1879.] L. Schwendler— On Electric Light Measurements. 105 
Dynamo-Electric Machine D produces owe weber at 920 meg-ergs 
per second. 
N. B.—Through certain external resistances which are sufficiently 
small to get currents above 24 webers. 
Regularity of the Electric Light — 
If the resistance external to the dynamo-electric machine is repre¬ 
sented by the resistance of the arc only, i. e., resistance of leading wires 
equal “ Nil,” then although the light is naturally the most powerful, it is 
the least steady, since any variation of the resistance of the arc has then 
evidently the largest influence on the current and on the light. 
By connecting across the electro-magnet of an electric lamp, another 
electro-magnet which acts as a shunt, and adjusting the two electro¬ 
magnets in such a manner that they produce equal extra currents when 
variations in the primary current take place, the regularity of the working 
of the lamp is greatly enhanced. 
(See Specification of Lamp, Appendix I, of Precis of Report.) 
An electro-static shunt will have a similar effect. For strong lights 
or strong currents, the electro-magnetic shunt is best ; for weak lights or 
weak currents the electro-static shunt is best. 
The lamp should be constructed mechanically so well and delicately 
that the carbon points run together with a minimum tension of the spring 
of the lamp. Then the lamp will be a balance with small weights. 
When making photometric measurements, to get more trustworthy 
results, it is best to use a flat carbon (2 to 3 thick) as the positive 
electrode, and a carbon of the usual form as the negative electrode. 
The light is to be observed in a line normal to the flat surface of 
upper carbon, and passing through the centre of the arc. In this manner 
the largest quantity of total light produced is measured, and, moreover, 
the ratio between total and external light is more constant. 
The lower carbon should be invariably of less section than the upper 
carbon. 
Further, when producing the light by a short arc, which it is always 
advisable to do, the lower carbon should be natural carbon. Coppering the 
carbon is advisable under all circumstances. Artificial as well as natural 
carbons appear to be very little hygroscopic, a great fortune for damp 
climates like India during the monsoons. 
When the arc is long, the flame by combustion of the carbons is large. 
This appears to be due to the fact that for a long arc the vacuum 
formed round the carbon points by expansion of the air by beat is less 
complete than in an arc of shorter length. 
