322 Ecological Restoration and Design
Within-field non-crop vegetation
A second characteristic of industrialized agriculture is active weed control, gener-
ally through the use of herbicides. Removal of weeds can be antagonistic to arthro-
pod pest management. Often these non-crop plants contribute towards the resource
concentration hypothesis effect. They may also favour natural enemies by provid-
ing non-host foods such as pollen and nectar, support non-pest alternative hosts or
prey, and provide shelter or a moderated microclimate. Reintroducing some weeds
in ways that are agronomically acceptable to farmers can restore some of the ‘eco-
system services’ they provide to natural enemies (Nentwig et al, 1998). In the latter
study, 1.5m-wide strips of weeds used at 24m intervals within cropped areas
favoured natural enemies such as syrphids, chrysopids and coccinellids. An alter-
native approach is to withhold all or some herbicide application in part of the crop
and allow growth of the existing weed community. The headland (the area of the
field in which farm machinery turns and crop productivity is often lower because
of the resulting soil compaction) has been used in this manner. In British work,
such ‘conservation headlands’ provided pollen sources for adult syrphids, Episyr-
phus balteatus (Degeer) (Cowgill et al, 1993). This resulted in a significant positive
relationship between number of syrphid eggs per aphid and the density of weeds.
Genetically modified herbicide resistant crops could increase within-crop
botanical diversity and invertebrate biodiversity if farmers used the technology to
apply herbicides after crop emergence and only if weed density justified spraying.
This approach contrasts with the currently common practice of prophylactic appli-
cation of pre-emergent herbicides, which is necessary because many weed species
are impossible to kill selectively after crop emergence. There will, however, be a
temptation to use herbicide resistant crops as an opportunity to exercise more
stringent levels of weed control than in conventional crops. This would have a
negative impact on biodiversity.
Diversification can apply also to perennial crop systems such as orchards and
vineyards, where vegetational structure can include a distinct understorey. A con-
siderable amount of work has taken place examining the effects of this relatively
common form of diversification and it is particularly widely practised in China.
The ground cover plant Ageratum conyzoides L. (Asteraceae) has been planted or
conserved in 135,000ha of citrus where it is claimed to stabilize populations of
Amblyseius spp., predators of the citrus red mite, Panonychus citri McGregor (Liang
and Huang, 1994).
Plants other than ‘weeds’ may also be introduced as strips within the area of
crops. ‘Beetle banks’, low ridges sown to perennial, tussock forming grasses have
been used in British and mainland European arable crops for over a decade to
provide overwintering habitat for natural enemies of aphid pests (Thomas et al,
1991; Thomas and Goulson, 2000). Migration of predatory beetles and spiders
from these habitats into the crop in the spring supplements those from the field
margin. Further, because beetle banks run through the centre of fields, predators
migrating from them have immediate access to regions of the crop remote from