Dr. Miller on the Spectrum of Thallium. 229 



Two pieces of stout thallium wire were then arranged as electrodes 

 to the secondary wire of an induction coil. A continuous torrent of 

 sparks was maintained without melting the wires or producing very- 

 rapid oxidation, or volatilization of the metal ; the light, however, 

 was much whiter than its ordinary monochromatic character would 

 have led us to expect. Mr. Crookes, who was with me during the 

 experiments, projected the image of the points by means of a lens 

 upon a distant white screen, when it was at once obvious that 

 the extremities of the spark were of a fine green colour, whilst the 

 flickering luminous arc, which filled up the interval, due chiefly to 

 ignited air, was much whiter. 



On viewing the sparks from the induction-coil by the spectroscope, 

 several new lines, independently of well-marked air-lines, made their 

 appearance. These lines were distinguished from air-lines by the 

 peculiar character which distinguishes most metallic lines, viz. the 

 much greater intensity of their extremities than of their central por- 

 tions. Besides the usual intense line in the green, five others were 

 particularly observable : first, a very faint one in the orange ; next, 

 two of nearly equal intensity in the green, more refrangible than Tla, 

 with a third much fainter, these three lines in the green being nearly 

 equidistant; whilst, 5th, in the blue was a bright well-defined line : 

 all these were strong at each extremity and evanescent in the central 

 portions. 



The induction- spark of thallium was then observed when produced 

 in a current of hydrogen gas. The air-lines disappeared, the pecu- 

 liar lines of hydrogen were very manifest, particularly the line in 

 the red and one of the lines in the blue ; whilst the new thallium 

 lines were preserved, with the exception, of the feeblest, though all 

 were reduced in intensity. 



Finally, a photographic impression of the thallium spectrum upon 

 collodion was obtained by the method which I have described in a 

 paper communicated to the Royal Society in June last. An impres- 

 sion extending to about division 154 of the scale then adopted was 

 obtained. This spectrum contains several very characteristic groups 

 of lines ; it recalls the features of the spectra of cadmium and zinc, 

 and less strongly that of lead. 



Measuring by the scale already adopted in my former paper, it is 

 found that there are two strong groups of lines at about 103 and 

 106. At 116, 121, and 126 are three groups — the first two less 

 intense than the third, which is of about the same strength as the 

 earliest two. Several feebler pairs of dots follow, and the spectrum 

 terminates rather abruptly with four nearly equidistant groups, com- 

 mencing respectively at 136, 141, 145, and 151. The first of these 

 groups is very strongly marked, the others are fainter, but of nearly 

 equal intensity. 



The remarkable way in which a spectrum at low temperatures so 

 simple becomes increased in complexity, both in the visible and in 

 the extra-visible portions, is of high interest considered in relation 

 to the physical cause of these phenomena; and it is not without 

 interest in a chemical sense, from its bearing upon the view sup- 

 ported by Dumas, that thallium belongs to the alkaline group » 



