Content uploaded by Rob Smith
Author content
All content in this area was uploaded by Rob Smith on Jun 14, 2016
Content may be subject to copyright.
© CAB International 2014. Rodent Pests and their Control, 2nd Ed
330 (A.P. Buckle and R.H. Smith)
Introduction
It is usually very difcult to control a pest
(the ‘target’) using chemicals without causing
some collateral damage to other (‘non-target’)
species. Non-target damage should, of course,
be minimized. Whether or not non-target
damage is regarded as signicant usually
depends on whether only a few individuals
are affected or whether there is an impact on
the wider population. A few deaths of indi-
viduals may quickly be compensated for by
density-dependent processes (see Chapters
1 and 5), for example by increased births,
fewer deaths or compensatory migration.
Effects on populations are, however, of con-
cern, especially if those effects continue into
the next generation. In the case of animals
or plants of conservation interest, it may be
that no accidental death or impairment is
acceptable. In addition, it is often regarded
as unacceptable to cause suffering to higher
organisms such as birds and mammals. All
of these concerns involve value judgements
that differ between contexts and societies,
and often within societies. Different interests
and values must be balanced and weighed
up in a benets–harm analysis. The trade-off
between benet and harm that is acceptable
is a function of the legislative framework, the
pressure to reduce pest damage, and what is
commercially and ethically acceptable within
a particular market/society.
Much of the public concern about the
potentially adverse effects of pesticides
derives from the well-documented impacts
of organochlorine insecticides on populations
of predatory birds in Western Europe and
North America (Newton, 1998). This example
describes an historic mistake with several
messages relevant to rodenticides. First, it
describes a situation that arose some 50 years
ago with the widespread application of novel
agrochemicals in order to increase food pro-
duction, but whose lethal effects on certain
bird species were not understood (Newton,
1979). This mistake would be less likely to
occur today because direct toxicity in birds
would be detected at an early stage of re-
search and development (see Walker, 1993),
although compounds that vary markedly in
the severity of the effects they cause in differ-
ent species may still cause problems. The
veterinary use of diclofenac in South-east
Asia is one such example (Oaks et al., 2004).
Secondly, the consequences extended over
many years after the problem of toxicity
was rst recognized because of the effects of
persistent metabolites on reproduction in
birds (Ratcliffe, 1967; Cooke, 1973; Newton,
1998). Severe sublethal effects of persistent
chemicals would be less likely to occur
16 Environmental Impacts of Rodenticides
R.H. Smith1 and R.F. Shore2
1School of Applied Sciences, University of Huddersfield,
Huddersfield, UK; 2NERC Centre for Ecology and Hydrology,
Lancaster Environment Centre, Lancaster, UK
Environmental Impacts of Rodenticides 331
today because the environmental fate of
pesticides and their metabolites is a major
concern to both the agrochemical industry
and legislators, though the sublethal effects
of long-term exposure may be difcult to
detect (Smith, 1993; Walker, 1993). Thirdly,
the effects of organochlorine insecticides on
bird populations were documented in a way
that would not be possible for other groups
of organisms. Dead birds attracted attention
and population-level effects were detected
because birds are relatively conspicuous,
most people consider them attractive, and
there are better data on population trends
for birds (e.g. O’Connor and Shrubb, 1986)
than for other vertebrates. While there is no
doubt that rodenticides kill a range of non-
target species, it is the accidental killing of
birds that attracts most attention.
In this chapter, we will rst consider why
we expect there to be problems with non-
target effects as a result of rodent control,
and then consider how better scientic
understanding of the underlying ecology
might enable us to assess, and perhaps to re-
duce, risk. We shall refer to the legislative ap-
proach in the European Union (EU), which is
currently the Biocidal Products Directive or
BPD (Directive 98/8/EC, recently superseded
by Regulation (EU) No 528/2012). Our main
focus is on understanding how biology, and
in particular population biology (Smith,
1999), affects the environmental risks result-
ing from the use of rodenticides. In this re-
spect, our chapter will be very different in
emphasis from Brown (1994) in the rst edi-
tion of this book, because we now have much
more data available on the underlying bio-
logical processes. We shall summarize the con-
ventional, tiered approach to environmental
risk assessment (ERA) but those who would
like more detail should refer to the chapter
by Brown (1994) in the rst edition, most of
which is still relevant to understanding the
approach to ERA that is taken in the EU.
We shall concentrate on anticoagulant
rodenticides, as they are the most widely used
chemicals in rodent control. Anticoagulants
are of particular concern to regulatory author-
ities for several reasons:
• They are seen as candidate PBT com-
pounds (PBT = persistent, bioaccumu-
lative and toxic).
• In the EU, they may also be classied as
CMR compounds (CMR = carcinogenic,
mutagenic and reprotoxic).
• The simplest ERA statistics place them
well outside acceptable risk levels at
the rst tier of assessment (see later).
Why Might Rodent Control Have
Environmental Impacts?
Rodent control using chemicals can affect
non-target species in a variety of ways. Most
attention is given to the adverse effects on
predators, especially predatory birds, which
are exposed to secondary poisoning through
eating poisoned prey. Predator populations
generally take longer to recover than their
prey because they have a lower reproductive
rate as a consequence of longer time to ma-
turity, longer intervals between reproductive
attempts and produce fewer offspring at each
reproductive attempt. Animals that compete
with rats for food are more directly affected
by primary poisoning through gaining access
to bait. These are mammals and birds that
are more comparable to the target rodent
pests in terms of body size and life- history
characteristics that contribute to reproduct-
ive rate (e.g. sparrows, Passer domesticus,
wood mice Apodemus sylvaticus, and voles).
Although rodent control is usually very
targeted compared with, say, spraying with
insecticides, there are specic factors that
contribute to both hazard and exposure, and
these are discussed below
The physiology of rodents is very
similar tothat of non-target mammals
and birds
Rodent control targets a particular order of
mammals (the Rodentia), but all warm-
blooded vertebrates (birds and mammals)
have very similar physiologies. Indeed, other
vertebrates, such as reptiles, amphibians
and sh are not so different either, especially
when compared with invertebrate animals
and plants. This is why mice and rats are the
most commonly used laboratory animals in
the development of medical and veterinary
pharmaceutical products; mice and rats are
332 R.H. Smith and R.F. Shore
used as surrogates in studies to help predict
any adverse effects of medicines, pesticides
and other chemicals precisely because their
overall physiology and their biochemical
processes are similar to those of humans and
domesticated vertebrates. It is, therefore, dif-
cult to develop chemical rodenticides that
do not adversely affect other vertebrates that
are exposed to them. Thus, many or most
chemical rodenticides are hazardous to
many other vertebrates, and whether or not
they pose a risk depends on the level of ex-
posure of the animal concerned. Environ-
mental risk is a function of both chemical
hazard and environmental exposure (see
equation), and controlling exposure is usu-
ally the key to minimizing non- target effects.
Risk=f(hazard, exposure)
The most effective and commonly used
chemical rodenticides are the anticoagulants
(see Chapter 6), which were developed from
their original role in cardiovascular therapy
because their delayed action prevents the de-
velopment of conditioned taste aversion or
‘bait shyness’ (see Chapter 1; also O’Connor
(1948) and Baker et al., (2007)). Birds, mam-
mals and other vertebrates share the same
blood- clotting mechanism and so they are all
more or less vulnerable to the toxic effects of
anticoagulants. There can, however, be sub-
stantial variability in susceptibility between
species, especially to some of the earlier, rst-
generation anticoagulants (see Table 6.2 in
Chapter 6). Furthermore, most anticoagu-
lants developed since the mid-1960s should
be considered hazardous to all mammals
and birds.
Delayed action of anticoagulants leads
tooverdosing of targets and exposure
ofpredators
Ingestion of rodenticides would ideally be
self-limiting such that the target pest took in
only sufcient to cause death. The delayed
action of anticoagulants, though, inevitably
leads to overdosing. Harmful effects of anti-
coagulants only appear after the manufac-
ture of clotting factors is blocked and the
clotting factors circulating in the blood are
used up, so rats and mice continue to feed for
≥2 days after ingesting a lethal dose. During
this time, the intoxicated rodents continue
moving around and feeding, effectively be-
coming parcels of poison on four legs, and
leading to the exposure of their predators
to the anticoagulant circulating in their
bodies.
It is true that, when the toxic effects of
anticoagulants take effect, rodents will re-
duce activity and feeding, so that body
levels of anticoagulants will begin to de-
cline, mostly through excretion. Rodent be-
haviour may also change in a way that may
increase the exposure of their predators to
anticoagulant. While the target rodents may
in due course huddle in their nests and die
underground, a study reported by Cox and
Smith (1992) using time-lapse video record-
ing of common rats, Rattus norvegicus, in
an enclosure showed that intoxicated rats
observed lost their normal tendency for a
nocturnal pattern of activity (described in
more detail later in the section Adaptive be-
haviour and environmental impacts).
Intoxicated rodents that die of internal
bleeding without being taken by predators
may then be eaten by scavengers, and the
rodent carcasses will contain amounts of
anticoagulant greater than was necessary
to kill them. Brakes (2003) observed how
captive red kites, Milvus milvus, selectively
chose to eat the viscera of (non-poisoned)
rat carcasses and also had a selective prefer-
ence ranking for different parts of the viscera:
small intestines > liver > urinogenital
organs
This sort of selective feeding behaviour
may, unfortunately, also act to increase ex-
posure to anticoagulants if it occurs in the
wild. This is because the guts of poisoned
rodents may contain undigested or partially
digested rodenticide bait and, once ab-
sorbed across the gut, much of the anti-
coagulant rodenticide is bound to liver tis-
sue. Thus, the main feature of anticoagulants
that makes them so effective (delayed ac-
tion), unfortunately leads to increased ex-
posure of non-target predators and scaven-
gers by overdosing rats beyond lethal doses.
Environmental Impacts of Rodenticides 333
Contract pest-control practice may
have led to higher than necessary
exposure of non-targets
Controlling rodents with chemical rodenti-
cides according to best practice is expensive
and requires a level of expertise. Many farms
develop rodent problems in the rst instance
because the farmers cannot nd the time to
carry out basic hygiene and proong measures
(Chapter 5). It is equally difcult for them to
nd the time needed to carry out rodent con-
trol, and they commonly use professional pest-
control operators (PCOs) to do this for them.
The business model followed by most
PCOs was developed decades ago by the major
players in the service-contract industry and
has not changed much, even though the chem-
ical rodenticides that they use are both more
persistent and more toxic than when the
business model was developed. In essence,
PCOs usually establish service contracts that
include a certain number of routine visits at
regular intervals, with additional visits if
necessary, for which there will be an extra
charge. The business model was developed
when rst-generation anticoagulants requir-
ing surplus baiting regimes (see Chapter 6)
were the only effective option. Competition
between pest-control companies has always
been intense, with small prot margins for
each contract, and prot could only be in-
creased by increasing the number of contracts
serviced by the pest-control technicians. All
of this led to bait being made available per-
manently at bait points that were inspected
rather infrequently (perhaps only once a
month), inevitably increasing the exposure
of non-target small mammals and birds to
toxic bait.
Recognition of the increased potential for
harm associated with second-generation anti-
coagulants has led to changes in instructions
on rodenticide labels, which every user is
obliged to follow. Bait points must be in-
spected regularly and bait removed at the end
of a treatment. There are, however, economic
pressures to do only as much as is required
rather than what is ideal. For example, the
pressure on pest-control technicians to cover
as many sites as possible makes it unlikely
that they will spend a lot of time searching
for poisoned rat carcasses, notwithstanding
label instructions to remove carcasses. Like-
wise, surveys of rodenticide use by farmers
on agricultural holdings also indicate that
searches for poisoned carcasses are rarely
carried out (Tosh et al., 2011)
A benecial development since the rst
edition of this book is the acknowledgement
by the rodenticide industry and PCOs that
there may be a risk to non-target wildlife from
rodenticides. This has been the result of
extensive research and monitoring studies
that have shown that there is extensive ex-
posure in a wide range of non-target species
(Chapter 17). A key consequence has been
the establishment of ‘best practice’ guidelines.
An example of an organization and a voluntary
accreditation process that aims to reduce
environmental impacts is the UK Campaign
for Responsible Rodenticide Use (CRRU:
http://www.thinkwildlife.org/crru-code/),
and many PCOs are signed up to the CRRU
Code of Practice (see later).
Audit schemes lead to prophylactic
use of rodenticides
A further factor here is the introduction of
audit or ‘passport’ schemes run by, for ex-
ample, AIB International, the British Retail
Consortium and major supermarket chains.
There is great pressure on farmers and other
suppliers to produce food that is not only
nutritious and safe but also free from any
detectable contaminants, such as rat hairs,
other tissues or faecal material. The nding
of unwanted contaminants in produce tends
to generate bad publicity and audit schemes
help to protect business as well as the health of
the consumer. The main aim of these schemes,
which cover the food supply and distribution
industries, is to protect the safety/qual ity of th e
human food chain (see Chapter 11). The
standards set by these schemes are compar-
able across the world and help to achieve
the protection of human food supplies.
There is, however, an unintended negative
consequence for non-target wildlife. Farm-
ers risk having their produce rejected if they
fall foul of audit schemes, and audit schemes
inevitably encourage routine perimeter baiting
334 R.H. Smith and R.F. Shore
and other prophylactic measures that verge
on permanent baiting ‘just in case’.
Game bird rearing and rat control take
poison well away from buildings
Farmers in many countries have diversied
to nd new sources of income. In countries
such as the UK, rearing game birds such
as pheasants for recreational hunting has
become an important source of income for
large farms and estates. Full-time gamekeep-
ers are employed to ensure a good supply of
birds for the shooting season. The birds are
encouraged to remain on the estate by sup-
plying large quantities of grain in hoppers,
known as ‘pheasant feeders’, and these attract
rats and other small mammals. In addition,
strips of ‘cover crops’, such as maize, are
grown to provide shelter for the birds. These
provide both shelter and food for rats, and
can lead to the build-up of large numbers of
rats that are detrimental to the game-bird
enterprise. Gamekeepers are typically skilled
at trapping or shooting vermin, but they use
substantial amounts of rodenticide bait in the
farm environment well away from buildings
(McDonald and Harris, 2000), presumably
because of the scale of their rat problems in
places where they are feeding game birds.
How Environmental Risk Is Assessed
and Managed by the Registration
Process
New pesticides and biocides are registered
for use in agriculture only if they satisfy safety
criteria for those applying them, the con-
sumers eating the crops and the environment
into which they are released. Environmen-
tal risk assessment or ERA (Brown, 1994;
Brown et al., 1988) is used to predict the
environmental fate and potential effects of
pesticides/biocides on non-target species by
a stepwise or tiered process. In the EU, a tiered
or stepwise approach has been designed
to make efcient use of resources, because a
chemical passes on to the next (more expen-
sive) tier only if it fails the safety criteria at
a lower level. The requirements of the US
Environmental Protection Agency (EPA) are
similar to those of the EU regulatory author-
ities, and are also based on a stepwise ap-
proach. There are four tiers to this regulatory
stepwise approach. The higher numbered
tiers are the most realistic (i.e. the closest to
eld conditions). The lower tiers are the
most conservative in terms of assessing risk.
Tier 1
The rst tier is an initial review of data based
on physico-chemical properties of the chem-
ical and its toxicological prole, allowing
an assessment of hazard (toxicity) and a pre-
liminary assessment of risk (likelihood of
exposure to the hazard). Some pesticides may
have such low toxicity (T) to vertebrates
that it is clear that no risk will result from
their use (Urban and Cook, 1986), and no
further investigation is necessary, but this
will not be the case with rodenticides or many
other broad-spectrum biocides. The likely
fate of the compound in the environment
is then examined to predict the maximum
expected environmental concentration and
consequent potential environmental expos-
ure (E). A rst-tier prediction of risk is then
made by comparing toxicity to exposure in
some form. In broad terms, a compound is
judged acceptable at the rst tier if E < T,
i.e. if the levels of exposure are sufciently
low not to cause harm.
In practice, the measures used in Tier 1
assessment are the predicted environmental
concentration (PEC) and the predicted no-
effect concentration (PNEC), combined as the
PEC/PNEC ratio, which should not exceed
one for acceptability. Uncertainty factors are
also typically applied to either the PEC and/
or PNEC to make the assessment conserva-
tive; their use is intended to account for un-
certainties in the assessment, such as whether
compounds are equally toxic to all species.
Not surprisingly, PEC/PNEC ratios for modern
anticoagulants are massively greater than
one (actually of the order of 103–105 or higher)
for characterizing the risk of both primary
and secondary poisoning for mammals and
birds at Tier 1 (Luttik et al., 1999).
Environmental Impacts of Rodenticides 335
Tier 2
The second tier uses supplementary studies
to examine routes of exposure of non-target
species (Hardy, 1990) and toxicological ef-
fects. These renements aim to incorporate
factors such as, for example, estimated food
consumption, in order to rene estimates of
exposure. The resulting PEC/PNEC ratios? are
still unacceptable by very wide margins for
modern anticoagulants (Luttik et al., 1999).
Tier 3
The third tier is nearer to realistic patterns of
use in the eld and is based on an examination
of ‘worst case’ scenarios in semi-eld (e.g.
large enclosure) or eld trials. Simulated eld
trials look at primary and secondary exposure.
An open-eld trial, with intensive studies
of residues in the eld, effects on non-target
animals and quantitative assessments of risk
is used to judge whether the effects of the test
compound/s are acceptable or not.
Tier 4
The nal tier is post-registration monitoring,
which provides a means of identifying prob-
lems that might not have been picked up
at earlier stages. In the UK, the responsible
government department, Department for the
Environment, Food and Rural Affairs (De-
fra), operates a pesticides incidents scheme
whereby any wildlife deaths reported that
may be attributable to pesticides are investi-
gated under the Wildlife Incident Investiga-
tion Scheme, known by the acronym WIIS.
Because the WIIS focuses on mortality inci-
dents where there is a suspicion that pesticides
are involved, it does not provide unbiased
information on how widespread exposure may
be, but it does give an overview of whether
pesticides may or may not be causing mor-
talities. In addition, in the UK, there is also a
non-statutory monitoring scheme, the Preda-
tory Bird Monitoring Scheme (PBMS: http://
pbms.ceh.ac.uk/), which does quantify the
extent of exposure to rodenticides in wildlife
by monitoring residues in three sentinel
predatory bird species. More details of this
and other residue-monitoring schemes are
given in Chapter 17.
Under the Biocidal Products Directive
in the EU, rodenticides are evaluated and
approved for use (‘listed’) across the EU.
Nevertheless, individual member states can
also impose local restrictions or conditions
of use based on perceptions of national
requirements. Over most of the EU, most
second- generation anticoagulant rodenticides
(SGARs) are registered for use in and around
buildings, which would preclude open eld
use or use around pheasant feeders away from
buildings. In contrast, the UK has, until re-
cently, used the more restrictive classication
‘indoors only’ for brodifacoum, difethialone
and ocoumafen, while bromadiolone and
difenacoum may be used ‘indoors and out-
doors’, with no specic restriction that they
should only be used around (i.e. close to)
farm buildings, as in the rest of the EU. The
indoors/outdoors classication used in the
UK is intended to protect non-target wild-
life. An unintended consequence is that use
of bromadiolone and difenacoum in large
quantities away from buildings on game es-
tates has been allowed in the UK, and may
have been a major source of wildlife con-
tamination by SGARs (Buckle et al., 2011).
The stepwise approach to ERA seems to
be reasonable and generally to work well,
although it only takes partial account of the
complex ecology of free-living vertebrates.
It is based almost entirely on effects observed
in individuals, and only in Tier 3 is there an
attempt to look at effects in populations.
Even there, the approach is based on a rather
static view of populations with little appre-
ciation of spatial effects and population
structure.
Population Biology and Environmental
Effects of Rodenticides
Smith and Sibly (1985) describe population
biology as a triad comprising population
dynamics, adaptive behaviour and evolu-
tionary genetics. This view of population
336 R.H. Smith and R.F. Shore
biology recognizes that changes in numbers
(population dynamics) affect and are affected
both by adaptive, behavioural responses of
individuals and by longer term, evolutionary
changes in a population. It is well known
that exposure to sublethal levels of pesticides
may affect behaviour (e.g. Cox and Smith,
1990) and may also lead to the evolution
of inherited resistance. Changes in both
behaviour and rodenticide resistance are
relevant to rodenticide ERA.
Individuals and populations
A population can be broadly dened as those
animals or other organisms that are present
in one place at one time. This denition is
not always straightforward to apply because
‘place’ may be difcult to dene precisely,
especially for animals such as birds that are
relatively settled for parts of the year but
may move or migrate over large distances at
other times.
The four processes that affect population
size are birth, deaths, immigration and emi-
gration (see Fig. 5.1a in Chapter 5). When
we think about the environmental effects of
pesticides or biocides, we usually think in
terms of effects on the death rate. However,
experience of how organochlorine pesti-
cides affected wildlife populations in the
past tells us that birth rate may also be af-
fected by sublethal levels of toxicants. Im-
migration and emigration rates might be af-
fected directly if the migration behaviour of
animals is changed by exposure to pesti-
cides, but more important than this is the
way that migration rates, and indeed birth
and death rates, respond to changes in local
population density in real populations
(Newton, 1998). Both the birth rate and the
death rate might change with population
density in natural populations. In general
terms, birth rate declines, whereas death
rate increases with increasing population
size; this is known as density dependence and
may lead to the regulation of a population
around some average level, known as the car-
rying capacity. If exposure to pesticides
simply leads to an increase in death rate for
a short period of time, we generally expect
compensation in the population-growth
rate, which will tend to restore population
numbers to the original level. This means
that the population-level effect of pesticide
exposure will be short lived unless either
exposure is repeated, or persistent residues
continue to have an effect, or the relation-
ship between death rate and population size
is fundamentally altered and thereby alters
the carrying capacity.
Spatially structured populations
Most populations of animals are not distributed
uniformly across the landscape. In many
cases, patches of suitable habitat are separ-
ated from other such patches by unsuitable
habitat, and movement between patches is
restricted. This introduces the concept of the
‘metapopulation’, a modern view of popu-
lation dynamics that is relevant to both tar-
get and non-target species (Smith, 1995).
In the classical metapopulation struc-
ture, local populations are concentrated in
patches of resource and there is limited
movement between local populations, mostly
between adjacent patches (see Fig. 5.1b in
Chapter 5). This is often described as a meta-
population structure, and the dynamics of
the metapopulation depend on both the
dynamics within local populations and the
movement between those local populations.
The rate of movement between local popu-
lations may be density dependent in a simi-
lar way to birth and death rates, and a large
effect on a local population might be compen-
sated for more or less rapidly by migration
from a nearby local population. The impact
of a rodenticide on the dynamics of the meta-
population would then depend on the extent
to which one or more patches are affected,
and also the synchronization of those effects
(patches might correspond in broad terms to
elds or forests or farms). Asynchronous
treatment of patches would be expected to
have less effect on the overall dynamics of the
metapopulation than synchronized treatment
of all the local populations, as a consequence
of density dependence and movement be-
tween patches.
Environmental Impacts of Rodenticides 337
This last point is relevant both to increas-
ing the effectiveness of rodent control and
to reducing impacts on non-target species.
Synchronized control of rodent pests over a
large area is almost always going to be more
effective than the same amount of control
effort used asynchronously (at different times
in different places). Unfortunately, synchron-
ized rodent control will also have a more
widespread, deleterious effect on some non-
target species, both through the direct effects
of rodenticide and through indirect effects
on predators of having their prey removed
over a large area all at once.
Many species of particular conservation
interest are, unfortunately, relatively slow
breeders, and recovery time may, therefore,
be relatively long, certainly compared with
the pests against which pesticides are targeted.
An interesting example of the recovery of a
bird species in the context of a spatially
structured population is provided by Newton
in his study of the sparrowhawk, Accipiter
nisus. The woodlands of Northamptonshire,
in the UK, have gradually been recolonized
by the sparrowhawk since the use of organo-
chlorine pesticides ceased, and Wyllie and
Newton (1991) provide an example of the
way that density-dependent migration has
contributed to the metapopulation dynam-
ics of the process. In this case, though, not
all patches are equal: some woodlands do not
provide sufciently good habitat for births
to match deaths and support only ‘sink’
populations. A sink population (where deaths
exceed births) can only be maintained by
migration from more productive woodlands
(where births exceed deaths), known as
‘source’ populations. The population dy-
namics of this sort of structure are known as
source–sink dynamics.
One problem with a source–sink struc-
ture is that, on a landscape scale, it can ap-
pear that a population is thriving because
many or all patches are colonized. Wiping
out a source population, however, would
eventually lead to extinction of the sink
populations that are maintained by migra-
tion from the source. Thus, a population
with a source–sink structure is not resilient
to perturbations that affect the source popu-
lations and can go into general decline if the
few source populations are wiped out or
have their productivity reduced, perhaps as
a result of exposure to rodenticides.
Adaptive behaviour and environmental
impacts
Most of the behaviour of wild animals can
be described as adaptive in the sense that it
has evolved because it increases the tness
of individuals showing that behaviour
(Smith and Sibly, 1985). Pest rodents show
a variety of behaviours that reduce their
chances of being predated, for example noc-
turnal activity, thigmotactic behaviour and
neophobia (Chapter 1). Predators have a
range of behaviours that help to maximize
their rate of prey intake and hence their sur-
vival and reproduction.
The behaviour of both prey and preda-
tor can be modied by exposure to pesti-
cides or biocides (Hart, 1990), although
rather little is known about the effect of sub-
lethal doses of these on the behaviour of
predators. In general, prey species exposed
to poisoning show lethargy, apparent lack of
awareness of surroundings and unwilling-
ness to y or to (otherwise) move more than
necessary. Cox and Smith (1992) have
described such changes in rat behaviour in
the days between the consumption of anti-
coagulant poison and death. Following ex-
posure to anticoagulant rodenticides, rats
seemed to lose their thigmotactic response
and to use the edges of their environment
less and open areas more; indeed, within
3 days of consuming a lethal dose of anti-
coagulant, rats were observed standing still
in the open for up to 20 min at a time, and
the time spent in the open as opposed to
around the edges of the environment changed
from 16 to 79% within 4 days. Diel activity
rhythms also changed such that, within 48 h
of intoxication, rats were active for a greater
percentage of time in the light and less
in the dark, which is the reverse of normal
behaviour.
Changes in the behaviour of prey can
substantially affect the risk of exposure
of predators to secondary poisoning from
338 R.H. Smith and R.F. Shore
consuming intoxicated prey. This change in
behaviour could have effects on potential
predators, in that rats moving in the open
are more easily seen and their reduced
motion would make them easier to catch. It
is known that nocturnal hunters such as
owls are more successful at hunting when
their prey is foraging in the open rather than
under cover (Southern and Lowe, 1968;
Kotler et al., 1988). Behaviour that makes
prey stand out as ‘unusual’, such as the stag-
gering shown by the rats described by Cox
and Smith (1992), has also been shown to
lead to selective predation (Rudebeck, 1951;
Mech, 1970). The effect of the rodenticide
on rats in changing their normal, adaptive
behaviour would seem to make them more
vulnerable to predators. The prole of pred-
ators might also alter; change in diel activity
could expose rats to diurnal hunters such as
weasels (Sleeman, 1989), kestrels, buzzards
and foxes, while nocturnal predators such
as owls might become less exposed.
Rodenticide resistance increases exposure
ofpredators and scavengers
Resistance to anticoagulant rodenticides
appeared in the UK within 7–8 years of the
introduction of warfarin, and has been
reviewed by Smith and Greaves (1987) and
Greaves (1994), and in Chapter 9. The main
practical problem of resistance is that rats
become difcult to control (Cowan et al.,
1995), but there could also be effects of
resistance on the environmental impacts of
rodenticides.
The problem arises as follows. In areas
with high levels of resistance to the roden-
ticides available for outdoor use (docu-
mented over much of the southern half of
the UK: see Chapter 14 and Fig. 3 in
Buckle and Prescott, 2012), large quan-
tities of bait can be used as part of outdoor
treatments that are either wholly or par-
tially ineffective. Resistant rats can feed
on the bait with few or no ill effects. Each
resistant rat potentially provides a sub-
stantial parcel of poison to any predators
that might attack it.
How large an effect is this? In a repli-
cated eld study comparing body loads of
the anticoagulant coumatetralyl in rats be-
tween areas with resistance (Berkshire, UK)
and areas with no resistance (Leicestershire,
UK), a non-metabolizable, chemical bait
marker was used to quantify bait consump-
tion by individual rats (MacVicker, 1998;
Smith, 1999). In addition, whole carcass body
residues of coumatetralyl were measured. It
was found that rats from populations with
resistance consumed on average around ve
times more bait and that their body loads of
anticoagulant were about ve times higher
than those of rats from populations without
resistance. Most of the rats from the popula-
tions with resistance survived these very
high doses and had to be trapped in order for
the analyses to be carried out. In contrast,
Atterby et al. (2005) found a smaller diffe-
rence in body loads (less than two times) be-
tween resistant and susceptible rats, but
their laboratory study looked at equilibrium
body loads rather than at the higher, transi-
ent body loads in rats encountered by pred-
ators in eld conditions. Hence, resistance
in rats may substantially increase the expos-
ure of their predators to anticoagulants at
any time during the treatment of an infest-
ation. Add to this the fact that resistance
increases treatment time, perhaps inden-
itely, and we see that anticoagulant resistance
is likely to increase substantially the risk of
accidental poisoning of those predators that
feed on rats and inhabit areas where resist-
ance is prevalent.
Environmental impacts mediated
by competitors of rats
It is easy to forget that there are several spe-
cies of small mammals and birds that eat the
same sorts of food as rats and mice. If they
can access poisoned bait, these competitors
may be killed, or they may act as vehicles
that carry poison to predators that might not
feed on rats. In practice, it is all but impossible
to design a bait box that a rat can access but
a smaller rodent cannot. Thus, non-target
rodents that are common in the agricultural
Environmental Impacts of Rodenticides 339
environment such as the wood mouse, bank
vole, Myodes glareolus, or eld vole, Microtus
agrestis, may all feed at bait boxes placed
around farm buildings.
Brakes and Smith (2005) carried out
exposure studies of non-target small mam-
mals alongside routine rat control at ve UK
sites. The three non-target species mentioned
above all fed on poisoned bait, and 49% of
individuals in local populations were known,
by use of markers, to eat the bait. Local popu-
lations declined by 56% on average, compared
with reference populations (no rat control),
which increased by 9% over the same period.
After a 3 month follow-up period without
rat control, small mammal populations had
still declined by 51% compared with a 5%
decline in reference populations. This dem-
onstration of substantial consumption of
bait by non-target small mammals may help
to explain why predators such as the kestrel,
Falco tinnunculus, that do not commonly
feed on rats show such high levels of expos-
ure to anticoagulants (see Chapter 17).
Modelling environmental exposure
Cox and Smith (1990) and Smith et al. (1990)
proposed a compartment model of rodenti-
cide ecotoxicology in order to combine the
main components of the rodenticide system.
Smith et al. (1990) presented some quantita-
tive predictions based on a simulation model
that did not include population dynamics or
genetic changes, but did model physiology
and behaviour based on laboratory and eld
data. Their generic model of rodenticide
ecotoxicology is shown in Fig. 16.1.
Smith et al. (1990) modelled two com-
pounds with contrasting LD50 values (25 mg kg–1,
typical of rst generation, and 0.25 mg kg–1,
typical of SGARs). For the more toxic com-
pound, most rats died with residues higher
than would be expected if they had ingested
and fully retained an LD50 dose, because the
delayed action of anticoagulants meant that
they continued to feed for a period of time
after the ingestion of a lethal dose. The simu-
lation results for the more toxic compound
Target
rodent
Target
carcases
Ver tebrate
scavenger
Non-target
carcases
Non-target
bait feeder
Sink
Compartments
Transfers
Bait Predator
1
2
6
6
6
3
2
2
3
4
4
5
4
Fig. 16.1. Conceptual model of rodenticide ecotoxicology. There are six transfer processes that need to be
quantified to make the model predictive: (1) primary feeding on poison bait by the target rodents; (2)
mortality due to primary and secondary poisoning of target and non-target organisms; (3) feeding on
carcasses by vertebrate scavengers; (4) predation of target and non-target vertebrates; (5) primary feeding on
poison bait by non-target vertebrates; (6) transfer of poison to soil from carcasses. (After Smith, Cox and
Rampaud, 1990.)
340 R.H. Smith and R.F. Shore
gave residue values very close to the mean
carcass residue level of 3.2 mg kg–1 reported
by Dubock (1984) for saturation baiting with
an anticoagulant (brodifacoum) that has a
similar LD50. Residue levels were substan-
tially higher for the less toxic compound, but
represented lower risk because they were not
100× fold higher, which was the difference in
toxicity (LD50 value) between the two com-
pounds. In addition, with the less toxic com-
pound, most carcass residues were lower
than would be expected if an LD50 dose had
been ingested and retained. This was because
of the rapid daily elimination (assumed to
be 30%), which also outweighed any accu-
mulation through continued feeding.
When modelled residues were expressed
as rat acute oral LD50 equivalents, the more
toxic compound was seen to represent a
greater potential risk to scavengers because
of their exposure to residue levels in car-
casses (Smith, 1999). The results suggested,
however, that the less toxic compound might
represent a greater risk to predators because
rats would be wandering around carrying
substantial levels of the compound for much
longer, and exposure would, therefore, be
increased. The model was found to be fairly
insensitive to uncertainties in exact values
of parameters even though its representation
of elimination and accumulation was quite
crude. In particular, the model assumed a
simple, one-phase elimination model (sum-
marized by a single- gure half-life param-
eter), whereas we know that anticoagulants
are eliminated in a biphasic process, with
the rapid, initial elimination of circulating
compound, followed by slower elimination
from binding sites (Huckle et al., 1988).
Further exploration of the model sug-
gested that the concentration of active in-
gredient in bait for the more toxic com-
pounds might safely be reduced without a
substantial reduction in efcacy and, in
consequence, the residue levels in the car-
casses would also be reduced. This sort of
mechanistic effect model based on certain
features of the population biology of rats
could prove to be a useful aid in optimizing
the concentration of active ingredient for
compounds with different levels of toxicity
and in helping to reduce the exposure of
non-target animals to rodenticides.
Risk Mitigation
Alternative chemicals to anticoagulants
There are currently no compounds on the
market that are as effective in rodent control
as anticoagulants, but it is worth consider-
ing whether there are alternatives that might
be good enough, yet avoid the known ad-
verse effects of anticoagulants.
Calciferols were hailed as candidate al-
ternatives in the 1970s because they also
have a delayed action, although they are
less stable than anticoagulants in damp con-
ditions and so have been used mainly in dry
situations (especially against mice). There
is, though, experimental evidence that the
symptoms of calciferol poisoning (hypercal-
caemia) are not sufciently delayed to over-
come bait shyness completely in rats (Pres-
cott et al., 1992). Calciferols have the major
benet that there is no evidence of inherited
resistance to them and they could, there-
fore, be used as a second line of attack when
anticoagulant resistance is present. Eason
etal. (2000) investigated the risks to com-
panion animals of non-target and secondary
poisoning from using cholecalciferol and
concluded that there is a risk of secondary
poisoning, but that it is lower than with
anticoagulants. The registration of calcif-
erols has, however, lapsed in the EU where
companies have chosen not to provide the
supporting information required by the
registration authorities for inclusion in
Annex 1 of the BPD.
Alphachloralose is a narcotic used for
mouse control and has the potential to cause
adverse effects in non-target organisms. In the
EU, environmental exposure to this rodenti-
cide is limited because alphachloralose is
only registered for indoor use against rodents.
Bait placement and composition
The way that rodenticide bait is applied is
a key element in minimizing the risk of
accidental poisoning on non-target organisms.
In the UK, the CRRU has promoted best
practice through its Think Wildlife campaign.
The CRRU web site should be consulted for
Environmental Impacts of Rodenticides 341
more detail, but the elements of the CRRU
code of practice are as follows:
• Always have a planned approach.
• Always record quantity of bait and
where it is placed.
• Always use enough baiting points.
• Always collect and dispose of rodent
bodies.
• Never leave bait exposed to non-target
animals and birds.
• Never fail to inspect bait regularly.
• Never leave bait down at the end of a
treatment.
The composition of bait includes both
the active ingredient and the medium that is
used to make the bait attractive to the target
pest species. Ideally, the bait should be more
attractive to the target than to non-target spe-
cies. In practice, other mammals, as well as
birds, are likely to eat bait, whether it is
presented as grain, formulated pellets or
wax blocks, if it is accessible, and local popu-
lations of non-target mammals may be affected
even if good practice is observed (Brakes
and Smith, 2005).
Alternative and supplementary
control measures
Alternative and supplementary measures
can mitigate adverse effects if they reduce the
exposure of non-target animals to chemical
rodenticides; such measures do not need to
replace anticoagulants entirely in order to
achieve this. Some non-chemical alternatives
are discussed in Chapter 5.
In general, habitats that are structurally
complex are attractive to rats and other eld
rodents because they provide nest sites and
reduce visibility and exposure to predators.
Lambert et al. (2008) describe a replicated
study that used radio tracking and population
estimation before and after habitat manage-
ment around farm buildings to study the im-
pact of reducing habitat complexity on the
ranging behaviour and survival of rats. Remov-
ing harbourage up to eld margins reduced
rat survival and activity around the farm
buildings and appeared to be cost-effective,
although it was seen as a supplement to (rather
than a replacement for) chemical control that
ought to reduce the quantity of anticoagulant
used on a farm, i.e. part of the integrated ap-
proach promoted by Singleton et al. (1999).
In a further development, the Defra sup-
ported eld studies in Yorkshire and Leices-
tershire, UK, aimed at reducing the use of
anticoagulant rodenticides through ecologic-
ally based rodent management (Brown, 2007).
The impetus for this work was concern about
both humaneness (Mason and Littin, 2003)
and the non-target effects of anticoagulants.
There were two elements to the study, which
was based on the concept of managing the
rat population across the landscape:
1. coordinating rat control over a large area
(400 ha) rather than poisoning rats around
farmyards in an uncoordinated way; and
2. incorporating systematic trapping into
the control strategy to reduce the numbers of
rats moving into farm buildings during the
autumn.
One of the aims was to reduce the overall rat
population in each of the 400 ha coordin-
ated blocks compared with the uncoordin-
ated reference areas, and this was achieved
over a 2–3 year period. Along with this re-
duction, there was a concomitant reduction
in the amount of anticoagulant applied. To
date, only part of the study has been pub-
lished, in conference proceedings (Ethering-
ton et al., 2009), though most of the results
can be found in a PhD thesis (Brown, 2007).
It does seem that the adoption of an integrated
pest management (IPM) approach could sig-
nicantly reduce the amount of rodenticide
put out into the environment at the same
time as controlling pest damage effectively.
Discussion
Whatever the process used in pesticide regis-
tration in order to avoid adverse environ-
mental effects, it will always be necessary to
monitor wild populations for unforeseen
effects associated with pesticide use. In the
UK, we are particularly lucky to have a large
band of enthusiastic volunteers who moni-
tor bird populations. Population monitoring
of the sort coordinated by the British Trust
342 R.H. Smith and R.F. Shore
for Ornithology (BTO) provides estimates of
average population size of common bird spe-
cies in different years in the UK.
Population monitoring alone, however,
generates many more questions than an-
swers. For example, there has been a decline
in barn owl, Tito alba, populations since the
1950s in the UK and other countries (Shawyer,
1987). It is known (and it is not surprising)
that the residues of several anticoagulant ro-
denticides are found in barn owls in the UK
(Chapter 17); but there is no evidence of sig-
nicant adverse effects on barn owl popula-
tions of harmful levels of rodenticides in their
mammalian prey. In fact, barn owl decline
appeared to precede the introduction of
SGARs into Britain. The exposure of barn
owls, and other non-target species, to rodenti-
cides may cause some concern, but the ob-
served decline in barn owl numbers is more
likely to be an indirect consequence of the
earlier use of organochlorine pesticides and
of subsequent changes in the agricultural
management of grassland.
The argument for well-controlled repli-
cated eld experiments as an invaluable aid
to understanding environmental effects has
been made many times and in many places
(e.g. Brown, 1990, 1994; Cox and Smith, 1990;
Hart, 1990). The ecotoxicology model sum-
marized in Fig. 16.1 can guide the design of
eld experiments and could be generalized
to other pesticides, but it is possible that the
system may become much more complicated
when herbicides or insecticides are con-
sidered. The main value of this sort of quan-
titative approach to prediction is that it
focuses attention on missing information and
can help to direct effort towards collecting
the important missing data. In addition, the
modelling approach can allow the inclusion
of what are now very basic and widely
accepted ecological principles in order to
move ecotoxicology away from effects on
individuals and towards genuine prediction
of the ecological effects of toxins.
The results for modelling rodenticide
ecotoxicology that are presented here demon-
strate the practical value of this approach,
although it has to be admitted that this model
is clearly only a rst step that does not even
include density-dependent effects in rat popu-
lations (e.g. Smith and Greaves, 1987). Future
advances will link the ecotoxicology model-
ling approach to the sorts of models now rou-
tinely used in formulating pest-management
strategies, and will include the incorporation
of spatial as well as temporal features.
In this chapter, we have attempted to
broaden the approach taken in the rst edi-
tion of this book (Brown, 1994), by moving
beyond a focus on regulatory requirements to
try to establish a conceptual framework based
around biology and, more specically, popu-
lation ecology. The aim is to develop an
ERA for rodenticides that is more realistic
than, for example, Tier 1 use of PEC/PNEC
ratios, which have little relevance. This
is still a work in progress. For example, we
need to be able to predict more accurately
the true level of mortality in non-target spe-
cies of interest that is attributable in whole
or in part to rodenticides. We need better
linkage of environmental risk assessment to
population dynamics (including spatial dy-
namics) in the eld, and we expect that this
will involve developing population models
that are both realistic and predictive. We would
hope eventually to see this better under-
standing guiding technological advances in
delivery systems that would lead to reducing
exposure to rodenticides to target species
only. In the meantime, we support initiatives
such as CRRU’s Think Wildlife campaign
that (in the words of Brown, 1994) help ‘to
ensure that successful rodent control is ac-
companied by safety to wildlife and the
environment’.
References
Atterby, H., Kerins, G.M. and MacNicoll, A.D. (2005) Whole-carcass residues of the rodenticide difena-
coum in anticoagulant-resistant and -susceptible rat strains (Rattus norvegicus). Environmental
Toxicology and Chemistry 24, 318–323.
Environmental Impacts of Rodenticides 343
Baker, S.E., Singleton, G.R. and Smith, R.H. (2007). The nature of the beast: using biological processes
in vertebrate pest management. In: Macdonald, D.W. (ed.) Key Topics in Conservation Biology.
Blackwell, Oxford, pp. 173–185.
Brakes, C.R. (2003) The exposure of non-target wildlife to rodenticides, with special reference to the
red kite (Milvus milvus). PhD thesis, University of Leicester, Leicester, UK.
Brakes, C.R. and Smith, R.H. (2005) Exposure of non-target small mammals to rodenticides: short-
term effects, recovery and implications for secondary poisoning. Journal of Applied Ecology 42,
118–128.
Brown, R.A. (1990) Contrasting approaches to eld trial design. In: Somerville, L. and Walker, C.H.
(eds) Pesticide Effects on Terrestrial Wildlife. Taylor & Francis, London, pp. 189–206.
Brown, R.A. (1994) Assessing the environmental impact of rodenticides. In: Buckle, A.P. and Smith, R.H.
(eds) Rodent Pests and Their Control, 1st edn. CAB International, Wallingford, UK, pp. 363–380.
Brown, M. (2007) Rats in an agricultural landscape: population size, movement and control. PhD thesis,
University of Leicester, Leicester, UK.
Brown, R.A., Hardy, A.R., Greig-Smith, P.W. and Edwards, P.J. (1988) Assessing the impact of rodenti-
cides on the environment. EPPO Bulletin 18, 283–292.
Buckle, A.P. (1994). Rodent Control Methods: Chemical. In: Buckle, A.P. and Smith, R.H. (eds) Ro-
dent Pests and Their Control, 1st edn. CAB International, Wallingford, UK, pp. 127–160.
Buckle, A.P. and Prescott, C.R. (2012) The Current Status of Anticoagulant Resistance in Rats and Mice in
the UK. Report from the Rodenticide Resistance Action Group of the United Kingdom to the Health and
Safety Executive. Rodenticide Resistance Action Group, British Crop Protection Association, Derby,
UK. Available at: http://www.bpca.org.uk/assets/RRAG-ReportAnticoagulantResistanceintheUK.pdf
(accessed 11 July 2014).
Buckle, A.P., Walker, L. and Shore, R.F. (2011) Wildlife at risk: can we stop anticoagulant rodenticide
contamination? Pest No. 13 (Jan–Feb 2011), 9–11. Available at: http://www.pestmagazine.co.uk/_
attachments/Resources/391_S4.pdf (accessed 11 July 2014)
Cooke, A.S. (1973) Shell thinning in avian eggs by environmental pollutants. Environmental Pollution
4, 85–150.
Cowan, D., Dunsford, G., Gill, E., Jones, A., Kerins, G., Macnicoll, A. and Quy, R. (1995) The impact of
resistance on the use of 2nd-generation anticoagulants against rats on farms in southern England.
Pesticide Science 43, 83–93.
Cox, P.R. and Smith, R.H. (1990) Rodenticide ecotoxicology: assessing non target population effects.
Functional Ecology 4, 315–320.
Cox, P.[R.] and Smith, R.H. (1992) Rodenticide ecotoxicology: pre-lethal effects of anticoagulants on rat
behaviour. In: Borrecco, J.E. and Marsh, R.E. (eds) Proceedings of the Fifteenth Vertebrate Pest
Conference, Held March 3, 4 and 5, 1992, at the Hyatt Newporter, Newport Beach, California.
University of California, Davis, California, pp. 165–170.
Dubock, A.C. (1984) Pulsed baiting –a new technique for high potency, slow acting rodenticides. In:
Dubock, A.C. (ed Proceedings of a Conference on the Organisation and Practice of Vertebrate Pest
Control, 30 August–3 September 1982, Elvetham Hall, Hampshire, England, pp. 105–142.
Eason, C.T., Wickstrom, M., Henderson, R., Milne, L. and Arthur, D. (2000) Non-target and secondary
poisoning risks associated with cholecalciferol. New Zealand Plant Protection 53, 299–304.
Etherington, T.R., Lambert, M.S., Quy, R.J. and Tricker, J.R. (2009) Using circuit theory to identify important
animal movement areas: a case-study using data from an evaluation of brown rat (Rattus norvegicus)
control practices in the agricultural landscape. In: Catchpole, R. et al. (eds) Proceedings of the
Sixteenth Annual ialeUK Conference, Held at Edinburgh University, 1st–3rd September 2009. Inter-
national Association for Landscape Ecology, UK Region, pp. 17–24. Publication available from
http://iale.org.uk (accessed 11 July 2014).
Greaves, J.H. (1994) Resistance to anticoagulant rodenticides. In: Buckle, A.P. and Smith, R.H. (eds)
Rodent Pests and Their Control, 1st edn. CAB International, Wallingford, UK, pp. 197–217.
Hardy, A.R. (1990) Estimating exposure: the identication of species at risk and routes of exposure. In:
Somerville, L. and Walker, C.H. (eds) Pesticide Effects on Terrestrial Wildlife. Taylor & Francis,
London, pp. 81–97.
Hart, A.D.M. (1990). Behavioural effects in eld tests of pesticides. In: Somerville, L. and Walker, C.H.
(eds) Pesticide Effects on Terrestrial Wildlife. Taylor & Francis, London, pp. 165–180.
Huckle, K.R., Hutson, D.H. and Warburton, P.A. (1988) Elimination and accumulation of the rodenticide
ocoumafen in rats following repeated oral administration. Xenobiotica 18, 1465–1479.
344 R.H. Smith and R.F. Shore
Kotler, B.P., Brown, J.S., Smith, R.J. and Wirtz, W.O. II (1988) The effects of morphology and body size
on rates of owl predation on desert rodents. Oikos 53, 145–152.
Lambert, M.S., Quy, R.J., Smith, R.H. and Cowan, D.P. (2008) The effect of habitat management on
home-range size and survival of rural Norway rat populations. Journal of Applied Ecology 45,
1753–1761.
Luttik, R., Clook, M.A., Taylor, M.R. and Hart, A.D.M. (1999) Regulatory aspects of the ecotoxicological
risk assessment of rodenticides. In: Cowan, D.P. and Feare, C.J. (eds) Advances in Vertebrate Pest
Management. Filander Verlag, Fürth, Germany, pp. 369–385.
MacVicker, H.J. (1998) The ecotoxicology of rodenticide use on farms. PhD thesis, University of Leicester,
Leicester, UK.
Mason, G. and Littin, K.E. (2003) The humaneness of rodent pest control. Animal Welfare 12, 1–37.
McDonald, R.A. and Harris, S. (2000) The use of fumigants and anticoagulant rodenticides on game
estates in Great Britain. Mammal Review 30, 57–64.
Mech, L.D. (1970) The Wolf: The Ecology and Behaviour of an Endangered Species. Natural History
Press, Garden City, New York.
Newton, I. (1979) Population Ecology of Raptors. Poyser, Berkhamstead, UK.
Newton, I. (1998) Population Limitation in Birds. Academic Press, London.
Oaks, J.L., Gilbert, M., Virani, M.Z., Watson, R.T., Meteyer, C.U., Rideout, B.A., Shivaprasad, H.L.,
Ahmed, S., Chaudhry, M.J.I., Arshad, M., Mahmood, S., Ali, A. and Khan, A.A. (2004) Diclofenac
residues as the cause of vulture population decline in Pakistan. Nature 427, 630–633
O’Connor, J.A. (1948) The use of blood anticoagulants for rodent control. Research London 1, 334–336.
O’Connor, R.J. and Shrubb, M. (1986) Farming and Birds. Cambridge University Press, Cambridge,
UK.
Prescott, C.V., El-Amin, M. and Smith, R.H. (1992) Calciferols and bait shyness in the laboratory rat.
In: Borrecco, J.E. and Marsh, R.E. (eds) Proceedings of the Fifteenth Vertebrate Pest Conference,
Held March 3, 4 and 5, 1992, at the Hyatt Newporter, Newport Beach, California. University of
California, Davis, California, pp. 218–223.
Ratcliffe, D.A. (1967) Decrease in eggshell weight in certain birds of prey. Nature 215, 208–210.
Rudebeck, G. (1951) The choice of prey and modes of hunting of predatory birds with special reference
to their selective effect. Oikos 3, 200–231.
Shawyer, C.R. (1987) The Barn Owl in the British Isles – Its Past, Present and Future. The Hawk Trust,
London.
Singleton, G.R., Leirs, H., Hinds, L.A. and Zhang, Z. (1999) Ecologically-based management of rodent
pests – re-evaluating our approach to an old problem. In: Singleton, G.R., Hinds, L.A., Leirs, H. and
Zhang, Z. (eds) Ecologically-based Rodent Management. ACIAR Monograph 59, Australian Centre
for International Agricultural Research, Canberra, pp. 17–29.
Sleeman, P. (1989) Stoats and Weasels. Whittet Books, Stansted, UK.
Smith, R.H. (1993) Terrestrial mammals. Chapter 17 In: Calow, P. (ed.) Handbook of Ecotoxicology, vol. 1.
Blackwell Scientic Publications, Oxford, UK, pp. 339–352.
Smith, R.H. (1995) Rodents and birds as invaders of stored-grain ecosystems. In: Jayas, D.S. White,
N.D.G., Muir, W.E. and Sinha, R.N. (eds) International Symposium on Stored-Grain Ecosystems.
Marcel Dekker, New York, pp. 2898–3323.
Smith, R.H. (1999) Population biology and non-target effects of rodenticides: trying to put the eco into
ecotoxicology. In: Cowan, D.P. and Feare, C.J. (eds) Advances in Vertebrate Pest Management.
Filander Verlag, Fürth, Germany, pp. 331–346.
Smith, R.H. and Greaves, J.H. (1987) Resistance to anticoagulant rodenticides: the problem and its
management. In: Donahaye, E. and Navarro, S. (eds) Fourth International Working Conference on
Stored-Product Protection. Agricultural Research Organisation, Bet Dagan, Israel, pp. 302–315.
Smith, R.H. and Sibly, R.M. (1985) Behavioural ecology and population dynamics: towards a synthesis.
In: Sibly, R.M. and Smith, R.H. (eds) Behavioural Ecology: Ecological Consequences of Adaptive
Behaviour. Blackwell Scientic Publications, Oxford, UK, pp. 577–591.
Smith, R.H., Cox, P.R. and Rampaud, M. (1990) Rodenticide ecotoxicology: systems analysis and simulation.
In: Davis, L.R., Marsh, R.E. and Beadle, D.E. (eds) Proceedings of the Fourteenth Vertebrate Pest Con-
ference, Held at the Red Lion Inn, Sacramento, California, March 6, 7 and 8, 1990. University of
California, Davis, California, pp. 47–54.
Southern, H.N. and Lowe, V.P.W. (1968) The pattern of distribution of prey and predation in tawny owl
territories. Journal of Animal Ecology 37, 75–97.
Environmental Impacts of Rodenticides 345
Tosh, D.G., Shore, R.F., Jess, S., Withers, A., Bearhop, S., Montgomery, W.I. and McDonald, R.A. (2011)
User behaviour, best practice and the risks of non-target exposure associated with anticoagulant
rodenticide use. Journal of Environmental Management 92, 1503–1508.
Urban, D.J. and Cook, B.S. (1986) Ecological Risk Assessment. EPA-504/9-85-001, US Environmental
Protection Agency Ofce of Pesticide Programmes, Washington, DC.
Walker (1993) Birds. Chapter 16 In: Calow, P. (ed.) Handbook of Ecotoxicology Vol. 1. Blackwell Scientic
Publications, Oxford, UK, pp. 326–336.
Wyllie, I. and Newton, I. (1991) Demography of an increasing population of sparrowhawks. Journal of
Animal Ecology 60, 749–766.
