Royal Society. 407 



time the F line of hydrogen becomes narrow and well defined. If 

 fresh gas be admitted, the line-spectrnm reappears, accompanied 

 by a widening of the F line. 



That these changes are not dependent on the density of the gas, 

 appears from the following experiment : — 



A sealed tnbe containing nitrogen, with traces of hydrogen at a 

 pressure of about 2 millims., was placed before the spectroscope. 

 A second tube was connected with the air-pump, and the current 

 passed through both tubes, a Leyden jar being placed in the circuit. 



When the pressure in the second tube was high, the line-spec- 

 trum of nitrogen appeared brilliantly in the sealed tube, and the F 

 line was broad and nebulous ; as the exhaustion proceeded these 

 lines faded out, and the F line became narrow, in precisely the same 

 manner as if the sealed tube had been in process of exhaustion. 

 The explanation appears to be that the production of the line- 

 spectrum of nitrogen, and the expansion of the F line of hydrogen, 

 depend entirely on the intensity of the charge communicated to 

 the Leyden jar. When the pressure of the gas between the elec- 

 trodes is high, the discharge does not take place until the jar is 

 fully charged ; but as the exhaustion proceeds a less and less charge 

 is communicated to the jar, and the discharge at last is virtually 

 not more than that of the simple current. 



The same effect may be produced by interposing a break in the 

 circuit, the length of which may be increased as the pressure in the 

 tube is reduced. Pliicker and Hittorf appear to have used a break, 

 as in their paper in the Philosophical Transactions, Nov. 1864, they 

 speak of the expansion of lines obtained by increasing the charge of 

 the jar by an interposed stratum of air. They do not, however, 

 appear to have noticed that the reduction of pressure in the tube 

 was only equivalent to a diminution of the charge of the jar, and 

 that to this cause many of the changes of spectra which accompany 

 the reduced pressure ought to be ascribed. 



We are continuing our experiments on the effect of temperature 

 on the spectrum, but prefer to reserve this portion of the subject 

 for the present. 



May 8. — Francis Sibson, M.D., Vice-President, in the Chair. 



The following communication was read : — 



" Eesearches in Spectrum- Analysis in connexion with the Spec- 

 trum of the Sun." — No. II. By J. Norman Lockyer, F.E.S. 



The observations in this paper are a continuation of those referred 

 to in the previous communication bearing the same title. They 

 deal (1) with the spectra of chemical compounds, and (2) with the 

 spectra of mechanical mixtures. 



I. Chemical Compounds. 



Several series of salts were observed ; these series may be divided 

 into two : — 1st, those in which the atomic weights varied in each 



