May 29, 1890] 



NATURE 



103 



hydrates '' and of " eutectic compounds " are equally 

 operative in the case of the separation of minerals from 

 a mixture of fused silicates ; and the same idea has been 

 elaborated by Lagorio. Guthrie has further shown that, 

 as water is added to a salt, the fusion-point of the mixture 

 is progressively lowered, and from this fact he concludes 

 that " the phenomenon of fusion is nothing more than an 

 extreme case of liquefaction by solution." 



That silicates, when they are mixed with water, fuse at 

 a lower temperature, was long ago recognized by geo- 

 logists—long, indeed, before any physical explanation had 

 been offered of the fact. Poulett-Scrope, Scheerer, Elie 

 de Beaumont, Daubrde, and many others who might be 1 

 mentioned, have insisted on the important part played I 

 by water in promoting the fusion of lavas and other 

 igneous masses. 



In the case of the volcanic glass known as viarekanite, 

 I have shown that at a comparatively low temperature : 

 the mass will, when heated, swell up and intumesce, the 

 escaping steam causing the molten glass to froth up and 

 assume the character of a true pumice {Geol. Majt^., Dec. 

 3, iii. 243). The brown glass ejected from Krakatab, j 

 during the great eruption of 1883, if heated, increases to 

 many times its original bulk, and passes into a substance 

 which, macroscopically and microscopically, is indistin- 

 guishable from the pumice thrown out in such vast 

 quantities during that great eruption {Geol. Mag., Dec. 

 3, V.6). _ I 



Many volcanic glasses contain an appreciable quantity | 

 of water, amounting in some cases, indeed, to as much as ; 

 10 per cent, of their mass. The glasses which contain 

 water fuse at a lower temperature than those which are ' 

 anhydrous. There is reason to believe that most lavas 

 are not masses in a state of simple fusion, but consist of 

 crystals floating in a mass of mixed silicates and water, 

 the magma being at a temperature above the fusion-point \ 

 of the mixture but below that of the crystals. 



VI. Mechanical stresses, which tend to overcome the at- ! 

 traction between the particles of a solid, promote chemical i 

 action at those parts of its mass luhich are in a condition ! 

 of intense strain. 



That a direct relation exists between mechanical 

 and chemical forces is shown by the fact that capillary 

 action is capable of overcoming weak chemical affinities. 

 Violent mechanical shocks will sometimes completely 

 overmaster chemical affinity, as was shown by Berthelot 

 in the case of acetylene, cyanogen, Sec, and more re- 

 cently by Prof Thorpe in the case of carbon disulphide. 



Carnelley and Schlerchmann endeavoured to show that 

 the solution of a copper wire by acid was promoted when 

 the wire was put into a condition of strain. These ex- 

 periments, it is true, yielded negative results, a circum- 

 stance which is, perhaps, hardly to be wondered at, when 

 we remember how feeble were the mechanical forces 

 employed. 



In the case of the curiously impressed limestone pebbles 

 of the Swiss Nagelflue, however, Sorby has shown that 

 there are grounds for believing that solution is- promoted 

 in masses which are subjected to intense mechanical 

 stresses, and he has confirmed this conclusion by an ' 

 ingenious experiment with rock-salt (Yorksh. Proc. Geol. ' 

 Soc, iv. 458-61). i 



Similarly impressed and faulted pebbles from the Old 

 Red Sandstone of Stonehaven, in Scotland, have afforded \ 

 what I think is indisputable evidence of the action of 

 ■strain in promoting solution. The sand-grains, of which 

 these pebbles are composed, are seen under the micro- 

 scope to be traversed by bands of liquid enclosures that 

 are clearly of secondary origin. Now, these bands of 

 ■enclosures are parallel to the actual faults that have been 

 produced in the pebbles, and the careful study of all the 

 facts renders inevitable the conclusion that when the 

 whole mass, under great statical pressures, was permeated 

 by fluids, solvent action was determined in parts of the 

 NO, 1074, VOL. 42] 



mass subjected to violent strain {Mineralogical Magazine^ 

 vii. 83). 



Similar bands of secondary liquid inclusions, which 

 have clearly been produced in the same way, abound in 

 the crystals of many rock-masses that have been sub- 

 jected to strain and movement, 



\TI, Pressure may supply the conditions required for 

 the renewal of the growth of crystals when their develop- 

 ment has been arrested for an indefinite period, and even 

 after they have suffered mechanical injuries. 



In 1856, Louis Pasteur published the results of his 

 interesting investigations upon the property exhibited by 

 bimalate of ammonia and other salts, the crystals of which 

 are able to repair injuries produced by fracture ; and this 

 experiment has been repeated and confirmed by Scharff 

 and other observers. 



This principle of the growth and repair of injured 

 crystals is one of great importance and wide application 

 in geological investigation. Sorby has shown that rounded 

 and water-worn sand-grains that have originally consti- 

 tuted a portion of granite or other igneous rock may, in 

 the presence of solutions of silica and under pressure, 

 renew their growth, and, in the end, acquire the faces and 

 angles characteristic of quartz-crystals The observations 

 of Becke, R. Irving, Van Hise, Bonney, and other micro- 

 scopists have shown that, not only fragments of quartz, 

 but portions of the crystals of felspar, augite, hornblende, 

 biotite, and other minerals, may undergo enlargement in 

 a similar way. It has further been shown that this re- 

 pairing and growth of crystals is continually taking place 

 in rocks under pressure ; that the composition of the 

 outer parts of a crystal may vary as growth goes on ; and 

 that the action can take place in solid rock-masses (Quart. 

 Journ. Geol. Soc, xlv. 175-86). 



I have found it possible to illustrate experimentally 

 some of the phenomena exhibited by zoned crystals in 

 rocks. An octahedral crvstal of chrome-alum of a dark- 

 purple colour was mutilated by having two opposite solid 

 angles broken off from it and then placed in a solution 

 of common ammonia-alum (see Fig. i). By more rapid 



growth in the injured portions, the crystal tended to 

 repair itself, but the regularity of this process was inter- 

 fered with by subjecting the crystal and solution to a some- 

 what wide range of temperature. As the coefficient of 

 expansion of chrome-alum appears to be different from 

 that of ammonia-alum, the shell of the latter material 



