3.6 COMPLEXITY AND STABILITY IN PRACTICE

20.3.6 Complexity and stability in practice: whole

simply reflect a diversity of forces acting on different webs.

communities

The prediction that populations in richer communities are less

stable when disturbed can also be investigated experimentally. One

classic study, for example, monitored the resistance in two grass-

Turning to the aggregate, whole community level, evidence is

land communities (McNaughton, 1978). In the first, plant nutri-

largely consistent in supporting the prediction that increased

ents were added to the soil of a community in New York State;

richness in a community increases stability (decreases variability),

in the second, the action of grazing animals was manipulated in

though a number of studies have failed to detect any consistent

the Serengeti. In both cases, the treatment was applied to

relationship (Cottingham et al., 2001; Worm & Duffy, 2003).

species-rich and species-poor plant communities, and in both, dis-

First, returning to McNaughton’s

(1978) studies of US and Serengeti grass-

turbance reduced the diversity of the former but not the latter

data support the

lands, the effects of perturbations were

(Table 20.1). This was consistent with the prediction, but the effects,

models: aggregates

while significant, were relatively slight.

quite different when viewed in ecosys-

are more stable in

richer communities

Similarly, Tilman (1996) pooled data for 39 common plant

tem (as opposed to population) terms.

species from 207 grassland plots in Cedar Creek Natural History

The addition of fertilizer significantly

increased primary productivity in the species-poor field in New

Area, Minnesota, over an 11-year period. He found that variation

in the biomass of individual species increased significantly, but only

York State ( + 53%), but only slightly and insignificantly changed

productivity in the species-rich field ( + 16%); and grazing in the

very weakly, with the richness of the plots (Figure 20.10a).

Finally, there have been a number of studies directed at the

Serengeti significantly reduced the standing crop biomass in

question of whether the level of ‘perceived stability’ of natural

the species-poor grassland ( − 69%), but only slightly reduced that

populations (interannual variation in abundance) varies with the

of the species-rich field ( − 11%). Similarly, in Tilman’s (1996)

Minnesota grasslands, in contrast to the weak negative effect found

richness or complexity of the community. Leigh (1975) for her-

bivorous vertebrates, Bigger (1976) for crop pests and Wolda (1978)

at the population level, there was a strong positive effect of rich-

ness on the stability of community biomass (Figure 20.10b).

for insects, all failed to find evidence that it did so.

(a)

1200

250

r

= 0.15**

1000

N

= 729

200

800

150

600

100

400

50

Coefficient of variation for species biomassr²

= 0.74

Standard deviation of CO2 flux (µl 18 h–1)

0 5 10 20

0 0

0 15

5 10 15 20

Realized species richness

Species richness

(b)

80 Field A

r

= –0.39**

90

Figure 20.11 Variation (i.e. ‘instability’) in productivity

80 Field B

r

= –0.32*

70

(standard deviation of carbon dioxide flux) declined with species

60

richness in microbial communities observed over a 6-week period.

40

Richness is described as ‘realized’ because it refers to the number

30

20

of species present at the time of the observation, irrespective of

10

the number of species with which the community was initiated.

0

20 0

4

2 6 8 10 12

(After McGrady-Steed et al., 1997.)

80 Field C

r

= –0.09(NS)*

80 Field D

r

= –0.53***

Studies of the response of a community to a perturbation

(e.g. McNaughton, 1978) or of variations in the community

in response to year-to-year variations in the environment (e.g.

10 20

Tilman, 1996), are focused largely on the resistance of com-

1214161820

2 6 8 10 22

14 16

munities to change. A quite different perspective examines the

Average species richness

resilience of communities to perturbations in ecosystem charac-

teristics such as the energy or nutrient levels contained within

Figure 20.10 (a) The coefficient of variation of population

them. O’Neill (1976), for example, considered the community as

biomass for 39 plant species from plots in four fields in Minnesota

a three-compartment system consisting of active plant tissue (P),

over 11 years (1984 –94) plotted against species richness in the plots.

heterotrophic organisms (H) and inactive dead organic matter (D).

Variation increased with richness but the slope was very shallow.

The rate of change in the standing crop in these compartments

(b) The coefficient of variation for community biomass in each

depends on transfers of energy between them (Figure 20.12a).

plot plotted against species richness for each of the four fields.

Inserting real data from six communities representing tundra,

Variation consistently decreased with richness. In both cases,

tropical forest, temperate deciduous forest, a salt marsh, a fresh-

regression lines and correlation coefficients are shown. *, P < 0.05;

water spring and a pond, O’Neill subjected the models of these

**, P < 0.01; ***, P < 0.001. (After Tilman, 1996.)

communities to a standard perturbation: a 10% decrease in the

initial standing crop of active plant tissue. He then monitored

the rates of recovery towards equilibrium, and plotted these as

McGrady-Steed et al. (1997) manipulated richness in aquatic

a function of the energy input per unit standing crop of living

tissue (Figure 20.12b).

microbial communities (producers, herbivores, bacterivores and

The pond system, with a relatively

predators) and found that variation in another ecosystem measure,

importance of the

low standing crop and a high rate of

carbon dioxide flux (a measure of community respiration) also

nature – not just

declined with richness (Figure 20.11). On the other hand, in an

biomass turnover, was the most resili-

the richness – of

ent. Most of its plant populations have

experimental study of small grassland communities perturbed

the community

by an induced drought, Wardle et al. (2000) found detailed com-

short lives and rapid rates of population

munity composition to be a far better predictor of stability than

increase. The salt marsh and forests had intermediate values, whilst

tundra had the lowest resilience. There is a clear relationship

overall richness.

cycling rather than energy flow. Here too, then, stability seems

more influenced by the nature of the species in the community

than by simple measures such as overall richness.

Net primary

production