42 
The American Geoloi^ist. juiy, looi. 
erosion has not removed any considerable part of the deposits 
they are generally found near the 900-foot level. 
(5) The prevailing- pisolitic structure of the bauxite. The 
process producing the same structure in calcareous and siliceous 
materials has been observed and described. In such cases, 
where the process has been observed the pisolitic structure re- 
sulted from the deposition of the material in solution or from 
suspension in the form of a fine ])recipitate. \''ery likely, as 
Hayes says, this applies to all material having a pisolitic struc- 
ture. 
(6) The grouping or occurrence of the ore-bodies along 
apparent lines of weakness and about certain centres. Refer- 
ence to the accompanying map makes this point clear. 
( 7 ) Xo eruptive rocks are yet known in the vicinity of 
the deposits, nor do any rocks occur in the region which by 
weathering could yield bauxite as a residual product. 
The theorv which best conforms to these premises is the 
one outlined by Dr. Hayes and briefly stated above, which as 
set forth by its author is based on the three essential features 
required of any satisfactory theory, namely : ( i ) the source 
from which the material was derived; (2) the means by which 
it was transported; and (3) the process of its local accumula- 
tion. Only the leading points under each one can here be 
stated. 
(I) llie Source of the Ahiuiina: The Knox dolomyte, 
with which the deposits are associated, contains, as shown by 
mmierous analyses, a very small percentage of alumina, which, 
bv the process of accumulation required, is thought to be entire- 
ly inadequate as the source of the material. 
Four analyses showed the composition of the Knox dolo- 
mvte to vary within the following limits : 
SiOo 3.75 to 7.252 percent. 
'^'sOs I 1236 to 1.76 " " 
FeaOs < 
CaCOs 34.07 to 53.44 " 
MgCOa 36.32 to 55.736 " 
The following analyses of the Knox dolomyte and its ac- 
' companying residual clay from Morrisville, Calhoun county, 
Alabama, w^ere made by Dr. W. F. Hillebrand* and described 
bv professor L C. Russell. f Thev are inserted here to illus- 
*Bullctin, U. S. Geol. Survey. No. 16S, 1900. pp. 25S and 295. 
■flljid. No. 52, 1889. pp. 24-25. 
