348 L. Bell — Rahiband Spectroscopy. 



wendung des Spectral-apparates ") or some modification, like 

 that described by the writer (American Chemical Journal, vii, 

 No. 1). The ideal scale therefore should be constructed so 

 that the absolute intensity of its various readings could be 

 found by the above or some other method. 



Janssen and others have tried various rude methods of esti- 

 mation, but the only scientific instrument for measuring ab- 

 sorption lines yet devised is that due to Mr. C. S. Cook, and 

 described by him in " Science," ii, 488. His plan consists, in 

 brief, of forming diffraction fringes in the field by a silk fiber 

 placed a trifle beyond the focus, and varying their intensity by 

 moving the fiber to or from the focus by means of a screw- 

 collar. The fringes thus produced resemble the lines of the 

 rainband quite closely, and form a very reliable and delicate 

 scale. Its successful use, however, requires a skilled observer 

 and much care, and its indications cannot be reduced to abso- 

 lute measure except by a comparison with some absorption 

 line which can in turn be evaluated with a spectro-photometer, 

 itself by no means an easy matter. While this method is very 

 effective in measuring the individual lines of the rainband 

 some plan for measuring the band as a whole seemed desirable, 

 and, after some casting about, the writer determined to give a 

 careful trial to the interference bands produced in the spec- 

 trum by a crystal in polarized light. So far as known, this 

 method has not hitherto been used. 



The instrument used was a direct-vision spectroscope by 

 Schmidt of Berlin, which gave a very bright spectrum with 

 moderate dispersion and beautiful definition. Its five prisms 

 polarized well enough to make it a very efficient analyzer. A 

 bit of selenite having a well defined natural edge was then 

 split down till the interference bands produced by it were 

 about as wide as the maximum rainband as seen in the above 

 instrument. The plate thus prepared was placed immediately 

 in front of the slit, covering, however, only half the field. 

 The usual cap of the spectroscope was removed, and replaced 

 by a tube graduated to five degrees and carrying a 9 mm Nicol 

 prism. Thus equipped, a long series of preliminary experi- 

 ments were made to settle upon the best method of observa- 

 tion and adjustment of instrument. 



Finally the following plan was adopted. The thickness of 

 the selenite plate was so adjusted that the bands appear as in 

 fig. 1, one of them being situated about the average width of 

 the rainband on the less refrangible side of D, i. e. in the red. 

 Then the Nicol prism was turned till the interference band was 

 of the same intensity as the rainband, when the field appeared 

 of the same brightness from the red side of the interference 

 band to D. In this method of observation it is quite as well 



