496 Proceedings of the Royal Society 
observations I have made, tbe sum of these for the two coils sepa- 
rately was approximately equal to that obtained when both currents 
were sent into the galvanometer circuit. To give an idea of how 
nearly this comes out, I may cite one observation repeated three 
times in succession with the same result. I found the angle of 
both together to be 47|°, that of the electric coil separately 40°, 
and that of the magnetic coil separately 34°. Now the square of 
the tangent of 47 J° is 1*1909, and the sum of those of the other 
two 1*15905. 
The theory of the machine, as I understand it, may be thus 
shortly summed up. In one case, namely, that of an easy common 
circuit, and it is likely to be more or less so in all cases, the two 
coils contribute each their full quota to the total electric fund of 
the armature. When the resistance of the circuits differ, this fund 
is divided inversely in some function of the relative resistance, but 
whether this takes place so as to excite the electro-magnet at no 
original expense of driving energy is still a matter for further 
determination. The results got from the machine would lead us to 
suspect as much, for they compare favourably with machines where 
a permanent battery of magnets is used; hut this test, though so 
far satisfactory, is far from exact. 
The interference of the coils seems to me to be a hopeful feature 
of the arrangement, as it does not make increased power simply 
dependent on increased velocity. There is a promise in it that by 
adjusting the relative sizes of the coils a powerful current may be 
got at a really practicable speed, and there would thus be obviated 
the serious objection to this class of machines, which, however 
astonishing in their power, are apt to wear themselves out by their 
rapid rate of motion when kept in action for days together. Even 
in the machine before you, if the collars were properly turned and 
centered, so as to give good contact with the springs at all rates of 
revolution, I have reason to believe that its effective speed of 
revolution would be very much diminished. 
In mentioning what a machine like this can do, considerable 
latitude must be understood in interpreting results. The strength 
or ardour of different workers may tell very differently. The only 
fair way would be to give the electric effect corresponding to a 
weight falling so far per second, hut this involves opportunities of 
