eLearning

high surface areas of the walls in even large meso-
cosms become sites where more and more nutrients
are sequestered as periphyton (Levine and Schindler
1992). If the enclosures remain in place for long
periods, the differences in nutrient concentrations
and phytoplankton and zooplankton communities
from those of whole lakes can be rather large (Lund
1972;. Smyly 1976; Levine and Schindler 1992).
A Bitter Lesson in Scaling Mesocosms
By the mid-1970s, we had performed a number of
whole-lake experiments with different nitrogen–
phosphorus (N:P) ratios. We were able to deduce
that the division between nitrogen and phosphorus
limitation occurred somewhere between ratios of
added nutrients of 5:1 and 14:1 by weight (11:1 to
31:1 by moles). Whole lakes fertilized at the lower
ratio produced massive blooms of nitrogen-fixing
cyanobacteria, which had the advantage of being
able to fix atmospheric nitrogen (Schindler 1977;
Findlay and Kasian 1987). The cyanobacteria were
seldom rare or absent in natural lakes of the area
(Kling and Holmgren 1972). In lakes fertilized with
higher N:P ratios, the green and chrysophycean
community consisted largely of species commonly
found in unfertilized lakes, but had simply increased
in abundance, with perhaps some change in relative
species composition (Kling and Holmgren 1972).
We believed that effective nutrient management
hinged on pinpointing the precise N:P ratio where
the switch to nitrogen-fixing cyanobacteria occurs,
for these bacteria are often associated with aestheti-
cally displeasing algal scums, taste and odor prob-
lems, and production of toxins (Kotak and others
1996). We decided to narrow down the nutrient
ratio where the switch in species occurred, by using
mesocosms that could be precisely replicated. We
installed several 10-m-diameter mesocosms in Lake
303, which has a large uniform area2min depth
(Brunskill and Schindler 1971) and a uniform bot-
tom composition of soft organic muds, into which
we could seal the bottoms of the mesocosms. We
installed themesocosms and sealed the bottomedge
of the plastic into the sediments by sewing lengths
of steel bars into the hems and pushing them 0.5 m
into the sediments. We then began fertilizing the
tubes with different nutrient ratios, from N:P of 4:1
to 33:1 (moles).
We were surprised to find that we could not
induce nitrogen deficits or blooms of cyanobacteria,
even at N:P ratios of 4:1. Even at the lowest ratio of
nutrient addition, the total N:P ratio remained high,
near the natural values of more than 100:1 found in
natural lakes of the area. The reason, we learned,
was that the lake muds were retaining almost all of
the sedimenting phosphorus in the form of algal
remains, whereas there was considerable nitrogen
return from the sediments. Subsequently, whole-
lake fertilization showed the importance of nitrogen
return from the sediments of such shallow lakes
(Levine and Schindler 1989).
We deduced that in deeper lakes, epilimnions did
not have such relatively large areas of sediment
contact. We tried to correct the problem by install-
ing a second set of mesocosms with sewn-in bot-
toms, so that the mud surface was not in contact
with overlying water. Remarkably, thismodification
did not correct the problem. The plastic bottom
surface was almost as efficient at retaining sedi-
mented phosphorus as surface sediments. After a
delay of a few weeks, it began returning nitrogen to
the water column as the muds had done. Once
again, we failed to produce nitrogen-fixing blooms
at any added nutrient ratio.
We were finally able tomakemesocosms simulate
responses to fertilization of thermally stratified lakes
by moving the mesocosm experiments to nearby
Lake 302, which was deeper. We simply installed
open-bottomed mesocosms long enough to reach
through the thermocline. This design prevented the
return of nutrients from the hypolimnion to the
epilimnion while allowing the downward passage of
nutrients in particulate form, as observed in pelagic
regions of lakes. This enabled us to lower the N:P
ratio enough to induce nitrogen-fixing blooms of
cyanobacteria similar to those observed in lakes
fertilized with low N:P ratios (Levine and Schindler
1992). However, another set of mesocosms in the
same lake, placed in shallow water and bottomed
with sediment, reacted similarly to the mesocosms
in Lake 303. In this case, large mats of filamentous
green algae attached to the plastic walls and littoral
sediments, but no nitrogen fixers were observed in
either attached or planktonic communities (Levine
and Schindler 1992). Over 80% of the algal re-
sponse in these enclosures was by periphyton rather
than phytoplankton (Levine and Schindler forth-
coming). Blumenshine and colleagues (1997) re-
port similar results.
Tritium additions to the same mesocosms re-
vealed still another problem. The tubes exchanged
some water, received nutrients from precipitation,
and, in the pelagic ones, received nutrients from
metalimnetic entrainment.When corrected for these
factors, the actual range of N:P supplied ranged from
8:1 to 50:1, substantially different from the afore-
mentioned values, which considered fertilizer alone
(Levine and Schindler forthcoming). While the
mass balance budgets of whole lakes are corrected