404 



Profs. Liveing and Dewar. 



[Nov. 27, 



reversed— 4305, 4865, 4892, 5223, 5225, 5228, 5480, 5503, 5522, and 

 momentarily only, 5155, 5238, 5256. Besides these, many dark bands 

 were observed in the less refrangible part of the spectrum, of which 

 three, with wave-lengths 5920, 6035, 6060, appear to be identical with 

 bright bands ascribed to strontia, and one, with wave-length 6597, is 

 identical with a bright line given by strontium chloride. 



Manganese, introduced as sulphate, gave with facility the violet 

 triplet, 4029, 4032, 4033 as dark lines on the continuous background. 

 The bright blue lines of manganese were not, however, reversed until 

 some metallic magnesium was introduced. This brought out the 

 reversal of the lines, wave-lengths 4753, 4783, and 4823, the last being 

 the most easily reversed of the three. 



Lead introduced in the metallic state gave a reversal of the violet 

 line, wave-length 4058, which Cornu had previously seen reversed, but 

 this reversal was far better seen, becoming a wide black band, when 

 the lead was introduced as an alloy with zinc. Probably the lead 

 vapour was not so rapidly oxidised when mixed with zinc, and a 

 thicker, if less dense, stratum interposed between the arc and the spec- 

 troscope. When lead ferrocyanide was used, not only the line above 

 mentioned was reversed, but also much less strongly, a line near it, 

 wave-length 4062. 



With zinc, only the less refrangible two of the three blue lines, 

 wave-lengths 4810, 4721, were seen reversed. The very bright lines, 

 wave-lengths 4924, 4911, seen in the spark between zinc poles, were 

 not seen by us at all in the arc, resembling in this respect the magne- 

 sium line, wave-length 4481, and the cadmium lines, wave-lengths 

 5377, 5336. 



When cadmium was put into the crucible the lines, wave-length 

 5085, 4799, and 4677, were reversed, not the line, wave-length 4415. 

 With a large dose of cadmium the red line, wave-length 6438, was 

 •once seen reversed for an instant only. 



With silver, besides the reversals before observed by us, the line, 

 wave-length 4053, showed a dark line in the middle of its expansion 

 ;as noticed by Mr. Lockyer, but we could see no reversal of the line, 

 wave-length 4208. Instead of the reversal of this line we observed that 

 a second bright line came out close to it, rather diffuse, and about mid- 

 way between the line 4208 and the calcium line 4215. By measuring 

 the intervals and interpolating, we deduced 4211 '3 for the wave- 

 length of this silver line, which we could see, though it was fainter, 

 in the less powerful current N of a De Meritens' machine. This second 

 line coming out near the other silver line gave the appearance of a 

 reversal in the middle of a diffuse line, but besides the measurements 

 made with a micrometer we assured ourselves of the fact of there being- 

 two lines by watching the fading of the second line as the silver 

 evaporated. The use of an alloy of zinc with silver did not alter 



