of the Elements. 435 



titanic, and stannic compounds. Moreover, by representing 

 boric cbloride as BCl"* instead of by BCP, the formulje and reac- 

 tions of most boric compounds become much simpUfied. Thus, 

 with regard to the five best-defined borates, wc have, — 



H-^O . BO^ Boric acid SH^O .2W0^ 



Mg^O . 2B02 Boracite SMg^O . ^WO^ 



Na'-^0.3B02 Borax ^-e^O .2W0^ 



4(Ca20 . SiO^) 3 (H^O . B02) BotryoUte 4(Ca20 . SiO^) . 3H^0 . 2W0^ 

 4(Ca20.Si02) 2H20.3B02 DatoUte 4.{Q??0 .^iO^) .2{W0 .2W0^) 

 in which the first column of formulae is altogether much simpler 

 than the other. Moreover, Hess's formula for hydroboracite 

 would become 



MgH°o}»BO' ■ «H'0 instead of ^^f^^] 4B' 0» . 6H»0. 



The two following equations also, representing respectively 

 the action of borax on carbonate of soda*, and the formation of 

 hydrofluoboric acid, are more simple with BO"^ than with W 0^: — 



Na'-^O . 3B02 + 2(Na20 . CO^) 

 = 3(Na20.BO^) +2C02 



instead of 



Na^O . 2B2 03 + 2(Na20 . CO^) 

 =3Na20.2B2 03 + 2C02. 



3BF^ + 3H20 = 

 2H2BF6+H20.B02 



instead of 



8BF3-f-6H20 = 

 SH^B^Fs + sH^O.B^O^. 



42 



Adopting for the formulae of the anhydrides BO^ SiO^ TiO^ 

 and SnO^ respectively; and for the chlorides BCl^ SiCl'*, TiCl^ 

 and SnCP respectively, the atomic weights will be as follow : — 



B = 14-5 91-4 _ 



Si = 28-5 -3- =30 46 



Ti = 48-4 



Sum 91-4 Mean diflPerence = 16 ? 



The numerical sequence is not very well marked. Calculating 

 from Berzelius's experiments, the number for silicon would be 

 29'6, which would answer the purpose better than that derived 

 from the results of Pelouze. Moreover, the number for titanium 

 is still open to question. The equivalent of tin, or 11 7' 6, pre- 

 sents a curious relation with that of boron, — rj— = 14-7, or 

 even with that of silicon, — -. — =29'4. 



Boron and silicon both occur as soft brownish powders, stain- 

 ing the fingers. They are both readily combustible ; they both 

 become dense by heat. They may both be obtained crystalline, 



* Arfwedson. Watts's translation of Gmelin's 'Handbook,' vol.ii. p. 87- 



