88 
L. Schwendler— On 'Electric Light Measurements. [March> 
It is, however, necessary to state that this relation only holds good as 
long as one dynamo-electric machine produces one electric light. 
The three dynamo-electric machines B, C, and D I found practically 
equal, i. e., they produce the same quantity of light by the same expendi¬ 
ture of power. Any difference which may be found is entirely within the 
limits of observation and other errors. 
But the dynamo-electric machine A gives a much stronger light for 
a comparatively smaller expenditure of power. 
In round numbers it may be said that dynamo-electric machine A 
gives about double the quantity of light given by any of the other three 
machines, and that only about half as much power is expended to produce 
the unit of light. 
This favourable result is principally due to the comparatively small 
internal resistance of A, and its low speed. 
With reference to mechanical construction and regular working, 
dynamo-electric machine C is highly efficient. 
My experiments necessarily have suggested alterations and improve¬ 
ments of the dynamo-electric machines, with which, however, I shall not 
deal in this precis .of the Report. 
II.— Constancy and Regularity or the Electric Light. 
This appears to be still the weak point, and many improvements in 
this respect are possible and desirable. 
Although it is not in the nature of dynamo-electric machines to 
produce a perfectly steady current, like, for instance, galvanic batteries, 
the irregularity of the light resulting from this source can be entirely 
neglected against those produced by the lamp itself. 
I have tried two different lamps :— 
(а) The Serrin Lamp, as supplied by Messrs. Soutter & Lemonnier, 
of Paris, and the British Telegraph Manufactory. 
(б) The, Siemens Lamp , as supplied by Messrs. Siemens, Brothers. 
The Serrin lamp, for any given adjustment, regulates the length of 
the arc only in one direction, i. e., it diminishes that length. 
The actual consumption of the carbon points regulates the length of 
the arc in the other direction, i. e., increases it. 
In the Siemens lamp the decrease of the length of the arc is effected 
exactly in the same manner as in the Serrin lamp, but the increase in the 
length of the arc is not only left to the consumption of the carbon points 
—a comparatively slow process—but is accelerated by the addition of a 
make and break arrangement, which separates the carbon points. 
Hence, from a theoretical point of view, the Siemens lamp is un¬ 
doubtedly superior, since the length of the arc is rapidly adjusted in both 
