eLearning

whole-lake isotope (tritiated water, radon) addi-
tions were being used to trace water movement.
After observing that the surfaces ofmesocosms were
often calm when the lake had waves, we expected
thatmesocosmsmight have weaker physicalmixing
than lakes. Tracers showed that the opposite was
true (Quay 1977). The difference was attributed to
transfer of wave energy in the lake to turbulent
energy in the mesocosm, and to bouncing of the
mesocosm flotation collars on the lake surface. In
addition, thermoclines were inexplicably shifted
slightly downward in mesocosms, possibly as the
result of precipitation or leakage through seams
(Quay 1977). In later experiments, tritiumwas used
as an epilimnion tracer to track water loss from
mesocosms. The water loss was often substantial,
even inmesocosms showing no physical evidence of
collapse or leakage. These studies provided critical
data for correcting nutrient addition rates to meso-
cosms (Levine and Schindler 1992, forthcoming).
Mesocosms generally underestimated air–water gas
exchange rates in lakes, presumably because of
weaker surface wave energy in the enclosures
(Schindler 1988).
Chemical Shortcomings
In the early years of ELA, we often performed
mesocosm experiments as pilot studies for whole-
lake experiments [Schindler and others (1971, 1980)
and other articles in Can. J. Fish. Aquat. Sci. 37(3)].
From such studies, we correctly deduced that addi-
tions of trace elements and base cations were not
necessary to produce algal blooms in even the
ultradilute waters of Precambrian Shield lakes. This
was not obvious from laboratory-scale experiments
or observational studies of the day (Provasoli 1969;
Goldman 1972). However, the short duration (less
than one season) of the mesocosm experiments did
not allow us to gauge properly the extent to which
the lake would be able to compensate for carbon
and nitrogen deficiencies. Development of a full
steady-state nitrogen concentration required 17
years, due to the short period each year that nitro-
gen fixation was able to supplement nitrogen fertil-
ization (Schindler and others 1987; Findlay and
others 1994; Hendzel and others 1994).Without the
latter result, we would have concluded from meso-
cosms that fertilization with nitrogen was required
to keep bloom-forming cyanobacteria fromdominat-
ing the phytoplankton (Schindler 1977).
It is noteworthy that standing crops of phytoplank-
ton in reference mesocosms often declined relative
to whole lakes, probably becausemesocosms are cut
off from catchment sources of nutrients (Schindler
and others 1971; Lund 1972). Also, the relatively