100 N.J. Fox et al.
(Hutchings and Harris, 1997). Badgers will also
enter farm buildings to forage on livestock feed
(Garnett et al., 2002; Tolhurst et al., 2009), and
proximity contact between badgers and cattle is
more frequent when cattle are housed than
when they are grazing (Tolhurst et al., 2009).
Such behaviours provide potential for direct (via
aerosols) and indirect (via environmental con-
tamination) transmission. As with possums,
M. bovis infection can alter badger behaviour,
and thus transmission risk, as terminally ill bad-
gers lose their fear of cattle and direct contact
rates can increase (Gavier-Widen et al., 2001;
Corner, 2006).
Badgers are a legally protected species in
the UK and the risk of transmission of M. bovis
from badgers to cattle remains controversial.
With the national control strategy failing and
M. bovis spreading and re-emerging in areas
where infection in cattle had previously been
eradicated, the UK undertook the Randomised
Badger Culling Trial (RBCT). With a budget of
£50 million, 10-year timeframe and covering
3000 km^2 (McDonald et al., 2008), the RBCT is
one of the largest ever controlled veterinary
epidemiology field experiments in UK history.
The aim of the RBCT was to quantify the impact
of culling badgers on the number of M. bovis
herd breakdowns in cattle (Bourne, 2007).
Somewhat counter-intuitively, the RBCT con-
cluded that while badgers do play a role in
M. bovis outbreaks in cattle, in some cases the
incidence in cattle could be increased by culling
of badgers (McDonald et al., 2008). In the cen-
tral core of areas where badgers were proac-
tively culled, the incidence of M. bovis in cattle
decreased, whereas the incidence increased at
the edge of these culled areas and increased if
patchy reactive culling of badgers was applied
following M. bovis breakdowns in cattle herds
(Donnelly et al., 2006). The inefficiency of cull-
ing is indicative of a perturbation effect, where
culling induces perturbations in the social
organization of badgers, disrupting territories,
enhancing movement and increasing badger-
to-badger, and badger-to- cattle transmission
(McDonald et al., 2008; Prentice et al., 2014).
The perturbation effect is potentially amplified
in disease systems with heterogeneous distribu-
tions of infection in highly structured host
populations (Prentice et al., 2014). M. bovis is
highly clustered in individual badger social
groups (Delahay et al., 2000). Incomplete pop-
ulation reduction may drive relatively rapid
increases in prevalence by increasing contact
between dispersing infected animals and sus-
ceptible populations. Although badger–cattle
transmission mechanisms remain poorly
understood, the large scale and controlled
nature of the RBCT experiment is the best avail-
able evidence that badgers can play a role in the
epidemiology of M. bovis in livestock. Interest-
ingly, it was the incomplete nature of the bad-
ger population reduction that enabled the
results of the RBCT, as removing all the badgers
would by definition remove the badger-to-cattle
transmission.
In the USA, deer are believed to play a role
in M. bovis outbreaks in cattle, and evidence of
spillback from deer to cattle has necessitated
implementation of control strategies (Palmer
et al., 2004b). Deer frequently visit feeding areas
in cattle farms and evidence of this has been
recorded in radio collared white-tailed deer in
Michigan (Berentsen et al., 2013). However,
transmission between cattle and deer is thought
to be indirect via shared resources, as direct con-
tacts between the two are rare. Due to the poten-
tial of deer to act as maintenance hosts in the
US, they have been a target of population reduc-
tion. In contrast to badgers, reducing deer popu-
lations in the US has been shown to reduce
M. bovis incidence with no sign of a perturbation
effect. As these deer exhibit a high degree of site
fidelity, occupancy of overlapping home ranges
and lack territorial defence behaviour, culling is
unlikely to alter home ranges of adjacent deer
groups and perturbation effects are not observed
(Palmer et al., 2015).
Before employing disease control based on
population reduction, it is important to under-
stand the idiosyncrasies of different host popula-
tions. However, these idiosyncrasies are often
unknown. Despite the investment of extensive
resources in trying to understand the role of
wildlife in livestock M. bovis outbreaks, live-
stock–wildlife transmission dynamics remain
poorly understood. However, the observations
that culling wildlife affects M. bovis incidence in
cattle further reiterates that disease processes
act at the host community and ecosystem level
rather than on isolated species.