400 Tests of -Glow-Lamps. 



contacts, and so to complete the circuit of an electromagnet ; 

 this, by closing one side of the mercury switch (fig. 6) , started 

 the motor, which increased the resistance in the main circuit 

 until the pressure was correct again. 



Exactly the opposite happened if the pressure fell, the 

 pointer touched the other contact, the other side of the mercury 

 switch was closed, and the motor working in the reverse 

 direction reduced the resistance between the cells and the lamps. 



The pointer was kept from sticking to its contacts by a 

 light wooden hammer which tapped the case of the galvano- 

 meter when the motor revolved. 



By means of this apparatus it was possible to keep the 

 pressure constant to within at least one tenth per cent. 



For many nights the action of this regulator was watched 

 pending the arrival of a recording voltmeter, but the recording 

 voltmeter not being forthcoming, and staying up all night, 

 even for four times a week, interfering much with our work 

 during the day, we decided to construct an automatic check 

 on the automatic regulator. 



In series with the first galvanometer was placed a second 

 working in exactly the same way, but with its contacts so far 

 apart that its pointer only touched them for a rise or fall of 

 half a volt in the pressure on the lamps. When for any 

 reason such an alteration took place, the pointer of this second 

 galvanometer, by touching one of its contacts, completed the 

 circuit of an electromagnet ; this attracted a piece of soft iron 

 which opened a mercury switch and put out the lamps, and 

 at the same time by stopping a clock indicated the time at 

 which the irregularity took place. 



Besides acting as a check on the regulating apparatus, this 

 automatic cut-out worked whenever the pressure rose on a 

 lamp breaking, and so made it possible to tell the exact life 

 of the broken filament. 



The lamps being tested were arranged in the special stand 

 shown in fig. 7. The circular plate A, which could be turned 

 on the centre pillar B, was divided into six compartments by 

 the radial divisions shown, and the nine lamps simultaneously 

 tested were placed in three of these compartments, which 

 were coloured a dead black inside. The whole was covered 

 up with an outer tin case, drawn in broken lines, which was 

 fixed to the uprights 0, D,E, F, and had an opening in front 

 through which any one of the groups of lamps could be seen 

 when turned into position. 



When photometric measurements were not being made, the 

 top of the case was left open to let out the heat. 



The arrangements for measuring the current and for 

 bringing it to the lamps are sketched in fig. 8. M and N 



