540 PROCEEDINGS OF THE AMERICAN ACADEMY. 



The absence of several of the strong lines of the spark spectrum is of 

 interest. (A 3854.0, A 4242.0, etc.) The strong pair of spark lines at 

 X 4245.0 and A 4387.0, which is entirely absent from the arc spectrum, 

 is present in the Wehnelt, but with greatly diminished intensity. This 

 pair is one like those already referred to under zinc and tin, differing 

 from the other lines of the spark spectrum in appearance by beino- 

 broader and more diffuse. 



The spectrum of a carbon point in various acids has also been photo- 

 graphed. (See Plate 1, Fig. 1.) None of the plates show any lines or 

 bands which are not directly traceable to impurities in the carbons used. 

 About 90 lines of iron, 12 of calcium, and 2 of aluminium were measured 

 in one case. These lines were superimposed on a continuous spectrum 

 of considerable intensity, but no bands were observed. 



Preliminary tests on salts containing complex ions have given results 

 of interest. A carbon or platinum point in a solution of potassium ferro- 

 cyanide gives lines of potassium and iron and no new lines. Potassium 

 chromate as electrolyte absorbs all but a small strip of actinic light, but 

 of a dozen lines measured in the vicinity of the D-lines, two were lines 

 of potassium and the rest were lines of chromium, the trij^let A 5204- 

 A 5208 being strong. 



Besides the metals described the following have been photographed 

 in the Wehnelt, both in the first and third spectrum : — copper, nickel, 

 iron, gold, palladium, platinum. 



As has been already stated, no effect of the anion has been observed 

 at any time. The breaking up of the chromate ion and the appearance 

 of the lines of metallic chromium in the case of potassium chromate, and 

 the fact that all the salts of the same metal give identical spectra, show 

 this clearly. The facts all point to the high temperature as the cause of 

 the luminescence, and to the probability that the electrolysis plays a 

 purely secondary part. 



Although no effect of the anion is observed, the spectra produced by 

 this arrangement are in many cases compound spectra, if we are to accept 

 the sharp distinction of compound and mefalUc or elementary spectra 

 as evidenced by the presence or absence of bands. Aluminium, marrne- 

 Slum, calcium, barium, give the bands which are ascribed to their oxides, 

 and zinc and tin give fluted spectra which appear to be of tlie same 

 general type. The presence of free oxygen and hydrogen at a tem- 

 perature above the point of dissociation of water affords the possibility 

 of a strong oxidizing action on any metal and the opportunity for the 



