66 T. S. Holland — Constitution of Laterite. 



competition, and the loosely held second and third molecules of 

 water are ejected to permit the crystallization of gothite, Feg O3 . H3 0, 

 or possibly even the whole of the water to form hematite, Feg O3. 

 It must be this action which develops the crust of turgite on nodular 

 masses of limonite ; for turgite itself has no individuality and is 

 probably only an intimate mixture of anhydrous and hydrous oxides. 

 In one sense the change from gibbsite to diaspore is analogous to 

 the dehydration of ferric hydrate. A comparison of the molecular 

 volumes of corundum, AI2 O3 ; diaspore, Alg O3 . Hj ; and gibbsite, 

 AI2O3 . 3 HjO, indicates a lower molecular volume for one molecule 

 of water in diaspore than for each of the three molecules in gibbsite. 

 The figures are as follows : — 



(u) AI2O3. H2 0-(«) Al2O3 = 350-25-5 = 9-5 . ... (a) 

 («) K\0^ . ?, R^O - {v) AUO3.H2 = 66-5 -35-0 = 31-5 . (6) 

 («) AlaOg. 3 H2 0-(«;) AI2 03= 66 -5 -25 -5 = 41-0 ... (c) 



Of these three equations, (a) shows the volume of one molecule of 

 water in diaspore to be 9*5, which is identical with the result 

 obtained by deducting (6) from (c), whilst (h) gives the volume of 

 the second and third molecules of water in gibbsite to be together 

 31*5, or an average of 15-75.^ Thus, the constitution of gibbsite is 

 more correctly expressed by the formula Alj O3 . H2O + 2 Hj O. 

 Now, the relation between the first molecule and the other two is 

 precisely analogous to the relation between Graham's ' water of 

 halhydration ' and the ' water of crystallization ' in soluble salts. 

 Thorpe and Watts have shown — with sulphates as examples — that 

 the first molecule has a volume near 10, whilst the water of 

 crystallization may be as much as 16. They have also confirmed 

 the conclusions of Grraham, and extended by Favre and Valson, to 

 the eifect that more heat is evolved in the act of combination of the 

 first molecule — that in which the greatest contraction occurs — than 

 is evolved in the combination of the remaining molecules. Similarly, 

 the water of crystallization is driven off more easily than the water 

 of constitution or halhydration.^ 



The evidence of the molecular volumes thus indicates a willingness 

 on the part of gibbsite to rid itself, on small provocation, of two 

 molecules of water. This willingness is shown by its ready trans- 

 formation into diaspore. ■ In the case of limonite the tendency to 

 dehydration is still more pronounced ; for it seems from the evidence 

 of laterite, that even a completely anhydrous condition can be 

 attained in a tropical climate which is at the same time humid for 

 much of the year. Whether limonite represents a chemical com- 

 pound or not it is impossible to say, but gibbsite, which forms 

 distinct crystals, is presumably a definite compound, and its loss 



1 The actual identity of results by these two equations is, of course, accidental ; 

 for it is not known that the molecular volume of AI2 O3 remains constant in all its 

 combinations. But, for reasons that need not be stated now, it is certain that its 

 variations in crystallized compounds is very much less than the difference between the 

 molecules of water. 



2 Thorpe & "Watts, "On the Specific Volume of Water of Crystallization": Journ. 

 Chem. Soc, 1880, vol. xxx^ii, p. 102. 



