GEOPHYSICAL LABORATORY. 157 



parallel to the Si02-Al203 side of the triangle, a fact that certainly is 

 connected wdth the already familiar fact that Si02 and AI2O3 resemble 

 each other in their chemical relations more than either resembles CaO — 

 another way of sajdng that Si02 and AI2O3 are "acidic" oxides, while 

 CaO is "basic." With this in mind we are not surprised to find that 

 the boundary curves in the system Al203-MgO-CaO trend parallel to 

 the MgO-CaO side; likewise that the boundary curves in the system 

 Si02-MgO-CaO also trend parallel to the MgO-CaO side. In the sys- 

 tem Si02-Al203-MgO, however, a definite trend of boundaries is not 

 so marked ; they tend rather to radiate from the peculiar solid-solution 

 field that occupies the middle of the triangle. 



The systems Si02-Al203-CaO and Al203-MgO-CaO are notably 

 free from the complications due to solid solution. A peculiar situation 

 arises in the system Si02-Al203-MgO from the fact that the only ternary 

 compound is unstable, and is represented only in a solid-solution field 

 whose boundaries do not include the compound. The system Si02- 

 jNIgO-CaO, finally, is considerably complicated by the existence of 

 solid solutions between the various binary and ternary compounds. 



Inversions. 



There is not space here to discuss the various polymorphic forms of 

 the oxides and compounds, beyond remarking that there must be a 

 fundamental connection between the inversions, both of the sluggish 

 and the prompt-reversible types, in the silicates on the one hand and 

 the similar inversions of the two types in silica itself on the other hand. 

 This relation forms a problem in molecular chemistry and physics 

 that remains for the future to solve. 



Temperature Range. 



The temperatures of the liquidus surfaces of these four ternary 

 systems lie between the limits 1165° C. (the eutectic of tridymite, 

 wollastonite, and anorthite) and 2800° C. (the melting-point of mag- 

 nesia). The great bulk of the quadruple and quintuple points, how- 

 ever, lie within the hmits 1300° C. to 1600° C. 



THE OXIDES OF IRON. 



The investigations of the high-temperature relations of the oxides 

 of iron, discussed in a previous Annual Report,^ were interrupted by 

 war conditions, but work on the low-temperatiu^e hydrated oxides, 

 as well as on methods for the determination of ferric and ferrous iron, 

 has been carried forward during the year.^ The previously existing 

 confusion with regard to the mineral hydrated oxides of iron has been 

 satisfactorily resolved by the microscopic and thermal study of a 

 large amount of material, and it has become clear that only one hy- 



1 Year Book 15, 137-139 (1916). 



2 See abstracts (16), (21), (22), (30), and (42), below. 



