454 EEPOET— 1880. 



stroDp; lines. In the sun, tlie absorption bands H and K are verj^ broad, but the 

 band 4226 is weak. Prof. Young, in liis observation of the lines of the chromo- 

 sphere, finds that H and K are strongly reversed in every important spot and in 

 solar storms ; but the line 4226, so prominent in the arc, was only observed three 

 times in tlie chromosphere. 



One of the most interesting features among the most recent researches in 

 Spectrum Analysis is the existence of rhythm in the spectra of bodies, as has been 

 shown by'M- Mascart, Cornu, and others, such as the occurrence and repetition of sets 

 of lines, doublets, and triplets in the spectra of different substances and in different 

 "parts of the spectrum of the same body. Professors Liveing and Dewar, using the 

 reversed lines in some cases for the more accurate determination of wave-lengths, 

 have traced out the rhythmical character in the spectra of sodium, potassium, and 

 lithium. They show that the lines of sodium and potassium form groups of four 

 lines each, which recur in a regular sequence, while lithium gives single lines, which, 

 including the green line, which they show really to belong to lithium, though it 

 was ascribed to caesium by Thalen, also recur in a similar way. In tliese three 

 metals the law of recurrence seems to be the same, but the wave-lengths show that 

 the whole series are not simple harmonics of one fundamental, although between 

 some of the terms very simple harmonic relations can be found. Between the lines 

 G and H are two triplets of iron lines, which, according to Mr. Loclryer, do not belong 

 to the same molecular grouping as most of the other lines. In many photographs 

 of the iron spectrum these triplets have appeared almost alone. Also the two 

 triplets are not always in the same relation as to brightness, the more refrangible 

 being barely visible with the spark ; combining this with Young's observations, in 

 which some short weak lines near G appear in the chromosphere 30 times, while 

 one of the lines of the less refrangible triplet only appears once, and with the fact 

 that in the solar spectrum the more refrangible triplet is much the more prominent 

 of the two, Mr. Lockyer is led to the conclusion that these two triplets are again 

 due to two distinct molecular groupings. 



There is one difficulty which must be taken account of in connection with 

 Mr. Ijockyer's theory with regard to the production of successive stages of disso- 

 ciation by means at our command. 



At each stage of the process there must be a considerable absorption of heat to 

 produce the change of state, and our supply of heat is limited in the electric arc 

 because of the dissociation of the conductors, and more limited still in quantity in 

 the electric spark or in the discharge through a vacuum tube, also we should 

 expect a recombination of the dissociated substances immediately after they have 

 been fust dissociated. Hence it seems easier to suppose that at temperatures whicli 

 we can command on the earth, the dissociation of molecules by the arc or the spark 

 is accompanied by the formation of new compounds, in the formation of which heat 

 and light, and especially chemical vibrations, would be again gi^en out, gi'\ing rise 

 to new spectra, rather than to suppose that we can reach the temperatures neces- 

 sary for successive stages of dissociation. 



To the lines C, F, the line near G, and h belonging to hydrogen, which have 

 a certain rhythmical character, Mr. Lockyer adds Dg and KirchofFs line ' 1474,' 

 regarding '1474 ' (wave-length 5315-0) as belonging to the coolest or most complex 

 form, rising to F at a higher temperature, which is again subdivided into C and G, 

 using the spark without a condenser, which again gives li with the spark and con- 

 denser, which is again split up and gives Dj, a more simple line than h, in the Chro- 

 mosphere. Professors Livemg and Dewar, on the other hand, trace a rhythmical 

 character or ratio between three of the brightest lines of the chromosphere, two 

 of which are lines ' 1474' and 'i' of Lorenzoni, similar to the character of C, F, 

 and h of hydrogen, and also trace a similar relation between the chromospheric line 

 Dj and ' 1474 ' to the ratio of the wave-lengths of F and the line near G. They 

 infer the probability that these four lines are due to the same at present unknown 

 substance as had been suggested by Yoxmg with regard to two of them. The 

 harmony of this arrangement is somewhat disturbed by the fact that D3 lies on the 

 wrong side of ' 1474 ' to correspond with the line near G of the hydrogen spectrum. 

 If we inquire what our sun and the stars have to say to tliese changes of 



