THE NAUTILUS. 0 
From the table the following conclusion may be drawn, quali- 
fied of course by the conditions under which the data is pre- 
sented: 
1. In all or the majority of cases discussed from the Upper Ohio 
Drainage, it appears that the thickness of the shell is positively corre- 
lated with the percentage of lime in the water. 
2. In all the cases of the species from Lake Erie, it appears that 
the thickness of the shell is negatively correlated with the percentage of 
CaCO, in the water. 
Why the shells of Lake Erie do not follow the type of corre- 
lation obtained for those from the Upper Ohio (should this ap- 
pear perfectly substantiated), must be largely speculative at 
present, but the following facts are offered in the light of affect- 
ing the ultimate explanation. Walker, (9), has already indi- 
cated the general differences between the shells of L. Erie and 
their parent forms of the Upper Ohio. L. Erie shells are com- 
paratively little eroded, shorter, greater relative degree of infla- 
tion, and in some species other characteristics indicating a de- 
pauperate type of growth. Certain characteristics of this type 
are so marked that it has been considered justifiable to assign 
certain L. Erie shells the rank of varieties (10). Possibly we 
may recognize the less relative thickness of L. Erie shells as a 
physiological variation keeping touch with the morphological 
ones. Dr. Walker in correspondence suggests that these differ- 
ences as above described may be due to different physical con- 
ditions present in L. Erie such as the freedom from disturbance, 
lower temperature and greater alkalinity of the water. Baker 
recorded Cardiwm thinnest where the water had the greater 
saltiness. Comparative and representative analyses of L. Erie 
and Upper Ohio water show that the former has twice as great 
alkalinity, and in addition to the greater amount of CaCO, as 
already pointed out, a greater proportion of sodium and _ potas- 
sium sulfates, and a large quantity of magnesium carbonate and 
sulfate which are not reported from the Upper Ohio Drainage. 
These latter elements occur in sea water to a higher degree than 
is usually ever reported for fresh water, and their presence may 
account in the light of the observations I have given, for the 
effect brackish water seems to have in malforming and depau- 
