8. Biodiversity and Ecosystem Processes
123
tern processes is pattern (ii) (Fig. 1 b), consistent with the rivet hypothesis. The
fundamental underpinning mechanism is that species differ in their ecologies, that
is, no two species have identical niches. But there are some important, and still
poorly understood, subtleties underlying this simple statement, because there are at
least three ways in which the different ecological requirements of different species
can generate relationships broadly similar to Fig. lb. They are, in order of increasing biological complexity:
(a) The sampling hypothesis. Species differ intrinsically in their potential maximum size, growth rate, and so on. Hence, as Huston (1997) points out, mixtures of many species are more likely to contain high-yielding species than
monocultures or low-diversity mixtures. In consequence, random samples of
species from a pool will, on average, show higher biomass, and higher productivity, as species richness increases (Tilman 1997) (Fig. 2a). In a sense, this is
the biological null hypothesis for the relationship between biodiversity and
ecosystem processes. It arises because species diffe:" in their ecologies, and
hence is due to 'hidden niche differences' between species. I refer to the niche
differences in this model as 'hidden' because they are not explicit in the model.
But clearly they are there, otherwise all species would behave identically. If all
species are identical, there cannot be any relationship between species richness
and ecosystem processes. Or, to rephrase this key point, I know of no theoretical mechanisms whereby a set of identical species could produce a relationship
between species richness and ecosystem processes (Lawton et al. 1998).
(b) The niche complementarity hypothesis. Here the niche differences are explicit
(Tilman et al. 1997a; Loreau 1998a). Niche-differences between species ensure that as species richness increases in ecological assemblages, so does the
range of 'functional space' occupied by the assemblage (Carlander 1955; Tilman
et al. 1997a). More diverse plant communities are more 'space-filling' above
ground (Naeem et al. 1994, 1995), have a greater variety of rooting depths
(Hooper 1998; Hooper and Vitousek 1997, 1998), a wider range of requirements for below-ground resources, and so on. Within anyone habitat, and
other things being equal, on average it is therefore almost inevitable that more
species rich assemblages have greater productivity, because they are able to
exploit a greater variety of limiting resources. Intercropping in agriculture and
agroforestry exploits this phenomenon (Ewel et al. 1991; Swift and Anderson
1993). In the fisheries literature, Carlander (1955) recognised over 40 years
ago that niche differences between species mean that species-rich fish assemblages sustain greater biomass than species-poor assemblages.
Although the sampling hypothesis and the niche complementarity hypothesis both rely on there being ecological differences between species, they differ
in a subtle and important way. Under the simple sampling hypothesis, the
maximum yield of a polyculture will not exceed that of the best-yielding monocultures. Under the niche complementarity hypothesis, it is possible for (but
not inevitable that) species rich mixtures will out-perform the best performing
monocultures (Fig. 2b). The phenomenon has been called 'overyielding' in the
123
tern processes is pattern (ii) (Fig. 1 b), consistent with the rivet hypothesis. The
fundamental underpinning mechanism is that species differ in their ecologies, that
is, no two species have identical niches. But there are some important, and still
poorly understood, subtleties underlying this simple statement, because there are at
least three ways in which the different ecological requirements of different species
can generate relationships broadly similar to Fig. lb. They are, in order of increasing biological complexity:
(a) The sampling hypothesis. Species differ intrinsically in their potential maximum size, growth rate, and so on. Hence, as Huston (1997) points out, mixtures of many species are more likely to contain high-yielding species than
monocultures or low-diversity mixtures. In consequence, random samples of
species from a pool will, on average, show higher biomass, and higher productivity, as species richness increases (Tilman 1997) (Fig. 2a). In a sense, this is
the biological null hypothesis for the relationship between biodiversity and
ecosystem processes. It arises because species diffe:" in their ecologies, and
hence is due to 'hidden niche differences' between species. I refer to the niche
differences in this model as 'hidden' because they are not explicit in the model.
But clearly they are there, otherwise all species would behave identically. If all
species are identical, there cannot be any relationship between species richness
and ecosystem processes. Or, to rephrase this key point, I know of no theoretical mechanisms whereby a set of identical species could produce a relationship
between species richness and ecosystem processes (Lawton et al. 1998).
(b) The niche complementarity hypothesis. Here the niche differences are explicit
(Tilman et al. 1997a; Loreau 1998a). Niche-differences between species ensure that as species richness increases in ecological assemblages, so does the
range of 'functional space' occupied by the assemblage (Carlander 1955; Tilman
et al. 1997a). More diverse plant communities are more 'space-filling' above
ground (Naeem et al. 1994, 1995), have a greater variety of rooting depths
(Hooper 1998; Hooper and Vitousek 1997, 1998), a wider range of requirements for below-ground resources, and so on. Within anyone habitat, and
other things being equal, on average it is therefore almost inevitable that more
species rich assemblages have greater productivity, because they are able to
exploit a greater variety of limiting resources. Intercropping in agriculture and
agroforestry exploits this phenomenon (Ewel et al. 1991; Swift and Anderson
1993). In the fisheries literature, Carlander (1955) recognised over 40 years
ago that niche differences between species mean that species-rich fish assemblages sustain greater biomass than species-poor assemblages.
Although the sampling hypothesis and the niche complementarity hypothesis both rely on there being ecological differences between species, they differ
in a subtle and important way. Under the simple sampling hypothesis, the
maximum yield of a polyculture will not exceed that of the best-yielding monocultures. Under the niche complementarity hypothesis, it is possible for (but
not inevitable that) species rich mixtures will out-perform the best performing
monocultures (Fig. 2b). The phenomenon has been called 'overyielding' in the