A tig itst 10, 1882] 



X. I TURE 



557 



The experiment I have ju^t shown illustrates the principle of 

 electric lighting by incandescence, which is briefly this — that a 

 while heat may be produced in a continuous solid con- 

 passing a sufficiently strong electrical current through it. 



A prin ciple, the importance of which cannot will be over- 

 estimated, underlies this method of producing light electrically — 

 namely, the principle of divisibility. By means of electric in- 

 candescence it is possible to produce exceedingly small centres 

 of light, even so small as the light of a single candle ; and with 

 1 power in proportion to the light pro- 

 duced, than is involved in the maintenance of light-centres 10 

 or ICO limes greater. Given a certain kind of wire, for 

 example a platinum wire, the 100th of an inch in diameter, 

 a certain quantity of current would make this wire white-hot 

 whatever its length. If in one case the wire were one inch long 

 and in another case ten inches long, the same current passing 

 through these two pieces of similar wire, would heat both to 

 precisely the same temperature. But in order to force the same 

 current through the ten times longer piece, ten times the electro- 

 motive force, or, if I may be allowed the expression, electrical 

 pressure, is required, and exactly ten times the amount of energy 

 would be expended in producing this increased electro-motive 

 force . 



Considering, therefore, the proportion between power applied 

 and light produced, there is neither gain nor loss in heating 

 these different lengths of wire. In the case of the longer wire, 

 as it had ten times the extent of surface, ten times more light 

 was radiated from it than from the shorter wire, and that is 

 exactly equivalent to the proportional amount of p 

 sorbed. It is therefore evident that whether a short pi 

 or a long piece is electrically heated, the amount of light pro- 

 duced is exactly proportional to the power expended in pro- 

 ducing it. 



This is extremely important ; for not only does it make it 

 possible to produce a small light v here n small light is required, 

 without having to pay for it at a higher rale than for a larger 

 light, 1 ut it gives also the great advantn- 



tributiou of light. As ihe illuminating effect of light is inversely 

 as the square of the distance of its source, ii follows that where 

 a large space is to be lighted, if the lighting is accomplished by 

 ■Dean-, of centres of light of great power, a much larger total 

 quantity of light has to be employed, in order to make the 

 spaces remotest from these centres sufficiently light, than would 

 be required if the illumination of the space were obtained by 

 numerous smaller lights equally distributed. 



In order to practically apply the principle of producing light 

 by the incandescence of an electrically heated continuous solid 

 conductor, it is necessary to select for the light-giving body a 

 material which offers a considerable resistance to the passage of 

 the electric current, and which is also capable of bearing an 

 exceedingly high temperature without undergoing fusion or other 

 change. 



As an illustration of the difference that exists among different 

 substances in respect of resistance to the flow of an electric 

 current, and consequent tendency to become heated in the act of 

 electrical transmission, here is a wire formed in alternate sections 

 of platinum and silver : the wire'is perfectly uniform in diameter, 

 and when I pass an electric current through it, although the 

 current i~ uniform in every part, yet, as you see, the wire is not 

 uuiformly hot, but white-hot only in parts. The white-hot sec- 

 tions are platinum, the dark sections arc silver. Platinum offers 

 a higher degree of re-istance to the passage of the electric 

 current than silver, and in consequence of this, more heat is 

 developed in the platinum than in the silver sections. 



The high electrical resistance f platinum, and its high 

 melting-point, mark it out as one of the most likely of the 

 metals to be useful in the construction of incandescent lamps. 

 When platinum is mixed with 10 or 20 per cent, of iridium, an 

 alloy is formed, which has a much higher melting-point than 

 platinum ; and many attempts have been made to employ this 

 alloy in electric lamps. But these attempts have not been suc- 

 cessful, chiefly because, high as is the melting-point of iridio- 

 platinum, it is not high enough to allow of its being heated to a 

 degree that would yield a sufficiently lar^e return in light for 

 energy expended. Before an economical temperature is reached, 

 iridio-platinmn wire slowly volatilises and breaks. This is a 

 tatal fault, because in obtaining light by incandescence there is 

 the greatest imaginable advantage in being able to heat the 

 incandescing body to an extremely high temperature. I will 

 illustrate this by experiment. 



Here is a glass bulb containing a filament of carbon. When 

 I pass through the filament one unit of current, light equal to 

 turn candles is pr iduced. If now I increase the current by one- 

 half, making it one unit and a half, the limit is increased to 

 thirty candles, or thereabout, so that for this one-half increase of 

 current (which involves nearly a doubling of the energy expended), 

 fifteen times more light is produced. 



It will readily be understood from what I have shown thai it 

 is essential to economy that the incandescing material should be 

 able to bear an enormous temperature without fusion. We 

 know of no metal that fulfils this requirement ; but there is a 

 non-metallic substance which does so in an eminent degree, and 

 which also possesses another quality, that of lore conductivity. 

 The substance is carbon. In attempting to utilise carbon for the 

 purpose in question, there are several serious practical difficulties 

 to be overcome. There is, in the first place, the mechanical 

 difficulty arising from its intractability. Carbon, as we com- 

 monly know it, is a brittle and non-elastic substance, possessing 

 neither ductility nor plasticity to favour its being shaped suitably 

 for use in an electric lamp. Vet, in order to render it service- 

 able for this purpose, it is necessary to form it into a slender 

 filament, which must possess sufficient strength and elasticity to 

 allow of its being firmly attached to conducting-v, ires, and to 

 prevent its breaking. If heated white hot in the air, carbon 

 burns away ; and therefore means must be found for preventing 

 its combustion. It must either be placed in an atmosphere of 

 some inert gas or in a vacuum. 



During the last forty year-, spasmodic efforts have from time 

 to time been made to grapple with the many difficulties which 

 surround the u-e of carl .on as the wick of an electric lamp. It 

 is only within the last three or four years that these difficulties 

 can be said to have been surmounted. It is now found that 

 carbon can be produced in the form of straight or bent filaments 

 of extreme thinness, and possessing a great degree of elasticity 

 and strength. Such filaments can be produced in various ways 

 — by the carbonisation of paper, thread, and fibrous wood 

 grasses. Excell-nt carbon filaments can be produced from the 

 bamboo, and also from cotton thread treated with sulphuric acid. 

 The sulphuric acid treatment effects a change in the cotton 

 thread similar t) that which is effected in paper in the procf 

 making parchment paper. In carbonising these materials, it is 

 of course necessary to preserve them from contact with the air. 

 This is done by surrounding them with charcoal. 



Here is an example of a carbon filament produced from parch- 

 mentised cotton thread. The filament is not more than the - oi 

 of an inch in diameter, and yet a length of three inches, having 

 therefore a surface of nearly the one-tenth of an inch, gives a 

 light of twenty candles when made incandescent to a moderate 

 degree. 



I have said, that, in order to preserve these slender carbon 

 filaments from combustion, they must be placed in a vacuum ; 

 and experience has shown that if the filaments are to be durable, 

 the vacuum must be exceptionally good. One of the chief causes 

 of failure of the earlier attempts to utilise the incandescence of 

 carbon, « as the imperfection of the vacua in which the white-hot 

 filaments were placed ; and the success which has recently been 

 obtained is in great measure due to the production of a better 

 vacuum in the lamps. 



In the primitive lamps, the glass shade or globe which enclosed 

 the carbon filament was large, and usually had screw joints, with 

 leather or india-rubber washers. The vacuum was made either 

 by filling the lamp with mercury, and then running the mercury 

 out so as to leave a vacuum like that at the upper end of a baro- 

 meter, in- the air was exhausted by a common air pump. The 

 invention of the mercury pump by Dr. Sprengel, and the publi- 

 cation of the delicate and beautiful experiments of Mr. Crookes 

 in connection with the radiometer, revealed the conditions under 

 which a really high vacuum could be produced, and in fact gave 

 quite a new meaning to the word vacuum. It was evident that 

 the old incandescent lamp experiments had not been made under 

 sui'able conditions as to vacuum ; and that before condemning 

 the use of carbon, its durability in a really high vacuum required 

 still to be tested. This idea having occurred to me, I communi- 

 cated it to Mr. Stearn, who was working on the subject of high 

 vacua, and asked his co-operation in a course of experiments 

 having for their object to ascertain whether a carbon filament 

 produced by the carbonisation of paper, and made incandescent 

 in a high vacuum was durable. After much experimenting we 

 arrived at the conclusion that when a well-formed carbon filament 

 is firmly connected with conducting wires, and placed in a hermeti- 



