874 



generate platinum-silicides, which afterwards are decomposed by 

 watervapoiir under generation of silico-hydrogencompounds. The 

 platinumcrucibles are strongly attacked by the Li^O^, which forms 

 from the orthosilicate, when melted in the air; the platinum is 

 coloured greenish or yellow, while the mass itself often looks 

 pale-pink. 



The fact of the attacking of the platinum by the lithiumperoxide, 

 made it impossible to obtain reliable measurements with mixtures, 

 richer in Li/) than the orthosilicate, because the thermoelements are 

 spoiled in the same way as the crucibles, and are thus losing their 

 calibration. This fact must be kept in mind also, when meltingpoint 

 determinations of the orthosilicate are made. 



The ortliosilicate, if brought into cold water, is momentaneously 

 hydrolised by it ; the solution shows a strong alkaline reaction. 

 Analogous decomposition tinds place by the carbon dioxide and the 

 humidity of the atmosphere. 



The heating of the orthosilicate in nickel crucibles, as already 

 mentioned, causes an intensive lilac colouring of the mass, probably 

 by diffusion of metallic nickel from the walls into the central part 

 of the substance. Tlie colour is the same, as many nickel compounds 

 with complex ions show; it is not improbable, that the nickel is 

 present in these products in the colloidal form, as observed e.g. by 

 LoRENZ ^), in the electrolysis of some molten salts. 



Microscopical investigation taught us, that the lilac product shows 

 the same polygonal or almost i-ound, very thin scales as the pure 

 orthosilicate ; however, they were now tinged homogeneously violet, 

 and also often locally darker than in other places. These coloured 

 crystals are noticeable dichroUic, as is often observed in crystals, 

 which take some dyes in solid solution. 



The refractive indices of this lilac substance did not differ appre- 

 ciably from those of the pure compound. By means of the immersion- 

 method, we found : n = 1,595 + 0,005 and n, = 1,610 ± 0,005 ; 

 the birefringence is fairly strong, and aliout : 0,015 to 0,020. Often 

 a twinning is present, which reminds of that of albite. The specific 

 gravity was determine dpycnometrically, in toluene, to be : (/40 = 2,392 

 at 25,°1 C. ; the coloured product showed a value slightly higher : 

 f/40 = 2,415 at 25°,1 C. Figure 8 may give a representation of 

 the compound between crossed nicols ; some twins are also shown 

 in it. 



1) Cf. among others Lorenz, Eleklrolyse der geschm. Salze II. S, 40. (1905); 

 Gedenkboek Van Bemmei.en. (1910). p. 395. 



