2C>() Dr. J. W. Nicholson on the 



We can then proceed to more elaborate deductions, and 

 arrive at the results of the last table. These results are of 

 course tentative in so far as integers are invoked. For 

 example, in the case of magnesium, we found p = 3*14, the 

 nearest integer being 3, and within the limit of experimental 

 error. We showed that there is nothing, from the optical 

 point of view, against the hypothesis that an atom of mag- 

 nesium has exactly three electrons free. But 3*14 X 7, for 

 example, is 22 almost exactly, and as this value maybe exact, 

 if the experiments are, it is possible that a solid magnesium 

 molecule contains 7 atoms, and has 22 electrons to spare. 

 Such aggregates of atoms are of course to be expected. 

 Sir J. J. Thomson, in his recent paper on rays of positive 

 electricity *, has found some complex aggregates even in 

 gases and vapours, of which we may mention the polymer 

 6 of oxygen, with a positive charge, and also a ray of an 

 electric atomic weight of 800 in mercury vapour, which can 

 be explained as an aggregate of four atoms of mercury with 

 only one negative electron missing. 



Information as to the presence of these aggregates can be 

 obtained by measuring the optical constants of fused metals. 

 In the process of fusion dissociation of these aggregates will 

 usually occur, and a different value of p would be found for 

 the liquid. Experiments on these lines would be very 

 valuable. 



[Note added June 2§th.~\ While this paper is in proof, my 

 attention has been directed to a paper by B. J. Spence, in 

 Avhich Drude's values of n for gold, silver, and platinum are 

 questioned. This paper will be discussed later. 



XXVII. Note on the Optical Properties of Fused J\Jetals. 

 By J. W. Nicholson, M.A., JD.ScA 



IN the preceding paper, concerned with the effective number 

 of free electrons in an atom of a metal, it was suggested 

 that a determination of tbe optical properties of metals in 

 the liquid state would afford valuable evidence of changes 

 in the state of molecular aggregation during the process of 

 fusion. 



The only metal for which such experiments have been 

 made appears to be tin, which was examined by Drude in his 

 paper of 1900 J. 



* Phil. Mag. February 1911. 

 f Communicated by the Author. 

 \ Annalen der Physik, Band xxxix. 



