[ 23 ] 200 



we are furnished with the means of measuring the rate of progress of the 

 great atmospheric wave. On the southern border of the United States, this 

 velocity varied from 17 to 29 statute miles per hour; and on the northern 

 border from 17 to 37 miles. The leading characteristics of this storm were 

 as follows: After a clear and cold interval, with barometer high, the wind 

 commenced blowing from the south. The barometer fell rapidly; the 

 thermometer rose; rain descended in abundance. The wind veered sud- 

 denly to northwest, and blew with great violence. The rain was succeeded 

 by hail or snow, which continues but a short time. The barometer rises 

 rapidly; the thermometer sinks as rapidly. These changes are experienced 

 progressively from west to east. 



This storm was not circular. The area of rain and snow was about 

 500 miles broad from east to west. Its length from north to south was 

 known to be 800 miles, and probably was not less than 1,500. For nearly 

 a day before the crisis, the wind blew from the southern Cjuarter, and gen- 

 erally for several hours h'om the southeast. After the mininmm of the 

 barometer, the wind blew with great violence from nearly the opposite point; 

 commonly the northwest. Here was clearly indicated a prevalent tendency 

 of the wind towards a central line: but, unfortunately, the observations em- 

 braced only one-half the area of the storm. Tlie oscillatioir of the barom- 

 eter showed a steady increase from latitude 25° to Quebec. The centre 

 of the storm, therefore, could not have been south of Quebec, and north 

 of this place we could obtain no observations. 



I. was now desirous of investigating a storm of marked characteristics 

 which could be entwely surrounded, so that more of its features would 

 need to be supplied by conjecture. Two storms which occurred in Febru- 

 ary, 1842, appeared tolerably well suited to my purpose, and were selected 

 for a new investigation. Great pains were taken to collect materials from 

 every part of the United States. I succeeded in obtaining barometric ob- 

 servations from 64 different stations. I also procured registers^ without 

 barometric observations, from 41 military posts, and 22 other stations, ma- 

 king 127 in all, not including registers from several stations too remote to 

 be of any service in the proposed investigation. 



I commenced the analysis of the storm of February 1-5, by following 

 the same course I iiad pursued with the storm of 1836, via: rep'esenting 

 the barometric observations by curves which should exhibit the fall and 

 rise of the mercury, with the time of minimum height; and from the time 

 of minima I proposed to deduce the progress of the storm. But I was 

 disappointed in my expectadons. 1 did not obtain such a uniform rate of 

 progress as I had anticipated. After some time I perceived that ray observa- 

 tions embraced two centres of disturbance; that I had got at least two 

 storms in close juxtaposition, and interfering with each other. I then dis- 

 covered that my lines of barometric minimum represented relations which 

 were extremely complex, and were not well adapted to my purpose, which 

 was the development of physical causes. I therefore sought for some mode 

 of graphically representing the observations which should be foimded upon 

 simple relations, and be better adapted to suggest the causes of the phenom- 

 ena. I at last settled upon a method wliich appears to me well suited to this 

 purpose, and substituted for lines of 'minimum, pressure, lines o{ equal pres- 

 sure. Having determined, as well as 1 was able, the mean height of the 

 barometer at each station, I compared each observation with the mean. I 

 then drew a line upon a map of the United States, passing through all 



