January i i, 1900] 



NA TURE 



263 



There is a line at A. 3905 8 which is associated in the spark 

 spectrum of silicium with the lines in set A, but while these are 

 entirely absent from the arc spectrum of silicium, 3905 "8 is a 

 strong line in the arc spectrum. This line differs from the 

 others, therefore, in not being enhanced in intensity in passing 

 from the conditions of the arc to those of the spark. So far as 

 is known, the lines in sets B and C have not been recorded by 

 any other observers of the silicium spectrum. 



The behaviour of the three sets of lines in terrestrial spectra 

 is shown in the following figure :— 



I 



.acuuin tube (capillary). 2. SiO^ spark with Spottlswoode 



3. Vacuum tube (bulb) SiOo spark with small coil. 



It is found, on investigating the occurrence of these silicium 

 lines in stellar spectra, that the three sets of lines respectively 

 attain a maximum intensity at the three different levels of stellar 

 temperature represented by /3, 7, and {■ Orionis. 



The accompanying figure shows the behaviour of the different 

 sets A, B, and C in the spectral, 7, C Orionis. 



I 



C, Orionis. 



Orioni 



3. Or 



We find thai set A is most prominent in the spectrum of )8 

 Orionis, that set C predominates in the spectrum of 7 Orionis, 

 and that set B is by far the strongest in that of C Orionis. 



That the stars named represent three different grades of tem- 

 perature, i Orionis being the hottest, and /3 Orionis the coolest, 

 has been previously deduced by the discussion of other lines m 

 their spectra. This result was embodied in a paper " On the 

 Order of Appearance of Chemical Substances at different 

 Stellar Temperatures," read to the Society in February, 1899. 

 In that paper o Crucis was given as a typical star represent- 



ing a stage of temperature between those of /5 Orionis antl 

 i Orionis. That star can be very well replaced for the purpose 

 of the present di.scussion by 7 Orionis, the two spectra being 

 nearly identical. 



It was recently shown that silicium made its appearance first 

 at the temperature represented by a Urste Minoris, and 

 strengthened at the higher temperature of aCygni and /3 Orionis, 

 afterwards weakening as we pass through the still higher tem- 

 peratures of C Tauri and 7 Orionis, until at the C Orionis stage it 

 IS bordering on extinction. 



In the same paper the behaviour of a 

 C line at A. 4089-2 was plotted, and at the 



I same time it was quoted as an "unknown" 

 line. 

 It is interesting to note that this line is 



now traced to silicium, and is the strongest 

 line in set B. It is apparently a short-lived 



— line in stellar spectra, as it only occurs be- 

 tween the stages of temperature represented 

 by 7 Orionis and C Orionis, being one of 



— the weakest lines in the spectrum of the 

 former star, and one of the strongest in that 

 of the latter. 



i November 23, 1899. — "Preliminary Table of Wavelengths 

 i of Enhanced Lines." By Sir Norman Lockyer, K.C.B., 

 F.R.S. 



The important part which the enhanced lines of the element 



play in the study of stellar spectra cannot be over estimated, 



but a great advance can only be made in this direction by a 



systematic examination of the spectra of 



^ all the elements. Such an undertakinjj 



as this involves considerable lime and 



labour. The author has been fortunate 



enough to have the use of the large 



142-inch Spottiswoode coil, made by Apps. 

 for a short space of time, and employed 

 , it in this work, for which it is specially 



adapted, as the brilliancy of its spark 

 shortens the time of exposure. 



The elements which have been dealt 



with in this investigation are the follow- 

 ing : — "Aluminium, bismuth, chromium, 

 copper, iron, magnesium, manganese, titanium, and vana- 

 dium." 



For each of these elements the spark and arc spectra were 

 photographed and compared, and the wave-lengths of the en- 

 hanced lines, that is, those lines which are intensified in 

 passing from the temperature of the electric arc to that of 

 the spark, were determined, and tables are given of their wave- 

 lengths. 



An illustration shows the enhanced and arc lines in the 

 spectrum of titanium. 



/> 



0*0 r^ 



orvQO 



OOCD 

 'O'O'O 



^ O) 



NO. 1576, VOL. 61] 



?? 



