454 



NATURE 



[December 23, 1915 



NOTES ON STELLAR CLASSIFICATION. 

 The Temperature Curve. 



IN the first article under this title — Nature 

 (November 12, 1914), and Bulletin i of the 

 Hill Observatory — I referred specially to the tem- 

 perature curve, and pointed 

 out that "to secure simplicity 

 I represented the two arms of 

 equal inclination," and to 

 save space I used a narrow 

 angle between them, I went 

 on to add that by all analogy 

 the descending arm should 

 fall less rapidly than the 

 ascending one rose. The analo- 

 gies to which I referred dealt 

 with the well-known curves 

 representing the various 

 changes in the light of new 

 stars and of such variable stars as Eta Aquilae, 

 to deal with one the light curve of which has been 

 previously given in these notes. 



Sequence or Stellar Temperatures. 



this would be indicated by a considerable re- 

 duction in the numbers. 



My recent publication of a catalogue of the 

 spectra of 354 of the fainter stars, combined with 

 my previous catalogue of 470 of the brighter ones, 

 has enabled me to test this method of inquiry. 



Humter of Stars 



2M 2M 300 3» 540 MO 



Fig. 2.— Temperature Curve based on the number of stars in each spectral group in the catalogue of 

 470 brighter stars. 



The only differences in the two arms I felt 

 myself justified in showing were the nebulous 

 condition in the ascending arm and the con- 

 densed condition in the de- 

 scending one. 



On my hypothesis the life 

 of a star consists first as an 

 increase of temperature due 

 to collisions in a condensing 

 nebula ; then, after the apex of 

 temperature has been reached, 

 the gradual cooling down of a 

 mass of gas and vapour. 



If we can deal with a large 

 number of stars, the true out- 

 line of a generalised tempera- 

 ture curve should be placed 

 before us in considering the 

 numbers of stars of the various 

 groups, because the longer a 

 star remained at about the same temperature the 

 larger would be the number of stars in that group ; 

 and if the rise of temperature were very rapid, 

 NO. 2408, VOL. 96] 



Although the number of stars in each catalogue 

 is comparatively small, it is perhaps large 

 enough to average the results fairly ; and as we 

 . can deal with two catalogues based on 

 the same scheme of classification under 

 as nearly equal conditions as possible, 

 the comparison of the results obtained 

 in both cases should show us, if they 

 are similar, that the base of inquiry 

 is a firm one ; and if minor departures 

 from similarity arise between the re- 

 sults obtained from the brighter and 

 the fainter stars, we may be enabled 

 to gather some important conclusions 

 from the similarities or dissimilarities 

 of the two curves. The illustrations 

 given show how far this inquiry has 

 already gone. In Fig. 2 the numbers 

 of stars in each group of the 470 

 brighter ones are indicated by the 

 figures shown opposite each group, 

 on a base line the length of which is 

 determined by the numbers of stars. 

 ^ In Fig. 3 the same process is followed in the case 

 of the fainter stars, the length of the base line 

 equally representing the number of stars dealt with. 



Number of Stars 



Fig. 3.— Temperature Curve based on the number of stars in each spectral group from the list in 

 Bulletin III. Numbers in brackets show the number of stars in the separate group. 



It will be seen that there is a very general 

 likeness between the curves of the brighter and 

 fainter stars, and to bring this out more clearly 



