606 



Proceedings of the Asiatic Society, 



[No, 5 



apparent. The value of Innar tables to the seaman, that of a geological 

 map to the miner, or of a chemical analysis to the manufacturer 

 depends upon their respective trustworthiness. If they cannot be trusted 

 they are worthless ; and the data, which it is the object of Meteoro- 

 logy to supply, in no way differ in this respect from those furnished by 

 the Astronomer, the Geologist, or the Chemist. Among the more 

 important indications of Meteorological data are the amount of rainfall, 

 and the variation which this undergoes as cultivation increases or as 

 forests are cleared ; the causes of local and epidemic disease, which 

 although much wrapped in obscurity, may not improbably be in part 

 dependent on the dampness of the atmosphere, the absence of ozone, 

 the prevalence of particular winds, &c. ; and the prognostication of 

 storms, or of seasons of drought or "unusual rainfall. Such phenomena 

 are indeed only in a few cases capable of control, but it is only neces- 

 sary to point to the results attained by Admiral Fitzroy, to prove 

 that, when forewarned, we may be able in a great number of cases 

 either to avoid or diminish their more disastrous effects. But in order 

 that any of the laws of these phenomena may be determined, so that 

 they may be acted upon with confidence, it is essential that the 

 observations from which they are deduced be reliable and accurate, 

 The observations of many successive years must, in most cases, be 

 recorded, in order that the laws of recurring atmospheric changes, and 

 the effects of those changes on. agriculture, health, &c,, may be ascer-r 

 tained, and the observations taken at different times and places must 

 be capable of strict comparison. It is clear that no loose system of 

 record will admit of this ; and indeed the very knowledge that a series 

 of observations had been made by an incompetent observer, or with 

 instruments not strictly trustworthy, would at once be sufficient to 

 warrant their rejection, when, as in Meteorology, the increments of 

 variation are so small, that the error of observation will in many cases 

 conceal or neutralize, if it does not absolutely invert, their true law of 

 succession. Even if, at one and the same station, the conditions of 

 error are sp constant that a result true in the main is obtained, when 

 the observation of different periods are compared, this will be the 

 Utmost attainable ; and the observations are neither comparable with 

 those tafcep. jindej* different conditions elsewhere, nor can they in any 

 pase be accepted with that con|idence which alone will give them value, 



