^256 M. Fraunhofer oil the Laxvs 



With the glass system of lines, where s z=z 0,0001223 the ray 

 (Er) ( + 1), is polarised, that is, the green part of this first spec- 

 trum, when, <r=:49°; (Er) (+"), or the green part in the se- 

 cond spectrum on the same side of the axis is polarised, when 

 <r = 40°. Lastly, (Er) (—i), or the green part of the first spec- 

 trum lying on the other side of the axis, when tf = 69°. 



When (Er) (— i) is completely polarised the remaining co- 

 lours of this spectrum are but imperfectly so. This is less 

 the case in (Er) ( + ii), and a may be perceptibly altered while 

 that colour still remains polarised. (Er) (— i) is under no 

 angle of incidence so completely polarised, as (Er) C + i). With 

 a system of lines, in which g is larger than in that just now 

 mentioned, the angles of incidence must be totally different, if 

 the above-mentioned spectra are to remain polarised.* It is 

 s^en from the equation (V.) that, when e becomes <£ . sin. c + 

 »«, cos. r (+0 > I, therefore is impossible. With vertically 

 received rays we must have £ > fw, if r (") shall be still vi- 

 sible, that is, possible. If s < m, no coloured ray remains 

 visible^ however the light may fall, and it remains only the 

 white light in the axis — namely cos. r(0) — sin. <s. Were s=rw, 

 there would be r (i)=90. If therefore a system of lines is 

 made, in which the distances of the intervals between any 

 two parallel lines is smaller than w, no spectrum can be pro- 

 duced by it in any case, but only a white ray (the axis) be- 

 comes visible. 



(It is not easy to imagine that the polish which art can pro- 

 duce upon glass, &c. should be mathematically correct. If 



• It would be premature, from an inconsiderable number of observations, 

 to form conclusions respecting the laws of this phenomenon, for it is only 

 by a number of systems of lines, in which « is greatly varied, that it can 

 be deduced with any certainty. Since it is not necessary in these experi- 

 ments that the fixed lines of the spectra should be well distinguished, 

 systems of lines may be made, in which « shall be still considerably smaller, 

 as in the finest glass system of lines that I have hitherto used. It is not 

 improbable that the principles of interference may perhaps yet lead to a 

 theory of the polarisation of light. This is not the place, nor is the time 

 arrived for communicating my view of this subject. Fortunately there 

 are still experiments of another kind possible, which seem to have a le* 

 ference to this object, but they are, like the greater part of all the experi- 

 ments relative to this subject, of a very delicate kind. , 



