CHEMICAL PROPERTIES OF CLAY. 63 



and 20.9 per cent, chemically combined water. The mixture 

 referred to above would therefore contain 15.1 1 per cent, of com- 

 bined water and 11.65 P er cent - °f sulphur trioxide (S0 3 ). 



In order to determine the loss that occurred in burning, two 

 of the bricklets which had been previously dried in the air bath 

 were carefully weighed and heated in succession to temperatures 

 of 86o° C. (1580 R), 1000 C. (1832 P.), 1100 C. (2012 

 F.). 1200 ° C. (2192 F.) and 1300 C. (2372 F.), respec- 

 tively. The total loss at each of these temperatures is given 

 below. 



Table showing loss in zveight by burning. 



Loss in weight, per cent. 

 Temperature. Sample No. I. Sample No. 2. 



86o° C. (1580 F.) 11.60% n.50% 



iooo C. (1832 F.) 13.18% 12.59% 



1100 C. (2012 F.) 19-93% 19.58% 



i200 3 C. (2192° F.) 23.15% 23.05% 



1300 C. (2372 F.) 23.21% 23.11% 



These figures are certainly interesting, for at 86o° C. the loss 

 only slightly exceeds the amount of combined water contained 

 in the white clay. At iooo C. the loss does not equal the sum 

 of the water contained in the clay and gypsum. A large loss 

 occurred between 1100 C. and 1200 C, while between the 

 latter temperature and 1300 C. the loss was exceedingly small. 

 Therefore, even at 1300 C, or slightly above the theoretic 

 melting point of cone 8, there is still over 3 per cent, of what 

 would be considered volatile material remaining in the mixture. 

 It is presumed that this represents sulphur trioxide which has 

 not been driven off. Furthermore, the clay behaves differently 

 from what it would if carbonate of lime had been the source of 

 the lime, for at cone 8 it had only begun to soften, while a mixture 

 with the same quantity of CaO derived from lime carbonate, as 

 shown by Madder's experiments, 1 warped at cone 1. 



The range of lime, as determined from a series of clays, is 

 as follows : 2 



1 See Magnesia. 



2 Bull. N. Y. State Museum, No. 35, p. 523. Owing to an error in the 

 analysis of one of the brick clays, the averages in this table have been recalcu- 

 lated. 



