I GRAHAM] DAWSONITE 173 



Friedel was the first to suggest a formula for the mineral, and Nos. 

 1 and 2 above are from his original paper; they are both given in many 

 subsequent references to dawsonite, by other authors. They are not iden- 

 tical, however, and the second should apparently be Al2(C03Na)j(OH)4. 

 The other formulae have been adopted to try and account for 

 the chemical behaviour of the mineral, and the only one which calls for 

 comment is the last, which differs from the rest, although in this case 

 the error may be due to a misprint for NaAl(OH)2C03. 



The formula Al2O3.Na2O.2CO2.2H2O accurately represents the 

 chemical composition of dawsonite; a constitutional formula would have 

 to account, among other things, for the fact that, as stated by Friedel, 

 and also Curie and Flamand, the mineral loses neither water nor carbon 

 dioxide at 140 °C, and that long heating at a high temperature is neces- 

 sary in order to bring about the complete expulsion of these constituents. 

 In one experiment, I heated a weighed quantity, placed on a boat in a 

 glass tube, for nearly half an hour over a full bunsen flame, and there 

 Avas a loss of only 50 9r of the total carbon dioxide. 



Several graphic formulas might be constructed which would explain 

 such a behaviour, but the two suggested seem to have many points in 

 their favour: — 



O /ONa O ^ /OH 



^C - O - Al ^C - O-Al 



HO ^OH NaO '^O H 



Either of these would account for the difficulty in expelling the 

 water, since this is shown to be all chemically combined, and not present 

 as water of crystallization. The first, perhaps, expresses the constitution 

 more accurately than the other, since the sodium is here in direct com- 

 bination with aluminium, which would be expected, the end product 

 after ignition being sodium aluminate: — 



I O^ I /ON«. /ONa 



i_^__X_-.Qj- Al — -Al +CO2+H2O 



iH_Oi _ _ ^Ofi") "^O 



If sodium aluminate is left after expelling all the carbon dioxide 

 and water, it might at first sight appear that the residue should be com- 

 pletely soluble in water, whereas it is not. An experiment was made 

 io test this point : a small quantity of the powdered mineral was strongly 

 ignited over the blowpipe for 45 minutes, at the end of which period it 

 appeared to be unaltered, as the mineral is infusible. The residue 

 •was lixiviated with water, and although it did not dissolve, it was yet 



