-1859] WRITINGS OP JOSEPH HENRY. 29 



statement may be made in regard to the other elements of 

 meteorology, and the result of all the observations may be 

 divided into two great classes, periodical and non-periodical, 

 though by a very long series of observations, it may happen 

 that a phenomenon which at first may appear entirely fit- 

 ful, will afterwards prove to be recurring ; and at all events 

 the non-periodic variations are found to be restricted within 

 definite limits, the maximum amount of which it is highly 

 necessary to obtain. 



The first element given in the tables is that of the mean 

 height of the barometer from month to month. This is 

 perhaps less immediately essential to the agriculturist than 

 any other meteorological element. It is however of much 

 importance in determining the progress of storms and the 

 area over which the commotions of the atmosphere con- 

 nected with them are perceptible, though no violent dis- 

 turbances may be observed. For example, if the barometer 

 on a given day is higher or lower than the average for the 

 month, we are then convinced that it is subjected to some 

 unusual perturbation; and by drawing a line on a map 

 through all the places at whicli a given amount of disturb- 

 ance is felt at a particular time, we are enabled to trace the 

 boundary of a storm, and to indicate its progress, develop- 

 ment, and end. For this purpose it is not necessary even 

 that the barometers should be strictly comparable with each 

 other ; it is only necessary that the results should be com- 

 parable among themselves. When the barometers have 

 been accurately compared with each other, (as in the case of 

 those of Green, of New York, constructed under the direc- 

 tion of the Smithsonian Institution,) they afford the data for 

 determining the relative elevation of different places of ob- 

 servation above the level of the sea. 



The indications of the barometer, compared with those of 

 the hygrometer, thermometer, and wind-vane, furnish us 

 with a method of predicting changes in the weather. These 

 however in many cases will be found to depend upon rules 

 applicable to particular places, and which can only be de- 

 termined by a long series of local observations. 



