518 On the Spectra of Irresolvable Star-Clusters. [Dec. 13, 



3. " The same bright lines appear to be common to all the nebulae 

 which give a bright-line spectrum. On Mr. Stone's view, differences in 

 the constitution of the enclosing atmospheres of different star-groups 

 would be probable." 



Upon the first point I must remark that, after an examination of the 

 spectra of the nebulae of Orion and rj Argus, I am not of Mr. Huggins's 

 opinion. 



The second point appears to me confirmatory of my views. The 

 irresolvability that I contemplate, as connected with bright-line spectra, 

 does not arise from the sensible vanishing of the angles subtended by the 

 distances between the isolated stellar masses of the cluster. In this case 

 the spectra would be continuous. In my paper in the ' Proceedings ' for 

 April 19, I state this point as follows : — " The linear spectrum can 

 only appear when the resolvability of the cluster is at least injuriously 

 affected by the light of the gaseous envelopes becoming sensibly propor- 

 tional to that from the stellar masses, and in the majority of such cases 

 it would only be in the light from the irresolvable portions of the 

 cluster that bright lines could be seen in the spectrum." By irresolva- 

 bility I here mean something independent of the power of any particular 

 telescope used. 



With respect to Mr. Huggins's third objection I must remark that the 

 difficulty is neither greater nor less, whether my views be accepted or 

 rejected. It cannot be more difficult to conceive that the same gas is 

 present and predominant in certain isolated stellar clusters than that it 

 is present and predominant in certain isolated irregular masses. 



With respect to the objection which appears to have been raised by 

 Professor Stokes, "that in a star-cluster in which the stars are surrounded 

 by self-luminous atmospheres, the proportion between the sum total of 

 the light from the stars and the light from the atmospheres will be in- 

 dependent of the distance of the cluster from us." Such, no doubt, 

 would be true in the kind of clusters Professor Stokes has in contempla- 

 tion. I have, rather unfortunately, in my paper of April 19, spoken of 

 a "close stellar- cluster," without, perhaps, explaining with sufficient 

 clearness that the comparative closeness to which I refer is not optical 

 but linear. The clusters I contemplate are only close in the sense that 

 the individual stellar masses must be close enough to destroy the isolation 

 of the vaporous surroundings, and thus give rise to a gaseous envelope 

 continuous over the whole or over portions of the stellar cluster. I am 

 quite aware, theoretically, as well as from experiment, that the spectra 

 of many clusters, however much they may be condensed to a centre, must 

 be continuous and not linear. In fact, according to my views, the 

 question whether a cluster which was irresolvable with our present 

 optical means would give a continuous or linear spectrum, would depend 

 on whether the irresolvability resulted from the vanishing of its linear 

 dimensions, or from the light of the stellar masses in any given direction 



