THE SMITHSONIAN INSTITUTION. 243 



want of having thus been elahorated that voluminous collections of 

 observations, the fruits of long years of toil, remain useless and for- 

 gotten in the dust of libraries, because the meteorologist finds it im- 

 possible to make use of them v/ithout first undertaking those calcula- 

 tions, the amount of which absolutely transcends the powers of an 

 individual, and would discourage the most ardent zeal, while they 

 would have cost the observer only an instant each day, if he had made 

 them at the time of the observations. 



The calculations desirable are a,s follows : 



1. Each barometrical observation must be reduced immediately to 

 the temperature of zero Centigrade, or 32° Fahrenheit, by means of 

 the tables, and the total correction of the barometer, if there is any, 

 will be applied. 



2. The diurnal means of the several instruments, resulting from the 

 sum of the three observations made at these different hours, divided 

 by three, must be entered each day in the respective columns, after 

 the observation of 10 p. m., [9 p. m.] It is needless to say that these 

 means should be drawn solely from observations reduced and corrected. 



3. The monthly means for each hour separately — that is, the monthly 

 mean of the observations of 6 a. m., [7 a. m.,] and that of 2 p. m., and 

 of the observations of 10 p. m., [9 p. m.] 



4. The monthly means drawn from the means of each day; the 

 monthly extremes of the instruments; the monthly amount of the 

 rain, hail, or snow; the mean cloudiness of the sky; the prevailing- 

 wind, &c. 



5. The annual means and amounts, and the respective extremes for 

 the civil year. 



It will be interesting to calculate also, if the observer is so disposed^ 

 the mean of the seasons of the meteorological year_, which begins De- 

 cember 1, to November 30, of the following civil year; 



The meteorological seasons are, then : 



Winter — December, January, February. 



Spring — March, April, May. 



(Summer — June, July, August, 



Autumn — September, October, November. 



In calculating all these difierent results, we should take, in order 

 to be very exact, the means of the sums of all the observations during 

 the period of time in question, by reason of the inecjuality of the 

 length of the months. 



The sums which form the basis of all these means should be in- 

 scribed in the tables in the place reserved for them. 



The preceding calculations, after a little practice, will not appear 

 difiicult, and may be quickly performed; but it can hardly be too 

 often urged upon the observer to make them without delay ; other- 

 wise, this task, which is slight if accomplished daily, would become 

 very heavy, if left to accumulate for several months. It is only by 

 maldng the correction himself that the observer can institute his own 

 comparisons, and really study the course of the meteorological phe- 

 nomena. His interest will increase still more with the feeling that 

 he is cooperating in a great work, which concerns at once his whole 



