158 



NATURE 



[Dec. 1 8, 1879 



possession of trigonometric measurements of the most 

 exact nature, extending from lat. 61° in the Shetland 

 Islands, to lat. 34 on the southern frontier of Algeria. 



The extension of this network southward and eastward 

 in Africa, desirable as it is for the elucidation of many 

 nice points in geodesy, is unfortunately scarcely possible 

 in the immediate future, and science must rest content 

 with gaining a foothold in the great continent. 



T. H. N. 



A NEW STANDARD OF LIGHT 1 



IN the pamphlet before us we have a proposal for a new 

 form of standard light, and the author has shown 

 some considerable skill in drawing out his method of 

 producing it. We cannot do better than quote his 

 opening paragraph as showing the requisites of a standard 

 that the author deems necessary. He says : — 



"No exact measurement of any quantity, even with the 

 most accurate and sensitive test measures available, can 

 reasonably be expected unless the standard by which the 

 unknown quantity is to be gauged is perfectly constant in 

 itself; or if nature does not permit of such a desirable 

 state of things, the causes to which the variation of the 

 standard are due should be known, and in addition also, 

 their quantitative effect on the standard, in order to be 

 able to introduce a correction whenever accuracy of 

 measurement should permit, and circumstances necessi- 

 tate it." 



The want of a standard of light has long been felt in 

 physical researches, and the British Association has 

 acknowledged the impossibility of obtaining scientific 

 measures with the ordinary standards, and has appointed 

 a committee to consider the question of fixing such a 

 standard of white light, that a unit of light may be 

 capable of accurate definition. It must not be forgotten 

 that up to quite recent times the principal necessity for a 

 standard at all has arisen through the introduction of gas 

 into our dwellings and streets, and it has only been 

 necessary to adopt one which should give the compara- 

 tive illuminating powers of any variable qualities of gas. 

 In fixing such a standard the points to be looked at were 

 (1) that the standard should be capable of easy and exact 

 reproduction ; (2) that the colour of the light should be 

 approximately the same ; and (3) that in varying states of 

 barometric pressure and temperature, proper corrections 

 in the results of the comparisons should be feasible. 

 It will be seen further on that a fourth desideratum 

 should be introduced for scientific work. Perhaps on no 

 subject has more attention been paid to small details 

 than in the production of a standard candle, and as a 

 result, when burnt under proper conditions, it gives fairly 

 correct values of the illuminating power of gases. 



In the record of Mr. Schwendler's experiments with 

 thp ctnnrlnrd randlp as against his new standard of light, 

 we have some startling variations in the light of a 

 standard candle, but we feel sure that, had the proper 

 conditions been observed, there would never have oc- 

 curred such a tremendous difference as 72 per cent. We 

 are more convinced that ordinary precautions could not 

 have been rigidly observed when we find that some of 

 the comparisons were made after the candle had been 

 freshly lighted. In gas photometry it is well known that 

 the standard candle should burn at least a quarter of an 

 hour before it can be considered to have settled down to 

 a steady light. The standard candle, however, is not a 

 nice unit of light; and two years ago Mr. Vernon Har- 

 court introduced to the notice of the British Asso- 

 ciation a gas standard which seems to meet every 

 requirement. By making a mixture in a small gas- 

 holder of one part of the most volatile spirit from Ameri- 

 can petroleum which distilled at 50 C. with 600 of air, 



1 On a New Standard of Light. By Louis Schwendler. From the 

 ymraalotHK Society of Bengal, vol. xlviii. Tart ii., 1879. 



or seven of the vapour with twenty of air, he produced a 

 gas which, whilst almost insoluble in water, was perma- 

 nent at all ordinary temperatures and pressures, and 

 which was of a known composition and easy of manufac- 

 ture. A jet of such a gas could be compared with the 

 ordinary coal-gas, and any variations affecting the one 

 would equally affect the other. The colours of the 

 standard and coal-gas lights are also approximately the 

 same. It seems that a standard of such a character 

 meets the requirements for comparing the illuminating 

 value of different coal-gases. Mr. Schwendler proposes 

 to use the light radiated from platinum foil, when raised 

 to incandescence by an electric current, as a new standard, 

 and we agree that a solid instead of a gaseous body as 

 the source of illumination is a step in the right direction. 

 The standards made, however, appear to have been used 

 for determining the illuminating value cf the light pro- 

 duced by dynamo-electric machines under varying con- 

 ditions of speed of armature and resistance in circuit, and it 

 is in reference to this that we will first judge of its probable 

 effectiveness, since for gas measurements the standards 

 already existent suffice. Some dynamo-electric machines 

 are advertised as generating the light of 50,000 candles, 

 and we will suppose for the moment we are comparing 

 such a light with Mr. Schwendler' s standard. 



Now it may be safely said that a standard candle, 

 farther away than twenty feet from the photometer, would 

 give too small a light to be practically of use as a standard, 

 whilst if approaching the photometer within one foot the 

 magnitude of the illuminating source would seriously 

 affect any accurate results. In the first case the electric 

 light would have to be about 4,500 feet away from the 

 photometer and in the last about 220 feet. For ordinary 

 photometric work even the least of these distances would 

 be objectionable. The platinum standard employed by 

 Mr. Schwendler is only about 7 of a standard candle, 

 and these distances would have to be increased nearly 

 20 per cent. 



For practical measurements of this description a candle- 

 power of fifty candles is a far preferable value, which it 

 would be difficult to attain by the method proposed. In 

 this case we have the distances reduced, and if the 

 electric lamp is fixed at a distance of ico feet, we have 

 the movable standard ranging between twenty feet and 

 three to four feet, and the readings become easy and are 

 not subject to be seriously affected by the magnitude of the 

 illuminating source ; in fact, the errors of observation 

 then become of larger magnitude than any error arising 

 from this cause. Another point which we have to note 

 is that as far as the colour of the light from the platinum 

 standard is concerned, it possesses very little advantage 

 over the ordinary gas or candle flame, and it would be 

 impossible, or at all events incorrect, to give the illu- 

 minating value of a light such as of that produced by 

 the electric arc in terms of the new standard ; some 

 recent experiments have demonstrated that the red light 

 emitted by one square mile of the hollow crater in the 

 positive carbon is equal to about the red light radiated by 

 40,000 standard candles, whilst the mean green light of 

 the former is equal to the mean green light of about 

 I35,coo of the latter, and until such a time as the relative 

 physiological values of green and red light are accurately 

 known it will be impossible to give any true estimate of 

 the illuminating power of the electric light by ordinary 

 photometric comparisons. Both in magnitude and colour, 

 then, the proposed platinum standard of light seems to 

 fail for measuring light produced by high temperatures. 



We now turn to the details of the lamp itself. We 

 have, firstly, a U-shaped piece of thin platinum foil cut 

 out about 20 mm. in total length, each limb of the U 

 being about 3 mm. in breadth, the tops of which are 

 clipped in thick metal clips. The usual arrangements 

 are made for passing a current through this foil, the 

 amount being registered by a galvanometer in circuit. A 



