THERMAL PERIODS. 



4. The annual thermometric period also varies with the latitude, 

 and with all the other conditions that affect the thermal phenomena. 



In order to he enabled to evolve the general thermal laws from 

 phenomena so complicated and shifting, it is above all things 

 necessary to define and ascertain those mean conditions or states, 

 round which the thermometric oscillations take place. 



5. The mean diurnal temperature is a temperature so taken 

 Dehveen the extremes, that all those temperatures which are 

 superior to it shall exceed it by exactly as much, as those which 

 are inferior to it shall fall short of it. 



This mean temperature may always be obtained by taking the 

 sum of the temperatures at sunrise, at 2 P.M., and at sunset, and 

 dividing the result by 3, or more simply still, by adding together 

 the maximum and minimum temperatures, and taking half their 

 sum. Whichever of these methods be adopted, the same result 

 very nearly will be obtained. 



6. The mean temperature of the month is found by dividing the 

 sum of the mean diurnal temperatures by the number of days. 



7. The mean temperature of the year may be found by dividing 

 the sum of the mean monthly temperatures by 12. 



It is found that in each climate there is a certain month of 

 which the mean temperature is identical with the mean tempera- 

 ture of the year, or very nearly so. This circumstance, when the 

 month is known, supplies an easy method of observing the mean 

 temperature of the year. 



In our climate this month is October. 



8. The mean annual temperature being observed in a given 

 place for a series of years, the comparison of these means, one with 

 another, will show whether the mean annual temperature is subject 

 to variation, and if so, whether the variation is periodic or pro- 

 gressive. All observations hitherto made and recorded tend to 

 support the conclusion, that the variations of the mean annual 

 temperature are, like all other cosmical phenomena, periodic, and 

 that the oscillations are made within definite limits and definite 

 intervals. 



But even though the period of these variations be not known, a 

 near approximation to the mean temperature of the place may be 

 obtained by adding together any attainable number of mean annual 

 temperatures, and dividing their sum by their number. The pro- 

 bable accuracy of the result will be greater, the less the difference 

 between the temperatures computed. 



Thus it was found by a comparison of thirty mean annual tem- 

 peratures at Paris, that the mean was 51'44, and that the 

 difference between the greatest and least of the mean annual 

 temperatures was only 5 '4. It may therefore be assumed that 



*2 67 



