89 



usually g:reater than /, but whoso relative value may by preference 

 be determined by referring each pair of conjugate thermometers to 

 an adopted standard pair. Ferrel's formula is especially devised 

 for thermometers having spherical bulbs, measures made by it at high 

 and low temperatures give results that are comparable with each 

 other; for absolute results the numerical coefficients may need some 

 modification, but as it stands it gives the values of q approximately 

 in calories per minute per square centimeter. 



Omitting for the present the factor /.' in FerreFs formula, which 

 must be specially applied for each thermometer, we have the values 

 of q in calories as given in the following table (see Ferrel, p. 37), 

 which also presents the corresponding values given by the formula of 

 Marie-Davy in actinometric degrees. In a critical study of observa- 

 tions reduced by these two methods we have to recall that Marie- 

 Davy's actinometric degrees are really fractions of a calorie, or units 

 of heat so small that 100 of them are equivalent to the absolute radia- 

 tion of the sun received at the outside of an atmosphere whose coeffi- 

 cient of transmission is 0.875; wdiereas Ferrel's calories have been 

 adopted without predicating anything as to the solar radiation or 

 atmospheric absorption, concerning wdiich his observations show that 

 the solar radiation constant is between 2 and 2.25 calories per minute 

 per square centimeter and the atmospheric coefficient of transmission 

 to be used with the conjugate bulbs is 0.72. 



Solar radiation deduced from oh ser vat ions loith the conjugate thennoiiteters. 



INTENSITY AND DURATION OF SUNSHINE AT MONTSOURIS. 



In order to have at hand data that will enable one to approximately 

 infer some of the relations between the temperature of the air and 

 of the soil and of the solar radiation, one may consult the tables for 

 the observations at Montsouris, given by Marie-Davy in his Annuaire 



for 1887. 



