Where The Wild Things Were: Re-wilding and the Sixth Extinction - Biodiversity Conservation and Environmental Change - page 43

Ecological observations show the association between herbivores and open areas in for-
ests. For example, deer maintain grassy glades in British upland forests (Peterken 1996); bea-
ver can create locally open patches, and elk ( Alces alces ) appear to prevent woodland
succession in fenland sedge communities in Poland (Svenning 2002). Cattle grazing in nature
reserves in the Veluwe in the Netherlands encourages the development of grazing lawns in
woodland gaps created by storms, through seed dispersal, grazing, and trampling (Bokdam
2003). Thus, whether fire, herbivory or both are involved, woodland composition can be
thought of as a dynamic, shifting mosaic that can maintain an overall steady state at large
spatial scales, because individual patch dynamics are out of phase, thus maintaining both
shade-tolerant and light-demanding taxa in the woodland mix (Watt 1947, Olff et al. 1999, Vera
2000, Rackham 2003).
There is good ecological and evolutionary evidence for a patch-mosaic, herbivore-and
disturbance-adapted flora in European woodlands. For example, many woodland plants
have herbivore defences like thorns, and the success of management strategies like hedging
and coppicing depends on the ability of trees to re-sprout after being browsing. Thorny plants
(e.g. Rubus , Ilex , Crataegus , Ulex , Prunus spinosa ) can shelter more palatable species. For
example, experiments have shown that blackthorn ( Prunus spinosa ) enhances the establish-
ment of oak seedling ( Quercus rober ) in grassland areas where herbivores were present (Bak-
ker et al. 2004), a finding consistent with the scrub phase of the wood-pasture hypothesis
(Figure 3.5). Studies of fossil beetles and pollen show evidence of the consistent presence of
light-demanding taxa in European woodlands, which would not have been present if forests
were closed canopies (Sandom et al. 2014).
As mentioned above, megaherbivore loss would have reduced open vegetation and mosaic
habitats, increased fires due to accumulated biomass, and caused a decline in coevolved
plants. Sandom et al. (2014) showed that mosaics of forest and wood-pasture were common
in the last interglacial, when herbivores were abundant, but closed forests became more
common from the early Holocene, associated with megafaunal attrition, and later due to
hunting of the remaining herbivores by people. Still, the Holocene pollen record from many
European woodlands shows consistent presence of light-demanding species such as Corylus
(hazel) and Quercus (oak) throughout the Holocene (Birks 2005), indicating heterogeneous
canopy cover.
Despite this compelling ecological evidence, there is still scepticism about the wood-
pasture hypothesis. A particularly ingenious study attempted to resolve the debate by com-
paring fossil pollen data from central Europe, Britain, and Ireland, which have different
histories of herbivory (Mitchell 2005). Whereas mainland Europe and Britain harboured sig-
nificant populations of large herbivores, Ireland only supported wild boar and red deer pop-
ulations during the early Holocene, and Mitchell therefore argues that Ireland is effectively a
'grazing exclosure' (Table 3.2). From c. 9,500-6,500 years ago, he found no significant differ-
ent in oak and hazel pollen abundance in all three areas, suggesting similar forest structure.
On this basis, Mitchell argues that herbivory therefore could not be an import driver of wood-
land structure, and that more open forests only became prevalent from about 3,000 years ago,
due to human management.