224 MR. S. BIDWELL ON CHANGES PRODUCED BY MAGNETISATION IN 



particularly with the smaller currents. Such variations as exceed the specified limits 

 of observational errors may be fully accounted for by small differences of temperature 

 affecting the susceptibility of the iron. 



(2.) The close agreement between the deflections obtained with ascending and with 

 descending currents. As before, changes of temperature would account for the 

 greater part of such differences as exist. 



(3.) In the thirty-three pairs of observations recorded in columns F and K, there 

 is only one instance in which the discrepancy exceeds a single scale division, and in 

 that case the full battery of 30 cells was in use. 



This shows that the degree of precision aimed at was practically attained, at least in 

 the final results, which are the means of two observations. 



The greatest reasonably possible combination of errors in the measurements of the 

 parts of the apparatus and of the distance of the mirror from the scale would not, 

 even assuming that they all conspired together in one direction, affect the results to a 

 greater extent than about 1 per cent. It may therefore be taken as exceedingly 

 probable that the figures recorded in the table correctly represent to one ten-millionth 

 part of the length of the rod the elongations and retractions which actually occurred. 

 The slight variations in the results obtained upon different days or when the order of 

 the currents was reversed are mainly due to differences in the physical condition of 

 the iron (thermal or magnetic), and not to instrumental or observational errors. 



Pig. 5. 



In order that an idea may be formed of the extent of such variations, the diagram 

 fig. 5 has been constnicted, in which the three series of observations have been plotted 

 together, the points belonging to each different series being distinguished by a different 

 kind of mark. Those of the first experiment with ascending currents are marked 

 with crosses ; those of the same experiment with descending currents are marked 

 with crosses surrounded by circles ; and those of the second experiment with ascending 

 currents are distinguished by dots. A smooth curve is drawn through all the marks. 



