Analysis of Magnesium and Carbon. 349 



flame ; and in such a flame there are variable quantities of the 

 uncombined gases, so that the rays in question may really be 

 emitted by oxygen or hydrogen, through mixed up in the 

 photograph with the rays of water. The ratio 3/5, which 

 expresses the relation between A in this group to the wave- 

 length of a corresponding ray in the spectrum of hydrogen, 

 according to the author's theory indicates the ratio between 

 the volume which a given quantity of b in free hydrogen 

 occupies and the volume which the same quantity of b in 

 carbon occupies. So that if the chemical density of b in free 

 hydrogen be 1, that of b in the state in which it occurs in 

 carbon will be 5/3. 



The second group of carbon-lines, for which X is 6577'5, 

 5638-6, 5379-0, 5150-5, 5144-2, 5133-0, 4266-0, 3919, fulfils 

 criteria III. for the element c in the same state as that in 

 which it occurs in oxygen. For 3/5 \ and 5/8 A. agree closely 

 with the observed wave-lengths of lines of the water-spectrum. 

 This group, like the second group of magnesium-rays, forms 

 but a part of the whole group of rays which c can emit when 

 in the state of chemical condensation in which it occurs in 

 oxygen and in magnesium. And this partial appearance of the 

 group depends on the intensities which are determined by the 

 influence of the other elements with which c is combined. 



The third group of carbon rays, for which A, is 3919, 3881"9, 

 3877, 3875-7, 3870-7, 3589-9, 3584°8, 3583-3, 3167-7, 2993-1, 

 forms, like the third group of magnesium-lines, a part of the 

 group of lines which b can emit when it is in the state in 

 which it exists in free hydrogen. It satisfies the same tests 

 as the corresponding magnesium group and with equal com- 

 pleteness. That b in carbon should give an entirely different 

 group of rays from that given by the same matter existing in 

 the same state of condensation in magnesium, the author ex- 

 plains, as before, by the variations of intensity induced by the 

 other elements in the two substances. 



The lines of the fourth group, for which \ is 3166, 2993*1, 

 2967-3, 2881-1, 2837-3, 2836-7, 2835-9, 2746-5, 2733-2,2640, 

 and 2523*6 (silicon-line), like the fourth group of magnesium- 

 lines, are also, within the limits of error of observation, lines 

 of the water-spectrum. Also 5/4 X satisfies, to the same 

 extent as in the case of the corresponding group belonging to 

 magnesium, the criteria II. for the wave-lengths of rays due 

 to b in the same chemical state as that in which it occurs in 

 the hydrogen in water-vapour ; and, as before, the author 

 thinks that the ratio 5/4 indicates the condensation of b, in 

 the form in which it exists in carbon and produces this group, 

 as compared with its condensation in free hydrogen. 



