

«. 



# 



f 



tf 



i'^r 



***** 



4 



P 



Sect. IX.] 



METEOHOLOGY 



c> 



95 



( 



of radiation var 



IPC witli 



A'^-fw' T^ i. VAX 



the glass used. 



To reduce provmonally a set of 



rfVt 



ohserva 



#^ 



tions 



If the set consist of orly four or five sun obser- 

 vations, with intermediate shades, take the mean of the 

 " changes per minute " in column 4, for all the sun and 

 for all the shade observations separately, attending duly 

 to the signs. Change the sign of the latter mean, and 

 add it to the former. The aggregate will be the un- 

 corrected radiation in parts of the scale. To correct it 

 for the unequal dilatability of the liquid, take the mean 

 of the temperatures shown by the internal thermometer at 

 the beginning and end of the set, and with it enter the 



Table V., which contains the factor by 



^^ppendix 



which the uncorrected radius is to be multiplied. 



If the 



series 



consist of more than a quadruple or sextuple 

 observation, it must be broken into quadruplets, or quin- 

 tuplets, and each reduced separately as above. 



The abstract unit of solar radiation to be adopted in 

 the ultimate reduction of the actinoractric observations 

 is the actine, by which is understood that intensity of solar 

 radiation which at a vertical incidence, and supposing it 

 wholly absorbed, would suffice to melt one millioTitli part 

 of a metre in thickness from the surface of a sheet of 

 ice horizontally exposed to its action per minute of mean 

 solar time ; but it will be well to reserve the reduction of 

 the radiations as expressed in parts of the scale to their 

 values in terms of their unit until some future and final 

 discussion of the observations. 



Meanwhile no opportunities should be lost of comparing 



fe. 



