Pietenpol—Visible Spectrum of Wisconsin Lake Waters 591 
gard from the other lakes shown in the same diagram and also 
from Waubesa (fig. 11) which more closely agrees with the 
general type in this part of its curve. 
It appears that the color of these lakes is due primarily to a 
stain—or a mixture of stains—which rapidly absorbs the rays 
at the blue end of the spectrum. In the less powerfully colored 
waters—rated from 6 to 35—there seems to be a direct rela¬ 
tion between intensity of color on the platinum-cobalt scale 
and the coefficient of absorption at 4700a and 5000a. Be¬ 
tween the limits of 6 and 35 on the color scale the coefficient 
of absorption at 4700a rises from about .275 to .825, rising in 
almost a straight line. Each degree on the color scale adds 
about .018 to the coefficient of absorption. This relation is 
shown by eleven lakes in six widely separated drainage basins. 
It would be interesting to learn whether a larger series would 
show the same relation. 
If the absorption curves for wave lengths below 5500a are 
studied, they disclose a very different set of facts from those 
shown by the same curves for shorter wave lengths. Consider 
first the lakes with marsh-stained water, as shown in fig. 12. 
The absorption curve for pure water begins to rise rapidly 
from about 5750a and in general rises steadily throughout the 
remainder of the spectrum shown in the diagram. There is a 
halt in the rise at about 6220^6300a. The same feature was 
found by von Aufsess in one of his observations on pure water 
and in one lake. No readings were made at this wave length 
by Pietenpol and his readings at 6180a and 6360a give no in¬ 
dication of such a feature as von Aufsess’ curves show. (’03, 
plate 1). 
In general these absorption curves for lake water show a 
minimum anywhere from 5400A to 5600A and a rather rapid 
and regular rise in absorption as the curve passes to greater 
wave lengths. The curve at wave lengths below 5750a is roughly 
parallel to that of pure water. If a constant sum is added to 
the coefficient of absorption of pure water the result will give 
with fair accuracy the form of this part of the curves. The 
observations of von Aufsess (’03, plate 1) show the same fact. 
Furthermore, there is no such correlation between intensity 
of color and coefficient of absorption as there is at the other end 
of the spectrum. The mean absorption at greater lengths 
than 6100a is lowest in lakes rated at 22. This is due in part 
