conservation, although the value of this practice is
subject to debate among ecologists.
Besides reducing and concentrating nutrients,
herbivores also influence vegetation structure and
succession. Aboveground herbivory does not nec-
essarily retard succession in all stages. Above-
ground herbivores may slow down succession in
some stages, but, initially, they may accelerate suc-
cession (Davidson 1993). Aboveground herbivores
interact, thereby facilitating or inhibiting each
other’s effects. In an exclosure experiment, small
herbivores (rabbits, voles) did not have consistent
effects along a productivity gradient, whereas cattle
increased plant diversity at high-production sites,
but reduced plant diversity at low-production sites
(Bakkeret al. 2006). Large vertebrate herbivores,
especially cattle, are easy to manage and their ef-
fects are relatively predictable. Invertebrate herbi-
vores can be controlled far less well and
interactions between vertebrates and invertebrates
have been rarely explored (Tscharntke 1997). In a
study focusing on sap-feeding insects, Schmitzet al.
(2006) showed that the sap feeders could be con-
trolled by both bottom-up and top-down forces,
which change the rate of succession due to abrupt
shifts in trophic control.
7.3.2 Secondary succession from an aboveground–belowground perspective
When secondary succession begins in post-agricul-
tural fields, the soil already contains a well-devel-
oped food web that includes decomposers,
bioturbators, symbionts, herbivores and pathogens.
Secondary succession, therefore, requires the trans-
formation of an arable soil food web into a soil food
web characteristic of a more natural system that
receives less soil disturbance, no mineral nutrients
and other fertilizers, no biocide spraying and no
monocultures of crop species. As a result, the soil
food web will transform from a bacteria-dominated
to a more fungi-dominated soil food web, although
this process is not necessarily linear over time (Van
der Walet al. 2006). The role of mutualistic sym-
bionts (especially of mycorrhizal fungi) and of bio-
turbators (earthworms) increases when succession
proceeds. The initial changes in community compo-
sition can be partly due to the absence of regular
mechanical soil disturbance (Van der Walet al.
2006). The dynamics of soil pathogens and root
herbivores will change from whole field oscillation
patterns coinciding with crop presence to finer spa-
tial and temporal patch dynamics as related to the
spatio-temporal dynamics in vegetation composi-
tion. Therefore, the soil community will acquire
(as well as generate) more spatial complexity,
whereas the food web will develop from a fast to a
slower cycling of nutrients and energy (Fig. 7.1).
Plants interact constantly with other organisms in
soil- and aboveground communities and these feed-
back effects can drive succession. Feedback interac-
tions with soil decomposer organisms proceed
indirectly, through root exudation, decaying roots
and litter. These interactions most likely take place
at a slower rate than feedback interactions with sym-
bionts and root pathogens, which have a more direct
and intimate association with plant roots. Feedbacks
between plants and their soil community can be
assessed by growth experiments and the net effects
range from negative to positive. These effects can
develop over several months. Negative feedback
leads to coexistence, whereas positive feedback re-
sults in dominance of plant species (Bever 2003). The
sign of plant–soil feedback varies along successional
sequences. In a 35-year-old chronosequence, early
successional plant species experienced negative soil
feedback, mid-successional species had neutral feed-
back and later successional plant species had posi-
tive feedback from the soil community (Kardolet al.
2006). This suggests that early succession soil com-
munities enhance and later succession soil commu-
nities slow the rate of secondary succession.
Moreover, plant–soil feedbacks of early successional
plant species make them less competitive and also
cause a legacy effect to mid-successional plant spe-
cies. Such feedbacks can result from soil microorgan-
isms (Kardolet al. 2007) as well as from soil fauna
(De Deynet al.2003).
In post-production grassland soils, the soil fauna
selectively reduces the abundance of early succes-
sional plant species, as well as that of the dominant
plants (De Deynet al. 2003). While soil fauna en-
hanced the rate of succession, it increased evenness
in plant community composition (De Deynet al.
2003). Nutrient addition increases dominance of
fast-growing plant species, but the soil community88 APPLICATIONS