Lifting Power of the Electro-Magnet. 307 



serious impediment to electro-magnetic engines), I give the set which first de- 

 cidedly convinced me of its influence. 



Time = 10" . . . Z = 20526 . . . F= 0-2751 

 6 19110 0-2779 



13 213 52 0-2655 



8 194-05 0-2603 



12 207-62 0-2568 



The increase of time more than compensates for a considerable diminution 

 of current. I have regulated the duration of each set according to -what I con- 

 ceived to be a sufScient allowance of time. 



4. These causes are uniform in their action, and can be avoided or corrected, 

 but there exists another, which is the chief source of error in these experi- 

 ments, namely, the molecular change which iron suffers when exposed to power- 

 ful magnetization. In consequence of this, however pure and soft it may be, 

 it becomes capable of retaining permanent magnetism, and in the same proportion 

 less susceptible of excitation by its helices. This magnetism (which I call X) is 

 variable ; it may, perhaps, be intense while the magnet is excited, but on lifting 

 the keeper it declines rapidly till it attains a certain amount, which is, however, 

 not invariable ; and it always increases during a set, though after a few hours 

 it returns to its ordinary quantity. It, however, occasionally happens that 

 when tlie magnet has been powerfully excited for many days, its iron becomes 

 disturbed in this respect, and then the values of L fall far short of their legiti- 

 mate magnitude. In such cases it is best to leave it at rest for a few weeks ; 

 but I have found that if the current be reversed the L becomes higher, and 

 have therefore in many instances performed this for the alternate measures. 

 The results thus obtained are tolerably uniform, but are always less than those 

 given by a magnet that has never been excited, or has been long in repose. 

 With excitation less than what is given to this magnet by a current = 1, this 

 cannot be done, because then it will be seen from the Table that there is a real 

 difference between the L's produced by the direct and reverse currents. One 

 consequence of this change deserves notice, which may be observed in almost 

 every series, — the gradual decrease of the successive measures of a set. Thus 

 in one taken with peculiar care, 



VOL. xxu. 2 s 



