June 27, 1895] 



NA TURE 



205 



associated with them, with gases at a very high temperature. 

 Now , in the hypothesis which I have ventured to jjut liefore the 

 world of science, I differ in this particular hoth from Laplace 

 and also from Vogel, who has most induslriotisly attempted to 

 establish a classification of the celestial bodies. I pointed out 

 that in accordance with thermo-dynamical principles, the 

 temperature must increase with condensation, and of course it 

 will depend, therefore, u]>on the condensation of the gas, whether 

 we have to deal with high or low temperatures in the bright-line 

 stars and the nebulx. I wish to take this occasion to state that 

 Prof. Darwin has recently shown, as the result of a most pro- 

 found inquiry, that swarms of meteorites in space will behave 

 exactly like a gas ; therefore, what can be said of the thermo- 

 ilynaniics of a gas may be said also of the tliermo-dynamics of a 

 meteoritic swarm. 



Now we come to a very interesting part of the inquiry, 

 because it lands us among phenomena which so far have been 

 considered to be exceptional. I refer to the phenomena of the 

 so-called variable stars. Vou will see in a moment that if there 

 is any truth in what has been brought before you, the Ijght of stars 

 as they pass from the nebulous to the more luminous stage must 

 change during the progress of that evolution. But remember, 

 that change will not be visible to one generation of men, prob- 

 ably not to a thousand generations of men. It is a change which 

 will require millions, and possibly-billions, of years for its accom- 

 plishment ; and therefore we must not associate the word 

 *' variable" with any change which dei^ends wholly upon the 

 evolution of these various stellar conditions. But in addition to 

 that, we can see almost in hours, certainly in days, frequently in 

 months, .sometimes in years, changes in the light of certain stars ; 

 and it is these short period changes which mark out and define 

 for us the phenomena of variable stars. 



Take a star like the sun. It is pretty obvious to you that any 

 change in the sun, such as we see it now, would require a very 

 ■considerable time for its accomplishment, so as to be obviously 

 vi.sible to us all ; hut if you take two bodies like the sun, you 

 might imagine a condition of things in which one body would 

 come e.xactly in the line between the earth and the other body, 

 and woidd so eclipse the further one. There yon have at once 

 the possibility of an eclipse due to the passage of one body in front 

 of another, and therefore of a variability which depends upon 

 eclipses. So much for two bodies like the sun ; but we know 

 that in various parts of celestial space some of the stars have 

 run through their life of light, and exist as dark bodies. Obviously 

 we should get the same eclijjse phenomena when dealing with 

 <me star like the sun and another dark body, provided always 

 that the dark body came and eclipsed the light one. That is a 

 very well known and accepted cause of variability, and one of 

 the most cjbvious cases of this kind we have in the star Algol. 

 There we have two bodies, a bright and a dark one, and a 

 diagram will give us what is called the light -curve, the curve in- 

 dicating the variability brought out by such a condition as that 

 I refer to. When we come to examine the light-curve of a body 

 like this, we find that the lumino.sity of the star remains constant 

 for some considerable time in relation to the period of variability, 

 and then it suddenly decreases. It almost at once — in an hour or 

 two — goes up again, continues then for another period, and 

 suddenly diminishes again (Fig. 29). 



.Spectroscopically we can inquire into the question as to 

 whether there is or is not any physical change connected with 

 this. Obviously, if it is merely an eclipse, there should be 

 no physical change, and therefore no change in the spectrum. 

 Here, t)y the kindness of I'rof. Pickering, I can show you two 

 photographs of the spectrum of this star, when it is most 

 luminous, .and when it is least luminous, and the spectra of these 

 two conditions are, you see, quite similar. The broad lines are 

 alike : in other dark lines also there is no change. Therefore, 

 spectroscopically, we are justified in saying that 'the theory that 

 variability is caused by eclipses is a perfectly justiliablc one'. 



But supposing we consider no longer two bodies like the sun, 

 or even one sun and another body more condensed and colder 

 than the sun, but two not completely condensed meteoritic 

 swarms ; various probabilities never before considered will lie 

 open to our inquiry. 



We inay take the remarkable case of variability presented to 

 us by one of the brighter stars in the constellation of the Lyre, 

 fl Lyra-. The spectrum i>f that star has been very carefully 

 studietl, and if you will look at the details now on this diagram, 

 you will see a series of the most marvellous spectral changes 

 -showing at once that we are not in the presence of phenomena 



NO. 1339, VOL. 52] 



at all similar to those presented in the last star examined. Fig. 

 30 shows the light curve of & Lyra.-, which when at its lowest 

 brightness is a 4A inagnituile star, and at its greatest brightness is 

 a 3i magnitude star, the changes going through one magnitude. 

 In this scale you see that the changes are run through in a period 

 of thirteen days. From the period of the greatest obscuration of 



light, in nearly three days we get to the highest lumino.sity, then 

 at the sixth day we get to what is called a secondary minimum, 

 i.e. the light has gone down a bit, but not so much as it had done 

 at the beginning of this light cycle : then it goes up again, so 

 that on the tenth day we get a maximum of light such as we 

 had on the third day ; after that it goes down, so that on the 



