10 JOURNAL OF THE WASHINGTON ACADEMY OF SCIENCES VOL. 11, NO. 1 



cake as a whole separates from the walls, there is often left a 

 narrow patch or series of patches around the upper edge of the 

 cake, strongly adherent to the crucible wall, though this is of less 

 consequence. I have frequently had the opportunity to observ^e 

 this annoying occurrence, with both platinum and iridium-plati- 

 num crucibles, and have elsewhere called attention to it.^ 



In the course of making many rock analyses it was found that 

 crucibles made of pure platinum (Heraeus) were very soft, and 

 apparently not sufficiently polishable internally, so that a ready 

 loosening of the cake from them was seldom accomplished; 

 although for general ignition purposes they are admirable, 

 because of the negligible loss of weight on ignition. Crucibles 

 made of platinum alloyed with a small amount of iridium (such 

 as were used in my laboratory from 1896 to 191 2), were found 

 to be much better for the carbonate fusion. They are stiff er 

 than pure platinum, and are therefore less liable to indentation, 

 and are also susceptible of a higher polish. The cake loosened 

 more often and more readily than from pure platinum, but still 

 adhered occasionally. In the spring of 191 8 I began using a 

 palau crucible, chiefly with the object of testing the material in 

 actual rock analysis, because of the constantly augmenting cost 

 of platinum. To my great gratification it was found that the 

 cold sodium carbonate cake always separated easily, quickly, 

 and almost or quite completely from the crucible, far better and 

 more surely than it had done from either platinum or iridium- 

 platinum crucibles of about the same size and shape. With this 

 crucible, which is now reserved for this purpose, I have made 

 many fusions with sodium carbonate, but have not once found 

 the cake to adhere — in each case it has freed itself rapidly and 

 completely, on gentle heating with enough water to cover it. 



This satisfactory behavior is the more noteworthy because 

 this crucible is slightly indented, at about the level of the upper 

 edge of the melt, around the zone of contact with the supporting 

 triangle.^ Apparently the palau softens rather more than plati- 



' H. S. Washington. Manual of the chemical analysis of rocks (New York, 

 1919), pp. 132 and 135. 



* A triangle of fused silica or of pipe stems must be used with palau crucibles. 

 Platinum triangles alloy with palau and ruin both crucible and triangle. 



