26. Stoichiometric Analysis of Pelagic Ecosystems: The Biogeochemistry of Planktonic Food Webs
403
(and thus clearer water and a brighter mixed layer)
as well as in reduced P loading. Thus, climatic
warming in this case simultaneously disrupted both
the light and phosphorus economies of the lake, a
perturbation whose ecological consequences might
best be understood via stoichiometric approaches.
Implications
What does the work synthesized in this chapter imply for the ways in which we study pelagic ecosystems? There is a mixture of good and bad news.
Some bad news is that it indicates that some familiar models of nutrient cycling and trophic interactions in pelagic food webs, as well as the intuition
about pelagic ecosystem dynamics we have developed via those models, are of restricted domain,
incomplete, or simply wrong. As noted by Andersen (1997), some of these models take liberties with
the first law of thermodynamics and nearly all completely ignore the effects of food quality on consumer success. Some good news is that incorporating ecological stoichiometry into models of trophic
interactions and nutrient cycling is reasonably
straightforward (though the outcomes are complex!) and the work of Andersen (1997) is a big
step forward. Really good news is the possibility
that many aspects of pelagic ecosystem structure
and function that have previously been puzzling or
unexplainable may ultimately be found to be linked
to stoichiometric influences once these effects are
incorporated into our thinking and study plans. For
example, trophic interaction models that incorporate only the effects of food quantity on herbivore
success invariably predict that alterations in top
trophic levels will alter primary producer abundance via the trophic cascade. However, observations suggest that the strength of trophic cascades
varies considerably among lakes (DeMelo et al.
1992; Brett and Goldman 1996; Elser et al. 1995b).
Ecological stoichiometry can reconcile this mismatch between existing theory and observation, as
the model of Andersen (1997) delineates the stoichiometric conditions for strong, stable trophic cascades. A recent whole-lake food web experiment
supports the existence of stoichiometric constraints
on trophic cascades (Elser et al. 1998). Additional
bad news is that several parameters that are apparently critical to understanding trophic dynamics
and nutrient cycling within the pelagic food web,
such as the elemental composition of suspended
particulate matter and of various important grazers,
are not routinely measured in pelagic ecosystem
studies. Fortunately, these parameters are relatively
easy and only moderately expensive to analyze. Of
greater difficulty and expense is the implication that
pelagic ecosystem scientists might understand their
systems better if they also measuredfiuxes of multiple elements in their study systems. Such measurements are in fact relatively laborious and expensive and achieving a complete picture of the
flows of multiple materials in pelagic food webs
will be difficult. Recent advances in network analysis (e.g., Gaedke and Straile 1994) may facilitate
such studies, as network analysis makes it possible
to obtain a reasonably complete picture of the flows
of multiple materials in a trophic network given a
restricted set of observations of key pools and pathways. Brave ecological theorists might be able to
take the next step by abducting more sophisticated
tools from theoretical chemistry where they are
used to analyze complex reaction networks constrained by chemical reality (Masuda 1990; Clarke
1992).
Despite Reiners' stimulus almost 15 years ago
(Reiners 1986), ecological stoichiometry is only
now becoming a self-identifying field. Many of the
most basic data have yet to be collected and we
have only a sketchy idea of how the stoichiometric
structure of various pelagic ecosystems varies. We
also have little idea whether or not patterns of ecological stoichiometry differ for pelagic versus benthic versus lotic versus terrestrial habitats. Discovering those patterns may an important next step in
producing an integrated and predictive ecosystem
science built on first principles rather than habitatspecific paradigms. To that end, I close this chapter
with a message for the adventurous ecologists from
beyond the pelagia who have braved this chapter to
its very end. The principles of stoichiometric analysis and understanding I have delineated here, such
as the links between elemental composition and
growth in both autotrophs and grazers or the role
of multiple mass balance in consumer production
and nutrient release, are broad and founded in basic
biology and thermodynamics. How do you imagine
these processes to unfold in your grassland, your
stream, your soil, your forest canopy? New methods of thinking accompanied by new methods of
403
(and thus clearer water and a brighter mixed layer)
as well as in reduced P loading. Thus, climatic
warming in this case simultaneously disrupted both
the light and phosphorus economies of the lake, a
perturbation whose ecological consequences might
best be understood via stoichiometric approaches.
Implications
What does the work synthesized in this chapter imply for the ways in which we study pelagic ecosystems? There is a mixture of good and bad news.
Some bad news is that it indicates that some familiar models of nutrient cycling and trophic interactions in pelagic food webs, as well as the intuition
about pelagic ecosystem dynamics we have developed via those models, are of restricted domain,
incomplete, or simply wrong. As noted by Andersen (1997), some of these models take liberties with
the first law of thermodynamics and nearly all completely ignore the effects of food quality on consumer success. Some good news is that incorporating ecological stoichiometry into models of trophic
interactions and nutrient cycling is reasonably
straightforward (though the outcomes are complex!) and the work of Andersen (1997) is a big
step forward. Really good news is the possibility
that many aspects of pelagic ecosystem structure
and function that have previously been puzzling or
unexplainable may ultimately be found to be linked
to stoichiometric influences once these effects are
incorporated into our thinking and study plans. For
example, trophic interaction models that incorporate only the effects of food quantity on herbivore
success invariably predict that alterations in top
trophic levels will alter primary producer abundance via the trophic cascade. However, observations suggest that the strength of trophic cascades
varies considerably among lakes (DeMelo et al.
1992; Brett and Goldman 1996; Elser et al. 1995b).
Ecological stoichiometry can reconcile this mismatch between existing theory and observation, as
the model of Andersen (1997) delineates the stoichiometric conditions for strong, stable trophic cascades. A recent whole-lake food web experiment
supports the existence of stoichiometric constraints
on trophic cascades (Elser et al. 1998). Additional
bad news is that several parameters that are apparently critical to understanding trophic dynamics
and nutrient cycling within the pelagic food web,
such as the elemental composition of suspended
particulate matter and of various important grazers,
are not routinely measured in pelagic ecosystem
studies. Fortunately, these parameters are relatively
easy and only moderately expensive to analyze. Of
greater difficulty and expense is the implication that
pelagic ecosystem scientists might understand their
systems better if they also measuredfiuxes of multiple elements in their study systems. Such measurements are in fact relatively laborious and expensive and achieving a complete picture of the
flows of multiple materials in pelagic food webs
will be difficult. Recent advances in network analysis (e.g., Gaedke and Straile 1994) may facilitate
such studies, as network analysis makes it possible
to obtain a reasonably complete picture of the flows
of multiple materials in a trophic network given a
restricted set of observations of key pools and pathways. Brave ecological theorists might be able to
take the next step by abducting more sophisticated
tools from theoretical chemistry where they are
used to analyze complex reaction networks constrained by chemical reality (Masuda 1990; Clarke
1992).
Despite Reiners' stimulus almost 15 years ago
(Reiners 1986), ecological stoichiometry is only
now becoming a self-identifying field. Many of the
most basic data have yet to be collected and we
have only a sketchy idea of how the stoichiometric
structure of various pelagic ecosystems varies. We
also have little idea whether or not patterns of ecological stoichiometry differ for pelagic versus benthic versus lotic versus terrestrial habitats. Discovering those patterns may an important next step in
producing an integrated and predictive ecosystem
science built on first principles rather than habitatspecific paradigms. To that end, I close this chapter
with a message for the adventurous ecologists from
beyond the pelagia who have braved this chapter to
its very end. The principles of stoichiometric analysis and understanding I have delineated here, such
as the links between elemental composition and
growth in both autotrophs and grazers or the role
of multiple mass balance in consumer production
and nutrient release, are broad and founded in basic
biology and thermodynamics. How do you imagine
these processes to unfold in your grassland, your
stream, your soil, your forest canopy? New methods of thinking accompanied by new methods of