45o 



NATURE 



[July 3, igi, 



LETTERS TO THE EDITOR. 

 [The Editor does not hold himself responsible for 

 opinions expressed by his correspondents. Neithei 

 can he undertake to return, or to correspond with 

 the writers of, rejected manuscripts intended for 

 this or any other part of Nature. No notice is 

 taken of anonymous communications.] 

 The lonisation of Gases in the Schumann Region. 

 In Nature for June 12, Prof. Lyman discusses the 

 evidence relating to the ionisation of air in the 

 Schumann region, and concludes that ionisation of air 

 can be produced by wave-lengths longer than A.1700. 



The reasons why I consider that A1350 is nearer 

 to the limiting wave-length at which the ionisation 

 of air sets in are as follows : — Using a discharge in 

 hydrogen as a source of ultra-violet light and trans- 

 mitting it through a quartz window (0-3 mm. thick), 

 I was unable to get any ionisation in filtered air. I 

 only obtained big effects with a certain piece of fluorite 

 as the window. Another piece of fluorite which did 

 not transmit the ionising light was transparent to 

 about A1350. Prof. Lyman's researches show that the 

 hydrogen discharge emits very intense ultra-violet 

 light distributed over a large number of wave-lengths 

 between X1200 and A.1600. Hence, with thin quartz, 

 there was plenty of light available down to M450, 

 but it produced no effect in my experiments. Similarly, 

 wave-lengths down to about A.1350 produced no appre- 

 ciable effects. 



Lenard and Ramsauer used a very intense 

 aluminium spark as their source of light, and found 

 that the light from it transmitted through fluorite 

 produced enormous ionisation in air. On the other 

 hand, the light when passed through quartz did not 

 produce any effect. According to Prof. Lyman's 

 photographs, the wave-lengths available from the Al 

 spark in air are a strong group of lines near M300, 

 some weak lines near M500, and strong lines near 

 A.1600 and A.1720-A.1800. Thin quartz cuts out the 

 group M300, but allows the others to pass. We are 

 not told explicitly whether the spark in Lenard and 

 Ramsauer's researches was ever placed close to the 

 quartz window to avoid air absorption ; if so, the 

 M500 and Xi6oo groups would be effective. Fluorite, 

 on the other hand, transmits the a 1300 group as well, 

 and Prof. Lyman considers that the ionisation ob- 

 served is probably due to these lines. He points out 

 that my interpretation of his remarks, viz. that A.1300 

 represents the longest wave-lengths which are effective 

 in ionising air, does not represent his views correctly. 

 He considers that Bloch's recent work on the ionisa- 

 tion of air by a mercury lamp indicates that wave- 

 lengths longer than M750 are effective. 



I expect it will be agreed that by air. we mean the 

 usual mixture of oxygen and nitrogen free from all 

 the more condensable gases. Lenard and Ramsauer 

 found that ordinary dust-free air was certainlv ionised 

 by the light transmitted through quartz from their 

 powerful source. It was only when verv drastic 

 methods of purification were adopted that the air was 

 no longer ionised by the light transmitted through 

 quartz. Although Bloch used dust-free air in his ex- 

 periments, there is no evidence that he took the 

 risrorous precautions which Lenard and Ramsauer 

 assert are necessary to get rid of all the impurities 

 which give rise to ionisation with comparatively long 

 wave-lengths. In Bloch's arrangement, the mercury 

 lamp was totallv immersed in the stream of air, and 

 consequently all the light emitted was available for 

 ionisation, and hence the traces of impurities have 

 everv chance to be ionised. Bloch does not eive any 

 details, but I think the supports, insulated wires, &c, 

 connected with the lamp inside the ionisation chamber 

 might act as sources of impurities in Lenard and 

 Ramsauer's sense. 



NO. 227Q, VOL. qi~| 



If we consider the quantum theory of radiation to 

 apply to ionisation of gases by light, then the energy 

 available in the quantum, hn, must exceed the work 

 V e required to separate an electron from a molecule. 

 Palmer's experiments (Phys. Rev., xxxii., p. 1, 191 1) 

 may perhaps be taken to indicate that the oxygen 

 accounts for most of the ionisation in air. Taking 

 the longest wave-length which ionises air to be M350, 

 and 7j = 6-55 x io -27 , and 6 = 4-65 x 10- 10 , we get V = 9-2 

 volts. Now the ionising potential for oxygen accord- 

 ing to Frank and Hertz is 9-0 volts. To maintain 

 that X1800 is nearer to the long wave-length limit 

 implies that the quantum theory is not applicable to 

 ionisation by light, for there is no reason to doubt the 

 accuracy of the experiments of Frank and Hertz. 

 A. Ll. Hughes. 



Cavendish Laboratory, Cambridge. 



The Microtropometer. 



Many roads to progress in physical investigation are 

 brought to an abrupt end through the lack of measur- 

 ing instruments of sufficient sensitiveness. In the 

 attempt to bridge over one of these disabling chasms 

 the writer was led to the following device, which 

 appears capable of some development. The principle 

 can be illustrated with reference to a particular case. 

 Suppose we have a Boys's radio-micrometer, which we 

 will call the secondary instrument. If we project on 

 to the vane of this an image of a Nernst lamp filament 

 the beam of light from the mirror of the instrument 

 may be deflected through one thousand scale divi- 

 sions. Suppose now that the image of the filament 

 is 1 mm. wide, and that it is projected by the mirror 

 of another radio-micrometer, which we may call the 

 primary one. 



It is evident that a movement of this primary image 

 through a distance of 1 mm. can produce a movement 

 of the beam of light from the secondary instrument 

 through 1000 mm. Hence, a movement of the image 

 of the filament cast by the primary instrument through 

 o-ooi mm. would give a deflection of about 1 mm., 

 and a movement of the primary image through 

 1 1, (.mi. (100 mm. would move the secondary image 

 through one-thousandth of a mm. If now the 

 secondary instrument be made to throw a similar 

 image on to a tertiary radio-micrometer, the motion 

 will again be magnified one thousand times, so that 

 an original movement of a millionth of a mm. pro- 

 duces a final movement of 1 mm. Evidently bv in- 

 creasing the number of instruments in arithmetical 

 progression we increase the magnification in geo- 

 metrical progression. 



I have applied the method to two radio-micrometers 

 with very satisfactory results. The principle, how- 

 ever, can be applied to any instrument in which a 

 beam of _ light is used as an indicator — e.g. the 

 primary instrument may be a galvanometer, an elec- 

 trometer, a double-thread-suspension mirror, or a 

 string-galvanometer (in the last case the image of the 

 string taking the place of the image of the filament). 

 The secondary instruments may be radio-micrometers, 

 thermo-couples, bolometers, selenium cells, or other 

 detectors of radiation. It will appear, therefore, 

 that the principle is one capable of wide applica- 

 tion to cases in which great sensitiveness of measure- 

 ment is required — from wireless telegraphv to 

 physiology. In fact, we may say that anv existing 

 instrument which uses light as an indicator can be 

 made more sensitive. 



Practical difficulties arise from the impossibilitv of 

 obtaining any instrument with absolutely constant 

 zero; moreover, fluctuations in the intensity of the 

 energy stream from the source of radiation, repre- 

 sented by the Nernst filament, would cause trouble 



