8 E. Loomis— Observations of the U. S. Signal Service. 
the summit of the mountain from May 12, 2 Pp. M., to May 16, 
2. P. M., the interval between the horizontal lines representing a 
difference of one-fiftieth inch pressure, and the interval between 
the vertical lines representing two hours of time. The second 
curve represents the barometric fluctuations at the base of the 
mountain for the same dates. The third and fourth curves 
represent the fluctuations from May 22, 8 P. M., to May 26,8 © 
P. M.; and the fifth and sixth curves represent the fluctuations 
from May 28, 10 a. M., , 10 A. M. minimum of 
May 18th is pretty sharply defined, and occurred on the sum- 
mit four hours later than at the base. On the morning of May 
15th occurred another minimum, which at the summit was 
sharply defined, but at the base was inconsiderable. The min- 
‘imum of May 24th was pretty sharply defined at both stations, 
and the same is true of the maximum May 26th. From May 
28th to June 1st occurred two minima and two maxima whic 
are not very sheen defined, but the critical points at the 
summit evidently lag behind the corresponding points at the 
base. The maximum of June 8th continued for twenty-four 
hours with but little change, and the highest reading occurr 
at the summit of the mountain earlier than at the base, which 
may be regarded as an exception to the general rule. If, how- 
maximum occurred six hours earliest at the base. The decided 
fall of the barometer began six hours earlier at the base than 
at the summit. The minimum of June 28th continued for 
eighteen hours without much change, but the absolute min- 
imum occurred thirteen hours earlier at the base than at the 
summit. 
These observations appear to me to prove that the maxima 
and minima of the barometer do not generally oceur simulta- 
neously at the top and bottom of Mt. Washington, but on an 
average occur more than three hours later at the summit than 
at the base, showing an average retardation of one hour for each 
900 feet of elevation. 
In order to test this result by independent observations, I 
compared the observations made on the summit of Mt. Wash- 
ington with those made at Burlington, Vt, and Portland, Me, 
from September, 1872, to January, 1875. Burlington and Port- 
land are distant from each other about 150 miles, and Mt. 
Washington is about midway between them, and we might 
expect that the barometric minima at Mt. Washington would 
occur at dates intermediate between those at Burlington and 
Portland. I therefore selected all those cases in which the 
barometer at Portland fell as low as 29°6 inches, and deter- 
