INTRODUCTION

The purpose of this document is two-fold. It aims to describe evidence
from the scientificliterature about the costs (financial and
ecological), benefits and role of grey squirrel cullingin red squirrel
conservation while at the same time examining whether grey squirrel
culling isa long-term solution to red squirrel conservation or whether
other solutions need to bedeveloped.

THE HISTORY OF RED SQUIRRELS IN BRITAIN

The Eurasian red squirrel (Sciurus vulgaris) is a widespread species
found in most woodedareas of Eurasia from Iberia to Kamchatka
Peninsula and Sakhalin Island (Russia), and GreatBritain south to
Mediterranean and Black Sea, north Mongolia, west and north-east
China(Gurnell, 1991; Lurz et al., 2005; Wilson & Reeder, 2005). In
Britain, the red squirrel wasthought to be distinct subspecies (S.
vulgaris leucorus) characterised by a distinctive annualbleaching of
hair, especially on the tail and ears (Lowe & Gardner, 1983). However,
recentDNA analysis indicated no evolutionary divergence between
British and continentalpopulations (Barratt et al., 1999). This is
unsurprising for two reasons; first populations inBritain have been
isolated for a relatively short time (8000 years) since the last ice
age andmay not have evolved sufficiently to acquire monophylognetic
status (Barratt et al., 1999);second, red squirrels were deliberately
reintroduced into Scotland and England fromcontinental stock at
various times during the last 150 years (Ritchie, 1920; Lowe &
Gardiner,1983), and many were also imported from the continent as a
food source and may haveescaped (Freethy, 1983). Further genetic
analysis supports this earlier hypothesis, although itwas suggested
that most British squirrels are descended from more recent
introductions fromScandinavia (Hale et al., 2004). Using museum
specimens, Kitchener et al. (2003) suggestedthat the Cumbrian
population of red squirrels remained a distinctive British subspecies,
andthat other populations were the result of introductions. However,
further museum workshowed a temporal loss of distinctive phenotypic
characteristics from the Cumbrianpopulation, which was caused by
increased gene flow into the population following increasedplanting of
woodland, which linked the Cumbrian and Scottish populations (Hale et
al.,2001; Hale & Lurz, 2003). As a consequence, all British red
squirrels represent a mixedgenetic stock (Barratt et al., 1999; Hale
et al., 2004).

Populations of red squirrels are notably volatile, undergoing marked
changes in numbersboth in the long- and short-term. Historically,
populations of red squirrel declined to nearextinction during the
18thcentury as a consequence of deforestation, eventually
necessitatingthe reintroduction of red squirrels from the continent
(Yalden, 1999). During the 19thcentury, forest plantations originally
planted at the end of the 18thcentury reached maturity.This led to an
increase in the squirrel population to the extent they were considered
a majorpest and necessitated culling (Tittensor, 1975). Locally they
became so numerous that redsquirrel destruction clubs were formed and
a bounty scheme implemented (Ritchie, 1920;Shorten, 1954). Large
numbers of squirrels were killed; over fifteen years, the Highland
Squirrel Club killed 60,500 red squirrels in just one area (Ritchie,
1920). Later populationdeclines during 1900-1930 have been linked to a
disease epizootic (Middleton, 1930). Sincethe 1940s population size
and distribution of the red squirrel in Britain have
contracted,largely in parallel with the increase in range of the grey
squirrel (Sciurus carolinensis).

THE POPULATION OF GREY SQUIRRELS IN BRITAIN

The grey squirrel is a native of eastern North America, inhabiting a
range of forestedhabitats, particularly deciduous woodland. The grey
squirrel has repeatedly been introducedto Britain between 1870-1930.
Notable introductions include 100 in Richmond Park, Surreyand 91 in
Regent’s Park, London (Shorten, 1954). Introductions were not confined
tosouthern England, but included releases all over Britain. With such
large numbers releasedand at so many points, it is perhaps
unsurprising that grey squirrel populations took hold inthis country.
Initially, populations spread locally; in 1920 populations were
focused aroundinitial release points, but by 1945 grey squirrels had
expanded their range to much ofsouthern and central England, although
spread into East Anglia and the southwest occurredlater. Later surveys
in 1970 and 1990 indicated that grey squirrel populations had
coveredmost of England and Wales, with two separate populations in
Scotland; one was basedaround the mid-Lothian, Fife, Tayside and
Glasgow, a second was limited to Aberdeen andthe surrounding area
(Lloyd, 1983; Gurnell & Pepper, 1993). The population in theAberdeen
area was thought to be isolated from the more southern population and
itsappearance was a result of a deliberate introduction. However,
Staines (1986) suggested thatthe Aberdeen population was the result of
natural colonisation from central Scotland, and thenon-contiguous
nature of some grey squirrel distribution maps in Scotland reflect
poorsurvey coverage of some areas.

CAUSES OF RED SQUIRREL POPULATION DECLINE

Causes of red squirrel populations declines can be categorised into
three factors: habitat loss,ecological replacement by grey squirrels
and disease.Habitat loss and red squirrel declines

Habitat loss has been an important factor in the decline of red
squirrels in the past. Thewidespread destruction of forests led to the
almost complete extinction of red squirrels inBritain (Ritchie, 1920;
Gurnell, 1987). Only through reintroductions from continental stockand
the increase in forestry plantations did the red squirrel recover.
However, the frequency with which certain tree species have been
planted are important. For example, pre-1920,40% of planted woodland
was oak; this habitat can lead to a competitive disadvantageagainst
grey squirrels (see section on ecological replacement). Currently,
habitat loss is notan important factor in red squirrel declines and
forest coverage is increasing (ForestryCommission, 2001; Scottish
Environment Statistics, 2006); however, the suitability of thisnew
forest for red squirrels is low. Habitat preferences of red squirrels
vary with locationand densities (Bryce et al., 2002; 2005).

 It is considered that ‘good’ habitat for red squirrels are those that
reds utilise, but grey squirrels avoid. These include Norway spruce,
larch andDouglas fir (Bryce et al., 2002). Poor habitat was either
selected by greys or avoided byreds; these include sitka spruce
(avoided by both species) and Scots pine (used by both).Overall, the
planting of forests in England and Scotland has not increased the area
of suitablered squirrel habitat (Table 1), and in fact the increase in
broadleaved woodland has been ofgreater benefit to grey squirrels
(Table 1).Table 1. The changes in woodland cover in England
(1980-1998) (Forestry Commission,2001) and Scotland (1980-1995)
(Scottish Environment Statistics, 2006), with the suitabilityof
habitat for squirrels based on definitions by Bryce et al.
(2002)SpeciesPercentage forestchange (England)Percentage forestchange
(Scotland)Suitability forsquirrelsSitka spruce+14+61Poor for both
speciesNorway spruce-21-28Good for bothDouglas fir+3+1Good for
redsEuropean larch-34-39Good for redsMixed/Japanese larch-3+21Good for
redsScots pine-5+8Poor for bothBroadleaved woodland+36+68Good for
greysWhether a habitat is ‘good’ for red squirrel is highly dependent
on tree seed productivity.Coniferous woodland can have red squirrel
densities equal to deciduous woodlands (Gurnell,1987; Lurz et al.,
1995; Wauters & Lens, 1995), but this is highly dependent on
speciescomposition and age structure of the forest; man-made
coniferous woodland in northernEngland have the lowest seed
productivity and lowest red squirrel density (Wauters et al.,2001). It
has been suggested that introduced red squirrels from Scandinavia are
better adapted to spruce dominated woodlands, and so are able to
tolerate these habitats more thanthe ‘native’ red squirrel (Hale et
al., 2004). However, quantitative data supporting thishypothesis are
absent. The idea that coniferous plantation are ‘good’ habitat for
reds is false;in many places they simply provide a refuge habitat into
which grey squirrels have difficultyin spreading.

Ecological replacement by grey squirrels
There are areas of Britain where red and grey squirrels coexisted for
a number of years(Teanganna et al., 2000; Bryce et al., 2002). In some
cases, red squirrels populations becomeextinct after many years of
coexistence (Harris, 1973/74; Reynolds, 1985). This contrastswith the
disease-induced declines of red squirrels which are followed by rapid
colonizationby grey squirrels (see section on diseases). Due to this
apparently slow pattern ofreplacement, it has been suggested that grey
squirrels are out competing red squirrels. Datafrom coniferous
woodland showed little direct interspecific competition with grey
squirrel(Wauters et al., 2000); there was no evidence that either
adult survival rates or the proportionof female breeding decreased
(Wauters & Gurnell, 1999; Wauters et al., 2000), althoughhowever, both
female red squirrel fecundity and juvenile recruitment were reduced in
thepresence of grey squirrels (Wauters & Gurnell, 1999; Gurnell et
al., 2004). In deciduouswoodland, grey squirrels are better able to
exploit acorn crops and have a competitiveadvantage over hazelnut
crops; this is predicted to lead to slow replacement of red
squirrelsby greys (Kenward & Holm, 1993). The different feeding
abilities of the two species makesred squirrels more vulnerable to
food shortages (Gurnell & Pepper, 1993), which may lead toa reduction
in the number of breeding females and subsequent recruitment into
thepopulation. The presence of grey squirrels also makes it hard for
juvenile red squirrels toestablish territories. This competitive
advantage is greatest in deciduous woodland; the rateof replacement is
thought to be related to habitat composition, with populations
coexistingfor longer in coniferous habitats (Bryce et al., 2002).
However, as suggested by Gurnell etal. (2004), even in these habitats,
it appears that red squirrel populations will decline and redsquirrels
will only persist where they have a competitive advantage over greys.

Diseases and squirrel population declines

Red squirrel populations seem particularly susceptible to diseases and
historically, diseasehas caused many population declines (Middleton,
1930; Edwards, 1960; Vizoso, 1968; Scottet al., 1981; Keymer, 1983).
It is now known that a disease, squirrel poxvirus (SQPV), is a
significant factor in red squirrel declines (Tompkins et al., 2002;
Thomas et al.,2003).Whether SQPV was introduced to Britain along with
grey squirrels is unclear (Daszak et al.,2000), but it is more likely
that SQPV was endemic within the native red squirrel population,as the
symptoms of this disease have been described from red squirrel
populations prior tothe introduction of the grey squirrel (Middleton,
1930). What is clear is that grey squirrelsact as hosts without
developing the disease (Sainsbury et al., 2000) and they aid the
spread ofdisease into previously unexposed red squirrel populations
(Daszak et al., 2000). They thentake over the vacant ecological niche
caused by the red squirrel’s extinction (Tompkins et al.,2002). The
speed of red squirrel loss is markedly faster in the presence of the
disease thansolely with ecological replacement (Ruston et al., 2006).
By tackling the spread of thisdisease, red squirrel population loss
will be markedly slower.

What will happen to red squirrel populations?There are two main
factors causing current red squirrel declines: ecological replacement
bygreys and disease. In the absence of SQPV, it is predicted that red
squirrels would becomerestricted to those habitats where grey
squirrels cannot out-compete reds. However, SQPV isan important
factor. If this disease enters novel red squirrel populations, it is
likely that redswill die out and be replaced by greys, even in these
poorer quality habitats.


CONTROLLING GREY SQUIRRELS

There are four approaches to conserving red squirrels: stop grey
squirrels spreading SQPV,stop greys colonizing new habitat and
replacing red squirrels, control of the disease andincreasing suitable
habitat for red squirrels. Intuitively, the simplest method would
appear tobe to stop grey squirrels spreading, and so this is the
approach most widely implemented.This is done through some form of
population control: culling (poisoning, shooting,trapping),
immunocontraception or biological control.Grey squirrel population
control: cullingThere are four methods of culling: poisoning, spring
trapping, shooting and live trapping.Of these, poisoning is banned in
areas with red squirrels, and for all other locations it isdiscouraged
due to the likelihood of non-target species poisoning. Spring trapping
(a killingtrap) is also discouraged due to the indiscriminate nature
of the traps (Hodge & Pepper, 1998;Mayle et al., 2004). Three
techniques predominate: poisoning (where not banned), shootingand live
trapping.

Poisoning
Poisoning with warfarin is favoured in most commercial woodlands,
where it is viewed asboth cheap and effective (Taylor et al., 1968;
Pepper, 1990; Mayle et al., 2004). There issome risk that poisoned
squirrels carcasses may be consumed by scavengers. Fifty-sevenpercent
of warfarin poisoned squirrels fell to the grounds and 25% of these
were scavenged(Kenward, 1988). Warfarin has a short half-life and is
not usually present at high levels inthe bodies of poisoned animals.
This is because death is relatively slow and there is ampletime for
much of the toxicant to be metabolized or excreted (Record & Marsh,
1988).Exposure to warfarin through scavenging is not thought to have a
large risk of mortality tonon-target species, but may have unknown
sub-lethal affects (Townsend et al., 1981).However, direct exposure to
warfarin can be important. Wood & Phillipson (1977) estimatedthat
55-65% of bait for grey squirrels was consumed by non-target species,
primarily rodents;this can cause significant local population declines
(Brakes & Smith, 2005). Poisoning ofnon-target species is another
route to secondary poisoning in predators and scavengers(Brakes &
Smith, 2005). Warfarin is the recommended poison to control squirrels,
but asurvey by McDonald & Harris (2000) showed that 46% of game
estates used poisons otherthan warfarin to control grey squirrels.
These second generation compounds have longer halflives than first
generation poisons such as warfarin, and so the risk of secondary
poisoning isincreased (McDonald & Harris, 2000). Species regularly
exposed to warfarin can evolvegenetic resistance to its toxicity (e.g.
Quy et al., 1992). If this trait evolves in grey squirrels(Taylor et
al., 1968), the use of other poisons may lead to even higher
incidences ofsecondary poisoning in rare and threaten species (Carter
& Burn, 2000; Mason & Littin,2003). Furthermore, anticoagulants such
as warfarin are described as “markedly inhumane”by the Pesticides
Safety Directorate and there is pressure from the EU to reduce the use
ofpesticides in the countryside (Anon, 1997; Forestry Commission,
2006).

Shooting
Shooting can take place in two forms: shooting of individuals located
whilst walking througha forest, or drey poking which involves
targeting dreys with poles to flush out squirrels to beshot. The
latter is considered more effective (Tittensor, 1975). Although
shooting is anotherwidely used method to control populations, it is
viewed as insufficient to reduce squirrelnumbers significantly (Hodge
& Pepper, 1998).

Live trapping
Live trapping is considered the most effective technique to catch grey
squirrels (Gurnell,1999), and is the main technique used where red
squirrels are present e.g. Thetford Forest(Gurnell & Steele,
2002).Traps are typically baited with whole maize and it isrecommended
that traps are checked twice daily (Tittensor, 1975; Mayle et al.,
2004).Captured grey squirrels can then be euthanased and any
non-target species released unharmed.

Grey squirrel population control: immunocontraception and biological
control
Immunocontraception relies on inducing an immune response which
attacks the host’sreproductive system. Delivery mechanisms are
typically through either bait or a vectoredmechanism such as a virus
(Barlow, 2000). The idea to control grey squirrel reproduction isnot
new (Johnson & Tait, 1983), and there have been attempts to create a
bait deliveredcontraceptive (Moore, 1997). Initial field trials
indicated a drop in pregnancy rate from 50%to 15% (Pepper & Moore,
2001). However, as yet no effective mechanism has beendeveloped
(Rushton et al., 2002) and recently funding into this topic has been
ceased(Forestry Commission, 2006; JNCC, 2006). Whilst
immunocontraceptives have beendeveloped for a range of species such as
elephants (Loxodonta africana) (Fayer-Hosken etal., 2000) and
brushtail possums (Trichosurus vulpecula) (Cowan, 1996), there are few
dataon their effectiveness as a management tool (Ramsey, 2005). For
some populations ofcertain species e.g. white-tailed deer (Odocoileus
virginianus) (Rudolph et al., 2000), fertilitycontrol has led to a
significant decrease in population size, whilst for other
species,demographic compensation (increased longevity, increased
reproductive rates or increasedimmigration) can negate the effects of
fertility control (Twigg & Williams, 1999; Ramsey,2005). As yet there
is no clear consensus on the effectiveness of immunocontraception as a
technique to manage large populations. It is also unclear how species
with contrasting lifehistory strategies will respond to fertility
control (Caughley et al., 1992; Hone 1992), andspecies, such as grey
squirrels, with increasing populations or high reproductive rates may
bemost resilient to fertility control. In addition, it is likely that
in species with highreproductive rates there will be selective
pressure to evolve resistance to theimmunocontraceptive (Magiafoglou
et al., 2003). Certainly, a lot more work needs to bedone before the
technique can be applied to a successful invasive species such as the
greysquirrel.

Biological control can be in two forms, control through a disease or
parasite (Rowe, 1983;Dobson 1988), or through the introduction of a
predator (Rowe, 1983; Simberloff & Stiling,1996). Examples of
biological control agents include rabbit haemorrhagic disease (RHD)
inAustralia (Kovaliski, 1998). Currently no biological methods are
available for grey squirrels,and there may be unknown ecological
effects of introducing a biological control agent intothe environment.
As a consequence the only other option would be natural predators
ofsquirrels such as pine martens (Martes martes) and goshawks
(Accipiter gentilis) (Rowe,1983). However, as a result of long term
persecution, current populations of both predatorsare most abundant in
areas containing red squirrels, not greys; predation on red squirrels
isnot thought to be a direct threat to population viability
(Halliwell, 1997; Petty et al., 2003),and it is highly unlikely that
these predators would have a significant impact on grey
squirrelpopulations. So whilst the use of natural predators as a
biological control of grey squirrelsappears an attractive option, it
is unlikely to have a significant impact on either the spread ofgrey
squirrels or population size.

EFFECTIVENESS AND COST OF CONTROL TECHNIQUES

It is not possible to eradicate the grey squirrel from Britain using
current control methods(Gurnell & Pepper, 1993). Grey squirrels are
resilient to control both through rapidrecolonization of vacated areas
and by increasing their reproductive rate to compensate forlosses
(Lawton & Rochford, 1999). There is widespread evidence suggesting
that controltechniques and programs have had little or no success in
lowering the grey squirrelpopulation. Recent calls for a grey squirrel
bounty (Horne, 2006) are misguided. Previousbounty schemes during the
1950s culled 1.5 million squirrels but did little to impact
thepopulation even though they were a lot lower (Sheail, 1999, 2003;
Forestry Commission,2006). More locally, trapping in Thetford Forest
did not lower the population of greysquirrels, despite trapping 2200
squirrels in 4600 ha over a three year period (Gurnell &Steele, 2002).
A long-term poisoning campaign in Lady Park Wood, Gloucestershire, did
notprevent woodland damage; in fact, it actually increased during this
period (Mountford, 1997).The level of squirrel control through
shooting had no effect on squirrel numbers (Fitzgibbon,1993), and
evidence from spring trapping suggests no reduction in the number of
squirrelscaught (E. Brun cited in Huxley, 2003). However, if selected
and designed correctly,effective squirrel control can be targeted
locally usually through poisoning rather thanshooting or trapping
(Kenward et al., 1988a). Kenward et al. (1996) showed that all
squirrelsin a marked population could be removed in five weeks of
poisoning when poisoning effort was widespread through a vulnerable
wood. However, Lawton & Rochford (1999) showedthat culling regimes
only caused a temporary reduction in density, and that recolonisation
canbe extremely rapid. Recolonisation time varies from three months in
an isolated area to onlyone month in an area that is not isolated
(Mayle et al., 2004). Poole & Lawton (2004)showed that following five
culls, populations recovered to pre-cull levels in 3-10 weeks. Inthe
short-term, culling may only have a temporary effect on the
population. In particular,localised culling will be offset rapidly by
immigration from unculled areas. If a source-sinksystem is created,
this may lead to higher populations in culled areas and may
actuallyincrease the speed of grey squirrel spread (Pulliam, 1988).The
cost of control must also be considered as well as the effectiveness
of culling squirrels.Cost of control varies primarily by method used.
In most cases, the intensity of effort isinsufficient for effective
control. Costs are generally lowest in poisoning, but costs of
othermethods vary with intensity of use (Table 2). Give its low cost,
poisoning is the preferred method of control; in 1992, 140,000 grey
squirrels were poisoned, compared to 11,000 trapped (Pepper,
1992).Table 2. The costs of different forms of grey squirrel control.
* based on mean costs fordifferent areas of woodland. ** indicates
cost estimated.MethodologyCost/ha/annumSourceLive trapping£2.61Gurnell
et al. (1997)Warfarin£0.21-£6.00*Kenward et al.
(1988a)Warfarin£3.92Kenward & Dutton (1996)Spring trapping£12.50E.
Brun cited in Huxley (2003)Immunocontraception£12.84**Ruston et al.
(2002)Warfarin£17.62Huxley (2003)Spring trapping &
shooting£23.75Huxley (2003)Live trapping£59.49**Ruston et al
(2002)Costs of squirrel control can vary markedly (Table 2); for
example, costs of warfarin controlper hectare decrease with increasing
wood size (Kenward et al., 1988a) and increase withintensity of effort
(Kenward & Dutton, 1996). The costs of culling techniques must be
offsetby the benefits, which include financial costs through reduced
damage to tree (see section: other perceived negative impacts on grey
squirrels) or whether it can prove an effectivemethod to conserved red
squirrels. Given this cost-benefit of squirrel control,
warfarinpoisoning is the favoured technique, despite the concerns over
humaneness and itsindiscriminate nature.  

There have been recent calls by a number of agencies including
Scottish Natural Heritage tocull grey squirrels to control spread into
areas inhabited by red squirrels (Horne, 2006).However, past
experience shows that there is no evidence to suggest this will be
successful.

The most successful eradication of grey squirrels has been on
Anglesey, Wales, where theremoval of 6,000 squirrels in five years has
allowed red squirrel populations to increase(Shuttleworth et al.,
2002). The success of this effort has been in part due to the
geographyof Angelsey; being an island with limited connections to the
mainland has meant greysquirrels were unable to repopulate culled
areas. In contrast, there has been considerable trapping effort in
Thetford Forest, East Anglia, without any discernible effect on
greysquirrel numbers (Gurnell & Steel, 2002). The cost of controlling
mainland populations islarge; Rushton et al. (2002) estimated that it
would cost £300,000 per annum to ensure thepersistence of red
squirrels in Thetford Forest. Similarly, in Redesdale
Forest,Northumberland the cost would be £200,000 per annum. The key
aspect of squirrel controlis immigration into areas following control,
and so culling would be more productive inisolated populations. Recent
plans by Aberdeen City Council to cull grey squirrels seem
particularly misguided, as it is suggested that grey squirrel
populations in Aberdeen City and Deeside are isolated from those
further south (Anon., 2006). Yet this assertion contradicts other
evidence; Staines (1986) suggested that Aberdeen’s grey squirrels
immigrated from thesouth, and maps in Poulsom et al. (2005) indicate a
contiguous grey squirrel population allalong the east coast of
Scotland. So culling grey squirrels in Aberdeen will have no effect
onthe grey squirrel population and will not aid red squirrel
conservation; whilst culling maylower population size in the short
term, the high dispersal abilities of grey squirrels meansthat
recolonization is inevitable. Modelling by Rushton et al. (2006)
suggested that during aten year period, culling of greater than 60% of
the population per annum was necessary tostop grey squirrels spreading
SQPV into the Cumbrian red squirrel population, and hence toensure red
squirrel persistence. Targeted or intensive culling in areas where
grey squirrelsare spreading may be effective, but must be very long
term and is thus expensive. However,this is a prediction based on
computer modelling and there are no field data to show whethersuch
intensive culling will prevent the spread of SQPV. So, what the costs
(ecological and financial) and benefits of intensive culling remains
unclear. In contrast to areas where greysquirrels are spreading, it is
not considered practical to exterminate grey squirrels where theyare
already established (Mayle et al., 2004).