( 298 ) 



To fuid out wlietlier these deviations are perhaps due to irregular 

 errors in the measurement, it will he useful to investigate wlietlier, 

 by addition of anotlier term, the chtferenees between these obser- 

 vations and tiie calculation might be reduced to within the limits 

 of the erroi's of observation. It succeeded indeed fairly well as may 

 l)e seen in column Obs. — Comp.ü of table VI. The calculated values 

 are derived by means of the fonnula 



w — J 10.045 (1 + 0.00391(^7 / — 0.000 000 3432 f + 

 -f 0.000 000 0020(59 f). 



The deviations tVoni forunda 11, with the exception of tJic last, 

 although they are iiol eulirely within the limits of tlie erroi'S of 

 observation wliicli were exj)ected, are oidy little in excess. 



In the case of methaue, where the deviation is 19 on 38674, an 

 eiToi- in the lemporalorc^ of 0°.04 C. is sufti''ient to explain this 

 amoiinl. 



