EF. Loomis— Observations of the U. S. Signal Service. 11 
force, and that the violent uprising of heated air, which is fre- 
quently witnessed in humid climates, especially during thunder 
storms, is mainly due to the presence of a large amount of 
aqueous vapor. 
I next made a comparison of the cases in which the tempe- 
rature at Denver was lower than at Pike’s Peak. These cases 
are thirty-nine in number, embraced in a period of twenty 
months’ observations, and they are shown in the preceding table 
which is arranged like that on pages 8 and 9. 
t will be noticed that thirty-one of these cases occurred in 
the month of January, and all occurred in the four months 
from November to February. The average relative humidity 
on Pike’s Peak at these dates was eighty-four, and at Denver 
seventy-one. On Pike’s Peak about half of the winds were 
from the southwest, and none of them were from any eastern 
point. - The average velocity of the winds was twenty-seven 
miles per hour. At Denver the wind never blew from any 
western point, and its average velocity was only four miles per 
hour. At Pike’s Peak the average cloudiness (counting fog as 
sky entirely overcast) was 0°71; and at Denver 0°36. The 
barometer at Denver was sometimes below the mean and some- 
imes above it, but the average was ‘06 inch above the mean. 
One of the most noticeable circumstances exhibited. by this 
table is the humidity of the air at Denver, showing that this air 
did not come from the west side of the Rocky Mountains, and 
the —_ fact is indicated by the observed direction of the 
winds. 
We thus learn that during periods of severe cold at Denver, 
the thermometer is frequently lower there than it is on Pike's 
Peak, and hence we must conclude that this cold did not result 
pet the subsidence of air from the upper regions of the atmos- 
ere. 
In order to test this conclusion more fully, I selected all 
those cases in which the thermometer at Denver sun ow 
as +5° from November, 1873, to June, 1878, and the lowest 
temperature at Pike’s Peak for the same date was entered in 
the same table. The number of these cases was ninety-nine. 
The average of these observations at Denver was —2°4°, and 
the average at Pike’s Peak was —9:0°, showing that it was 
only 6°6° colder at Pike’s Peak than at Denver. 
next made a similarcomparison for Mt. Washington and 
two neighboring stations. I selected all the cases from Novem- 
ber, 1878, to June, 1878, in which the thermometer at Burling- 
ton, Vt., sunk as low as +5° and determined the lowest tempe- 
rature on Mt. Washington for the same dates. The number of 
these cases was 145. The average of these observations at Bur- 
lington was —2-7°, and at Mt. Washington —18-9°, showing 
