6^6 



NA TURE 



[October 27, 1904 



the relative intensities of different parts of tfie continuous 

 spectrum. While this was in reality what had already been 

 done in the case of spectra photographed with glass prisms, 

 tlie new apparatus permitted comparisons over a much longer 

 range. In each case the limits of the spectra towards the 

 ultra-violet in the photographs are determined by the 

 intensities of the spectra in that region and the duration 

 of the exposures. 



To eliminate as far as possible the varying effects of 

 atmospheric absorption, to which the ultra-violet rays are 

 specially sensitive, and the errors which might arise from 

 differences in photographic treatment, each selected pair of 

 stars was photographed on the same plate when the stars 

 had appro.ximately the same altitude, and if any change in 

 the atmospheric conditions were suspected the result was 

 discarded. In each case an attempt was made so to expose 

 the photographs that in every pair the intensity of the 

 .spectrum was as nearly as possible the same in both stars 

 in the region between H;8 and H7. This condition was 

 very difficult to fulfil in actual practice, owing to the different 

 magnitudes and declinations of the stars compared, and the 

 consequent need for very careful adjustment of the clock 

 rate. The difficulty was further increased on account of the 

 different actinism of the stars in this part of the spectrum. 

 The work, however, has resulted in a series of comparison 

 photographs from which all variable conditions except the 

 natural variations in radiation have, so far as possible, been 



less e.xtended towards the ultra-violet than that of n Urs;e 

 Majoris, and the maximum intensity is much nearer the 

 red end. 



The general result of this research is thus stated : — 

 " Taking the stars assumed to be hottest in the chemical 

 classification, we find that in all cases the relative length of 

 the spectrum is reduced, and the relative intensity of the 

 red is increased, as a lower temperature is reached. That 

 is to say that where two spectra having their intensities 

 about the region H/3-H7 equal are compared, we find that 

 in the cooler stars, according to the chemical classification, 

 the emissions in the red preponderate, whilst in the hotter 

 star the ultra-violet is more extended and intense." 



In other words, the sequence of the various groups of stars, 

 as determined by this more extended study of the continuous 

 radiations, so far as the investigation has yet been carried, 

 is identical with that previously arrived at from a discussion 

 of the line spectra. It follows, therefore, that the classifi- 

 cation which was based upon the chemical differences in- 

 dicated by the successive appearances of metallic, proto- 

 metallic, and gaseous lines, in all probability reveals also 

 the true temperature sequence of the different varieties of 

 stars. 



As pointed out by Sir Norman Lockyer, this result is at 

 variance with that arrived at by Sir William and Lady 

 Huggins. While also basing their conclusions as to re- 

 lative stellar temperatures on the comparative intensities 



(2) 



: spectra of pairs of 

 rvatory. South Kens: 



Enlarged 3A times from the original negatives taken 

 age 2. (2) T\ Ursa; Majoris, stage 3 ; Capella, stage 2. 



eliminated. Nine pairs of spectra are reproduced in the 

 paper, and detailed descriptions of these, and other photo- 

 graphs not reproduced, are given. A sufficiently clear 

 idea of the results may be gathered from Fig. 8, showing 

 two of the selected pairs of stars. It is necessary to point 

 out that as the plates employed were but little sensitive to 

 the green rays, there is a break in each spectrum from 

 about A 486 to A 550, followed on the less refrangible side 

 by a further portion of the spectrum having its centre about 

 " D. " The numbered " stages " in the description refer to 

 the ten horizons of mean temperature already shown in 

 Fig. 6, stage i corresponding to the fluted spectra of the 

 Antarian and Piscian stars, and stage 10 to the simplified 

 spectra of the 7 Argus type. 



A glance at the photographs will suffice to show that in 

 the case of each pair the star at tlie higher stage of tempera- 

 ture, as previously determined from the investigation of the 

 line spectra, has the greater development of the violet end 

 of the spectrum, and that the difference is more marked 

 the greater the temperature difference. In the first example 

 it will be seen that, while the nia.Kimum intensity of the 

 spectrum of Vega is in the blue, that of Arcturus is obviously 

 much further towards the red end ; the differences at the 

 extremities of the spectrum are also very marked, Vega 

 haviiig the greater extension into the ultra-violet, and a 

 relatively reduced intensity at the red end as compared 

 with Arcttirus. Again, in the second pair, in spite of 

 relative over exposure, the spectrum of Capella is notably 

 NO. 1826, VOL. 70] 



of different parts of the continuous spectra, and recognising 

 that some of the stars must be getting hotter, these observers 

 have concluded that the highest temperature is to be found 

 not in the white stars, but in stars resembling the sun. 

 They write ' : — " If the relative intensity of this part of the 

 spectrum, from about K onwards to about A 3400, be re- 

 garded as an indication of temperature, we should have to 

 consider Procyon as at a hotter stage than Vega, and that 

 the highest stage of temperature is reached in the true solar 

 stage, of which Capella is typical. Then a fall of tempera- 

 ture sets in, as is shown in the advancing enfeebleinent of 

 this part of the spectrum in Arcturus, Betelgeuse, and 

 Aldebaran." Special stress is laid by these observers on 

 " the rather sudden fall of intensity of the continuous 

 spectrum at about the place of the end of the series of dark 

 hydrogen lines " in such stars as Vega (a feature which is 

 well brought out in the photograph of the spectrum of this 

 star reproduced in Fig. 8), but Sir Norman Lockyer states 

 that this in no way affects his results, and promises another 

 paper dealing with this and similar points. The precautions 

 taken by Sir Norman Lockyer to secure equal treatment for 

 the stars compared would seem to give his results greater 

 weight than those of the other observers, whose photographs 

 appear to have been obtained in the course of more or less 

 routine work on the spectra of individual stars. 



It should also be noted that the occurrence of proto- 



metallic lines has not been accepted as evidence of the 



1 " Atlas of Representative Stellar Spectra," p. 85. 



