Undlffracted Long-wivel Heat Rays by Wire Gratings. 323 



light from such metallic layers is also polarized at right 

 angles. On the other hand, the ordinary diamond-ruled 

 glass gratings polarize, in the Hertzian sense, up to 6 per 

 cent.; they are, of course, not so well denned as the regular 

 wire gratings. The latter also produce ellipticity. 



In continuing this work Dr. Franz Braun* investigated 

 platinum, copper, iron, and gold wire gratings immersed in 

 air, carbon bisulphide, methyl alcohol, and benzene. The 

 ratio of transmissibility n 2 = Q s /Qp diminished with decreasing 

 wave-length, no matter whether as a consequence of change 

 of colour or of the refractive index of the surrounding medium. 

 Increase of temperature up to 250° had no perceptible in- 

 fluence. The sense of the polarization was always the same 

 also when, for example, the gold wires were brought as near 

 to one another as possible ; on the other hand, no effects of 

 any sort were observed with quartz threads. 



Similar experiments were carried out by Prof. Ambronn f 

 with gratings made of sewing-needles, prepared in the same 

 way as by v. Mohl and Hof meister. He repeated Fizeau's 

 experiment with rulings less than 0*1 //, in width on a very 

 thin silver film ; he found, among other results, that many 

 of these rulings readily 1 1 polarize red and yellow light, while, 

 on the other hand, blue or violet light is transmitted J_ 

 polarized. 



§ 2. In 1893 we investigated, in the infra-red, eight 

 gratings of platinum, copper, iron, gold, and silver wire, the 

 apparent slit- widths b' of which varied between 10 and 60 /j,; 

 since wave-lengths from 0'8 to 5/a were used, this investiga- 

 tion thus extended over the region 



2<6'/\<75, 



while in Hertz's analogous investigation b'/X was only 0*05. 



The ratio of transmissibility n 2 = Q s /Q p could, in pursuance 

 of the optical measurements, be determined as a function of 

 the wave-length. In the case of each of the five metals it 

 very soon reached a maximum in the infra-red (n 2 >l, up to 

 nearly 1*5) with wave-lengths 0'8-l*3yLt: an inversion 

 (n 2 = l) occurred later of such a kind that from that point 

 onwards the polarization of longer waves followed according 

 to the Hertzian sense (n 2 <l). 



For each metal the ratio of the characteristic wave-lengths 

 corresponding to the above-mentioned maximum and inversion 

 points was about 1 : 2'35, and appeared independent of the 



* Franz Braun, Inaug.-Diss., Berlin 1896. 



t H. du Bois and H. Rubens, TVied. Ann. xlix. p. 593 (1893), further 

 on referred to as /. c. 



Y 2 



