at the Royal Institution, 1908-1916 743 



from recent researches and the one solid argument in favour of the 

 view that the elements are compounds is still Mendeleeff's great 

 s:eneralization, The Periodic Law, so-called : the only explanation of 

 the inter-relationship which this discloses has always been that some 

 more or less close genetic connexion, such as that known to hold 

 ^amongst carbon compounds, also exists among elements. 



For the present, as chemical methods of inquiry cannot be brought 

 to bear, the subject must remain outside the ken of the chemist : he 

 being unable to assist so long as a change in composition cannot be 

 observed. 



The lecture on April 1, 1887, on "Light as an Analytic Agent," 

 also belongs to this section and may be referred to as a special 

 illustration of the lecturer's genius as an experimenter. The observa- 

 tions described were made to ascertain whether the advance of 

 an explosive wave could be detected by the shifting of lines in the 

 spectrum of the flash corresponding to the displacements observed 

 by astronomers in the case of stars, which have made it possible to 

 -deduce values for the rate of approach or recession of stars. The 

 primary object was not realized. But on using a bent tube for the 

 <ixplosion and reflecting the light from either limb on to the sht of 

 the spectroscope, so that the two images were superposed, it was 

 observed that one of the two images of the lithium line was reversed. 

 Such reversal is a proof that the temperature of the gas in the 

 explosive wave does not reach its maximum immediately but is lower 

 in the front of the wave. 



Observations were described made with various metals and atten- 

 tion was specially drawn to the remarkable fact that, in the case of 

 metals so slightly volatile as iron, nickel and cobalt, many lines are 

 visible in the flash of the explosion, whilst few are seen when more 

 volatile metals are present. No explanation is suggested. May it 

 not be that the fact that these metals, in particular, are capable of 

 forming volatile compounds, such as nickel carbonyl, Ni(CO)^, is the 

 cause of their special behaviour ? 



High Temperature Dissociation Studies. 



The spectroscopic studies already discussed in part belong to this 

 series. Two other Friday evening discourses, however, may be 

 specially referred to in this connexion : that delivered on April 21, 

 1882, on the '* Experimental Eesearches of Henri Ste. Claire Deville," 

 to whom we owe the introduction of the term "dissociation" as 

 connoting cases of chemical change which occur reversibly, to dis- 

 tinguish these from cases of irreversible decomposition ; and that 

 given on June 8, 1883, on "The Electric Arc and Chemical 

 8vnthesis." Of neither is an abstract to be found in the Proceed- 



