BY MEANS OF THE RADIOMICROMETER. 43 



evidence that the neodymium salts were solvated. The photographic 

 results were hardly more than qualitative. It was desired to study 

 these " solvent" bands quantitatively. 



The photographic work was done with a grating which gave consid- 

 erable dispersion. It was hardly to be expected that the prism in the 

 prism spectroscope which we used would disperse the light sufficiently 

 to enable us to study the two "solvent" bands. We could, however, 

 by means of the prism spectroscope and radiomicrometer, learn some- 

 thing of the difference in position and intensities of the "solvent" 

 bands, and at least gain an idea as to how to proceed when, in the near 

 future, we shall study this problem in detail with a large new grating 

 spectroscope now completed, and with a still more improved radiomi- 

 crometer. 



Seven solutions were prepared, each containing neodymium chloride 

 of the concentration 0.141 normal. The two solvents were water and 

 absolute ethyl alcohol. As recorded in table 15, the solvents contained 

 the following amounts of water: 100, 50, 25, 13, 9, 3, and per cent. 



The neodymium chloride was of course dehydrated. This was 

 effected by heating for a long time at about 150 in a current of dry 

 hydrochloric-acid gas. All the solutions used were perfectly clear. 

 The depth of layer employed in determining the percentage transmis- 

 sions for each solution was 10 mm. The results recorded in the table 

 enable us to study three neodymium bands for each solution. It is 

 evident that the greatest changes in the percentage transmissions take 

 place in those solutions which contain more than 90 per cent alcohol. 

 The bands for only three of the solutions have been plotted (fig. 19). 

 They are the bands for neodymium chloride in 100 per cent water, in 

 97 per cent alcohol, and 100 per cent alcohol. The heavy curve nearest 

 the violet end of the spectrum is for 100 per cent water. The dotted 

 curve is for 97 per cent alcohol, while the curve represented by dashes 

 is for 100 per cent alcohol. 



It is evident that the spectrum does not show any indications of a 

 hydrate and an alcoholate band, both bands appearing as one. This 

 was anticipated, since the dispersive power of the glass prism is so 

 slight. Near the centers of the bands, on both the descending and 

 ascending arms and at the center of the bands, the wave-lengths of 

 light recorded in the table are those which would be obtained by turning 

 the calibrated head of the spectrometer one division. This increased 

 or decreased the wave-lengths of light through this region by about 

 40 A. u., according to the direction in which the drum-head was turned. 

 Increasing the wave-length to this extent would make it impossible to 

 detect the two "solvent" bands. Accordingly, the drum-head was 

 turned through increments of about 5 A. u. and the transmission deter- 

 mined. This was done for the mixed solvents for all three bands given 

 in the curves, and also for a few of the bands which showed best the 



