592 Wisconsin Academy of Sciences^ Arts^ and Letters. 
to the singular behavior of lake Waubesa in this region, but if 
this lake were omitted the others would be at least no higher 
than the lakes rated 6 and 8. We must, therefore, conclude 
that the colors of these waters do not depend in any essential 
way on the selective absorption of the longer waves of light. 
The same relation is found in the lakes of the second type as 
shown on fig. 13. 
It appears, therefore, that all of these lakes possess stains in 
general which raise the coefficient of absorption in the longer 
wave lengths much above that of pure water, but which do not 
essentially modify the selective absorption within the range of 
the spectrum included in the observations. 
Certain exceptions must be made to the above statement. 
Pine lake alone among those observed has a curve below 6005A 
which is practically identical with that of pure water, in spite 
of its color rating of 14. This case strongly confirms the rule 
given above, so far as selective absorption in this region is con¬ 
cerned. Lake Waubesa (see fig.. 11 and table 4) is a very con¬ 
spicuous exception. The coefficient of absorption rises in the 
region from 5390a to 6180a and thence declines rapidly and 
steadily to 6630a where it is but little above that of pure 
water, being .287 while pure water is .245. This condition 
seems to show that Waubesa lacks in whole or in great part a 
stain which is present in the other lakes whose absorption at 
the other end of the spectrum is quite similar to that of Wau¬ 
besa. Minocqua lake also shows a decline of absorption from 
6005a on, but it is always much greater than that of pure 
water. 
While these two cases stand alone in the series of lakes that 
have been examined, there are hints of something of the same 
kind elsewhere. Oeonomowoc lake shows a decline between 
6480a and 6630a in the observation of 1914, and perhaps the 
corresponding observation on Okauehee lake may bear a similar 
interpretation. Round lake in both years shows an absorp¬ 
tion coefficient nearly stationary between 6480a and 663 OA 5 
though a very high one. 
When we pass to consider lakes of high color, 47 and above, 
on the platinum-cobalt scale,' the absorption curves below 5750 a 
are quite different from those discussed above (see fig. 7). 
The waters concerned are partly swamp and bog waters from 
the northern part of the states viz., Bog, Swamp, and Pond 
