200 



NATURE 



[June 26, 1890 



I may preface all by saying that much of what I have to 

 show you will be seen to be closely related to the 

 phenomena studied by Mr. Crookes in his splendid and 

 classical researches on radiant matter. Our starting- 

 point for this purpose is a discovery made by Mr. Edison 

 in 1884, and which received careful examination at the 

 hands of Mr. Preece in the following year/ and by 

 myself more recently. Here is the initial experiment. 

 A glow-lamp having the usual horseshoe-shaped carbon 

 (see Fig. 3) has a metal plate held on a platinum wire 

 sealed through the glass bulb. This plate is so fixed that 



Fig. 3. — Glow-lamp having insulated metal middle plate m sealed into bulb 

 to exhibit " Edison effect." 



it stands up between the two sides of the carbon arch 

 without touching either of them. We shall illuminate 

 the lamp by a continuous current of electricity, and for 

 brevity's sake speak of that half of the loop of carbon on 

 the side by which the current enters it as the positive leg, 

 and the other half of the loop as the negative leg. The 

 diagram in Fig. 4 shows the position of the plate with 

 respect to the carbon loop. There is a distance of half 

 an inch, or in some cases many inches, between either leg 

 of the carbon and this middle plate. Setting the lamp in 

 action, I connect a sensitive galvanometer between the 



Fig. 4.— Sensitive galvanometer connected between the middle plate and 

 positive electrode of a glow-lamp, showing current flowing through it 

 when the lamp is in action (" Edison eflect "). 



middle plate and the negative ter7ninal of the lamp, and 

 you see that there is no current passing through the in- 

 strument. If, however, I connect the terminals of my 

 galvanometer to the middle plate and to the positive 

 electrode of the lamp, we find a current of some milli- 

 amperes is passing through it. The diagrams in Fig. 5 show 

 the mode of connection of the galvanometer in the two 

 cases. This effect, which is often spoken of as the 

 " Edison effect," clearly indicates that an insulated plate 



I Mr. Preece's interesting paper on this 

 "Proceedings" of the Royal Society for 

 Electrician, April 4, 1885, p. 436. 



NO. 1078. VOL. 42] 



subject is published in the 

 1885, p. 219. See also the 



so placed in the vacuum of a lamp in action is brought 

 down to the same potential or electrical state as the nega- 

 tive electrode of the carbon loop. On examining the 

 direction of the current through the galvanometer we find 

 that it is equivalent to a flow of negative electricity taking 

 place through it frofti the middle plate to the positive 

 electrode of the lamp. A consideration of this fact shows 

 us that there must be some way by which negative 

 electricity gets across the vacuous space from the negative 

 leg of the carbon to the metal plate, whilst at the same 

 time a negative charge cannot pass from the metal plate 

 across to the positive leg. Before I pass away from this 



Fig. 5 — Mode of connection of galvanometer g to rniddle plate m and car- 

 bon horseshoe-shaped conductor c in the experiment of the " Edison 

 effect." 



initial experiment, I should like to call your attention to 

 a curious effect at the moment when the lamp is ex- 

 tinguished. Connecting the galvanometer as at first, 

 between the middle plate and the negative electrode of 

 the lamp, we notice that though made highly sensitive the 

 galvanometer indicates no current flowing through it 

 whilst the lamp is in action. Switching off the current 

 from the lamp produces, as you see, a violent kick or 

 deflection of the galvanometer, indicating a sudden rush 

 of current through it. 



In endeavouring to ascertain further facts about this 

 effect one of the experiments which early suggested itself 

 was one directed to determine the relative effects of 



Fig. 6.— Glow-lamp having negative leg of carbon inclosed in glass tube t, 

 the " Edison effect " thereby being annulled or greatly diminished. 



different portions of the carbon conductor. Here is a 

 lamp (see Fig. 6) in which one leg of the carbon horse- 

 shoe has been inclosed in a glass tube of the size of a 

 quill, which shuts in one-half of the carbon. The bulb 

 contains, as before, an insulated middle plate. If we pass 

 the actuating current through this lamp in such a direction 

 that the covered or sheathed leg is the positive leg, we 

 find the effect existing as before. A galvanometer con- 

 nected between the plate and positive terminal of the 

 lamp yields a strong current, whilst if connected between 

 the negative terminal and the middle plate there is no 

 current at all. Let us, however, reverse the current 



