MLL  C M.L C M.L M L Obviously the fractions contain salts which increase in solu. bility as one passes from the

left to right, and with sufficient care and patience this method permits a complete separation . The salts which have been used include the sulphates, nitrates, chromates, formates, oxalates and malonates . R . J . Meyer (Zeit. anorg . Chem., 1904, 41, p . 97) separates the cerium earths by forming the double potassium carbonates, e.g . K2Ce2(CO,)4 . 12H20, which are soluble in potassium carbonate solution, being precipitated in the order lanthanum, praseodymium, cerium and neodymium on diluting the solution; C . A. von Welsbach (Chem . News, 1907, 95, p . 196; 1908, 98, pp .

223, 297) separates the metals of the

ytterbium group by converting the basic nitrates into double ammonium oxalates and fractionating; C . James (ibid., 1907, 95, p . 181; 1908, 97, pp . 61, 205) formed the oxalates of the yttrium earths and dissolved them in dilute ammonia saturated with ammonium carbonate; by boiling this solution the earths are precipitated in the order yttrium, holmium and dysprosium, and erbium; he also fractionally crystallized the bromates (see, e.g . Jour . Amer . Chem . Soc., 1910, 32, p . 517, for thulium) . Complex organic reagents are also employed . Neish (Jour . Amer .

Chem . Soc., 1904, 26, p . 78o) used

meta-nitrobenzoic acid ; O . Holmberg separates neodymium, praseodymium and lanthanum (and also thorium) with meta-nitrobenzene sulphonic acid, and has investigated many other organic salts (see Abs . J . C . S., 1907, ii. p . 90), whilst H . Erdmann and F . Wirth (Ann., 1908, 361, p . 18o) employ the 1.8 naphthol sulphonates . In order to determine whether any chosen method for separating these earths is really effective, it is necessary to analyse the fractions .

For this purpose two processes are available . We may convert the

salt into the oxalate from which the oxide is obtained by heating . A weighed quantity of the oxide is now taken and converted into sulphate by evaporating with dilute sulphuric acid . The sulphate is gently dried until the weight is constant, and from this weight the equivalent of the earth can be calculated . When repeated fractionation is attended by no change in the equivalent we may conclude that only one element is present . This process, however, is only rough, for the elements with which we are dealing have very close equivalents . A more exact method employs the Didymia f Praseodidymia Neodidymia spectra—spark, arc, phosphorescence and absorption; the evidence, however, cannot in all cases be accepted as conclusive, but when taken in conjunction with chemical tests it is the most valuable method . Chemical Relations.—The rare earth metals were at first regarded as divalent, but determinations of the specific heats of cerium by Mendeleeff and Hillebrand and of lanthanum and didymium by Hillebrand pointed to their trivalency; and this view now has general acceptance . They are comparatively reactive: they burn in air to form oxides of the type Me203; combine directly with hydrogen at 2000–3000 to form hydrides of the formula MH2 or MH3; nitrides of the formula MN are formed by passing nitrogen over the oxides mixed with magnesium; whilst carbides of the type MVIC2 are obtained in the electrolytic reduction of the oxides with carbon . In addition to the oxides M203, several, e.g. cerium, terbium and neodymium, form oxides of the formula MO2 . The sesquioxides are bases which form salts and increase in basicity in the order Sc, Yb, Tm, Er, Ho, Tb, Gd, Sm, Y, Ce, Nd, Pr, La; the latter hissing with water like quicklime . The placing of these elements in the periodic table has attracted much attention on account of the many difficulties which it presented .

The simplest

plan of regarding them all as trivalent and placing them in the third group is met by the fact that there is not room for them . Another scheme scatters them in the order of their atomic weights in the last four groups of the system, but grave objections have been urged against this plan . A third device places them in one group as a bridge between barium and tantalum . This was suggested by Benedick in 1904 (Zeit. anorg . Chen., 1904, 39, p . 41), and adopted in Werner's table of 1905 (Bey . 38, p . 914), whilst in 1902 Brauner (ibid . 32, p . 18) placed the group as a bridge on a plane perpendicular to the planes containing the other elements, thus expanding the table into a three-dimensional figure . The question has also been considered by Sir William Crookes (Jour . Chem .

Soc., 1888, 53, p . 487; 1889, 55, pp . 257 et seq.), whose inquiries led him to a new conception of the chemical elements .