324
large scale when crops are planted (e.g., experiments with nitrogen-fixing bacteria) or when areas
are reforested. Transplants of this type occur in
many managed ecosystems, but unfortunately these
are only rarely viewed as experimental opportunities. Recent attempts to restore degraded habitats
may provide special opportunities to determine the
effect of transplants on systems, especially if analyses are conducted prior to as well as after transplants take place. Smaller-scale, highly controlled
transplants can be especially useful for identifying
specific changes in ecosystem processes. These
types of experiments can be used to determine the
effect of plant species composition and diversity on
belowground ecosystem processes (e.g., Monz et
al. 1994). In some manipulation experiments, cylinders are driven into the soil, the cylinders containing the soil removed, and used in growth chamber experiments. Monz et al. (1994) removed more
than 50 cylinders of soil and plants from the shortgrass steppe in Colorado to the Duke University,
North Carolina, phytotron to test the impact of
global change (elevated CO 2 , moisture, and temperature) on soil mycorrhizae under controlled
conditions.
A variation of transplanting plants is the reciprocal movement of organic material such as litter
or thatch (Knapp and Seastedt 1986). This type of
experiment increases the quantity of litter that the
soil food web processes. In these studies, litter is
raked from one location, added to another site, and
then the decomposition and soil biotic results are
compared with unmanipulated controls in which
the litter is present. The shading factor of litter manipulation can be accomplished by using screens of
different mesh size to filter light.
Among the most difficult types of ecosystem
studies is evaluating the impact of aboveground
animals (soil compaction, grazing, nutrient input)
on soil, the soil biota, and on processes. Predator
or prey densities can be "transplanted" or altered,
by adding or subtracting either, to determine the
effect on belowground processes. Vertebrates are
often fenced or confined in some manner to facilitate the manipulation. For example, bison have
been reintroduced into some prairie settings, and
their effect on certain belowground ecosystem processes has been analyzed (James 1992; Seastedt
1995).
Diana H. Wall and O. James Reichman
As appealing and productive as these approaches
can be, they also carry significant liabilities. There
is very little control over the independent variables,
which also limits the opportunity for adequate replication. In addition, if comparisons are made over
relatively large scales, such as between regions or
continents, or along latitudinal or longitudinal gradients, the covariation of many variables can confound the results.
Resource Manipulations
Manipulation of water, sun, and nutrients or other
chemicals can have a direct effect on ecosystems,
but also indirectly influence processes by stimulating or inhibiting biotic variables in the system. For
example, fertilizing or irrigating plants with overhead sprinklers to supplement rainfall may increase
plant density and root biomass, and change root
architecture and the resource quality of the plant.
This can indirectly increase soil biotic competition
for other resources, or the manipulation may provide additional food for consumers, which in their
own right will impinge on ecosystem processes
(Verhoef and Brussaard 1990; Swift and Anderson
1994).
Laboratory Studies
The ultimate biotic manipulation involves bringing
the soil of the studied ecosystem into the laboratory
or Ecotron (Naeem et al. 1994), isolating or manipulating the soil biotic organisms, and more effectively controlling climatic variables. There are obvious problems with such approaches, such as the
inherent artificiality, and large systems are particularly difficult to simulate accurately. Yet, investigators familiar with the complexity of the soil system and its inhabitants frequently test hypotheses
using a combination of experimental approaches:
microcosm, mesocosm, greenhouse, phytotron, and
field manipUlations (see Moore et al. 1996; Naeem
et al. 1994). However, although seemingly simpler
and with fewer variables, laboratory microcosms
and growth chambers are not without problems.
Microscopic soil invertebrates and microorganisms
can be introduced through the water source,
through soil contamination (sweeping or splashing
large scale when crops are planted (e.g., experiments with nitrogen-fixing bacteria) or when areas
are reforested. Transplants of this type occur in
many managed ecosystems, but unfortunately these
are only rarely viewed as experimental opportunities. Recent attempts to restore degraded habitats
may provide special opportunities to determine the
effect of transplants on systems, especially if analyses are conducted prior to as well as after transplants take place. Smaller-scale, highly controlled
transplants can be especially useful for identifying
specific changes in ecosystem processes. These
types of experiments can be used to determine the
effect of plant species composition and diversity on
belowground ecosystem processes (e.g., Monz et
al. 1994). In some manipulation experiments, cylinders are driven into the soil, the cylinders containing the soil removed, and used in growth chamber experiments. Monz et al. (1994) removed more
than 50 cylinders of soil and plants from the shortgrass steppe in Colorado to the Duke University,
North Carolina, phytotron to test the impact of
global change (elevated CO 2 , moisture, and temperature) on soil mycorrhizae under controlled
conditions.
A variation of transplanting plants is the reciprocal movement of organic material such as litter
or thatch (Knapp and Seastedt 1986). This type of
experiment increases the quantity of litter that the
soil food web processes. In these studies, litter is
raked from one location, added to another site, and
then the decomposition and soil biotic results are
compared with unmanipulated controls in which
the litter is present. The shading factor of litter manipulation can be accomplished by using screens of
different mesh size to filter light.
Among the most difficult types of ecosystem
studies is evaluating the impact of aboveground
animals (soil compaction, grazing, nutrient input)
on soil, the soil biota, and on processes. Predator
or prey densities can be "transplanted" or altered,
by adding or subtracting either, to determine the
effect on belowground processes. Vertebrates are
often fenced or confined in some manner to facilitate the manipulation. For example, bison have
been reintroduced into some prairie settings, and
their effect on certain belowground ecosystem processes has been analyzed (James 1992; Seastedt
1995).
Diana H. Wall and O. James Reichman
As appealing and productive as these approaches
can be, they also carry significant liabilities. There
is very little control over the independent variables,
which also limits the opportunity for adequate replication. In addition, if comparisons are made over
relatively large scales, such as between regions or
continents, or along latitudinal or longitudinal gradients, the covariation of many variables can confound the results.
Resource Manipulations
Manipulation of water, sun, and nutrients or other
chemicals can have a direct effect on ecosystems,
but also indirectly influence processes by stimulating or inhibiting biotic variables in the system. For
example, fertilizing or irrigating plants with overhead sprinklers to supplement rainfall may increase
plant density and root biomass, and change root
architecture and the resource quality of the plant.
This can indirectly increase soil biotic competition
for other resources, or the manipulation may provide additional food for consumers, which in their
own right will impinge on ecosystem processes
(Verhoef and Brussaard 1990; Swift and Anderson
1994).
Laboratory Studies
The ultimate biotic manipulation involves bringing
the soil of the studied ecosystem into the laboratory
or Ecotron (Naeem et al. 1994), isolating or manipulating the soil biotic organisms, and more effectively controlling climatic variables. There are obvious problems with such approaches, such as the
inherent artificiality, and large systems are particularly difficult to simulate accurately. Yet, investigators familiar with the complexity of the soil system and its inhabitants frequently test hypotheses
using a combination of experimental approaches:
microcosm, mesocosm, greenhouse, phytotron, and
field manipUlations (see Moore et al. 1996; Naeem
et al. 1994). However, although seemingly simpler
and with fewer variables, laboratory microcosms
and growth chambers are not without problems.
Microscopic soil invertebrates and microorganisms
can be introduced through the water source,
through soil contamination (sweeping or splashing