38 



KNOWLEDGE 



[FKnurARv 1, 1890. 



historical aspect, taking the various fabrics, as far as 

 possible, in order of date, and pointing out the special 

 characteristics of each. It may, however, be advisable 

 first to say a few words on the chronology of the dittbreut 

 periods. 



Roughly speaking, we may divide Greek vases into four 

 great classes, as follows : — 



A. — Vdsi's of tlw rn'i/iitiri' Period, down to about (iOO B.f. ; 

 decoration in brown or blacl; on a groimd varying from 

 white to pale red, often unglazed ; human figures and 

 mythological scenes rare. 



R. — IHack-ii(iurid Vokcs, from 60Q to 500 b.c. ; figures 

 painted in black, like silhouettes, on glazed ground 

 varying from cream colour to bright orange-red, with 

 background of lustrous black varnish ; mostly mythological 

 scenes. 



C. — Red-fif/ured Vases, from 520 to 350 n.c. ; figures 

 outlined on red clay, and background filled in round them 

 with black varnish ; inner details painted in afterwards ; 

 scenes from daily life or mythological. With these are 

 included vases with figures in polychrome on white ground. 



J). — Wistsof the r>eeiidenee, from 360 to 200 n.c; method 

 of class C, but careless and inferior ; scenes sepulchral 

 or fanciful ; general striving after effect. With these 

 are included various forms of moulded vases or with 

 decoration in relief. 



This classification, though serviceable, is of course far 

 from exhaustive, but will form a convenient di\'ision for 

 the four succeeding article;; which will be necessary to 

 complete the subject. It must, however, be borne in mind 

 that the transition from one to the other is by no means 

 clearly marked, and any such distinctions as these must 

 perforce be more or less conventional. 

 (To be continued.) 



WHAT IS A NEBULA? 



By E. Waltkr Maundek, F.E.A.S. 



IN my previous paper* it was my wish to lay stress on 

 the thought that our solar system must not be taken 

 as the type of all systems ; that the simplicity of its 

 construction — practically the entire mass of the 

 system being locked up in one central body— must 

 not be supposed to represent, certainly does not represent, 

 the condition of things in other regions of space. To my 

 own mind, indeed, it seems the more probable that such 

 centralized systems are the exception rather than the rule. 

 Tossibly the isolation of the solar system has had much to 

 do with this marked autocracy of the sun ; if so, we might 

 expect a greater amount of decentralization where two or 

 more systems are in closer neighbourhood to each other. 



I wish on the present occasion to suggest that, just as 

 we must not take the mode in which matter is distributed 

 in our system as a type of its distribution in other systems, 

 so neither must we take its distribution in our sun as a 

 type of its distribution in other suns. 



In our own sun, as I have before had occasion to point 

 out in Knowledge,! by far the greatest portion of the 

 sun's mass lies below the photosphere. From the point of 

 view of mass, the chromosphere prominences and coronre 

 may be neglected. They are probably less important rela- 

 tively to the sun than our atmosphere is to the earth. 



What is even more germane to our present point, they 

 are practically negligeable as to the amount of light which 

 they emit. In spite of the fact that the corona has been 

 traced to a distance of six solar diameters from the sun, as 



* Knowledge for ISovember, 1895. 

 'The Temiity of the Sun's Surroimdiugs," March, 1894. 



by Newcomb in 1878, and that prominences of a total 

 elevation of three hundred thousand or four hundred 

 thousand miles have been recorded, it still remains the 

 fact that the corona can only be seen when the sun itself 

 is hidden, and the prominences only by the artifice of the 

 spectroscope. Were the sun removed to stellar distances 

 we should have no hint of their presence. 



r>ut though these appendages are not sufficiently 

 important in the case of the sun to reveal themselves 

 from stellar distances, we are not without Eome direct 

 indications of similar formations in the case of some 

 stars. As to prominences, the continual increase in the 

 number of stars showing not only the lines of hydrogen 

 bright, but the helium lines as well, the recent discovery 

 of a bright (' line in some well-known stars, and the 

 confidence with which we look for i.hese lines in the 

 spectrum of any Nova which flashes out upon our view, 

 point unmistakably to vastly developed prominences and 

 chromospheres as inseparable appendages of a large pro- 

 portion of stars ; whilst C^ampbell's remarkable observation 

 of a positive hydrogen disc to the star DM -f 30°, 

 No. 3059 {Aslron. <ind Asirojili., XII., p. 913), stands out 

 as an example more striking in its character than we could 

 have possibly expected. 



The indications of stellar corouii' are much slighter. 

 We could not expect it to be otherwise, since as yet we 

 have secured no clear and indisputable record of the corona 

 of our own sun, except during a total eclipse. And the 

 evidently composite character of the coronal spectrum 

 renders it all but hopeless that the spectroscope should 

 assist us here. In effect, the coronal light is chiefly 

 reflected sunlight and white light from incandescent dust 

 particles. Neither of these sources could give us any sign 

 by which to recognize them m the surroundings of a star. 



The only available means of detecting a corona is in its 

 faint bright-line spectrum. This consists mainly of the 

 hydrogen and helium lines, and the so-called " coronal " 

 line — 1474 on Kirchhofl's scale. But even here it is a 

 moot point whether we should not rather consider these 

 bright lines as properly belonging to the chromospheric 

 and prominence region than to the corona itself. At best, 

 then, our one chance of any hint of a stellar corona is 

 the occurrence of the bright coronal line— 1474 K. 



It seems to be sufficient, then — it is the utmost the 

 probabilities of the case would lead us to expect — that on 

 two occasions Xorn' should have shown the 1474 K. 

 line : Nova Cygni in 1870, as observed by Cornu, Vogel, 

 and Backhouse, and Nova Aurigie in 1892. 



These hints as to the existence of stellar chromo- 

 spheres and coromp do more than warrant us in concluding 

 that the stars in general possess a structure which in 

 these features resembles that of the sun. They prove — 

 what, indeed, I claim that we ought on i! priori grounds to 

 expect — that in some instances the stellar chromospheres 

 and coronas are of vastly greater relative importance than 

 in the sun. The scheme of the general structure is 

 probably alike in all, whilst the proportion which the 

 different details bear to one another varies indefinitely. 



These considerations seem to me to have a distinct 

 bearing on the problem of the gaseous nebulie. It is 

 hard enough to understand how we can have gaseous 

 masses of such enormous extent. The difficulty is 

 increased when we bear in mind how extreme must be 

 their average tenuity. It will be remembered that Mr. 

 Ranyard showed that in the case of the Orion nebula, 

 evidently one of the densest, there was good cause to 

 think that its mean density could not exceed one ten- 

 thousand-millionth of that of our atmosphere at sea-level. 

 To this we have to add the yet further difficulty that the 



