30 Penjield and Stanley — 



chemical structure. In part it goes with oxygen to form 

 hydroxyl (OH), which radical has the character of an mi- 

 metallic, acid-forming element, as shoAvn by its isomorphism 

 with fluorine ; in part also hydrogen is basic like a metal, H 2 

 being isomorphous with Na 2 , Ca or Mg, or, in accordance with 

 our conventional method of expressing analyses, H 2 is 

 isomorphous with Na 2 0, CaO, MgO, etc. From the results of 

 the determination of water in analyses, there seems at the pres- 

 ent time no way of determining how much hydrogen is to be 

 ascribed to hydroxyl and how much is basic. In both cases 

 the resulting water is driven off only on intense ignition. In 

 the present analyses so much confidence is felt in the purity 

 of the materials analyzed that the question of hydroscopic 

 water or water resulting from alteration does not enter into 

 the calculations, nor is water of crystallization present. 



From the supposed mode of formation of amphibole by pneu- 

 matolitic action, both hydroxyl and fluorine enter into the 

 composition of the mineral and must be reckoned with the 

 same as other constituents : Water as a base, isomorphous 

 with JSTa 2 0, CaO, MgO, etc., must also be taken into considera- 

 tion, and this is such an important matter that it seems best to 

 refer to a simple yet striking example encountered by one of 

 the writers* in the investigation of a very pure variety of 

 anthophyllite. In the analysis water was first estimated as 

 loss at a low red heat and found to be only 0'19 per cent. On 

 completing the analysis the summation was unsatisfactory, as 

 was also the ratio of Si0 2 to the bases, FeO,MgO with traces of 

 MnO and CaO. In endeavoring to find an explanation for the 

 defects of the analysis and its failure to give a good ratio, 

 some of the powdered mineral was heated with a blast-lamp in 

 a closed tube, when at a very high temperature abundant 

 water was given off. This water, amounting to 1*67 per cent, 

 brought the summation of the analysis to 99'99 and the ratio 

 of Si0 2 toFe04-MnO+MgO+CaO + H 2 became 1-00 : 0-997. 

 Therefore in this comparatively simple combination of silica 

 with protoxide bases there can be no question but that H 2 is 

 isomorphous with MgO and FeO. 



In addition to the ratio of silica to the protoxide bases, it 

 will be interesting to note the relative proportion of the several 

 bases present : This is conveniently expressed in per cents and 

 is given with each analysis. 



The analyses are as follows : 



Tremolite. 



Of this material, representing the simplest type of amphi- 

 bole, two analyses have been made. The specimens selected 



* This Journal (8), xl, p. -394, 1890. 



