M e t o p o p u l o t i o n
E c o l o g y , G e n e l i c s , q n d E v o l u t i o n
BIOLOGY
Edired by
llkka Hanski
Department of Ecology and Systematics University of Helsinki
FIN-00014 Helsinki. Finland
Michoel E. Gi'lpin
Department of Biology
University of Califomia, San Diego La Jolla. Califomia
ACADEIUIK PRTSS
San Diego London Boston New York Sydney Tokyo Toronto
The Melopopulotion Approoch, lfs History, Concepluol Domuin, ond Applicotion to Conservulion
llkka Honski Doniel Simberloff
I. INIRODUOION
At no period in the history ofecology has the spatial structure ofpopulations and communities been entirely ignored, but the role that space plays in forming ecological pattems and in molding pfocesses has been viewed very differently in different times (Mclntosh, l99l). In the 1960s and 1970s, theoretical ecology was largely focused on issues other than spatial dynamics (May,l916a), with notable exceptions (MacArthur and Wilson, 1967), and field ecologists tended to follow suit. Today, space is in the forefront and space is introduced in various ways into all fields of ecology and population biology more generally. Whether one is interested in processes occurring at the level of genes, individuals, popu- lations, or communities, spatial structure is widely seen as a vital ingredient of better and more powerful theories, and good empirical work involving space is seen as a great challenge (Kareiva, 1990).
Five years ago, before the publication of the predecessor of this volume (Metapopulation Dynamics: Empirical and Theoretical Investigations, Gilpin and Hanski, 1991), the metapopulation concept was new to most biolo-eists. Since then, literature on metapopulations has grown exponentially' with adoubling time of less than 2 years (Fig. 1). The metapopulation concept has by now been firmly established in population biology and beyond; we review and analyze il this
Metapopulation BioIo$
Copy.igtrt O 199? by Academic Press, Inc All rights of reprcduction in any fom reseryed'
lcuJelut ol fIe{II ,(lenbe spnpr^rpu ile qlp\'srnlJruls uortepdod cllJnuued g elunssu leql uoqceretur flrunruuoc pue 'scrleue8 'quro-d uorlelndod ',(qderSoruep Jo slapou prepuers Jo esoql qlrl\ lsBIuoJ sesrruerd asarll 'uorlruuxe 8urzno11o;
lueuqsqqelseer uorlelndod yo ,{lqrqrssod eql Surpnlcur 'srrureu,(p lecol uo lceJJe euos seq suorlepdod Ierol orll Suoue uorlur8ru 1eql pue suorlulndod Surpeerq leool Jo seEelquesse olur poJnlcluls ,(leqeds ere suouelndod teqr ere ,(3o1orq uopelndod o1 qceorddu srql w sesrurerd fe1 om1 eqt lloqstnu e ur lnq (^\oleq uerrr8 sr uorlucrldxe alelduoc arotu v iqceordde uorlelndodetau eql sr leet\
'suortecrldde puu ,(8o1orq uorlulJesuoc o1 ldacuoc uoqelndodeler.u eql Jo peards eql reldeqc
eql ur sredud Jo reg.,,nu p1o1 e,rqcedser eqr fq pezrprepuuls 's66r-0r61 rn eseq
"* J;tT:"t:
ur..uor1e1ndode1eu,, pu" ..{qderSooSorq pu€lsr,, spro^r ,{e{ eq1 or suoqetn Jo s.laqunN I ilngH
00'0
z0'0 r0'0
90'0 80'0 0 r ' 0z r ' 0
t I o 9 t ' 0 8 t ' 0
00'0
z0'0 r0'0 90'0 80'0 0 r ' 0
z|o
v L ' 0 9 t ' 0 8 l ' 0
o - a
N q )
o o .
J O )
U , a
No o-
fc
3u o
o
o. gtt
qt q,
o o - :t o) v | a
F'o o-u
c3
Ito o
|l0il0qu5 Pru00 Puo Dlsuo|.| Dllll
I lhe Meiopopulotion Approoch
with any others. Population biology has made productive use of such models for at least 100 years, but today there is a distinct need to account for the position of individuals and populations in space. This need has arisen from the intrinsic development of population biology as a science, but the trend has clearly been strengthened by the demand for professional advice on environmental issues typ- ically involving space.
In the past few years, metapopulation studies have shed new light on such phenomena as patterns of distribution and population tumover dynamics in frag- mented landscapes (Hanski, this volume; Harrison and Taylor, this volume;
Thomas and Hanski, this volume; Smith and Gilpin, this volume; van der Meijden and van der Veen-van Wijk, this volume), landscape ecology (Wiens, this vol- ume) and community structure (Holt, this volume), population viability and time to extinction (Gyllenberg et al., this volume; Foley, this volume), coexistence of competing species, and of prey and theirnatural enemies (Nee et al., this volume), evolution of migration rate and other life-history traits (Olivieri and Gouyon, this volume), ecological consequences of migration (Ims and Yoccoz, this volume;
Stacey and Taper, this volume), unexpectedly high levels of inbreeding and low heterozygosity in natural populations (Hedrick and Gilpin, this volume), patterns of genetic differentiation (Giles and Goudet, this volume), adaptation (Barton and Whitlock, this volume), and coevolutionary processes (Frank, this volume). As is apparent from the citations, these developments are well represented in the chapters in this volume, which provide an excellent entree to the literature at large.
There are many advantages of a metapopulation approach, but success may also breed problems. As in any scientific field experiencing rapid growth, there is the danger of blurring of concepts. There is the temptation to view any system with any kind of pafchiness at any spatial or even temporal scale as a "metapop- ulation." Harrison (1991,1994b; Harrison and Taylor, this volume) cautions us about this tendency. Anticipating the kind of verbal entropy that has enveloped many terms in population biology, Hanski and Gilpin sketched in the 1991 vol- ume the meaning of the term "metapopulation," highlighting issues of scale, hi- erarchy, and a requirement for some population turnover. We feel a need to dwell on the same issues in this chapter, and we provide a revised succinct glossary of the commonly used terms in the literature. First, however, let us examine briefly the history of the metapopulation concept.
II. BRIEI HISTORY ()F MEIAPOPULATION STUDIES
The metapopulation concept has a pedigree dating back to the early part of this century, but until recently this tradition played only a minor and episodic role in the intellectual advance of population biology. For a long time. the pre- vailing view was one emphasizing persistence and stability of local populations.
or as Mclntosh (1991) put it, "the great tradition of balance of nature. going back
',{3o1orq uopelJesuoc pue suoutlndodeletu uo uollJes eql uI './(oleq ,(e1ep srqt roJ suoseeJ elqlssod eql 01 untal 3-\\ 'r I '3rg) ssecer ;o sree,(
97 Klteau;o poued e,{q pe.,rro11oy se,l pepr,ro:d sul\:_1 lEql perl
^lrea eql leql 'q8noql '3ur1zznd sr t1 ''(3o1orq uorlelndoduleru .uerodueluol -Io 3uruurSeq eql s>IJeru {Jo.^A s.sul^eT '(serpnls uorlelndodeletu ..leclssel-1.. uI lr \otxnl uorlelndod uo srsuqdura eqt acueq) uortelndod Iecol P uI spnpl.\Ipur -lLr \qlEap pue squlq ol peue{q uorlelndodeletu € uI suoll€lndod 1eco1 Jo uollJull\e puE uollezluoloJ qu,t\
'suorleyndod;o uorlelndod e slseSSns JIeslI prolr\ eql'0/61 ptre '86961) 6961 ul surleT pJeqJIU Jo s>lJol\ eql uI pecnpoJlul sen\ .,uor1e1ndodelaui.. trlel rql
'.\\olaq ssnJSlp a.\\ sP 'sellluntutuoc sercedstllnu qll,\\ peuecuoc,(peurrd eJeA\ uoslli[ pue lnquyJ€tr\
Jr ue^e 'ldacuoc uorlelndoduteru eql qll^\ uoruruoc uI qcnu seq .iqde:ioaSorq puelsr Jo,(roeql crureufp (tget'9961) uos1t16 puB rnquvs€I tr aqJ'sreqlo \\3-J e ,{lpelqnopun puu '(1161) IFpBD 'GSOD ue^el{ pue qcllrqg '(SSOtl roog uap '(gSOt) re>leJJnHJo s{ro,4A ul's096I pue s096I eql uI eJuotslxe }etnb E p€q ssel -eqteuouldecuoc uorlelndodeleu luerdrcul eqJ 'uollo8ro; ,{1pn1ue^o pue parouit fle8rel sum eceds ut sctrueu,(p uorlelndod lnoqe uollou qcJlg puu sque,l.rolpuY eq; 'uorluyn8er uorlelndod luepuedep-fllsuep Jo ldecuoc eql Jo uollJeier leorroS -s1ec ,(1reeu Jleql sI uoseoJ eql eAeIIoq e11 iacueldecJe Jeplt\ urc8 ol IIBJ seepl rreql plp fq,tr 're.ternoH '(t8'd'tg6I 'qc4g pue eqile.^derpuy) ,.ree,( lxeu ro lea,t lxeu perdnccoeJ pup . Aorloruol luecan etuooeq ,{eu fepol perdncco ere 1eq1 s1ods,.
:uoueruoueqd uolu(uoc a s€,ld uollcurlxe uotlelndod pcol leql .4AeI^ eql peleo -o,rpe (yg61) I{crlfl pue eqUB^\eJpuV Jerrq u1 'sellllecol pel€c€^ 1e suorlelndod Jo lueruqsqqelseeJ Jo ,(lqrqrssod eql pezruSoceJ osle lnq 'suollcullxe 1eco1 luenb -eg pelueuncop 'suorlelndod yo suotlullltso plll'r pezrseqdtue ,,{eq1 'elqeldecceun ..sJolJBJ luepuedep-,{llsuep Jo eur3op,, eql punoJ qorlg pue equ€^\erpuv '.{8o1oce uorlelndod lcesul tuor; ecuerredxe eplA Jrel{l uo Sutrtrerg'f3o1oce lerulue uo 1xe1 peqsrnSurlsrp rreql peqsrlqnd qcrrg pu? eqlre.{\erpuv ueq^\ 6t96I eroJeq pere -prsuoc lou ere,^A ldecuoc uorlelndodeleu eql Jo suotlecrldtut pcrSoloce eq;
'(t66t 'eeN lV66t "to p uolwel:1661 ',(u6) .{1luecer Illun luau -tuoc lnoqlr,4A peuletuer uollceuuoJ leql rnq 'selpnls uorlelndodeleru o1 ,{lpcrler -oeql pue flenldecuoc pelul ,{1eso1c s€ uees ldou eJE (eunlorr srql ''lD /, eeN :1661 ',(eyq puB uosrepuv eas:LZ6I'>lcupue;c141 pue {renuox '606I 'ssou) ,{Solorrueprde uI selpnls e,rtlelttuenb Surreeuor4 '@g6el 'I{sueH ees 'uorssnc -srp q8noroql erolu e rog) suorlelndod peus;o scrleue8 pue ,(3o1oce uo (196I) elloru€T pue '(8€6I) relrq '(9961) ttoc,(og Jo selpnls ,{q ecuelsut roJ palues -erder ',ftnluec sr{l Jo JI€q lsrg eql ur suorl€lndod perntcnrls ,(1pr1eds uI tserelul peleln{urls e,req ,{eu ryom s.1q3rr16 '(eurn1o,t slql '{coltlqi[ pue uouug) eJuls Ja^a Iapou pelsel elilll pue 'poolsrapun ,(llcepedut 'SutnSulut uB peulEtuer seq ,(roeql ecuuleq Surgrqs s.lq8lJ,t 'suollezluolocer pue suollJullxe Iscol eru erel{]
yr flprcedse 'suorlulndod porntcnrls ,{lpueds ur ,{lprder peecord lq8lru uollnlo^e ler1t lq8rsur eql p€q (Ofet 'tgOt) 1q8rr16 IIed\eS',(3o1otq,{reuot1n1o,r.e u1
..'sseuuopupr pue uollcullxe ;o slq8noql peroq -qB pue urnrrqrlnbe ro ',(cuelsuoc 'flreueSouroq ernluu ol pelndrut '.{ltnbrlue ol
iloF0qurs l0ru00 PUO rlsu0H D)Plll
-_--
I The MeioPoPulolion APProoch q
III. (()NGPIUAL DOMAIN AND MEIAPOPUI.AIION APPROAffES
A fundamental assumption of the original metapopulation concept (Levins, 1969a) is that space is discrete and that it is possible and useful to distinguish between habitat patches that are suitable for the focal species and the rest of the environment, often called the matrix. In this respect the metapopulation approach is closely akin to the dynamic theory of island biogeography (MacArthur and wilson. 1967) but differs from landscape ecology (wiens, this volume). The metapopulation concept also presumes that the habitat patches are large enough to accommodate panmictic local populations, but not larger. other fields of ecol- ogyareconcernedwithspatialpatchiness'buteitheratasmaller(foragingthe- oiies; Krebs and Davies, 1gg4) or at a larger scale (e.g., much of landscape e c o l o g y : F o r m a n a n d G o d r o n , l g 8 6 ; G A P a n a l y s e s : S c o t t e t a l " l 9 9 l ;
geogr:aptri"at ecology: Ricklefs and Schluter, 1993) than the scale of (panmictic) io"a-t ptpulations. The concept of an ideal metapopulation a la Levins includes three tther simplifying assumptions: habitat patches have equal areas and isola- tion, local populations in the metapopulation have entirely independent (uncor- related)dynamics,andtheexchangerateofindividualsamonglocalpopulations is so low that migration has no real effect on local dynamics in the existing populations: local dynamics occur on a fast time scale in comparison with meta- population dynamics'
No real metapopulation completely satisfies all these requirements' However, the more specific assumptions, such as equal patch areas and isolation, can be relaxed wiihout need for a major conceptual amendment. This is not unlike how the population concept is used in population biology: no real population com- ptetelysatisResallthecriteriaofanideal'closedandpanmictic,population.What reallymatters is the notion of discrete local breeding populations connected by m i g r a t i o n . W e s u g g e s t t h a t i f t h i s a s s u m p t i o n c a n n o t b e d e f e n d e d , s o m e o t h e r approach should be used instead of the metapopulation approach; and conversely' ttre more distinct and smaller the local breedin-e populations are, the more useful t h e m e t a p o p u l a t i o n a p p r o a c h i s l i k e l y t o b e . H a n s k i a n d G i l p i n ( 1 9 9 1 ) u s e d p o p - ulation turnover, local extinctions and colonizations. as the hallmark of true meta- populations. By this definition, the mainland_island SyStemS studied in the dy- namic theory of island biogeography and in recent metapopulation models (Gotelli, 1991; Hanski and Gyllenberg. 1993) would not count as metapopula- tions. Following the cunent usage of the term. we now include mainland-island structures among other metapopulation structures'
It has been suggested that "much" mi-eration among local populations makes the metapopulation approach less useful (Harrison, 1994b). While it is true that the classical concept (Levins, 1969a) implicitly assumes a low migration rate' so l o w t h a t m i g r a t i o n p l a y s n o r o l e i n t h e d y n a m i c s o f e x i s t i n g l o c a l p o p u l a t i o n s , more recent theoretical (Hassell el a/.. l99la, l994,Gyllenberg and Hanski' 1992;
N e e e r a / ' , t h i s v o l u m e ) a n d e m p i r i c a l w o r k ( H a n s k i e t a l ' , | 9 9 5 a . b ) h a s m a d e g o o d u s e o f t h e m e t a p o p u l a t i o n c o n c e p t e v e n w h e n S o m e t e n s o f p e r c e n t s o f
rpr^\ spep lepour eqt reqteq^\ uo peseq 'sesselo o.,ru ottl .{ltuetue,ruoc pepl^Ip eq uec suorlulndod pernlcruts flprpds o1 seqceordde Sqtapolri 'suortelndod pernl -crulsun ,(lpqeds sournsse ,(8olorq uorlulndod o1 qceorddu IBUoIIIpBJI eqJ
s€lPDoiddv 6ur;epoyl'y
'(arun1o,t srqt'fe1og) ea,tltsod eJB ezls uorlelndod runrrqrlrnbe patradxa aqt pue ,{ltsuap ,trol le oler qltor8 pelcedxe eql JI ue^e 'uollcullxe Jlls€qJols Jo lsg qirq e e,req suorlelndod 11eurs 11 elq€I ro (SSO1) tuBIIInd Jo esuas eql uI ..s)ilns.. -iluessecau lou ore ,{eqt tnq 'uorlcurlxe Jo >lslJ q8rq e eireq .{lpctd,t seqeled fieuls ul suollsl -ndo4 'seqcled 1e11qeq e8rel pue Ipus Jo ernlxllu ? JoJ sI pesn uego st tdaruoc {urs-ecrnos eql qcrq^\ ur esuos Surpeelsrur ,(fieuuelod pue prrl{l V '(E661 -puelle -qlns pue uosul>lle71[ iuotlerSturut Jo ecuesqe etll ur enp^ urnFqqrnbe a.rtltsod tnq ,^^o[ B ol eullcep ,(eur uorlelndod lurs 3 6uolllugep s.uBIIInd fq) uorterfnuun Jo eruesqe aql ur lcurlxe oB ppom leql suoneFdod ere s>1urs 'eseo renel eq1 uI 'sesofund lecrSoloce rog eyqeregerd eq uaryo ,(eur 'ecuepuedep flrsuep cgtcedsertur Jo eJuesqe eql ur ',trsuep
^\ol 13 eler ql,uor8 uotlqndod pelcedxe eql uo poseq sl qJlqa 'I elqeJ ul ueurrS uorlrugep er{J '(eurn1o,r sql'lepnoc pue sellc letunlo,r sql '>lcoprql11 pue uoueg) uotlBldepe pue ernlcruls ctleue8 ro; secuanbesuoc lueuodur eleq ,(uru qcq,la. 6A\oU eue8 ut ,{rleulu,(se Surzrsuqdrue ur 'sasofund crleue8 uorlelndod JoJ InJosn sr uorlruuep sltlJ 'IunIJqIIrnbe le 'ps;eit ectl Jo 'uoll -er8ruur spoecxo uorler8rue Jeqleql\ Jo slseq eql uo s{uIS pue secJnos peugep (SSe t) ruslllnd'erruBrelll eql ur uolsnJuoc osnec ol senulluoJ ldecuoc {us-eJrnos srLL'peluerreA ere slueluuroc ,$eJ e 1nq
",(roleueldxe-31as
f1e3re1 st 1st1 stql '(1 elqeJ) suuelJo 1sr1 pepuudxe prru pesrler e r{lr.,'r 'eJeI{ eslcroxe slql leeder e16 ',{3o -louruuel uuoJrun eJor.u € Surloruord;o edoq eql uI suual uorlelndodeleur fe1 go las u peugep (tOOt) urdpg pue DIsueH '{ooq uolteFdodeleru snot,te:d eqt u1
'(arunlo^
srql) 1coprr111 pue uoueg pue (erunyorr slql) uofnog pue IroI^ITO fq peqgcs -ep su ldecuoc uorleyndodeleru eql Jo ule{uop lecrSolorq eqt e8rulua seJnlJruts uorlelndodeleu eseqlJo secuenbesuoc ,fueuor1n1o,te pue clloueD 'suorlelndoduletu
;o sergedord lecrSoloce uoqepdod eql lp pelceJlp ueeq e^eq s>lJetueJ JnO 'lou Jo uorleutxordde InJesn E sI uoIlBIn -dod pco; eqlJo scrrueufp eqt ot ecuereJJrp ou se>lulu uorler8tru leql uorldurnsse trclldur el{l JOIIleqa sr uorlsenb eqt ',(3o1orq uorlulndod Puolllperl ;o e,rrlcads -red eql uroJC '1(oI ro q8rq sI elBJ uoll€J8nu reqleq,tr lou'palserelut eq o1 ued -deq e,r qcrq,ry\ ur suorlsenb eql Sulleplcnlo uI InJesn sr qceordde uorlelndoduleur eql lou Jo JeLIleq,^A sI 'Je^eaoII 'uoueluc IBlueLUBpunJ eql '1n;d1eq eq ol ,{1e41un sr qcuordde uorlelndoduleu B 'pelltull lou ere seou€lslp uotlerSrul Jr pue 'ob1g <
fes 'q8rq ,{re,r sr eler uotler8tu gr ',(pee13 '(y661 'ueuqcu51) uorler8ltu ecuelslp -uoqs Jo elur q8rq u se sacuenbesuoc etues eql lnoqe ueuo sBtI uorler8ru ecuelslp -3uo1 go aler ,trol e 1eq1 1se33ns selpnls leclleroet{J 'uotler8tru Jo elscs prleds aql sr ereq enssr luegodul uv 'qc1ed pleu rleql e^BeI uotlereue8 red syenpr,rrpur
lJo|loqus pru00 puo Dlsu0H 0llll 0 l
The Metopopulofion Approorh
t l
IABTI I Melopopulofion Terminology'
Term Synonyms and definition
E
r
r
s c
Patch
Local population
Metapopulation
Metapopulation structure
Levins metapopulation
Mainland-island metapopulation
Source-sink metapopulation
),Jonequilibrium metapopulation
Tumover
\'Ietapopulation persistence time
Synonyms: Habitat patch, (habitat) island, (population) site, locality Definition: A continuous area of space with all necessary resources
for the persistence of a local population and separated by unsuit- able habitat from other patches (at any given time, a patch may be occupied or empty)
Synonyms: Population, subpopulation, deme
Definition: Set of individuals that live in the same habitat patch and therefore interact with each other; most naturally applied to "pop- ulations" living in such small patches that all individuals practi- cally share a common environment
Synonyms: Composite population, assemblage (of populations) [pop- ulation (when "local populations" are called "subpopulations")l Definition: Set of local populations within some larger area, where
typically migration from one local population to at least some other patches is possible (but see nonequilibrium metapopulation) Synonyms: Metapopulation type
Definition: Network of habitat patches which is occupied by a meta- population and which has a certain distribution of patch areas and interpatch migration rates
Synonyms: Classical metapopulation
Definition: Metapopulation structure assumed in the Levins model: a large network of similar small patches, with local dynamics oc- curring at a much faster time scale than metapopulation dynamics;
in a broader sense used for systems in which all local populations, even if they may differ in size, have a significant risk of extinction Synonyms: Boorman*Levitt metapopulation
Definition: System of habitat patches (islands) located within dis- persal distance from a very large habitat patch (mainland) where the local population never goes extinct (hence mainland-island metapopulations do not go extinct)
Definition: Metapopulation in which there are patches in which the population growth rate at low density and in the absence of im- migration is negative (sinks) and patches in which the growth rate at low density is positive (sources)
Definition: Metapopulation in $hich (long-term) extinction rate ex- ceeds colonization rate or Vice versal an exfeme case is where local populations are Iocated so far from each other that there is no migration betseen them and hence no possibility for recoloni- zalron
Synonyms: Colonization-extinction events (or dynamics) Definition: Extinction of local populations and establishment of new
local populations in empty habitat patches by migrants from ex- isting local populations
Synonyms: Expected life-time of a metapopulation
Definition: The length of time until all local populations in a meta- population ha\e gone extinct
( tontinues )
'B6g6D suIAe'I 13ql Ispotu eql 'tuelqord xelduor ,(lssaledoq e sreedde lsrg 13 leq^\ Jo uorpcgllduns I?JIlIrc e uo peseq uago ere slt{3ISuI luecgru8ts '(1ru1
saqreorddY p11dru1,{P;1ud5'1 ' (tVeel'HsueH) seqceordde cIlsIIeeJ,(lerpds pue'trcrldxe'{1pr1eds'lcqd -ur,(lprteds uee^Ueq qsrnflurlsrp an 'seqceordde esaql Suoury 'suortulndod pcol pue seqcled prlqeq ,{uuru Sulunsse seqcuordde Suqepou 01 uoser o1 {resseceu st 1r .re^ouJru uonelndod sr eJsqt ueqa 'osues ,rorJeu 3r$ uI selpnls uoqelndodeleu uI '(g66I ''p Q Snguell{D :S86I 'lloH:?L6I 'utne1) rnrro lou op suollrullxe ter{l pe$ln8er ,(ye,rrlce;;e os eJP suorlBlndod ieqt elunssB ol 3utllt'n st euo puu scftueufp I8col uo uorlerS[u3o lceJJe er{l uo,(lecgrceds st,{pn1s alpJo snroJ eql uer{/vr InJesn sr qceordde JaIruoJ oql 'suollslndod 1eco1 ,{ueur Suoure suollrerelul rlll^d Jo 'uotler8rur ,(q pelceuuoc suorlelndod cgrcedsuoc onl Suoure suorlceJalul
'€966I 'rlsu€H pue '166I 'utdpg puu DISUeH uro+ PeUlPohl"
Iepou uollJunJ eJuepIJuI eql pu€ sle -poru uorlelntuls epnlJul sleporu rr1s1lue: ,{lPueds :slJo,^alou qJl€d IeeJ qlr/$ lueruee:8e ur 'saqclud lellquq 01 selnqlrll€:aqlo ,{lqrssod pue'suolturol lutleds'see.m .relnlrged su8rsse luql lepol tr :uoqluge(I
(slapou cr1s11ue: ,(lprreds pue ltclldxe flprleds uea.{r1eq uoururlsrp e eIBIu e/t\ luql etou) lapour lrcqdxa flpnud5 :sur,(uou,(5
(elqeIJE^
snonurluoJ e sr qrlud E uI ezrs uotlelndod JI Jepou ecqlel deur peldnoo e pelpc st yepour eqr) poJeplsuoJ sI IIoJ € ut serceds eql Jo oJuesqe :o eruese:d ,(1uo puu 'pr.r8 re1n8a;3 uo sllet leJrluepl ,{1pord(1 e:e seqcled aq1 iseqcled lulrqeq lsorceu eql ol pelJusel uelgo'luepuadep-eJu€lslp st uot1er8rur qf,Iq^\ w Iepontr :uolllugaq
Iepo{u euols -Surddets 'lepour eteruolnu relnller 'lepoul (pu8) eJnle'I :stu,(uou,{5
slepou 1ter1dut,{lPtteds oJe slepou uotlulndoduteru pelnlJruls puu sloporu qJled :pelteu -uoc ,{1pnbe e.re suorlelndod I€Jol IIe qJrq,'v\ uI IepotrAI :uolllugaC
lePou Pu?lsl :stu,(uou,{5 saqcled eqt;o suolt€losl Puu sezls eql Jo suollcunJ eq 01 pounsse .{1preue3 'paldncco Sureq seqcted yo (secueprcur) seullqeqo:d freuorlels eql Jo Iapou V :uouluge(J
pelepou sI sezrs uortelndod leJol Jo uounqlJlsp eql qclql\ ur Iepou V :uoqluge(
(etunlo^
srql'DlsueH ees :e696I) sutnel fq peluese:d Iepour aql :uouruUe6l pelepotx sl seqclud lelrqeq perdncco;o (uouJe4 Jo) loqunu oql pu? perou8t st ozrs uoguyndod I€Jol qJlqil\ uI Iepou uorlelndodeleru V:uollruge(l
Iepour eJuesqe/ecuese:d'1epotu,{cuedncca :sur,{uou'{5
leporu uortepdod4au cllsrper ,(1pr1ed5
lepour uorlulndodetau lrcrldxo ,(1prred5
yeporu uonulndodetaur lrolydurr ,{lJuued5
Iepoul uorl)unJ a3ueptJul yepour uouelndodetaru
perntJruts
Iapour sur^e-l
Iopou qrlDd
uolllugep pue sur,{uou,{5 r|IJeI
(panuuuor) lfl[Ul
|J0pqurs laruo0 puo DlsuDH DPlll 7,1
I lhe Metopopulotion Approoch 13
1970) constructed to caricature metapopulation dynamics is an excellent example.
Instead of attempting to extend a model of a single population to many popula- tions connected by migration, Levins modeled the changes in the number of such populations, effectively ignoring what happens in each one of them and where in space they happen to be located (Hanski, this volume). For the latter reason, the Levins model and other related patch models (Table I) are spatially implicit; the habitat patches and local populations are discrete (and are generally assumed to have independent dynamics), but they are assumed to be all equally connected to each other. In spite of this simplifying assumption, which can be generally de- fended only for metapopulations close to steady state and with no strong spatial aggregation, the patch models allow us to analyze many interesting questions about metapopulation dynamics, starting with the conditions of metapopulation persistence in a balance between local extinctions and colonizations. The advan- tage of the spatially implicit approach is that it greatly facilitates the mathematical and conceptual analysis; the disadvantage is that it can be used to study only a subset of all interesting questions.
Thinking about the restricfive assumptions of the Levins model and other patch models, ecologists have asked what happens when local dynamics are in- cluded in the metapopulation model. What happens.when the habitat patches are of different sizes and when the local populations have different extinction prob- abilities? What if migration rate is high enough to "rescue" local populations before extinction? What are the consequences of real spatial locations of local populations? What if extinction events are correlated over the entire metapopu- lation? What if there is spatial asymmetry and source and sink populations? Some of these questions have been explored in the context of spatially implicit models (Pulliam, 1988; Hanison and Quinn, 1989; Hanski and Gyllenberg, 1993; Gyl- lenberg et al., this volume), but it comes as no surprise that at some point we have to tum to models that incorporate specific information on the spatial loca- tions of populations. Incidentally, most analyses of metapopulation genetics (Bar- ton and Whitlock, this volume; Hedrick and Gilpin, this volume) have been based on the Levins model, which is essentially equivalent to what population geneti- cists call the "island model." As in ecology, there is an increasing need to add space in a more explicit manner to metapopulation genetic models.
2. Spatially Explicit Approaches
Under the rubric of spatially explicit approaches are several related modeling frameworks, such as cellular automata models (Caswell and Etter, 1993), inter- acting particle systems (Durrett, 1989), and coupled map lattice models (Hassell et al., l99la). These modeling approaches assume that "local populations" are arranged as cells on a regular grid (lattice), with population sizes modeled as either discrete or continuous variables. The key feature that distinguishes spatially explicit approaches from spatially implicit approaches is localized interactions:
populations are assumed to interact only with populations in the nearby "cells."
Localized interactions can have profound dynamic consequences. such as very
'(966I'Ile^pury) Suorlplndod Suoue uoqer8tur S€ qcns 'sctueufp uortelndodeleur 8ur1ce;;e sessec -ord.(e1 olur lq8rsur elqunl?^ eprrrord sseleqleuou deur qceordde Sutlepou peseq ,(lpnprrrrpur uy 'ldecuoc uorlelndodeleu eqt Jo eSslue,rpe e4el ,t1per lou seop 'eJnlcruls uoqelndod ,{ue rog pesn oq uec qrlq/( 'qceordde slql lnq '(7661 'ssorg pue stle8uyaq) spnpr,rrpul Jo l[eep pue 'uollrnpordelsluarue^ou 'qulq eql elelnrurs o1 sr qcuordde e(uerxe eqJ 'elsp qcnur setunsse slspotu sseql Jo uollec -r1ddu 1n;Suruueur 'flSursrrd.rns 1o11 '(966I ',(eU pue a^lnD ue:39!5 '.y661 'e(e -{"5{V) eIqBIIB^e ^,(puerp ere ed,(l sII{1 Jo slepotu cueue8 IBra,\eS '(t661 'e,{e{e5 -\v :V661.sBIuoqI pue DISusH) uorler8tu ,(q pelceuuoc suorlelndod pcol ,{uutu Jo uorlslnuls ellsuelxo uo peseq eru seqceordde crlstper ,{lpueds a^tletue1lv
.uorlezruoloseJJo sorlrlrqeqord rreql l3eJJ€ r1crq,r'suorlecol prleds rgnads erreq seqcled oql pue lseqlpqeqord uollcullxe Iesol lceJJB ol peulnsse sI qJrqa'33J?
uI reJJIp ot pe^\olle ere seqcled oql :lepou suIAe-I cllslutlruelsp oql 01 lserluoc ur crlsBqcols sr Iepou aql ecueq pue 'seqcled lsllqBq Jo rsqunu elruu e sI eJeql :secueJeJJlp IecIlIJc 8ur,tro11o; eql qlIA lnq .lapotu suIAe'I eqt ol petelar ,{1pn1 -decuoc sI Iepou dI eqJ 'lepour uorlelndodeleu cllsll€er ,(ler1eds lsaldulrs eql sdeqred st (eun1o,t sgt'q'et66l 'I{sueH) Iepotu (CI) uollcunJ ecueptcu aq.L
'selpnls pleg pcurdrue qllA pe{uq,(1eso1c sr qceordde cllsqeer,(Ipueds oql'suoseer snol^qo Jod 'Iepou cIlsIIeeJ ,{1pr1eds e peeu e,r '1rolr1eu qcted e ur seqcled ruyncruud eruos fiut,{orlsep ;o secuenbasuoc ,{1e41 eI11 ssosse ol lue^\
e,Lr gr 'ecuelsur Jod 'Suollelndodeleru I€oJ Jo Scltueuf,p eqt lnoqe suollclperd alrlel -puenb cgrceds 3ur1eu uI palseJelul sI euo Jl{ressacau sI Iepou oql ur uollutu -JoJur srqt 1e 8urpn1cu1 apelecol ,(eql ere ,{llcexe eJoq^\ pus '.{eql ere e?re1 noq '{JolyUeu eql uI eJeI! ere seqcled ,{uuu rrroq :s>lro^ueu qclud repcrged;o ,ft1etuo -eB cgrcods aqt Iepou eqt ur opnlcu ttl euo ^\olls slspolu cIlsIIBeJ flenedg
qePoIAI Jllslpeg fPJledS 'g '(y661'ef,e4e5lV :gO6t ''P ta uwil -8rng) sts.{1eue ctueu,(p B JoJ Iepou uorlelndodele{u pesBq-qcl8d e ol Suruerr.er snl11 'lslxe ,(eur suorlelndod pcol qclq^\ uo seqcled olur 6,sllec,, elefie:3fie uec euo',{111tqe1rns lBlIqBq 3o uorldrrcsep peseq-Jels8r IUorC '(^Aoleq) slepou cBSIIe -ar ,{gerluds pue 1rc11dxe ,(leueds ueealeq uollcullslp eql Jnlq leqt slepou xeld -uroc dolo^ep o1 elqrssod sl lI snqJ 'ecuds Surluesarder 3o lBIruoJ etues eql eJeqs ,{8o1oce edecspuel ut suorlducsep SIC peseq-Jels€r puu slepotu peseq-erllle'I
'scrureu,(p eql Iepou ol ruerSord relnduroc e elu,{\ ol ,{see sr 1l pu?'IIe3 01 IIoJ r.uoJJ olues eql ars Jol^eqeq pcol ure,ro8 lBql selru pcllslrrelllBlrr eql'Surcpds luulsuoc prre seJe luelsuoc u seq pu8 eql uo ilac qcee aJUIS '1eql sl eSulue,rpe uy'uollElndluoc elq€Joplsuoc eJlnber slepotu qcns 'pepeeu sI socuesqB puu secuaserdJo Jolce^ eJllue uB iperdncco silecJo uollcBq eqt ,tq fldruts pequcs -ep eq louusc uortelndodeletlI eql Jo e131s eql 13ql sI e8?luB^peslp aqa '(eurnlo'tr sr.r+ ''lo la eeN ie1661 ''1o la gesseg) scrureu,(p cIloBIIc f11er1eds pue (7661 'sut8 -8IH pue s8urlseg) epts ,(peals € ol sellles uorlelndodeleu oql eJoJeq seutl Euol
Jl0pqu6 pru00 Puo Dlsu0H 0)|)lll tl
1 lhe Metopopulolion Approoch
t 5 B. Empiricol Approoches
In a standard ecological metapopulation study, a key initial task is to make a practical distinction between habitat and nonhabitat and to delimit the suitable habitat patches in the study area. Suitable habitat may be defined subjectively or with the help of statistical methods (Lawton and Woodroffe, 1991). An experi- mental approach may be used to test the accuracy of an existing habitat classifi- cation: experimental introductions to empty habitat should succeed (Harrison, 1989; Oates and Warren, 1990; Thomas, 1992: Massot et a1.,1994); introducrions to nonhabitat should fail. Metapopulation studies focused on assemblages of ex- tinction-prone local populations typically proceed to record the presence or ab- sence of the focal species in the habitat patches and then to analyze the effects of various environmental factors on patch occupancy (Verboom et al., l99lb;
Thomas and Hanison,1992; Hanski et al., l995a,b) and on the rates of extinction and colonization (Sjcigren,l99l; Eber and Brandl, 1994; Hanski et al., 1995b).
Other field studies have been concerned with more permanent local populations, but ones whose dynamics are significantly affected by migration (Stacey and Taper, this volume). Experimental studies have attempted to demonstrate the pre- dicted temporal stability of local populations in a metapopulation as opposed to that in isolated local populations (Murdoch et al.,1996; Harrison and Taylor, this volume).
Landscape ecology (Forman and Godron, 1986; Tumer, 1989; Wiens, this volume) and metapopulation ecology share a common focus on space and patchiness. The difference is primarily in the complex mosaic structure of real landscapes that is the object of landscape ecology (Wiens, this volume).
In contrast, metapopulation studies typically assume that the patches which are used by the focal species are of the same type, though this assumption is made primarily for the sake of keeping the models reasonably simple (see Holt, this volume, for metapopulation models with two patch types). Empirical re- search in landscape ecology has been reluctant to use the population dynamic theory that metapopulation ecology purports to provide, even if in a rudiment- ary form, and consequently the two fields have developed largely independ- ently. One trend that is beginning to change this situation is the use of GIS- based landscape descriptions in generic metapopulation simulation models (Akgakaya, 1994). Today, ecologists have access to huge data bases of digi- tized information about landscape structure, and the imminent arrival of low-cost global positioning systems will greatly facilitate further empirical research in this atea.
It should come as no surprise that the bulk of current empirical research that is conceptually related to the metapopulation notion is conducted in consen'ation biology. We therefore devote the rest of this chapter to a more thorough scrurin).
of the past and present links between metapopulation biology and consen'ation biology.
uouepdod leql JI ue^o 'uollelndodelelu u ueql raqlel uollelndod e18uts 3 qll^\
Suqeep Sr euo leql luenbery os sI luelue6o1ll el{l Jaqleq.$ sI uorlueluoc Jo euoq eql 'suorlulndodelau roJ 'lnJasn uletuer 01 slepour aql roJ salls usaaleq sI ereql lueu -e^ou qcnu 1(oq lsnf sI sseurleJllqre Jo lue{uele IrB 'slapou;o sadfl qloq uI
perdncco sells Jo Jeqlunu eql se qcns scllsllels Sursn 'setceds o,trl ro euo Jo uorlepdod -uleru errlue eql uo sesncoJ '.{ltrcqdxe pelepou lou eru suortelndod Iecol pue seus lueJeJJrp eql qclq^\ uI euo ue^e 'lepotu uollBlndodelaul v 'pu?ISI leql uo sel8r uorlcurlxe pue uotlur8ruru4 pue satceds Jo Jequnu eql lnoqe suotlsenb uo pue puulsr e18urs e uo pasncoJ IIIIs sI 'secJnos IeJeAes uor; uotlerirru qll^\ euo uele pue 'ed,{1 slr{l Jo Iepou pessq-serceds E uele 'JOAe^\oH '(eun1o,l sqt 'qsueg) ldecuoc uorppdodelelu s.suIAe'I Jo uoIsJeA puelsl-pueluleru u lsnl sI sJnluu ur Suruedduq sr ter{^\ go uorldecuoc Surfpepun eql 'slepotu reuel aseql q '(I86I 'puorrr€Iq pue urdpg :tgOt 'eget'S;opeqrurg) seler cgrceds-satcads aatlradser eql Jo sluns fldrurs Sureq seler uollcullxe pue uotle.rSrurut eprm-&runuuroJ eIIl qlr.tr 'serceds Ienpl^Ipur JoJ slepotu Jo sollsodtuoc ,(leruro; ere sleporu crqder8 -oeSorq prrBISr etuos 'JeAe^\oH 'sseuqclJ serceds sI cllsllels fe1 sl1 'sarceds pn -plllpq lou'selllunturuoc s1ee4 fuoeql crqderSoeSorq pueISI - ecueJegp tuered -du ue oslu sI eJeqJ 'suotlelndod Iecol olqelsun ,{1errt1eler Suorue sPnpI^IpuI Jo lueuelorrr qlr/t\ 'serlrlue eleJJSIp olul eJnluu Jo uoISI^Ip eql-slepotu uorlelndod -uleur qll.ir^ s8uruurdrepun ,(e1 sereqs froeql crqderSoeSorq puBIsI eql ,(1ree13
',(1rcr1suqco1s crleue8 pue crqderSouep ol 3ur,to ,(preruFd 'puulsr aqr uo suorlcurtxe ptol Suro8uo ,(q pecuuleq 'setceds 1e;o uo4ur8rtutur Sutnuuuoc ,(q peurelureu urnuqqrnbe crtueu,(p B se puelsl I{cea uo sseuqJlJ sercads slrsod froaqt eq1 '(B8t6l 'tL6l 'JJopequrS) spuelst leuqeq ro leor ,(1aueu'sllun lletus olu pephlp sI 1I esnBceq poolsJepun ,(ppeer sseleqleuou sI l€ql aJru?u cru'reu,(p e Surlcrdep pue (sseuqcrr serceds) cps11els peulelqo ,(1rsee ue uo sncoJ ol srlletue -qturu eldrurs 3ursn,(q (t '8tS) slsrSoyoca uroq uollueilB l{cntu pel3erle ,(plcrnb USet '$et'uosp11 pue rnquvc€ytr) ,(qdurSoeSorq puulsr ;o ,ftoeqt eqa
{qdorOoe0o1g puolsl l0 tuoeql eq l0 ll0l puD osr1 aql 'V
'spsodordpesncoJsllu etuos ol pel osIB seq lI'slq8tiur luugodurr peunds seq ,(3o1o -rq uorlu^Jesuoc olul sleporu uotlelndodelelu Jo uollerodrocur eql q8noqtly 'lg elqeuoJr.uoc erolu E Surprlord sr Suquql uouelndodeleur ,(q .{8o1otq uoue^Jes -uoc Jo lueuodruoc crqdurSoeSorq puEISI eql Jo lueuecelder e '.{11ueun3 '1erre1 uorlepdod eql le pe{ulp Ja11el eql elrqlt 'serlrun1llluoc Jo sseuqcu satcads qlrzrr ll€ep relruoJ eql su 'lle1Y\ reqleSol lg lou plp ..f3o1otq uollB^resuoc ^\au,, sII$ Jo se^pq o^u er{J '(886I 'gopeqrurg) Surpeerqur pue UIrp ol Sullrro uollerouelep crleueS eqt puu ,{qderSoeSorq puelsl Jo ,{roeql clueu,(p eqt ,{q pepm8 'u8tsep e8n;er uo sncoJ B ol serceds pnpl^Ipul Jo sdrqsuorluler $llquq uo srseqdrue .{,\ueq e trrog 'gL6I 'ec Suruufaq ', 11ect1uuterp pe8ueqc ,Golotq uolle^JosuoJ
Ieot()t[ N0llv u]sN0l oNv sN0[flnd()dvl]ltl| ',\l
|+opqurs pruo0 puD !)lsu0H o)plll 9 l
1 The Metopopulotion Approoch
l 7
may be so large that individuals are likely to interact with their neighbors only (Harrison, 1991). For island biogeographic models, the argument is whether groups of conspecific individuals on sets of islands are separate populations or just transient parts of one widely ranging population (Smith, 1975; Simberloff.
1976). The latter argument has recrudesced recently with important conservation implications. Pimm et al. (1988), extrapolating from records of breeding birds on small British islands, suggested guidelines for how many individuals should con- stitute propagules for reintroduction efforts based on the body size ofthe species.
Aside from problems with the statistics of extrapolation (Tracy and George, 1992),Haila and Hanski (1993) saw a more fundamental flaw: the assumption that the birds on each island constituted a population, and their disappearance an extinction. In their view, all these birds are parts of wide-ranging, large popula- tions and their disappearances from specific islands within the range are not pop- ulation phenomena. Diamond and Pimm (1993) retorted that the birds of each island could be viewed as a population within a metapopulation.
Within about a decade (Fig. 1), the theory of island biogeography came to dominate much of conservation biology, with a series of nearly simultaneous papers (Terborgh, 1914, 1975; Diamond, 1975a: Wilson and Willis, 1915) all advocating a set of "rules" of refuge design ostensibly based on the theory (Fig.
2). The rules each suggest a refuge configuration that would maximize species richness, and the papers describing the rules apparently stemmed from lectures given by E. O. Willis beginning in 1971 (Willis, 1984). Although some of the rules in fact were not based on the theory (references in Simberloff, 1988); they became popular in conservation circles, particularly after their publication in 1980 in the first synthetic plan for dealing with a perceived disastrous wave of extinc tions, World Conservation Strategy, jointly authored by the International Union for the Conservation of Nature and Natural Resources, the United Nations, and the World Wildlife Fund. With this imprimatur, it is unsurprising that these rules, and the theory that supposedly suppofied them, became the governing paradigm in conservation biology, reproduced in textbooks and published in newspapers.
The dominance of the island biogeographic paradigm was so strong that even studies that today would be seen as metapopulation research were published as island biogeographic studies, with no mention of the term "metapopulation" (e.g..
Fritz, 19'19).
It was noted early that most ecological publications citing island bio-seo- graphic theory simply interpreted a species-area relationship in terms of the the- ory, when altemative explanations were also possible (Simberloff, 1974). and that there was little empirical evidence for continuing local extinctions of the son envisioned by the theory (Lynch and Johnson, 1974; Simberloff, 1974). Funher.
as noted by Smith (1975) and Simberloff (1976), by defining the comings to and goings from local sites of individuals within continuous populations as "immi- gration" and "extinction," one could almost always claim that e\tinctions and colonizations were occurring, even if the theory really envisioned most recruit- ment to local populations as being by itt situ reproduction rather than immigration.
ur lsorelur uorlu^resuoJ peuessel .(urerer{l seoueraJer puu ,996I Uolraqults prrE elnog) ,{roeqt eql ot p3leler lou ere (qZ .BtS llluurs lere.res ro a8rey el8urs .g5g15 r euo peleqep flepr,r lsoru aql Surpnlcur .selru er{l Jo oruos ler{l uorlru8ocel aqf '(Z '8tg) u8rsap e8ngar;o selru eql (E) pue ,(eVOAl JJolrequrs :9g6I ,JJolreqr.urs puu glnos) ,(lqr8er; srql Jo sesnec pue seSn;ar pnpr^rpur Jo elorq eqt yo ,{lrg8urg oql ut lsorelur (7)'drqseceds ro pueysr ue se e8n;er e;o roqdelaru eql (t.l erea ,(qderSoeSorq puelsr Jo froeql cruuu,{p eqt yo serce8el uorle^lesuoc fe1 eq1
'(7961 'ueur. re1 pue epeg) drqsuorlular ueru-serceds ar{l se qJns 'surelled Jo Jeqr.unu e .r\er^ 01 qcrq,r uory elrlcedsred Ieclleroeql u pepr,rord ,{roeqt crqderSoeSorq puelsr .etnluu Jo r{f,nr.u roJ Iepou s sB uees re8uol ou sr lr q8noql .froeqt eql ur qtluJ Bururlcep eql slJeger (t .Stg) ..,,{qderSoeSorq puelsr,, Jo suorlslrc ur eurlcep eql .(qtg6I .repdg pue reueoqcg) .,lueuerured Sureg 're8rey qcnur fllsoru 'sJeqlo fueu qlr,l 'suorleyndod e.r.rilBng Jo losqns € se^Io,\ur Je^orunl,, 'sruelsr(s lsou uI .l3rl1 sr ,r\ou A\erA Surpe,rerd eq1 '(6961 'uosruelllll6 :qlg6I 're11rd5 pue reueoqcs :0g6I ,ueqllg) e,trse,ued euec -eq uorlcurlxe pcol puerdsepl,!\ Jo ecuelsrxe eql lnoq€ lqnop ueq,tr .0g6I puno&
plun ,,(yqero,tey .{roeqt eql ,t\er^ ol pepuel aloq,^A eql uo slsrSoloce .sseleque^eN
'tq8u eqt uo e^rtsualp eqt ot rorradns se ueas sr ga1 aql uo u8rsap aql'elnJ qrBe JoC'(0g6I 'seilnoseu IBrnteN pw arnl€NJo uorlE^resuoJ eqt roJ uorun puortetxalul :916I 'sllll16 puu uosli6 rege) u8rsep e8n;ar:o3 sa1ru,.crqde:Soe8orq pueysr,, eq1
Z ilnCll
o o o o o o
o o o o
-
o o o o o o o
p
c o
o o
Jj0[oqurs l0ru00 puo Dlsu0H D)plll
oo q
oo
8 r
I lhe Metopopulotion Approoch
t 9
the theory, while documented exceptions to some of the rules, including SLOSS, led to their fall from status of conventional wisdom. For example, the third edition of one of the most widely used introductory ecology textbooks, Ecology (Krebs, 1985, p. 559), reprinted the figure of the rules as popularized by the IUCN and described them as flowing from island biogeographic theory. The fourth edition (Krebs, 1994) omits the figure, makes no mention of the rules, and cites the criticism of the theory as "true but trivial" by Williamson (1989).
B. Porodigm Shift
The waning of the theory of island biogeography as a dominant conservation paradigm in the late 1980s coincided with the burgeoning interest among biolo- -eists in the metapopulation concept (Fig. 1). As does Hanski (1989), Merriam (1991, p. 134) explicitly claims a paradigm shift: "Metapopulation models have largely replaced equilibrium island biogeography as a way of thinking about terrestrial habitat islands, fragmented habitats and heterogeneous terrestrial en- vironments in general For other conservation biologists, the shift is tacit and consists simply of an assertion that nature is structured as metapopulations followed by discussion of what actions are required to preserve metapopulations (e.g., Noss, 1993). Perhaps most telling isThe Diversit;^ of Lfe, by Wilson (1992), a founder of the theory of island biogeography, in which species are typically seen as structured as metapopulations and the consequences of this structure for conservation are explored.
The causes of the shift are many. One must be the growing ecological lit- erature, described above, doubting the verisimilitude of island biogeographic the- ory. However, scientific data and the weakness of a prevailing paradigm alone are unlikely to precipitate a paradigm shift (Kuhn, 19701' Ha1la, 1988), and we must seek other prevailing currents. It is worth recalling that, fundamentally, the theory of island biogeography can be construed as just a multispecies version of an analogous metapopulation theory, so it is hard to imagine objective scientific reasons for accepting one while rejecting the other.
One possible explanation is a shift among conservation biologists and ecol- ogists from the conception of nature as an equilibrium world to that of a non- equilibrium one (Wiens, 1977,19841' Chesson and Case, 1986). Island biogeo- graphic theory is dynamic, of course, but the emphasis is on equilibrium species richness, hence the nickname, "equilibrium theory," and even the underlving im- migration and extinction rates are seen as constant. Though metapopulation the- ories are not any more, or less, "equilibrium" theories than the theory of island biogeography, the emphasis in the latter on equilibrium species richness and in the former on population turnover may have created the sense of a conflict be- tween an equilibrium and a nonequilibrium theory. The key point. of course, is that in both theories there is no equilibrium at the population level. However, the modus operandi of the island biogeographic theory is to ignore the changes in the presences and absences of individual species and to focus on the equilibrium
're,lerrro11 '(066I 'rellei(\ ''3'e) ,(1ere1q oot ,{,roaqr aqt Suqel ]sure8e suolln€c se pecnppe ere^\ uluualllur roJ suollelndod lptus .{:e,t se polsrsrad puq leql suollel -ndod euog '(986I JJolrequls pu€ uellceog) paululsns eq tou ppoc esdelloc IUUneJ lUeUIIUIuI JO UoIUaSS€ UB leql SIIuII eCuepguoJ apl.t\ qJns e^sq ol u,toqs su,Lr drqsuorleleJ BeJe-serceds eq1 's,(ezrr lerelas uI paIIElsaJoJ S€1rr SeAJeSeJ Il€rus Jo ecueueluleur aqt ot froel{l crqderSoeSorq pUEISI tlror-I leerql sIqI
'(7961'uolteclunuruor leuosrad'leuo'd) serJads esol ,{lqelr,reur p,t ,{eql leql s,^Aol{s 'selnr u8lsap afn;ar aqt ur pegtpoc ',{ro -eqt crqderSoeSorq puelsr esnereq lnq ulaql ulqll^\ suollelndod Sururlrep pc,^^oqs r{oreeser cgrceds esneJeq tou 'se8nJor ll€tus oluos uopu?qe ol s0861 eql Suunp oJnSSeJd learfl repun su,Lr, ,(1rroq1ny oql pue 'X1uoq1ny uoIlB^JeSuoJ alnlu\i aql ,{q seer8ep snoIJeA ol pe3€uetu pue pelcalord eJe esoq; 'llstus ,fue^ eft qrlqa Jo ,{ueu 'se,lreseJ 002 e1uos Jo slslsuoc IeBJsI Jo uolleu llutus eql Jo tllals.is eEnJaJ eql .Oldr.uexo JoC 'uorle^Jesuoo lsurBS€ suorlelndod II€uS elgel^ut uo stseqdua srql rrJnl plnoc slueuoddo elnlse pq1 pezruSoce-r .{11en1uele $sluolle^resuoJ
'esde11oc ol lcelqns eq plno,t 'salels perlun eqt ur sryud IBuoIl€u 1se3re1 oI{1 Jo ezIS aql 'se8n;el snouuoue uI suollel -ndod ue,,r,e ter{1 pelseSSns Sullepotu pcllstueqletu 'suollelndod lpus qstn8utlxe serqdorlseluc puu ,(lrorlseqcols luluouuoJr^ue teql luelxe eql ot 'peepul '(6[6I
''p p g1nos,:ZL6I'pvovrqq ''3'e) ol€redn€dep rueql repueJ ,(1qe1r,leut IILtr seuo IIBrus ur seleJ uorlsurlxe pels^ele eq1 pu€ popeeu ere seSnJer a3re1 leqt enSrE 01 pesn ,(1eprm ere,tr ',{roeql eql pue '.selru oql 's:euueld uoll€AJesuoo JoJ elepuelu
€ se drqsuorleler srqt sesserdxe (eZ '3W) u8lsop e8n;er ;o elru lsrg eqJ 'seuo Iprus u€ql serceds erolu eAuI{ ol puot selrs e8rel 'pnbe Sureq s8urq1 reqlo IIe 'leql Sur,troqs 'sdrqsuorleler eere-serceds ,(ldturs ere,t ,fuoeql crqderSoeSorq puelsl Jo suust ur pelerdralur elep pcrSoloce uletu eqa ''fuoeqt crqderSoeSorq puelsl ,{q uorlenlea.ep Jleql tuoq sells lletus penJser slepotu uorlelndodeleu 'flpurg
'luelolor ssel ueos froeql crqderSoeSorq pu€lsl e{eu ol punoq sr serlrunuuoJ uo ueqt reqleJ suollelndod uo sncoJ € 'uru8y '(3361 lopeq -rurg) eldruexe erutrd eql sI sezls uorlelndod elqel^ runtulutru Sutlles ul {tlclrseqc -ols lelueuuorr^ue pu€ crqderSouep Jo eloJ eql - serceds u€ql rel{leJ suorlelndod Jo suuol ur uees eq 01 osle ue8eq uorle,uesuot ;o slcedse lecrSolocg 'suolleln -dod puu serceds.ol selllunlutuoc uor; slsrSololq uoll€^resuoc Jo snroJ eql UIqs 01 pepuot qJr€eser sIqJ '(886I 'gopaqurr5) suorlulndod lletus uI Sutpeerqut pue UIrpJo fpnls eqt ,{pulncrlred 'sctleua8 uorlelndod sr .,,(8o1otq uollelresuoc
^\eu,, eqt go lueuodtuoc uIBuI reqto eql ',{roeql crqderSoeSorq puelsl ol uolllppe uI
'rellel eql uI lou lnq reuuoJ aql uI uollelndod puelureu tueuetu;ed e 3o ecueserd eqt st (e6961) sut.r.e1 pue (ISAD uoslll& pue rnI{uYcBW Jo sleporu eql uee^\leq ecuereJJlp I€cIlIrc e 'peepul 'slserelul eseql qll^\ ile,$ slg ,ftoeql uorlelndodeleu eql 'seuo leuotsseccns ,(pee pue serceds e,r4t8ng s€ qcns 'lereureqde fleco1 aq ol peullsep ere 1€ql serceds 3o ecusuetur?tu puu'slcJo esn eql puu €l€p lerluds ur asuerJul ue ,{q perelsoy 1ue1xa e8rel u ol'slepotu lrcrldxe ,(gerpds eJB uollalJesuoo uI slseJelut Sut,a,ror8 o,ln,l la1 'fruouoxul rno ut ltctldrur ,(1pr1eds sr ,{roeql slql :sessouqcu serceds 3o ureled
l]0|laq{xrs loruD0 puO DlsuoH 0llll
01,
I The Metopopulotion Approoch 2l
the main salvation of small sites was the shift by conservationists to the meta- population paradigm. In the Levins model, at least, small sites containing small populations were the only homes of a species and thus the proper locus of con- servation concem. The model even suggests that a certain number of unoccupied sites is required for metapopulation persistence (Lande, 1988a; Hanski, this vol- ume), thus relieving beleaguered conservation biologists from having to justify a refuge for a given species by confirmed residence. A famous example in which local extinction rates are high and a supply of suitable empty sites is necessary is Pedicularis furbishae, the Furbish lousewort (Menges, 1990).
In sum, from a conservation standpoint, it is not surprising that citations of metapopulation studies increase exactly when those of island biogeography de- cline (Fig. l). These trends represent a paradigm shift.
C Use ond Misuse of fie Melopopulolion (oncepl in (onservotion Biology
Hanski and Gilpin (1991) observed that "metapopulation ideas have recently become the vogue in consenvation biology." and numerous conservation strategies are explicitly based on metapopulation models (references in Harrison, 1994b).
The general effect has been to draw attention to landscapes and networks, as opposed to individual reserves in isolation, for which the island metaphor of island biogeographic theory is appropriate. This is a salutary development. Even if there were no significant interactions among populations in different refuges, it would be good to have multiple refuges simply as insurance against local catastrophes (Sou16 and Simberlofl 1986). Doak and Mills (1994) and Harrison (1994b) argue that, even if most species are not structured as Levins-type metapopulations in nature, the rise of the metapopulation paradigm has served and continues to serve a useful function by forcing conservation biologists to gather data that are im- portant to effective conservation strategies of individual species-movement rates from site to site, relative reproduction and mortality rates at different sites, and the like. This is precisely the view of Haila and Jiirvinen (1982) for island biogeographic theory.
The problems arise, for metapopulation models as for island biogeographic theory, when it is assumed without empirical evidence that all species, or all species in a large class, conform to some particular model (Doak and Mills, 199'1;
Harrison, 1994b).lf a conservation strategy is based on metapopulation dynamics that do not exist, it can misfire. Thus, for example, Murphy et al. (1990) suggested that small-bodied, short-lived species with high reproductive rates and high hab- itat specificity typically constitute Levins metapopulations, but there are simplr insufficient data to evaluate this claim (Harrison, 1991,1994b). To t-cn--us auto- matically on metapopulation dynamics for such species would not constitute ef- fective science. No wide-ranging generalizations are yet possible. because teu' data really demonstrate the existence of classical metapopulatirrns. Harrison (1991) could cite only pool frogs (Rana lessoniae) in Sueden snd naterflies (Daphnial in rock pools as unequivocal cases (for some ne\\ eramples. see Har-
-e8olq puslsr eql uuql InJdleq eJoru eq ol suorlrrpard Jo suos JellBI eql lcedxe ol suoseer o,rl eru ereqa '(q9661 'l{s{r"H 'g '3lg) slepoiu rrlsll€er flprluds 1nq e1d -rurs ,,(1e,rr1e1e,r uo pessq seducspuel pelueur8er; relncrued ur serceds relncrged go ecueuuoyed elrtelal or{l lnoqe suorlcrperd sr ,ftoeqt uortelndodeleu tuoq uorlnq -rrtuoc snoSopue eql'Q,'3rg) e.r,oqe pessnrsrp u8rsap a8n;a: ;o salnr eql perrds -ur ,,{qderSoeSorq puelsr 3o f:oaqt eLm'(qZ'3rg 'e1ru SSO'IS eql ot uoltcouuoo eql elou :966I ''tD la 3req1qe16 :cg66l ''tD la r{sueH :q'Bt66l 'nlsueg) seqcled pJeles Jo seer€ eql Surcnper ueqt uorlelndodeleur u ot 1eluatuutep eJoru sr qcled leltqeq e8rul euo Jo lelorueJ eJrlue eql Jeqlaql\ 'ecuelsur roJ '{se ,ieru au6 'serceds IecoJ e Jo acuelsrsred;o suuel ur e8ueqc edecspuel ;o soFeuaJs e^uetuelp {ueJ ol slepotu uorlelndodeleur esn ol sr qceorddu elrlcruJsuoJ oJo{u V .'sloqtunu cr8eu,, snorceds olur elereue8ep,(eu sldecuoc eseqlJo esn 'ecrtcerd q '(q9661 "lD ta I11s -ueg) seqcled lellqeq elqelrns Jo fillqelpr'e eqt puu suorlepdod go reqr.unu elqer^
unrururu er{l Woq sollolur WAW leql erueraJJrp IBcrluJ aqt qlr,$ tnq'(lg6l 'JeJJeqS) ezrs (4n6) uorlelndod elqer^ tuntururru eqt ot snoSopuu sr ldacuoc slqJ '(q966I ''p p r>gueH) suorldrunssu umuer repun ozrs (Wnytl uorlelndod -elau olqur^ runlururru eql ornpep ol pesn ueeq e^eq sleporu uopeyndodera4
'(886I JJolrequrs) selrs rreql tsur^res -erd pue suorlulndod ,(qtpeq .{lluenddu 3ur1eco1 ;o ,(ldurrs slsrsuoc tueurqsllqelse e8ryer;o ,fto1srq er1l Jo qcnry 'tetlqeq perdnccoun flluerrnc 3o ecueuodur aql;o uotltuSocer eql ol pel e^Bq 1ou plno,4A ornleu Jo 1(er^ uorlelndodeleuuou 'rrtels V 'luleJ oq pFol( ,{1uo selrs perdncco ol seJnseoru uorlelJesuo3 Jo uortJulseJ '(erun1o,r srqt ''lD /a eeN :ounlo^ srql 'DISusH legg6l 'apue1) leraua8 ur oueu -acs surlo'I eql ur se serceds srql ur .Joqilnd 'uorlenlrs eql eleqJecexe ,(1a>11 11,n erur8er (,,rrog) ecueqrnlsp eqt ;o Surredruel luql pu€ urnuqryrnbe le ueql roqter eurlcep ur sr serceds eq1 lsql slseSSns uorlezruoloceJ pu€ uorlJurlxe IBJoI uo suorl -elresqo Surpnlcur (9661 'sa8uetr1) srsfleuu uorleyndodeletu V'pezruolor oq 01 pelelosr ool tou eJe leql selrs elqelrns .{lrreroduel 3o flddns e salnber uoqelndod -Blelu eql os 'Surdtunls >luuq pue rnoJs ecr ,(q pe.,(ortsep fletururtln sr uorlelndod 1uco1 ,,{uu 'rerr.e,tro11 '(9661 'se8uery) peq-rnlsrp ,(Frreeq ool rou el]lrl oot Jeqlreu sI l€qr lullqell eplsJe^rr e serrnber 'elduexe roy 'uomesnol qsrqrnC pere8uepue eq1 'scruuufp uoqelndodelou Jo uorldecuoc sur^e-I aql tncrepun .,(lplueuupun;
lou seop lc€J srql lnq 'crlseqJols uer11 JeqlBJ orlsruruuetap sr suorlelndod 1eco1 go uollcullxe eql 'serceds qcns ,(ueru u1 '(eunlorr srr11 'DISueH pr.re setuoql osle ees) uorsseccns Jo esnuceq e8ueqc ,{lqelrneur leql stelrqeq ,(dncco lepotu sur^el oql o1 Suruuo;uoc .(13urcur.tuoc lsoru serceds oql 1eql slseSSns (qr66D uosursH
'(qsoor
''p ;zrXSueH:t661 ''lD /, sutuoql'C'J) uorlulosr pu€ azrs qc1ed,(q pelcege eJB eseql ,4AoI{ pu€ seler uorler8rtu pue uorlcurlxe Iucol uo pelnber sr qcJ€esog 'uorlelresuoc e^rlceJJe 01 luroruc sr scrureu,(p uo4elndodeletu;o Surpuelsropun ue 'seJuetsur qcns uI '(eurn1o,r sql 'DISueH pu" s€uoqJ :S66I 'ueBSSnn) pue r{sueH :t66I 'ssruorlJ pue rlsueg) serceds fgrallnq reqlo fueur lueserde; ,(eru puu (q'e9661 ''lo p r11sueH) elduexe poo8 reqloue sr pueprC ur (otxurc oaolt -p741)
'lgte11nq,ftBllplr; ellr^UBIC pereSuepue eq; '(eurn1o,r srql'ro1,(e1 pue uosrr
l+o|lE[]rs l0ru00 puD rlsu0H 0)plll tz
a
I
I
I
r ' l
I lhe Metopopulotion Approoch
23
c
r lr'l
FIGURE 3 Fou. examples ofthe same landscape fragmented in different ways (scenarios a to d).
Spatially realistic metapopulation models, such as the incidence function model (Hanski, 1994a), can be used to rank the altemative scenarios in terms of the persistence time of the focal species. Note the conceptual link to the SLOSS rule (Fig. 2b).
ographic rules of refuge design. First. the rules of refuge design ( Fig. 2) are static, even those actually flowing from the theory. For example, the fundamental con- cept in rule a (Fig. 2a) is the species-area relationship, which in applications is seen as meaning a fixed number of species in a fixed area. In contrast, the meta- population predictions explicitly address the dynamics of species survival. Sec- ond, the rules ofrefuge design contrast fixed general alternatives (such as in Fig.
2b), whereas the spatially realistic metapopulation models practically force one to compare specific fragmented landscapes (Fig. 3).
We now turn to potential misuses of the metapopulation concept in conser- vation. To start with, if a species is structured as a mainland-island (Levitt- Boorman) metapopulation, population turnover in the peripheral "island" popu- lations may be irrelevant to the persistence of the metapopulation as a whole.
though the dynamics are crucial to the persistence of the peripheral populations (Doak and Mills, 1994; Harrison, 1994b, Simberloff, 1994b). More generalll.
emphasis on metapopulation models can potentially harm conservation b1'dra*'- ing attention away from single populations on the grounds that no one of these is crucial to a species' persistence and it is the ensemble that matters (Harrison.
r994b).
Another example of the metapopulation concept used in misguided attempts
r t I
I
d
r o
r l
posuq tunturunu JgrJads e JoJ e8uustu ol ldueue .ieru auoauros 'sr Iepou tr{.\I\
aql se IeJeueB pur a.rrletuet se '.snqJ ';o pu ta3 uec .{eql eJnleu Jo qJntu .\\otl ees 01 1r esn uorlE,\Jesuoc o1 pesoddo secro; 'elqBuB,\ ,.iue ro3 les sI tunlulurtu e su uoos su :(9661 'eruorJ :8861 'JJolrequrg) eururru pasnsru;o seldurexe ,{ueru ,{q pelretu sr ,{8o1orq uorJullesuor;o ,(rolsrq eql '(t661 'slltll{ pue ryoq) suuld lueue8eueu cgrrads aprnS 1q3ru teqt slepotu crlsrleer f1ler1uds qtltt ere^as erou uele eruoJeq sualqord eqt pue'suorlcrperd e,rtleltluenb le tule leql slepou uolteln -dod 1p qlr^\ suelqord 'esrnoc Jo 'er€ eseql 'reqte8 ol llnrgJrp ets teql etup pue .{yrre'r o1 preq eJe 1uq1 suorldunsse ,(q peredrueq eJe uorlelndodetetu elqel^ e uIEl -ureu ol ,{ressacau suorlelndod }o Jeqrunu runlururtu eql Iepou ot stdruelly
'sederspuel lueserd ur lsrsred ol setcsds ;o ,{}r1rqe aql Jo arnlcrd fsor flSurpeelsrtu e sn e,rr3 plnotn qcrq,l.r '(eun1o,r. srqt'rlsueg) qcns se ureql ezruSo -cal lou lq8ru eA\ 1eq1 sr seducspuel pelueu8erg ,(lSurseercul uI suollelndodeteu urnuqqrnbeuou lnoqe .{rron relnrrlred y 'serceds e Sururelututu o1 ,{a>1 eq osle uuc uorlelndodeleu >1urs-eJJnos ur slurs pue secJnos uo €leql 'uorlulndoduleu eJrluo aql tceJJe plnoc sroprrroc puu suortelndod relncrlred Jo ssol ,toq pe1se33ns Iepou uortelnrurs e qtr,,tA peurquoc Jueruonolu uo elep ,,{rleruelelotpeJ slq pue 'sropuror uerredrr fq pe1ur1 ,{1esoo1 suorleyndod iletus Jo uol}cellor e sp slslxe .{llueunc serceds oql leql pe^\oqs eruroJrleJ Jo suletunotr tr euv uluus eqtlur ("tolo:) -uot ryag) sre8nocJo (8661) rereg,(q srs,{1eue uorlelndoduleu eql'elduruxe loC 'lBArAJns Jreql ol luIJnrJ eq ,{uu suorlJelolul reql 'seuo II€us leJeles sI Uel e^eq e,4 IIB Jr lnq 'reileq ueeq eleq tqSru uortelndod e8rel e18urs u '(qZ '8tg) ,{3o1ou -ruret SSOTS uI '(I66I 'uosrr-re11) edecspuel pe8ueqc oslmrel{lo ro peluauSe1y e ur seroeds er{l Jo ecueuelumtu eql ur ple u€J scttueu,{p luerlnJ s1t Surpuelsrep -un 'esec srgl ur ue^g 'GVeet'gogequrg) uoltenlls uorlulndodeleu runrrqrltnbe -uou s se 'eurlJep eql yo sgud eruos Surrnp ':eedde plno,,lr uollrurlxa lenperi ,{ue lsoruls 1eq1 ,(1e>11 sluees 1I 'uorlcurtxe o1 Sururlcep fl,trols 11u suorlelndod 3o e3elq -ruessB ue,(1ereu'peleeJc se,u uotlelndodeleur leuorlcunJ ou lnq'suollJullxe leJol luenbesuoc qll,$ 'sllun Jelletus olul paprnlp seuoceq uorlelndod snonuttuoJ eJotu ,{lsnor,terd s qcrqa ur 'suorlenlrs urnuqrpnbeuou 'luolsueJl lueserde: ,{eu sadecs -puel petuaru8er;-uerunq ur suorlulndodeleru lecrssulc ;o selduexa eruos
'(ZeOl ''to la JJolrequls) 1ro,tr1eu e ;o ged tou sI teql eSnger cgrceds e errnber leql suonelndod relncrgud lcelord ol suoJJe uoJ.] trertep UPJ pue enis -uedxe eq uec sropruoc ur lueu4selur'8ur1qnor1 eroru uelg '(2661 "tD J, JJolreq -urls :2661 'sqqoH) uortJurtxe llelsoroJ fypnlce pyno,r,r ,{eql teql ro luatualoru JoJ pesn eq plnod sJoprJJoJ oql leql eJuepr^e luecs sI eraql 'suels,{s JopIuoJ pesodord lsoru ro; 're,te,tro11 'aler uotlur8truut Surseercur ,{q ctlstlets slqt aseeJJul ol pounsse eJea sJoprJJoc pue 'sseuqcrr serceds 'cllsllels 1e,te1-flrunuruoc e uo se,/r snJoJ eql 'elnr srr.{l Jo uorl€lnuroJ leurSrro eqt u1 '$7 '3rg) sropuor lnoqll^\
les luJrluepr osr,r\Joqlo uu uptll sotcads erou ureluoJ lll/( sJopuror,(q petcou -uoc se8n;er Jo les € leql sr 'fuoeq1 crqderSoeSolq puelsl qtl^\ pel€Icosse u8tsep e8n;er;o elnJ euo '(sroprrroc 1ou 'seqcled tullq€q Suoue eJu€treuuoc ezrseqd -rue ol puol sloporu uortulndodeleu 'Suqeads fllculs 'let{l o^resqo ert q8noqt) sJopruoc luotueloru Surpunorrns ed,(q eqt st suotlelndod ey8urs ezrsuqdtueep o1
l]ofoqurs l0ru00 puo rlsu0H 0}lll nz
1 lhe Metopopulotion Approoch
25
on this model. As discussed above, a less controversial use of spatially realistic metapopulation models is simply to rank alternative management scenarios (Fig.
3) and to recognize that making long-term predictions about (meta)population persistence time in our rapidly changing world is practically hopeless.
The shift from island biogeographic to metapopulation thinking united ecol- ogists and geneticists in focusing on populations. Genetic concerns have been prominent in the interest in metapopulations. Kimura and Crow (1963) first pointed out that occasional migration between local populations can maintain genetic variation better than would a single large population, essentially because drift is likely to fix different alleles in different populations. However, the situ- ation becomes more complicated when we allow for local extinctions and recol- onizations, and recently much effort has gone into modeling the way that such population turnover affects the maintenance of genetic diversity in metapopula- tions (Wade and McCauley, I 988; Hastings and Harrison, 1994; Barton and Whit- lock, this volume). On theoretical grounds, one might expect species thatnaturally exist in metapopulations not to be prone to inbreeding depression because they lack genetic load (Harrison, 1994b), while local populations in a recently frag- mented large population might be particularly susceptible to inbreeding depres- sion because heterozygosity would quickly decline (Simberloff, 1988; Hedrick and Gilpin, this volume). Under the latter circumstances, maintaining movement among populations might seem particularly important, and indeed many man- agement plans for declining populations call for measures to enhance population interaction, such as translocation and corridors, specifically to avoid inbreeding depression (e.g., U.S. Department of Agriculture, 1995). However, field evidence for inbreeding depression or other problems in recently fragmented populations is scarce (Harrison. 1994). Lande (1988b) argues more generally that the impor- tance of genetic threats in conservation has been overblown. His view is that, in naturally small populations, the genes causing threatening inbreeding depression would have been selected out, while in recently reduced populations, ecological threats are more immediate. Despite this widely cited statement, genetic principles still underpin many viability analyses and management plans (Harrison, 1994b\.
Perhaps the very fact that genetic modeling is feasible ensures that it will be done.
particularly if ecological modeling, even if potentially more useful, is more prob- lematic. The latest round of papers (e.g., Lynch et al.,1995) appears to strengthen the genetic argument, but the most urgent need is for relevant field studies.
Thompson (1996) contends that metapopulations may be crucial to the con- servation of various coevolutionary interactions, such as those between pathogens or parasites and their hosts. In models, locally unstable population dynamics can be stabilized by the addition of metapopulation structure (Hassell et al..l99la1' Nee e/ al., this volume). In other cases, the metapopulation structure stabilizes evolutionary dynamics of the interaction. For example, under certain circum- stances, the coevolutionary dialog between a pathogen and its host can lead to the extinction of the host, if a new virulence gene in the pathogen spreads rapidly enough. A metapopulation structure can then prevent the gene from eliminating
'pauua\
-un flrelnsrued sr uorlerouelep crleueS 3ur1p1saro3 ur scrureu,(p uogelndodetaur Jo alor eqJ 'slstlqeq l€uorsseccns Jo cllslJelcerEq3 asoql .iq peleururop '{lqeqord .serceds;o e8uur pelrunl 3 ol lelctuc aq feur ocusuelurEul uortelndodelew '3^UJnp -oJdJelUnOC eq ueJ esuepr^e Suruoddns ou sr eJaql qJnl,.t\ JoJ surels{s 01 ,(lpu[q qo€ordde uollepdodeleu eql flddp ol uolleldt1;el eql'sselaqleuoN 'lueua^oul pIIB ,(qderSouep uo Btep ,{e1 8uqce11oc uo pesnJoJ serpnls plag parrdsur e,req,{eqt pue .uorle^resuoc o1 slr{8rsur lueuodur pelnqrluoc e^Bq slepou uollelndodeletr tr
'sJeD€IlI 1r
"{1UO uolldeJ -Jed ur sr ecuereJJrp eql qSnoqJ 'sedecspuel ,fteurpro Jno Jo uollslueurSerJ qltm pel€rcosse sl tdecuoc uollslndodgletu eql s?oJeq^\'sepcs lep€ds e8rel le sulel -1ed ureldxe o1 pedole,rep flleurSrro sen ,{qderSoeSorq puelsr;o ,fuoeql clueu,{p eql.eleJs lerleds eql sI enssl lu€lrodIul(llueredde euo 'erueroJJlp eql epetu seq lsrp lxeluoc repr,^A el11 ,(1e3re1 sI tL Aoq reldeqc slql ur pessncslp a^Bg e16 ',fuoeqt Jeqlo eq1 ol lcedser qll.t\ uollcunJ eurcs eql e^res ol f1ery1 sr ,{roeql euo 'lsute8e Jo .JoJ peJelsnu eq u€c ecuopr^o Je^elBqld l8r{1 percIeJ ,(1eso1c os aJ8 selJoeql elrlcedseJ aqt esneceq 3ur1u1s eroru eql sI eldtuexe srq; ','(8o1orq uorlelndod uI UHS ur8rpured e ;o eldruexe Surleurcse; e slueserder ..uoltBlndod€1etu,, pue ..,{qdurSoeSolq pu"ISL, sprolr .'(e1 el11 Jo suoll€llc Jo luau€d Sur8ueqc eq;
sN0lsnDNol 't\
'eJnlnJ elf uI su€ld luerueSeue11r cgrceds ur ocuenuur pue Surpun; r{sreeser qloq JoJ eledluoc ,{lerns p,tr ,(eqt pue ,{lelueuepunJ reJJIp seqceordde o1(1 eseql Jo sl€o8 '&euFd pue SuEJSoJd qcJ€eSeJ eql 'SseIeqUeAeN 'sessecoJd ,{reuorlnlo,teoc eql lnoqe Sluetuoc elupo1utuocce uuc (9661 $opeqrurS) selceds ueql Jeqler sessecord Suturelureu uo luetueSeuetu uelsfsoco ur srseqdue eql 'osIV .,{lsnoeuellnruls suollelndod -e1eur ,{ueur ur€}urE(u plno,t\ 13q1 seqcl8d FlrqBq Jo uollnqlrlslp e qll^\ edecspuel e eurSurur plnoc euo pue .sSurlles pcol ,(1q3rq ueql reqler'suorSer pue sedecs -puel qll^\ tuecuoc € erel{s slopou uollelndoduleru puu luerue8sueru tuels^soce qloq .esrnoc JO 'r{creeser l{cns uo pesuq luerue8euetu IIL4A os pue 'luelcgJeul pu€
erusuedxe.{lsseledoq eq IIII* selcods JoUB solceds uo qcluoseJ l€ql sI luetuo8eueu ruetsfsoce Jo uolte^Itou .(e1 e '1ce; uI '(966I .;;opequr5) luerueSeuelu setceds -elSurs 01 pue suollelndod uI lseJelul frerurrd B ol clleqlllue elrnb sI lue1lloSe -ueru uels,(soce w lseJelur ;o uorsoldxo lueceJ eql l€ql 3ur4u1s sI lI 'uoIlBAJes -uoc ur rrecuoc Jo sle^el 1u€urtuop eql Sururoceq selceds eql pue uorlelndod eql ol pel s€q iscrleue8 uo ]€ql qll^\ peulqtuoc 'suollelndodeleru uo sncoJ eql
'senssl eseql Jo uolssncslp q8noroql e stueserd (aunlon srql) 4uerg '(E661 'uosdruoqJ pus uoprng) uollcunJ srql se^Jes eJntcruls uorlelndodBleu lsoq eql qclq^\ ut elduexe leJnluu u eq ,(eu (tur1 otosduto1a741) su xeg pue (a1out7tow wnurT) x€U pIlA. 'setceds eJllue eql
Jl0|l0qu5 pruD0 puD DlsuDH rylll
