Rickett—The Larger Aquatic Plants of Green Lake, 385 
Experience with Lake Mendota profited to make the collection 
more systematic and more evenly distributed. An effort was made 
to collect from every region at the time of flowering, thus obtaining 
nearly the greatest weight attained by the plants, and making 
quantitative comparisons between different species and different 
localities more valuable. Plants in shallow water flower first; the 
time of flowering varies directly with the depth of the water. The 
early summer was therefore spent in collecting samples from the 
shallow water of the lake; then collections were made in water of 
medium depth; finally the deepest flora was sampled, using the 
diving hood. In the latter case it was not possible in the time re¬ 
maining to visit all the stations; but this does not seriously impair 
the accuracy of the results, for the flora is more uniform in deeper 
water and the plant belt might here be divided into fewer stations. 
One station was therefore taken as representative of a group of 
several adjacent ones. 
Because of the previous experience in this method of collecting 
it was possible, in spite of much stormy weather, to collect a larger 
number of samples than in Mendota—309 as against 221. 
It soon became evident that the character of the flora varied at 
different depths. For convenience in handling the data, the plant 
belt was divided, as in Mendota, into three zones, within each of 
which the flora may be taken as fairly uniform, but between which 
there are great differences. The limits of the zones were the same 
as those used in Mendota, namely: Zone 1, 0-1 meter; Zone 2, 1-3 
meters; Zone 3, 3 meters to the deepest limit of plant growth. 
Samples were brought back to the lakeside laboratory (impro¬ 
vised from a boathouse), and there each was separated into its 
component species, the latter being numbered as sub-samples. 
These were weighed and spread out to dry. When air-dry, the 
smaller ones were dried in an oven at 60 °C. for 48 hours. A few 
trials served to show that loss of water beyond this point was 
negligible. 
From the wet and dry weights of a number of samples was ob¬ 
tained the percentage of moisture of each species. About twelve 
determinations were made for each species and averaged. In 
many cases, differences between the averages of different species' 
were not significant, as shown by their probable errors; such spe¬ 
cies were therefore averaged together and considered as having the 
same percentage of moisture. Between some species, however, 
there were marked differences. The values for all species are shown 
