of the oxides of Didymium and Erbium, etc. 203 



or even by Faraday wax. Between the end of the spectroscope 

 and the end of this cylinder is a lens d of 3 inches focal length for 

 condensing the light on the slit, and a similar one f placed at the 

 further end of the cylinder. There is a screen g, having a hole 

 of about ^ of an inch diameter, to allow the light to pass through 

 from a Welsbach incandescent lamp h placed a few inches away 

 from it. There is also a tube G of exactly the same internal 

 diameter as the tube B, and whose lower end is closed by a glass 

 end in a similar way to those fixed at the ends of B. This tube 

 G is held vertically by means of a clamp and stand Z, and it has 

 a scale carefully marked on it or attached to the outside. In the 

 tube G there is another tube X of about f of an inch diameter, 

 the upper end of which is open, and the lower end is also 

 closed by plain glass. The tube X is so fixed to the stand that it 

 can be easily raised or lowered either by a clamp or by a rack and 

 pinion. There are lenses at k and o for condensing the ray of 

 light in its passage through the circular hole in the screen p, and 

 its reflection by means of the right-angled prism m. s is a 

 Welsbach incandescent light similar to that at h ; and at e 

 there is a right-angled prism to reflect the ray of light into the 

 spectroscope issuing from the solution in G. The incandescent 

 lights at s and h are so arranged that their spectra are equally 

 bright as seen through the spectroscope when the tubes B and 

 G are removed. 



Method of experiment. 



The method is to have a definite volume of a definite strength 

 of a pure salt of either of the earths of Didymia or Erbia placed in 

 the perpendicular tube G. The chlorides are preferred because there 

 is no change in the bands when their strong solutions are diluted. 

 As the amount of the oxide contained in 1 c.c. of the solution is 

 therefore known, the scale attached to G enables one to read off 

 directly the amount dissolved in any fractional part of the volume 

 of this solution. 



The solution to be examined, or an aliquot portion of it, is 

 placed in the tube B, whose volume is also exactly known, and so 

 that its level at the lower end of the side tube Y is parallel with 

 the inside wall of the cylinder. There will be a meniscus, of 

 course, but it is so small that its effect may be neglected. The 

 parallel spectra are then examined through the eye-piece of the 

 spectroscope and the cylinder G is moved upwards or downwards 

 until the intensities of the absorption bands of the two solutions 

 are identical. As the intensities and widths of the bands depend 

 upon the amount of the oxides dissolved, it follows that, if the 

 bands in the solution in G measured from the bottom of the 



