[23] 



204 



laws, and that a knowledge of these laws cannot fail to contribute to the 

 weahh and happiness of mankind. 



I proceed, therefore, to inquire — 



IV. Upon what plan the observations should be conducted to secure the 

 object proposed. 



Violent winter storms appear most suitable for investigation, because 

 they are of longer duration, and their features are more strongly marked. 

 The oscillation of the barometer affords the surest criterion for identifying 

 a storm in its progress from day to day; and these oscillations are greatest 

 in winter. 



1. How large an area shonld be covered by our observations, to enable 

 us to investigate advantageously the phenomena of our winter storms ? 



Our observations ought plainly to embrace the entire region of the 

 storm, and even extend somewhat beyond its margin, for in no other way 

 could we be sure that we had found its limits. Now, the great storms 

 which are experienced between the parallels of 40 and 45 degrees are fre- 

 quently felt as far south as latitude 30'^, and sometimes to 25°. South of 

 this line, the oscillations of the barometer, except in a few rare instances, 

 are quite small; and it is probable that the limiting parallel of the trade 

 winds forms a dividing line between the ordinary storms of the torrid and 

 temperate zones. We should therefore extend our system of observations 

 to the southern margin of the United States. The northern limit of our 

 ordinary winter storms remains unknown. In the storm of December 21, 

 1836, the oscillation of the barometer increased uninterruptedly with the 

 latitude as far as the most northerly station, (Quebec. Only the southern 

 half of the storm was included within the United States. This storm was 

 probably experienced as far northward as to the northern shore of Hudson's 

 bay. VVe should therefore extend our system of observations to the north- 

 ern margin of the United States; and to render our system complete, re- 

 quires the co-operation of the British government to extend the observa- 

 tions to the entire region of Hudson's bay. 



If we merely wished to embrace in our observations the area of a storm 

 for a single hour, then the interval between the Mississippi and the Atlan- 

 tic would ordinarily be sufficient. But we could not thus expect to dis- 

 cover the origin of a storm. We must trace it in its progress from its com- 

 mencement to its greatest violence, and thence to its decline. By observing 

 under what circuntstmices it takes its rise, we may hope to be able to dis- 

 cover the cause of its activity. AVe should therefore aim to trace every 

 storm from its origin to its close. At the ordinary rate of progress a great 

 storm would travel from the Rocky mountains to the Atlantic in two days. 

 "What influence this range of mountains may have upon our storms is un- 

 known, except from conjecture. VVe cannot suppose that a storm could 

 travel from tiie Pacific to the Atlantic without experiencing some modifi- 

 cation in passing over a range of mountains rising into the region of per- 

 petual snow. What this modification is, can be determined by corres- 

 ponding observations on both sides of the mountain. I therefore conclude 

 that it is important to embrace in our system of observations the entire 

 continent from the Pacific to the Atlantic, and from the gulf of Mexico to 

 the northern shore of Hudson's bay. 



2. At what distance from each other should stations of observation be 

 selected ? 



If we were investigating the phenomena of a summer shower, it would 



