ABOUT THE 20tH OF DECEMBER, 1836. 149 



nervations have been received, this appears to be the smallest fraction ever re- 

 corded. Add to this that the observations are generally taken but twice a day, 

 and the precise hour to which the observation corresponds is seldom given, and 

 we shall be prepared to expect considerable irregularity in comparing the ob- 

 servations at a hundred different stations. Finally, it is but reasonable to bear 

 in mind that in the table of winds which I have given there may be actual 

 errors of transcribing. The materials from which the table was prepared are 

 copies, most of them at second hand, from the original records. In transcribing 

 such a list of unmeaning letters, where the copyist can receive no assistance 

 from the context to guide his judgment, slips of the pen are particularly liable 

 to occur. 



I trust it will not be inferred from these remarks that I have a theory of 

 winds to which all the observations are to be made forcibly to conform, and 

 that whatever seems obstinate is to be pronounced an error of observation. I 

 wish merely to show that we should look at the general or average tendency of 

 the winds at the several stations, and when one observation differs essentially 

 from all the rest, it is not unphilosophical to regard it with distrust. That 

 these remarks are not uncalled for will appear evident from comparing the dif- 

 ferent registers kept at almost identically the same station; for example, the 

 two at St. Louis, and that at Jefferson Barracks; those at St. Augustine and 

 Fort Marion; and those at New York, and its immediate vicinity. 



It may be well to enumerate here the principal known causes of wind. It 

 is believed that winds may commonly be referred to one of the three following 

 causes, viz: 



I. Inequality of atmospheric pressure. 



II. Unequal specific gravity of air. 



III. Rotation of the earth. 



Conceive two vertical columns of air connected by a horizontal canal. If 

 the weight of one column exceeds that of the other, it must preponderate. The 

 wind, therefore, must hXow from places where the barometer is highest, towards 

 those where it is most depressed. Moreover, if the two columns of air sup- 

 posed be of equal weight, but unequal specific gravity, there will be effected a 

 new distribution of the particles of the two portions. The denser will flow 

 under and displace the lighter. This inequality of specific gravity may arise 

 either from a more elevated temperature or an excess of aqueous vapour. Even 

 VII. — 2 N 



