218 



KNOWLEDGE 



[October 1, 1894. 



compared ; no sign of resolution into lines could be seen, 

 even with a highly dispersive train of five prisms. 



Next, a ray of white light from the flame of a paraffin 

 lamp was made to pass through the glowing tube and enter 

 the spectroscope, after having suffered absorption by the 

 heated vapour. The dark bands characteristic of cooler 

 iodine were seen to be unchanged, and there was no sign 

 of any continuous absorption. Thus, at the temperature 

 of the experiment, iodine emits a continuous spectrum, and 

 does nut emit only those rays which it absorbs. 



The second question, as to whether this curious phe- 

 nomenon of a gas glowing with continuous light like a 

 solid was a property special to iodine, was soon answered 

 in the negative. ThiTS bromine was tried with exactly the 

 same results, the glow being quite as conspicuous as that 

 of iodine. Then chlorine was attacked, and after several 

 unsuccessful attempts, this also was found to give a perfectly 

 distinct glow of the same character as the others. The 

 experiment was arranged as follows ; chlorine from a 

 generating flask was thoroughly dried by passing slowly 

 through a long tube of calcium chloride ; the dry gas was 

 then passed along a straight piece of combustion tube heated 

 strongly in the middle, this tube being connected with a 

 larger glass tube, in which was placed a small reflecting 

 prism, which allowed one to look along the central bore of 

 the hot tube. With the room in which the experiment was 

 performed made as dark as possible, and the eye screened 

 from the light of the Bunsen, a faint but perfectly distinct 

 glow was seen to gradually fill up the tube, beginning on 

 the lower or hottest portion. With a spectroscope of very 

 small dispersion and wide slit, the light appeared to be 

 continuous, but it could only be traced between the positions 

 of D and E of the solar spectrum, probably because of 

 the greater sensitiveness of the eye for this region. There 

 seems to be no more reason to suppose any chemical 

 " luminescence " in the case of dried chlorine than in that 

 of iodine or bromine. 



After these experiments it was thought probable that all 

 colonri'd vapours would be found to glow in the same way. 

 Accordingly the vapours of sulphur, selenium, and arsenic 

 were tried in an atmosphere of nitrogen, and all were 

 found to glow faintly. Phosphorus, however, which gives 

 a colourless vapour, appeared to give no light when free 

 from every trace of oxygen. 



The conclusion, therefore, is that all the elements 

 ■which have been named, except phosphorus, can be made 

 to emit Hght by the mere application of heat, but that 

 probably in every case Kirchoff's law is entirely disregarded, 

 the emission not corresponding in any way with the 

 absorption spectra, and for this reason one cannot compare 

 these luminous vapours with the chromosphere radiations, 

 or with discontinuous radiation generally. It might still 

 be contended that line spectra can only be generated by 

 chemical or electrical luminescence. 



My next endeavours were, therefore, directed to the pro- 

 duction of metallic line spectra by heat, rigidly excluding 

 chemical action. Sodium was the element chosen for the 

 initial experiments ; the powerful absorption produced by 

 the vapour of this metal on yellow " D " hght seemed, on 

 Kirchoff's hypothesis, to give the best chance of success at 

 the very moderate temperatures at my command." 



The apparatus used in this research, although apparently 

 rather complicated, is really simple enough and can easily 

 be rigged up by anyone at all famUiar with chemical mani- 

 pulations. It was designed with the object of excluding 



* My heating outfit consists of a single Bunsen burner with an 

 oblongaperture, supplied from a lialf-incli gas main, and giving a flat 

 flame about six inches high and two inches wide. 



absolutely from the tube which was to be heated any gas 

 that might be expected to act chemically on the vapour of 

 the metal. In the diagram Fig. 1, A and B are two 

 similar gas holders, improvised by inverting a couple of 

 two gallon tins mouth down into larger iron vessels 

 containing strong brine ; guides are provided, so that 

 the tins will float up perpendicularly when full of gas, 

 a stopcock is soldered into the bottom plate of each 

 tin (at the top in the position used), and a rubber tube 

 leads from A to a couple of wash bottles (S' S") con- 

 taining strong sulphuric acid ; from S" a long tube of hard 

 glass (Pi containing a little phosphorus leads into the 

 drying tube (C), which is packed with calcium oxide and 

 calcium chloride, the former to remove carbonic acid and 

 the latter the last traces of water which may remain in 

 the gas used after passing the sulphuric acid bottles. The 

 drying tube connects on to the porcelain heating tube (H) 

 through a metal T-piece, one end of the T having a glass 

 plate carefully cemented in so that one may look along the 

 inside of the heating tube, to which the T is connected by 

 rubber tube tightened with wire, both connections being 

 also buried in sealing wax. At the other end of H, which 

 is covered in the centre by a tire-clay arch, is a second T 

 of glass, one limb connecting with a glass gland or stufllng- 

 box (G) with pierced rubber ends and filled with mercury. 

 A long steel rod passes through the gland, the end being 

 flattened to a spoon shape ; this can be pushed along to 

 the centre of H or drawn out past the entrance of the 

 side tube of the T. This side tube is closed by a perforated 

 rubber stopper, through which a small glass tube passes 

 bearing a small reflecting prism cemented to the end, 

 which is thus closed up ; but in order to allow of the 

 escape of the gases, so that a current may be set up in the 

 apparatus, a hole is blown in the s'de of this tube near the 

 prism. The outer end of the tube is connected by rubber 

 tubing to another wash bottle (S'") containing sulphuric 

 acid, and from this again a tube leads to the gasholder B. 

 Thus the entire apparatus forms a closed circuit and has 

 no inlet or outlet. The gasholders have each a Y tube 

 attached, one branch of the Y leading to the apparatus 

 and the other connecting A with B by means of a rubber 

 tube carrymg a clip. The S at the foot of the diagram is a 

 direct vision spectroscope, and L is a lens focussing the 

 central parts of the tube H on to the slit of the instrument ; 

 both are attached to a strip of hard wood movable hori- 

 zontally about an axis placed between the lens and the 

 slit. This enables one to instantly push aside the spectro- 

 scope into the position shown by dotted lines, and observe 

 the glowing tube directly. 



The modm operandi of an experiment is as follows : 

 First, one fills the gasholder A with some neutral gas 

 containing no oxygen, or only a trace of that element, 

 such as nitrogen or hydrogen, or ordinary coal gas. 

 When full it is disconnected with the gas generator or 

 gas main, as the case maybe, and connection is made with 

 the apparatus. Next, weights are put on the inverted tin 

 until sufficient pressure is obtained to drive a current of 

 gas thi-ough the wash bottles, drying tube, &c., into B, 

 which will rise as A sinks. Then the Buuseu is lighted 

 under the porcelain tube, which it presently heats up to a 

 white heat for about two inches of its length. After 

 sufficient dry gas has passed through, and all trace of 

 moisture has gone, the current is stopped by closing the 

 stopcock on A, and a small pellet of sodium is dropped 

 into the steel spoon through the side tube of the glass T, 

 the stopper with the inner tube and prism being removed 

 for tills purpose and quickly replaced ; the current is tLen 

 restarted in order to drive away any oxygen that will ha^e 

 diflused in by the operation, and at the same time the tube 



