Dr. J. H. Gladstone on the Violet Flame of many Chlorides. 417 



the centre towards the side where the preponderance is. This 

 is in no case found to be true, very minute quantities being- 

 neglected. The amount of gravity, moreover, at the two spots 

 above described would not be the same, although they would be 

 in the same latitude, one north and the other south. But 

 neither is this result found to be a fact in nature. The only 

 change perceived in the amount of gravity is a very slight 

 increase in proceeding northwards from the equator, an effect 

 which accords precisely with another observed fact, that the 

 number of miles in a degree of latitude increases in proceeding 

 northwards, and therefore the globe is somewhat flattened 

 towards the poles, which would produce this very effect of an 

 increase in gravity, as the ordinary principles of attraction show. 

 I think it is clear, then, apart from all theory, that the mass of 

 the earth is made up of nearly spherical strata arranged about 

 the earth's centre. The law of density of these strata we have 

 no means of determining by this kind of reasoning, nor of ascer- 

 taining even whether the strata first • increase in density and 

 then decrease in approaching the centre, or increase the whole 

 way. It is only the fluid theory which can help us here. 



I would observe that this arrangement of the earth's mass is 

 assumed, avowedly or tacitly, in every calculation I have seen 

 on the subject. Thus, in Professor Stokes's valuable paper " On 

 the Variation of Gravity" (Cambridge Philosophical Transac- 

 tions, 1849), he assumes it in the step where he states that the 

 series 



Y Y 



V = -\ s" + • • » • 



would reduce itself to its first term, if the surface were spherical 

 and the earth had no motion of rotation. This would not be 

 the case unless the sphere were homogeneous, or consisted of 

 concentric spherical shells of matter ; for in that case only does 

 the potential of a sphere for an external point vary inversely as 

 the distance from the centre. 



Calcutta, October 1, 1862. 



LVI. On the Violet Flame of many Chlorides. 



To the Editors of the Philosophical Magazine and Journal. 



Gentlemen, 

 N a recent Number of PoggendorfFs Annalen there is an im- 

 portant paper by Alexander Mitscherlich, in which he shows 

 that the prismatic spectra of the flames of certain compounds of 

 the metals are different from those of the metals themselves. 

 He gives drawings of the spectra obtained from four chlorides 



