2o6 



NATURE 



[June 27, 1895 



thineenth day, or thereabouts, we get to another minimum, and 

 then the cycle begins again. Associated with these changes we 

 have considerable changes in the spectrum. We have U'en 

 fortunate enough to get a spectrum of this mar\cllous star for 

 every day included in this |ieriod of change, although of course 

 the photographs have not been taken in a period of thirteen 

 days or in ten periods of thirteen days ; but by knowing this 

 period, we have been able to place the different pholograjihs 



Kli#, 3i. L.iu.sc uf %ari;ibility in uucondcn^cd Nw;iriu>, 



together so as to see exactly what happens. We get bright lines 

 and dark lines, and bright lines changing their places : but the 

 main jxjint we have lieen able to make out so far, is that we are 

 dealing with two stars very njuch like a number of stars that we 

 see in the constellation of Oriim. In Kig. 31 we have photo- 

 graphs of the s|)ectra of two of the stars in the constellation 

 of Orion, and a.ssociated with them, three photographs of 

 the spectnmi of /3 Lyrx- ; from the change in the position and 

 coincidence of ihcse lines we are able lo make out that the 



minimum the spectnmi of fl Lyr.v (3) becomes more like that of 

 Rigel (4), the differences at these times being mainly in the 

 intensities of the lines. The photograph of the spectrum about 

 the time of second ma.ximum (5) shows that there are two spectra 

 displaced with respect to each other. The spectrum displaced 

 to the less refrangible side is shown to resemble that of Rigel, 

 while that displaced to the more refrangible side closely re- 

 sembles Bellatrix. I do not profess for one moment to imagine 

 that all the conditions of varialnlity in that 

 star have been thoroughly explained, but 

 we know enough to say that it is something 

 quite different from the condition which 

 obtains in such a star as .\lgol. Also, 

 from the fact that we are dealing wi,th stars 

 like those in Orion, we know that we have 

 to do with more or less condensed bodies, 

 bodies not so condensed as the son is, but 

 still condensed enough to be called stars, 

 without fear of making any great mistake. 

 But in this class of condensed bodies 

 we have only really touched one part of 

 the subject, because if that condition holds 

 for bodies which are condensed, it will not 

 have held good for then> and for others 

 when they were less condensed than they 

 are now. How, then, can we explain the 

 variability of uncondensed swarms ? Fig. 

 32 shows this. 



Here we are dealing with two swaniis 

 so sparse that they may Ik- almost con- 

 sidere<l as nebulx ; antl we will suppose 

 that round the denser and larger one a 

 smaller one is moving in the orbit repre- 

 sented on the diagram. You will see that 

 for a considerable part of the orbit the 

 smaller swarm can perlorm its movement 

 along the orbit without any chance of 

 running U]i against any of the constituents 

 of the greater swarm : but when that little 

 swarm has got to go round what is called 

 the periastron, i.e. the region nearest the 

 centre of gravity, which is occupied by the 

 densest portion of the primary swarm, it 

 is impossible that it can get through with- 

 out a considerable number of collisions 

 between its own constituents and the con- 

 stituents of the majority (1 am not talking 

 [xilitics). What will haiipen ? \'ou will 

 get light and heat produced, forming a variable star, which 

 will give the greatest amount of light w hen those two swarms are 

 closest together, and the least amount of light when they are 

 furthest apart. 



You can imagine also, that, instead of dealing with a highly 

 elliptic orbit such as imagined in Kig. 32, we may have one 

 in which the main mass is very much nearer the centre of the 

 orbit of the smallest swarm, that orbit being much more 

 circular than in the former case. There you will get a chance 



Fi(i 33. — Spcclriiin of o Ccli (I'ickcring). 



variability of /9 I.yne is produced by the revolution round each 

 other of two Mars like certain stars in the conslellaliim of 

 Orion, and that |nrt of the light i.s proliably cut off l;y some kind 

 of eclipse ; alv) that a certain amount of light which writes 

 out for us these bright lines is produced at a certain part of the 

 light curve. The photographs show that alMiul the lime of 

 prinrifKil minimum, the dark line spectrum of /3 Lyne (2) is very 

 similar lo that of Kellalrix (I), while about the time of secondary 



NO. 1339, VOL. 52] 



of a greater number of collisions in one part of the orbit than 

 in another; but there will not be anything like so great a 

 difference Ijctween the number of collisions at the two ends of 

 the major axis of the orbit as there would have been in the first 

 ca.se supposed. In that way, therefore, we can explain the 

 variability of these uncondensed swarms, and not only the 

 variability, but a very consiilerabic differciii c in llu' lime of Ihe 

 cycle occupied by the changes and in the inlensily of the greatest 



