RESULTS OF ANALYSES. 263 
The increase in the amount of alumina and iron oxide is in harmony 
with the general rule observed in cases of weathering under the normal 
oxidizing conditions. 
In regard to the behavior of titanic oxide in weathering rocks, no very 
specific data are available, but from what is known of titanium minerals 
in general and from the insoluble nature of many artificial compounds 
of the element, a marked resistance to weathering might be expected. 
This idea finds support in the work of Clarke* and of Dunnington.t 
In his table of the percentage composition of the earth’s crust, based 
upon 880 analyses, Clarke gives 0.33 per cent as the estimated amount 
of titanium. ‘This is equivalent to 0.55 per cent of the oxide. Dun- 
nington has determined the titanic oxide in seventy-two soils of varied 
character, representing widely separated regions, and has found in them 
an average of 1.10 per cent of titanic oxide. Whilea comparison of these 
results does not furnish positive proof of the behavior of titanium min- 
erals during weathering, it affords, in view of the character of the work . 
in both cases, a strong indication that weathering produces a concentra- 
tion of titanium. Lyons,} however, finds a large loss of titanium in the 
incipient weathering of some Hawaiian lavas; but as lime remained con- 
stant and magnesia actually increased, it would seem that there must be 
something unusual in the process, perhaps the action of sea water, as sug- 
gested by the author. In other cases, where the weathering is extreme. 
with a large loss of most constituents, he shows that titanium has suffered 
but slight decrease; and this is probably the more normal result. 
In the case under consideration the titanic oxide has evidently been 
one of the most resistant constituents of the rock, its behavior being almost 
identical with that of alumina. Indeed, the ratio of increase for the two 
constituents is so nearly the same that if alumina be assumed as constant 
the titanic oxide has lost less than 0.3 per cent, an amount well within the 
limit of error of analysis. The relative increase of these two constituents 
is slightly greater than that of the iron oxide, and for this reason,together 
with’ their close agreement in ratio, the titanic oxide and alumina are 
assumed to have remained constant and are taken together as the basis 
for comparison of the two analyses. 
Recalculating the analyses to totals of 100 and estimating the percent- 
age loss of each constituent for the whole rock, the figures shown in col- 
umus III, IV, and V are obtained. 
*P. W. Clarke: The relative abundance of the chemical elements. Bull. 78, U.S. G.S., pp. 34-42. 
+F. P. Dunnington : Distribution of titanic oxide upon the surface of the earth. Am. Jour. Sci. 
(3), xli, pp. 491-495. 
{ A. B. Lyons: Chemical composition of Hawaiian soils and the rocks from which they have been 
derived. Am. Jour. Sci. (4), ii, pp. 421-429. 
