118 Allen and Clement — Role of Water in Tremolite. 



Summary. 



1. A stud} 7 of live different specimens of natural tremolite, 

 two of them of exceptional purity, proves that all contain water 

 ranging from 1'7 to 2*5 per cent. This water is lost gradually 

 with rising temperatures without any loss in homogeneity and 

 with very slight change in the optical properties. The water 

 is therefore not chemically combined, although the mineral in 

 the powdered state is not completely dehydrated under 900°. 

 It is to be regarded as dissolved water, and tremolite as a solid 

 solution. A diopside from a metamorphosed limestone con- 

 tained 1 per cent of water and behaved in practically the same 

 way, though presumably the diopside of eruptive rocks is 

 anhydrous. 



The amphibole kupfferite and a specimen of beryl contained 

 respectively 3*8 per cent and 2 - 5 per cent of water, which they 

 lost very slowly at comparatively high temperatures (100°-800°) 

 and still retained their homogeneity. With them, however, 

 the loss of water appeared to progress so slowly at these tem- 

 peratures that the total water could not be driven off in any 

 reasonable time. The beryl lost at the same rate for a long 

 period, both in dry air and in an atmosphere containing water 

 vapor at the partial pressure of about 23 mm , even though this 

 rate appeared to show that the mineral possessed a vapor 

 pressure of only about 0"5 mm of mercury. The kupfferite 

 showed a similar behavior, but the fact that it suffered a sec- 

 ondary change in composition at the higher temperatures (prob- 

 ably due to the absorption of oxygen) made the experiments on 

 it less satisfactory. 



All these minerals show important points of resemblance 

 with the zeolites, with which they may broadly be classed, but 1 

 in one important particular they differ, — at least, this is true 

 of kupfferite and beryl, — they do not give true equilibrium 

 with water vapor at low pressures, while the zeolites under 

 similar conditions do so (Friedel). Diopside and tremolite 

 seem to give off their water continuously, but not indefinitely, 

 with rising temperatures, though it is quite possible the curves 

 represent cases of " false equilibria." 



2. Recent analyses indicate that all the amphiboles contain 

 water. Actinolite, glaucophane, and pargasite contain 1-3-3 

 per cent, mostly retained above 100°. The hornblendes also 

 contain water, though usually in smaller quantity. These facts, 

 taken in connection with the above work on tremolite and 

 kupfferite, lead to the suspicion that the amphiboles generally 

 contain dissolved water as a characteristic constituent, and are 

 solid solutions. 



The authors wish to express their hearty thanks to Mr. F. E. 

 Wright, to whom they are indebted for all the microscopic 

 data found in this paper. 



Geophysical Laboratory, 



Carnegie Institution of Washington, April 23, 1908. 



