Systematics and global biodiversity strategy

Abstract

This article is based on a paper by C.J. Humphries, R.I. Vane-Wright and P.H. Williams, “Biodiversity reserves: setting new priorities for the conservation of wildlife”, which appeared in Parks (IUCN) 2(2): 34–38, 1991. Dick Vane-Wright presented some of these ideas in a talk to the Royal Entomological Society of London on November 6th, 1991, entitled “Entomology, systematics and biogeography: a global strategy for the conservation of biodiversity.”

Full text

Guest Article
Systematics and the global biodiversity strategy
R.l. \/ane-Wright, Biogeographt'& Coruervation Project. Biodiversity programme, The
Natural Historl' Museum, Cromv.ell Road, London SW7 SBD, UK.
This article is based on a poper bt'C.J. Humphries. R./. Vane-Wright and p.H. William,
"Biodiversit,- reserves: sening new'priorities for the consert,ation of wildlife", whiclt
appeared in Parks (IUCN) 2(2): 3a-38, 1991. Dick Vane-Wright presented some of these
ideas in a talk to the Societv on November 6th, 1991 . entitled "Entomolog!,. systematics and
biogeographv: a global strategv for the consen,ation of biodiversitv.'.
The most universol Emlitt b diversin.
(Michael Eyquem de Montaigne. 1580)
Biodiversitv is the sheervarietyof life-the varietv of species. including the genetic variation
betu een and within them. The number of living species is estimated to be in the range 5-50
million. of which less than 2 million have been named or even recognised (May. lggl).
Roughll' half of all those that have been named are insects, u'rth beetles alone perhaps
making up one fifth of the total.
Faced with such huge numbers. and the rapid ecological changes affecting all areas of the
world. many entomologists and other biologists are convinced that a penod of massive
extinction is imminent (Mvers, 1989). In the next feu decades \\,e may iyitness the loss of
millions of the species that make up the web of life on our planet - the Lnique svstem from
which we evolved, and upon which wc are totally dependent.
Some plant species can be consen'ed in botanic gardens or seed banks. A few large and
exciting vertebrates, such as the tieer. could be maintained in zoos. Hower,er. it is incon-
ceivable. with current resources. to sustain more than a tiny fraction of invertebrates sepa-
rate from their natural environments. Sunival of the majoritv of species will depend on
bettermanagement of allecosystems, and on judiciousselection of nationalparksand other
wildlife reserves. The problem of how to select reserues to maximise global biodiversity
protection is the topic of this article.
Protected areas are currently chosen on many different, often conflicting criteria. These
include the needs of indigenous peoples and for recreation and material resources, as well
as species conservation. The recognition of this urgent new need, the consen,ation of biodi-
versity itself, demands rapid development of efficient means of evaluation for this specific
purpose (Pressey & Nicholls. 1989: Margules & Austin. 1991; Rebelo & SiegfrieA, tggO,
and in press).
Adequate protection forbiodiversitywillrequire a globalstraregy (Reid, Barber & Mil-
ler, 1992), including the establishment of a worldwidi network of reserves to provide re-
fuges for as wide a variety of animal and plant life as possible. The task involves deciding

hou to recognise and set priorities for the selection of reserves so that. as sites are added to
the network, they progressively'protect the maximum possible diversity (Margules et al.,
1988). To do this we need to measure biodiversity'in a wav which will allou'us to compare
areas on both absolute and relative scales. These requirements can be met by measuring
three properties of floras and faunas: species richness. complementarity and taxonomic
difference (Vane-Wri-eht er a\.,1997: Williams et a\.,1991).
Complementarity
Imagine being given the problem of picking tu'o areas from three. as biodiversity reserves.
You can choose any two. but you must put them into priority order because vou can only
afford to buy one of them at present. The second area of vour choice will remain unpro-
tected. while the third area willbe redeveloped immediately. as an olvmpicsportsstadium.
The onl,v other information you are given is a list of beetle species for each area, and that the
cost of conserving the areas would be equal. Usine species richness as a working measure of
biodiversitv. vourfirst choice will be the area with the most species: it offers the potentialof
consen'inq the largest "amount" of biodiversitv.
What willbe your second choice -a choice u'hich u'ill result in the immediate extermina-
tion of virtually all wildlife in the area not selected? Will it be the area with the second
highest number of species? In terms of setting priorities. the correct answer is "not necessar-
ily". We need to identi! as second choice the site containing the highest number of additio-
nal species, i.e. those not represented in the first area.
Putting figures to the example. imagrne that area t has 60 species. area2has 55 species of
which 10 are not found in areas I or 3. and area 3 has ,15. of which 30 are not found in either
area 1 or2. Areas I and 2 are woodlands, area 3 is a heath. The numberof species in all areas
combined is called the complement. a total of 100 in this case. If we take the richest area
(area 1.60 species) as our first choice. this determines the residual cornplemenr (the 40
species not represented in area 1). The idealsecond choice is then the site giving the greatest
additional biodiversit!'to the first - the site having the highest proportion of the residual
complement. Area 2 offers a l0% increment. but area 3 offers 30qc . So we would buy the
richer and sacrifice the poorer of the two woods. and hope that funds might be found to
protect the heath.
Taxonomic difference
In order to rank sites in this way it is necessary to measure their diversities in terms of
absolute values. and also in terms of their relative contributions to residual complements.
In the example. species richness was employed for both purposes. However, thissimplest of
measures, which treats all species as equally valuable, is not alu'ays appropriate. Suppose
the 10 species unique to area 2 belonged to 10 separate and highly distinct beetle families
unrepresented in the other areas. while all but two of the species restricted to area 3 were
members of a single species complex belonging to a family well represented in area 1.
Armed with this information, you might want to reverse your decision about the second
priority area-or even the first. 'Megadiversity', as measured simply by species richness, is
by no means always best.
50

We thus have an intuitive idea of taxonomic distinctness or difference.ltis based on an
appreciation of the taxonomic hierarchy. Some species are very closely related to each
other (low relative rankl, while others are very distinct. Single species can represent whole
superfamilies (Fig. 1). orders or even phyla. In terms of biodiversity, to equate one of these
with a single member of a sibling species complex seems quite wrong. itrir i, one of the
reasons why it is easier to get money to protect the giant panda than an equally threatened
species of rat. If there were many species of pandas, but onlv one sort ol rut. the reverse
would probably be true. In species conservation terms we value rarity; in biodiversitr terms
we value difference.
In practice' real faunas and floras consist of a wide varietl,of more or less distinct species.
How can we devise an acceptable measure that willreflect taxonomic difference? Mv col-
leagues Chris Humphries and Paul Williams are appraising a variety of measures designed
to do this. andwe have some promisingcandidates (Williamse rat.,1991. and in press). The
measures can also be made sensitive to species richness. We call these measures taxic diver-
siq, indices.
Priority areas
Using a taxic diversitf index in a prioritv area analvsis applies differentialweiehting to the
species. \\'eighting can be fixed or relative. Species richness is a special tr.pe of fixed u eight-
-i
,:--\
Fig' l' Taxonomic difference. The Australian dragonfly Hemiphtebia mirabitis. the only
extant member of the superfamily Hemiphlebioidea. Althougtrnot impressive bv dragon-
fly standards in terms of size or colour. this little species is abundantly distinct (Davies,
1985)' A recent study by John Trueman (Australian National University) suggests rhar the
Hemiphlebioidea evolved before a number of extinct groups of dragonflies known only
from Permian fossils. and thus would appear to have a minimum age of 240 million years. As
a result' it should score heavilf in favour of conserving the wetlands of Wilson,s promon-
tory. victoria, the only ecosystem in which Hemiphtebia now occurs.
51

ing: all species are simply given unit weight. Root weighting ( Fig. 2) is another fixed weight
index. species being valued for difference according to their position in the taxonomic
hierarchy (Vane-Wright er al.,l99l). For priority area analvses. the wei-qhts allocated are
simpl-v.' summed for all species represented within an area. to sive an area-value. Calcula-
tion of increments based on residual complements is equall-t" straightforward (Fi-e. 2).
Relativeweights. such asthe dispersion measureswe nowfavourforthistask (Williamser
al., 1991, and in press). are more complex to apply because they are not simplv additive at
each step - but the fundamental principle remains the same. The ideal first choice is asses-
sed by the maximal score: subsequent areas are assessed by'calculating increments based
C
9.3 3.3 4.3 7.3
35 46 78
22
Fig. 2. The basics of priority analy'sis - species. areas, complementaritv and taxonomic
difference. A theoreticalexample forfive species (A-E).and three areas (R1-R3) each with
three species (dots). Taxonomic differences calculated by the root weight method gives a
set of additive weights (column W) reflecting the position of each species in the taxonomic
hierarchy. Total diversity for A-E (complement) and each area is given in rowT. Scores as
percentages of the complement are given in row P1 . Row P2 grves diversity increments for
Rl and R2, based on residual complement after selecting R3 as first choice. Note that
p3max > Plmax, and that Plmax * P3max = 1O0Vc. indicating priority is R3 then R1, and
that Rl + R3 form critical faunas set. (Reproduced from Vane-Wright et al., 1991: fig. 5,
with permission of Elsevier Science Publishers Ltd.; see also text.)
52
W R1 R2R3
A1o
a
o
oo
oa
1.3
2
11
P1
P2

on their representation of the residual complement in combination with the species already
selected. In practice, foranythingother than afew areas and species, a computer programis
essential. Williams' WORLDMAP program, originally written for his work on bumble
bees. is being developed for this purpose.
A priority analvsis is completed b1'the step which finally accounts for the entire comple-
ment (Figs 3' 4)' This identifies a list of sites identrcal to the critical fauna (or florajser
(Ackery & Vane-Wright. 1984). and places them in an optimised step-wise sequence. The
critical faunas set is the minimum selection of particular areas containing all species under
analysis. In realitv it will rarely be possible to establish reserves in areas corresponding to
the optimal sequence. Rebelo & Siegfried (1990) have described this as the conflict be-
tween "ideal and real-world options." Without precise calculation it can be very difficult to
assess the best of manv alternative options. orthe furtherconsequencesof taking particular
decisions. The WORLDMAP program enables suboptimalchoices to be assessed rapidly
and interactively. The second range of data plotted in Fig. 4 gives an idea of how this can
work.
Fig. 3. Priorities for bumblebees of the Bombus sibiricus group - a critical faunas set and
priority analysis. The 43 species of the sibincu.s group occur in t j0of the grid squares shown
on the map (Williams, 1991). To complete the priority areas analysis riing root weighr as
the taxic divenity index requires 13 areas (Williams et a\.,1991). ihe direisity accumula-
tion sequence is shown in Fig. 4. (Reproduced from Vane-Wright et aI.,1991: fig. 7, with
permission of Elsevier Science publishers Ltd.)
53

This world, where much is to be done and little to be kno*n
(SamuelJohnson, l17O)
The svstem outlined is under development. A further essential step will involve finding a
satisfactory way to combine data for different groups of organisms. We envisage analysing
100 or more well-known groups of animals and plants. including several groups of insects.
on aglobalbasis. These willinclude. amongst the Lepidoptera. swallowtailbutterflies (the
first insect group to be the subject of an IUCN Action Plan: Neu'& Collins. 1991). milk-
weed and glasswing butterflies, and hawk moths.
Much of this work on Lepidoptera will be carried out bv myself and colleagues in London
Fig. 4. Optima and suboptima-complementaritv versus megadiversity in the search for an
effective solution. The main data range shows taxic diversity accumulation at each step of a
priority area analvsis (using root weight) for bumblebees of the Bombtrs sibiricus group,
starting with the Chinese region of Gansu (1). and finishing with northern California (13)
(from Williams et al.,l99l; see also Fig. 3). The second range of data (five columns) shows
stepwise accumulation of diversitv (on same root weight scale) if regions are chosen to
maximise local species richness without using complementarit,v-. tn this suboptimalsequ-
ence.area I : Ecuador(l0spp.).2: Kashmir(9spp.),3: Gansu(9spp.),4: northern
Peru (8 spp.),5 : central Bolivia (8 spp.); all other areas have less than eight species. Note
that by third step optimised sequence exceeds 50% taxic diversity. whereas aggregate for
the tive 'megadiversity' sites does not even reach 50%.
54

(including Campbell Smith. George Beccaloni. and Ian Mitching). Other systemarists ar
the Natural Histon' lv{useum will study'and analvse groups of plants. However. to get a
large sample of groups for u'hich the hieher classification. species limits and distributional
ranges can be sufficientlv u ell understood. we u'ill need the help of many more scientists.
Currentlv we are tning to formulate plans to bring together an international eroup of
specialists willing and able to contribute to the task of slobal systematic consen,ation eva-
luation.
A somewhat different approach u'ill inr olve an analvsis of the global distributions of all
familiesof flou'erine plants and beetles. Thisu'ould be a rapirJ u,avof eyaluating one third of
all knou'n terrestrial species diversitr'. and u'ould help test assuffiptions about ho*.repre-
sentati\.e smaller groLlps are of hiodiversitv as a ri,hole.
The basic fle.ribiliry' of the method. applving weights and assessing suboptimal choices.
means that it could also be extended to gir e a general planning svstem. Other factors such as
raritv. ende'mism or minimum viable population size. or el.en non-bioloeicalfactors such as
consen'ation costs per unit area. could be built into an increasingly elaborate prosram.
Such svstems can be refined and extended indefinitelv. For example. WORLDMAp has
clear potential for site monitoring. and as part of an earlv-u'arning system to alert us to ne*,
threats. as called ft r in slobal strateq\.(Reid er al..l9g2). Rather than develop WORL_
DMAP in isolation. it is more likeh'to be incorporated u,ithin LS. technologv as a tool
dedicated to the assessment and evaluation of global or regional biodiversity.
Iner"itablr'. debate over methodolosical detail has alreadv besun (e.g. N{av. 1990:
Altschul & Lipman. 1990: Faith. in press). While this is healthv scientificallv. *,e musr
remember. as Samuel Johnson understood. that rvhat is needed most is action. Efficient
selection procedures are ven important to ensure value for moner.. but action is e'en more
vital. The proposal that the UN should inaugurate an International Biodiversitl,Decarle,
1994-2003 (Reid er al.. l99l). recognises that our options for choice are decreasing da1,' by
dav' In the future. the resen'e network u'ill be called upon to supplr-the raw materials to
regro\r and restore to the Earth that "most universal of all qualities", diversitr'. S1,'stematists
have a considerable responsibilitl,' to ensure that the netu,ork is capable of delivering as
much diversitt'as possible. In truth. to misquote Rov Tav-lor. we need a diversitv of di'ersi-
ties.
References
Ackery. P.R. & \'ane-Wright. R.I. ( 1981) Milkv,eed Butterflies. Cornell Universin, press.
New York
Altschul. s.F. & Lipman. D.J (1990) Equal animals. Nature.34g: .193-494.
Davies. D.A.L. (1985) Hemiphlebia mirabili.s Sely,'s: some notes on distribution and con-
servation status (Zvgoptera: Hemiphlebiidae). O donatologica. 7 4: 33 1 -339.
Faith' D.P. (in press) Conservation evaluation and phvlogeneticdiversity. BiologicalCon-
serv,atiott.
Ivlargules. C'R. & Austin. l\{.P. ( 1991 ) lt'ature Conservatio,.t: cost effecriv,e biological sur-
v,evs and data analt'sr. CSIRO. East Melbourne.
55

lVlarsules. C.R., Nicholls. A.O. & Pressey. R.L. (l9SB) Selecting net*'orks of resen'es to
maximise biological diversity. B io lo gic al C o ns e n' ario n . 13 63-7 6.
Ma1'. R.M. ( 1990) Taxonomy as destiny' Nature.3lT:129-130'
Mal.. R.M. ( l99l ) How many species? Philosophical Transactiorts of the Royal Sociery of
London (B). 330: 293'304.
Myers, N. (1989) A majorextinction spasm: preclictable and inevitable? Pp 109-120 in
Western. D. & Pearl. M. (eds) Conservation for the Twentv-first Centurl'. Oxford
Universitv Press. Nerv York.
New. T.R. & Collins. N.M. (1991) Swallowrctilbutterflies: an action planfor their consen'a-
rron. IUCN. Gland.
presser,. R.L. & Nicholls. A.O. ( 1989) Eficiencv in conservation evaluation:scorinqversus
interactive approache s. Biological Conser,-atiort. -50: I 99-21 8.
Rebelo. A.G. & Siegfried. \l'.R. ( 1990) Prorection of Fynbos vesetation: ideal and real-
world options. Biological Conservation. 51 15-31'
Rebelo. A.G, & Siegfried. W. R. (in press) Where sht-ruld nature resen'es be located in the
Cape Floristic Region. South Africa'l lv{odels for the spatial configuration of a re-
sen,e network aimed at maximising the protectic-rn of floristic diversiN'. Conservation
Biologt'.
Reid. W.V.. Barber. C. & Miller. K.R. ( 1993) Gtobat biodiversitl' strareg\:: gidelines for
ttction to sove, stud-r- and use Earth's bioic wealth sustainably and eqttitabl.r. World
Resources Institute. New York'
Vane-Wright. R. I.. Humphries. C. J. & Wiliams. P.H. ( 1 99 I ) What to protect? - s,vstema-
tics and the agony of choice. Biological Consen'ation,55l.235-251.
Williams. P.H. (1991)The bumble beesof the Kashmir Himalaya (Hymenoptera: Apidae,
Bombini). BulletinoftheBritbhMrueumofNaruralHistory,Entomology,60: l-203.
Williams. p.H., Humphries, C.J. & Vane-Wright, R.l. (1991) Measuring biodiversitv:
taxonomic relatedness for consen'ation priorities. A wtralian Sy stematic Botany, 4:
665-679.
Williams. p.H.. Vane-Wright, R.I. & Humphries. C.J. (in press) Measuring biodiversity
for choosing conservation areas. In LaSalle. J. (ed.) Hymenoptera and biodiversity.
CABI.
56
Suggested citation:
Vane-Wright, R.l. 1992. Systematics and the global biodiversity strategy. Antenna,
London 16:49-56.