390 PELTIER ON THE CAUSES OF THE 
the index of my electrometer is less considerable when the in- 
strument descends than when it is raised; the difference is very 
obvious, and the amount ought to be stated in order to ascer- © 
tain the direct tension for the purpose of forming correct tables. f 
The cause of this difference is that the action of the globe in- | 
creases more on the lower than on the upper end of the stem, 
as may be seen in the table. 
Value decreasing as | Ratio between the 
the square of the superficies of each 
distance. stratum. 
| Equal and super- 
posed strata. 
0-°0204 
ae} a) 
0-027 ) ss 
| 
| 
mt WO OCI ST © 
> 
= 
Ca 
i) 
= 
The difference obtained with an instrument held in the hand 
is less than the table indicates; and, in fact, the armatures com-_ 
municating with the ground preserve an almost equal action on ; 
the gold leaves during the descent of the instrument. It is | 
necessary to make some previous experiments in order to de- | 
duce the proportion from them, if it be desired to take them — 
into consideration in tables of atmospheric electricity. When | 
I develop these principles I will give examples of the formation — 
of these tables and of the corrections to be made to the rough 
observations. 
We see by what precedes, that in nature the same takes place 
as under the induction of an electrified body ; in the atmosphere 
we obtain only an unequal distribution of electricity arising 
from the resinous action of the earth. These experiments well - 
understood, those of Saussure and Erman no longer present 
any difficulty, we find in them all the results of electric induc- 
tion, and not those of communication and distribution. 
29. If it be desired to keep to the simplicity of the primitive 
fact, a greater length than five decimeters must not be given to 
the stem of the electrometer ; we even prefer to give it only three 
or four ; by taking a longer stem the sensibility of the instru- j 
ment diminishes. The reason of this is simple: the electricity of 
induction, coerced at the extremity of the stem, allows that of 
the contrary name to distribute itself over the remainder of the 
