2. In the afternoon the bullet induces a current in the 

 transformer in the same direction in which it is flowing, and 

 opposite to what it did in the morning. Since the Indict did 

 not act like a moving negative charge it was acting like a mov- 

 ing positive charge, thus showing that a moving positive charge 

 induces a moving negative charge in the same direction in 

 which it is moving. 



The result obtained in 2 is contrary to the generally ac- 

 cepted laws of induction, and we are broughl in contact wirh 

 a new condition and must add another law to the already 

 known one. viz.: A moving positive charge will induce a posi- 

 tive charge in the opposite direction and a negative charge in 

 the same direction in which it is flowing. 



The above data indicates that the bullet charges positively 

 in the afternoon. This is contrary to what we could expect. 

 Since the air is highly charged with free electrons in the after- 

 noon, one would think that the moving bullet would collect 



electrons from the atmosphere and become negatively charged 

 instead of positively charged. Our conclusion that the bullet 

 charges positively in the afternoon rests upon the assumption 

 that the current of electricity flows from the copper to the 

 zinc of a battery in the outside circuit. The above result leads 

 us to think that the assumption as to the direction of the cur- 

 rent in a battery may he wrong, and the following experiment 

 was devised in order to demonstrate the direction of the elec- 

 tric current coming from a battery. The same galvanometer 

 was used as in the previous experiment. 



As a preliminary, rubbed glass and gutta-percha rods 

 were brought near the galvanometer, when it was unconnected 

 to any circuit. The galvanometer deflected to the black 

 strongly in either case owing to the induced charge upon the 

 case. 



When the static machine was connected to the galvanom- 

 eter and rotated very slowly the same result was produced. 

 After considerable experimenting the following scheme was 

 adopted: In Fig. 5 let M be a metal plate carried on an in- 



67 



