124
J.H. Lawton
earlier botanical literature (Hector 1998), although detecting it statistically in
experiments can be tricky (Loreau 1998b).
(c) The positive species interactions hypothesis. Species do not just compete with
each other in communities. It is becoming increasingly clear that some plant
species benefit from the presence of other species in the community. That is,
some species interactions are ++, or +0 in Williamson's (1972) classification,
for a variety of reasons (provision of shelter, other beneficial habitat modification, etc.) (e.g. Bertness and Leonard 1997). Now overyielding in mixed-species assemblages compared with monocultures is almost inevitable, and could
be substantial (Fig. 2b).
The relative importance of models a, band c in explaining empirical results (see
Section 2.2) is currently unresolved, and presents a major challenge for future research.
2.1.3 Consequences
Several points follow from this theoretical framework. First, there is currently some
debate in the literature about the relative importance of plant species richness vs.
plant functional types, or plant species identity, in determining ecosystem processes
(e.g. Grime 1997; Hooper and Vitousek 1997; Tilman et al. 1997b,c; Wardle et al.
1997a,b). A consideration of the underlying theory helps to clarify the debate. Because no two plant species within a functional group (however this is defined) have
identical niches, species richness will influence ecosystem processes, by the arguments laid out above. However, if the niche differences between plant species are
small, then the effects will be small. By these same arguments, deliberately selecting species to be as different as possible, by selecting different functional types is
likely to have an even bigger effect on ecosystem processes. Because functional
types are arbitrary divisions of continuous niche-space, deciding whether species
richness or functional types has a bigger impact on ecosystem processes is to arbitrarily divide a continuum; the bigger the niche-differences between species in the
assemblage, the bigger the effect we expect to observe on ecosystem processes.
Second, if there are no, or only very small differences between species influencing some ecosystem process of interest, then we are more likely than not to see data
consistent with the redundant species hypothesis (Fig. 1a) - that is, statistically,
we may be unable to detect any effect of species richness on the ecosystem process
in question, except perhaps in very depauperate communities. An example might
be leaf-litter decomposition. If litter quality is similar across a set of species, plant
species richness and litter decomposition rates may be unrelated (e.g. Wardle et al.
1997c).
Third, if only a few species (or functional groups) are involved, and/or if there
are major changes in dominants with diversity, the effects of species' identities (the
idiosyncratic hypothesis) will be paramount. These ideas, particularly that species
identities matter, are discussed in Lawton (1994), Naeem et al. (1996), Chapin et
at. (1997), and Hooper and Vitousek (1997, 1998).
J.H. Lawton
earlier botanical literature (Hector 1998), although detecting it statistically in
experiments can be tricky (Loreau 1998b).
(c) The positive species interactions hypothesis. Species do not just compete with
each other in communities. It is becoming increasingly clear that some plant
species benefit from the presence of other species in the community. That is,
some species interactions are ++, or +0 in Williamson's (1972) classification,
for a variety of reasons (provision of shelter, other beneficial habitat modification, etc.) (e.g. Bertness and Leonard 1997). Now overyielding in mixed-species assemblages compared with monocultures is almost inevitable, and could
be substantial (Fig. 2b).
The relative importance of models a, band c in explaining empirical results (see
Section 2.2) is currently unresolved, and presents a major challenge for future research.
2.1.3 Consequences
Several points follow from this theoretical framework. First, there is currently some
debate in the literature about the relative importance of plant species richness vs.
plant functional types, or plant species identity, in determining ecosystem processes
(e.g. Grime 1997; Hooper and Vitousek 1997; Tilman et al. 1997b,c; Wardle et al.
1997a,b). A consideration of the underlying theory helps to clarify the debate. Because no two plant species within a functional group (however this is defined) have
identical niches, species richness will influence ecosystem processes, by the arguments laid out above. However, if the niche differences between plant species are
small, then the effects will be small. By these same arguments, deliberately selecting species to be as different as possible, by selecting different functional types is
likely to have an even bigger effect on ecosystem processes. Because functional
types are arbitrary divisions of continuous niche-space, deciding whether species
richness or functional types has a bigger impact on ecosystem processes is to arbitrarily divide a continuum; the bigger the niche-differences between species in the
assemblage, the bigger the effect we expect to observe on ecosystem processes.
Second, if there are no, or only very small differences between species influencing some ecosystem process of interest, then we are more likely than not to see data
consistent with the redundant species hypothesis (Fig. 1a) - that is, statistically,
we may be unable to detect any effect of species richness on the ecosystem process
in question, except perhaps in very depauperate communities. An example might
be leaf-litter decomposition. If litter quality is similar across a set of species, plant
species richness and litter decomposition rates may be unrelated (e.g. Wardle et al.
1997c).
Third, if only a few species (or functional groups) are involved, and/or if there
are major changes in dominants with diversity, the effects of species' identities (the
idiosyncratic hypothesis) will be paramount. These ideas, particularly that species
identities matter, are discussed in Lawton (1994), Naeem et al. (1996), Chapin et
at. (1997), and Hooper and Vitousek (1997, 1998).