222 Mr. W. H. Preece. Behaviour of Glow-Lamps [Mar. 26, 



The direction of the current is shown in the figure. 



Towards the end of the experiment, when the characteristic diffused 

 blue effect in the globe was very marked, a bright arc was observed 

 to be playing about the bottom of the limb at c, and it was quite 

 clear that a bridge of conducting material was formed between e and 

 c, which, together with the galvanometer, made a shunt to the fila- 

 ment. 



This experiment was repeated upon different lamps, and the results 

 were so similar that it is not necessary to reproduce the observa- 

 tions. 



In all cases an intimate connexion was observed between the blue 

 effect and the appearance of the shunt current. 



4. One of Mr. Edison's assistants showed me in Philadelphia that 

 while the effect was very perceptible when the connexions were made 

 as shown in fig. 1, a being in connexion with the positive and e with 

 the negative pole, no shunt current or a very slight one could be 

 observed when the direction of the current was reversed. I did not 

 find this invariably the case. One lamp (No. 8) only — a long filament 

 lamp (150 mm. long) — which gave a marked shunt current and the 

 blue effect, when the connexions were made as in fig. 1 failed to 

 give any current or blue effect when the current was reversed within 

 the limits of the electromotive force at my command. Doubtless I 

 should have got both effects if I could have raised the electromotive 

 force. In all cases, however, the effects appeared sooner and were 

 more marked when the connexions were as shown in fig. 1 than when 

 the direction of the current was reversed. The effect of placing 

 the galvanometer between e and c was the same as reversing the 

 current. 



5. As the effect might be due in some way to the material of the 

 conducting plate (de) inserted between the limbs of the filament, 

 Mr. Edison made for me lamps with copper, iron, and carbon 

 plates. 



The following experiments (pp. 223 and 224) were then made, the 

 connexions being the same as fig. 1. 



No marked difference was thus observable. 



If we examine the shunt current when the faint blue tinge appeared, 

 it was : with carbon, 3'42 ; with iron, 5*85 ; and with copper, 3*80 

 milliamperes. 



6. It might be affected by the extent of surface of the metal plate, 

 therefore lamps were made with a plate of fine wire, and also of a 

 very broad surface, but no difference was observable between these 

 and the normal plate used. 



