points of the curve. These two series are separated by two series of electronegative elements : those non-metals on the descending branches are difficultly fusible ; those elements of a non-metallic character on the ascending branches, on the contrary, are easily fusible and volatile.
Besides these very important properties a number of others have found their expression in L. Meyer's classifica- tion. Amongst these are : the co-efficient of expansion by heat, the refractive index, the thermal and electric con- ductivity, &c.
This subject may now be concluded by a consideration of a general nature. In the neighbourhood of the minima of the curve of atomic volumes the contrast in chemical properties of the elements is but little marked, whilst in the neighbourhood of the summits the contrary is the case, and the passage from the metallic to the non-metallic nature takes place in the most abrupt manner. It would therefore appear that the accumulation of a large mass of matter in a small space is not conducive to the development of a well-pronounced chemical character.
��APPENDIX
With regard to the table on page 29 we see that certain of the periods are quite complete, whilst there are some gaps in others. These gaps correspond with elements which are still unknown, and the periodic law permits us to predict not only their existence but also their properties. This was done by Mendel6eff d propos of three elements which he termed eka-boron, eka-aluminium, and eka- silicon, which have since been discovered by Nilson, Lecoq de Boisbaudran, and Winkler, and are now known as scandium, gallium, and germanium.
Such a confirmation certainly tends to show that this periodic classification is an expression of a natural law. It
�� �
