The lowest temperature was — 84° C. or 189° A. 
The following table gives the mean values of H, and T, for certain 
places determined from the monthly mean values. The thirteen stations 
are those enumerated above and Guadaljara, Milan and Pavia, and 
Hamburg :— 
= 
Mean of 
(ot © E : : | 
13Stations onion England st Paris | noes | as Milan | Vienna | Berlin | 
| 
H. 106 | 109 | 108 | 108 | 104 |. 96 | 106 | 107 | 102 | 107 | 
ies). |, 16° 16> .|. 18° | 15° «| 18° |. ag0 | wae | aze | 45° | qe° | 
No.of Cases} 336 53 32 67 57 28 | 18 25 | a4 | 32 
Latitude .|  — 4g? | 52° | 49° | 49° | Goo | see | 45° | 48° | 52° 
There is very little variation for places in Europe between lat. 45-55°: 
the more Continental stations give a slightly lower value for T, than the 
others. For Pavlovsk the value of H, is 1 km. below the average for the 
other stations and the value of T, is 2° above the average value. 
The results indicate that there must be a comparatively rapid increase 
in H, in crossing the limit of the trade-wind region, and it appears probable 
that the equatorial currents and the trade winds form a closed system 
without very much interchange of air with higher latitudes. 
Schmauss‘ has pointed out that the value of H, is greater in summer 
than in winter.> The following table shows the annual variation in H,, T, 
' The upper region has been usually described as ‘ the isothermal layer.’ Recently 
T. de Bort has introduced the terms ‘stratosphere,’ ‘troposphere,’ to denote the 
upper and lower regions. 
2 Compt. rend., 145, 1907; Met. Zeit., 1907. 
8 Quart. Journ. Royal Met. Soc., 1909. 
+ Registrierballonfahrten, Miinchen, 1908. 
> Rotch states that in America the conditions are reversed, and the minimum 
value of H. occurs in the summer. This is probably due to the inclusion of the 
large values he found in October in the winter series. Met. Zeit., January 1909. 
