64 
While the subdivision of the materials of one large 
machine into a number of small ones was attended by 
several important advantages, yet it was found necessary to 
adopt some means to secure the synchronous rotation of 
the armatures, when the combined direct current from 
several machines was required. 
As the high speed at which the armatures were driven 
precluded the employment of toothed gearing, the only 
method which seemed at all feasible for producing the 
requisite synchronism was, to place several machines in a 
straight line, and connect them together by means of a 
clutch fixed on the end of each armature spindle. It was 
while experimenting with a pair of machines so geared 
together that the author first observed the phenomenon 
which forms the subject of his communication. 
The armatures of these machines were 4 inches in dia- 
meter, and each of them was coiled with a copper wire 
conductor, 280 feet long and one-eighth of an inch in 
diameter. The currents were taken from the armatures by 
means of copper brushes rubbing against metal rings con- 
nected respectively with the ends of the armature coils, 
and were therefore in alternate directions for producing the 
electric light. 
The clutch bv which the armatures were connected con- 
•j 
sisted of two iron discs, about four inches in diameter, 
having in the face of one, two iron pins which could be guided 
into two corresponding holes in the face of the other. 
These discs could be engaged, or disengaged, either when the 
machines were at rest or in motion. The relative positions 
of the pins and holes in the discs were such that the 
armatures might be engaged in reversed positions of half a 
revolution, when required. 
Each of the machines, when making about 2000 revolu- 
tions per minute, was, of itself, capable of producing a very 
efficient electric light ; and when the two armatures were 
