ON OUR KNOWLEDGE OF SPECTRDM ANALYSIS. 259 



is sometimes adopted ; it presents no advantages, but, on the contrary, 

 may give rise to a good deal of misconception. We shall not use the 

 expressions. A spectrum is called a continuous spectrum when it extends 

 over a wide range, and is not broken up into separate lines. It is, how- 

 ever, not necessary that it should extend through all the colours. We 

 may have a continuous spectrum in the green without an admixture of 

 red and blue, and we often have continuous spectra which are confined to 

 one end of the spectrum, either to the red or to the violet. 



The spectra of fluted bands or channelled spaces generally appear, 

 when seen in spectroscopes of small dispersive power, as made up of 

 bands, which have a sharp boundary on one side and gradually fade away 

 on the other. When seen with a more perfect instrument, each band 

 seems to be made up of a number of lines of nearly equal intensity, which 

 gradually come nearer and nearer together as the sharp edge is approached. 

 This sharp edge is generally only the place where the lines are ruled so 

 closely that we cannot distinguish any more the individual components. 

 The edge is sometimes towards the red, sometimes towards the violet end 

 of the spectrum. Occasionally, however, the bands of channelled space 

 spectra do not present any sharp edge whatever ; but are simply made up 

 of a series of lines which are, roughly speaking, equidistant. In small spec- 

 troscopes these bands appear to be altogether homogeneous, presenting a 

 fairly sharp edge on both sides. A body, as we shall see, may have more 

 than one spectrum of the same kind. 



Variations in the spectra of gases are generally obtained by a sufficient 

 alteration in the intensity of the electric discharge, which renders them 

 luminous. We shall call the discharge which passes, when the electrodes 

 are connected directly with the terminals of the induction coil, ' the 

 ordinary discharge,' in contradistinction to the ' jar discharge,' in which 

 each terminal is also connected with one of the coatings of the Leyden 

 jar. In order to get the best effect with the jar discharge, it is generally 

 necessary to interrupt the circuit in some part, so that a spark is forced 

 to break through the air whenever the discharge passes. 



II. Nitrogen. 



o 



Angstrom: ' Pogg. Ann.' xciv. p. 158 (1865). 



Pliicker: ' Pogg. Ann.' cv. p. 76 (1858) ; cvii. p. 519 (1859). 



V. d. Willigen: ' Pogg. Ann.' cvi. p. 618 (1859). 



Huggins : 'Phil. Trans.' cliv. p. 144 (1864). 



Pliicker and Hittorf : 'Phil. Trans.' civ. p. 1 (1865). 



Brassak: ' Abh. Nat. Ges. Halle,' x. (1866). 



Wiillner : ' Pogg. Ann.' cxxxv. p. 524 (1868) ; cxxxvii. p. 356 (1869); 



cxlvii. p. 326 (1872) ; cxlix. p. 103 (1873). 

 Salet : ' Ann. Ch. Phys.' xxviii. p. 52 (1873) ; C.R. Ixxxii. p. 223 : 

 „ 274(1876). 

 Angstrom and Thalen : ' Nov. Act. Ups.' (3), ix. (1875). 



The Line-spectrum. — This spectrum appears whenever a strong spark 

 (jar discharge) is taken in nitrogen gas. It is always present when 

 metallic spectra are examined by the ordinary method of allowing the jar 

 discharge to pass between two metallic poles. A good knowledge of this 

 spectrum, which is very rich in lines, is important in all cases where 

 an electric discharge is used for spectroscopic analysis. The spectrum 

 has been studied especially by Huggins and Thalen. The latter has 



S2 



