Observations of the United States Signal Service. 11 
Only those cases were selected which contained an isobar as 
low as 740 millimeters. The isobar selected for measurement 
was seldom the lowest isobar drawn on the maps, but the 
largest isobar which was complete (or nearly so) about the storm 
cente number of cases employed was 70, and the aver- 
age ratio of the two axes was 1°60. Column 6th in the precedin 
table shows the number of cases for each 10° of azimuth; an 
column 7th shows the averages of the same numbers taken in 
sets of three. 
In a similar manner the isobars about areas of maximum 
pressure were measured. Only those cases were selected which 
contained an isobar as high as 775 millimeters, and frequently 
this was the isobar selected for measurement; but if the map 
rs taken in sets of three. 
The following table presents a summary of the preceding re- 
sults both for low and high pressures in the United States and 
urope. 
Summary of Results for Isobars. 
LOW BAROMETER. || HIGH BAROMETER. 
| United States.) Hurope. || Unite States.| Europe. 
Ratio of the two axes, 1:94 | 1-60 i 191 | 182 
Prevalent direct. of major axis,’ N.39°R. |N.3i°E.!| N.44°E |N. 76° B. 
This table suggests some obvious reflections, but I prefer to 
withhold them until I have obtained a longer series of observa- 
tions from Europe. 
Relation of rainfall to variations of barometric pressure. 
In former articles (this Jour., vol. viii, p. 4 and vol. x, p. 5) T 
have shown a close connection between the rain-fall and the 
direction and_ velocity of a storm’s progress. I have also en- 
have discovered a decided connection between the amount of 
rain-fall and the pressure at the center of the storm. For the 
purpose of comparison, storms were divided into three classes ; 
one class including those cases in which the barometric depres- 
sion at the center of the storm was the same on two successive 
