PI
a,b,
?
c R. Dudash d, Mark D.B. Eldridge b, Charles B. Fenster d,
Australian Museum, 6 College Street, Sydney, NSW 2010, Australia
cCenter for Conservation and Evolutionary Genetics, Smithsonian Conservation Biology Institute, Washington, DC 20008, USA
dDepartment of Biology, University of Maryland, College
eChicago Zoological Society, Brook?eld, IL 60513, USA
f Zoo Atlanta, 800 Cherokee Ave., SE Atlanta, GA 30315,
g School of Biology, Georgia Institute of Technology, 301
h Saint Louis Zoo, One Government Drive, St. Louis, MO
i San Diego Zoo Institute for Conservation Research, 156
4.2. Empirical data on rapid attainment of diagnosable differences between populations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Abbreviations: BSC, biological species concept; CE, critically endangered; E, endangered; ESC, evolutionary species concept; DFSC, differential ?tness species concept; ID,
inbreeding depression; OD, outbreeding depression; PSC, phylogenetic species concept; TSC, taxonomic species concept; V, vulnerable.
? Corresponding author at: Department of Biological Sciences, Macquarie University, NSW 2109, Australia. Tel.: +61 2 9850 8186; fax: +61 2 9850 8245.
E-mail addresses: richard.frankham@mq.edu.au (R. Frankham), ballouj@si.edu (J.D. Ballou), mdudash@umd.edu (M.R. Dudash), mark.eldridge@austmus.gov.au (M.D.B.
Eldridge), cfenster@umd.edu (C.B. Fenster), rlacy@ix.netcom.com (R.C. Lacy), jmendelson@zooatlanta.org (J.R. Mendelson III), Porton@stlzoo.org (I.J. Porton), rallsk@the-
Biological Conservation 153 (2012) 25?31
Contents lists available at SciVerse ScienceDirect
Biological Conservationgrid.net (K. Ralls), oryder@sandiegozoo.org (O.A. Ryder).Outbreeding depression
Species concepts ulations and/or ?xed chromosomal differences) are required to satisfy conservation issues. Species delin-
eations based upon the biological and differential ?tness species concepts meet the above requirements.
Conversely, if species are delineated using the diagnostic phylogenetic species concept, genetic rescue of
small genetically isolated populations may require crosses between species, with consequent legal and
regulatory rami?cations that could preclude actions to prevent extinction. Consequently, we conclude
that the diagnostic phylogenetic species concept is unsuitable for use in conservation contexts, especially
for classifying allopatric populations.
 2012 Elsevier Ltd. All rights reserved.
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2. Minimizing harm and maximizing potential conservation benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3. How do excessively broad species delineations occur? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4. How does excessive splitting of small populations occur? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.1. Theory predicting generations to attain reciprocal monophyly or no shared alleles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28a r t i c l e i n f o
Article history:
Received 16 January 2012
Received in revised form 17 April 2012
Accepted 28 April 2012
Available online 29 June 2012
Keywords:
Fragmented populations
Genetic rescue
Inbreeding depression
Loss of genetic diversity0006-3207/$ - see front matter  2012 Elsevier Ltd. A
http://dx.doi.org/10.1016/j.biocon.2012.04.034Park, MD 20742, USA
USA
Ferst Dr, Atlanta, GA 30332, USA
63110, USA
00 San Pasqual Valley Road, Escondido, CA 92027, USA
a b s t r a c t
The 26 de?nitions of species often yield different numbers of species and disparate groupings, with
?nancial, legal, biological and conservation implications. Using conservation genetic considerations, we
demonstrate that different species concepts have a critical bearing on our ability to conserve species.
Many species of animals and plants persist as small isolated populations suffering inbreeding depression,
loss of genetic diversity, and elevated extinction risks. Such small populations usually can be rescued by
restoring gene ?ow, but substantial genetic drift effects can lead them to be classi?ed as distinct species
under the diagnostic phylogenetic species concept. Minimum harm to ?tness is done and maximum
potential ?tness and evolutionary potential bene?ts accrue when reproductive isolation (pre- and/or
post-zygotic) is used as the criterion to de?ne distinct species. For sympatric populations, distinct species
are diagnosed by very limited gene ?ow. For allopatric populations, both minimal gene ?ow and evidence
of reduced reproductive ?tness in crosses (or effects predicted from adaptive differentiation among pop-Robert C. Lacy e, Joseph R. Mendelson III f,g, Ingrid J. Porton h, Katherine Ralls c, Oliver A. Ryder i
aDepartment of Biological Sciences, Macquarie University, NSW 2109, Australia
bRichard Frankham , Jonathan D. Ballou , Micheleerspective
mplications of different species concepts for conserving biodiversityjournal homepage: www.elsevier .com/ locate /bioconll rights reserved.
. . .
. . .
. . .
. . .
p
th
2
Conet al., 2011; Sexton et al., 2011; see Supplementary material).
The incongruities indicate that no species concept is without
roblems. Points 1 and 2 cause severe dif?culties for the BSC; fur-
2. Minimizing harm and maximizing potential conservation
bene?ts
The ideal species concept for conservation purposes would min-and/or ?xed chromosomal differences (reviewed by Frankhamusually accompanies genetic adaption to different environ-
ments (via natural selection, as proposed by Darwin (1859)),
cepts that are most bene?cial for conserving global biodiversity.6. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix A. Supplementary material . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction
There are at least 26 de?nitions of biological species (see Wil-
kins, 2009; Hausdorf, 2011). Use of different species concepts to
classify species has potential ?nancial, legal, biological and conser-
vation implications (Hey et al., 2003). It leads to much confusion
and controversy, and to potential problems of inappropriate delin-
eation of species for conservation purposes. Mace (2004) con-
cluded that ?taxonomists and conservationists need to work
together to design some explicit rules to delimit the units included
as species for the purposes of conservation planning and assess-
ment.? Thus, there is an urgent need to evaluate the suitability of
the different species concepts for conservation purposes.
The three concepts most widely used by the systematic and
conservation communities are the biological species concept
(BSC; Mayr, 1942, 1963), the evolutionary species concept (ESC;
Simpson, 1951, 1961; Wiley, 1978) and the phylogenetic species
concept (PSC; Eldredge and Cracraft, 1980; Cracraft, 1997), as de-
?ned in Table 1. We also discuss the recently proposed differential
?tness species concept (DFSC; Hausdorf, 2011), as it is highly rele-
vant to conservation. This concept is most similar to the BSC, but
BSC uses mating isolation and/or sterility to delineate species
while DFSC is broader, encompassing any pre- or post-zygotic ?t-
ness decrement following attempted crossing.
An alternative to the use of de?ned species concepts is to rely
upon the judgment of taxonomists, sometimes referred to as the
taxonomic species concept (TSC; Mayden, 1997). This corresponds
to the de?nition that species are ?whatever a competent taxono-
mist chooses to call a species? (Wilkins, 2009). This seems to be
widely practised, as papers on new species delineations, or revi-
sions usually fail to specify what species concept has been used
(see McDade, 1995).
As the literature on species concepts is massive, we can only re-
fer to a sample of references. We favored key references, reviews,
recent publications, and studies addressing conservation concerns.
All of the commonly used species concepts suffer from incon-
gruencies with biological reality (Hausdorf, 2011), namely:
1. ?reproductive barriers are often semipermeable to gene ?ow?
(Hey and Pinho, 2012);
2. ?species can differentiate despite ongoing inter-breeding? (sym-
patric speciation; Papadopulos et al., 2011);
3. ?parallel speciation can occur due to parallel adaptation or
recurrent polyploidizations?, and
4. ?uniparental organisms are actually organized in units that
resemble species of biparental organisms?;
In addition, we conclude that:
5. development of reproductive isolation between populations
26 R. Frankham et al. / Biologicalermore, it does not apply to asexual organisms. For PSC, Points 1,
and 3 cause dif?culties. Point 3 may cause dif?culties for ESC, butESC copes with the other points. Point 5 partly counters some of
the problems, as it makes it feasible to predict reproductive isola-
tions for diagnosing species under BSC and DFSC (see below)
(Frankham et al., 2011). A serious concern with PSC is that techno-
logical advances (e.g. those lessening DNA sequencing costs) and
increased effort lead to increased resolution among lineages, such
that even individuals within populations can be diagnosably differ-
ent (Avise and Ball, 1990; Groves, 2004; Winkler, 2010).
Despite the disparate de?nitions, species concepts typically
indicate that species are cohesive clusters of individuals that have
at least partially different evolutionary paths representing differ-
ent lineages (see Avise and Ball, 1990; Knowlton and Weigt,
1997; de Queiroz, 1998; Hey et al., 2003; Coyne and Orr, 2004;
Hausdorf, 2011). The differences among concepts are typically in
how far evolutionary population differentiation needs to proceed
before the populations should be considered distinct species. All
serious concepts recognise that populations inherently incapable
of gene exchange are distinct species, while those exhibiting ran-
dom mating in sympatry are conspeci?c. However, there are major
differences in the treatment of partly diverged allopatric popula-
tions capable of gene ?ow without adverse ?tness consequences,
or with bene?cial consequences. In allopatric populations, espe-
cially those with small population sizes, genetic drift and mutation
will lead to diagnosably different units that are not intrinsically
reproductively isolated (see below) that may be ephemeral under
natural patterns of population separation and re-connection.
De?ning such units as species for conservation purposes may
accelerate extinction of broader BSC species rather than preserve
adaptive differences (see below).
Scientists working in different disciplines or on disparate major
taxa often favor alternative species concepts (Claridge et al., 1997).
For example, evolutionary geneticists generally favor BSC (see
Noor, 2002; Coyne and Orr, 2004) because it relates to the ?tness
consequences of gene ?ow between populations and the process
of speciation. In contrast, some taxonomists now favor PSC (Cra-
craft, 1997; Groves, 2004), because it is considered easier to imple-
ment. Use of PSC results in more splitting: it yielded 49% more
species than BSC on the same group of organisms (Agapow et al.,
2004). In some cases, the groupings according to BSC and PSC were
discordant, with PSC species not nested within BSC species, or vice
versa. Such inconsistencies can often lead to different manage-
ment, some resulting in adverse consequences for conservation
of biodiversity.
We evaluate methods for de?ning species from the perspective
of conservation biology, advocating that de?nitions used in conser-
vation biology should maximize conservation bene?ts. We show,
from a population genetics perspective, that current methods for
species? delineation often lead to species? classi?cation that is too
narrow, or too broad, both of which can compromise the conserva-
tion of the taxon?s biodiversity. We then recommend use of con-. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
servation 153 (2012) 25?31imize potential harm and maximize potential bene?ts, as mea-
sured by reproductive ?tness and sustaining adaptive
evolutionary processes. Loss and fragmentation of habitat stem-
ming from human population growth is one of the most severe
threats to biodiversity (Millennium Ecosystem Assessment,
2005). Fragmentation of populations that were once widely dis-
tributed results in small, isolated populations potentially subject
to loss of genetic diversity, inbreeding depression and increased
risk of extinction (see Allendorf and Luikart, 2006; Frankham
et al., 2010). Conservation of these populations often requires re-
of large genetic drift effects in small populations (see below), clas-
si?es the small a4 population as a distinct species, without any
populations within its species that can be used to rescue it genet-
ically or reinforce it demographically. This means that splitting,
sometimes in an attempt to promote greater conservation of biodi-
versity, can actually prevent conservation actions necessary to pre-
?ts, species de?nitions and delineations need to de?ne and identify
Table 1
Species de?nition according to different species concepts.
Species concept Species de?nition Reference
Biological (BSC) ?Groups of actually or potentially interbreeding natural populations, which are reproductively isolated from other such groups? Mayr
(1942)
Evolutionary (ESC) ?A species is a lineage of ancestral descent which maintains its identity from other such lineages and which has its own
evolutionary tendencies and historical fate?
Wiley
(1978)
Phylogenetic (PSC)
(diagnostic)
?A species is the smallest diagnosable cluster of individual organisms within which there is a parental pattern of ancestry and
descent?
Cracraft
(1983)
Differential ?tness
(DFSC)
?Species can be de?ned as groups of individuals that are reciprocally characterized by features that would have negative ?tness
effects in other groups and that cannot be regularly exchanged between groups upon contact?
Hausdorf
(2011)
R. Frankham et al. / Biological Conservation 153 (2012) 25?31 27establishment of gene ?ow between them (Frankham et al.,
2010). Further, it has been proposed that populations be translo-
cated into new habitats to cope with global climate change. For
populations with low genetic diversity, the best strategy is often
to translocate admixed populations into new habitat (Weeks
et al., 2011).
Managers that advocate either the transfer of organisms be-
tween fragmented populations or restriction of such transfers need
to consider the potential impacts of both outbreeding depression
(OD ? de?ned to include any deleterious consequences of crossing
on mating preference, pre- or post-zygotic reproductive ?tness),
and inbreeding depression (ID ? de?ned as the relative reduction
of ?tness in offspring of related mates compared to matings be-
tween unrelated individuals). Thus, de?nitions and delineations
for taxa with fragmented populations should lead to units that
simultaneously (a) minimize OD, whilst (b) allowing maximum
opportunities to outcross small inbred populations with low genet-
ic diversity to reverse inbreeding depression and loss of genetic
diversity (genetic rescue) (Frankham et al., 2010, 2011).
The consequences of different species delineations for six hypo-
thetical populations are illustrated in Fig. 1. Too broad a delinea-
tion of species in case 1 leads to a high risk of OD when
populations a and b are crossed. Over splitting in case 2, as a result
Populations a1 a2 a3 a4Case Species delineations
a1 a2 a3 a41
a1 a2 a3 a42
a1 a2 a3 a43
Fig. 1. Consequences of crossing populations following different species delineations in
rescue. Populations a and b are reproductively isolated (show OD on crossing), but popula
effective population size, is inbred and has low genetic diversity.populations that have or have not yet become reproductively iso-
lated to a substantial degree. A possible approach to de?ning ??sub-
stantial degree?? is to compare the degree of reproductive isolation
with that for well researched and widely accepted BSC species.
Reluctance to test reproductive isolation may exacerbate the dif?-
cult process of implementing managed gene ?ow.
We did not attempt to address de?nitions of units within spe-
cies (e.g. sub-species and evolutionarily signi?cant units) due to
space constraints.
3. How do excessively broad species delineations occur?
Excessively broad species delineations arise primarily from the
use of characters (mainly morphological) with insuf?cient resolv-
ing power to delimit cryptic species. For example, the endangered
grassland daisy Rutidosis leptorrhynchoides has been found to con-
b1 b2serve taxa with a small population size, and thereby result in
greater loss of existing biodiversity. In case 3, use of reproductive
isolation (de?ned as any adverse effect on pre-zygotic or post-zy-
gotic ?tness and equivalent to outbreeding depression) to delin-
eate species a versus species b both minimizes the risk of
outbreeding depression, and allows genetic rescue of small popula-
tions within species. Thus, to minimize harm and maximize bene-Consequences of crossing
b1 b2 OD in a x b crosses
b1 b2
ID in a4 and no rescue,
no OD
b1 b2
No OD and rescue of a4
possible
relation to outbreeding depression (OD), inbreeding depression (ID) and genetic
tions within them do not show reproductive isolation. The a4 population has a small
Papadopulos et al. (2011) described 13 potential instances of spe-
many mammal, bird, ?sh, lizard and plant species in Australia, Eur-
Conope and the Americas show evidence of the merging of previously
isolated and differentiated populations following climatic cycles
(see Frankham et al., 2011 Supporting information).
From a conservation perspective, such small populations are
susceptible to being classi?ed as different species according to
the diagnostic version of PSC, especially when maternally inherited
markers (mtDNA and cpDNA) and/or highly mutable genetic mark-
ers (microsatellites and animal mtDNA) are used in delineations.
Relying on neutral markers is also problematic since they have
been shown to be poor predictors of reproductive isolation, com-
pared to adaptive differentiation in a diverse array of taxa (Nosil
et al., 2002; Zigler et al., 2005; Stelkens and Seehausen, 2009;
Thorpe et al., 2010; Wang and Summers, 2010).
Below we discuss theory and empirical observations bearing on
the problem of excessive splitting of small threatened populations.
4.1. Theory predicting generations to attain reciprocal monophyly or
no shared alleles
The issue of diagnosability under PSC has as its purpose to
delineate populations where gene ?ow has ceased through either
intrinsic (e.g. failure to mate, or F1 sterility) or extrinsic factors
(e.g. geographic isolation, rivers, and mountains). Fixed gene differ-
ences (populations homozygous for different alleles) and reciprocal
monophyly are required under different implementations of theciation with gene ?ow for plants on Lord Howe Island, Australia.
Use of neutral genetic markers (or organelle DNA) may result in
such populations lacking ?xed differences or reciprocal monophyly
being classi?ed as a single species under PSC, ESC or the BSC.
Third, populations that diverged in allopatry may later come
into contact and form hybrid zones with some introgression of al-
leles in each direction. If such populations do not show reciprocal
monophyly or ?xed differences they may be classi?ed as a single
species (Eldridge and Close, 1992). For example, several rock wal-
laby species in Australia that exhibit hybrid zones and lack recipro-
cal monophyly for mtDNA and allozymes were the subject of
con?icting taxonomic delineations. Combined evidence from
mtDNA, allozymes and chromosomes eventually led to resolution
of their taxonomy, largely following chromosomal discontinuities
(Eldridge and Close, 1992).
4. How does excessive splitting of small populations occur?
Small isolated populations of conservation concern are subject
to large genetic drift effects that can quickly result in genetic dif-
ferentiation, without adaptation to different environments or the
evolution of reproductive isolation. Further, fragmented popula-
tions that are now geographically isolated (allopatric), but not
reproductively isolated may later come into contact and merge,
as has happened many times in nature through environmental
change (especially expansion and retreat of glaciers). For example,sist of diploid, tetraploid and hexaploid forms that are highly ster-
ile upon crossing (Murray and Young, 2001). Further, well studied
African elephants have recently been separated into savannah and
forest species despite regions of contact, based upon genome wide
deep sequence divergence between the two forms (Rohland et al.,
2010).
A second cause of excessive lumping occurs when speciation oc-
curs in the face of gene ?ow, as may occur when strong adaptive
differences drive reproductive isolation in sympatry. For example,
28 R. Frankham et al. / Biologicaldiagnostic PSC (Cracraft, 1997; Groves, 2004; see Supplementary
material). Lack of shared alleles between populations at one or
more loci is suf?cient to diagnose clusters that have experienceda long history without gene ?ow (see Supplementary material).
Fixed gene differences are one form of unshared alleles, but more
stringent than necessary to delineate lack of gene ?ow with multi-
ple alleles. Confusingly, different authors use diverse de?nitions for
?xed gene differences (see Supplementary material).
The relevant theory on generations required for populations to
be diagnosably different is couched in terms of reciprocal mono-
phyly, no shared alleles or ?xed gene differences. We present the
?rst two estimates in the main text and the third in the Supple-
mentary material, as the theoretical studies consider different sce-
narios, often with different assumptions. For reciprocal
monophyly, it takes about 4Ne generations from the time that
two populations separate for there to be a high probability of their
having reciprocally monophyletic alleles for mtDNA (Niegel and
Avise, 1986; Moritz, 1994; Hudson and Coyne, 2002), where Ne is
the effective population size (de?ned in the Supplementary mate-
rial; Frankham et al., 2010). Since the Ne for autosomal nuclear loci
is four times that for mtDNA loci under the conditions of the mod-
els, it takes approximately 16Ne generations to attain reciprocal
monophyly for nuclear autosomal loci (Hudson and Coyne, 2002).
The probability of shared alleles/haplotypes for DNA sequences
at a neutral nuclear autosomal locus approaches zero for divergence
times greater than 10Ne generations (Hey, 1991; Supplementary
material), and by extension 2.5Ne generations for mtDNA. The num-
berof generations is alsopartiallydependentuponallele frequencies
in the common ancestral population (Kimura and Ohta, 1971).
The number of generations to diagnosability will be less if mul-
tiple independent (unlinked) nuclear autosomal loci are genotyped
(Hudson and Coyne, 2002). For example, if the probabilities that
two populations are diagnosably different at each locus are all
0.5 (at the tth generation), then with 1, 2, 3 and 10 loci the proba-
bilities of diagnosing populations as different are 0.5, 0.75, 0.875
and 0.999, respectively.
Since we have a conservation focus, we ask how long it takes for
reciprocalmonophyly or no shared alleles to be detectable for threa-
tened species (Table 2). For species with stable population sizes, the
critically endangered (CE) IUCN (World Conservation Union, 2011)
RedList category criterionD is de?nedbyanadult censuspopulation
size for the entire species (N) <50, the endangered category (E) by
N < 250and thevulnerable category (V)byN < 1000 (Table 2). If each
of these categories has two equally sized isolated populations, then
they will be half the above numbers. We translated these census
numbers into genetically effective sizes (Ne) using empirical esti-
mates of the Ne/N ratio. Frankham (1995) and Palstra and Ruzzante
(2008) reported average ratios of 0.11 and 0.15 for nuclear loci,
respectively. Using the mid-point of this range (0.13), reciprocal
monophyly for autosomal loci is achieved in <52 (=16  25  0.13)
generations for populations of CE species, <260 generations for pop-
ulations of E species and <1040 generations for populations of V spe-
cies (Table 2). Somewhat fewer generations are required for almost
all populationpairs tobediagnosableusingno sharedalleles at a sin-
gle locus. Given the stochastic nature of differentiation, therewill be
many diagnosable population pairs in even fewer generations. If the
extent of fragmentation is greater, the numbers of generations will
be correspondingly reduced.
If such isolated populations inhabit similar environments, they
will be diagnosably different long before they show outbreeding
depression (especially for CE and E species) because OD has not
evolved in populations isolated for up to 6000 generations under
these conditions (Frankham et al., 2011). Full reproductive isola-
tion typically takes on the order of millions of years to evolve (Coy-
ne and Orr, 2004). In contrast, populations in different
environments show the ?rst signs of outbreeding depression with-
servation 153 (2012) 25?31in a few dozens of generations (Hendry et al., 2007).
Random mating populations of threatened species, maintained
at the above effective population sizes for the number of genera-
ono
.
st ca
dan
Contions required to achieve reciprocal monophyly will be highly
inbred, will suffer substantial inbreeding depression, and exhibit
large genetic rescue effects upon crossing (Frankham et al.,
2010). The expected inbreeding coef?cient (F) for a closed diploid
random mating population of size Ne over t generations is (Frank-
ham et al., 2010):
F ? 1  ?1  1=?2Ne?t ?1?
For example, a CE population with an Ne of 3.25 maintained
with random mating for 52 generations has an expected inbreed-
ing coef?cient of >0.99, and similar calculations for Ne and genera-
tions to achieve reciprocal monophyly in E and V populations also
yield inbreeding coef?cients of >0.99. At these inbreeding levels,
populations of naturally outbreeding species have high probabili-
ties of extinction from their inbreeding (Frankham et al., 2010).
Mitochondrial DNA is extensively used to delineate species (see
Karl and Bowen, 1999; Hebert et al., 2003; Craig et al., 2009). For
example, DNAbarcoding, based on sequencing a650 base pair sec-
tion of themtDNA cytochromeoxidase I locus (CO1), is being used to
discover new species and to estimate the approximate number of
animal species on Earth (see Hebert et al., 2003; Rubinoff et al.,
2006). mtDNA is maternally inherited in most animal species and
its effective population size (NemtDNA) is less than 3% of potentially
breeding adults, due to the combined effects ofmode of inheritance,
deviations from the idealized population structure and selection
(see Supplementary material). With such low ratios of Ne mtDNA/N,
even populations with reasonable sizes rapidly exhibit reciprocal
monophyly. For example, isolated populations of CE, E and V species
are expected to show reciprocal monophyly in mtDNA within less
than 3, 15 and 58 generations, and require even fewer genera-
tions to achieve no shared haplotypes (Table 2). Further,mtDNA has
a nucleotide mutation rate about 10 times higher than for nuclear
Table 2
Census (N) and effective population sizes (Ne), and predicted generations to reciprocal m
critically endangered, endangered and vulnerable species with stable population sizes
Item IUCN red li
Critically en
N (IUCN criterion D) <50
N when split into 2 equally sized fragments <25
Autosomal loci
Ne autosomes <3.25
Generations to reciprocal monophyly <52
Generations to no shared alleles <32.5
mtDNA
Ne mtDNA <0.73
Generations to reciprocal monophyly <2.9
Generations to no shared haplotypes <1.8
R. Frankham et al. / Biologicalloci in animals (see Ballard and Whitlock, 2004; Frankham, 2012).
Consequently, mtDNA shows large divergence due to mutation
and drift effects in animals, and excessive splitting of populations
into species based on mtDNA is a serious conservation problem
(Rubinoff et al., 2006; Frankham, 2012). Chloroplast DNA is widely
used in plant taxonomy, and is also expected to suffer low Ne/N ra-
tios, but there is very limited informationonwhich toestimate ratios
(Frankham, 2012).
Microsatellites have higher mutation rates than for other DNA
sequences or allozymes (Frankham et al., 2010), and their use will
be more likely to result in excessive splitting of populations (other
things being equal).
4.2. Empirical data on rapid attainment of diagnosable differences
between populations
Empirical data support the above theory indicating that isolated
populations of threatened species will be diagnosably differentwithin a small number of generations (well before they are likely
to evolve reproductive isolation). First, eight replicate populations
of Drosophila melanogaster derived from the same wild source pop-
ulations and maintained in isolation for 48?49 generations at
effective population sizes of 25 in the same environment were
diagnosably different in 27 of 28 comparisons (did not share alleles
at one or more of the eight loci; see data in the Supplementary
material). Note that these were diagnosably different in 2Ne gen-
erations, many fewer than estimated in Table 2, as a result of using
multiple loci, rather than a single locus. These populations suffered
30% and 88% inbreeding depression in reproductive ?tness in be-
nign and stressful conditions and showed 59% and nearly sixfold
genetic rescue effect in the two environments, respectively (Wood-
worth et al., 2002). These diagnosably different populations would
be classi?ed as distinct species according to the PSC if only two
populations survived, or if a taxonomist only sampled two popula-
tions and typed them as done in this study. It would be inadvisable
to classify interfertile threatened populations with such poor ?t-
ness as distinct species after such a short period without gene ?ow.
Second, six island populations of black-footed rock-wallabies
(Petrogale lateralis) in Australia are all diagnosably different from
each other (?xed gene differences) based on genotypes for eight
autosomal microsatellite loci (Eldridge et al., 1999), and all Island
populations have distinct mtDNA haplotypes (Eldridge et al.,
2001; Eldridge, unpublished data). However, the island popula-
tions were isolated by sea level rises only 1600?3000 generations
ago and the threatened mainland population contains almost all
of the alleles present in the combination of the island populations.
Some island populations have been crossed to the mainland popu-
lation and there are no indications of reproductive isolation (Close
and Bell, 1997). In the absence of the threatened mainland popula-
tion, the six island populations would be classi?able as six PSC spe-
phyly, and no shared alleles/haplotypes for autosomal and mtDNA genetic markers for
tegory
gered Endangered Vulnerable
<250 <1000
<125 <500
<16.25 <65
<260 <1040
<162.5 <650
<3.65 <14.6
<14.6 <58.3
<9.1 <36.5
servation 153 (2012) 25?31 29cies. The Barrow Island population (the only population
investigated) is suffering inbreeding depression (Eldridge et al.,
1999). Each of these island populations would probably bene?t
from augmented gene ?ow, so it would be counter-productive to
designate them as separate species that cannot be crossed.
Third, inbred strains of laboratory animals, such as mice (Mus
musculus) (typically maintained using one pair of parents per gen-
eration) show ?xed differences at many molecular loci and are
diagnosable different within 20 generations (Falconer and Mac-
kay, 1996), but strains within species cross readily with bene?cial
effects on ?tness (Atchley and Fitch, 1991). These inbred mice pop-
ulations would be delineated as separate species if we applied
diagnostic PSC to pairs of populations in the absence of their ances-
tral wild species. While the population sizes used in these popula-
tions are continuously smaller than in most threatened species, the
same processes occur in threatened populations at a slower rate. In
fact, several wild species, including Mauritius kestrels (Falco punct-
atus) and Chatham Island black robins (Petroica traversi) have expe-
major taxa (algae/protozoa, fungi, plants and animals), while dif-
ferences predict outbreeding depression.
and methods of biological conservation.
ConGiven the important implications of using different species con-
cepts in delineating species, it is critical that the species conceptrienced single pair bottlenecks (Ardern and Lambert, 1997; Groom-
bridge et al., 2000).
5. Discussion
The arguments above lead us to recommend that substantial
reproductive isolation (pre- and/or post-zygotic) be used to de?ne
species of outbreeding sexual organisms for conservation pur-
poses. In this way genetic rescue efforts will be possible and the
risk of outbreeding depression minimized. DFSC satis?es these cri-
teria, whilst BSC captures large components of it, especially if it is
used in a ?relaxed? form that accepts limited gene ?ow. Consider-
ation of these issues for taxa (mainly plants) that are not outbreed-
ing diploids is given in the Supplementary material.
We are strongly opposed to the use of a non-de?ned approach
to species delineations, as used in the taxonomic species concept.
The current situation in taxonomy is clearly unsatisfactory with
use of different (and often contradictory) species concepts, traits,
sampling regimes, methods of analysis, etc. that results in many
controversies about the taxonomic status of populations (Mace,
2004). Current practices often fail the usual scienti?c requirements
of robust sampling, repeatability and adequate statistical support.
How then should we distinguish species for sympatric, parapat-
ric and allopatric populations? Lack of shared alleles at one or more
autosomal loci (given suf?cient sampling) is suf?cient to establish
lack of gene ?ow and reproductive isolation between two sympat-
ric or parapatric populations and for them to be to be classi?ed as
separate species (even evidence of very limited gene ?ow should
be acceptable). For example, samples of Onychophora from the
same log in the Blue Mountains west of Sydney, Australia showed
?xed gene difference at 70% of loci and were reclassi?ed from
belonging to the same morphologically de?ned species into two
species (Briscoe, DA pers. comm.; Briscoe and Tait, 1995). When
there is lack of gene ?ow between sympatric or parapatric popula-
tions, classi?cations with BSC, PSC, ESC and DFSC should be concor-
dant (Knowlton and Weigt, 1997).
Allopatric populations will be classi?ed appropriately using the
DFSC, as such species are delineated using pre- and/or post-zygotic
isolation, while BSC uses mating isolation and sterility to delineate
species. However, allopatric diagnostic PSC species may not be
reproductively isolated. In such cases, genetic rescue of small pop-
ulations may only be possible by carrying out crosses between dis-
tinct PSC species. If this is to be done to save biodiversity,
regulatory and legal hurdles will need to be removed (O?Brien
and Mayr, 1991; Haig and Allendorf, 2006; Ellstrand et al., 2010).
Given the dif?culties involved, we recommend that allopatric PSC
species that require conservation management have their taxon-
omy reassessed on the basis of outbreeding depression following
DFSC or BSC.
It has been argued that it is dif?cult or impractical to determine
whether populations are reproductive isolated, as experimental
crossing through multiple generations is frequently impractical.
However, ?xed chromosomal differences and/or adaptation to dif-
ferent environments are good predictors of outbreeding depression
if populations are crossed (see Frankham et al., 2011; Sexton et al.,
2011). Further, Coleman (2009) reported that DNA sequence simi-
larity in the 50 region of helix III of ITS2 of nuclear rRNA predicts
ability to successfully cross populations within a broad range of
30 R. Frankham et al. / Biologicalused in delineations be routinely provided in the conservation lit-
erature. Worryingly, this is rarely the case. We recommend that
journals require that the species concept used be identi?ed in pa-References
Agapow, P.-M., Bininda-Emonds, O.R.P., Crandall, K.A., Gittleman, J.L., Mace, G.M.,
Marshall, J.C., Purvis, A., 2004. The impact of species concepts on biodiversity De?nitions that are too broad will often lead to outbreeding
depression when populations are crossed, while those that split
excessively may preclude genetic rescue of small inbred popu-
lations with low genetic diversity.
 Minimum harm is done and maximum potential bene?ts in
terms of ?tness and adaptive evolutionary potential accrue
when reproductive isolation (pre- and/or post-zygotic) is used
as the criterion to de?ne distinct species.
 Many ?good? species exhibit low levels of gene ?ow from other
taxa.
 Use of the differential ?tness or biological species concepts will
typically yield a classi?cation appropriate to conservation
concerns.
 Conversely, use of the diagnostic phylogenetic or the taxonomic
species concepts will often lead to inappropriate classi?cations.
 For sympatric or parapatric populations, distinct species are
diagnosed by any genetically based distinctiveness that indi-
cates lack of (or very limited) gene ?ow, and different species
concepts typically yield concordant delineations.
 For allopatric populations the appearance of detectable diver-
gence due to lack of recent gene ?ow alone is not necessarily
diagnostic of intrinsic reproductive isolation. In addition, such
diagnoses require evidence of reduced reproductive ?tness in
population crosses, or inferences from ?xed chromosomal dif-
ferences, or adaptive differentiation among populations that
predict reproductive isolation.
 It is critical that the species de?nition being used be stated
explicitly when new species are designated, and conservation
decisions prescribed or proscribed by taxonomy need to iden-
tify the species concept that was used in the delineation.
Acknowledgments
RF thanks R. Lacy, J. Ballou and I. Porton for inviting him to
speak on this topic at a meeting at Saint Louis Zoo in 2011 and
to the Saint Louis Zoo for providing travel funds. JRM thanks C.
Feldman for helpful discussion. We thank J. Avise, D. Colgan, G.
Cooke, K. de Queiroz, S. Donellan, B. Hausdorf, G. Mace, S. Smith,
A. Stow, P. Sunnucks, R. Waples and four anonymous reviewers
for comments on the manuscript, and J. Hey for answering queries
relating to his published work.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.biocon.2012.04.
034.pers where new threatened species are described, where taxo-
nomic disputes are resolved, or where boundaries for threatened
species are described.
6. Conclusions
 Conservation biology is bedevilled by the existence and use of at
least 26 de?nitions of species. Consequently, it is important to
determine which de?nition(s) best accord with the purposes
servation 153 (2012) 25?31studies. Quart. Rev. Biol. 79, 161?179.
Allendorf, F.W., Luikart, G., 2006. Conservation and the Genetics of Populations.
Blackwell, Malden, MA.
R. Frankham et al. / Biological Conservation 153 (2012) 25?31 31Ardern, S.L., Lambert, D.M., 1997. Is the black robin in genetic peril? Mol. Ecol. 6,
21?28.
Atchley, W.R., Fitch, W.M., 1991. Gene trees and the origin of inbred strains of mice.
Science 254, 554?558.
Avise, J.C., Ball, R.M.J., 1990. Principles of genealogical concordance in species
concepts and biological taxonomy. Oxford Surv. Evol. Biol. 7, 45?67.
Ballard, J.W.O., Whitlock, M.C., 2004. The incomplete natural history of
mitochondria. Mol. Ecol. 13, 729?744.
Briscoe, D.A., Tait, N.N., 1995. Allozyme evidence for extensive and ancient
radiations in Australian Onychophora. J. Hered. 114, 91?102.
Claridge, M.F., Dawah, H.A., Wilson, M.R., 1997. Species: The Units of Biodiversity.
Chapman & Hall, London.
Close, R.L., Bell, J.N., 1997. Fertile hybrids in two genera of wallabies. Petrogale and
Thylogale. J. Hered. 88, 393?397.
Coleman, A.W., 2009. Is there a molecular key to the level of ??biological species?? in
eukaryotes? A DNA guide. Mol. Phylogenet. Evol. 50, 197?2003.
Coyne, J.A., Orr, H.A., 2004. Speciation. Sinauer, Sunderland, MA.
Cracraft, J., 1983. Species concepts and speciation analysis. Curr. Ornithol. 1, 159?
187.
Cracraft, J., 1997. Species concepts in systematics and conservation biology ? an
ornithological viewpoint. In: Claridge, M.F., Dawah, H.A., Wilson, M.R. (Eds.),
Species: The Units of Biodiversity. Chapman & Hall, London, pp. 325?339.
Craig, M.T., Graham, R.T., Torres, R.A., Hyde, J.R., Freitas, M.O., Ferreira, B.P., Hostim-
Silva, M., Gerhardinger, L.C., Bertoncini, A.A., Robertson, D.R., 2009. How many
species of goliath grouper are there? Cryptic genetic divergence in a threatened
marine ?sh and the resurrection of a geopolitical species. Endangered Species
Res. 7, 167?174.
Darwin, C., 1859. On the Origin of Species by Means of Natural Selection, or the
Preservation of Favoured Races in the Struggle for Life. John Murray, London.
de Queiroz, K., 1998. The general lineage concept of species, species criteria, and the
process of speciation. In: Howard, D.J., Berlocher, S.H. (Eds.), Endless Forms:
Species and Speciation. Oxford University Press, New York, pp. 57?75.
Eldredge, N., Cracraft, J., 1980. Phylogenetic Patterns and the Evolutionary Process.
Columbia University Press, New York.
Eldridge, M.D.B., Close, R.L., 1992. Taxonomy of rock-wallabies, Petrogale
(Marsupialia: Macropodidae). I. A revision of the eastern Petrogale with a
description of three new species. Aust. J. Zool. 40, 605?625.
Eldridge, M.D.B., King, J.M., Loupis, A.K., Spencer, P.B.S., Taylor, A.C., Pope, L.C., Hall,
G.P., 1999. Unprecedented low levels of genetic variation and inbreeding
depression in an island population of the black-footed rock-wallaby. Conserv.
Biol. 13, 531?541.
Eldridge, M.D.B., Wilson, A.C.C., Metcalfe, C.J., Dollin, A.E., Bell, J.N., Johnson, P.M.,
Johnston, P.G., Close, R.L., 2001. Taxonomy of rock-wallabies, Petrogale
(Marsupialia: Macropodidae). III. Molecular data con?rms the species status
of the purple-necked rock-wallaby (Petrogale purpureicollis Le Souef). Aust. J.
Zool. 49, 323?343.
Ellstrand, N.C., Biggs, D., Kaus, A., Lubinsky, P., McDade, L.A., Preston, K., Prince, L.M.,
Regan, H.M., Rorive, V., Ryder, O.A., Schierenbeck, K.A., 2010. Got hybridization?
A multidisciplinary approach for informing science policy. Bioscience 60, 384?
388.
Falconer, D.S., Mackay, T.F.C., 1996. Introduction to Quantitative Genetics, fourth
ed. Longman, Harlow, England.
Frankham, R., 1995. Effective population size/adult population size ratios in
wildlife: a review. Genet. Res. 66, 95?107.
Frankham, R., 2012. How closely does genetic diversity in ?nite populations
conform to predictions of neutral theory? Large de?cits in regions of low
recombination. Heredity 108, 167?178.
Frankham, R., Ballou, J.D., Briscoe, D.A., 2010. Introduction to Conservation Genetics,
second ed. Cambridge University Press, Cambridge, UK.
Frankham, R., Ballou, J.D., Eldridge, M.D.B., Lacy, R.C., Ralls, K., Dudash, M.R., Fenster,
C.B., 2011. Predicting the probability of outbreeding depression. Conserv. Biol.
25, 465?475.
Groombridge, J.J., Jones, C.G., Bruford, M.W., Nichols, R.A., 2000. ?Ghost? alleles of the
Mauritius kestrel. Nature 403, 616.
Groves, C., 2004. The what, why and how of primate taxonomy. Int. J. Primatol. 25,
1105?1126.
Haig, S.M., Allendorf, F.W., 2006. Hybrids and policy. In: Scott, J.M., Goble, D.D.,
Davis, F. (Eds.), The Endangered Species Act at Thirty. Conserving Biodiversity in
Human-dominated Landscapes, 2. Island Press, Washington, DC, pp. 150?163.
Hausdorf, B., 2011. Progress toward a general species concept. Evolution 65, 923?
931.
Hebert, P.D.N., Cywinska, A., Ball, S.L., de Waard, J.R., 2003. Biological identi?cation
through DNA barcodes. Proc. Roy. Soc. Lond. B: Biol. Sci. 270, 313?321.
Hendry, A.P., Nosil, P., Rieseberg, L.H., 2007. The speed of ecological speciation.
Funct. Ecol. 21, 455?464.
Hey, J., 1991. The structure of genealogies and the distribution of ?xed differences
between DNA sequence samples from natural populations. Genetics 128, 831?
840.
Hey, J., Pinho, C., 2012. Population genetics and objectivity in species diagnosis.
Evolution. http://dx.doi.org/10.1111/j.1558-5646.2011.01542.x.
Hey, J., Waples, R.S., Arnold, M.L., Butlin, R.K., Harrison, R.G., 2003. Understanding
and confronting species uncertainty in biology and conservation. Trends Ecol.
Evol. 18, 597?603.Hudson, R.R., Coyne, J.A., 2002. Mathematical consequences of the genealogical
species concept. Evolution 56, 1557?1565.
IUCN, 2011. IUCN Red List of Threatened Species. <http://www.redlist.org/>
accessed December 2011.
Karl, S.A., Bowen, B.W., 1999. Evolutionary signi?cant units versus geopolitical
taxonomy: molecular systematics of an endangered sea turtle (genus Chelonia).
Conserv. Biol. 13, 990?999.
Kimura, M., Ohta, T., 1971. Theoretical Aspects of Population Genetics. Princeton
University Press, Princeton, NJ.
Knowlton, N., Weigt, L.A., 1997. Species of marine invertebrates: a comparison of
the biological and phylogenetic species concepts. In: Claridge, M.F., Dawah, H.A.,
Wilson, M.R. (Eds.), Species: The Units of Biodiversity. Chapman & Hall, London,
pp. 199?219.
Mace, G.M., 2004. The role of taxonomy in species conservation. Philos. Trans. Roy.
Soc. B: Biol. Sci. 359, 711?719.
Mayden, R.L., 1997. A hierarchy of species concepts: the denouncement in the saga
of the species problem. In: Claridge, M.F., Dawah, H.A., Wilson, M.R. (Eds.),
Species: The Units of Biodiversity. Chapman & Hall, London, pp. 381?424.
Mayr, E., 1942. Systematics and the Origin of Species. Columbia University Press,
New York, NY.
Mayr, E., 1963. Animal Species and Evolution. Harvard University Press, Cambridge,
MA.
McDade, L.A., 1995. Species concepts and problems in practice. insights from
botanical monographs. Syst. Bot. 20, 606?622.
Millennium Ecosystem Assessment (Ed.), 2005. Ecosystems and Human Well-being:
Current State and Trends. Findings of the Conditions and Trends Working Group
of the Millenium Ecosystem Assessment, vol. 1. Island Press, Washington, DC.
Moritz, C., 1994. De?ning ?evolutionary signi?cant units? for conservation. Trends
Ecol. Evol. 9, 373?375.
Murray, B.G., Young, A.G., 2001. Widespread chromosomal variation in the
endangered grassland forb Rutidosis leptorrhynchoides F. Muell. (Asteraceae:
Gnaphalieae). Ann. Bot. 87, 83?90.
Neigel, J.E., Avise, J.C., 1986. Phylogenetic relationships of mitochondrial DNA under
various demographic models of speciation. In: Karlin, S., Nevo, E. (Eds.),
Evolutionary Processes and Theory. Academic Press, Orlando, FL, pp. 515?534.
Noor, M.A.F., 2002. Is the biological species concept showing its age? Trends Ecol.
Evol. 17, 153?154.
Nosil, P., Crespi, B.J., Sandoval, C.P., 2002. Host-plant adaptation drives the parallel
evolution of reproductive isolation. Nature 417, 440?443.
O?Brien, S.J., Mayr, E., 1991. Bureaucratic mischief: recognizing endangered species
and subspecies. Science 251, 1187?1188.
Palstra, F.P., Ruzzante, D.E., 2008. Genetic estimates of contemporary effective
population size: what can they tell us about the importance of genetic
stochasticity for wild population persistence? Mol. Ecol. 17, 3428?3447.
Papadopulos, A.S.T., Baker, W.J., Crayn, D., Butlin, R.K., Kynast, R.G., Hutton, I.,
Savolainen, V., 2011. Speciation with gene ?ow on Lord Howe island. Proc. Natl.
Acad. Sci. USA 108, 13188?13193.
Rohland, N., Reich, D., Mallick, S., Meyer, M., Green, R.E., Georgiadis, N.J., Roca, A.L.,
Hofreiter, M., 2010. Genomic DNA sequences from mastodon and woolly
mammoth reveal deep speciation of forest and savanna elephants. PLoS Biol. 8
(12), e1000564.
Rubinoff, D., Cameron, S., Will, K., 2006. A genomic perspective on the shortcomings
of mitochondrial DNA for ??barcoding?? identi?cation. J. Hered. 97, 581?594.
Sexton, J.P., Strauss, S.Y., Rice, K.J., 2011. Gene ?ow increases ?tness at the warm
edge of species? range. Proc. Natl. Acad. Sci. USA 108, 11704?11709.
Simpson, G.G., 1951. The species concept. Evolution 5, 285?298.
Simpson, G.G., 1961. Principles of Animal Taxonomy. Columbia University Press,
New York.
Stelkens, R.B., Seehausen, O., 2009. Phenotypic divergence but not genetic distance
predicts assortative mating among species of a cichlid ?sh radiation. J. Evol.
Biol. 22, 1679?1694.
Thorpe, R.S., Surget-Groba, Y., Johansson, H., 2010. Genetic tests for ecological and
allopatric speciation in anoles on an island archipelago. PLoS Genet. 6,
e1000929.
Wang, I.J., Summers, K., 2010. Genetic structure is correlated with phenotypic
divergence rather than geographic isolation in the highly polymorphic
strawberry poison-dart frog. Mol. Ecol. 19, 447?458.
Weeks, A.R., Sgr?, C.M., Young, A.G., Frankham, R., Mitchell, N.J., Miller, K.A., Byrne,
M., Coates, D.J., Eldridge, M.D.B., Sunnucks, P., Breed, M.F., James, E.A.,
Hoffmann, A.A., 2011. Assessing the bene?ts and risks of translocations in
changing environments: a genetic perspective. Evol. Appl. 4, 709?725.
Wiley, E.O., 1978. The evolutionary species concept reconsidered. Syst. Biol. 27, 17?
26.
Wilkins, J.S., 2009. Species: The History of an Idea. University of California Press,
Berkeley, CA.
Winkler, K., 2010. Is it a species? Int. J. Avian Sci. 152, 679?682.
Woodworth, L.M., Montgomery, M.E., Briscoe, D.A., Frankham, R., 2002. Rapid
genetic deterioration in captivity: causes and conservation implications.
Conserv. Genet. 3, 277?288.
Zigler, K.S., McCartney, M.A., Levitan, D.R., Lessios, H.A., 2005. Sea urchin bindin
divergence predicts gamete compatibility. Evolution 59, 2399?2404.