98 N.J. Fox et al.
without spillover from other species (Naranjo
et al., 2008). This is supported by evidence from
estates fenced off from other domestic livestock
for over two decades, where M. bovis has been
found circulating in wild ungulates (Gortazar
et al., 2005). Areas where wild boar are believed
to act as maintenance hosts include the Mediter-
ranean regions of south-western Spain and
south-eastern Portugal. In such regions, reduc-
tions in numbers of wild boar have led to declines
in M. bovis levels in sympatric cattle and deer
(Boadella et al., 2012; García-Jiménez et al.,
2013). In contrast, intra-specific transmission is
low in areas with low pig population densities,
and outside of the afore-mentioned Mediterra-
nean regions feral pigs are considered to be
dead-end hosts. However, spillover from other
host species can be frequent. In areas with low
pig population densities but high levels of infec-
tion in sympatric wildlife, prevalence in feral
pigs can reach 100%, yet prevalence in these
pigs rapidly declines following implementation
of intensive lethal control targeted at other host
species (e.g. brushtail possums in New Zealand
[Nugent et al., 2015b] and bovids in Australia
[McInerney et al., 1995; Corner, 2006]). In such
regions, suids are useful sentinel hosts for deter-
mining M. bovis presence in target areas, due to
their large home range size and omnivorous pro-
clivity for scavenging carcasses of infected hosts
(Nugent et al., 2002). This example highlights
the importance of understanding the role of dif-
ferent host species in infection maintenance,
rather than merely targeting those showing
high disease prevalence.
The role of individual species in multi-host
systems is complex, and not simply density
dependent. Although high density maintenance
hosts play a driving role in disease dynamics,
low-density spillover hosts are an important
component of host communities. Spillover spe-
cies with large home ranges and long-lived, sub-
clinical infections can act as spatial and temporal
vectors, spilling back M. bovis to true mainte-
nance hosts as their expansive ranges overlap
with naïve populations, or as maintenance host
densities recover following population reduc-
tion. Wild deer provide an example of this phe-
nomenon. Outside of the US, the main role of
deer in disease maintenance is attributed to their
longevity and large home ranges (Nugent et al.,
2015a). After M. bovis has been eliminated in its
true maintenance hosts through control of
short-lived carriers (e.g. possums), subclinically
infected deer can act as temporal vectors surviv-
ing for many years and carrying M. bovis. They
can then spill infection back to more susceptible
yet short-lived host species as populations of
these hosts recover rapidly after culling efforts.
The large home ranges of wild deer also provide
a mechanism for M. bovis to be translocated to
new areas, prior to spillback into more common
host species. If this phenomenon is known to be
present, it is straightforward to take into account
in the development of control strategies.
In multi-host systems it is often difficult to
isolate the role of individual species. Such com-
plexity is epitomized in the species-rich reserves
of Southern Africa. M. bovis has been found in
16 different species in Southern Africa, with evi-
dence of inter- and intra-species transmission
(Hlokwe et al., 2014). There is evidence that
M. bovis distribution is increasing, and the num-
ber of known host species is on the rise (Hlokwe
et al., 2014). Thought to be first introduced into
the national parks by cattle in nearby farms,
African buffalo are now thought to be the pri-
mary maintenance hosts of M. bovis, capable
of retaining it in ecosystems in the absence of
domestic cattle (Michel et al., 2006). However,
other species (e.g. greater kudu and lechwe)
have also been shown to act as maintenance
hosts (Michel et al., 2006; Mwacalimba et al.,
2013). M. bovis is also found in predator species,
and high prevalences are found in lion prides
(Keet et al., 2010). As predators preferentially
target weak and debilitated prey, wild ungulates
succumbing to clinical infection are clear targets
for predation, and large predators and scaveng-
ing omnivores are exposed to high quantities of
infectious tissue. Due to the high species rich-
ness and inter-species exposure levels it has not
been possible to determine whether such species
(e.g. lions) can be maintenance hosts, and the
true role of individual species in the multi-
faceted, multi-host complex is not fully
understood.
As has been demonstrated in New Zealand,
culling of wildlife hosts can be an effective way
of controlling M. bovis. However, possums are an
alien species in New Zealand and a similar con-
trol approach is not feasible if culling could
threaten the survival of protected and endan-
gered indigenous species. Southern African