16 THE NAUTILUS. 
erosion of shells to the presence of CO, in the water. This is 
also confirmed in a way from the interpretation of geologic data, 
which gives evidence of the solvent power of ‘‘ carbonic acid.”’ 
Not only is CO, being continually liberated in nature in other 
ways, but there is hardly any doubt but that the interaction of 
humic acid often present in streams with lime may also produce 
CO,. Thus the observation of Shrubsole, whose shells were 
collected from a drainage containing a diversified material, may 
plausibly fitin here. Of course the fact must never be excluded 
that coarser material carried along by the current also plays a 
part in the erosion of shells, but the consequences of such a 
factor may be intensified by the chemical reactions which already 
may have taken place. Most of the eroded shells I have, ex- 
amined come from streams having an abundance of gravel. 
Again, it is probable that in some cases an abundance of lime 
in a stream may neutralize the humic acid before the latter can 
produce any marked effect. 
Later on, some evidence will be presented in support of 
March’s contention to the effect that high CaCO, content of the 
water somehow inhibits absorption of material, preventing the © 
shell from becoming as thick as it might. This, however, is 
only a phase of the well-established principle that living cells 
are able to control the absorption of substances used in their 
metabolism. Sinceit is admitted that the lime of shells comes 
from the water in which they live, there is reason to think there 
may be some correlation—positive or negative—between the 
amount of lime present and the thickness of the shells. Several 
investigators have indicated their probable attack of this prob- 
lem, but so far there does not seem to be any published results. 
Having already secured data on the thickness of the shell and 
reduced it to a convenient factor, (the thickness just superior 
to the pallial line directly beneath the umbo, divided by the 
height), I found a publication of the U. S. Geological Survey 
(8) which fortunately gave analyses of the water at the same 
or what seem to be reasonably adjacent points to where my 
material had been collected. All the localities concerned—col- 
lecting, and points where analysis of water was taken, are indi- 
cated in the data which appear to correlate for my conclusions 
in the table. 
