4 Wilde, Resolution of Elementary Substances. 



atomic weights of the intermediate members of the same 

 series. The accepted valencies of these members, as I 

 have already stated, are still open to revision, and they 

 will be found ultimately to agree with the values shown in 

 my Tables. The constitution and position of carbon 

 at the head of the series H3n also give support to the 

 tetrahedral concept of van't Hoff* and Le Belf in 

 regard to this element, as the tetrahedron is necessarily 

 correlated with tridimensional space in the stereo-piling of 

 the carbon molecule 4X H3 = C. 



An examination of the atomic weights of the 

 positive and negative series Hn and H2n in Table I. 

 will show that the differences between the theoretic and 

 experimental atomic weights (excluding radium pro- 

 visionally), when distributed among the twenty-four 

 members of these series, amount to less than half of i per 

 cent, of the actual determinations. It is also interesting 

 to note (i)the common numerical difference of 4 between 

 the monatomic halogens and the alkaline metals in 

 homologous positions, and (2) the common difference of 

 8 between the diatomic oxygen series and the alkaline 

 earth metals in similar positions, (3) the natural arrange- 

 ment and extension of the well-known triads in which the 

 sum of the atomic weights of the extremes is equal to 

 double the atomic weights of the means in these series, 

 (4) that the highest places in my complete Table are filled 

 up by well-known elements of the greatest atomic weights. 



From the definite multiple relations which the atomic 

 weights of mercury, lead and thorium have to those of the 

 lower members of their respective series, for the same 

 number of elementary condensations, there is abundant 

 reason for concluding that no elements exist of higher 



* Archives Neerlandaises des Sciences, Tome IX., pp. 445-454; i874- 

 t Biill. de la Soc. Ckiiit., 1874. 



