450 KNIGHT DUNLAP 



be introduced. This set-up is shown in figure 1. In this lamp- 

 board four 100 or 120 watt lamps are satisfactory, and all are 

 left permanently screwed down. Four 120 watt lamps reduce 

 the voltage to about 50, when the circuit is broken. 



Storage cells may be used as a source of direct current, the 

 voltage of the battery depending on the fork somewhat. In 

 some cases a standard 12 volt battery will serve, but a twenty 

 to twenty-five volt set, with adjustable resistance capable of 

 carrying 1 to 1.5 amperes and a maximum of 12 to 18 ohms is 

 preferable. Edison storage cells are recommended on account 

 of their requiring less care; and they may be used in the same 

 room with the apparatus, as they have no fumes deleterious to 

 steel as do the acid cells. The variations in voltage of the Edison 

 cells do not affect the measurements. 



If the fork-magnet is high-wound, it may be necessary to 

 arrange a shunt around it, to provide sufficient current for the 

 motor field without forcing too much through the fork. I have 

 not found this necessary with any of the standard wound forks. 



The fork contacts should be of dental gold alloy (gold and 

 platinum), which is cheaper than platinum and far more satis- 

 factory. Tungsten contacts might possibly be employed, but 

 I am unable to say definitely that tungsten is satisfactory. Much 

 trouble will be found if platinum be used and I do not find that 

 the motor operates satisfactorily with a platinum contact fork. 

 I find dental wire of 18 gauge (Brown and Sharpe) works well: 

 19 or 20 might perhaps be used. The table on which the wire 

 strikes should be surfaced with a section (disc) of large gauge 

 wire. 



With any fork interrupting a large current (more than a small 

 fraction of. an ampere) a condenser ought to be used. The 

 small size 2 microfarad condenser made by the Western Electric 

 Company is cheap in price and quite satisfactory. The con- 

 denser must be connected across the gap of the fork contact, 

 and must not include the electro-magnet (of the fork) between 

 its terminals. In other words, when the fork-contact is broken, 

 the circuit must be traceable unbroken from one contact of the 

 condenser through the magnet windings. For convenience, an 



