P a p e r r e a d b e f o r e t h e R o y a l S o c i e t y 1899 M a y 4 ) . In the attempts made to classify the stars by means of their spectra, from Rutherford's time to quite recently, the various criteria selected were necessarily for the most part of unknown origin ; with the exception of hydrogen, calcium, iron, and carbon, in the main chemical origins could not be assigned with certainty to the spectral lines. Hence the various groups defined by the behaviour of unknown lines were

more »

... ed to by numbers, and as the views of those employed in the work of classifying differed widely as to the sequence of the phenomena observed, the numerical sequences vary very considerably so that any co-ordination becomes difficult and confusing. Recent work has thrown such a flood of light on the chemistry of the stars that most definite chemical groupings can now be established, and the object of the present communication is to suggest a general scheme of classification in which they are employed, in relation to the line of cosmical evolution which I have developed in former papers communicated to the Society. The fact that most of the important lines in the photographic region of the stellar spectra have now been traced to their origins renders this step desirable, although many of the chemical elements still remain to be completely investigated from the stellar point of view. The scheme is based upon a minute inquiry into the varying intensities, in the different stars, of the lines and flutings of the undermentioned substances: Certain unknown elements (probably gaseous, unless their lines represent ,principal series<) in the hottest stars, and the new form of hydrogen discovered by Professor Pickering (which I term proto-hydrogen for the sake of clearness)