430 



NA TURE 



{April 2, 1874 



earth, and this being so, we can chemically examine them as 

 easily as substances which have consolidated here. The^e bodies 

 may be roughly divided into iron meteorites and stone meteorites, 

 and roughly again into meteorites which contain elementary 

 metallic molecules, and others which contain compound mole- 

 cules, that is, mixtures of metals with the metalloids. 



The observations of the spectra of nebulre we owe mainly to 

 Mr. Huggins, a distinguished English observer, who has given 

 very much attention to this branch of inquiry. One thing we 

 know for certain about the nebuU-e from these observations is 

 this : we are dealing either with gases alone or with solids bang- 

 ing about in gases, and one of the bright lines which we observe 

 tells us that we are dealing with hydrogen gas ; so that the same 

 method of inquiry which, applied to comets, tells us that pro- 



Fig. 3.— Spectrum of the nebula. — i, 2, 3, lines observed. Above, the solar 

 spectrum is shown from (^ to F ; below, the bright lines of magnesium, 

 nitrogen, barium, and hydrogen, in the corresponding part oi the 

 spectrum. 



bably in some cases we are dealing with a compound of carbon, 

 tells us that we aie chiefly dealing, in the case of nebulce, not 

 with a ccmpound of carbon or with a compound of anything, but 

 with a true element — hydrogen. 



That, you see, is a very great step ; and a very few years ago 

 it would have been considered presumptuous almost to think 

 that man could ever tell what substances were building up the 

 nebulEe which lie at such infinitely remote distances from us. 



So much then for comets and meteors and nebula;. These ex- 



haust the bodies in space which shine by their own light, the 

 light not being subsequently absorbed by an atmosphere through 

 which it pa<ses ; we will now therefore pass to the class of Stars. 



With regard to stars, I have a diagram to bring before you, 

 which summarises in a convenient way a good deal of the 

 work which has been very carefully done by Father Secchi of 

 Rome. I shall have to refer to several other diagrams after- 

 wards ; but this, I think, is the best one to place before you in 

 the first instance. 



Tlie spectra in the diagram are the spectra of various stars. 

 You will at once see that there is a difference between those 

 spectra, and you will see that there is a double difference be- 

 tween some of them. In the first place, you have an extreme 

 simplicity in some cases and complication in the others. But 



Fig. 4. — Ring nebula in Lyra, with its spectrum. 



you see, also, that the question of simplicity and complica- 

 tion is not the only question, that is a question merely of de- 

 gree ; but there is a difference in kind. For instance, you will 

 at once acknowledge a difference in kind between the spectrum 

 of a Herculis and .Suius. 



The diagram of the former, although it has been made in no 

 laboratory, and although it deals with no metalloids or no com- 

 pounds which have been got upon the earth, is as good a diagram 

 as I can put before you to explain what I mean by the channelled 

 structure of the spectrum of the metalloids as opposed to the line 

 spectnun of the metals. 



\ Bearing in mind the great simplicity of the spectra of stars like 

 a Lyrae and Sirius, and the greater complexity of the spectral hnes 

 in a star like the sun ; bearing in mind, also, the difference in 

 kind between the spectra that I have referred to, we can 

 divide all the stars which shine in the heavens which we have 

 already observed, into three classes. This has been done by 

 Rutherfurd and Secchi. 



Let me strengthen what I have to say by showing you rather 

 more elaborate drawings of three stars belonging to these difte- 

 rent classes. You will now see the importance of the considerations 

 which I have brought before you regarding the spectra of the 



metals and the metalloids. There is the channelled space spec- 

 trum of the star a Herculis ; there is the banded spectrum of 

 the star a. Orionis ; and there is the equally banded spectrum 

 of a star in the constellation of the Scorpion. In all these 

 cases you will see we are dealing, not with the first class in 

 which we have simple spectra, but with the second and third 

 classes. 



Now let me contrast 'on another' diagram the spectra of two 

 stars, one in the first class and the other in the second. Let us 

 contrast the spectrum of Sirius with the spectrum of the star in 

 Orion, to which I have already referred. In the spectrum of 



