334
responses across a gradient of pH conditions in 10liter containers incubated in situ for 7 days. Species
known to respond to acidification through indirect
pathways did not respond to acidification in these
short-term experiments. In general, short-term experiments isolate the direct effects of experimental
manipulations that happen quickly. They are less
effective, however, in incorporating slower-acting
indirect mechanisms that can operate within a lake.
Thus, we view the combination of experiments conducted at different scales (e.g., beakers to whole
lakes) as an effective approach for examining the
effects of acidification on aquatic ecosystems (Frost
et al. 1988).
Large-Scale Experiments
to Evaluate the Effects
of Acidification
As with other whole-ecosystem experiments (Carpenter et al. 1995), tremendous insights have been
gained into the effects of acidification and into basic ecosystem processes through whole-lake acidification experiments. Such experiments can be designed to ask different basic questions about
acidification. Extrapolation from any single experiment must be based on a fundamental understanding of the way that a manipulated lake interacts
with its surrounding landscape, as well as the ways
that it compares with other lakes. Detailed comparisons of a few whole-lake acidification experiments revealed a fair number of common features
in their results (Schindler et al. 1991).
The most fundamental element of any largescale experiment is the availability of a reference
system or systems against which the conditions in
a treatment system can be compared. Such comparisons allow inferences of manipUlation effects
against the background of natural variability (e.g.,
Schindler et al. 1991; Frost et al. 1988). Treatment
lakes are paired against nearby reference lakes or
are divided into treatment and reference basins using vinyl curtains (available from, for example,
Environetics, Inc., Bridgeview, Ill. and FabriCraft,
Syracuse, NY). Such situations have obvious limitations in terms of their replication, of course, and
require the use of special statistical techniques
Thomas M. Frost and Janet M. Fischer
(e.g., Carpenter et al. 1989; Rasmussen et al. 1993)
that have some important limitations (StewartOaten et al. 1992).
A whole-ecosystem acidification experiment can
be designed to test either how much a lake would
respond to a predetermined amount of acid loading
(Rudd et al. 1990; Schindler 1988) or the effects
of acidification to a particular target level on the
organisms and processes within a lake (Schindler
et al. 1985; Brezonik et al. 1993). The former experimental design revealed the importance of acidneutralizing processes other than those associated
with the immediate chemical response of a lake's
water to the effects of added acid, a phenomenon
that is responsible for the much less extensive occurrence of lake acidification than might have been
anticipated based upon standard limnological measures (Schindler 1986). The latter design can be
used to explore the consequences oflevels of acidification that have occurred, such as might be important in developing management programs for
regions effected by acidification.
Experiments to test the consequences of acidification must consider a wide range of factors
when evaluating the consequences of decreasing
pH levels. The proportional occurrence of nitric
and sulfuric acids used in a manipulation must be
considered because they have fundamentally different effects as they acidify an ecosystem (Rudd
et al. 1988). The full range of biogeochemical responses to changes in pH conditions must also be
factored into experimental designs. Some of the
most dramatic organismal responses to reduced pH
conditions have been associated with changes in
the levels of aluminum (AI) (Cronan and Schofield
1979) and mercury (Hg) (Wiener et al. 1990) that
have been generated by acidification. Increased Al
levels result from changes in its solubility with
acidification, whereas Hg increases have been
driven by microbial processes coupled with biomagnification up a lake's food web. Likewise, the
fundamental cycling of nitrogen was shut down by
acidification to around pH 5.4, causing a progressive accumulation of ammonium (Rudd et al.
1988). The addition of even a single acid can induce multiple stresses and generate a wide range
of changes that propagate from the initial stress
(Frost et al. 1999). These changes not only include
the biogeochemical shifts described above but also
changes in physical lake properties such as light
responses across a gradient of pH conditions in 10liter containers incubated in situ for 7 days. Species
known to respond to acidification through indirect
pathways did not respond to acidification in these
short-term experiments. In general, short-term experiments isolate the direct effects of experimental
manipulations that happen quickly. They are less
effective, however, in incorporating slower-acting
indirect mechanisms that can operate within a lake.
Thus, we view the combination of experiments conducted at different scales (e.g., beakers to whole
lakes) as an effective approach for examining the
effects of acidification on aquatic ecosystems (Frost
et al. 1988).
Large-Scale Experiments
to Evaluate the Effects
of Acidification
As with other whole-ecosystem experiments (Carpenter et al. 1995), tremendous insights have been
gained into the effects of acidification and into basic ecosystem processes through whole-lake acidification experiments. Such experiments can be designed to ask different basic questions about
acidification. Extrapolation from any single experiment must be based on a fundamental understanding of the way that a manipulated lake interacts
with its surrounding landscape, as well as the ways
that it compares with other lakes. Detailed comparisons of a few whole-lake acidification experiments revealed a fair number of common features
in their results (Schindler et al. 1991).
The most fundamental element of any largescale experiment is the availability of a reference
system or systems against which the conditions in
a treatment system can be compared. Such comparisons allow inferences of manipUlation effects
against the background of natural variability (e.g.,
Schindler et al. 1991; Frost et al. 1988). Treatment
lakes are paired against nearby reference lakes or
are divided into treatment and reference basins using vinyl curtains (available from, for example,
Environetics, Inc., Bridgeview, Ill. and FabriCraft,
Syracuse, NY). Such situations have obvious limitations in terms of their replication, of course, and
require the use of special statistical techniques
Thomas M. Frost and Janet M. Fischer
(e.g., Carpenter et al. 1989; Rasmussen et al. 1993)
that have some important limitations (StewartOaten et al. 1992).
A whole-ecosystem acidification experiment can
be designed to test either how much a lake would
respond to a predetermined amount of acid loading
(Rudd et al. 1990; Schindler 1988) or the effects
of acidification to a particular target level on the
organisms and processes within a lake (Schindler
et al. 1985; Brezonik et al. 1993). The former experimental design revealed the importance of acidneutralizing processes other than those associated
with the immediate chemical response of a lake's
water to the effects of added acid, a phenomenon
that is responsible for the much less extensive occurrence of lake acidification than might have been
anticipated based upon standard limnological measures (Schindler 1986). The latter design can be
used to explore the consequences oflevels of acidification that have occurred, such as might be important in developing management programs for
regions effected by acidification.
Experiments to test the consequences of acidification must consider a wide range of factors
when evaluating the consequences of decreasing
pH levels. The proportional occurrence of nitric
and sulfuric acids used in a manipulation must be
considered because they have fundamentally different effects as they acidify an ecosystem (Rudd
et al. 1988). The full range of biogeochemical responses to changes in pH conditions must also be
factored into experimental designs. Some of the
most dramatic organismal responses to reduced pH
conditions have been associated with changes in
the levels of aluminum (AI) (Cronan and Schofield
1979) and mercury (Hg) (Wiener et al. 1990) that
have been generated by acidification. Increased Al
levels result from changes in its solubility with
acidification, whereas Hg increases have been
driven by microbial processes coupled with biomagnification up a lake's food web. Likewise, the
fundamental cycling of nitrogen was shut down by
acidification to around pH 5.4, causing a progressive accumulation of ammonium (Rudd et al.
1988). The addition of even a single acid can induce multiple stresses and generate a wide range
of changes that propagate from the initial stress
(Frost et al. 1999). These changes not only include
the biogeochemical shifts described above but also
changes in physical lake properties such as light