873 
compared with the ascending part in the oscillographie record, we 
see that they correspond fairly well. The constructed curve shows 
f deter Pues 
is Rl eas opi fi 
1 
Fig. 3. 
a somewhat more rapid ascent in its first part, and also some difference 
in the last part. But this can readily be explained. If we bad calcu- 
lated the current strength, taking into account that the selfinduction 
was greater at the beginning and at the end of the make period 
and smaller in the middle, the curves might have agreed better 
numerically : theoretically this point is of little or no interest. 
The descending part of the curve, which embraces the two last 
quarter periods, represents the magnetic attraction during the break 
period. A quantitative explanation is as yet not possible, though 
qualitatively there seems to be no difficulty. We know that the less 
reluctance there is in the magnetic circuit, the longer will an electro- 
magnet keep its magnetism after breaking the current. Immediately 
after breaking the current the air-path is rather large and conse- 
quently the reluctance is great and the magnetism disappears rapidly. 
As the armature approaches the core, the magnetic circuit improves 
and the magnetism disappears more slowly. The slope of the curve 
is indeed least at the end of the 3'¢ quarter period. From then to 
the end of the 4" quarter period the reluctance grows and the 
descent becomes more rapid again, becoming nearly as fast as in the 
commencement of the 3'¢ quarter period, though not quite, as at that 
moment the direct action of the magnetomotive force is taken away. 
57 
Proceedings Royal Acad. Amsterdam. Vol XXIII. 
