SOLAR ECLIPSE, MAY 0, 1883. 



69 



It first becomes neces«ary to compute the constant c. From (3) we have — 



e "' 



*' 0, 



M — M 



111 order to oI)tain Mie value of c under varying' conditions, observations were selected IVoiii 



difierent days and at dillerent times of tlie day. Tlic, following table contains the observations 



used and the resulting values : 



Computation of c. 



Diito. 



April 27 

 28 

 30 

 30 

 30 

 I 

 2 

 2 

 2 

 2 

 2 



M.ay 



Time. 



9. 00 a. 111. ; Vi. ir> p. ni. 

 9.35 a. 111. to 2. (IS p. m. 

 3. OOji. in. to .'•>. (lOp. n;. 



7. 00 a. III. ; 5. 00 p. in. 

 9. 00 a. 111. ; 3. 00 p. ni. 



(i. ri.^> a. 111. to 10. 30 a., m. 

 10.30 a. ni. to 2. 30 p. in. 

 11.30 a. ni. to 12.3(1 p. in. 

 7.00 a. in. ; 5.011 p. in. 

 10.00 a. 111. ; 2. 00 p. ni. 



8. 00 a. III. ; 4. 00 p. in. 



The resulting values of c arc quite discordant, and have a marked progression according to 

 the values of 0, or according to the hour of observation. Tliey show that in the observations 

 under discussion c is not strietlj^ a constant, but varies with the time of day. Tiiis may be 

 true of the particular instruments used, or be due to some circumstance connected with the 

 exposure of the instruments in these observations, or it may be inherent in this method of meas- 

 uring solar radiation. Without further experiment it is impossible to discover the cause. It is 

 therefore necessary to inquire what error is introduced by assuming a constant value of c. For 

 his investigation the value c = 1.75 was assumed, and a computation of o' made for observations 

 on May 1 and 2. The following table gives the result of the computation and a comparison with 

 the observed values of the same quantity : 



Comparison of computed with observed shade temperatures {S'). 



MAY 1. 



