Mr Filon, On double refraction in strained glass. 337 



Note on the measurement of Hie dispersion of double refraction 

 in strained glass. By L. N. G. Filon, B.A., King's College. 



[Received 23 December 1903.] 



In a note published in the Annalen der Physik (1903, Series 4, 

 Vol. xi. part (3)), Herr F. Pockels takes objection to some state- 

 ments made by myself in a paper " On the variation with the 

 wave-length of the double refraction in strained glass" (Gamb. 

 Phil. Soc. Proceedings, Vol. xi. pp. 491 — 492). 



He points out that for the lightest glass he has examined, 

 viz., the Jena glass S 205, the change in his quantity n (p — q)/^ 1 

 which is proportional to my stress-optical coefficient C amounts to 

 3 per cent, as we pass from red to green light, whereas in one of 

 the heavier glasses, 500, the dispersion of double refraction is 

 sensibly nil. 



Accordingly my statement (loc. cit. p. 491) that Herr Pockels 

 had found no such dispersion for light glasses, but an appreciable 

 one in heavier glasses, requires qualification, and I withdraw it 

 therefore, with many apologies to Herr Pockels. I may, however, 

 plead for my excuse that it was his own general statement 

 {Annalen der Physik, April, 1902, p. 765) that the dispersion of 

 double refraction was " bei den leichteren Glasern sehr gering " 

 which led me into this error. 



With regard to my other criticism, concerning the significance 

 of apparent dispersion of double refraction introduced by the 

 correction applied for dispersion of double refraction in the quartz 

 plates of the compensator, I quite agree with Herr Pockels that 

 in his case the correction should be applied, since it is theoretically 

 necessary. At the same time he himself {loc. cit. p. 751) admits 

 a possible error of 1 or 2 per cent, in the observations, which is 

 precisely the order of the correction applied. 



Now it appears to me that the significance of numerically 

 equal differences in the shift of the fringes is not the same when 

 the difference is due to the correction, and when it is the outcome 

 of direct observation. A difference observed directly may not be 

 accurately measured, but it nearly always represents a real shift 

 which the observer has noticed. A positive result of this kind 

 has more weight than the negative result that there is no shift of 

 the fringes visible; for an observer is more likely to overlook a 

 shift which exists than to record one which does not. Of course 

 this is, to a certain extent, a matter of opinion ; for my own part 

 I still feel that differences found in this way are less reliable than 

 those directly observed. 



1 See P. Pockels, " Ueber die Aenderung des optischen Verhaltens verschiedener 

 Glaser durch elastische Deformation," Annalen der Piiysik, April, 1902. 



