138 



SCIENCE 



[N. S. Vol. XLIX. No. 1258 



the saiae way. The other group, typified by 

 U Geminorum and SS Cygni, are normally 

 very faint, but at irregular intervals (of two or 

 three months for the last named star) in- 

 crease rapidly by some four magnitudes, to 

 fade away again after a few days. It is a 

 curiously suggestive fact that the stars of 

 each of these singular classes are very similar 

 to one another ia spectrum, while the spectra 

 of each class as a whole are quite unlike one 

 another or anything else in the heavens. Here 

 is indeed a riddle for the future to solve. 



Finally, we come to the temporary stars — those 

 most spectacular of all celestial objects. To 

 discuss them fully would require another ad- 

 dress comparable in length to this. The mer- 

 est outline must suffice. Before the outbxirst, 

 several of them are knovm to have existed as 

 faint stars, often slightly variable. Without 

 warning and within a very few days at most, 

 their light increases at least a thousand fold — 

 sometimes fully a hundred thousand times. 

 The happy chance by which the recent great 

 Nova in Aquila was caiight midway in the 

 rise indicates that its whole ascent occupied 

 about two days. The maximum brightness is 

 sometimes very great. — Tycho's star of 1512 

 equalled Venus, — ^but a rapid decline sets in 

 almost at once, followed by irregular oscilla- 

 tions with a general downward tendency, 

 merging into a slower but steady decline, till 

 within a decade or so the star has lost eight 

 or ten magnitudes and returned nearly to its 

 original brightness. 



The spectroscopic changes are meanwhile of 

 the most extraordinary character. The three 

 stars which have been caught on the rise 

 showed dark line spectra, roughly resembling 

 familiar types, but no two alike, and with the 

 lines greatly displaced, as if by a huge velocity 

 of approach. As the star goes " over the top " 

 its spectrum explodes, so to speak, in a few 

 hours into a flamboyant afiair of bright and 

 dark lines, enormously widened and displaced, 

 and undergoing continual changes. Lines of 

 hydrogen, helium, and enhanced metallic lines 

 have been recognized. Besides the bright hy- 

 drogen lines in Nova Aquilse there were at 



times two sets of sharp dark lines — apparently 

 due to hydrogen, but displaced by amounts 

 corresponding to velocities of approach of 

 about 1,800 and 2,600 km./sec. Complicated 

 changes occur as the light fades, the most im- 

 portant being the appearance of the character- 

 istic nebular lines, which at some stages are 

 the most conspicuous feature of the spectrum, 

 and remain visible for a long time. After 

 some years, however, they begin to fade, and 

 the last state so far recognized is spectro- 

 scopically identical with the Wolf-Rayet stars. 



Novse show a very strong galactic condensa- 

 tion. Nothing is known of their proper mo- 

 tions, or (for obvious reasons) of their peculiar 

 velocities; but direct measures of parallax in- 

 dicate that the distances of some of the 

 brighter ones are of the order of at least 100 

 parsecs. They must therefore be exceedingly 

 bright objects at maximum; but how bright 

 we do not know. 



These objects bear very remarkable rela- 

 tions to nebula;. They appear to be related 

 spectroscopically to the gaseous nebulae. The 

 unique moving nebula near Nova Persei was 

 admirably explained by Kapteyn as due to the 

 illumination of a sheet of diffuse ' matter, 

 nearly at rest, by the outgoing light of the 

 great outburst — a hypothesis confirmed by 

 Slipher's recent discovei-y that two variable 

 nebulae appear to shine by reflected light from 

 their nuclei, which show spectra very similar 

 to novse. 



Most remarkable of all is the recent dis- 

 covery that novaB appear in the spiral nebulae 

 so fast that it would take intensive observa- 

 tions to catch them all. 



It is obvious that in these temporary stars 

 we are in the presence of catastrophes, which 

 in magnitude utterly transcend all other known 

 physical phenomena. And these catastrophes 

 are not of rare occurrence, but happen every 

 few years, or oftener, in the galaxy, and ap- 

 parently every few weeks in the Andromeda 

 nebula. Two possibilities suggest themselves 

 at once — a collision or an internal explosion. 

 Collisions between two stars are quite out of 

 the question — owing to the frequency and 



wr.ir*-' 



