Lines of Low Temperature. 243 



lines observed have been identified by means of comparison with 

 Geissler's tubes showing known spectra. The wave-lengths of 

 the lines in the tubes so used for comparison I afterwards 

 measured with a grating, and verified by Dr. Watts's catalogue. 

 The light of the thermometer itself is too feeble to permit of 

 direct measurement with a grating. 



X= Dr. Watts's No. 

 a a . . orange 5768 53*6 



c c . . green 5460 63*3 



hh . . violet 4352 124-2 



These are three of the brightest mercury-lines. The tube of 

 comparison was one in which, previously to exhaustion, some 

 mercury had been vaporized. I mention the grounds of my 

 conclusion, because in Wullner's experiments it was shown that 

 mercury required considerable heat before its' lines could be 

 made to appear (Pogg. Ann. vol. cxxxv. p. 512). No one, how- 

 ever, who examines these lines will doubt that they are mercury- 

 lines; and they certainly are produced without any appreciable 

 rise of temperature. They appear in every thermometer I have 

 examined. There was a fourth line (dd) in one of Casella's 

 thermometers, apparently coincident with mercury-line No. 84*8 

 of Dr. Watts; but it was so feeble that the cross-wires could 

 not be brought upon it with any certainty. 



But in chemical thermometers purchased from six makers of 



slightly less repute than those above named, four other lines 



made their appearance in addition. These do not coincide with 



the mercury-lines, nor with the lines of a hydrogen-tube, nor 



with the lines of air ; but they coincide exactly with some of the 



strongest lines of tubes containing carbon compounds. I find 



them in carbonic acid, in paraffin-oil vapour, in hydrocyanic 



acid, in cyanogen, in olefiant gas, in methyl, in coal-gas, and in 



carbonic oxide. Their wave-length has been verified with a 



grating as before. 



Thermometer-line. Dr. Watts's carbon spectra. A. 



bb (j) 58, yellow 5602 



ee (k) 74, green 5195 



// (/) 92, blue 4834 



gg (m) 112, violet 4505 



Many observers have been plagued by the intrusion of car- 

 bon-lines in various tubes where they had no place. Wiilln el- 

 even assigns to carbon one of the various spectra which oxygen- 

 tubes have been found to present. These carbon-lines are at- 

 tributed by him to the grease surrounding stopcocks ; by others 

 (I think, by Mr. Crookes) to caoutchouc joints. These explana- 

 tions are clearly not applicable to thermometers; and I think a 



R2 



