[GRAHAM] DAWSONITE 173 



Here again SM is a larger angle than 2V, for which, in this case, 

 the calculated vahie is 76'' 44'. 



The values for the angle 2V calculated in this manner from 

 three independent sets of determinations are well within the limits of 

 experimental error. 



Knowing a, /?, and y, a further independent calculation oi 2V 

 may be made by means of the usual formula, 



tan V =' ^~J~ ^[, which gives ^T = 76° 46'. 



or \ p^ — y~ 



This value is the mean of those calculated by the previous method, 

 and may be taken to be the true internal optic axial angle for dawsonite 

 in yellow light. The fact that the value is the same, as calculated by the 

 two methods, proves that a (determined by total reflection) is concor- 

 dant with fi and y (by minimum deviation). 



The following are the optical constants for yellow light : — • 



a = 1.466, /3 = 1.542, y = 1.596. 



Birefringence, negative, very strong; y — a = 0.130. 



2Va= 76^ 46'; 2E = 146'' 27'. 



The single measurement in red (Li) light gives 2V = 76" 33', 

 showing that there is only a very slight dispersion, in the sense 

 P < V. 



Chemical Composition. 



As explained at the outset, Harrington's analyses did not lead to 

 any simple formula for the mineral, and always showed the presence of 

 lime. Although in his first two analyses, given in columns I and II 

 below, the percentage of the latter seemed to be constant, in others it 

 varied considerably, and was sometimes as high as 16^. By assuming 

 the lime to be present as admixed calcite, the recalculated analyses 

 brought to light the fact that the ratio of the remaining constituents 

 was always constant, and agreed well with the formula, Al^Oj.NojO- 

 2C02.2HjO, which Friedel had calculated from his analyses of the 

 Tuscan dawsonite. 



I thought it advisable to make a new analysis of the Montreal 

 crystals, with a view to determining whether carefully selected material 

 would not be found to be free from lime; or, on the other hand, if this 

 were not the case, to see whether som'e, at least, of the lime might not 

 be isomorphously replacing an equivalent amount of soda — a fairly com- 

 mon replacement among minerals. About a quarter of a gram was 

 carefully selected from several specimens, rejecting all material which 

 was not perfectly transparent. The specific gravity determined with 



