OEIGIN OF CEYSTALLINE EOCKS. 43 



experiment, mentioned by Danbrée in this connection, is important. By heating in a glass 

 tube with water a refractory clay (probably under similar conditions to the preceding 

 experiments) this became filled with white pearly hexagonal scales, resembling a mica. 

 They were fusible, attacked by hydrochloric acid, and contained both silica and alumina, 

 being seemingly a product of the action of the alkaline silicate from the glass npon 

 the infusible kaolin."" 



Daubrée recalls in this connection the observations of Frémy, who found that colloidal 

 silicates of soda, (water-glass), made at low temperatures, and containing a large excess of 

 silica, give up, when heated, a portion of their silica, which separates in a form having the 

 insolubility of quartz." Daubrée well remarks that we appear to have, in his own experi- 

 ments at high temperatiires with water, a similar breaking-iip of the silicate of soda, 

 which had separated from the glass, into quartz and a more basic silicate. 



§ 89. In connection with this apparent solubility of alumina, under certain conditions, 

 in watery solutions of alkaline silicates, the observations of Ordway are very important. 

 In his extended studies of the alkaline silicates in 1861, he notes that Bolley had shown 

 that magnesia and lime are slightly soluble in solutions of water-glass, and that Kuhlmann 

 had obtained a double silicate of potash and manganese as a violet-colored vitreous mass, 

 giving a brown solution with water, and had also observed a similar combination of 

 cobalt. Ordway found that in the manufacture of water-glass, if care be not taken, a por- 

 tion of iron passes into the compound, Avhich is not separated from the solution by peroxy-' 

 dation, and but imperfectly by sulxihids. The solvent power of the water-glass is dimin- 

 ished by dilution, but the liquid thus rendered turbid, becomes clear again on concentration. 

 He observed that when a few drops of a weak solution of a metallic salt are added to a 

 solution of water-glass, the precipitate at first formed is redissolved by agitation. " A 

 liquid silicate thus takes up no inconsiderable amount of the oxyds of iron, zinc, mangan- 

 ese, tin, antimony, copper and mercury." By agitating a solution of ferrous sulphate 

 with one of water-glass, in a vessel partly filled with air, a liquid is got which, after filtra- 

 tion, has a very deep blue color.'* This solubility of metallic oxyds in aqueous solutions 

 of alkaline silicates will help to a rational explanation of many obscure facts in mineralo- 

 gical chemistry, as, for example, the presence of iron, manganese and copper-oxyds, and 

 of metallic copper, with the zeolites and other minerals secreted from basic rocks. 



§ 90. We may now consider the observations of Daubrée and others on the contempo- 

 raneous formation of crystalline zeolites, and many other mineral species, by the slow action 

 of various thermal waters on the bricks and mortar of ancient Roman masonry in France 

 and Algeria. It was at Plombières, in the Vosges, that his first observations were made. 

 The hot water, here rising from a fissure in a granitic rock, penetrates a layer of gravel, and 

 to protect it from the superficial waters, the Romans had capped the spring with a mass of 

 concrete, resting partly upon the granite and partly upon the gravel. From beneath this 

 concrete, extending over a length of more than a hundred metres, and in parts, three 

 metres in thickness, the waters were led to the surface through vertical channels of cut 

 stone. The water, having at its outlet a temperature of '70'^ C, fills the gravel beneath 

 the roof of concrete, and a portion filters slowly upward through this. The concrete 



™ Daubrée, Géologie Expérimentale, pp. 159-179. 



" Frémy, Comptes RendiLs de l'Académie des Sciences, (1856) xliii, p. 1146. 



'* Ordway, Amer. Journal Science, (1861) xxxii, 338. 



