Conservation Biology

Conservation BiologyStanford Encyclopedia of Philosophy

Jay OdenbaughMarch 29, 2015

1 IntroductionOne way of conceptualizing the structure of a science is that at any time it containsontologies, theories, and values (Kuhn, 2012; Lakatos, 1976; Laudan, 1978, 1986).In this entry, we explore questions regarding the ontology, theoretical structure,and values central to conservation biology.

1. What is biodiversity?2. What is the structure of conservation biology?3. What are the aims of conservation biology?

First, conservation biology as a discipline has expended a great deal of intellectualeffort in articulating exactly what is its object of study and has settled on biodiver-sity as the answer. However, there is a debate concerning what biodiversity is, andwhether it should be center stage in the discipline. Second, conservation biology,though a scientific discipline with its theories, models, experiments, and field-work also utilizes a variety of tools, which are do not fit into the above categories.That is, some of its core tools are not theories or models in the ordinary sense.Thus, as discipline it is appears interestingly unique and ultimately unlike onesthat philosophers of science have typically examined. Third, conservation biologywas originally articulated by the early pioneers as a "crisis discipline" where cer-tain types of normative concerns were at the forefront. In fact, many have likenedit to medicine insofar as it has ethical foundations. In the medical case, it would be

1

the well-being of patients and in conservation biology it is the conservation of bio-diversity. Thus, we should ask what are the discipline’s aims or values explicitlyethical or sociopolitical and to what extent are they controversial?

2 What is Biodiversity?Conservation biology clearly concerns conserving something biological or ecolog-ical, but what is or should be conserved? Work has focused on a variety of units. Asexamples, some have focused on species such as the spotted owl Strix occidentalis(Yaffee, 1994) and loggerhead turtle Caretta caretta (Bolten and Witherington,2003); some have focused on populations and sub-species such as wild salmonOncorhynchus spp. (Quinn, 2011); some have focused on biomes or eco-regionssuch as tropical coral in the Great Barrier Reef (Hutchings et al., 2008); and finallysome have focused on genotypes or genetic features such as heterozygosity (Aviseet al., 1996). However, in the 1980s, conservation biologists united and arguedthat biodiversity should be the focus of the discipline (Wilson, 1988). What thenis biodiversity? Here is a standard definition from an influential textbook.

Biological diversity, or biodiversity, is the sum total of all living things- the immense richness and variation of the living world. Biodiver-sity can be considered at many levels of biological variation, rangingfrom genetic variability within a species, to the biota of some selectedregion of the globe, to the number of evolutionary lineages and thedegree of distinctness among them, to the diversity of ecosystems andbiomes on Earth. (Groom et al., 2006, 27)

Informally then, biodiversity is variation at all levels of the biological hierar-chy. To conserve biodiversity would then to be to conserve all of this variation.However, there are two fundamental issues to be resolved. First, there is an issueof practicality. As philosopher and conservation biologist Sahotra Sarkar writes,

Conserving biodiversity, and construing the term intuitively to referto all the biological diversity that there is, at every level of both hi-erarchies amounts to saying that "biodiversity" refers to all biologicalentities. "Biodiversity" in effect becomes all of biology. Conserva-tion would be an impractical proposal if biodiversity were construedin this way: everything biological would become a goal of conserva-tion. (Sarkar, 2005, 180)

2

In fact the issue is more complicated than even this. In order to estimate the amountof biodiversity with regard to a place, we must find some true surrogate (Margulesand Sarkar, 2007, Ch. 2). That is, we must find some feature that strongly corre-lates with biodiversity. Suppose we used species as a true surrogate. We cannotsimply observe the species present in some locale. We must use some estimatorsurrogate for the true surrogate too; e.g. soil, vegetation, precipitation, etc. Thus,we estimate biodiversity twice removed. Second, since cannot conserve all bio-logical variation, we need a principled way to decide what to conserve. As wesaw above, conservation biologists have refine they account to include variation ingenes, species, and ecosystems. This is still far too broad since there is a great of"junk DNA"; often we want to conserve wild populations (e.g. wild salmon ver-sus hatchery salmon); and rare habitats are often considered more important thancommon ones (Morgan, 2009).

In the remainder of this section, we consider three recent proposals concerningwhat biodiversity is (and is not) that try to characterize biodiversity in a practicaland principled manner.

2.1 Biodiversity as MultidimensionalOne attempt to provide a principled and practical way to characterize biodiversitycomes from Maclaurin and Sterelny (2008). James Maclaurin and Kim Sterelnystart with the suggestion that find the relevant "natural" units and then providemeans for measuring the variation with regard to those units. By ’natural’ units,they mean ones that identify properties that are recognizable and are causallysalient (Maclaurin and Sterelny, 2008, 10). As the history of phonetic taxon-omy has shown, measures of "overall similarity" were fraught with subjectivity.This subjectivity depended on choice of distance metric and on how one articu-lates character traits (Ereshefsky, 2000, 60-66). Contemporary taxonomists grouptaxa according to patterns of evolutionary descent, which is called "cladistics,"precisely to avoid such subjectivity given that the evolutionary history of life isnatural in the above sense (Baum and Smith, 2013).

Maclaurin and Sterelny contend that the fundamental unit of biodiversity isspecies and specifically species richness. Notoriously, biologists and philosophershave debated the nature of species. However, Maclaurin and Sterelny defend anevolutionary concept of species which states that "Evolutionary species are lin-eages of organisms with their ’own evolutionary tendencies and historical fate’"((Maclaurin and Sterelny, 2008, 33), see Wiley (1978)). They suggest that this isthe more natural approach to species because,

3

The living world is organized into phenomenological species: rec-ognizable, reidentifiable clusters of organisms. This fact makes theproduction of bird and butterfly field guides, identification keys forinvertebrates, regional floras, and the like, all possible... There aremany routes through which one population can become demographi-cally isolated from, and hence evolutionarily independent of, popula-tions that were once sources and sinks of its own genes. But the fact ofisolation and evolutionary independence is of immense importance tothe fate of local adaptation in such populations. So the phenomenolog-ical species richness of a region is, in an importance sense, a catalogueboth of phenotypic variety and of the potential evolutionary resourcesavailable in that region. (Maclaurin and Sterelny, 2008, 40)

That is, though a variety of evolutionary processes lead to the formation and co-hesion of organisms into species, there nevertheless is a uniformity of pattern inthe natural world.

Species richness with regard to a given community, a group of interactingspecies in a given place, is the number of species present. We might also addspecies evenness to species evenness (Magurran, 1988, 2013). With respect to agiven community, there will be some number of species. However, suppose wehave two species in which the first has 300 members but the second has 700. Sup-pose again we have two species where the first has 500 members and the second500 as well. The second community of species is far more "even" than the first.One formal representation of evenness is as follows (Magurran, 1988). We startwith the Shannon-Weiner diversity index.

H ′ = −S∑

i=1pi ln pi

Here pi is the proportion of organisms belonging to the ith species and S is thetotal number of species. If we define the following,

H ′

max = −S∑

i=1

1Sln 1S= lnS

We now have a measure of species evenness,

J ′ = H ′

H ′max

4

J ′ takes a value between 0 and 1 and the closer to the value of 1, the less variationbetween the number of individuals in a species.

The first way in which we might augment species richness is through disparity,morphological diversity, or more simply, phenotypic variation ((Maclaurin andSterelny, 2008, Ch. 3, 4) see Gould (2000) for a discussion of disparity in thehistory of life). Disparity, morphological diversity, or phenotypic vacation in gen-eral, concern the variation in phenotypic traits exhibited in and between species.Suppose we have one suite of species that are very similar phenotypically; how-ever, we might have another suite of species that vary a great deal phenotypically.A measure of such variation would weight species richness with some parame-ter(s) representing the variation in phenotypic traits. Unfortunately, if we includesuch phenotypic variation per se we find ourselves exactly with the problems thatdogged phenetic taxonomy and the attempt to capture some notion of overall simi-larity. Additionally, if we focus on disparity per se as opposed to mere phenotypicvariation, we must try to capture some notion of basic "body plan," (Arthur, 2000).

Maclaurin and Sternly suggest that one way to operationalize such variation isthrough the notion of a morphospace (Maclaurin and Sterelny, 2008, Section 4.4).Suppose we have a n-dimensional hyperspace where each dimension i is a variableconcerning the ith phenotype. Thus, a point in the space represents a specificorganism and their phenotype trait. Associated with any given species would be ahypervolume representing the phenotypic variation of that species. Though thereare not many worked out hyperspaces of this sort, a famous one is patelontologistDavid Raup’s analysis of the coiling patterns in mollusks that includes variablesfor whorl expansion rate, uncoiling rate, translation of the generating curve, andshape of the growing edge (Raup, 1966).

5

Figure 1: Raup’s coiling patterns in mollusks morphospace (Raup, 1966)

The morphospace describes actual coiling patterns and merely possible ones, and

6

raises explananda for evolutionary biologists. Why did certain coiling patternsevolve and not others? Maclaurin and Sterelny claim morphospaces will be oflow dimensionality and specifically limited to lineages or clades (Maclaurin andSterelny, 2008, 75-9). Thus, biologists must be assured that the variables includedare representative of the phenotypic traits of the organisms and hence species un-der study. They hypothesize that the variables will correspond to developmentalmodules or mosaics, which is a relatively independent developmental system thatcontributes to fitness (Wagner, 1995). This approach of course assumes that thereexist independent developmental systems, and ideally it would help articulate howa lineage would fair over time as a function of its plasticity.

Another component of biodiversity according to Maclaurin and Sterelny con-cerns the ecological interactions between species (Morin, 2011). This additionalstructure is given by interspecific competition, predator-prey relations, and mu-tualism along with possible higher-order ones like trophic cascades. In the ab-sence of such interactions, a group of species is nothing more than the "sum ofthe parts" (Clements, 1916; Eliot, 2011; Gleason, 1926; Odenbaugh, 2007; Whit-taker, 1967). In the presence of such interactions, they hypothesize the existenceof emergent properties that are not a "simple reflection" of the species themselves(Maclaurin and Sterelny, 2008, 120). Additionally, even if there were such emer-gent properties, we would like to know their causal contribution is; how to do theydrive community patterns? Infamously, some ecologists have hypothesized thatas the diversity of a community increases, so does the stability of the community(Egerton, 1973; Elton, 2000; MacArthur, 1955; May, 2001; Tilman et al., 1996).’Diversity’ can be understood as the number of species present, the number ofinteractions, or intensity of those interactions. ’Stability’ can be understood in avariety of ways from notions of Lyapunov stability, return time after a perturba-tion, biomass stability, etc (Justus, 2008; Pimm, 1984). Ecologists have theorized,experimented, and collected a great deal of data regarding this hypothesis, but itremains elusive as to its truth. Maclaurin and Sterelny suggest we can articulate anecospace including at least the boundedness, degree of emergent properties, andinternal regulation as variables to describe ecological communities (Maclaurin andSterelny, 2008, Ch. 6). One serious worry is whether any of these features can bemeasured in an informative way. But, insofar as there are no community bound-aries, no emergent properties, and no internal regulation, there are no communities.The robustness of communities would require each of these take non-zero values.Additionally, these variables are not independent and hence the ecospace wouldhave to take account of the interactions between each variable. For example, in-ternal regulation might be a function of the emergent properties possessed by the

7

community, and the boundaries of a given community determined by those veryproperties (Sterelny, 2006).

Finally, we can consider how to include phylogenetic information in our mea-sure of biodiversity. One method comes from Daniel Faith (Faith, 1992, 1994,2002). In cladistics, we use character traits to divide organisms into groups. Webegin with the traits shared by all of the organisms and then articulate smallergroups by traits unique to them where each then is a clade. A cladogram thendisplays the distinct character traits of each group. Cladograms differ from phy-logenetic trees insofar as the latter includes ancestor-descendent relations apartfrom merely relations of nested similarity. For any phylogenetic tree, there aremany cladograms consistent with it. Faith writes,

The phylogenetic diversity (’PD’) of s is equal to the sumof the lengthsof all those branches that are members of the corresponding minimumspanning path. (Faith, 1992, 4)

Consider the following cladogram.

Figure 2: The phylogenetic diversity represented in a cladogram (Faith, 1992, 3)

In Figure 2, we have taxa 2, 6, 8, 10. The minimum spanning path connecting eachrequires 28 character changes and thus its PD is 28. As Maclaurin and Sterelnynote, this assumes that we have an adequate sample of character traits and giventhat many are molecular they do not tell us much about phenotypes (Maclaurin

8

and Sterelny, 2008, 141). Thus, they suggest we stay with stick with local mor-phospaces.

Maclaurin and Sterelny’s approach is very suggestive but as we have seen itsexecution by biologists requires making good on using species richness augmentedwith information regarding phenotypic variation, ecology, development, and phy-logeny into am empirical tractable hyperspace. The extent to which this can bedone is unclear (Justus, 2010; Morgan, 2009).

2.2 Operationalizing BiodiversitySahotra Sarkar attempts to define biodiversity "implicitly" as opposed to "explic-itly" contrary toMaclaurin and Sterelny (Margules and Sarkar, 2007; Sarkar, 2002;Sarkar and Margules, 2002; Sarkar, 2005). A concept is given an explicit defini-tion if, and only if, there is a set of necessary and sufficient conditions for thesatisfaction of the concept. For example, suppose we define the concept SPECIESphylogenetically in the following way: necessarily, something is a species if, andonly if, it is a lineage between two speciation events or a speciation event and anextinction event (Ridley, 1989). An implicit definition is different; the concept isdefined by some procedure say rules or axioms in which the concept appears. So,Sarkar characterizes BIODIVERSITY as follows,

[B]iodiversity should be (implicitly) operationally defined as what isbeing optimized by the place prioritization procedures that prioritizeall places on the basis of their biodiversity content using true surro-gates (Sarkar, 2005, 182).

To understand this account, we need to articulate what a place prioritizationprocedure is (Sarkar et al., 2002; Sarkar, 2005, 159-68). In the abstract, place pri-oritization starts with a list of places ("cells") and a list of surrogates that mustbe represented at a certain target in a conservation area network. We either wantto select the smallest set of cells such that every one of our surrogates meets ittarget, or select those cells subject to a given size, which maximizes the numberof surrogates at or above the relevant target. Place prioritization is a componentof a consensus framework called systematic conservation planning that we willdiscuss below. However, in effect, it is a procedure for prioritizing places for their"biodiversity content" for creating conservation area networks in which we incor-porate information regarding the design of the area. This can include the shapeof the network, its size, dispersion, and connectivity. It is typical that this is put

9

more formally as follows (Sarkar, 2005, 161). First, suppose we begin with a set∑ of places or cells �j; thus, ∑ = {�j|�j ∈

∑

} for j = 1, 2, ..., n. Second, wehave a set Λ of surrogates �i; thus, Λ = {�i|�i ∈ Λ} for i = 1, 2, ..., m. Third, wehave targets for the representation of their expected coverage representation of �ifor i = 1, 2, ...., m. Place prioritization then takes the form of two problems:

• MinimumArea: selection the smallest set of cell Γ such that every surrogatemeets their target of representation;

• Maximum Representation: given the cardinality � ≤ n of Γ, select thosecells that maximize the number of surrogates for which the expected cover-age exceeds the targets of representation.

The former problem has economic constraints but none on the number of placeswhereas the second has a constraint on the total number of places.

As we saw above, true surrogates consist in species, characters or traits, life-zones, or environmental parameters. Estimator surrogates consist in environmen-tal parameter composition, soil-type composition, dominant vegetation composi-tion, subsets of species composition, and subsets of genus or other higher taxoncomposition. So, biodiversity is preserved by preserving true surrogates throughpreserving estimator surrogates (Sarkar, 2005, 168-73). To a first approximation,a place prioritization procedure optimizes the number of estimators and thus truesurrogates with respect to place, and what is so optimized is biodiversity. Biodi-versity just is what is optimized by this procedure.

There are several concerns worthmentioning with regard to how this procedureimplicitly defines BIODIVERSITY. First, since we are operationalizing the conceptBIODIVERSITY, it is simply defined by what a place prioritization procedure op-timizes. Thus, it cannot be in error with regard to whether it has captured somepre-theoretic notion. It rules out the possibility of mistake (Hempel, 1966, Ch. 7).We do not think that the concept of TEMPERATURE is just whatever is measuredby thermometers (Chang, 2004). Second, insofar as there are different place prior-itization procedures, they will implicitly define distinct concepts. They will not bemeasuring the same magnitude. Third, concept explication requires that the expli-catum be distinct but sufficiently similar to the explicandum for the analysis to beof theoretical value (Carnap, 1962, Ch.1). There are two ways of seeing how theconcept(s) articulate by Sarkar differ from our "ordinary" ones. The concept mostconservation biologists articulate is categorical and not relational. The concept issimilarly imbued with much moralizing as we shall see and one can be compe-

10

tent with the one and not the other. Hence, one might argue that they are differentconcepts (Odenbaugh, 2009). As David Takacs writes,

Biodiversity is the rallying cry currently used by biologists to drawattention to this crisis and to encapsulate the Earth’s myriad speciesand biological processes, as well as a host of values ascribed to thenatural world. (Takacs, 1996, 9)

At first blush, one might wonder if this is the same thing articulated by place pri-oritization procedures.

2.3 Eliminating BiodiversityRecently philosopher Carlos Santana has argued that we should eliminate the con-cept of biodiversity from conservation biology (Santana, 2014). He writes,

Biodiversity is generally the assumed target of conservation biology,but the biological world is composed of a number of distinct types ofdiversity, which only loosely correlate with each other and with bio-logical value. Since the function of the biodiversity concept in conser-vation science is to help us preserve or increase biological value, weshould therefore consider eliminating biodiversity from its privilegedposition in conservation theory and practice. (Santana, 2014, 778)

As we have seen, we begin with a notion of biodiversity which can be estimatedby a true surrogate but which actually is empirically tracked through a estimatorsurrogate. However, Santana points out that we are concerned with biodiversitybecause of the "biological values" associated with it. This leads to two sequences(Santana, 2014, 765):

1. estimator surrogate→ true surrogate → biodiversity → biological value2. estimator surrogate→ true surrogate → biological value

Santana contends that (2) is preferable to (1). Why does he think this? He arguesthat both Sarkar’s and Maclaurin and Sterelny’s respective attempts at defining’biodiversity’ fail. Moreover, since according to Santana, using the concept inconservation biology does more harm than good, we should eliminate its use.

11

First, Santana argues that there are many different sorts of diversity of biologi-cal value that cannot be subsumed by any single true surrogate like species richness(Santana, 2014, 768-9). For example, we often wish to conserve biomes such asold growth forests or coral reefs independent of their composition of species. Or,we might want to conserve distinct salmonoid populations even if they are of thesame species. Second, assuming that is a single or small number of true surrogatessuch as species richness ignores the fact that not all species are of equal biologicalvalue. A species of pollinator or top predator like the grey wolf is of greater thanvalue than some species ofminnow. Additionally, there are values that are only ten-dentiously connected to species richness like "wildness" or aesthetic value. Third,some species are of great disvalue. For example, many would argue that taxa likeEbola would be better extinguished. If this is correct, Santana contends that thereis no single true surrogate through biodiversity can be articulated. We must havevery pluralistic notion of biodiversity.

The second part of Santana’s argument is this pluralism is unproblematic onlyif there some true surrogate that covaries with the other surrogates and hence bi-ological value (Santana, 2014, 768-73). However, he complains that the availableevidence suggests this is false. If we augment species richness with abundancewe will not capture certain biological values since for example the rarity of a birdspecies increases its value and does not reduce it. Phenotypic variation requiresthe sort of morphospaces discussed above, but this as we saw are local and cannotbe compared across very different taxa. Hence, even if we augment species rich-ness with phenotypic variation we are unable to determine the biodiversity presentin comparable ways. Likewise, evolutionary diversity of the sort that Faith de-scribed does not necessarily correlate with species richness. In one study of theSouth Africa’s cape region, it was determined that the eastern region of the capehad greater evolutionary diversity as measured by the number of speciation eventsbetween clades versus the western region of the cape which had greater speciesrichness (Mooers, 2007). Speciation events are often unrecoverable from the fos-sil record and so this sort of diversity cannot typically augment species richness.Lastly, as we noted above, the connection between species richness and ecologicalproperties like stability is extremely complicated and controversial. It has not beenshow in a robust manner that greater species richness leads to greater ecosystemor food web stability.

Thus, there aremany different types of diversity encoded in the sorts of true sur-rogates we would like to conserve and they do not positively correlate in straight-forward ways. If this is right, Santana contends that the concept of biodiversityin effect is not tracking a feature of the biological world. At most it is a place-

12

holder for whatever conservation biologists choose to conserve. The concept hascertainly made its way into public discourse and one might worry that this wouldbe extremely disruptive if we removed it from conservation practice. However,Santana suggests we should not try to decide what to conserve based on someelusive property of biological systems, which really is just a disjunction of verydifferent diversities. Rather, the tools through which value conflicts are best artic-ulated is from political science and economics (Frank and Sarkar, 2010; Colyvanet al., 2011). Regardless of the tools used, Santana argues that we should bypassthe concept of biodiversity and consider the biological values that we share andevaluate true surrogates on those grounds. He writes,

Biodiversity is generally the assumed target of conservation biology,but the biological world is composed of a number of distinct types ofdiversity, which only loosely correlate with each other and with bio-logical value. Since the function of the biodiversity concept in conser-vation science is to help us preserve or increase biological value, weshould therefore consider eliminating biodiversity from its privilegedposition in conservation theory and practice. (Santana, 2014, 778)

The discipline of conservation biology has settled on biodiversity as its pur-ported object of study. However, there is a serious debate over whether biodiversityis some natural feature possibly multdimensional or simply whatever is arrived atthrough procedures proposed in designing conservation area networks. And, someeven have argued that it should be eliminated from scientific practice.

3 What is Conservation Biology?Regardless if there is single thing that conservation biology as a discipline purportsto conserve, there are philosophical questions to be raised regarding the disciplineitself. Specifically, is it a collection of theories or models or something else? Toexplore this question, we will consider central examples of conservation biology’spractice. We can then turn to how best to understand what conservation biologyis. But first, it useful to consider what theories and models are.

Traditionally, philosophers of science have typically focused on theories, whichthey assumed were sets of natural laws (Hempel, 1966). Examples of laws ofnature including the following. Newton’s universal law of gravitation is Fab =G(mamb)∕r2 which states that the gravitational force between two objects a and b

13

is proportional to their masses and inversely proportional the square of their dis-tance. Newton’s second law of motion is F = ma where the vector sum of forcesF equals mass m times acceleration vector a. Finally, there is the Boyle-Charlesideal gas law which is PV = nRT which says that the pressure P and volume V ofa gas is proportional to its temperature T . Philosophers of science and metaphysi-cians debate what makes a statement like those above laws. The answers rangefrom they are generalizations concerning empirical regularities (Ayer, 1999), rela-tions between universals (Dretske, 1977), the best system that trades simplicity inprimitives for explanatory power (Lewis, 1983), encode various symmetries (vanFraassen, 1989), and so on. One powerful worry with thinking that theories aresets of laws of nature is this. If, at a minimum, laws are true universal general-izations, then there are few such laws at least in the species sciences since thosegeneralizations have exceptions (Beatty, 1995; Cartwright, 1983). That is, theyare false. As the result of this sort of argument amongst others, philosophers ofscience have focused on models (Giere, 2010; Odenbaugh, 2005;Weisberg, 2012).

Models are a type of representation but they involve abstractions and idealiza-tions (Cartwright, 1999; Chakravartty, 2001; Weisberg, 2012). Abstraction in arepresentation involves not including all of the properties of the thing representedand idealization involves distorting those features that are represented. That is,models delete and distort. Another feature of models that is important is that theytypically represent the world indirectly – models are developed and after this, sci-entists determine the respects and degrees to which they are similar to some sys-tem of interest. Theories as sets of laws are thought to represent the world directlysince their propositional content concerns natural spatiotemporal systems. Theequations describing models have as their proportional content mathematical orabstract objects, and these objects are more or less similar to particular objects inthe world (Weisberg 2012, but see Hughes 1997; Odenbaugh 2014). There are anumber of different types of models used which include physical, scale, computa-tional, and mathematical models.

In the biological sciences especially, one rarely hears talk of laws; rather, it ismodels that are discussed. We can begin by considering a simple model used inconservation biology, a metapopulation model (Hanski and Gilpin, 1997; Levins,1970). A metapopulation is a population of populations that are subdivided spa-tially, but are causally connected through migration. Let P be the proportion ofpatches occupied by a species, c is the rate of patch colonization, and e is therate of patch extinction. Thus, the instantaneous rate of change in the proportionof occupied patches is the proportion of patches colonized minus those in whichextinctions occur. Mathematically, we have,

14

dPdt

= cP (1 − P ) − eP

When dP∕dt = 0, the non-trivial equilibrium proportion is P ∗ = (c − e)∕c =1 − e∕c. Thus, if the equilibrium proportion is greater than zero, then the rate ofcolonization must be greater than the rate of extinction; namely, P ∗ > 0 just incase c > e. Given that this is a model, there are abstractions and idealizations in-cluding the assumption that the local populations are identical in their chances ofcolonization and going extinct, rates of colonization and extinction are constants,and distance between patches does not matter. Models can be characterized bythree features – their parameters, variables, and laws (Otto and Day, 2007). Pa-rameters describe properties that are constant in value, variables describe prop-erties that change, and laws simply describe relationships between parametersand variables. Thus, P is a variable, c and e are parameters, and the equationdP∕dt = cP (1 − P ) − eP is law. Note that the notion of law used by modelers isnot necessarily the same as that of a law of nature.

Conservation biologists use models all the time in almost every facet of theirwork. So we now consider several case studies of conservation biology.

3.1 Northern Spotted OwlsMetapopulation models are extremely important and have been used in one of themost difficult of environmental debates, that surrounding the northern spotted owl(Strix occidentalis caurina) and its habitat (Durbin, 1996; Norse, 1989; Yaffee,1994). The northern spotted owl is subspecies that lives exclusively in old-growthconiferous forests in northern California, westernWashington and Oregon. North-ern spotted owls are monogamous and each breeding pair uses one to three squaremiles of forest that is at least two-hundred and fifty years old. Farms, logging andother forms of development reduced northern spotted owl habitat and so scien-tists and policymakers began to ask what must we do if we are to prevent themfrom going extinct? One way of answering that question is to use metapopulationmodels.

In 1985, environmental lawyer Andrew Stahl presented mathematical biologistRussell Lande with just this question (Yaffee, 1994, 98). He showed Lande the USForest Service’s "regional guide" which was their plan to protect the northern spot-ted owl and asked Lande to determine if it would be sufficient. Using a variety ofdata, he devised the following metapopulation model to try to answer this question

15

(Lande, 1988a but see Lande, 1988b as well). Let � be the probability that a juve-nile female inherits her mother’s territory,m be the number of territories a juvenilecan disperse through before dying, ℎ be the proportion of habitable territory, andp be the proportion of occupied habitable sites. Therefore, the probability of afemale not finding habitable territory in m trials is,

(1 − �)(pℎ + 1 − ℎ)m

This just is the probability of not inheriting the territory of one’s mother and notsuccessfully dispersing to an unoccupied territory before death. Female northernspotted owls reproduce only when they are three years old and we can representtheir growth rate by � = Nt+1∕Nt. When � = 1, the species is in a demographicequilibrium. Assuming they are in such an equilibrium, then,

[1 − (1 − �)(pℎ + 1 − ℎ)m]R′

0 = 1

R′

0 =∑∞

x=0 l′

xfx where l′x is the probability of surviving to age x assuming shehas found habitable territory and fx is the mean lifetime offspring production perfemale assuming she finds habitable territory. Finally, we can solve the following,

p̂ ={

1 − 1−kℎ

for ℎ > 1 − k0 for ℎ ≤ 1 − k

In the above, k is the the equilibrium occupancy of the territory. Given this model,the northern spotted owl can persist if, and only if, ℎ > 1 − k. Lande determinedthat in 1987, 38% of forest in western Washington and Oregon was older than twohundred years. So, ℎ = 0.38. Fieldwork suggests that 44% of sites were occupied.Thus, p = 0.44. We can determine k from this equation, k = 1− ℎ(1 − p). Hence,k = 0.79. The forest plans suggested leaving between 7−16% of 200 or more yearold forest. But given that 1−k = 0.21, Lande argued 7−16% was insufficient andthe forest plans should be revised. Lande and others’ work led to the US ForestService withdrawing six old growth forest timber sales in Oregon and Washingtonand became a component of the 1994 adoption of the Northwest Forest Plan thatprotected 8 million acres of old-growth forests. This was an increase to protecting80% of the owl’s remaining habitat in comparison to 5% protected by the ForestService’s original owl plan.

Models like Lande’s, and population viability analysis (PVA) more generally,often are susceptible to several criticisms (see Beissinger and McCullough 2002a discussion of PVA, and Wimsatt 2012 for a discussion of the robustness and

16

fragility ofmodels). They exhibit parameter and structural sensitivity; if we changethe parameter values or change the assumptions of the models, we often find verydifferent results. For example, Patrick Foley examined amodel of grizzly bearsUr-sus arctic in Yellowstone National Park that incorporated environmental stochas-ticity along with intrinsic growth rates and carrying capacity (Foley, 1994). Themodel for estimated values of the above predicted the expected time to extinctionwas 12, 000 years. However, he later examined a model which included demo-graphic and environmental stochasticity but set demographic stochasticity to zero(Foley, 1997). The expected time to extinction was approximately 50 years. Thus,PVA models can be very fragile with respect to their parameter values and func-tional forms.

Of course, Lande’s metapopulation model is clearly abstracted and idealized.However, he argued that his model was consistent with the data and that his predic-tions were more conservative than that of a more realistic model. First, he arguedthat the data showed that the annual geometric growth rate of the northern spot-ted was not significantly different from equilibrium (e.g. 0.96 ± 0.03, see Lande,1988a, 605). Second, if we assume that owls have difficulty finding mates, or dis-persal is more limited, or there is demographic or environmental stochasticity, thiswould reduce k even further.

3.2 Island BiogeographyIsland biogeography concerns the distribution and abundances of species on is-lands. One of the fundamental patterns the biologists have attempted to explain isthe species-area effect. That is, why do larger islands support more species thansmaller ones? It seems that as the area of an island increases, so does speciesrichness. The species-area effect is the product of the area effect namely thanmore species are found on larger islands rather than smaller, and the distance ef-fect namely that islands closer to the mainland have more species than those thatare further. E. O. Wilson and Robert H. MacArthur provided one of the famouspotential explanations of the species area effect. They did so through their equilib-rium model of island biogeography (MacArthur and Wilson, 1963, 1967). Theybegan with several assumptions including the following. First, species richness onan island has a stable equilibrium. Second, the stabile equilibrium is due to theimmigration rate from the mainland to the island the extinction rate on the island.Third, the distance of the island from the mainland by itself determines the rate ofimmigration. Fourth, the extinction rate is determined solely by the island’s size.Fifth, the stable equilibrium is a dynamic one since while the number of species

17

is constant, which species are on the island changes. This rate of change is thespecies turnover rate.

MacArthur and Wilson formulated a mathematical model with these assump-tions (Wilson and Bossert, 1971). Suppose P is a parameter representing a pool ofspecies. Let the total immigration rate �S be the number of new species colonizingan island per unit time. Let the total extinction rate �S be the number of specieson the island going extinct per unit time. Thus, we have the instantaneous rate ofchange of species on the island is,

dS∕dt = �S − �SIf �S = �S , then dS∕dt = 0, and we have an equilibrium S∗. Assume that Sspecies are present on our island. We characterize the total immigration rate asfollows. Let �A be the average immigration rate of news species colonizing theisland per unit time, and then the total immigration rate is �S = �A(P − S). Thetotal extinction rate can be characterized as well. Let the average extinction rateper species per unit be �A. Then �S = �AS. Therefore, dS∕dt = �A(P −S)−�A.Finally, we can derive dS∕dt evaluated at the equilibrium S∗,

dSdt

|

|

|S=S∗= �S − �S = �A(P − S) − �AS = 0

This is equivalent to the following expression

S∗ =�AP

�A + �ABy integrating the differential equation, we have,

S∗ =�AP

�A + �A(1 − e−(�A+�A)t)

As t increases, e−(�A+�A)t approaches zero.We can also derive an expression for the species turnover rate. MacArthur and

Wilson as we saw thought that though the number of species on an island would beconstant, their identities would be constantly changing. Suppose we are interestingin some fraction say 90% of the equilibrium number of species S∗. Then, wemultiply both sides of the above by 0.9, and have 0.9S∗ = �AP∕(�A + �A) × 0.9.From the above, it follows that (1 − e−(�A+�A)t) = 0.9. With some alegebra andtaking the natural logarithms, we arrive at the species turnover rate,

18

t0.9 =2.3

�A + �AFinally, let us assume that for each species Si’s size is proportional to the is-

land’s area and the probability that a species goes extinct decreases as island sizebecomes smaller.

Figure 3: Equilibriummodel and explanation of the species area effect (MacArthurand Wilson, 1967, 26)

Where the immigration rate I and extinction rate E cross is the species equilib-rium S∗ (Figure 3B). Moreover, the island whose immigration rate is Inear andwhose extinction rate is Elarge has the largest species equilibrium. The one withthe smallest is the one with Ifar and Esmall (Figure 3A).After constructing their model, E. O. Wilson and his graduate student DanielSimberloff tested the model by devising natural experiments in the Florida keys(Simberloff and Wilson, 1970). They had islands defaunated of varying sizes anddistances from the mainland. They could then evaluate the predictions and as-sumptions of themodel. They determined that species abundances returned to theirprevious equilibria and species richness on an island appeared to be determined bythe size of the island and its distance from the mainland. However, Simberloff wasable to show that the equilibriummodel predicted that the species turnover rate wasmuch higher than what was measured (Simberloff, 1976). In some ways, it is nota surprise that the model was predictively inaccurate since it was highly idealized.MacArthur and Wilson had assumed distance alone determines the immigrationrate and size alone determines the extinction rate. MacArthur and Wilson wrote,

We do not seriously believe that the particular formulations advanced

19

in the chapters to followwill fit for very long the exacting results of fu-ture empirical investigation. We hope instead that they will contributeto the stimulation of new forms of theoretical and empirical studies,which will lead in turn to a stronger general theory... (MacArthur andWilson, 1967)

But, conservation biologists quickly embraced the model and began to apply it tonew topics. The most important of which was reserve design.

Several conservation biologists but especially Jared Diamond argued that sin-gle large reserves were better several small ones (SLOSS) (Diamond, 1975b,a;May, 1975; Terborgh, 1975; Wilson and Willis, 1975).

Figure 4: Principles as depicted in the SLOSS debate (Diamond, 1975b, 174)

The principles debated included (A) larger reserves are better than smaller ones,(B) a large reserve is better than a few small ones of the same total area, (C) severalreserves close to one another are better than several further apart, (D) when there

20

are several reserves they should be grouped equidistantly rather then linearly, (E)reserves connected by corridors are better than unconnected ones, and (F) compactshapes are better for minimizing boundary length (Diamond, 1975b, 143-5).

Often the the relation between species and area is represented by the followingequation, S = kAz whereS is species richness,A is area, and k and z are constants(z is usually between 0.2 and 0.35). Clearly, as A increases other things beingequal, S should increase as well. So, it was argued that larger reserves are betterfor conserving species than smaller ones. Note that if the reserve is isolated fromother such reserves then only the extinction rate matters according to the equilib-rium model. In 1976, Daniel Simberloff and Lawrence Abele argued that this lineof argument was incorrect (Simberloff and Abele, 1976). Suppose we have a singlereserve of size A1 and juxtapose its against two with size A2 = A1∕2 each and thatz = 0.263 (this precise value is not important to the argument). Each of the twosmall reserves would have S2 = kAz

2 species. If all of the species P in the pool arecapable of dispersing in and surviving in the refuges, then the total expected num-ber of species in both refuges is 2S2 − S22∕P ; namely the species in either smallreserve minus those in both. How many species S1 would be in the large reserveA1? Simberloff and Abele determined that S1 = kAz

1 = k(2A2)z = 1.200 × S2.They pointed out that this this is less than the total expected number of species2S2 − S22∕P in the two smaller reserves when S1∕P < 0.96. Another importantcriticism of the use of the equilibrium model in reserve design was that reserveswere unlike islands in an ocean (Margules et al., 1982). Specifically, the equilib-rium model assumes that between the mainland and island no species of the poolcan exist; i.e. the immigration rate is zero. However, when we consider areas be-tween reserves, it often is possible species to occur even if at lower abundances.Thus, the equilibrium model could not be extended without violation of its as-sumptions to other habitat types such mountain tops (Brown, 1971).

The SLOSS debate was vociferous and the arguments continued for quite sometime. The importance of the equilibrium model of island biogeography was muchdiscussed and hotly contested (Whittaker and Fernández-Palacios, 2007).

3.3 Systematic Conservation PlanningThough the types of models we have just discussed have an important place in theconservation biology and its history, it has become an explicitly socio-ecologicaldiscipline in which sophisticated computational tools are used for the purpose ofdesigning conservation area networks. From work involving population genet-ics and ecology for population viability analysis and the equilibrium model of

21

MacArthur and Wilson, with the associated SLOSS debate, we see a disciplinetransformed. Work on the genetics of inbreeding, habitat fragmentation, metapop-ulation dynamics, and so on continues but in the guise of something more social;namely, systematic conservation planning (SPC) (Margules and Pressey, 2000;Margules and Sarkar, 2007; Watson et al., 2011). SPC involves a variety of stepsincluding the following:

1. Choose and delimit the planning region.2. Identify all stakeholders.3. Compile and assess all data.4. Treat data and construct models as necessary.5. Identify and evaluate biodiversity constituents and surrogates.6. Set explicit biodiversity goals and targets.7. Review existing conservation areas for performance with respect to targets.

additional areas for conservation management.8. Assess biodiversity constituent and selected area vulnerabilities.9. Refine the network of selected areas.

10. Carry out multi-criteria analysis.11. Implement conservation plan.12. Monitor network performance (Sarkar, 2012, 100-3)There are a variety of philosophical questions that are raised by SCP. First, who

are the stakeholders in determining the boundaries of the analysis, the relevant cri-teria to be used, and what is the focal biotic unit? Second, what is biodiversity andwhat are the biotic constituents we are trying to conserve and ensure persist? Givenproblems with current definitions of ’biodiversity’ as we saw above, we must artic-ulate what biodiversity constituents and surrogates are, along with other diversityand persistence concepts. Third, given resource scarcity, efficiency is one of thegoals of SCP, and thus how do we solve the minimum area and maximum repre-sentation problems mentioned above? What is striking is how much the disciplineincludes over and above models like the Lande metapopulation and MacArthurand Wilson equilibrium models. As Sahotra Sarkar writes,

22

Thus, in conservation biology, like computer science and unlike mostof ecology, theoretical research consists of devising algorithms ratherthan formulating models and theories. In fact, because a variety ofalgorithms can be used to solve these problems, a lot of theoretical de-bate in conservation biology has been about the choice of algorithms.(Sarkar, 2012, 124)

We have seen the role of place prioritization, which is key to SCP. I want to con-sidered one other formal tool that is used to help make decisions about how toaccommodate different values in choosing conservation area networks.

One way to represent values (i.e. preferences) would be through the tools ofneoclassical economics. Ideally we would reduce these values to a single scalevia a utility function. However, this is possible only if these values are com-mensurable (orderable on single scale) and we can measure then through peo-ple’s willingness-to-pay for them or willingness-to-be compensated for their loss.Multi-criteria analysis does not make these two assumptions (Arrow and Raynaud,1986). Rather, the preferred method is determining a set of non-dominated alter-natives (Moffett and Sarkar, 2006). Suppose have a set of criteria K = {�i|i =1, 2, ..., n} and a set of feasible alternatives A = {�j|j = 1, 2, ..., m} (for details,see Sarkar 2005, 196-203). We will further assume that each criterion �i induces aweak linear order ≤∗ onA; that is, for any two �m, �n ∈ A, �m ≤∗i �n or �n ≤∗i �m orboth. Letwim be the value of the mth alternative according to the ith criterion. Wecan say then that an alternative �m ∈ A is non-dominated by another alternative�n ∈ A with respect to K if, and only if:

∃p, kp ∈ Ksuch that wpm < wpn

∀q, kq ∈ K,wqn ≤ wqm

This implies that �n is strictly better than �m by at least one criterion in kp ∈ K ,and it not worse than �m by any criterion ki ∈ K . An alternative �i ∈ A is non-dominated if, and only if, it is not dominated by any other element of A. Thus,using multi-criteria analysis, we attempt to find a non-dominated set of alterna-tives from which to design conservation area networks. An example from (Sarkar,2012, 204-6) is briefly considered. We begin with place prioritization with 100attempts to create a prioritized lists of hexagons for a target of 10 representativesfor each vertebrate species present in Texas. There were thirty-two feasible so-lutions – ones that satisfied the conservation target. However, two other criteria

23

were used for mutli-criteria analysis, economic cost and social cost. The first wasestimated by the cost of the conservation plan per hexagon and the second was es-timated by the human population size of each hexagon included. From the thirty-two feasible solutions, there were only two non-dominated solutions selected bythe multi-criteria analysis.

Figure 5: Place prioritization and multi-criteria analysis in Texas (Sarkar, 2012,206)

I thus now want to reflect on the case studies and then ask what exactly isconservation biology?

3.4 Conservation Biology and the Philosophy of SciencePhilosophers of science have thought of the core questions of the discipline asthese. What is the structure of scientific theories? What is the logic of confirma-tion of theories by evidence? What is a scientific explanation? What is scientificobjectivity and does the presence of values in science thwart that objectivity? With

24

regard to the Lande’s metapopulation model and MacArthur andWilson’s equilib-rium model, we can ask the first three questions. First, we can recognize that theseare models that abstracted and idealized representations of objects like northernspotted owls and fauna on Mangrove islands in Florida. Likewise, we can look atevidence that confirms and disconfirms them. Lande argues that the assumptionthat northern spotted owls were at a demographic equilibrium was supported bythe available evidence; Simberloff argued that measures of species turnover ratesdisconfirmed the equilibrium model as applied to the fauna on the Mangrove is-lands. We can argue whether there is sufficient evidence that the assumptions ofthe equilibrium model apply to mountain tops as opposed to islands. One mightalso argue that the equilibrium model explains the species area effect. For thosecase studies the traditional questions are present and philosophical answers to themcan be articulated. We can also ask whether Lande’s work was sufficiently objec-tive given that he did the analysis for "political" purposes. However, when weconsider SCP, things look very different. Consider place prioritization proceduresand multi-criteria analysis. The former considers a set of cells, surrogates, andtargets and tries to solve to find the smallest set of cells such that every surro-gate is met at their target and find those places subject to a constraint for whichcoverage exceeds the target. These are optimization problems to be solved by analgorithm via a computer program. The latter considers a set of alternatives and us-ing a suite of criteria representing our values, which determines a non-dominatedset of alternatives to inform design of conservation networks. This brings toolsfrom economics, policymaking, and operations research to help identify solutionsto problems. In fact, much of the work that is done in systematic conservationplanning is done using software packages like the following:

• Reserve design (Marxan, Zonation, Sites)• Species distribution modeling (Maxent, DesktopGARP, DIVA-GIS, open-

Modeller, BioMapper),• Connectivity (Linkage Mapper, Corridor Designer, CircuitScape, Connec-

tivity Analysis Toolkit, Conefor, UNICOR), threat assessment (GeoCAT,Mirada),

• Population viability analysis (Vortex, BIOMOD, RAMAS),• Multi-criteria analysis (EXPERTCHOICE, MultiCSync).

25

Thus, with regard to SCP, it would make little sense to ask are these programs andtheir associated algorithms "true"? Are they confirmed or disconfirmed? How andwhat do they potentially explain?

Consider what philosopher of science Bas van Fraassen wrote in his The Sci-entific Image,

Science aims to give us, in its theories, a literally true story of what theworld is like; and acceptance of a scientific theory involves the beliefthat it is true. (van Fraassen, 1980, 8)

Of course, he is no scientific realist. For the constructive empiricist,Science aims to give us theories which are empirically adequate; andacceptance of a theory involves as belief only that it is empiricallyadequate. (van Fraassen, 1980, 12)

If SCP is science, them many of its core tools do not concern the aim of scienceas articulated by prominent philosophers of science. If so much of conservationbiology involves developing algorithms and computer programs and articulatingvarious conventions, then truth and empirical adequacy are relevant for parts of thediscipline but irrelevant for large swaths. Moreover, if theory structure and con-firmation are irrelevant to those swaths because they concern truth or empiricaladequacy, then the topic of scientific explanation will be irrelevant as well. Sincethere are few theories and models, then the questions regarding Bayesian versusfrequentist methods of scientific inference will find less purchase in conservationbiology practice. Additionally, and maybe most important, socioeconomic valuesare inputs into conservation biology and it simply cannot be done without them.Or, if it is to be done, it will be the values of small set of scientists thus privilegingthe wrong people (Guha, 1998). If SCP is central to conservation biology, muchof traditional philosophy of science is irrelevant to conservation biology. Conser-vation biology increasingly looks like a pragmatic or instrumental endeavor.

Finally, if we inspect the SCP list from above, we see so many different thingsthat have little to do with biology. The tools includes ones drawn from the naturalsciences, but the social sciences and humanities as well. This raises the questionto what extend is conservation biology biology? Should it simply conservationscience (Kareiva and Marvier, 2011)? It is to issues concerning values that wenow turn.

26

4 What Are the Aims of Conservation Biology?In this section, we will consider three questions. Is conservation biology "value-laden"? What values are to found in conservation biology? Should conservationbiologists be advocates of these values?

Scientific decision-making involves values. Some however have argued thatthe values present are merely epistemic values (Laudan, 1986). For example,Thomas Kuhn argued that scientific theories are evaluated in terms of accuracy,consistency, scope, simplicity, and frutifulness (Kuhn, 1977). Moreover, he con-strued these as values because they are so vague as to require interpretation andthey are many and thus require weighting. If they were merely methodologicalrules, then these would not be problems. Nevertheless, they appear to be epis-temic. Conservation biology serves as an interesting case because it appears thatmoral and sociopolitical values are to be found at its core. That is, they are notonly epistemic.

Let consider an argument from inductive risk (Douglas 2000; Hempel 1965;Rudner 1953, and in the context of conservation biology see Shrader-Frechetteand McCoy, 1993, Ch. 6). In standard Neyman-Pearson hypothesis testing weformulate a null hypothesis H0 and an alternate H1 (Gotelli and Ellison, 2012).The null states that, "Some cause C does not have an effect E." The alternateis the negation, which states that "Some C has an effect E." Thus there are tworelevant probabilities of error,

H0 is true H0 is falseAcceptH0 Correct Type II errorRejectH0 Type I error Correct

Table 1: Type I and Type II errors

We can also represent them as,Pr(RejectingH0|H0 is true) = �Pr(AcceptingH0|H0 is false) = �

Suppose we have the following null and alternate hypotheses.H0: Logging old growth forest will not reduce northern spotted owlpopulations

27

H1: Logging old growth forest will reduce northern spotted owl pop-ulationsThus there are two errors again. We rejectH0 and log old growth forests will notreduce northern spotted owl populations. We accept H0 and logging old growthforests will reduce northern spotted owl populations. For mathematical reasons,we cannot minimize both Type I and Type II errors, and thus we must minimize �or �. By custom, scientists attempt to minimize Type I errors; that is,

RejectH0 if, and only if, Pr(RejectingH0|H0 is true) < 0.05However, this requires that we evaluate which would be worse – allowing loggingthough it reduces the owl population or not allowing logging though it would nothave reduced them anyways? That is, we have a moral or sociopolitical question.Given habitat protection affects both the owls and the livelihoods of loggers livingin the Pacific Northwest, this is not merely a technical scientific question. Somehave argued that when environmental harms are at stake, we should minimize TypeII errors (Shrader-Frechette and McCoy, 1993). This mode of reasoning is some-times called the "precautionary principle" (Steel, 2014).

Neyman-Pearson statistical hypothesis testing is a very common methodol-ogy taught in statistics courses and used by practicing biologists. However, somephilosophers and statisticians reject it for an alternative, Bayesianism (Howson andUrbach, 2006). According to Bayesianism, we neither accept nor reject hypothe-ses (though one can work out a notion of acceptance in the framework, see Levi1974; Maher 1993). Rather, we compare the probability of hypothesis with regardto some evidence and determine how it affects its probability. That is, we comparethe prior probability of the hypothesis Pr(H) to its posterior probability Pr(H|E)using Bayes’ theorem. In its simplest form, it says,

Pr(H|E) =(Pr(H) × Pr(E|H)

Pr(E)In addition to the prior and posterior probabilities, we also have Pr(E) is the prob-ability of the evidence, and Pr(E|H) is the probability of the evidence E giventhe hypothesis. Put very simply, if Pr(H|E) > Pr(H), then E confirms H and ifPr(H|E) < Pr(H), the E disconfirms H . We then update probabilities througha rule of conditionalization Prnew(H) = Prold(H|E). Whether the problem ofinductive risk disappears in a Bayesian inferential frameworks is an interestingquestion. Insofar as credences (i.e. subjective degrees of belief) are affected byattitudes towards risk, then arguably the problem arises in a different guise.

28

Whether one finds the argument from inductive risk persuasive, it is clear thatethical and sociopolitical values are brought into the science itself by various scien-tists. Conservation biologists such as Paul Ehrlich, E. O.Wilson, Thomas Lovejoy,etc. often advocated ethical values as scientists in order to raise awareness regard-ing ecological degradation and its effects on species including our own (Takacs,1996). Supposing that there are epistemic and moral values in conservation biol-ogy, what are the specific ones present? One way to taxonomize those values isthrough Helen Longino’s work (Longino, 1990). She draws a distinction betweenconstitutive and contextual values in science. Constitutive values are those that areessential to the scientific inquiry. For example, if Kuhn is correct that accuracy,simplicity, and scope are values and they are essential to scientific methodology,then these would be constitutive values. You simply cannot do science withoutthem. Contextual values are those that accidental to science; one can do sciencewithout endorsing them. Generally, they enter scientific practice through scien-tists, scientific research groups, or from culture or society more generally.

Some of these contextual values arise from individuals because of their owncommitments. E. O. Wilson has written a great deal arguing that we should workagainst anthropogenic species extinction for example (Wilson, 1999). Group con-textual values include would include codes of ethics. For example, the Society forConservation Biology has a Code of Ethics. It includes obligations like,

Actively disseminate information to promote understanding of and ap-preciation for biodiversity and the science of conservation biology.Recognize that uncertainty is inherent in managing ecosystems andspecies and encourage application of the precautionary principle inmanagement and policy decisions affecting biodiversity.

Though one might argue about specific obligations included in the Code of Ethics,generally we the public agrees that one should act with integrity in one’s researchand treat non-human animals with respect. Still, there are more controversial so-cietal values that play a part in conservation biology. Conservation biologist ReedNoss writes,

Conservation biology has been described throughout its history as"value-laden," "mission-oriented," "normative," and sometimes in lessflattering terms. The entire field rests on the value assumption thatbiodiversity is good and ought to be conserved. Human actions that

29

protect and restore biodiversity are good; those that destroy or degradebiodiversity are bad. (Noss, 2007, 18)

Recently, many human lives have been lost to the virus Zaire ebolavirus in thegenus Ebolavirus. To exterminate this virus or the genus one might argue is notmorally wrong but morally permissible if not obligatory. Noss’ claim is that thisis morally wrong, which many would find controversial.

Consider the claim that biodiversity has intrinsic value. Should conservationbiologists accept this? Michael Soulé famously claimed that conservation biologyincluded a variety of "normative postulates" including: diversity of organisms isgood, ecological complexity is good, evolution is good, and biotic diversity hasintrinsic value (Soulé, 1985, 730-1). He writes,

This normative postulate is the most fundamental. In emphasizingthe inherent value of nonhuman life, it distinguishes the dualistic, ex-ploitive world view from a more unitary perspective: Species havevalue in themselves. a value neither conferred nor revocable, but spring-ing from a species’ long evolutionary heritage and potential or evenfrom the mere fact of its existence. (Soulé, 1985, 731)

First, what does it mean to say biodiversity has intrinsic value (Vucetich et al.,2015)? Ethicists debate what is intrinsic value and whether there is such a "thing"(O’Neill, 1992). For example, is intrinsic value something has in virtue of itsintrinsic properties (Moore, 1993)? If extrinsic properties are irrelevant to intrinsicvalue and beauty is a relational property, then it is irrelevant to intrinsic value. Isintrinsic value what we value for its own sake? But this is says nothing aboutwhat the attitude of valuing is (Ewing, 2012; Lewis, 1989). Is intrinsic value thatwhich would have value even if there were no valuers? This appears to assume astrong form of moral realism about value, which some metaethicists would deny(Blackburn, 1993; Gibbard, 1992).

Second, conservation biologists themselves do not agree on the matter. Hereare representative quotations from conservation biologists on the subject.

David Ehrenfeld: "For biological diversity, value is. Nothing moreand nothing less... Well, I couldn’t prove it, I guess. I just believe it."Paul Ehrlich: "...I just can’t have the feeling that the only value they[species] might have is what they might mean to us. But you can’t

30

possibly defend that scientifically."Jerry Franklin: "Oh, I basically think so, yes. But I haven’t given awhole lot of thought to it."Daniel Janzen: "The word value is anthropocentric... That’s a contra-diction in terms."S. J. MacNaughton: "I don’t see how anything can have value outsideof a value that human beings place on it, because value is really some-thing uniquely human, isn’t it?"David Pimmentel: "[I]n trying to protect or conserve nature, to usethe argument of intrinsic value gets you – well, I don’t think it sellsvery well." (Takacs, 1996, 249-52)

If prominent conservation biologists disagree whether biodiversity has intrinsicvalue and what it means to claim it, one might ask how it could be axiomatic inthe first place.

Recently, Soulé’s vision has been called into question by the "new conserva-tionists." Peter Kareiva and Michelle Marvier have argued for a different founda-tion for conservation biology or what they call "conservation science" (Kareivaand Marvier, 2012). They argue that the science needs a new normative directionfor a variety of reasons. First, conservation biology has traditionally focused onconserving natural systems for the intrinsic value of these systems. However, theycontend ecological systems can only be successfully managed by embracing theinstrumental values provided to humans by these systems. Traditional conserva-tion biology they allege has been at the expense of third world stakeholders andcontrary to the interests of women. Second, we live on planet that has been deeplyinfluenced by humans; so much so that many suggest we live in a new geologicalepoch, the Anthropocene. Human population stands at approximately 7 billion andour energy consumption has only increased over time. Humans have deforested,urbanized, polluted, extirpated, and overharvested to such a degree that there isno "pristine" nature available. "Fortress conservation" is an out-of-date modelof what we should and can do. Moreover, environmental attitudes have changedover time – we value environmental goods and services less and less in the UnitedStates. Thus, we must focus on what people care about and focus on motivatingthem through these valuations. Third, ecological systems are not as fragile as tra-

31

ditional ecology suggests. Rather, they are remarkably resilient. Marine systemsovercome oilspills, deforested areas rebound, bird species dwindle but are rarelyeliminatedwhole cloth, and coral reefs can even come back from a hydrogen bomb.Fourth, we can avoid tragedy of the commons, which has been a major theoreticalstructure for interpreting environmental decision-making. Following the work ofElinor Ostrom, local buy-in can eliminate the self-interested behavior that leads tosuch zero-sum behavior (Ostrom, 1990). To move forward, they recommend werecognize that conservation can only occur within human-affected landscapes, wemust work with corporations and not merely oppose them, and avoid threateningthe human rights of the disenfranchised.

As onewould expect, the new conservationists have been forcefully challenged.For example, David Doak, Victoria Bakker, Bruce Goldstein, and Ben Hale havechallenged their claims across the board (Doak et al., 2014). First, they contendthat Kareiva andMarvier focus on human well-being in the most narrow economicterms. Moreover, conservation has never been solely for the "more-than-humanworld" but has a rich tradition of managing for human interests as well. Second,there are many relatively undisturbed ecological systems that exist, and others havesuffered irreversible effects by humans (e.g. anthropogenic global climate change).Third, they suggest that a narrowly anthropocentric approach to conservation haslittle by way of evidence of success. And, traditional approaches have much thattestifies in their favor. Additionally, though some conservation projects have dis-placed indigenous peoples, conservation organizations have been working hardover the last few decades to try to avoid these environmental injustices. Fourth,they contest the social science that claims environmental attitudes have fundamen-tally changed and notes that actors are rarely self-motivated and apathetic in theways the new conservationists suggest. Finally, if one were solely committed tohuman well-being with little regard to the moral standing of ecological systems,then we should be investing not in the Nature Conservancy but Save the Children,Oxfam, and Water for the People.

Regardless, of how we view these contemporary debates, it is clear that thereare constitutive and contextual values present in conservation biology and any suc-cessor discipline moving forward. These include the relatively uncontroversial re-garding inductive risk and SCB’s Code of Ethics to those regarding the intrinsicvalue of organisms, species, and ecosystems. Furthermore, there are very impor-tant questions regarding how these values, non-controversial and controversial,should be advocated in the public sphere. If the values are broadly shared by sci-entist and non-scientist alike, then it appears that advocacy is simply on the public’sbehalf. However, as the values become specific to the scientists themselves, their

32

advocacy can appear to be in tension with regard to their objectivity as scientists.They appear as "tree huggers" and "environmentalists" and less than impartial.Moreover, if their advocacy detracts from their role as scientists, they one mightworry that this all things considered renders them impotent in policymaking. Theyseem to be just another "talking head" (Odenbaugh, 2003). Thus, conservation bi-ologists must discern how to balance the values at work in their discipline and thegreater good to which they are committed.

Some policy analysts argue that if scientific work is to impact policy, it mustbe objective (Pielke, 2007). However, advocating personal contextual values isalleged to undercut that objectivity (Burke and Lauenroth 2009, but see Strong2008). If one claims that biodiversity is intrinsically valuable, this value isn’twidely shared, and it cannot be reasonably defended by its proponents, then con-servation biologists risk their work being ignored or dismissed. On the other hand,if one advocates the conservation of biodiversity on the basis of ecosystem servicesincluding carbon sequestration, waste decomposition, purification of air and wa-ter, pollination, etc., and loss of these services would negatively impact humans,then this advocacy would be far less controversial (Brussard and Tull, 2007). Inaddition, one might reconsider how we think about science itself. No one regardsmedicine as less objective because it advocates on behalf of patients. We recog-nize that the enterprise is not "value-neutral." Should the same thing be said ofconservation biology? How might we educate the public to think more carefullyand clearly about the role of values in science?

5 ConclusionIn this entry, we have consider three philosophical issues in conservation biology.First, we began with the ontology of the discipline – what is biodiversity? Weexamined three proposals: biodiversity should be multidimensional, it should beoperationalized, and that the concept should be eliminated from conservation prac-tice. Second, we considered several different important cases studies in the science– modeling northern spotted owls, the equilibrium model of island biogeographyand its impact on the SLOSS debate, and systematic conservation planning. Whatis striking is how traditional philosophical question appear to disappear or be re-oriented in recent work. That is, though there are theories and models, the coretheoretical structures are different from other sciences since it involves variousconventions and algorithms. Third, we consider the axiological foundations ofthe discipline. Many conclude that the sciences involve considerations of value.

33

However, conservation biology was thought to be founded on the intrinsic valueof ecological systems though the recent new conservationists have challenged thisfoundation as resting on soft sand. Critics of course have come out in full forcearguing this critique misses its mark and the foundations, though apparently shaky,are only apparently so.

ReferencesArrow, K. J. and H. Raynaud (1986). Social choice and multicriterion decision-making. The MIT Press.

Arthur, W. (2000). The origin of animal body plans: A study in evolutionarydevelopmental biology. Cambridge University Press.

Avise, J. C., J. L. Hamrick, et al. (1996). Conservation genetics: case historiesfrom nature. Chapman & Hall.

Ayer, A. (1999). What is a law of nature? Laws of Nature, Causation, and Super-venience 1, 52.

Baum, D. A. and S. D. Smith (2013). Tree thinking: an introduction to phyloge-netic biology. Roberts.

Beatty, J. (1995). The evolutionary contingency thesis. Concepts, theories, andrationality in the biological sciences, 45–81.

Beissinger, S. R. and D. R. McCullough (2002). Population viability analysis.University of Chicago Press.

Blackburn, S. (1993). Essays in quasi-realism. Oxford University Press.Bolten, A. B. and B. E.Witherington (2003). Loggerhead sea turtles. SmithsonianBooks Washington, DC.

Brown, J. H. (1971). Mammals onmountaintops: nonequilibrium insular biogeog-raphy. American Naturalist, 467–478.

Brussard, P. F. and J. C. Tull (2007). Conservation biology and four types ofadvocacy. Conservation Biology 21(1), 21–24.

34

Burke, I. and B. Lauenroth (2009). ?environmentalist? label not in our best inter-ests. Frontiers in Ecology and the Environment 7(5), 240–240.

Carnap, R. (1962). Logical foundations of probability.Cartwright, N. (1983). How the laws of physics lie. Cambridge Univ Press.Cartwright, N. (1999). The dappled world: A study of the boundaries of science.Cambridge University Press.

Chakravartty, A. (2001). The semantic or model-theoretic view of theories andscientific realism. Synthese 127(3), 325–345.

Chang, H. (2004). Inventing temperature: Measurement and scientific progress.Oxford University Press.

Clements, F. E. (1916). Plant succession: an analysis of the development of veg-etation. Number 242. Carnegie Institution of Washington.

Colyvan, M., J. Justus, and H. M. Regan (2011). The conservation game. Biolog-ical Conservation 144(4), 1246–1253.

Diamond, J. (1975a). Assembly of species communities. á in: Cody, ml and dia-mond, jm (eds), ecology and evolution of communities.

Diamond, J. M. (1975b). The island dilemma: lessons of modern biogeographicstudies for the design of natural reserves. Biological conservation 7(2), 129–146.

Doak, D. F., V. J. Bakker, B. E. Goldstein, and B. Hale (2014). What is the futureof conservation? Trends in ecology & evolution 29(2), 77–81.

Douglas, H. (2000). Inductive risk and values in science. Philosophy of Science,559–579.

Dretske, F. I. (1977). Laws of nature. Philosophy of Science, 248–268.Durbin, K. (1996). Tree huggers: Victory, defeat & renewal in the northwestancient forest campaign. Seattle, WA: Mountaineers.

Egerton, F. N. (1973). Changing concepts of the balance of nature. Quarterlyreview of biology, 322–350.

35

Eliot, C. (2011). The legend of order and chaos: Communities and early commu-nity ecology. Handbook of the philosophy of ecology, 49–108.

Elton, C. S. (2000). The ecology of invasions by animals and plants. Universityof Chicago Press.

Ereshefsky, M. (2000). The poverty of the Linnaean hierarchy: A philosophicalstudy of biological taxonomy. Cambridge University Press.

Ewing, A. C. (2012). The Definition of Good (Routledge Revivals). Routledge.Faith, D. P. (1992). Conservation evaluation and phylogenetic diversity. Biologicalconservation 61(1), 1–10.

Faith, D. P. (1994). Phylogenetic pattern and the quantification of organismalbiodiversity. Philosophical Transactions of the Royal Society of London. SeriesB: Biological Sciences 345(1311), 45–58.

Faith, D. P. (2002). Quantifying biodiversity: a phylogenetic perspective. Conser-vation Biology 16(1), 248–252.

Foley, P. (1994). Predicting extinction times from environmental stochasticity andcarrying capacity. Conservation Biology 8(1), 124–137.

Foley, P. (1997). Extinctionmodels for local populations.Metapopulation biology:ecology genetics and evolution, 215–246.

Frank, D. M. and S. Sarkar (2010). Group decisions in biodiversity conservation:implications from game theory. PloS one 5(5), e10688.

Gibbard, A. (1992). Wise choices, apt feelings: A theory of normative judgment.Harvard University Press.

Giere, R. N. (2010). Explaining science: A cognitive approach. University ofChicago Press.

Gleason, H. A. (1926). The individualistic concept of the plant association. Bul-letin of the Torrey Botanical Club, 7–26.

Gotelli, N. J. and A. M. Ellison (2012). A Primer of Ecological Statistics. SinauerAssociates, Inc.

36

Gould, S. J. (2000). Wonderful life: the Burgess Shale and the nature of history.Random House.

Groom, M. J., G. K. Meffe, C. R. Carroll, et al. (2006). Principles of conservationbiology. Sinauer Associates Sunderland.

Guha, R. (1998). Radical american environmentalism and wilderness preservation(1989). The great new wilderness debate, 231.

Hanski, I. and M. E. Gilpin (1997). Metapopulation biology: ecology, genetics,and evolution, Volume 1. Academic press San Diego.

Hempel, C. (1965). Science and human values. InAspects of Scientific Explanationand Other Essays in the Philosophy of Science, pp. 81–96. The Free Press.

Hempel, C. G. (1966). Philosophy of natural science.Howson, C. and P. Urbach (2006). Scientific reasoning: the Bayesian approach.Open Court Publishing.

Hughes, R. I. (1997). Models and representation. Philosophy of Science, S325–S336.

Hutchings, P. A., M. Kingsford, and O. Hoegh-Guldberg (2008). The Great Bar-rier Reef: biology, environment and management, Volume 2. Csiro publishing.

Justus, J. (2008). Complexity, diversity, and stability. A companion to the philos-ophy of biology, 321–350.

Justus, J. (2010). The diversities of biodiversity. Metascience 19(2), 247–250.Kareiva, P. and M. Marvier (2012). What is conservation science? Bio-Science 62(11), 962–969.

Kareiva, P. M. andM.Marvier (2011). Conservation science: balancing the needsof people and nature. Roberts and Company Greenwood Village, Colorado,USA.

Kuhn, T. S. (1977). The Essential Tension: Selected Studies in Scientific Teaditionand Change. University of Chicago Press.

Kuhn, T. S. (2012). The structure of scientific revolutions. University of Chicagopress.

37

Lakatos, I. (1976). Falsification and the methodology of scientific research pro-grammes. Springer.

Lande, R. (1988a). Demographic models of the northern spotted owl (strix occi-dentalis caurina). Oecologia 75(4), 601–607.

Lande, R. (1988b). Genetics and demography in biological conservation. Sci-ence 241(4872), 1455–1460.

Laudan, L. (1978). Progress and its problems: Towards a theory of scientificgrowth, Volume 282. Univ of California Press.

Laudan, L. (1986). Science and values: The aims of science and their role inscientific debate. Univ of California Press.

Levi, I. (1974). Gambling with truth: An essay on induction and the aims ofscience.

Levins, R. (1970). Extinction. Lectures on mathematics in the life sciences 2,75–107.

Lewis, D. (1983). New work for a theory of universals. Australasian Journal ofPhilosophy 61(4), 343–377.

Lewis, D. (1989). Dispositional theories of value. Proceedings of the AristotelianSociety 63, 113?37.

Longino, H. E. (1990). Science as social knowledge: Values and objectivity inscientific inquiry. Princeton University Press.

MacArthur, R. (1955). Fluctuations of animal populations and a measure of com-munity stability. ecology 36(3), 533–536.

MacArthur, R. H. and E. O. Wilson (1963). An equilibrium theory of insularzoogeography. Evolution, 373–387.

MacArthur, R. H. and E. O. Wilson (1967). The theory of island biogeography,Volume 1. Princeton University Press.

Maclaurin, J. and K. Sterelny (2008). What is biodiversity? University of ChicagoPress.

38

Magurran, A. E. (1988). Ecological diversity and its measurement, Volume 168.Princeton university press Princeton.

Magurran, A. E. (2013). Measuring biological diversity. John Wiley & Sons.Maher, P. (1993). Betting on theories. Cambridge University Press.Margules, C., A. Higgs, and R. Rafe (1982). Modern biogeographic theory: arethere any lessons for nature reserve design? Biological Conservation 24(2),115–128.

Margules, C. and S. Sarkar (2007). Systematic conservation planning. CambridgeUniversity Press.

Margules, C. R. and R. L. Pressey (2000). Systematic conservation planning. Na-ture 405(6783), 243–253.

May, R. M. (1975). Islands biogeography and the design of wildlife preserves.Nature 254, 177–178.

May, R. M. (2001). Stability and complexity in model ecosystems, Volume 6.Princeton University Press.

Moffett, A. and S. Sarkar (2006). Incorporating multiple criteria into the designof conservation area networks: a minireview with recommendations. Diversityand Distributions 12(2), 125–137.

Mooers, A. Ø. (2007). Conservation biology: The diversity of biodiversity. Na-ture 445(7129), 717–718.

Moore, G. E. (1993). Principia ethica. Cambridge University Press.Morgan, G. J. (2009). Themany dimensions of biodiversity. Studies in History andPhilosophy of Science Part C: Studies in History and Philosophy of Biologicaland Biomedical Sciences 40(3), 235–238.

Morin, P. J. (2011). Community Ecology. John Wiley & Sons.Norse, E. A. (1989). Ancient forests of the Pacific Northwest. Island Press.Noss, R. (2007). Values are a good thing in conservation biology. ConservationBiology 21(1), 18–20.

39

Odenbaugh, J. (2003). Values, advocacy and conservation biology. EnvironmentalValues 12(1), 55–69.

Odenbaugh, J. (2005). Idealized, inaccurate but successful: A pragmatic approachto evaluating models in theoretical ecology. Biology and Philosophy 20(2-3),231–255.

Odenbaugh, J. (2007). Seeing the forest and the trees: realism about communitiesand ecosystems. Philosophy of Science 74(5), 628–641.

Odenbaugh, J. (2009). Sahotra sarkar, biodiversity and environmental philosophy:An introduction. Biology & Philosophy 24(4), 541–550.

Odenbaugh, J. (2014). Semblance or similarity? reflections on simulation andsimilarity. Biology & Philosophy, 1–15.

O’Neill, J. (1992). The varieties of intrinsic value. The Monist, 119–137.Ostrom, E. (1990). Governing the commons: The evolution of institutions forcollective action. Cambridge university press.

Otto, S. P. and T. Day (2007). A biologist’s guide to mathematical modeling inecology and evolution, Volume 13. Princeton University Press.

Pielke, R. A. (2007). The honest broker: making sense of science in policy andpolitics. Cambridge Univ Press.

Pimm, S. L. (1984). The complexity and stability of ecosystems. Na-ture 307(5949), 321–326.

Quinn, T. P. (2011). The behavior and ecology of Pacific salmon and trout. UBCPress.

Raup, D. M. (1966). Geometric analysis of shell coiling: general problems. Jour-nal of Paleontology, 1178–1190.

Ridley, M. (1989). The cladistic solution to the species problem. Biology andPhilosophy 4(1), 1–16.

Rudner, R. (1953). The scientist qua scientist makes value judgments. Philosophyof Science 20(1), 1–6.

40

Santana, C. (2014). Save the planet: eliminate biodiversity. Biology & Philoso-phy 29(6), 761–780.

Sarkar, S. (2002). Defining ?biodiversity?; assessing biodiversity. TheMonist 85(1), 131–155.

Sarkar, S. (2005). Biodiversity and environmental philosophy: An introduction.Cambridge University Press.

Sarkar, S. (2012). Environmental philosophy: From theory to practice. JohnWiley& Sons.

Sarkar, S., A. Aggarwal, J. Garson, C. R. Margules, and J. Zeidler (2002). Placeprioritization for biodiversity content. Journal of Biosciences 27(4), 339–346.

Sarkar, S. and C. Margules (2002). Operationalizing biodiversity for conservationplanning. Journal of biosciences 27(4), 299–308.

Shrader-Frechette, K. S. and E. McCoy (1993). Method in ecology: strategies forconservation. Cambridge University Press.

Simberloff, D. (1976). Species turnover and equilibrium island biogeography. Sci-ence 194(4265), 572–578.

Simberloff, D. S. and L. G. Abele (1976). Island biogeography theory and conser-vation practice. Science 191(4224), 285–286.

Simberloff, D. S. and E. O.Wilson (1970). Experimental zoogeography of islands.a two-year record of colonization. Ecology, 934–937.

Soulé, M. E. (1985). What is conservation biology? BioScience, 727–734.Steel, D. (2014). Philosophy and the precautionary principle: Science, evidence,and environmental policy. Cambridge University Press.

Sterelny, K. (2006). Local ecological communities*. Philosophy of Science 73(2),215–231.

Strong, D. R. (2008). Ecologists and environmentalism. Frontiers in Ecology andthe Environment 6(7), 347–347.

Takacs, D. (1996). The idea of biodiversity. Johns Hopkins University Press.

41

Terborgh, J. (1975). Faunal equilibria and the design of wildlife preserves.Springer.

Tilman, D., D. Wedin, J. Knops, et al. (1996). Productivity and sustainabilityinfluenced by biodiversity in grassland ecosystems. Nature 379(6567), 718–720.

van Fraassen, B. C. (1980). The scientific image. Oxford University Press.van Fraassen, B. C. (1989). Laws and symmetry. Oxford: Clarendon.Vucetich, J. A., J. T. Bruskotter, and M. P. Nelson (2015). Evaluating whether na-ture’s intrinsic value is an axiom of or anathema to conservation. ConservationBiology 29(2), 321–332.

Wagner, P. J. (1995). Diversity patterns among early gastropods: contrasting tax-onomic and phylogenetic descriptions. Paleobiology, 410–439.

Watson, J. E., H. Grantham, K. A. Wilson, and H. P. Possingham (2011). System-atic conservation planning: Past, present and future. Wiley-Blackwell, 136–160.

Weisberg, M. (2012). Simulation and similarity: using models to understand theworld. Oxford University Press.

Whittaker, R. H. (1967). Gradient analysis of vegetation*. Biological Re-views 42(2), 207–264.

Whittaker, R. J. and J. M. Fernández-Palacios (2007). Island biogeography: ecol-ogy, evolution, and conservation. Oxford University Press.

Wiley, E. O. (1978). The evolutionary species concept reconsidered. SystematicBiology 27(1), 17–26.

Wilson, E. O. (1988). Biodiversity. National Academy of Sciences SmithsonianInstitution.

Wilson, E. O. (1999). The diversity of life. WW Norton & Company.Wilson, E. O. and W. H. Bossert (1971). A primer of population biology. SinauerAssociates Sunderland.

Wilson, E. O. and E. O. Willis (1975). Applied biogeography. Ecology and evo-lution of communities, 522–534.

42

Wimsatt, W. C. (2012). Robustness, reliability, and overdetermination (1981). InCharacterizing the Robustness of Science, pp. 61–87. Springer.

Yaffee, S. L. (1994). The wisdom of the spotted owl: policy lessons for a newcentury. Island Press.

43