322
nematicides have been employed to exclude large
numbers of individuals (Santos and Whitford 1981;
Ingham 1985; Seastedt et al. 1987; Brown and
Gange 1990; Yeates et al. 1991; Beare et al. 1992).
The nontarget effects may have been unknown or
very reduced in those soils when the study was conducted. Additionally, biocides used in these studies
may have been fairly effective across broad taxonomic groups, but would be less effective if a single
species or group were the target. Yet, there is
evidence that even within a broad taxonomic group,
species can differ in their response to the biocide. For example, Freckman et al. (1980) found
species within nematode trophic groups (plant parasites, bacterial feeders, fungal feeders, predators,
omnivores) varied in their response to 1,3dichloropropene, a nematicide that was used primarily to exclude plant parasitic nematodes. Some
species were killed faster than others, irrespective
of trophic group, and some species survived the
chemical. Specific biocides can have broad impacts
across the soil food web. For example, the fungicide Benomyl, which was widely used in ecosystem
experiments, has a detrimental effect on earthworm
populations (Edwards and Thompson 1973; Tomlin
and Gore 1974; Greig-Smith 1992) and some species of protozoa. Insecticides including naphthalene
and the nematicide carbofuran, reduce earthworm
populations (Tomlin and Gore 1974; Ingham 1985;
Colinas et al. 1994; Ingham et al. 1994). The
nematicide carbofuran has also decreased mites,
Rhizobium populations, field crickets, and springtails (see Ingham 1985). Of the biocides presently
used in ecosystem studies for biotic manipulations,
naphthalene "moth balls" (Seastedt and Crossley
1980), an insoluble volatile compound used since
the early 1960s, is considered to have less impact
on nontarget organisms. However, it also has to be
carefully considered. When placed on the soil surface, it has been shown to affect not only moths,
but fungi, protozoa, and earthworms as well (Ingham et al. 1994).
Compounds resulting from degradation of biocides can lead to other spurious results. For example, a compound applied to kill soil herbivores may
break down into nitrogen or phosphorus compounds that cause a "fertilizer effect," or increased
plant growth. Unless an investigator had information on the degradation products, the increased NPP
Diana H. Wall and O. James Reichman
could erroneously be attributed solely to exclusion
of the herbivores.
Chemical manipulations have proven to be questionable for examining the role of soil biota in ecosystem processes for many reasons. We have no
further recommendations and now elaborate on the
important directions for future ecosystem studies of
soil biota.
Natura! Gradients as Treatments
Ecologists often use the range or natural variation
in environmental and biotic conditions as independent variables to determine the effect of key factors
on ecosystem processes. While these are not truly
manipulative, they do have the effect of providing
large-scale differences in the variable of interest.
For example, gradients of temperature or moisture can be employed to understand how these parameters impinge on the entire array of ecosystem
processes. Even climatic variation can be used to
analyze characteristics of ecosystems. Similarly,
soil differences over short distances (Freckman and
Vrrginia 1997) or large differences in soil features
(Schimel et al. 1985; Schlesinger et al. 1990; Bragg
1995) have been used effectively to understand the
relationship between abiotic and biotic processes.
The soil features of interest may include parent material and soil nutrients such as nitrogen or carbon
(see Schlesinger et al. 1990).
Geography can also be used to exclude various
taxa from soil as a factor in soil ecosystem processes. This can occur on continental scales, such
as comparisons between African and North American habitats (Reichman and Jarvis 1989), or for
habitat differences within a region.
Another type of "natural" experiment that is useful for evaluating the effect of a species on an ecosystem process is the invasion of nonindigenous
species. Exotic species, whether plants or animals,
have invaded huge areas of the globe, and thus offer
an unfortunate experiment for the effect of alternate
conditions on an ecosystem. For example, in the
northeastern United States, the spread of the exotic
earthworm species (Amynthas hawayanus and A.
agrestis) has decreased forest floor litter, changed
soil properties and increased erosion (Burtelow et
al. 1998). The introduction into Europe of the New
Zealand flatworm, an obligate predator of lumbri-
nematicides have been employed to exclude large
numbers of individuals (Santos and Whitford 1981;
Ingham 1985; Seastedt et al. 1987; Brown and
Gange 1990; Yeates et al. 1991; Beare et al. 1992).
The nontarget effects may have been unknown or
very reduced in those soils when the study was conducted. Additionally, biocides used in these studies
may have been fairly effective across broad taxonomic groups, but would be less effective if a single
species or group were the target. Yet, there is
evidence that even within a broad taxonomic group,
species can differ in their response to the biocide. For example, Freckman et al. (1980) found
species within nematode trophic groups (plant parasites, bacterial feeders, fungal feeders, predators,
omnivores) varied in their response to 1,3dichloropropene, a nematicide that was used primarily to exclude plant parasitic nematodes. Some
species were killed faster than others, irrespective
of trophic group, and some species survived the
chemical. Specific biocides can have broad impacts
across the soil food web. For example, the fungicide Benomyl, which was widely used in ecosystem
experiments, has a detrimental effect on earthworm
populations (Edwards and Thompson 1973; Tomlin
and Gore 1974; Greig-Smith 1992) and some species of protozoa. Insecticides including naphthalene
and the nematicide carbofuran, reduce earthworm
populations (Tomlin and Gore 1974; Ingham 1985;
Colinas et al. 1994; Ingham et al. 1994). The
nematicide carbofuran has also decreased mites,
Rhizobium populations, field crickets, and springtails (see Ingham 1985). Of the biocides presently
used in ecosystem studies for biotic manipulations,
naphthalene "moth balls" (Seastedt and Crossley
1980), an insoluble volatile compound used since
the early 1960s, is considered to have less impact
on nontarget organisms. However, it also has to be
carefully considered. When placed on the soil surface, it has been shown to affect not only moths,
but fungi, protozoa, and earthworms as well (Ingham et al. 1994).
Compounds resulting from degradation of biocides can lead to other spurious results. For example, a compound applied to kill soil herbivores may
break down into nitrogen or phosphorus compounds that cause a "fertilizer effect," or increased
plant growth. Unless an investigator had information on the degradation products, the increased NPP
Diana H. Wall and O. James Reichman
could erroneously be attributed solely to exclusion
of the herbivores.
Chemical manipulations have proven to be questionable for examining the role of soil biota in ecosystem processes for many reasons. We have no
further recommendations and now elaborate on the
important directions for future ecosystem studies of
soil biota.
Natura! Gradients as Treatments
Ecologists often use the range or natural variation
in environmental and biotic conditions as independent variables to determine the effect of key factors
on ecosystem processes. While these are not truly
manipulative, they do have the effect of providing
large-scale differences in the variable of interest.
For example, gradients of temperature or moisture can be employed to understand how these parameters impinge on the entire array of ecosystem
processes. Even climatic variation can be used to
analyze characteristics of ecosystems. Similarly,
soil differences over short distances (Freckman and
Vrrginia 1997) or large differences in soil features
(Schimel et al. 1985; Schlesinger et al. 1990; Bragg
1995) have been used effectively to understand the
relationship between abiotic and biotic processes.
The soil features of interest may include parent material and soil nutrients such as nitrogen or carbon
(see Schlesinger et al. 1990).
Geography can also be used to exclude various
taxa from soil as a factor in soil ecosystem processes. This can occur on continental scales, such
as comparisons between African and North American habitats (Reichman and Jarvis 1989), or for
habitat differences within a region.
Another type of "natural" experiment that is useful for evaluating the effect of a species on an ecosystem process is the invasion of nonindigenous
species. Exotic species, whether plants or animals,
have invaded huge areas of the globe, and thus offer
an unfortunate experiment for the effect of alternate
conditions on an ecosystem. For example, in the
northeastern United States, the spread of the exotic
earthworm species (Amynthas hawayanus and A.
agrestis) has decreased forest floor litter, changed
soil properties and increased erosion (Burtelow et
al. 1998). The introduction into Europe of the New
Zealand flatworm, an obligate predator of lumbri-