ANIMAL BEHAVIOUR, 2007, 74, 5e13

doi:10.1016/j.anbehav.2007.04.002

ESSAYS
Ethical regulation and animal science: why animal
behaviour is special
CHRIS BARNARD !

Animal Behaviour & Ecology Research Group, School of Biology, University of Nottingham, U.K.
(Received 23 February 2007; initial acceptance 7 April 2007;
nal acceptance 16 April 2007; published online 5 June 2007; MS. number: E-2)

Like other areas of animal science, the study of animal behaviour is becoming increasingly subject to ethical regulation and legislation. Sensible and well-informed regulation is to be welcomed both on compassionate grounds and because misuse of animals is likely to compromise the science itself. However, it is
evident that much of the impetus and direction in the regulation debate is coming from one particular
corner of animal science: that concerned with utilitarian and commercial interests taking place in controlled laboratory environments. This is sustained by an overemphasis on potential medical benets as
the perceived key selling point for animal science to a hostile public. Such a bias risks potentially unfortunate consequences for wider aspects of animal science, and is unlikely to serve the best welfare interests
of animals when viewed from their own biological perspective. I review some of the concerns that arise
from this, and suggest that the study of animal behaviour has a uniquely important role to play both
in the development of animal welfare science itself and in the public debate about the regulation of animal
science as a whole.
2007 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

Keywords: animal behaviour; animal experimentation; animal welfare; regulation; research ethics; the 3Rs

Like many other animal scientists, we who study animal

behaviour are becoming increasingly enmeshed in regulatory frameworks aimed at safeguarding the ethical probity of our experimental and observational procedures.
Institutional, professional and national and, increasingly,
international governmental bodies impose near forensic
scrutiny on applications for animal research and the
subsequent progress of work (e.g. Hudson & Bhogal
2004; Horgan & Gavinelli 2006; Laszlo & Csaba 2006;
Laurance 2006). In principle (and leaving aside the cost
of the unwelcome bureaucratic consequences), this is as
it should be. No scientist who is seriously interested in
the behaviour of animals wishes to see them used in a cavalier and thoughtless fashion; compassionate grounds
apart, such abuse would be likely to compromise the science itself (e.g. Russell & Burch 1959; Barnard & Hurst
1996; Bateson 2005). The Association for the Study of Animal Behaviour (ASAB) and its North American partner the
Animal Behavior Society (ABS) have long had their own,
regularly reviewed, ethical guidelines (http://www.elsevier.

Died 1 June 2007.

0003e 3472/07/$30.00/0

com/framework_products/promis_misc/ASAB2006.pdf)
through which they oversee work published in Animal
Behaviour and presented at ASAB and ABS conferences,
and both societies support fully the proper regulation of
research using animals. However, at least within the U.K.
and Europe, there are disquieting signs that the debate
about regulation and legislation is fast becoming parochialized within one sector of animal science: that to do
with broadly biomedical and commercial research. The
argument I want to make here, based largely on the situation within Europe, is that this is a potentially distorting
inuence that is likely to have unfortunate unintended
consequences for animal science in the wider sense that
mainstream biologists would recognize, and, moreover,
is unlikely to serve the welfare interests of animals when
considered from the animals point of view.
THE NATURE OF ETHICAL CONCERNS ABOUT
ANIMAL SCIENCE
All animal science brings ethical concerns in its wake
if it has an impact on the animals concerned. However,
it is useful to make a distinction between two

5
2007 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

ANIMAL BEHAVIOUR, 74, 1

fundamentally different reasons for using animals. The

rst, which encompasses various forms of utilitarian
research (e.g. Bateson 1986, 2005; Singer 1990; Reiss
1993), exploits animals as resources, and includes studies
for medical, consumer commercial, pharmaceutical, toxicological, military and other purposes which confer
health, social/political and/or nancial benets on human society at large. This is overtly self-interested exploitation which we justify ethically in terms of the
magnitude of resulting societal benets. Ethical concerns
are given added political impetus by the pressure of public opinion (especially from the extremist animal rights
wing), which is perhaps felt more acutely by the exploitative biomedical/commercial sector than by others
within animal science (Berg & Hammarstrom 2006; Signal & Taylor 2006).
The second category, which I shall refer to as curiositydriven research, uses animals to satisfy an intellectual
desire to understand the world around us (e.g. Reiss 1993).
It is thus no less self-interested than utilitarian research
(Cuthill 2007), but simply measures the worth of studies
in a different currency. Curiosity-driven research also
sees good welfare as a prerequisite for good science and
is subject to a measure of public pressure. The intrinsic
self-interest of utilitarian and curiosity-driven research
provides shaky ground on which to build a moral case
for exploiting animals, and will never cut much ice with
hardcore abolitionists because no amount of cost/benet
relativism will change the extreme moral view that we
have no right to exploit other animals for our advantage.
However, there is another argument that provides a rmer
moral case for research on animals, which I refer to as
a moral need to know.

The Case for a Moral Need to Know

That we are having a serious impact on our planetary
cohabitants, whether ecologically or through managing
them for food or the whims of the commodity market, is
hardly in doubt (e.g. Clark et al. 2006; Kim & Byrne 2006;
Mattson & Angermeier 2007). Much of that impact,
whether deliberate or incidental, arises from indifference
in the face of nancial or wider societal benets, but it is
exacerbated by our as yet shadowy grasp of what can be
done from the organisms point of view to alleviate it
(Barnard & Hurst 1996). As witting agents of the problem,
and the only ones capable of planning and effecting a response to it, we can be argued to have a moral obligation
to respond in a properly informed fashion. (While Cuthill
2007 argues that even studies designed to benet animals
themselves can be construed in terms of a self-satisfying
altruistic feel good factor, and so are also essentially
self-serving, this is a criticism that can be levelled at any
course of action based on moral principles and is thus
an argument against ethical scruples generally rather
than moral attitudes to animal science per se.) A sophisticated understanding of how other species function, and
thus of their welfare in the rounded sense of being able
to function on their own terms d what I shall call evolutionarily salient welfare (see below) d is thus essential if

the problem is not simply to be compounded by wellintentioned, but misguided, shots in the dark. Key elds
of animal biology here are behavioural and population
ecology, conservation biology, veterinary biology and animal welfare itself, but all aspects of whole-organism biology are likely to play an important role. Understanding in
these elds cannot be gained simply by staring passively at
animals in their natural state, but requires manipulative
and sometimes invasive experimentation. So what do I
mean by evolutionarily salient welfare?

WHAT IS WELFARE AND WHAT SHOULD IT

MEAN IN PRACTICE?
There is little consensus on either what welfare really
means in other species or how it should be measured (e.g.
Mendl 1991; Mason & Mendl 1993; Fraser 1995; Dawkins
2006). As has been noted by several authors (e.g. Moberg
1993; Barnard & Hurst 1996; Clark et al. 1997; Skutch
1997; Timberlake 1997; Serpell 2003), traditional concepts
and measures of welfare have relied on a suite of essentially
anthropomorphic criteria, which owe more to how we
think we might feel in the animals circumstances than to
a deep understanding of how the animal is actually likely
to feel. These broadly emphasize comfort (the extent to
which the animal is maintained with adequate space and
basic requirements for subsistence; e.g. Wolfensohn &
Lloyd 1994), health (the extent to which it is maintained
hygienically and free from infection, injury, stress and behavioural abnormality; e.g. Hurnik 1988; Broom 1991;
Burn et al. 2006; Dawkins 2006), normality of opportunity
(maintenance in circumstances that allow a perceived natural range of behaviours and provide a perceived natural social, sexual and reproductive environment; e.g. Thorpe
1965; Martin 1975; Bracke & Hopster 2006), or philosophical stances on ethics and animal rights (e.g. Rollin 1981,
1989, 1993; Midgley 1983; Regan 1984; Berg & Hammarstrom 2006), including arguments for respecting the
species-specic dignity of other animals (e.g. Rollin
1998; Wickins-Drazilova 2006). Such criteria form the basis
of the so-called ve freedoms (freedom from (1) hunger
and thirst, (2) discomfort, (3) pain, injury or disease, (4)
fear and distress, and freedom to (5) express normal behaviour) proposed by the U.K. Farm Animal Welfare Council
(Harrison 1988), and inform the ubiquitous three Rs (Replacement, Renement, Reduction) originally set out by
Russell & Burch (1959) (see also Smyth 1978) and now
the guiding credo of welfare regulation in the U.K. and
Europe. Welfare concerns arise when infringements of
these criteria are deemed sufciently serious to risk the
animal suffering, where suffering is a putative negative
subjective state that is rarely dened, either in terms of
what it comprises or of where on a presumably sliding scale
of negativity it lies. I shall return to this problem later.
Arguments in favour of anthropomorphic criteria include our obvious evolutionary continuity with other
species, with its potential for homologous psychological
responses to circumstance, at least in moderately close
relatives, and the fact that our own subjective experiences
provide the only model we have of suffering and so,

ESSAY

perforce, act as a best of a bad job comparator. Arguments

against hinge on the point that evolution is as much
about adaptive specialization as it is about continuity
(McFarland 1989); thus, while there may be similarities
between species at a general level that may have implications for welfare (homology in central nervous and sensory system structure, physiology and cognitive processes,
for example), there may be adaptive differences of far
greater signicance, especially, as we shall see, in life history priorities and the importance attached to individual
well-being (Barnard & Hurst 1996). Indeed, the very concept of welfare could be regarded as an anthropomorphic
conceit imposed on other species by one with a long life
span and long period of parental care in which attaching
a high priority to future survival (and thus maintaining
well-being) yields reproductive dividends. The conceit is
exacerbated by the fact that the comfort and health criteria owing from it arguably do not even serve our own
species well in terms of individual welfare (see Barnard
& Hurst 1996). That species, and individuals within species, are likely to differ in the priority attached to future
survival is an axiom of life history strategy theory, and
should ring alarm bells in the context of the above. So
how can we accommodate such fundamental differences
within a concept of welfare?

Adaptive Cost Gauging, Decision-making

Rules of Thumb and Suffering
The problem with suffering is that it is a private subjective state inaccessible to an external observer. Allowances for it therefore have to be made on some kind of
benet of the doubt basis, but on what should this
benet of the doubt be based if not the anthropomorphic criteria referred to above or other comparisons with
humans, such as those involving apparent cognitive processes (e.g. Mendl & Paul 2004; Paul et al. 2005)? Barnard
& Hurst (1996) have argued that it should be founded on
a thorough understanding of the animals decisionmaking rules of thumb in the context of its adaptive life
history strategy and thus trade-off between survival and
reproduction. Their point is that positive and negative
subjective states can be viewed as proximate mechanisms
gauging the canonical (reproductive) cost of the animals
current state. Thus pleasant sensations, such as a nice taste
or orgasm, gauge the likelihood that the animals actions
(eating a food item or copulating) will reap reproductive
benets, while negative ones, such as increasing pain or
fatigue, gauge the likelihood that they will inict reproductive costs (e.g. from overzealously defending a territory). Barnard & Hurst (1996) argued that such adaptive
cost gauging, where it involves negative subjective states,
should not merit the label suffering because the animal is
simply trading off activity costs against their reproductive
benet. As long as it is free to make the trade-off on its
own adaptive terms (so as to maximize lifetime reproductive success), the animal is not under any imposition or in
a situation into which it has not willingly entered. Thus,
any clinical sequelae, such as raised stress hormone
levels, reduced immunocompetence or weight loss, will

not be indicators of likely suffering but of what Barnard

& Hurst (1996) refer to as adaptive self-expenditure.
This leads to a very different view of welfare from the
traditional one in which homeostatic coping mechanisms
are designed to maintain the well-being of the individual,
and suffering is invoked when the animal fails to cope
(Broom 1986; Broom & Johnson 1993). In the adaptive
self-expenditure view, welfare is equated, not with individual well-being (the fallacy of individual preservationism,
as Barnard & Hurst (1996) put it), but with the animals
ability to spend itself in the manner dictated by its adaptive life history strategy. If suffering is possible at all, and
whatever the form it may take in different species, it is reserved as a subjective state of emergency experienced by
the animal when it is forced to act outside the adaptive
bounds of its life history strategy. Since any such suffering remains privately invisible, however, the most reasonable basis for recognizing when it is likely (and thus giving
the benet of the doubt) is by understanding when the
animal is constrained to behave beyond the competence
of its decision-making rules of thumb, since these have
evolved to serve its adaptive needs within the framework
of its life history strategy and environment of evolutionary adaptation (EEA) (Bowlby 1969; Mace 1995; Barnard
& Hurst 1996). Such violations might arise, for example,
when the animal is put into a novel or impoverished environment that is a poor analogue of its natural one,
and might include the squashing of normally resilient behaviours within the animals time budget, normally adaptive responses to a situation having an inappropriate
outcome, decision-making paralysis (limbo) caused by
a lack of opportunity to perform a next-in-priority behaviour, and other infringements of the animals ability to behave adaptively (McFarland 1989; Barnard & Hurst 1996).
Of particular importance here are violations of the animals perceptual rules of thumb (how it perceives and
models the world around it for its own adaptive ends),
which may lead to vastly different perceptions of objects
and events from our own (e.g. Ewert 1980; Milligan
et al. 1993; Dawkins 1998; Barnard 2004; Evans et al.
2006), and conditionality rules governing the modulating
inuence of developmental experience on potential subjective states and decision making (e.g. Burman & Mendl
2003; Paul et al. 2005). It is where such rules are infringed,
then, that we might reasonably look for clinical corroboration of suffering and be in a position to distinguish it
from similar changes that simply reect adaptive cost
gauging. Rooting welfare concerns in the context of the
animals adaptive function like this is the basis of what I
refer to as evolutionarily salient welfare. Two brief examples, both involving laboratory rodents, illustrate the
point.

Immunity trade-offs in laboratory mice

The rst concerns the regulation of testosterone secretion in relation to immunocompetence in male laboratory
mice, Mus musculus, and highlights the importance of individual differences in adaptive self-expenditure.
Males put together in novel groups usually polarize
into two easily identiable social classes on the basis of
aggressive and submissive behaviour (e.g. Mackintosh

ANIMAL BEHAVIOUR, 74, 1

1981; Gosling et al. 2000): dominant males that initiate

signicantly and disproportionately more aggressive behaviour than their companions, and subordinates that
show much less aggression and more submissive behaviour and do not differ signicantly in this regard among
themselves (e.g. Barnard et al. 1996a). By analogy with
their commensal wild counterparts (e.g. Hurst 1990,
1993; Barnard et al. 1991), dominants and subordinates
can be seen as representing different life history strategies
centred on competitive ability (Barnard & Behnke 2001,
2006). Competitive dominant males can command access
to limited resources, such as food and nesting sites, and invest heavily in the reproductive opportunities this brings.
Less competitive males cannot compete for resources so
have to make do with sneaking opportunistic matings as
and when they can.
A prediction that follows from this is that dominant
males will be more likely to trade-off future survival for
short-term reproductive gain than subordinates, which
would do better to safeguard future survival and maximize
the likelihood of chance matings. One way of regulating
the trade-off might be to link the secretion of the
potentially immunodepressive sex hormone testosterone
(e.g. Grossman 1985; Folstad & Karter 1992) to current immunocompetence. Barnard et al. (1996a) found that subordinate males did exactly this by regulating the
secretion of testosterone over the period of grouping in relation to their current circulating antibody (total IgG) titre. As a result, their resistance to a subsequent infection
of the blood protozoan Babesia microti was unaffected by
testosterone. Dominants, in contrast, did not regulate their
secretion of testosterone in relation to antibody levels,
with the result that testosterone reduced their subsequent
resistance to infection. Thus, within social groups, individual males responded in apparently different adaptive
ways to the reproductive implications of their social environment. Further evidence that this might reect an adaptive reproductive trade-off comes from a later study of
experimentally immunodepressed males (Barnard et al.
1997a, b). Males of both rank categories that had had their
cellular immunity temporarily depressed by antithymocyte serum downregulated both testosterone and behaviours such as aggression that were generally associated
with reduced immunocompetence, but performed more
of those behaviours, such as sleep, that were associated
with enhanced immune function (Barnard et al. 1997a).
The really telling point, however, was that the regulation
of both testosterone and behaviours affecting immunity
in immunodepressed males was abolished when female
odours were present in the environment (Barnard et al.
1997b).
Taken together, therefore, these results suggest that
males bring costly physiological mechanisms into action
only if it is reproductively worthwhile to do so. Thus,
dominants are more prepared than subordinates to risk
their future survival by secreting testosterone indiscriminately because they are in a better position to secure
matings in the short term. However, both classes of male
are prepared to risk it if current reproductive opportunity
seems high enough (there is a lot of female odour about).
These are important distinctions if one is to attempt to

impose welfare requirements on these animals, because

they demand different criteria if they are to recognize
the respective self-expenditure choices of different
individuals.
The studies above also offer a cautionary tale with
respect to welfare initiatives. Environmental enrichment
is an increasing enthusiasm in the world of welfare
regulation (e.g. Markowitz 1982; Chamove 1989; Patterson-Kane 2004; Buchanan-Smith et al. 2005) and stems
from the laudable aim of providing heterogeneity and
naturalistic diversity in otherwise monotonous captive
environments. Enrichment, however, is a loaded subjective term, designed to equate increased complexity with
enhanced living conditions. While this equation appears
rto be supported by at least some lines of evidence (Wu
bel 2001), just like the concept of welfare itself, enrichment requires contextualizing in relation to the animals
adaptive rules of thumb and EEA and cannot simply be
assumed to equate to increased complexity. A case in
point comes from some of the subtle conicts it can generate. Barnard et al. (1996b), for example, studied the
effects of adding nestboxes and shelving (henrichment)
to the cages of groups of male mice on the testosterone/
immunity trade-off discussed above. They found that
mice in enriched cages were signicantly more
aggressive, something that has been found by other
authors (e.g. McGregor & Ayling 1990; Haemisch &
Gartner 1994), and less resistant to a later infection of
B. microti than mice from nonenriched cages. Barnard
et al. (1996b) also found negative partial regression relationships between aggression received and both circulating antibody titre (total IgG) and resistance to B. microti.
Thus, the enriched cages appeared to provide a more hostile environment, probably because the discrete physical
features of the furnished cages provided a focus for aggressive defence by dominant males, an adaptive feature
of their reproductive behaviour (Hurst 1987a, b). Did
mice care about this in the sense of attempting to offset
the apparently deleterious effects? Evidence from the regulation of testosterone secretion suggests they did. In the
enriched cages, and in contrast to Barnard et al.s (1996a)
results earlier, both dominant and subordinate males
modulated testosterone secretion in relation to total
IgG. As a result, neither category of male showed a testosterone-dependent reduction in resistance to B. microti. In
the less hostile nonenriched cages, testosterone was decoupled from total IgG and mice did suffer a testosterone-dependent reduction in resistance (Barnard et al.
1996b). So, in the enriched cages, mice appeared to
take precautionary measures to safeguard immunocompetence in the face of apparently environmentally induced immune depression, whereas they were prepared
to tolerate an impact of testosterone in the more benign
nonenriched cages. Interestingly, time spent in the nestboxes and on the shelves correlated positively with total
IgG concentration and resistance, so enrichment did
appear to offer benets in terms of well-being, but only
in the context of an overall drop for the environment
as a whole. Once again, therefore, effects of conditions
imposed on animals need to be evaluated in the context
of different individual strategies of self-expenditure.

ESSAY

Social housing in laboratory rats

The second example comes from a study of housing
conditions in laboratory rats, Rattus norvegicus (Hurst et al.
1997, 1998). Although rats are social animals, they are often housed singly in laboratories to prevent problems with
aggression and other sources of social stress or standardize
social experience for experimental purposes (e.g. Brain &
Benton 1983; Mormede et al. 1990). That this works in
terms of classical clinical yardsticks of stress is suggested
by the fact that singly housed rats have reduced circulating levels of corticosterone and less early stage organ pathology than socially housed counterparts (Hurst et al.
1997). But which kind of environment would rats choose
if given the option? By housing singletons and groups either side of partitions allowing different degrees of detectability across them (ranging from solid metal sheets to
open mesh), Hurst et al. (1997, 1998) were able to show
that singletons spent signicantly more time at the partition, attempting to interact with grouped animals on the
other side, than vice versa, and that time at the partition
increased with the opportunity for interaction across it.
Similar preferences for groups have been obtained in other
choice experiments with rats (e.g. Patterson-Kane et al.
2002, 2004). Thus rats appeared to seek a social environment in which clinical indicators suggest their welfare,
in the traditional sense of individual well-being, will be
compromised, but an environment, of course, in which
their life history priorities have evolved and their reproductive opportunities lie (Barnett 1975; Lore & Flannelly
1977). The rats apparent choice thus makes functional
sense in terms of adaptive self-expenditure.

PAROCHIALIZING THE REGULATION DEBATE:

IMPLICATIONS FOR SCIENCE AND WELFARE
While evolutionarily salient welfare should inform animal
welfare policy in any circumstances, it self-evidently does
not, and one reason for that, I suggest, is because welfare
policy and regulation within animal science is being
driven largely by the laboratory utilitarian interests referred to earlier. It is not difcult to see why this should be.
Medical benets for humans are seen by many scientists as
the strongest selling point for animal experimentation in
the public domain; a tack that is also, not surprisingly,
taken by the powerful pharmaceutical industry. This cuts
a lot of ice with legislators and politicians and is the
reason for the explicitly biomedical line taken by the
European Coalition for Biomedical Research (http://
www.ecbr.eu) recently established to coordinate European
scientic interest responses to the draft revision of EU
Directive 86/609 on animal experimentation, a Directive
that affects Europe-wide legislation on scientic work
with animals. While the Directive applies to all animal science, lobbyists see medical benets from animal work as
the key to getting the ear of European politicians, and
thus couch their responses to the draft in terms of the laboratory science that underpins it. Of the 34 professional
societies and associations so far making up the Coalition
28 are wholly or partly concerned with biomedical research, so mainstream organismal biology is vastly

underrepresented in the process. Tellingly, the recent survey of animal scientists and special issue articles on animal experimentation by Nature on 14 December 2006
(Vol. 444, No. 7121) also reects almost exclusively the interests and procedures of laboratory biomedical science.
This bias is becoming institutionalized. For example, in
its mission statement, the recently established U.K.
National Centre for the Three Rs (NC3Rs) (http://www.
nc3rs.org.uk) declares that The NC3Rs is dedicated to the
3Rs - replacing, rening and reducing the use of animals
in research and testing licensed under the Animals (Scientic Procedures) Act 1986, that Optimal laboratory animal
welfare is critical for scientic, legal and ethical reasons,
and that its aims will be achieved by Supporting the UK
scientic communitys commitment to best practice in all
aspects of laboratory animal science and welfare. It also
states that Replacement is the ultimate aim for the Centre.
The inception of the NC3Rs and its general raison detre
are, of course, to be welcomed. One can also understand
the inherent bias towards laboratory animal science given
the organizations roots in the U.K.s Medical Research
Council. However, the overtly limited scope of its mission
leads to some serious concerns in terms of applying the
three Rs across animal science as a whole, only one aspect
of which is appropriate to the conditions and aspirations
of the NC3Rs statement. Confusion arises immediately,
for example, because U.K. Home Ofce licensing extends
beyond the boundaries of laboratory/utilitarian research,
and because ethical concerns, to which the three Rs apply,
extend beyond procedures coming under Home Ofce
scrutiny (see, for example, the ASAB/ABS ethical guidelines
referred to earlier). Thus casting its mission in terms of laboratory procedures falling within the purview of Home
Ofce licensing is doubly limiting.
There is also an issue with Replacement as the organizations ultimate aim. This is partly because Replacement
cannot be a logical objective where animals themselves
are the object of study (as opposed to exploitable resources; see also Grifn & Gauthier 2004), which is the
case for most mainstream animal biologists (the argument
that Replacement refers to procedures rather than animals
[Russell & Burch 1959] makes no difference here, since the
aim in either case is to remove the need to experiment on
animals), but also because Replacement requires a degree
of condence in the current state of knowledge about
a system that may not be warranted. The animal biology
literature is replete with examples of models and conclusions, long held to be the last word on a particular problem, being overturned and radically revised some years
later when new ideas or discoveries have prompted review
(the classic textbook model of the neural control of the
escape swimming response in the sea slug Tritonia [see
e.g. Willows & Hoyle 1969; Katz & Frost 1995] is a good
example). An overenthusiastic pursuit of Replacement in
response to ckle, short-term pressures of public and political opinion thus carries a serious risk of premature bridge
burning. While this may not be much of an issue if, say,
cells or tissue are being cultured simply to harvest an enzyme or hormone, it is very likely to be one if what is at
stake is the holistic functioning of a system in relation
to, say, drug efcacy. This leads to a more general

10

ANIMAL BEHAVIOUR, 74, 1

consideration of what the three Rs should mean when it

comes to animal science as a whole.

What Should the Three Rs Mean?

Although Russell & Burchs (1959) three Rs are the chief
guiding principle in welfare regulation, there is a considerable lack of clarity and consensus over what they should
mean in practice, even within relatively standardized
laboratory environments (Buchanan-Smith et al. 2005;
Schuppli & Fraser 2005; Vorstenbosch 2005). The reasons
for this are several, and often stem from different opinions
or assumptions as to what effects and actions fall within
the denition of a particular R (Schuppli et al. 2004;
Buchanan-Smith et al. 2005), or even as to how the particular R should be dened in the rst place (Buchanan-Smith
et al. 2005). The three Rs can also conict with one another
in terms of ethical objectives, leading to dilemmas in regulatory decision making (de Boo et al. 2005). From the point
of view of animal science as a whole, however, it is what the
three Rs should mean in the broad tranches of animal biology that lie beyond the clinical/commercial laboratory
setting that gives most cause for thought.
As Grifn & Gauthier (2004) have pointed out in their
excellent consideration of the three Rs in wildlife research,
criteria for the respective Rs in curiosity-driven animal biology may differ fundamentally from those in a highly
controlled laboratory environment. Replacement per se
is usually not an option (see above), and criteria for Reduction and Renement may depart substantially from the
classical experimental design and statistical protocols of
laboratory work (Festing 2003, 2004a; Grifn & Gauthier
2004; see also Schuppli & Fraser 2005). This may be especially so in the case of eld studies, where what Albrecht
(1999) has referred to as an ecocentric focus (concentrating
on the ecological integrity of the system) often replaces the
individualist one of traditional welfare considerations
(Grifn & Gauthier 2004). Albrechts distinction has resonance with that between adaptive self-expenditure and individual well-being, in that the adaptive self-expenditure
view also argues for departure from traditional yardsticks
for the legislative quartet of pain, suffering, distress and
lasting harm on the basis that these may not reect impositions on the animal as much as adaptive choice (see
above). This has important implications for what one
might want to regard as Renement in a welfare context,
where arguably the emphasis should be on tailoring environments and procedures to harmonize with the animals
adaptive rules of thumb rather than on maintaining individual well-being. Of particular importance is that Renement should embrace individual differences in adaptive
rules of thumb and recognize diversity in chosen strategies of self-expenditure within any given context (e.g.
Clutton-Brock et al. 1982; J.L. Hurst et al. 1996, 1999;
Barnard & Behnke 2001; see above), something that is
also likely to have consequences for decisions about sample size and approaches to Reduction. Arguably, therefore,
the most urgent requirement under Renement is for fundamental research on species- and individual-specic
behaviour.

The Biomedical/Laboratory Science Trap

Arguments based on welfare as adaptive self-expenditure rather than well-being highlight the hazards inherent
in the current over-focus on laboratory utilitarian science.
Internalizing the loop between the development of regulatory policy and the practices it is intended to regulate
within the highly controlled environments of laboratory
utilitarian research risks unintended, counterproductive
consequences when policy and perceived good practice
are extended, as they already are, into the much less
controlled (and often less controllable) environments of
open-ended curiosity-driven studies.
Much good practice in laboratory settings is characterized by a desire for standardization (e.g. in stocking
densities, laboratory strains, environmental enrichment,
temperature and humidity, air exchange rate; e.g. Festing
2001, 2004b; Marashi et al. 2004; Patterson-Kane 2004;
Brandstetter et al. 2005). Not only is this inimical to the
environmental requirements of much evolutionary and
ecological curiosity-driven research, where natural variation in physical, social and genetic environments is an essential backdrop to the biology under consideration, but it
also constrains the external validity, and thus general explanatory power, of the very laboratory protocols that em rbel 2000, 2002; Cuthill 2007). Furthermore,
brace it (Wu
the criteria for standardization are largely based on either
the anthropomorphic welfare considerations discussed
earlier, or on practical (husbandry or experimental) or analytical (statistical) convenience, often with uncritical
generalization across contexts (e.g. environmental enrichment). Thus standardization is likely to conict with
the principles of evolutionarily salient welfare, and itself
be a source of welfare concern.
In short, therefore, the domination of biomedical/
commercial interests and a laboratory animal mindset
risks regulation that is inappropriate to large swathes of
animal science, and encourages concepts of welfare that
are at odds with a properly animal-centred approach. So
what can be done about it?
WHERE NOW?
Organismal animal biology, in the sense that evolutionary biologists, ecologists and zoologists recognize it, is at
the moment a tiny voice in the debate about the
regulation of animal science. If it does not become
louder, and participate much more actively in the media,
political consultations and national and international
forums that drive these issues, it will nd itself pushed
into an ever tighter corner in terms of its freedom to
operate. This will be bad news for animal science and bad
news for animal welfare. Animal behaviour has a prime
role to play here since it is the branch of animal science
that is concerned with the functional design of animals
and thus provides the evolutionary lens through which
underlying anatomy, physiology, biochemistry and genetics should be viewed. Indeed, one might almost
paraphrase Dobzhansky (1973) and say that: Nothing
in biology makes sense except in the light of the functioning whole organism. Even notable exceptions to

ESSAY

the maxim, such as selsh genetic elements (Hatcher

2000), extended phenotypic effects (Dawkins 1999) and
the genetic conicts inherent in genome evolution
(L.D. Hurst et al. 1996; Burt & Trivers 2006) emerge by
comparison with the adaptive interests and mechanisms
of individual organisms. But animal behaviour has
much more to say than this. In its concern with the
decision-making propensities of animals, it provides the
only avenue we have for establishing a meaningfully animal-centred concept of welfare, and for contextualizing
clinical, physiological and other changes that on their
own can provide only equivocal evidence from a welfare
point of view. The science of evolutionarily salient welfare is barely in its infancy; a vigorously expanding programme of behavioural studies is essential if we are ever
to have a sound framework for animal welfare policy.
This means much more experimental animal biology
and an approach to regulation that is properly inclusive
of the branches of animal science and founded on robust
evolutionary understanding. We who study behaviour
have a particular responsibility in this regard and should
make sure our voices are heard both in the ongoing public and political debate about animal experimentation,
and in the organizations, conferences and journals concerning themselves with animal welfare. If we do not,
and we wind up saddled with inappropriate legislation
based on skewed notions of welfare and good practice,
we shall have only ourselves to blame.
Acknowledgments
I thank Lucy Barnard, Tom Reader, Olivia Curno, Jerzy
Behnke and Angela Turner for helpful comments on the
manuscript and Gilly Grifn for equally helpful correspondence. The views expressed are entirely the authors
and should not be taken to represent those of any
institution, professional society or other body. The essay
has its origins in two talks given at the workshop on The
Appropriateness and Application of the 3Rs in Animal
Research organized by the ASAB Ethical Committee on 28
November 2006; my thanks to Charlotte Nevison for
organizing the meeting.
References
Albrecht, G. 1999. A case for and some consequences of employing
an ecological ethic to guide wildlife research. In: The Use of Wildlife
for Research (Ed. by D. Mellor & V. Monamy), pp. 115e118. Glen
Osmond: ANZCCART.
Barnard, C. J. 2004. Animal Behaviour: Mechanism, Development,
Function and Evolution. London: Pearson.
Barnard, C. J. & Behnke, J. M. 2001. From psychoneuroimmunology to ecological immunology: life history strategies and immunity trade-offs. In: Psychoneuroimmunology. Vol. 2. 3rd edn. (Ed.
by R. L. Ader, D. J. Felten & N. Cohen), pp. 35e47. New York:
Academic Press.
Barnard, C. J. & Behnke, J. M. 2006. Behaviour, life history strategies and parasite infection in rodents. In: Micromammals and
Macroparasites: from Evolutionary Ecology to Management (Ed.
by S. Morand, B. R. Krasnov & R. Poulin), pp. 475e511. Tokyo:
Springer-Verlag.

Barnard, C. J. & Hurst, J. L. 1996. Welfare by design: the natural selection of welfare criteria. Animal Welfare, 5, 405e433.
Barnard, C. J., Hurst, J. L. & Aldhous, P. 1991. Of mice and kin: the
functional significance of kin bias in social behaviour. Biological
Reviews, 66, 379e430.
Barnard, C. J., Behnke, J. M. & Sewell, J. 1996a. Social status
and resistance to disease in house mice (Mus musculus): statusrelated modulation of hormonal responses in relation to immunity
costs in different social and physical environments. Ethology, 102,
63e84.
Barnard, C. J., Behnke, J. M. & Sewell, J. 1996b. Environmental enrichment, immunocompetence and resistance to Babesia microti in
male mice. Physiology and Behavior, 60, 1223e1231.
Barnard, C. J., Behnke, J. M., Gage, A. R., Brown, H. & Smithurst,
P. R. 1997a. Modulation of behavior and testosterone concentration in immunodepressed male laboratory mice (Mus musculus).
Physiology and Behavior, 61, 907e917.
Barnard, C. J., Behnke, J. M., Gage, A. R., Brown, H. & Smithurst,
P. R. 1997b. Immunity costs and behavioral modulation in male
laboratory mice (Mus musculus) exposed to the odour of females.
Physiology and Behavior, 62, 857e866.
Barnett, S. A. 1975. The Rat: a Study in Behavior. Chicago: University
of Chicago Press.
Bateson, P. 1986. When to experiment on animals. New Scientist,
20 February 1986, 30e32.
Bateson, P. 2005. Ethics and behavioral biology. Advances in the
Study of Behavior, 35, 211e233.
Berg, C. & Hammarstrom, A. 2006. The process of building a new
governmental authority based on public demands for improved
animal welfare. Livestock Science, 103, 297e302.
de Boo, M. J., Rennie, A. E., Buchanan-Smith, H. M. & Hendriksen,
C. F. M. 2005. The interplay between replacement, reduction and
refinement: considerations where the Three Rs interact. Animal
Welfare, 14, 327e332.
Bowlby, J. 1969. Attachment and Loss, Vol. 1: Attachment, New York:
Basic Books.
Bracke, M. B. M. & Hopster, H. 2006. Assessing the importance of
natural behavior for animal welfare. Journal of Agricultural and
Environmental Ethics, 19, 77e89.
Brain, P. F. & Benton, D. 1983. Conditions of housing, hormones,
and aggressive behavior. In: Hormones and Aggressive Behavior (Ed.
by B. B. Svare), pp. 351e372. New York: Plenum.
Brandstetter, H., Scheer, M., Heinekamp, C., Gippner-Steppert,
C., Loge, O., Ruprecht, L., Thull, B., Wagner, R., Wilheim,
R. & Scheuber, H. P. 2005. Performance evaluation of IVC
systems. Laboratory Animals, 39, 40e44.
Broom, D. M. 1986. Indicators of poor welfare. British Veterinary
Journal, 142, 524e526.
Broom, D. M. 1991. Assessing welfare and suffering. Behavioural
Processes, 25, 117e123.
Broom, D. M. & Johnson, K. 1993. Stress and Animal Welfare.
London: Chapman & Hall.
Buchanan-Smith, H. M., Rennie, A. E., Vitale, A., Pollo, S.,
Prescott, M. & Morton, D. R. 2005. Harmonizing the definition
of refinement. Animal Welfare, 14, 379e384.
Burman, O. H. P. & Mendl, M. 2003. The influence of pre-experimental experience on social discrimination in rats (Rattus norvegicus). Journal of Comparative Psychology, 117, 344e349.
Burn, C. C., Peters, A., Day, M. J. & Mason, G. J. 2006. Long-term
effects of cage-cleaning frequency and bedding type on laboratory
rat health, welfare and handleability: a cross-laboratory study.
Laboratory Animals, 40, 353e370.
Burt, A. & Trivers, R. L. 2006. Genes in Conflict: the Biology of Selfish
Genetic Elements. Cambridge, Massachusetts: Belknap.

11

12

ANIMAL BEHAVIOUR, 74, 1

Chamove, A. S. 1989. Environmental enrichment: a review. Animal

Technology, 40, 155e178.
Clark, J. A. M., Potter, M. & Harding, E. 2006. The welfare implications of animal breeding and breeding technologies in commercial
agriculture. Livestock Science, 103, 270e281.
Clark, J. D., Rager, D. R. & Calpin, J. P. 1997. Animal well-being
I. General considerations. Laboratory Animal Science, 47, 564e570.
Clutton-Brock, T. H., Albon, S. D. & Guinness, F. E. 1982. Red Deer:
Behaviour and Ecology of Two Sexes. Cambridge: Cambridge
University Press.
Cuthill, I. C. 2007. Ethical regulation and animal science: why animal behaviour is not so special. Animal Behaviour, 74, 15e22.
Dawkins, M. S. 2006. A users guide to animal welfare science.
Trends in Ecology and Evolution, 21, 77e82.
Dawkins, R. 1998. Unweaving the Rainbow. London: Allen Lane.
Dawkins, R. 1999. The Extended Phenotype: the Long Reach of the
Gene. 2nd edn. Oxford: Oxford University Press.
Dobzhansky, T. 1973. Nothing in biology makes sense except in the
light of evolution. American Biology Teacher, 35, 125e129.
Evans, J. E., Cuthill, I. C. & Bennett, A. T. D. 2006. The effect of
flicker from fluorescent lights on mate choice in captive birds.
Animal Behaviour, 72, 393e400.
Ewert, J.-P. 1980. Neuroethology. New York: Springer Verlag.
Festing, M. F. W. 2001. Experimental approaches to the determination of genetic variability. Toxicology Letters, 120, 293e300.
Festing, M. F. W. 2003. Principles: the need for better experimental
design. Trends in Pharmacological Sciences, 24, 341e345.
Festing, M. F. W. 2004a. Good experimental design and statistics
can save animals, but how can it be promoted? Alternatives to
Laboratory Animals, 32, 133e135.
Festing, M. F. W. 2004b. Refinement and reduction through
the control of variation. Alternatives to Laboratory Animals, 32,
259e263.
Folstad, I. & Karter, A. J. 1992. Parasites, bright males and the
immunocompetence handicap. American Naturalist, 139,
603e622.
Fraser, D. 1995. Science, values and animal welfare: exploring the
inextricable connection. Animal Welfare, 4, 103e117.
Gosling, L. M., Roberts, S. C., Thornton, E. A. & Andrew, M. J.
2000. Life history costs of olfactory signalling in mice. Behavioral
Ecology and Sociobiology, 48, 328e332.
Griffin, G. & Gauthier, C. 2004. Incorporation of the Three Rs in
wildlife research. Alternatives to Laboratory Animals, 32 (Suppl. 1),
215e219.
Grossman, C. J. 1985. Interactions between gonadal steroids and
the immune system. Science, 277, 257e261.
Haemisch, A. & Gartner, K. 1994. The cage design affects intermale
aggression in small groups of male laboratory mice: strain specific
consequences on social organization, and endocrine activations in
two inbred strains (DBA/2J and CBA/J). Journal of Experimental
Animal Science, 36, 101e116.
Harrison, R. 1988. Special address. Applied Animal Behaviour Science,
20, 21e27.
Hatcher, M. 2000. Persistence of selfish genetic elements: population structure and conflict. Trends in Ecology and Evolution, 15,
271e277.
Horgan, R. & Gavinelli, A. 2006. The expanding role of animal
welfare within EU legislation and beyond. Livestock Science, 103,
303e307.
Hudson, M. & Bhogal, N. 2004. An analysis of the home office statistics of scientific procedures on living animals, Great Britain
2004. Alternatives to Laboratory Animals, 34, 85e103.
Hurnik, J. F. 1988. Welfare of farm animals. Applied Animal Behaviour
Science, 20, 105e117.

Hurst, J. L. 1987a. The functions of urine marking in a free-living

population of house mice, Mus domesticus Rutty. Animal Behaviour,
35, 1433e1442.
Hurst, J. L. 1987b. Behavioural variation in wild house mice (Mus domesticus Rutty): a quantitative assessment of female social organization. Animal Behaviour, 35, 1846e1857.
Hurst, J. L. 1990. Urine marking in populations of wild house mice
(Mus domesticus Rutty). I. Communication between males. Animal
Behaviour, 40, 209e222.
Hurst, J. L. 1993. The priming effects of urine substrate marks on interactions between male house mice, Mus musculus domesticus,
Schwarz and Schwarz. Animal Behaviour, 45, 55e81.
Hurst, J. L., Barnard, C. J., Hare, R., Wheeldon, E. L. & West, C. D.
1996. Housing and welfare in laboratory rats: status-dependent
time budgeting and pathophysiology in single-sex groups maintained in open rooms. Animal Behaviour, 52, 335e360.
Hurst, J. L., Barnard, C. J., Nevison, C. M. & West, C. D. 1997.
Housing and welfare in laboratory rats: welfare implications of isolation and social contact among caged males. Animal Welfare, 6,
329e347.
Hurst, J. L., Barnard, C. J., Nevison, C. M. & West, C. D. 1998.
Housing and welfare in laboratory rats: the welfare implications
of social isolation and social contact among females. Animal
Welfare, 7, 121e136.
Hurst, J. L., Barnard, C. J., Tolladay, U., Nevison, C. M. & West,
C. D. 1999. Housing and welfare in laboratory rats: effects of
cage stocking density and behavioural predictors of welfare.
Animal Behaviour, 58, 563e586.
Hurst, L. D., Atlan, A. & Bengtsson, B. O. 1996. Genetic conflicts.
Quarterly Review of Biology, 71, 317e364.
Katz, P. S. & Frost, W. N. 1995. Intrinsic neuromodulation in the
Tritonia swim CPG: the serotonergic dorsal swim interneurons
act presynaptically to enhance transmitter relase from interneuron
C2. Journal of Neuroscience, 15, 6035e6045.
Kim, K. C. & Byrne, L. B. 2006. Biodiversity loss and the taxonomic
bottleneck: emerging biodiversity science. Ecological Research, 21,
794e810.
Laszlo, V. & Csaba, C. 2006. Legal aspects of animal experiments.
Magyar Allatorvosok Lapja, 128, 239e245.
Laurance, C. 2006. Animal welfare legislation: where have we got
to? Journal of Small Animal Practice, 47, 489.
Lore, R. L. & Flannelly, K. 1977. Rat societies. Scientific American,
236, 106e116.
Mace, R. 1995. Why do we do what we do? Trends in Ecology and
Evolution, 10, 4e5.
McFarland, D. J. 1989. Problems of Animal Behaviour. Harlow: Longman.
McGregor, P. K. & Ayling, S. J. 1990. Varied cages result in more
aggression in male CFLP mice. Applied Animal Behaviour Science,
26, 277e281.
Mackintosh, J. H. 1981. Behaviour of the house mouse. Symposium
of the Zoological Society of London, 47, 337e365.
Marashi, V., Barnekow, A. & Sachser, N. 2004. Effects of environmental enrichment on males of a docile inbred strain of mice.
Physiology and Behavior, 82, 765e776.
Markowitz, H. 1982. Behavioral Enrichment in the Zoo. New York:
Van Nostrand.
Martin, G. 1975. Uber Verhaltensstorungen von Legehennen in
Kafig: ein Beitrag zur Klarung des Problemes tierschutzgerechter
Huhnhaltung. Angerwandte Ornithologie, 4, 145e176.
Mason, G. J. & Mendl, M. 1993. Why is there no simple way of
measuring animal welfare? Animal Welfare, 2, 301e309.
Mattson, K. M. & Angermeier, P. L. 2007. Integrating human impacts and ecological integrity into a risk-based protocol for conservation planning. Environmental Management, 39, 125e138.

ESSAY

Mendl, M. 1991. Some problems with a cut-off point for determining when an animals welfare is at risk. Applied Animal Behaviour
Science, 31, 139e146.
Mendl, M. & Paul, E. S. 2004. Consciousness, emotion and
animal welfare: insights from cognitive science. Animal Welfare,
13, S17eS25.
Midgley, M. 1983. Animals and Why They Matter. Harmondsworth:
Penguin.
Milligan, S. R., Sales, G. D. & Khirnykh, K. 1993. Sound levels in
rooms housing laboratory animals: an uncontrolled daily variable.
Physiology and Behavior, 53, 1067e1076.
Moberg, G. P. 1993. Using risk assessment to define animal welfare.
Journal of Agricultural and Environmental Ethics, 6 (Supplement 2),
1e7.
Mormede, P., Lemaire, V., Castanon, N., Dulluc, J., Laval, M. M. &
Le Moal, M. 1990. Multiple neuroendocrine responses to chronic
social stress: interaction between individual and situational factors.
Physiology and Behavior, 47, 1099e1105.
Patterson-Kane, E. P. 2004. Enrichment of laboratory caging for
rats: a review. Animal Welfare, 13, S209eS214.
Patterson-Kane, E. P., Hunt, E. G. & Harper, D. 2002. Rats demand
social contact. Animal Welfare, 11, 327e332.
Patterson-Kane, E. P., Hunt, E. G. & Harper, D. 2004. Rats demand for group size. Journal of Applied Animal Welfare Science, 7,
267e272.
Paul, E. S., Harding, E. J. & Mendl, M. 2005. Measuring emotional
processes in animals: the utility of a cognitive approach. Neuroscience and Biobehavioural Reviews, 29, 469e491.
Regan, T. 1984. The Case for Animal Rights. Berkeley, California:
University of California Press.
Reiss, M. 1993. Introductory overview: an ethical framework for the use
of animals in research and education. In: Ethical Issues in Biomedical
Sciences: Animals in Research and Education (Ed. by D. Anderson,
M. Reiss & P. Campbell), pp. 3e8. London: Institute of Biology.
Rollin, B. E. 1981. Animal Rights and Human Morality. Buffalo:
Prometheus.
Rollin, B. E. 1989. The Unheeded Cry: Animal Consciousness, Animal
Pain and Science. New York: Oxford University Press.
Rollin, B. E. 1993. Animal welfare, science and value. Journal
of Agricultural and Environmental Ethics, 6 (Supplement 2), 44e50.
Rollin, B. E. 1998. On telos and genetic engineering. In: Animal Biotechnology and Ethics (Ed. by A. Holland & A. Johnson), pp. 156e
171. London: Chapman & Hall.

Russell, W. M. S. & Burch, R. 1959. Principles of Humane Experimental Technique. London: Methuen.
Schuppli, C. A. & Fraser, D. 2005. The interpretation and application of the Three Rs by Animal Ethics Committee members. Alternatives to Laboratory Animals, 33, 487e500.
Schuppli, C. A., Fraser, D. & McDonald, D. 2004. Expanding the
three Rs to meet new challenges in humane animal experimentation. Alternatives to Laboratory Animals, 32, 525e532.
Serpell, J. A. 2003. Anthropomorphism and anthropomorphic
selection: beyond the cute response. Society and Animals, 11,
83e100.
Signal, T. D. & Taylor, N. 2006. Attitudes to animals in the animal
protection community compared to a normative community sample. Society and Animals, 14, 265e274.
Singer, P. 1990. Animal Liberation. New York: Avon.
Skutch, A. F. 1997. Three biological heresies. Revista de Biologia
Tropical, 45, 1e11.
Smyth, D. 1978. Alternatives to Animal Experiments. London: Scholar
Press.
Thorpe, W. H. 1965. The assessment of pain and distress of
animals in livestock husbandry systems. In: Report of the Technical
Committee to Enquire into the Welfare of Animals Kept under Intensive Livestock Husbandry Systems . Command paper 2836, pp.
71e79. London: HMSO.
Timberlake, W. 1997. An animal-centred, causal-system approach
to the understanding and control of behaviour. Applied Animal
Behaviour Science, 53, 107e129.
Vorstenbosch, J. M. G. 2005. The ethics of the Three Rs principle:
a reconsideration. Animal Welfare, 14, 339e345.
Wickins-Drazilova, D. 2006. Zoo animal welfare. Journal of Agricultural and Environmental Ethics, 19, 27e36.
Willows, A. O. D. & Hoyle, G. 1969. Neuronal network triggering
a fixed action pattern. Science, 166, 1549e1551.
Wolfensohn, S. & Lloyd, M. 1994. Handbook of Laboratory
Animal Management and Welfare. Oxford: Oxford University
Press.
rbel, H. 2000. Behaviour and the standardization fallacy. Nature
Wu
Genetics, 26, 263.
rbel, H. 2001. Ideal homes? Housing effects on rodent brain and
Wu
behaviour. Trends in Neurosciences, 24, 207e211.
rbel, H. 2002. Behavioral phenotyping enhanced: beyond
Wu
(environmental) standardization. Genes, Brains and Behavior, 1,
3e8.

13