MICROSCOPIC CHARACTERS OF CORE MATERIAL 435 



In most of the glass there is a moderate amomit of recr3'stallization. 

 The separated crystals are not definitely identifiable, but appear to con- 

 sist largely of. minute needles of sillimanite. Close to the iron pipe there 

 is a fairly conspicuous development of fayalite in small but definitely 

 determinable grains of very high birefringence and minimum refractive 

 index about 1.83. 



Chemical Character oe the Core Material 



The general appearance of the core and the microscopic characters of 

 its material leave little room for question as to the fusion of sediment in 

 place. Any lingering doubt is completely dispelled by a chemical study 

 of the fused and unfused portions. 



SamjDles for analysis were obtained by selecting representative frag- 

 ments of the broken part of the fused plug and by taking a chisel cut 

 about a foot long in the unfused material immediately above the plug. 



The ordinary methods of silicate analysis, as advocated by Hillebrand 

 and Washington, were emplo3'ed, except that in the determination of 

 metallic iron in the fused portions recourse was had to the method given 

 by Treadwell and Hall, which involves treatment of the sample with 

 mercuric chloride solution and subsequent titration of the dissolved iron 

 in the presence of manganous sulphate.^ 



The determination of metallic iron in the presence of iron oxides can 

 not be carried out with great precision by any method adaptable to our 

 problem. The results are to be regarded as close approximations. In 

 the presence of metallic iron the determination of ferrous iron by Pratt's 

 method is unreliable. The core contains only a slight amount of metallic 

 iron, and therefore the oxide determinations, both by Pratt's method and 

 as total iron oxides, were in very close accord. A series of determinations 

 was made and the figures given are the average of these determinations. 

 As the average values for FeO plus metallic iron considered as FeO and 

 total oxides considered as FeO differed only by 0.03 per cent and as the 

 fusion took place in the presence of metallic iron, it must be considered 

 that there is no ferric iron present. Moreover, we have been unable to 

 detect magnetite in the fused product. At the same time it can not be 

 proved by chemical means that ferric iron is entirely absent. 



In the determination of water lost at 110 degrees from the unfused 

 sediment the weighing had to be carried out with despatch. The powder 

 is hygroscopic and regains a large projoortoin of this water on the 

 balance pan. 



Treadwell and Hall: Analjtical Chemistry, vol. ii. ".tli ed.. 1019, p. 611. 



