446 FE. Loomis—Barometric Gradient in great storms. 
tive. The numbers in column 12th denote the gradients com- 
puted by using only the first term of the formula. The mode 
of obtaining the other columns will be explained hereafter. 
This table readily suggests several important conclusions, but 
I will defer considering them until I have obtained the corre- 
sponding results for the United States. : 
The lower part of Plate I gives a graphic representation of 
the numbers in the preceding table. At the center of low 
of high pressure where the barometer stands at 792™. e 
ceding shows the corresponding gradients from the low center 
to the high center. The distance from the low center to the 
high center is 29°31 degrees. The entire figure represents @ 
length of 58°62 degrees, or 4,045 English miles. 
In order to obtain the average elements of the violent storms 
of the United States, I employed the Weather Maps of the 
Signal Service. I have a complete set of the tri-daily weather 
from Nov., 1871, to the close of 1880; and I have also 
one daily map from the beginning of 1881 to the present time, 
making in the aggregate about 11,000 maps. isobars on 
these maps show much greater irregularities than those upon 
Hoffmeyer’s charts for the Atlantic Ocean. This difference 
may be ascribed in part to the greater number of observations, 
and the attempt to inv the isobars so as to be consistent with 
all the observations; but a large part of the difference appears 
to be due to local disturbing influences which are more nume 
rous over a continent than over the ocean. In order to obtain 
suitable data for testing Ferrel’s formula, we must in some way 
eliminate the influence of these local disturbing causes. If we 
should take an indiscriminate average of the gradients, wind 
