OBSERVATIONS IN FOUR BALLOON ASCENTS. 335 
exhibited, in a less marked manner, on November 10. On October 21, the departure 
from a uniform decrease is very decidedly shown in connection with the stratum of 
dense cloud passed throug-h. The temperature had been uniformly decreasing until 
the thick cloud was reached, when a decided rise commenced, which continued 
through the cloud, and for a space of about 600 feet above it ; after which height the 
decrease was resumed, at first slowly, and afterwards with more rapidity. 
The disturbance in the variation of the temperature now noticed, is in each series 
exhibited in such a systematic manner, that the hypothesis of a regular progression 
at all heights can scarcely be maintained. In order therefore to arrive at some 
approximate value of the normal variation of temperature in the atmosphere, it 
appears necessary to make abstraction of the disturbing cause. This I have en- 
deavoured to do by dividing each series into two divisions; 1st, between the earth 
and the height where the diminution of temperature appears to be arrested ; 2nd, 
above the point where the regular diminution of temperature seems to be resumed^ 
omitting the space which is under the influence of the disturbance. The divisions 
adopted for the four series are as follows : — 
Aug. 17. 
Feet. 
0 to 4000 
Aug. 26. 
Feet. 
0 to 7000 
10,000 to 19,000 
Oct. 21. 
Feet. 
0 to 2700 
5000 to 13,000 
Nov. 10. 
Feet. 
0 to 4000 
Lower division 
Upper division 7000 to 20,000 1 u,uuu to 19,000 5000 to 13,000 9000 to 23,000 
These partial series have been examined by the same methods as the entire series ; 
the number of groups in the lower division being, however, with the exception of 
August 26, too small to admit of the application with any advantage of equation (2.). 
The results for the different series are as follows : — 
By Equat ion (1.). By Equation (2.). 
X 
Y 
Mean error. 
X 
y 
Z 
Mean error. 
Aug. 17 . ^ 
f Lower division 
. 71-62 
-3°-598 
0-00 
1 Upper division 
. 76-68 
- 3*371 
0-73 
7°9-l7 
- 3-771 
+ 0-01484 
00 
Aug. 26 . j 
f Lower division 
. 67-46 
- 3-549 
0-39 
66-75 
- 2-969 
-0-08220 
0-30 
L Upper division 
. 76-36 
— 3-355 
0-60 
81-68 
— 4-104 
+ 0-02552 
0-58 
Oct. 21 .j 
r Lower division , 
. 57-77 
— 3-581 
0-26 
L Upper division . 
. 68-77 
-3-376 
0-82 
67-81 
- 3-162 
- 0-01119 
0-80 
Nov. 10 . j 
f Lower division 
50-21 
-3-760 
0*26 
L Upper division . 
. 59-45 
-3-046 
1-22 
48-05 
-1-516 
- 0-04791 
1-02 
The values of the constants in equation (2.), deduced from the higher divisions, 
show that, in the two series of August 17 and 26, the temperature decreases less 
lapidly as we ascend ; whilst the values for October 21 and November 10 indicate a 
contrary result. Tbe value of the second term (^r) is, with the exception of the series 
of November 10, very small, and tbe amounts of the mean errors show that the 
observations are little better represented than by the single constant of equation (1.). 
MDCCCLIII. 2 Y 
