2^0 



NATURE 



[July 3, 1890 



by Mr. Preece as far back as 1885, and I have myself 

 abundantly confirmed his results. 



We should expect also to find that the larger we make 

 our plate, and the nearer we bring it to the negative leg 

 of the carbon, the greater will be the current produced in 

 a circuit connecting this plate to the positive terminal of 

 the lamp. I have before me a lamp with a large plate 

 placed very near the negative leg of the carbon of a lamp, 

 and we find that we can collect enough current from these 

 molecular charges to work a telegraph relay and ring an 

 felectric bell. The current which is now working this 

 relay is made up of the charges collected by the plate 

 from the negatively charged carbon molecules which are 

 projected against it from the negative leg, across the 

 highly perfect vacuum. I have tried experiments with 

 lamps in which the collecting plate is placed in all kinds 

 of positions, and has various forms, some of which are 

 here, and are represented in the diagrams before you ; 

 but the results may all be summed up by saying that the 

 greatest effects are produced when the collecting plate is 

 as near as possible to the base of the negative end of the 

 loops, and, as far as possible, incloses, without touching, 

 the carbon conductor. Time will not permit me to make 

 more than a passing reference to the fact that the magni- 

 tude of the current flowing through the galvanometer 

 when connected between the middle plate and the posi- 

 tive terminal of the lamp often "jumps " from a low to a 

 high value, or vice versd, in a remarkable manner, and 



Fig. 9. — Collecting plate placed at end of : 

 bulb. 



elbow tube opening out of the 



part of the circuit of the incandescent carbon loop. This 

 experiment with the condenser discharged by the lamp 

 may be then looked upon as an arrangement in which the 

 plates of a charged condenser are connected respectively 

 to an incandescent carbon loop and to a cool metal plate, 

 both being inclosed in a highly vacuous space, and it 

 appears that when the incandescent conductor is the 

 negative electrode of this arrangement the discharge 

 takes place, but not when the cooler metal plate is the 

 negative electrode of the charged condenser. The nega- 

 tive charge of the condenser can be carried across the 



Fig. 10.— Charged condenser c discharged by middle plate m, when the 

 positively charged side of condenser is in connection with the plate and 

 other side to earth e. 



vacuous space from the hot carbon to the colder metal 

 plate, but not in the reverse direction. 



This experimental result led me to examine the con- 

 dition of the vacuous space between the middle metal 

 plate and the negative leg of the carbon loop in the case 

 of the lamp employed in our first experiment. Let us 

 return for a moment to that lamp. I join the galvano- 

 meter between the middle plate and the negative terminal 

 of the lamp, and find, as before, no indication of a current. 

 The metal plate and the negative terminal of the lamp 



that this sudden change in the current can be produced 

 by bringing strong magnets near the outside of the bulb. 

 Let us now follow out into some other consequences 

 this hypothesis that the interior of the bulb of a glow- 

 lamp when in action is populated by flying crowds of 

 carbon atoms all carrying a negative charge of electricity. 

 Suppose we connect our middle collecting plate with some 

 external reservoir of electric energy, such as a Leyden 

 jar, or with a condenser equivalent in capacity to many 

 hundreds of Leyden jars, and let the side of the condenser 

 which is charged positively be first placed in connection 

 through a galvanometer with the middle plate (see Fig. 10), 

 whilst the negative side is placed in connection with the | 

 earth. Here is a condenser of two microfarads capacity j 

 so charged and connected. Note what happens when I 

 complete the circuit and illuminate the lamp by passing [ 

 the current through its filament. The condenser is at ' 

 once discharged. If, however, we repeat the same ex- 

 periment with the sole difference that the negatively 

 charged side of the condenser is in connection with the 

 middle plate then there is no discharge. The experi- 

 mental results may be regarded from another point of 

 view. In order that the condenser may be discharged as 

 in the first case, it is essential that the negatively charged 

 side of the condenser shall be in connection with some 



NO. 1079, VOL. 42] 



Fig. II. — Current from Clark cell Ck being sent across vacuous space 

 between negative leg of carbon and middle plate m. Positive pole of 

 cell in connection with plate m through galvanometer g. 



are at the same electrical potential. In the circuit of the 

 galvanometer we will insert a single galvanic cell having 

 an electromotive force of rather over one volt. In the 

 first place let that cell be so inserted that its negative pole 

 is in connection with the middle plate, and its positive 

 pole in connection through the galvanometer with the 

 negative terminal of the lamp (see Fig. 1 1). Regarding 

 the circuit of that cell alone, we find that it consists of 

 the cell itself, the galvanometer wire, and that half-inch 

 of highly vacuous space between the hot carbon conductor 



