7i6 



NATURE 



[August 5, 1920 



" indirect " group the variations agree in time, 

 biit are opposite in character; in the third, the 

 " indifferent " group, there is no regular corre- 

 spondence. Sir Norman and Dr. W. J. S. 

 Lockyer have shown that a region may for years 

 belong to the " direct " group, then suddenly 

 become "indirect," and later return to the 

 "direct" group. Drs. Helland-Hansen and 

 Nansen accept this frequent inversion, and also 

 their explanation of the phenomenon. 



The authors' instructive study of North Atlantic 

 temperatures therefore strengthens the case for 



solar variations acting through the atmospheric 

 circulation as the main cause of meteorological 

 changes. To what extent the ocean helps by 

 regulating the air temperature and circulation the 

 authors do not discuss in the present memoir; 

 that and other questions are to be dealt with after 

 further investigations in a series of memoirs to 

 which the present is introductory. The useful- 

 ness of the promised memoirs would be increased 

 (should they have as many appendices and sup- 

 plementary notes as the present) if each were 

 provided with an index. 



The Thermionic Valve in Wireless Telegraphy and Telephony. ^ 



By Prof. J. A. Fleming, F.R.S. 



THE thermionic valve is an invention which has 

 vastly increased the powers and range of 

 wireless telegraphy. Like many other inventions, 

 the telephone, for instance, it is simple in its 

 essential construction. It consists of a little elec- 

 tric lamp comprising a glass bulb, very highly 

 exhausted of its air, containing a filament of 

 carbon, or better tungsten, which can be rendered 

 incandescent by an electric current. Within the 

 bulb and around the filament are fixed certain 

 metal plates or cylinders, and, it may be, spirals 

 of wire or metal networks called the grid. To 

 explain its origin in its simplest form I shall have 

 to take you back in thought to the days when 

 the physical effects taking place in in9andescent 

 electric lamps were first beginning to be con- 

 sidered carefully. In 1883 M^- Edison for some 

 purpose placed in the glass bulb of one of his 

 carbon filament lamps a metal plate which was 

 carried on a platinum wire sealed through the 

 glass. When the filament was rendered incan- 

 descent by a current from a battery, he found 

 that if the plate was connected by a wire, external 

 to the lamp, with the positive terminal of the 

 filament, a small electric current flowed through 

 it, but if connected to the negative terminal no 

 current, or at most a very feeble current, flowed. 

 This new and interesting effect became known as 

 the " Edison effect " in glow lamps, but Mr. 

 Edison gave no explanation of it, and made no 

 practical application of it. 



Edison supplied some lamps with plates in the 

 bulb to the late Sir William Preece, and the latter 

 found that the current called the Edison effect 

 current increased very rapidly as the filament was 

 heated to higher and higher temperatures, and 

 that the collecting plate could be placed a long 

 way from the filament, even at the end of a side 

 tube, without altogether causing it fo vanish. At 

 a little later date I took up the subject, and one 

 of the first things discovered was that the Edison 

 effect was greatly reduced if that side of the 

 carbon loop filament in connection with the nega- 

 tive pole of the battery was enclosed in a glass 

 or metal tube, or if a sheet of mica was inter- 

 posed between the filament and the collecting 

 plate. This seemed to indicate that the effect was 



' From a discourse delivered at the Royal Institution on Friday, May 21. 



NO. 2649, VOL. 105] 



due to some material emission from the hot fila- 

 ment. 



Another fact I observed very soon was that the 

 filament was giving off torrents of negative elec- 

 tricity, and could discharge a positively electrified 

 conductor connected to the plate, but not one 

 negatively charged. Furthermore, I found that 

 the vacuous space between the filament and the 

 plate possessed a curious unilateral electric con- 

 ductivity for low-voltage direct electric currents, 

 and that even a single cell of a battery could pass 

 a current from the hot filament to the collecting 

 plate if the negative pole of the battery was in 

 connection with the hot filament, but not in the 

 opposite direction. This fact had, however, been 

 previously noticed in another manner by W. 

 Hittorf. These experiments were made in 1888 or 

 i88g, and at that time were not satisfactorily ex- 

 plained. 



It was not until nearly ten years later that your 

 distinguished professor of natural philosophy, Sir 

 Joseph Thomson, published accounts of his epoch- 

 making and important researches, in which he 

 proved that the agency we call negative electricity 

 is atomic in structure, and exists in indivisible 

 units now named electrons, which carry a certain 

 electric charge and have a certain mass. These 

 negative electrons are constituents of all chemical 

 atoms. An electrically neutral atom which has 

 lost one or more electrons is called a positive ion, 

 and neutral atoms which have lost or gained elec- 

 trons are said to be ionised. There are arguments 

 in favour of the view that the majority of the 

 atoms in metals and other good conductors of 

 electricity are in a state of intermittent ionisation, 

 and that intermingled with the atoms or positive 

 ions, say in a wire of copper, tungsten, or car- 

 bon, there are electrons which are jumping from 

 atom to atom with great velocity. If we apply 

 to the wire an electromotive force, this causes a 

 drift of these electrons at the instant they are 

 free in the opposite direction to the force (on 

 usual conventions), and this drift or unidirectional 

 motion is superimposed on the irregular motion, 

 and constitutes an electric current. The drift 

 velocity may be very slow compared with the 

 velocity of the irregular motion. The drift motion 

 of the electrons superimposed on the irregular 



