648 



NATURE 



[July 21, 192 1 



Letters to the Editor. 



\The Editor does not hold himself responsible for opinions ex- 

 pressed by hts correr.pondents. Neither can he undertake to 

 return, or to correspond with the writers of, rejected manu- 

 scripts intended for this or any other part of Nature. 

 No notice is taken of anonymous communications.'] 



Radiation from the Carbon Arc. 



An application of Merton and Nicholson's form of 

 the wedge method has been made to the study 

 of the intensity distribution in typical stellar spectra. 

 In the late type spectra (including the sun), which are 

 suflficiently bright at the red end, the energy curves 

 give a marked depression from o-5o-o-67/jt — a result 

 not obtained by other observers. As the spectrum of 

 the carbon arc, assumed to radiate as a black body 

 at a temperature of 3750° Abs., is used to remove the 

 tfolour curve of the plate, a possible cause of this 

 depression, which is common to all stellar spectra 

 that have been observed in the red, would lie in an 

 intensity distribution in the carbon arc differing from 

 that assumed. 



In order to determine the intensity distribution in 

 the spectrum of the positive crater of the carbon arc. 



OJtXf*. 0«.b oSo oSif. aS% eit^ 



Fig. I. — Intensity distribution in spectrum of positive crater 



the acetylene flame was used as a standard. The 

 intensity distribution in a cylindrical flame of specified 

 dimensions burning acetylene generated from com- 

 mercial calcium carbide has been carefully determined 

 by Coblentz with the bolometer (Bureau of Standards, 

 Scieritific Papers No. 279, 1916, and No. 362, 1920). 

 To reproduce as closely as possible the conditions em- 

 ployed by Coblentz, a burner was used giving a cylin- 

 drical flame of the specified dimensions, and the 

 acetylene used was obtained from three different 

 sources : two cylinders containing acetylene com- 

 pressed in acetone and a simple gasometer which 

 generated gas from commercial calcium carbide. 

 From o-40-o-6o/x the intensit}^ distributions in the 

 flames from these three sources were in good agree- 

 ment. From o-6o-o-675/x there were systematic dif- 

 ferences with a range of 8 per cent., due probably to 

 the varying quality of the acetylene. In this part of 

 the spectrum mean intensities were used as giving 

 the closest approximation to Coblentz 's conditions. 

 Some initial experimental difficulties were encountered 

 owing to non-uniform illumination of the wedge by 

 the flame which could not be detected by the usual 

 methods. It was found that reversal of the wedge 



NO. 2699, VOL. 107] 



on the slit furnished a very sensitive test of uniformity, 

 and by the use of this method an arrangement of 

 apparatus was secured which gave uniformity of 

 illumination within the limits of the accidental errors 

 of the wedge method. 



When the initial difficulties had been overcome a 

 series of eight spectra of the acetylene flame was 

 obtained on Ilford panchromatic plates. At the same 

 time five spectra of the positive crater of the carbon 

 arc were secured, cored carbons being used with a 

 current of 57 amperes. The heights of these wedge 

 spectra were measured to two different densities at 

 various wave-lengths by means of a simple form of 

 microphotometer devised for that purpose. The two 

 series of measures differed by less than 2 per cent. 

 The final mean absolute intensities of the carbon arc 

 for various wave-lengths are shown in Fig. i by black 

 circles. The mean probable error of these intensities 

 is 17 per cent., the mean probable error of a single 

 plate being 57 per cent. These mean probable errors 

 would be considerably reduced if the intensity at 

 0-675/X were not used, the intensity at this point being 

 subject to large accidental errors on account of the 

 rapidly changing colour-curve of the plate. 



It will be noticed that the observed intensities 

 depart considerably from the intensi- 

 ties computed on the assumption of 

 black-body radiation at a temperature 

 of 3750° Abs. (shown by the dotted 

 curve). The use of the observed 

 values of the intensity distribution 

 gives results for stellar spectra more 

 in accord with those obtained bv pre- 

 vious observers. There is still out- 

 standing, however, in the case of the 

 sun a depression requiring further 

 investigation. 



Two hypotheses may be advanced to 

 account for the observed intensitv 

 distribution :- — 



(i) The carbon arc radiates as a 

 black body at a temperature of 3750° 

 Abs., but there is an absorption band 

 with its centre at o-50yu, due possiblv 

 to the incandescent carbon particles in 

 the arc flame. Coblentz has shown 

 that at 2360° Abs. these carbon par- 

 ticles have an absorption band with 

 centre at 0-60/* (Bureau of Standards, 

 Scientific Paper No. 156, 191 1). At 

 the temperature of the arc flame this 

 absorption band vvouM suffer a shift to the violet, 

 bringing its centre approximately in the observed place. 

 The advantage of this hypothesis is that it is in accord 

 with previous determinations of the arc temperature by 

 such various methods as (a) the calorimetric method 

 used by Violle, {b) the wave-length of maximum 

 energy used by Lummer and Pringsheim, (c) F^r5''s 

 determination from the total radiation, and (d) various 

 determinations by optical pyrometers using an ap- 

 proximately monochromatic band in the red. This 

 hypothesis is represented graphically in Fig. i by the 

 dotted curve showing the intensity distribution of a 

 black body at 3750° Abs. computed from Wien's law, 

 so as to bring it into approximate agreement with the 

 observed values from o-6oo-o-675//. 



(2) The carbon arc radiates as a black body at a 

 temperature of 3325° Abs. This hypothesis is repre- 

 sented by the full curve in Fig. i, which fits the 

 observed values fairly well. It can be brought into 

 accord with previous work only by supposing that 

 the commercial cored carbons used burned at a lower 

 temperature than those carbons used by other inves- 

 tigators. Wanner has found differences of 200** 

 depending upon the carbons used. 



M>.4(* O.JO 



carbon arc. 



