Pietenpol—Visible Spectrum of Wisconsin Lake Waters 573 
zinc tube. He used double-distilled water and water from 
which the dust particles had been precipitated with zinc 
chloride. Experiments with turbid distilled water and water 
from some Bavarian lakes show that turbidity changes the 
intensity of the color but leaves the absorption bands practi¬ 
cally unchanged. The addition of dissolved lime salts to dis¬ 
tilled water however changes the color into green. He there¬ 
fore believes the color of water to be due to chemical compo¬ 
sition and not to diffraction in the sense of Soret and Rayleigh. 
The observations of the writer upon distilled water were 
made after great care had been taken to obtain water of the 
highest possible purity. After distillation the water was passed 
through a Berkefeld filter, which under right conditions takes out 
all bacteria. Care was taken to use bottles which for a long 
time had contained distilled water only. In every case the 
tube was thoroughly rinsed and the run made immediately 
after filtering. The water was almost optically clear since 
but a small amount of scattering was noticeable even when 
observed at a slight angle with the direction of the oncoming 
ray. 
5. Effect of Turbidity 
For the determination of the absorption by lake water the 
question came up whether the previously described laboratory 
method could be adopted or whether a method of observing 
water in its natural state would be necessary. An accurate 
determination of the selective absorption by lake water in the 
unfiltered state offered almost insurmountable difficulties, due 
to the high turbidity. It was accordingly necessary first 
to determine whether or not the particles in suspension ab¬ 
sorbed selectively. If not, the object sought could be accom¬ 
plished by bringing in water, filtering, and taking observa¬ 
tions in the laboratory according to the method used for dis¬ 
tilled water. Preliminary observations on Mendota lake 
water when different methods of filtering were used, indicated 
that'the absorption by particles was not noticeably selective. 
Results, taken somewhat later, of the absorption coefficients of 
water from lakes Mendota and Waubesa under different con¬ 
ditions are given in Table 2. The filtered water was entirely 
free from particles in suspension, having been forced through 
a Berkefeld filter. The ^‘effluent” was water which had been 
