NIPHEK MAGNETIC DETERMINATIONS IN MISSOURI, 79. I3I 



•which values have been used in the reduction of the deflection 

 series. 



In reducing the magnetic moment to the mean temperature of 

 the series, a mistake vv^as made in the former report (p. 93) in 

 using the temperatures of the oscillation series ; that of the de- 

 flection series, to which the values had been reduced, should of 

 course have been used. Making the proper correction, the mag- 

 netic moment of Cg, at the mean temperature of 90.3 for the 

 summer of 1878, is represented by the formula 



log. ?« = 9-87402 +O-000064 ^5 



where d represents the number of days, estimated from Aug. 3d. 

 The general remarks made upon this result are therefore still 

 applicable. 



During 1879 the intensity magnet seems to have been in a more 

 satisfactor}^ condition than in 1878. The change (a) in the value 

 of log. 7n being assumed constant, we have 



log. m = log. m,) _ ad . 



Here m =^ the magnetic moment on days of observation being 

 reduced to the mean temperature 84.5.. 

 m^ = the same for Aug. 10. 

 d =z the time in days, estimated from Aug. 10. 



Putting log. z«Q — log. m := y , the residuals take the form 



e = y ^ ad . 

 The most probable value of a, as deduced from the 22 obser- 

 vations, is 



" - + [d^] 



The calculations are given in the adjoining table. 



For reducing the observations to the mean temperature 84.5, 

 the following formula was vised : 



log. m = log. m, [i 4- (/_84.5) ^] , 



where g (the temperature constant) is o.^oQ^g , as determined in 

 1878. From this formula the following table was calculated, and 

 was used in the reductions : 



