268 



Dr, W. J. Russell. 



hydrated crystals and their aqueous solutions give similar spectra ; but 

 with cobalt chloride this is not the case, the hydrate is not the one 

 existing in the solution. The readiness with which the pink crystals 

 of the hydrated chloride become blue arises from the ease with which 

 it parts with water, the blue compound in all cases giving the 

 spectrum of the anhydrous chloride. 



There is another band which is generally visible in moderately strong 

 aqueous solutions at 653 to 638, and is often well seen with a depth of 

 about 3 inches of a 10 per cent, solution. This band, which, from its 

 origin and its constant occurrence appeared to belong to all the cobalt 

 salts, is really due to the presence of nickel. Cobalt salts which have 

 been carefully purified do not give this band, even when very strong 

 solutions are used. 



The oxide of cobalt is spectroscopically an interesting body. It 

 gives, as is well known, a very characteristic spectrum (fig. 10). 

 This is readily obtained by examining the precipitate formed on adding 

 an excess of caustic potash or soda to a solution of any cobalt salt. This 

 blue precipitate (which is said to be a basic oxide) rapidly loses its 

 colour, and the spectrum disappears. The spectrum consists of two 

 very strongly marked bands, one about 665 to 638, and the other from 

 600 to 585. These are very characteristic of the oxide, for on 

 increasing the thickness of absorbing layer, these bands always join 

 together, forming thus a single band, before they increase in other 

 directions. Besides these two bands, when the oxide is precipitated by 

 potash or soda, there is always an absorption at the blue end, begin- 

 ning at about 550, and having very much the appearance of a band, 

 but rendered somewhat indistinct by the general absorption taking 

 place in the blue. At the red end of the spectrum the absorption ends 

 at about 675. If ammonia be used as the precipitant in place of potash 

 or soda, then there is clearly no band in this part of the spectrum 

 and no absorption at 540 (fig. 11). This difference between the two 

 spectra seems clear and definite, and is of much interest, and the 

 obvious deduction is that the two bands at 640 and 585 are produced 

 by the oxide, and that the absorption at 540 is produced by a com- 

 pound of oxide and potash or soda. 



Vogel has pointed out the similarity existing between the spectrum 

 of the precipitated oxide and that given by ordinary cobalt glass. The 

 spectrum of the glass is seen at fig. 10a, and it will at once be seen 

 how nearly this spectrum agrees with that of the oxide precipitated by 

 potash or soda ; the two bands are of the same character as before and 

 very nearly in the same position, but apparently a little nearer to the 

 red, and now clearly a third band from 550 to 525 is visible. The 

 conclusion seems inevitable that the oxide and potash compound which 

 exist in the precipitated oxide exist also in the blue glass. Again, 

 these same spectra can be obtained by another process, and it is easy 



