Quantifying Biodiversity: a Phenotypic Perspective

1014

Conservation Biology, Pages 1014–1022Volume 14, No. 4, August 2000

Quantifying Biodiversity: a Phenotypic Perspective

IAN P.F. OWENS*‡ AND PETER M. BENNETT†

*Department of Zoology and Entomology, University of Queensland, Brisbane, Queensland 4072, Australia,email [email protected]†Institute of Zoology, Zoological Society of London, Regent’s Park, London NW1 4RY, United Kingdom

Abstract:

With recently derived algorithms, it is possible to calculate the relative phylogenetic distinctivenessof taxa with respect to patterns of phylogenetic branching. We extended this approach by exploring the rela-tive extent to which taxa represent phenotypic biodiversity. We devised a method, based on the use of inde-pendent contrasts, that measures the amount of phenotypic change that occurs when lineages diverge. We usethis method to quantify the extent to which a taxon represents the phenotypic diversification that has oc-curred in the past. We applied our method to an analysis of variation in clutch size across 133 avian fami-lies. All families did not contribute equally to representing clutch size diversification. The top 10 avian fami-lies in terms of representing clutch size diversification were the mesites (Mesitornithidae), cranes (Gruidae),bustards (Otidae), new world quail (Odontophoridae), seriemas (Cariamidae), finfoots (Heliornithidae),swallows (Hirundinidae), megapodes (Megapodiae), and guans (Cracidae). The 217 species in these 10 fam-ilies (2.3% of all bird species, 7.5% of families) represented 19.3% of diversification in clutch size. Seventeenpercent of overall clutch size diversification was represented by taxa threatened with extinction. The 10 fam-ilies that represent the greatest proportion of overall clutch size diversification threatened by extinction werethe mesites (Mesitornithidae), kagu (Rhynochetidae), cranes (Gruidae), kiwis (Apterygidae), new worldquail (Odontophoridae), megapodes (Megapodiae), cassowaries (Casuariidae), finfoots (Heliornithidae),guans (Cracidae), and logrunners (Orthonychidae). The 42 threatened species (0.5% of all bird species, 3.8%of all threatened bird species) in these 10 families encompassed 53% of the clutch size diversification whoserepresentation was threatened with extinction. Our results suggest that this type of analysis could potentiallyhelp prioritize species-based conservation efforts by identifying those taxa that contribute most toward repre-senting the evolutionary processes that lead to current phenotypic biodiversity.

Cuantificación de la Biodiversidad: una Perspectiva Fenotípica

Resumen:

Con algoritmos recientemente calculados es posible estimar el grado de distinción filogenética detaxones con respecto a patrones de ramificación filogenética. Extendimos esta estrategia al explorar el gradorelativo al cual los taxones representan a la biodiversidad fenotípica. Diseñamos un método basado en eluso de contrastes independientes que mide la cantidad de cambio fenotípico que ocurre cuando los linajes sebifurcan. Utilizamos este método para cuantificar el grado al que un taxón representa la diversificaciónfenotípica que ha ocurrido en el pasado. Aplicamos nuestro método a un análisis de variación en el tamaño denidada en 133 familias de aves. No todas las familias contribuyeron de igual manera a la representación de ladiversificación del tamaño de nidada. Las principales 10 familias de aves, en cuanto a la representación de la di-versificación del tamaño de la nidada fueron los monias (Mesitornithidae), grullas (Gruidae), avutardas(Otidae), perdices del nuevo mundo (Odontophoridae), cariamas (Cariamidae), pájaros cantil (Heliornithidae),golondrinas (Hirundinidae), megápodos (Megapodiae) y pavones (Cracidae). Las 217 especies en estas 10 fa-milias (2.3% de todas las especies de aves, 7.5% de las familias) representaron el 19.3% de la diversificacióndel tamaño de nidada. El diecisiete por ciento de la diversificación total del tamaño de nidada estuvo repre-sentado por taxones amenazados de extinción. Las 10 familias que representan la mayor proporción de ladiversificación total del tamaño de nidada amenazadas de extinción fueron los monias (Mesitornithidae),kagu (Rhynochetidae), grullas (Gruidae), kiwis (Apterygidae), perdices del nuevo mundo (Odontophoridae),

‡

Current address: Department of Biology, Imperial College, Silwood Park, Ascot, Berkshire SL5 7PY, United Kingdom,email [email protected] submitted October 27, 1998; revised manuscript accepted November 3, 1999.

Conservation BiologyVolume 14, No. 4, August 2000

Owens & Bennett Identifying Distinctive Threatened Taxa

1015

megápodos (Megapodiae), casuarios (Casuariidae), pájaros cantil (Heliornithidae), pavones (Cracidae) ylog-runners (Orthonychidae). Las 42 especies amenazadas (0.5% de todas las especies de aves, 3.8% de todaslas especies de aves amenazadas) dentro de estas 10 familias abarcaron el 53% de la diversificación deltamaño de nidada, cuya representación estuvo amenazada por la extinción. Nuestros resultados sugierenque este tipo de análisis podría potencialmente ayudar a priorizar los esfuerzos de conservación basados enespecies al identificar aquellos taxones que más contribuyan a la representación de los procesos evolutivos

que conduzcan a la biodiversidad fenotípica actual.

Introduction

Biodiversity refers to variation in the organic world. Suchvariation may be expressed in many ways, such as phyloge-netic, molecular, or phenotypic variation. For example, re-cent investigations of phylogenetic variation have led to thedevelopment of algorithms that rank taxa according to de-grees of “phylogenetic distinctiveness.” In essence, the rela-tive distinctiveness of a taxon is inversely proportional tothe relative number and closeness of its phylogenetic rela-tives (for a range of approaches, see May 1990, 1994; Vane-Wright et al. 1991; Williams et al. 1991, 1994; Crozier 1992;Faith 1992

a

, 1992

b

, 1993, 1994

a

, 1994

b

, 1994

c

; Williams &Humphries 1994, 1996; Walker & Faith 1995).

Our primary aim was to develop a method that can beused to explore the evolutionary history of phenotypicvariation. We did this by measuring phenotypic diversifi-cation relative to patterns of phylogenetic branching.Previous methods implicitly assume that all phyloge-netic divergences are equal. Of course, different diver-gences are known to be associated with very differentamounts of phenotypic change: the amount of pheno-typic diversification between two sister lineages mayvary anywhere between huge and minimal. We soughtto rank lineages on the basis of how much phenotypicdiversification they represent because we are interestedin the events that lead to current phenotypic variation aswell as the process of phylogenetic branching itself (fordiscussion of the role of phenotypic diversity in contrib-uting to overall biodiversity, see Williams et al. 1994;Williams & Humphries 1996). The method we introduceabandons the assumption that all phylogenetic diver-gences produce equal amounts of phenotypic change.

We applied our method to data on clutch size variationin birds to illustrate its potential value to conservation bi-ology. We stress throughout that (1) our analyses ofclutch size variation is presented largely for illustrativepurposes, and we do not propose that conservation ef-fort should be focused solely on the maintenance of anysingle trait; and (2) the method is not designed to mea-sure evolutionary distinctiveness. With respect to thissecond point, we agree with Faith (1992

a

, 1992

b

, 1993,1994

a

, 1994

b

, 1994

c

), Harvey (1994, 1995), and May(1994) that the best way to measure true phylogeneticdistinctiveness is to calculate maximum phylogenetic

path length and discount for shared paths. Here we arenot attempting to measure phylogenetic distinctiveness,however, but rather we are looking for a way to explorethe processes that lead to phenotypic diversity.

Methods

Quantifying Phenotypic Diversification

Phenotypic diversity is created by differences betweentaxa in their phenotype. We assumed that such differ-ences arose from phenotypic diversification betweenlineages during the evolutionary past. To quantify this di-versification process, we adopted a phylogenetic per-spective based on Felsenstein’s (1985) idea of evolution-arily independent comparisons. Differences betweentaxa in their phenotype were quantified by identifyingand measuring evolutionarily independent comparisonsbetween lineages. For example, we considered four taxa(T1, T2, T3, and T4) that varied with respect to theirvalue of a phenotypic trait (X1, X2, X3, and X4, respec-tively) and that were related to each other (relationshipdepicted in Fig. 1). The variation among these four taxawith respect to the phenotypic trait was encapsulatedby three evolutionarily independent comparisons A, B,and C. The value of each of these comparisons was thedifference between the phenotypic trait values of the twodaughter lineages, equivalent to unstandardized contrasts inthe comparative analysis by independent contrasts methodof Harvey and Pagel (1991), Pagel (1992), and Purvis andRambaut (1995) (Owens & Bennett 1995).

The recognition of evolutionarily independent con-trasts allowed us to quantify the relative contribution ofeach taxon to representing the phenotypic diversifica-tion that occurred at a particular phylogenetic diver-gence. For the phenotypic diversification at a compari-son to be represented, a taxon must be present fromeach of the daughter lineages. If there is only one taxonin one of these lineages, then this taxon must be presentfor the comparison to exist and this taxon is crucial forthe representation of that phenotypic diversification. Ifthere are several taxa present in one of the daughter lin-eages, then the presence of any one of these taxa is suffi-cient for the comparison to exist and the importance of

1016

Identifying Distinctive Threatened Taxa Owens & Bennett


each taxon in representing this phenotypic diversifica-tion is devalued by a proportion relative to the numberof other taxa present in the daughter lineage. In the caseillustrated in Fig. 1, for instance, for the phenotypic di-versification at comparison C to exist, both taxon T3and taxon T4 must be present. Thus, both of these taxaare equally important for the representation of the phe-notypic diversification at comparison C, and both ofthem are crucial. For the representation of the pheno-typic diversification at comparison B, taxon T2 must bepresent and so must either taxon T3 or taxon T4. Thus,with respect to representing the phenotypic diversifica-tion at comparison B, taxon T2 is crucial and is twice asimportant as T3 and T4. Finally, for the representation ofthe phenotypic diversification at comparison A to exist,taxon T1 must be present and so must either taxon T2,T3, or T4. So, with respect to representing the pheno-typic diversification at comparison A, taxon T1 is crucialand is three times as important as taxa T2, T3, and T4.

To obtain an overall index of the contribution of eachtaxon toward the representation of overall phenotypicdiversification, we added together the separate mea-sures of the relative contribution of the taxon in repre-senting phenotypic diversification at each independent

comparison. We referred to this overall index as the di-versification score of the taxon (

D

i

, where

i

is the

i

thtaxon). The magnitude of a taxon’s diversification scorewas determined by the number of comparisons towhich it contributed, the amount of phylogenetic diver-sification at each of these comparisons, and the relativecontribution of the taxon in representing the diversifica-tion at each comparison. In Fig. 1, for example, taxonT1 was crucial for representing the phenotypic diversifi-cation at comparison A but was irrelevant to compari-sons B and C. The diversification score of taxon T1 (

D

1

)was therefore equal to the size of comparison A. TaxonT2, on the other hand, contributed the same amount astwo other taxa in representing the phenotypic diversifi-cation at comparison A, crucial in representing the phe-notypic diversification at comparison B and irrelevant atcomparison C. The diversification score of taxon T2 (

D

2

)was therefore equal to one-third the size of comparisonA plus the size of comparison B. Finally, taxon T3 (andtaxon T4) contributed the same amount as two othertaxa in representing the phenotypic diversification atcomparison A, contributed the same amount as oneother taxon at comparison B, and was crucial at compar-ison C. Thus, the diversification score of taxon T3 (

D

3

and of taxon T4 (

D

4

) was equal to one-third of the size ofcomparison A plus half the size of comparison B plus thesize of comparison C (Table 1).

Phenotypic Diversification under Threat

We also used the diversification scores (1) to measurethe extent to which each taxon’s representation of phe-notypic diversification is due to the presence of speciesthat are currently classified as being threatened by ex-tinction and (2) to examine the proportion of overallphenotypic diversification that is due to the presence ofthreatened species.

The diversification score of each taxon (

D

i

) was multi-plied by the proportion of species within the taxon thatwere classified as being threatened by extinction (

P

i

) togive an index of the extent to which the taxon’s repre-sentation of phenotypic diversification was jeopardizedby current extinction risk. This index was called the di-versification-under-threat score (

E

i

; Table 1; Fig. 1).

Figure 1. Branching pattern illustrates the method for calculating the diversification score and diversifica-tion-under-threat score ( for full details see Table 1 and text).

Table 1. Method for calculating the diversification score and diversification-under-threat score.

Taxon( Ti )

Trait value( Xi )

Relativediversification score

*

(

D

i

)

Proportion of species in familythreatened by extinction

(

P

i

)

Diversification-under-threatscore(

E

i

)

T1 X1

D

1

5

A

P

1

E

1

5

D

1

3

P

1

T2 X2

D

2

5

A/3 + B

P

2

E

2

5

D

2

3

P

2

T3 X3

D

3

5

A/3 + B/2 + C

P

3

E

3

5

D

3

3

P

3

T4 X4

D

4

5

A/3 + B/2 + C

P

4

E

4

5

D

4

3

P

4

*

A, B, and C refer to the nodes illustrated in Fig. 1.


Owens & Bennett Identifying Distinctive Threatened Taxa

1017

The proportion of overall phenotypic diversificationrepresented by threatened taxa was calculated by divid-ing the sum of the taxon-specific diversification underthreat scores by the sum of the taxon-specific diversifica-tion scores (

S

i

E

i

/

S

i

D

i

).

Clutch Size Diversification among Birds

To illustrate our method for quantifying the representa-tion of phenotypic diversification, we used the exampleof clutch size variation among extant species of bird.Data on the mean clutch size for a species of bird fromeach of 133 avian families were obtained from Bennett(1986). Where there were data for more than one spe-cies within a family in Bennett (1986), a species was se-lected at random. Among birds, variation within familiesaccounts for

,

5% of overall variation in life-history traits(Bennett 1986; Owens & Bennett 1995), so a randomlychosen species will, on average, be representative ofother species in the same family. Clutch size was usedbecause (1) data were available for a large number andvariety of species, (2) it varies by 18-fold among extantbirds, and (3) variation in clutch size is correlated withvariation in other life-history traits (Bennett 1986). Weused families rather than species as the taxonomic levelfor analysis because we had already shown elsewherethat over 95% of overall variation in avian life histories isdue to comparisons between ancient avian lineages cor-responding to differences between families and differ-ences between orders (Bennett 1986; Owens & Bennett1995). Thus, differences in clutch size within familiesare largely irrelevant to overall avian life-history diver-sity. Data on which species are threatened by extinctionwere taken from Collar et al. (1994), who use Mace and

Stuart’s (1994) criteria for quantifying extinction risk. Todescribe the phylogenetic relationship between fami-lies, we used Sibley and Ahlquist’s (1990) tapestry phy-logeny, and we followed Sibley and Monroe (1990) forthe number of species in each family and the phyloge-netic position of families not included in the tapestryphylogeny. We assumed that all branch lengths areequal. Where Collar et al.’s (1994) taxonomy differedfrom Sibley and Monroe’s (1990), we followed the latter.We classified species as either threatened (categories ofvulnerable, endangered, and critical according to Collaret al. 1994) or not threatened by extinction (Bennett &Owens 1997).

Using these data, we calculated a diversification scorefor each of 133 avian families and ranked them accord-ing to this index (example in Table 2; Fig. 2). We thencalculated the diversification-under-threat score for eachfamily and used these figures to estimate the proportionof overall clutch size diversification that is due to taxacurrently being classified as threatened by extinction.

Results

All results refer to our analyses of clutch size variationamong birds. The sum of the diversification scores acrossall 133 families was 346. The 20 highest ranking familiesrepresented significantly more phenotypic diversificationthan would be expected if they had been selected at ran-dom. For example, the top 10 families (7.5%) encom-passed 19.3% of the sum of the diversification scores (

x

2

5

33.8, df

5

1,

p

,

0.0001), and the top 20 ranking families(15%) encompassed 33.2% of the sum of the diversifica-tion scores (

x

2

5

15.7, df

5

1,

p

,

0.001). In terms of spe-

Table 2. Example of a calculation of the diversification score and diversification-under-threat score based on data on clutch size variation among birds.

Taxon( Ti )

Trait value( Xi )

*

Relativediversification score

(

D

i

)

Proportion of species in familythreatened by extinction

(

P

i

)

Diversification-under-threatscore(

E

i

)

Struthionidae 6.0 2.50 0.00 0.00Rheidae 4.0 2.50 0.00 0.00Casuaridae 6.0 3.80 0.50 1.88Apterygidae 1.5 3.80 1.00 3.76Tinamidae 4.0 1.04 0.15 0.15Cracidae 3.0 5.58 0.29 1.58Megapodidae 12.5 5.54 0.42 2.33Odontophoridae 4.0 7.22 0.33 2.41Numidae 10.0 4.47 0.17 0.74Phasianidae 11.0 4.47 0.25 1.14Anhimidae 5.5 3.56 0.00 0.00Anseranatidae 7.5 3.56 0.00 0.00Dendrocygnidae 11.0 4.06 0.11 0.45Anatidae 6.0 4.06 0.16 0.66Outgroup 3.75

*

Trait value refers to clutch size.

1018

Identifying Distinctive Threatened Taxa Owens & Bennett


cies, the 217 species (2.3%) in the top 10 families encom-passed 19.3% of the sum of the diversification scores (

x

2

5

36.0, df

5

1,

p

,

0.0001), and the 1701 species (17.4%) inthe top 20 families encompassed 33.2% of the sum of thediversification scores (

x

2

5

11.5, df 51, p , 0.001).The sum of the diversification-under-threat scores was

58.6. Thus, 17% of the sum of the diversification scoresacross all families is due to species being classified as beingthreatened by extinction. The 20 highest ranking avianfamilies encompassed significantly more of the clutch sizediversification under threat than would be expected ifthey had been selected at random (Table 3; Fig. 3). For ex-ample, the top 10 families (11.6% of families that containat least one threatened species) encompassed 53.1% of thesum of the diversification-under-threat scores (x2 5 11.2,df 51, p , 0.001, respectively), and the top 21 rankingfamilies (23.3% of families that contain at least one threat-ened species) encompassed 70.8% (x2 5 12.4, df 51, p ,0.001, respectively). Or, in terms of species, the 42 threat-ened species (3.8%) in the top 10 families encompassed

53.1% of the sum of the diversification-under-threat scores(x2 5 17.7, df 51, p , 0.001, respectively), and the 190threatened species (17%) in the top 21 families encompass70.8% (x2 5 16.5, df 51, p , 0.001).

Discussion

Phenotypic Diversification

All taxa are not equal with respect to representing theevolutionary processes that lead to current phenotypicbiodiversity. Our method is based on the observationthat all phylogenetic divergences have not led to similaramounts of phenotypic change. In the extreme case, cer-tain phylogenetic divergences can be viewed as majortransitions in evolution (Maynard Smith & Szathamáry1995), whereas other divergences represent the forma-tion of a phenotypically homogeneous taxon swarm. Ourmethod represents an attempt to quantify this view.

Our analyses of clutch size variation among birds con-firm that it is possible to quantify differences among taxain their relative representation of phenotypic diversifica-tion with respect to a specified phenotypic trait. For ex-ample, the 217 species (2.3% of all species) in the top 10families (7.5% of all families) encompass over 33% of thetotal diversification in clutch size (Table 3).

This degree of heterogeneity in the representation ofevolutionary phenotypic diversification is exciting be-cause it stresses the extraordinarily different evolutionaryhistories of different avian lineages. It is also worrying be-cause, as Tables 2 and 3 suggest, a lot of evolutionary phe-notypic diversification is currently represented by onlysmall families, many of which are disproportionately threat-ened by extinction (Bennett & Owens 1997; Owens et al.1999). The next phase of this work is to repeat these anal-yses for a number of phenotypic traits to test whether cer-tain families are unusually good representatives of otherforms of phenotypic diversification.

Our method differs from previous approaches (e.g., May1990, 1994; Vane-Wright et al. 1991; Crozier 1992; Faith1993) because we have explicitly attempted to account fordifferences in the amount of phenotypic change occurringat each branching point in the phylogeny. Previous ap-proaches concentrated exclusively on branch lengths perse as criteria for evaluating the distinctiveness of taxa, butthey have assumed that the amount of change in a trait ateach branching point is equal. Although using branchlengths to calculate maximum phylogenetic path length is,we believe, the best way to measure “phylogenetic distinc-tiveness,” it does not take into account the differentamounts of evolutionary change that took place at differentpoints of cladogenesis. Conserving phylogenetic distinc-tiveness may not, therefore, conserve the processes thathave led to phenotypic diversity.

But do these alternative methods actually result in dif-

Figure 2. Phylogeny ( from Sibley & Ahlquist 1990) used to illustrate the calculation of the diversification score and diversification-under-threat score based on data on clutch size variation among birds. The figures in circles are the estimated ancestral character states at each node, calculated according to the correspond-ing formulas in the circles in Fig. 1. The estimates of the diversification score and the diversification-under-threat scores for these families are shown in Table 2.


Owens & Bennett Identifying Distinctive Threatened Taxa 1019

ferent conclusions about species conservation priorities?We examined this question by using estimates of branchlengths (DT50H from Sibley & Ahlquist 1990) for the fami-lies listed in Table 2. We multiplied these branch lengthsby the proportion of species in each family threatenedwith extinction (Pi) to obtain another index of evolution-ary distinctiveness under threat. We then compared thisvalue with the diversification-under-threat score (Ei) esti-mated by the methods we describe in this paper. Therewas a strong correlation between these two scores (r 50.94, p , 0.0001, n 5 14). Thus, both methods yield sim-ilar rankings for the relative importance of particular fami-lies in representing evolutionary distinctiveness. This re-sult is not surprising because we used a trait—clutchsize—that we know varies more among ancient than re-cent lineages (Owens & Bennett 1995; Owens et al.1999). We are currently investigating whether thesemethods yield different results when phenotypic traits areused that vary more among recent than ancient lineages.

Implications for Conservation Biology

Although large numbers of species are threatened by ex-tinction, the resources available for their conservation

are severely limited. Hence, if conservation effort is tobe expended on species-based approaches, it is neces-sary to decide which species should be the focus of con-servation effort (Atkinson 1989; Vane-Wright et al. 1991).The first question is whether some species are morevaluable than others, and there are two sides to this de-bate (May 1990). One side states that all species shouldbe valued equally because each one is, by definition, anindependent evolutionary product. The other side statesthat certain species are more valuable than others be-cause, evolutionarily speaking, some species are moreunusual than others. In recent years this second point ofview has led to the development of a series of algo-rithms that can be used to rank taxa with respect to phy-logenetic differentiation: higher ranks are given to spe-cies that are the only extant representative of an ancientphylogenetic lineage than to species that are representa-tives of recently diverged and speciose lineages (Atkin-son 1989; Vane-Wright et al. 1991).

Our method could also be used to prioritize conserva-tion effort. For example, Fig. 3 illustrates how familiesvary both in their representation of phenotypic diversifi-cation and in their vulnerability to extinction. Familiesin quadrant I are more vulnerable to extinction than av-

Table 3. Top 21 avian families ranked by diversification-under-threat score.

Rank* Family Common name

Number ofthreatened

speciesin family

Totalnumber

of speciesin family

Proportion ofspecies in family

threatened byextinction

( Pi )

Relativediversification

score( Di )

Diversification-under-threat

score( Ei )

1 Mesitornithidae mesites 3 3 1.00 8.1 8.12 Rhynochetidae kagu 1 1 1.00 4.1 4.135 Apterygidae kiwis 3 3 1.00 3.8 3.835 Gruidae cranes 7 15 0.47 8.1 3.85 Odontophoridae new world quail 2 6 0.30 7.2 2.46 Megapodiae megapodes 8 19 0.42 5.5 2.37 Casuariidae cassowaries 2 4 0.50 3.8 1.985 Heliornithidae finfoots 1 4 0.25 6.4 1.685 Cracidae guans 14 49 0.29 5.5 1.6

10 Orthonychidae logrunners 1 2 0.50 3.0 1.511 Otididae bustards 4 25 0.16 7.8 1.2125 Phasianidae pheasants 45 177 0.25 4.5 1.1125 Zosteropidae white-eyes 21 96 0.22 5.1 1.1145 Ciconiidae storks 6 26 0.23 4.5 1.0145 Pelecanidae pelicans 2 9 0.22 4.5 1.0165 Rallidae rails 32 141 0.23 4.0 0.9165 Fregatidae frigatebirds 2 5 0.40 2.3 0.9165 Callaeatidae New Zealand

wattlebirds 1 3 0.33 2.7 0.9165 Procellariidae petrels 32 115 0.28 3.2 0.9205 Numididae guineafowl 1 6 0.17 4.2 0.7205 Menuridae lyrebirds 2 4 0.50 1.5 0.7

Total for top 10 families 42 31.1Total for top 21 families 190 41.5Overall total 1111 58.6

Overall total in top 10 families (%) 3.8 53.1Overall total in top 21 families (%) 17.1 70.8

*Equals sign indicates that the families have tied ranks.

1020 Identifying Distinctive Threatened Taxa Owens & Bennett


erage and represent more phenotypic diversificationthan average. Families in quadrant II are also more vul-nerable to extinction than average but represent lessphenotypic diversification than average. Families inquadrant III are less vulnerable to extinction than aver-age but represent more phenotypic diversification thanaverage, and families in quadrant IV are less vulnerableto extinction than average and represent less pheno-typic diversification than average. Given this variationand assuming for illustrative purposes that the represen-tation of avian clutch size diversification is an aim ofconservation, it seems logical to give highest priority tospecies in families in quadrant I. Our analyses of clutchsize variation among birds also supports this view: de-spite the fact that the top 20 families in quadrant I con-tain only 42 threatened species (only 0.5% of all species,or 3.8% of threatened species), these 20 families repre-sent 53% of the overall diversification in clutch size thatis currently threatened by extinction. Thus, by targetingunusual families it would be possible to secure the rep-

resentation of significantly more clutch size diversifica-tion than would be expected by chance. Preliminaryanalyses suggest that the same is true for other pheno-typic traits and that, for life-history traits at least, theranking order of families is highly repeatable. This ap-proach can also be integrated with regional analyses toreveal nonintuitive differences between areas in the dis-tribution of unusual families. In the case of clutch sizevariation, for example, more of the “high priority” fami-lies in quadrant 1 are concentrated in Australasia thanany other continent (Australasia, 12 families; Asia, 11; Af-rica, 8; South America, 4; North America, 4; Europe, 3;Antarctica, 1). This highlights the disproportionatelylarge number of small and distinctive endemic lineagesin that region which are now threatened by extinction.

But what of the other side of the debate, the one thatstates that all species are equal? What would happen ifconservation priorities were assigned randomly amongequally endangered taxa? Stating that all taxa are equallyimportant in terms of representing biological diversity

Figure 3. Indices of phenotypic di-versification illustrate assignment of conservation priorities. Diversifi-cation score ( Di ) against the pro-portion of species in the family cur-rently classified as threatened by extinction. Calculations are based on data on clutch size variation among 133 avian families. The av-erage proportion of species in a family threatened by extinction is 0.11. The average diversification score is 2.26. Filled circles show the top 20 species in terms of diversifi-cation under threat score: 1, mes-ites (Mesitornithidae); 2, kagu (Rhynochetidae); 3, kiwis (Aptery-gidae); 4, cranes (Gruidae); 5, New World quail (Odontophoridae); 6, megapodes (Megapodiae); 7, cas-sowaries (Casuariidae); 8, finfoots (Heliornithidae); 9, guans (Cracidae); 10, logrunners (Ortho-nychidae); 11, bustards (Otididae); 12, pheasants (Phasianidae);13, white-eyes (Zosteropidae);14, storks (Ciconiidae); 15, pelicans (Pelecanidae); 16, rails (Rallidae); 17, frigate-birds (Fregatidae);18, New Zealand wattlebirds(Callaeatidae); 19, albatrosses (Procellariidae); 20, guineafowl (Numididae). Empty circles show other families.


Owens & Bennett Identifying Distinctive Threatened Taxa 1021

implicitly assumes that all taxa have an equal number ofclose phylogenetic relatives and that all phylogenetic di-vergences led to similar amounts of phenotypic diversifi-cation. We suggest that both these assumptions are in-valid. If it is assumed that all threatened taxa are equallyimportant, there is a risk that conservation effort will bedirected toward rare but mundane forms that are notnovel. This danger is characterized by the plight of manyoceanic island–dwelling taxa, which are often the tar-gets for conservation effort because they are dispropor-tionately threatened (Diamond 1984, 1989a, 1989b;Pimm et al. 1995). In phenotypic terms, unfortunately,oceanic islands dwellers are often little more than re-cently derived forms of an otherwise common mainlandlineage. They may, for example, have a slightly differ-ently colored head, but they are basically the same crea-ture as on the mainland. Thus, assigning high conserva-tion priority to an island form that simply has a differenthead color from its mainland sister species is implicitlythe same as deciding that diversification in head color isthe key element of biological diversity that must be con-served. Such a decision would be hard to defend on bio-logical grounds because it is the equivalent to sacrificingthe diversification that occurred over several tens orhundreds of millions of years in favor of preserving thechange that occurred in the last few thousand years. Theendemic creatures of the “great islands” such as NewZealand and Madagascar are exceptions to this rule be-cause they usually represent ancient lineages that splitaway from their mainland counterparts many tens of mil-lions of years ago (see the high proportion of such lin-eages in Table 3).

Of course, the biggest problem in assigning conserva-tion priorities is not in developing new algorithms but indeciding what exactly should be the target of conserva-tion; when people say that they wish to conserve biodi-versity what exactly do they mean? This is the subject ofongoing debate. The approach that we propose is basedon the assumption that unusual forms should be con-served because unusual things contribute more towarddiversity than do usual things. But this is only one schoolof thought. Another contends that the more specioseforms should be given priority because these currentlysuccessful forms will act as species dynamos for the pro-duction of future forms (e.g., Erwin 1991; Brooks et al.1992), whereas more ancient and unusual forms repre-sent evolutionary dead ends (Williams et al. 1994; Williams& Humphries 1994). Thus, the major limitation to usingmethods such as the one we outline is that conservationistsdo not know which traits to put into the algorithms.

Acknowledgments

We are grateful to E. Bermingham, M. Bruford, M. Cunning-ham, P. Dwyer, A. Goldizen, P. Harvey, W. Holt, G. Mace,C. Moritz, M. Minnegal, and A. Smyth for helpful discussion

or comments on the manuscript. This work was begunwhile I.P.F.O. was a National Environment Research Coun-cil, U.K., Postdoctoral Fellow at the Institute of Zoologyin London and was completed while P.M.B. was a Visit-ing Zoologist at The University of Queensland, Australia.I.P.F.O. thanks National Environment Research Council,U.K., for support and P.M.B. thanks the Department ofZoology at the University of Queensland for financial assis-tance and hospitality under its Visiting Zoologist Scheme.

Literature Cited

Atkinson, I. 1989. Introduced animals and extinctions. Pages 54–69 inD. Western and M. C. Pearl, editors. Conservation for the twenty-first century. Oxford University Press, Oxford, United Kingdom.

Bennett, P. M. 1986. Comparative studies of morphology, life history andecology among birds. Ph.D. thesis. University of Sussex, Falmer, U.K.

Bennett, P. M. and I. P. F. Owens. 1997. Extinction-risk among birds:chance or evolutionary predisposition? Proceedings of the RoyalSociety of London Series B 264:401–408.

Brooks, D. R, R. L. Mayden, and D. A. McLennan.1992. Phylogeny andbiodiversity: conserving our evolutionary legacy. Trends in Ecologyand Evolution 7:55–59.

Collar, N. J., M. J. Crosby, and A. J. Stattersfield. 1994. Birds to watch 2:the world list of threatened birds. Birdlife International, Cam-bridge, United Kingdom.

Crozier, R. H. 1992. Genetic diversity and the agony of choice. Biologi-cal Conservation 61:11–15

Diamond, J. M. 1984. “Normal” extinctions of isolated populations.Pages 191–246 in M. U. Nitecki, editor. Extinctions. University ofChicago Press, Chicago.

Diamond, J.M. 1989a. Overview of recent extinctions. Pages 824–862in D. Western and M. C. Pearl, editors. Conservation for the twenty-first century. Oxford University Press, Oxford, United Kingdom.

Diamond, J. M. 1989b. The present, past and future of human-causedextinction. Philosophical Transactions of the Royal Society of Lon-don Series B 325:469–478.

Erwin, T. L. 1991. An evolutionary basis for conservation strategies.Science 253:750–752.

Faith, D. P. 1992a. Conservation evaluation and phylogenetic diver-sity. Biological Conservation 61:1–10.

Faith, D. P. 1992b. Systematics and conservation: on predicting the fea-ture diversity of subsets of taxa. Cladistics 8:361–373.

Faith, D. P. 1993. Biodiversity and systematics: the use and misuse ofdivergence information in assessing taxonomic diversity. PacificJournal of Conservation Biology 1:53–57.

Faith, D. P. 1994a. Genetic diversity and taxonomic priorities for con-servation. Biological Conservation 68:69–74.

Faith, D. P. 1994b. Phylogenetic diversity: a general framework for theprediction of feature diversity. Pages 251–268 in C. J. Humphriesand R. I. Vane-Wright, editors. Systematics and conservation evalua-tion. Clarendon Press, Oxford, United Kingdom.

Faith, D. P. 1994c. Phylogenetic pattern and the quantification of or-ganismal biodiversity. Philosophical Transactions of the Royal Soci-ety of London Series B 345:45–58.

Felsenstein, J. 1985. Phylogenies and the comparative method. Ameri-can Naturalist 125:1–15.

Harvey, P. H. 1994. South Africa evolving? Trends in Ecology and Evo-lution 9:416–417.

Harvey, P. H. 1995. The WORLDMAP debate. Trends in Ecology & Evo-lution 10:82–83.

Harvey, P. H., and M. Pagel. 1991. The comparative method in evolu-tionary biology. Oxford University Press, Oxford, United Kingdom.

Mace, G., and S. Stuart. 1994. Draft IUCN redlist categories. Species21:13–24.

May, R. M. 1990. Taxonomy as destiny. Nature 347:129-130.

1022 Identifying Distinctive Threatened Taxa Owens & Bennett


May, R. M. 1994. Conceptual aspects of the question of the extent ofbiological diversity. Philosophical Transactions of the Royal Societyof London Series B 345:13–20.

Maynard Smith, J., and E. Szathamáry. 1995. The major transitions inevolution. Freeman, Oxford, United Kingdom.

Owens, I. P. F., and P. M. Bennett. 1995. Ancient ecological diversifica-tion explains life-history variation among living birds. Proceedingsof the Royal Society of London Series B 261:227–232.

Owens, I. P. F., P. M. Bennett, and P. H. Harvey 1999. Species-richnessamong birds: body size, life history, sexual selection or ecology?Proceedings of the Royal Society of London Series B 266:933–939.

Pagel, M. D. 1992. A method for the analysis of comparative data. Jour-nal of Theoretical Biology 156:431–442.

Pimm, S. L., M. P. Moulton, and L. J. Justice. 1995. Bird extinctions in thecentral pacific. Pages 75–87 in J. H. Lawton and R. M. May, editors.Extinction rates. Oxford University Press, Oxford, United Kingdom.

Purvis, A., and A. Rambaut. 1995. Comparative analysis by independentcomparisons (CAIC): an Apple Macintosh application for analysingcomparative data. Computer Applications Biosciences 11:247–251.

Sibley, C. G., and J. E. Ahlquist. 1990. Phylogeny and classification ofbirds. Yale University Press, New Haven, Connecticut.

Sibley, C. G., and B. L. Monroe. 1990. Distribution and taxonomy ofbirds of the world. Yale University Press, New Haven, Connecticut.

Vane-Wright, R. I, C. J. Humphries, and P. H. Williams. 1991. What toprotect? Systematics and the agony of choice. Biological Conserva-tion 55:235–254.

Walker, P. A., and D. P. Faith. 1995. DIVERSITY-PD: procedures forconservation evaluation based on phylogenetic diversity. Biodiver-sity Letters 2:132–139.

Williams, P. H., and C. J. Humphries. 1994. Biodiversity, taxonomic re-latedness, and endemism in conservation. Pages 269–287 in P. L. Fo-rey, C. J. Humphries, and R. I. Vane-Wright, editors. Systematics andconservation evaluation. Clarendon Press, Oxford, United Kingdom.

Williams, P. H., and C. J. Humphries. 1996. Comparing character diversityamong biotas. Pages 54–76 in K. J. Gaston, editor. Biodiversity: a biol-ogy of numbers and difference. Blackwell, Oxford, United Kingdom.

Williams, P. H., C. J. Humphries, and R. I. Vane-Wright. 1991. Measur-ing biodiversity: taxonomic relatedness for conservation priorities.Australian Journal of Systematics and Botany 4:665–679.

Williams, P. H., K. J. Gaston, and C. J. Humphries. 1994. Do conserva-tionists and molecular biologists value differences between organ-isms in the same way? Biodiversity Letters 2:67–78.

Quantifying Biodiversity: a Phenotypic Perspective

Documents

Transcript of Quantifying Biodiversity: a Phenotypic Perspective