June io, 1922] 



NA TURE 



745 



(The Blue Flame produced by Common Salt on a 

 Coal Fire. 



The blue flame produced by sprinkling salt on a 



)wing coal fire is a good example of common know- 



ige, which, not finding a niche and an explanation 



text-books, becomes a recurrent topic of inquiry 



id discussion in scientific journals. It may perhaps 



of interest if I add a historical note to what Prof. 



srton has stated in Nature of May 27, p. 683. 



The blue flame in question appears to have been 



5t treated from the spectroscopic standpoint by 



le late Dr. J. H. Gladstone in 1862 in a letter to 



le Philosophical Magazine (ser. iv., vol. 24, p. 417). 



rithout being quite conclusive he seems to have 



yarded copper chloride as the source. The matter 



is raised again by an anonymous letter to Nature 



in 1876 (vol. xiii. p. 287), and a discussion has recurred 



from time to time in these columns from that date 



until 1890. Full references to this are to be found 



in Kaysev's " Handbuch der Spectroscopic, " vol. v. 



p. 391. A communication to Nature by T. N. 



^iiiller in 1876 (vol. xiii. p. 448) seems to have hit the 



lark. He recognised the flame as being like that of 



cupper chloride, and, surmising that the source of 



the copper lay probably in the pyrites of the coal, 



found that the blue flame did not appear when salt 



was sprinkled on a glowing fire of charcoal. The 



matter was clinched by Salet in 1890 {Comptes 



rend, no, p. 282), who identified the spectrum with 



that of copper chloride as carefully mapped by Lecoq 



de Boisbaudran, and he actually isolated metallic 



copper from the fuel ash. 



The blue flame given by salt always seems to me 

 distinguishable from that of carbon monoxide, and 

 appears very bright by contrast with the yellow-red 

 glow of the fire. It is somewhat surprising to see 

 how far the yellow sodium flame is suppressed. 



Arthur Smithells. 

 The University, Leeds, May 28. 



Optical Resolving Power and Definition. 



In Nature of May 27, p. 678, Mr. A. Mallock 

 suggests as a quantitative measure of " definition " 

 in an optical instrument " the angular or Unear size 

 of the field of view compared with the smallest 

 corresponding quantity which can be clearly distin- 

 guished," and proceeds to extend " definition " on 

 an equivalent general basis to a number of other 

 instruments. 



Whether or not this proposal will serve a useful 

 l)urpose in other directions need not be discussed 

 liere, but in the case of optical instruments the 

 measure proposed will not commend itself to opti- 

 cians, for it involves a radical change in the accepted 

 meaning of " definition " in this connection. The 

 suggestion in fact amounts to nothing more than the 

 measurement of the angular field of view in terms of 

 a unit which varies with the aperture of the lens 

 and the wave-length of the hght which is used, a 

 proposal which surely carries its own condemnation 

 in its enunciation. That the ratio in question is 

 worthless as a measure of " definition " is obvious 

 from the consideration that in many instruments, 

 at say the centre of the field, the resolving power and 

 the "definition" — that is the degree to wliich 

 details of an object are clearly discernible in its 

 image — may remain unaffected while the field of 

 ' icw is greatly changed by an alteration in the size 

 t a suitably placed stop. Conversely in apparently 

 -miliar instruments the " definition " may vary 

 appreciably from one instrument to another while 

 the field of view and the resolving power are alike in 

 all cases. 



NO. 2745, VOL. 109] 



The distinction between resolving power and 

 definition is real but not easily defined in a few 

 words. The former deals with the discernment of 

 separate sources of such apparent minuteness that it 

 cannot be claimed that the image indicates with 

 any accuracy the shape of the source itself. The 

 latter is concerned with the sharpness of the apparent 

 image outline of larger objects. The former depends 

 primarily on the dimensions of the first dark ring 

 in the image of an apparently point source, and the 

 conditions of observation require the range of wave- 

 lengths of light forming the focussed image to be 

 Hmited. The latter depends more upon the broad 

 light distribution in the diffraction pattern than 

 upoa the alternations of light and darkness, and the 

 range of wave-lengths is not an important factor. 

 As the size of the rings is not greatly affected by 

 small amounts of aberration, .the resolving power is 

 not a suitable measure of the correction of a lens 

 system, but it is precisely upon the degree to wliich 

 aberrations are removed that definition depends. 

 Of two photographic lenses with the same resolving 

 power, and the same field of view, one may give 

 brilliant pictures because the definition is good and 

 the other comparatively flat pictures because the 

 definition is poor. To the user of simple instruments 

 definition is of great importance, resolving power does 

 not concern him. 



This is not a suitable occasion on which to- discuss 

 the measurement of " definition " or the standards 

 which are suitable for application to various types of 

 instrument. The subject is one of great difficulty 

 particularly in view of our ignorance of the extent 

 to which it is possible to ehminate aberrations in 

 systems of simple construction. Lest, however, 

 readers of Nature should be misled it cannot be too 

 emphatically stated that in omitting from " defini- 

 tion " its most essential factor and substituting 

 therefor an independent conception, Mr. Mallock's 

 attempted generaUsation is Hkely to prove only a 

 cause of confusion to those who hope to measure the 

 merits of optical instruments by its means. 



T. Smith. 



The Difference between Series Spectra of Isotopes. 



Prof. P. Zeeman mentioned to me recently some 

 new measurements of the absorption spectrum of 

 lithium which he undertook in order to prove the 

 presence of both isotopes. It seems to me, that 

 at the present time it is not certain what one 

 should expect here theoretically. Bohr's formula 

 for the change in the frequency v due to the motion 

 of the nucleus lias been applied by him only to the 

 cases in which a single electron moves around the 

 nucleus ; namely, to H and He"^. Recently the 

 formula has been also applied by various authors 

 (see F. W. Aston, " Isotopes," p. 123 — London 1922) 

 to the calculation of the difference between series 

 spectra of isotopes ; this means to atoms in which 

 several electrons move around the nucleus. So far as 

 I know there are as yet no investigations on the 

 equation which must for these cases replace Bohr's 

 equation 



Mj_ . Ml 

 "^•"I'M-i+rn'M, +m • * ' ^ ' 



(Ml, M,, w are respectively the masses of the nucleii 

 of the "isotopes and of the electron ; v^, v^ are the 

 frequencies of cprresponding lines). 



In the case of one electron only, (i) follows immedi- 

 ately from the well-known transformation of the 

 " problem of two bodies " from absolute to relative 

 co-ordinates (see, e.g., Whittaker, " Analytical 



2B 2 



