268 /. C. BRANNER 



county deposits, but they were collected by prospectors for iron 

 ore, and their exact locations are not known. 



In other cases the bauxite contains so much silica and so 

 little water that it is not to be distinguished by analysis from 

 ordinary kaolm. And these varieties grade so insensibly into 

 each other that no line of demarkation can be drawn between 

 them. The following analyses show the similarity in composi- 

 tion between a kaolin and a variety of bauxite, which, properly 

 speaking, is simply a pisolitic kaolin. 



ANALYSES OF PISOLITIC AND ORDINARY KAOLINS. 



Silica, SiOg 



Alumina, AlgOg . . 

 Ferric oxide, Fe^O 



Lime, CaO 



Magnesia, MgO . . . 



Potash, KgO 



Soda, NasO 



Water 



Total 



Pisolitic kaolin 



I S. 13 W., 



sec. 10 N. E. 



of N. E. 



48.05 



38.92 



1. 19 



0.58 



0.45 



0.18 



0.28 



10.86 



100.51 



Pisolitic kaolin 



Same sec, S. E. of 



N.W. 



13 



100.98 



Washed kaolin from 

 Brandywine Summit, Pa. 



47.24 



37-27 

 1.94 

 0.52 



trace 

 0.22 

 0.13 



13.62 



100.94 



Water at II0°-II5°C 0.46 % 1.4 



In the following table are brought together all the analyses 

 of the Arkansas bauxite, and a few analyses of representative 

 bauxites from other parts of the world. ^ These analyses are of 

 individual examples, however, and must not be' accepted as if 

 made of car-load lots. 



^ In the article published in the A7nerican Geologist for March 1891 I gave an 

 average of fourteen partial analyses of bauxite from France, Austria, and Ireland. I 

 wish here to express my disapproval of such a method of making comparisons. Owing 

 to the extreme variability of bauxites, even from the same 156ds, such an average can 

 scarcely fail to be misleading. If material is wanted with a low percentage of silica 

 the high percentage in some of the samples will so increase the silica in the average 

 that one may be led to infer that none of it is low enough to be available — a conclu- 

 sion entirely unwarranted. 



