706 



SCIENCE. 



[N. S. Vol. XXIII. No. 592. 



the rainfall stands for a considerable time 

 after falling. jjo^^ rj,_ -^^^^^ 



111 Broadway, New Yoek, 

 April 20, 1906. 



SPECIAL ARTICLES. 



THE AVAILABILITY OP CELLULOID IN ILLUSTRATING 

 CHROMATIC POLABIZATION. 



1. It is not unusual to find that celluloid 

 shows brilliant colors between crossed nicols 

 on the cut edges. This observation suggested 

 the use of the material to illustrate the prop- 

 erties of plates cut parallel to the optic axis, 

 when seen in polarized light in the usual way. 

 In fact, if ■ a strip of celluloid is evenly 

 stretched, fields of color vying in brilliancy 

 with those of the natural crystal, may be ob- 

 tained quite uniformly over an area an inch 

 or more square, and variable at will through 

 two or more well-defined orders; or the color 

 of any given crystal may be similarly in- 

 creased or decreased continuously in order. 

 The well-knovm complicated figures seen in 

 compressed or annealed glass are thus simpli- 

 fied, in a way that is at once interpretable in 

 terms of elementary optical theory. 



2. In the following experiment I used strips 

 of clear celluloid, about 20 cm. long, 1 cm. or 

 less broad (to avoid the need of excessive 

 traction) and but .025 cm. thick (for flexi- 

 bility). They were mounted between rollers, 

 very much in the manner \ised in film cameras, 

 except that one roller was rigidly fastened 

 while the other could be rotated by the aid of 

 a lever and clamped. If many strips are to 

 be simultaneously stretched, it is advisable to 

 secure one end of the strip under a plate of 

 brass, bringing the ends around the remote 

 edge and holding them down under a second 

 smaller plate, in order that the maximum of 

 friction may be encountered. The roller in 

 this case is preferably a strong hollow brass 

 cylinder (say 2 cm. in diameter) with a cen- 

 tral longitudinal slot. Through this the ends 

 are passed and wedged in place with a conical 

 rod forced into the inside of the tube. About 

 one complete turn should be taken to insure 

 friction. For special purposes instanced be- 

 low, a similar adjustment for stretching at 



right angles to the preceding should be added. 

 Screw apparatus or uniform loads are also ; 



useful in particular cases. ' 



3. As the two directions of vibration are ; 

 parallel to the strain and at right angles to it, j 

 respectively, the nicols may be adjusted at 45 i 

 degrees to the vertical and the pulls be either 

 horizontal or vertical. The phase difference 



ip for a thickness of strip, d, being ; 



<l>^2ird(n — n')\, 



where n and n' are the two indices of refrac- 

 tion for light of normal wave-length, X. The 

 plan of experiment consists in varying n — n' 

 continuously from by increasing stress as ; 



far as the breaking point of the strip, and to i 



increase d successively from the thickness of , 



1 to that of four strips {d = .025 to .1 cm.). i 

 The following results may be recorded: ': 



One strip, d = .025 cm. Colors whitish to j 



middle of the first order. When the strip 

 breaks, the efficiency does not much exceed 

 a quarter wave-length plate. Such a plate of ■ 



mica, where mean and minimum elasticities 

 are involved, is but d = .0032 cm. thick ; a 

 similar plate of selenite, where maximum and '. 



minimum elasticities occur, is .0027 cm. thick j 



for mean wave-lengths. Hence the efficiency j 



of a strip of celluloid stretched nearly to the j 



breaking point is for like thicknesses, about ] 



13 per cent, of that of mica and 10 per cent. i 



of that of selenite. ] 



Two strips, d = .050 cm. The earlier whit- i 



ish colors now become more and more satu- j 



rated and the strips break about at the end i 



of the order. ' 



Three strips carry the phenomenon (when 

 stress is gradually increased) from colorless to 

 the middie of the second order, extremely vivid 

 colors overlying the whole visible area of the 

 strip ; four strips complete the first two series. • 



If more strips are to be used the machine ' 

 must be strong and quite perfect in its clutch, j 



otherwise there is slipping and abrasion, by ; 



which the strength of the strips is decreased. : 



Apart from this the experiments may be car- \ 

 ried into higher orders of color at pleasure. ' 



4. Special Experiments. — If three or four 

 strips of successively decreasing width over- 

 lie each other symmetrically, so as to form a ' 



