DISCUSSION OF THE VERTICAL FORCE. 



College building and at the observatory. This value being known, I considered 

 that the value of the scale of the new reflecting magnetometer could best be ascer- 

 tained by comparison with the former. The result of this, continued at intervals, 

 was, that two divisions of the new scale were equivalent to one of the old, or that 

 a change of one division of the reflecting instrument corresponded to a change of 

 vertical force of 0.000033 parts. This was after the instrument had been finally 

 adjusted. 



The only disadvantage in the new instrument was the large effect of changes of 

 temperature upon it; by direct observations it was found that a change of 1 (F.) 

 of temperature produced a corresponding change of 13.5 + 0.25 scale readings, 

 whereas in the Lloyd instrument the corresponding change was but 3.12 scale 

 divisions. We have accordingly for the Lloyd instrument 5=0.0000515, and for 

 the reflecting instrument 5 0.000446. The values actually used in the reduction 

 of the observed reading to a standard temperature will be seen further on. 



The importance of ascertaining the most correct and suitable coefficients of tem- 

 perature for the two series of observations, demands a more detailed statement and 

 elaborate discussion of the observations themselves independently of the special 

 trials. Experience has shown that the value for q deduced from the differential 

 intensity observations themselves, with the magnet subject generally to gradual 

 and small changes of temperature, is smaller by a considerable fraction than the 

 value found by direct and special observation during which the temperature changes 

 are necessarily more violent. There is no doubt that in the reduction to a standard 

 temperature that value of q should be used which was obtained while the magnet 

 was under its ordinary influences and condition. The same view is taken by Gene- 

 ral Sabine, and was also carried out in the discussion of the horizontal component 

 of the magnetic force; for which see the preceding paper (Part IV). 



Determination of the Effect of a Change of Temperature on the Readings of the 



Vertical Force. 



(A.) Results of special observations made for determining the temperature coeffi- 

 cient. The correction for temperature of the Lloyd vertical force magnetometer was 

 ascertained by the usual method of vibrating the bar when suspended horizontally, 

 and when alternately heated and cooled artificially. The thermometer was placed 

 with its ball near the axis of the magnet. The changes of the horizontal force 

 magnetometer, while these experiments were going on, were noted and allowed for. 



which is equivalent to 3.15 scale divisions; in the first reduction of the record 3.12 

 was used. 



