Ou comes o B ood Pa asi e–Hos In e ac ions Media ed
by Egg Ma ching: Common Cuckoos
Cuculus cano us
e sus
F ingilla
Finches
Johan Reine Vikan
1
*, F ode Fossøy
1
, Esa Huh a
2
, A ne Moksnes
1
, Ei in Røska
1
,Ba
˚ d Gunna S okke
1
1Depa men o Biology, No wegian Uni e si y o Science and Technology (NTNU), T ondheim, No way, 2Ro aniemi Resea ch S a ion, Finnish Fo es Resea ch Ins i u e,
Ro aniemi, Finland
Abs ac
Backg ound:
An agonis ic species o en in e ac ia ma ching o pheno ypes, and in e ac ions be ween b ood pa asi ic
common cuckoos (Cuculus cano us) and hei hos s cons i u e classic examples. The ou come o a pa asi ic e en is o en
de e mined by he ma ch be ween hos and cuckoo eggs, gi ing ise o po en ially s ong associa ions be ween i ness and
egg pheno ype. Ye , empi ical e o s aiming o documen and unde s and he esul ing e olu iona y ou comes a e in sho
supply.
Me hods/P incipal Findings:
We used a ian colo space models o analyze pa e ns o egg colo a ia ion wi hin and
be ween he cuckoo and wo closely ela ed hos s, he nomadic b ambling (F ingilla mon i ingilla) and he si e idelic
cha inch (F. coelebs). We ound ha he e is p onounced oppo uni y o dis up i e selec ion on b ambling egg colo a ion.
The co esponding cuckoo hos ace has e ol ed egg colo s ha maximize i ness in bo h sympa ic and allopa ic
b ambling popula ions. By con as , he cha inch has a mo e bimodal egg colo dis ibu ion consis en wi h he
e olu iona y di ec ion p edic ed o he b ambling. Whe eas he b ambling and i s cuckoo hos ace show li le
geog aphical a ia ion in hei egg colo dis ibu ions, he cha inch’s dis ibu ion becomes inc easingly dissimila o he
b ambling’s dis ibu ion owa ds he co e a ea o he b ambling cuckoo hos ace.
Conclusion:
High a es o b ambling gene low is likely o cool down coe olu iona y ho spo s by cancelling ou he
selec ion imposed by a pa chily dis ibu ed cuckoo hos ace, he eby p omo ing a ma ching equilib ium. By con as , he
si e idelic cha inch is mo e likely o espond o selec ion om adap ing cuckoos, esul ing in a ma kedly mo e bimodal egg
colo dis ibu ion. The geog aphic a ia ion in he cha inch’s egg colo dis ibu ion could e lec a his o ical g adien in
pa asi ism p essu e. Finally, ma ked cuckoo egg polymo phisms a e unlikely o e ol e in hese sys ems unless he hos s
e ol e e en mo e exquisi e egg ecogni ion capabili ies han cu en ly possessed.
Ci a ion: Vikan JR, Fossøy F, Huh a E, Moksnes A, Røska E, e al. (2011) Ou comes o B ood Pa asi e–Hos In e ac ions Media ed by Egg Ma ching: Common
Cuckoos Cuculus cano us e sus F ingilla Finches. PLoS ONE 6(4): e19288. doi:10.1371/jou nal.pone.0019288
Edi o : Robe C. Fleische , Smi hsonian Ins i u ion Na ional Zoological Pa k, Uni ed S a es o Ame ica
Recei ed Oc obe 11, 2010; Accep ed Ma ch 31, 2011; Published Ap il 29, 2011
Copy igh : ß2011 Vikan e al. This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s
un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed.
Funding: This wo k was suppo ed by he Eu opean Communi y - LAPBIAT p ojec RITA-CT-2006-025969. The unde s had no ole in s udy design, da a collec ion
and analysis, decision o publish, o p epa a ion o he manusc ip .
Compe ing In e es s: The au ho s ha e decla ed ha no compe ing in e es s exis .
* E-mail: [email p o ec ed]
In oduc ion
Nume ous an agonis ic species in e ac ia ma ching o
pheno ypes [1,2,3,4,5]. Some o he p ime examples o his kind
a e ound among a ian b ood pa asi es, such as he common
cuckoo (Cuculus cano us) (he ea e cuckoo), and hei hos s. In hese
in e ac ions, he ou come o a pa asi ic e en is o en de e mined
by he ma ch be ween cuckoo and hos eggs [6]. Since his
si ua ion can gi e ise o s ong associa ions be ween i ness and
egg pheno ype in bo h species, and because he he i abili ies
in ol ed a e high [7,8,9], much o he coe olu iona y dynamics o
hese in e ac ions is likely o in ol e egg pheno ypes. Indeed,
compa a i e e idence implica es coe olu ion as he main d i e
behind he high egg pheno ype di e si y ound in some hos s o
specialized b ood pa asi es [10,11]. Howe e , ou p esen insigh
in he ou comes o egg pheno ype coe olu ion es s almos
exclusi ely on ma hema ical models and heo e ical a gumen s
[12,13,14,15,16,17,18,19]. Theo e ical p edic ions co e a wide
ange o scena ios, om ma ching equilib ia (equal mean
pheno ypes) o non-ma ching equilib ia, coe olu iona y cycles
and s able poin polymo phisms, depending on he speci ic
assump ions made abou he s uc u e o gene ic a iance, he
le els o inhe i ance, and he p esence o s abilizing selec ion. This
di e si y o heo e ical ou comes wa an s de ailed empi ical
in es iga ions o egg pheno ype dis ibu ions in di e en hos
pa asi e-sys ems [18]. Once a ained, such da a can be used o
assess he oppo uni y o ecip ocal selec ion, de e mine he mode
o selec ion ac ing on hos and pa asi e, and e alua e pa asi e and
hos pheno ypes in ela ion o p edic ed op ima. Such app oaches
a e impo an because hey will acili a e mo e in o med
discussions abou e olu iona y di ec ions.
In his s udy, we compa e pa e ns o a ia ion in egg colo
dis ibu ions wi hin and be ween wo closely ela ed cuckoo hos s,
he b ambling (F ingilla mon i ingilla) and he cha inch (F. coelebs).
PLoS ONE | www.plosone.o g 1 Ap il 2011 | Volume 6 | Issue 4 | e19288
The cuckoo is known o comp ise specialized emale lineages
(called gen es) which in many cases ha e e ol ed eggs ha end o
mimic hose o hei espec i e hos s, esul ing in an as onishing
di e si y o egg ypes [20,21,22,23]. The wo F ingilla inches show
he same wide ange o egg colo s, including pu e blue, g een,
eddish-g ey and da k oli e-b own clu ches (Figu e 1A), and a e
hus ap o in es iga ing i ness in ela ion o a ia ion in egg
pheno ypes. Since egg ejec ion p obabili y is s ongly in luenced
by he ma ch be ween hos and cuckoo egg [24,25] and
accep ance o he cuckoo egg leads o a massi e educ ion in
nes ling p oduc ion [6], bo h inches show su icien a ia ion
be ween clu ches o gene a e signi ican i ness di e ences among
indi iduals in pa asi ized popula ions. Mo eo e , bo h species
possess all he basic ecological ea u es ha cha ac e ize p ime
cuckoo hos s. In pa icula , bo h build a bo eal and shallow nes
cups and o en make up 20-50% o hei espec i e b eeding
passe ine communi ies, and aise hei chicks on a p o ein die
sui able o cuckoos [26,27,28,29,30]. In addi ion, bo h show
highly de eloped egg ecogni ion abili ies h oughou a all si es
whe e hey ha e been es ed and bo h espond agg essi ely
Figu e 1. S udy popula ions, pheno ypic a ia ion o b ambling and cha inch eggs, and geog aphical dis ibu ion o museum
clu ches. A: ange o egg colo s o cha inch ( ows 1-3) and b ambling clu ches ( ows 4-6). Each egg comes om a di e en clu ch. The bo om ow
gi es examples o cuckoo eggs ound in b ambling nes s in Ki ila
¨. B: loca ion o s udy popula ions. C: geog aphic o igin o museum clu ches. The
poin s a e sligh ly ji e ed o plo ing pu pose. The maps we e d awn using he maps package o R [88].
doi:10.1371/jou nal.pone.0019288.g001
Egg Colou Coe olu ion
PLoS ONE | www.plosone.o g 2 Ap il 2011 | Volume 6 | Issue 4 | e19288
owa ds adul cuckoos [25,31,32,33,34]. Finally, he exis ence o a
cuckoo egg mo ph esembling F ingilla eggs has been well
documen ed [20,22,25,35,36]. Pa simony he e o e sugges s ha
he wo inches ha e a his o y as impo an hos s o he cuckoo in
Eu ope [34,37]. A he same ime, hei na u al his o ies imply
ha hey di e ma kedly in hei abili y o e ol e in esponse o
adap ing cuckoos, which leads o di e en p edic ions ega ding
coe olu iona y ou comes [12,13,17,38,39,40].
The b ambling is he no he n bo eal ecological coun e pa o
he cha inch [28]. I employs an oppo unis ic b eeding s a egy
o exploi a ion o a ou able bu unp edic able b eeding oppo -
uni ies [29,41], and he e o e has a e y low b eeding si e ideli y
[42,43,44] which is indica i e o high amoun s o gene low
possibly swamping local selec ion [45]. Ou o 1945 b eeding
adul s ma ked in Finnish and Swedish Lapland, only 11
indi iduals (0.6%) ha e subsequen ly been ecap u ed on he
same b eeding si e, whe eas none ou o 2300 ma ked ju eniles
ha e been ecap u ed [42,43, A
˚. Linds o¨m, unpublished da a].
Mo eo e , eco e ies a di e en b eeding si es indica e ha adul
b amblings may b eed a si es up o 600 km apa in di e en
yea s [43]. By con as , he cha inch exhibi s ma ked b eeding si e
ideli y [42,44,46], possibly e lec ing he mo e p edic able
ecological condi ions o a sou he n dis ibu ional ange [47]. In
Finnish Lapland, 39.3% o ma ked b eeding cha inches we e
ound o nes again a he same si e he ollowing yea [42]. The
po en ial e olu iona y consequences o hese di e ences a e
mani es ed in he subspecies-s uc u e o he wo species. While
he cha inch has a ma ked geog aphical a ia ion comp ising
se e al subspecies (7 in he coelebs g oup), no subspecies has so a
been ecognized in he b ambling [27,28].
Thus, heo y p esc ibes ha he ex en o egg pheno ype
e olu ion should di e signi ican ly be ween he wo species
because, all else being equal, he abili y o e ol e in esponse o
adap ing cuckoos should be ma kedly s onge in he cha inch.
In e ac ions ha a e media ed by pheno ype-ma ching a e likely o
in ol e e olu ion o inc easingly bimodal hos pheno ype dis i-
bu ions as a esponse o dis up i e selec ion imposed by mime ic
pa asi e pheno ypes [17]. We should he e o e (1) expec o ind
s onge bimodali y in he egg colo s o he cha inch han in he
b ambling. The i s aim o his s udy is o in es iga e he possible
occu ence o such di e ences. In o de o achie e his, we use a
e ahed al colo space model o compa e he egg colo
dis ibu ions o he wo species, e alua e he cu en oppo uni y
o dis up i e selec ion on b ambling egg colo s, and isualize he
di ec ion o p edic ed e olu iona y change in colo space.
Secondly, we e alua e cuckoo eggs in ela ion o p edic ed
pheno ypic op ima o h ee dis an b ambling popula ions in
Fennoscandia (one cu en ly pa asi ized and wo non-pa asi ized).
Gi en he nomadic na u e o b amblings, we p edic (2) ha
cuckoos should be equally well adap ed o he egg pheno ype
dis ibu ions o he b ambling in he h ee a eas. In o he wo ds,
we expec li le geog aphical a ia ion in he eggs o cuckoos
specializing on b amblings. In gene al, he geog aphical scale a
which we ind no iceable a ia ion in b ambling and cuckoo egg
pheno ypes should be e y la ge in his sys em.
Finally, (3) we examine i he ange o he coe olu iona y
ajec o y is likely o ex end beyond he s age o inc easing hos
a iance. One s ep in his di ec ion is o e alua e whe he e olu ion
on pa o he hos (bimodal egg colo dis ibu ion) can gi e ise o
co esponding adap i e peaks o cuckoo egg appea ance [16,48].
To his end, we es ima e a e age ejec ion p obabili ies o a wide
ange o po en ial cuckoo egg pheno ypes, and use hese
p obabili ies o ske ch he o ms o selec ion imposed by he hos s’
egg colo dis ibu ions and ele an subse s o hese dis ibu ions.
Me hods
E hics s a emen
Ou esea ch ollowed guidelines o he Animal Beha iou
Socie y o e hical use o animals in esea ch, and pe missions o
he p o ocols in ol ed in he ieldwo k we e p o ided by Lapland
Regional En i onmen Cen e (pe mi numbe s LAP-2005-L-41-
254, LAP-2005-L-666-254/1.3.2006, LAP-2008-L-191-254), The
Finnish Fo es y agency (pe mi numbe s 1737/42/2005, 1207/
41/2006, 2296/662/2008), The Swedish Commi ee o Animal
Resea ch (pe mi numbe A 35-08), and he No wegian Di ec o a e
o Na u e Managemen (pe mi numbe s 2008/1524 A -VI-
ID,05/2580 ART-VI-ARES, 2007/1177 ART-VI-JAA).
Field da a
The b ambling was s udied in h ee a eas in Fennoscandia
(Figu e 1B): Tydal in Cen al No way (63uN, 12uE, 2006-2008),
Amma na¨s in No he n Sweden (66uN, 16uE, 2008), and Ki ila¨in
No he n Finland (68uN, 25uE, 2005–2008). The Ki ila¨ popula-
ion is he only one cu en ly pa asi ized. Da a on hos egg
colo a ion we e ob ained om 47 clu ches in Tydal, 97 clu ches in
Amma na¨s, and 88 clu ches in Ki ila¨, whe eas da a on cuckoo egg
colo a ion we e ob ained om i e cuckoo eggs measu ed in
Ki ila¨. Gene ic analyses on ou o hem and olume and shape
analyses show ha hese eggs s em om a leas ou di e en
emales (own unpublished da a). The cha inch was s udied in an
unpa asi ized popula ion in S jø dal, Cen al No way (63uN,
11uE, 2007–2008), whe e da a om 157 clu ches we e ob ained.
Museum da a
In addi ion o ield da a, we also collec ed da a on cha inch,
b ambling, and cuckoo egg appea ance om clu ches held in he
collec ions o B i ish Museum (Na u al His o y), T ing, G ea
B i ain; Museum o Na u al His o y, Go henbu g, Sweden;
Museum o Na u al His o y, Copenhagen, Denma k; Finnish
Museum o Na u al His o y, Helsinki, Finland, and Swedish
Museum o Na u al His o y, S ockholm, Sweden. Da a on 343
b ambling and 625 cha inch clu ches collec ed du ing he pe iod
1839–1991 we e included in he analyses. Figu e 1C shows he
geog aphical dis ibu ion o he clu ches.
Based on inspec ion, cuckoo eggs we e classi ied o belong o a
F ingilla mo ph [22] i he egg was clea ly wi hin he ange o
a ia ion o b ambling and cha inch eggs (N = 72). These cuckoo
eggs we e laid in clu ches o b ambling (N = 47), cha inch
(N = 10), willow wa ble (Phylloscopus ochilus) (N = 5), obin
(E i hacus ubecula) (N = 2), spo ed lyca che (Muscicapa s ia a)
(N = 2), eed bun ing (Embe iza schoeniclus) (N = 1), ee pipi (An hus
i ialis) (N = 1), us ic bun ing (Embe iza us ica) (N = 1), chi cha
(Phylloscopus collybi a) (N = 1), yellow wag ail (Mo acilla la a) (N = 1),
and one unknown hos species. In all o he non-F ingilla hos s, he
sampling locali y was wi hin No he n Fennoscandia and
indica ed sympa y wi h ei he cha inch o b ambling.
Cuckoo eggs ound in he clu ches o cha inches o b amblings
ha ob iously belonged o a di e en cuckoo egg mo ph (such as
An hus o Mo acilla [22]) we e no included in he analyses. The e a e
wo main easons o his delimi a ion. Fi s ly, in he con ex o egg
pheno ype coe olu ion, cuckoo eggs ha a e ou side he ai space
o he hos a e likely o con ibu e li le o selec ion on hos egg
appea ance, because such eggs a e nea ly always ejec ed by bo h
hos s, i espec i e o he hos ’s own egg ype. We ha e conduc ed a
o al o 66 egg ejec ion expe imen s whe e he pa asi ic egg came
om a non-F ingilla species, and 92% o hose eggs we e ejec ed
(19/19 expe imen s wi h b amblings and 42/47 expe imen s wi h
cha inches). Secondly, he occu ence o such cuckoo eggs may
Egg Colou Coe olu ion
PLoS ONE | www.plosone.o g 3 Ap il 2011 | Volume 6 | Issue 4 | e19288
di e be ween he wo F ingilla hos s o easons ha a e comple ely
un ela ed o a coe olu iona y p ocess, o example because o
di e ences in hos communi y composi ion (hos -speci ic gen es may
acciden ally lay eggs in o he hos s’ nes s [22]).
Da a on sampling loca ion was a ailable o 581 cha inch
clu ches, 341 b ambling clu ches, and 68 cuckoo eggs. In 771 cases
(438 cha inch, 270 b ambling, and 63 cuckoo), he clu ch label
con ained in o ma ion abou ci y/municipali y o a speci ic si e
wi hin a ci y/municipali y. In he emaining 218 cases (5 cuckoo,
143 cha inch, 70 b ambling), he label con ained in o ma ion
abou he coun y/p o ince/shi e in which he clu ch was
collec ed. In hese cases, he si e o collec ion was de ined as he
cen e o he coun y/p o ince/shi e (chosen by isual inspec ion o
he map). Da a on he yea o collec ion was a ailable o 600
cha inch clu ches, 333 b ambling clu ches, and 72 cuckoo eggs.
Egg Expe imen s
We ca ied ou egg exchange expe imen s ac oss all s udy
popula ions in o de o ob ain he hos disc imina ion unc ion
which bes desc ibes he ela ionship be ween he colo con as
be ween hos and pa asi ic eggs and he p obabili y o egg ejec ion.
In his s udy, he hos disc imina ion unc ion is used o es ima e
a e age su i al p obabili ies o hos and cuckoo eggs. As
expe imen al pa asi ic eggs, we used eal b ambling and cha inch
eggs. Hence, he hos disc imina ion unc ion ob ained applies o
di e ences ha occu wi hin he bounda ies o he ai space o he
wo hos s. A o al o 288 success ul expe imen s we e eco ded (137
wi h b amblings (14 o he pa asi ic eggs whe e cha inch eggs) and
151 wi h cha inches (all pa asi ic eggs we e cha inch eggs)). The
esul o each expe imen was classi ied as ei he ejec ion (pa asi ic
egg ejec ed) o accep ance (pa asi ic egg incuba ed o a leas i e
days). We ha e shown in wo sepa a e s udies ha p e ious
expe imen al pa asi ism does no a ec he p obabili y o ejec ion
o a pa asi ic egg added la e in he same b eeding a emp [25,49].
We he e o e included se en y- ou expe imen s ha we e eplica es
a he indi idual le el. In all cases whe e wo expe imen s we e
ca ied ou on he same indi idual, wo di e en pa asi ic eggs we e
used (one o high con as and one o low-medium con as ). Fo
u he de ails abou he egg expe imen s see [24,25,49].
Measu emen s o Egg Shell Re lec ance
We used a spec opho ome e o ob ain e lec ance spec a o
he g ound colo o he eggs. The measu emen p ocedu es a e
de ailed elsewhe e [24,25,49]. Eggs collec ed in he ield we e all
esh when measu ed. One andom egg was measu ed in each
clu ch, which is jus i ied by he ex ao dina y low in aclu ch
a ia ion ound in hese wo species [34,50]. Fou (n = 1080) o
eigh (n = 126) measu emen s we e aken om each egg, which is
su icien o desc ibe backg ound colo a ion adequa ely in hese
species [49]. We he e o e calcula ed an a e age spec um om
he ou (eigh ) measu emen s, and used hese spec a o he
subsequen analyses.
Colo Space, Egg Colo Con as s, and Egg Colo
Dis ibu ions
In o de o analyze egg colo s and egg colo dis ibu ions we
applied a e ahed al colo space model which has been s ongly
ad oca ed o use in s udies o colo e olu ion in e ach oma s
[51]. All o mulas used o he colo space calcula ions a e de ailed
in [51,52]. The main ad an ages o he his colo space model a e
ha i makes e y ew assump ions, i is p agma ic and
quan i a i ely p ecise, and i p o ides a anspo able scale o
measu emen ha can be compa ed among independen analyses
(unlike he p incipal componen analyses which hi he o has
domina ed egg colo esea ch) [51]. The posi ion o any colo in
he colo space is de e mined by he ela i e s imula ion o he ou
e inal cone ypes by he e lec ance spec um unde idealized ligh
condi ions. Thus, he posi ion o each egg colo is gi en by a se o
ela i e cone s imula ion alues {u , s, m, l}. The ou e ices o
he e ahed on co espond o exclusi e s imula ion o he
ul a iole -sensi i e (u ), sho -wa eleng h-sensi i e (s), medium-
wa eleng h-sensi i e (m), and long-wa eleng h-sensi i e cone
pho o ecep o s (l) (see Figu e 1 in [51]). Fo hese calcula ions,
we used he a e age o spec al sensi i i y cu es o UVS – ype
e inas om Endle and Mielke [53, a ailable in hei supple-
men a y online ma e ial]. Following [51], he cone s imula ion
alues we e no malized o sum o 1 and hen ans o med in o
Ca esian coo dina es {x, y, z}. The chosen ans o ma ion places
he ach oma ic poin o equal cone s imula ion a he o igin and
he u - e ex along he z-axis (see Figu e 1 in [51]).
The colo con as (DT
C
), i.e. le el o mimic y, be ween any wo
eggs was calcula ed as he Euclidean dis ance be ween he wo egg
colo s in e ahed al colo space. Hue and sa u a ion, which a e
undamen al colo aspec s, we e ob ained by de i ing he sphe ical
coo dina es h,w, and o he colo ec o {x, y, z} (see Figu e 1 in
[51]). The hue o a colo is de ined as he di ec ion o he colo
ec o , and is he e o e gi en by he angula displacemen o he
colo ec o om he posi i e x-axis (h[½{p,zp), which uns
be ween he m (g een) and l ( ed) e ices o he e ahed on, and
he angula displacemen om he ho izon al xy-plane (W
[½{
p
2,z
p
2), which equals he u -componen o hue [51]. hand
Wa e analogous o longi ude and la i ude, espec i ely. A
con enien and heu is ic way o isualize he dis ibu ion o hues
o a sample o clu ches is o map he hues on o a uni sphe e
cen e ed a he ach oma ic o igin and de i e hei wo-
dimensional Robinson p ojec ions (sensu [54]).
Sa u a ion desc ibes how di e en a colo is om ach oma ic
whi e/black. The leng h o he colo ec o is he e o e a measu e
o sa u a ion. This means ha colo s ha di e in sa u a ion bu
no in hue a e posi i e scala mul iples o he same colo ec o .
B igh ness is no pa o he e ach oma ic colo space, bu
con as in b igh ness can be an impo an ecogni ion cue o he
hos in some ci cums ances [55,56]. Howe e , his does no appea
o be he case wi h b amblings and cha inches [24,25, see Resul s
sec ion], and we he e o e only ocus on colo (i.e. a ia ion
be ween e lec ance spec a independen o in ensi y) in his s udy.
Es ima ing he Fo m o Selec ion on Cuckoo and Hos
Egg Colo
In o de o e alua e he di ec ion o selec ion on b ambling egg
pheno ypes we calcula ed he a e age colo con as o each
clu ch based on pai ings wi h eal cuckoo eggs. Fo b ambling
clu ches measu ed in he ield we used he i e cuckoo eggs ound
in b ambling nes s in Ki ila¨. Fo museum clu ches we used 72
F ingilla ype cuckoo eggs (see abo e). In o de o e alua e how
close cuckoo eggs a e o he op imal pheno ype o a popula ion,
we calcula ed pai wise colo con as s be ween all clu ches in a
popula ion, and he bes achie able mime ic egg (i.e. op imal
cuckoo egg ype) o ha popula ion was aken o equal he clu ch
wi h he lowes a e age colo con as . The a e age ejec ion
p obabili y o op imal cuckoo egg ypes was es ima ed using he
hos disc imina ion unc ion.
In o de o e alua e he po en ial o ms o selec ion imposed on
cuckoo eggs by he hos s, we sampled 3000 po en ial cuckoo egg
colo s om wi hin he e ach oma ic colo space o each hos ha
we e e enly dis ibu ed o hue and sa u a ion and hen calcula ed
Egg Colou Coe olu ion
PLoS ONE | www.plosone.o g 4 Ap il 2011 | Volume 6 | Issue 4 | e19288
he a e age accep ance p obabili y o each o hese eggs. In o de
o assess he e ec s o inc easing hos bimodali y on he adap i e
landscape o he cuckoo, we also simula ed ha he po en ial
cuckoo eggs we e acing app op ia e subse s o he hos ’s cu en
dis ibu ion (i.e. inc easingly bimodal hos dis ibu ions).
S a is ical Analyses
All calcula ions and analyses we e ca ied ou in R2.8.1. We used a
binomial logis ic eg ession (logi link) o ob ain pa ame e es ima es
o p edic ion o egg ejec ion p obabili y. We used dis ance based
pe mu a ion es s o compa isons o egg colo dis ibu ions wi hin
and among species, and o e alua ing changes in egg colo
dis ibu ions along la i udinal/longi udinal g adien s. Man el es s
[57] we e used o examine he deg ee o conco dance be ween
geog aphical dis ances and colo dis ances (DT
C
). Man el es s we e
pe o med using he egan package [58]. Yea o collec ion was
con olled o in pa ial Man el es s. The e was a weak empo al
end in he egg colo s o museum eggs (explaining 1.2% o he o al
a ia ion), and he e ec was simila in all h ee species (own
unpublished esul s). Fu he mo e, he e was no ela ionship be ween
yea o collec ion and geog aphical coo dina es o sampling locali y
(own unpublished esul s). Finally, he di ec ion o he empo al end
was in close ag eemen wi h wha one should expec ega ding he
e ec s o s o age and ageing (JR Vikan, unpublished esul s).
Howe e , since he aim o his s udy is no o de e mine wha ac o s
con ibu e o he empo al e ec , he c ucial poin o emphasize is
ha museum eggs can be used o conduc meaning ul compa isons o
geog aphical a ia ion in he h ee species.
Geog aphical dis ances be ween locali ies we e es ima ed using
he dis .ea h unc ion in he ields package [59], which e u ns
G ea ci cle dis ances om longi ude/la i ude da a. We used
pe mu a ional MANOVAs ( unc ion adonis in egan package) o
examine e ec s o la i ude/longi ude on colo space loca ion. This
p ocedu e pa i ions he a ia ion inhe en in dis ance ma ixes (in
ou case ma ixes in which elemen s a e colo dis ances be ween
eggs), and uses pe mu a ion es s o inspec he signi icances o
hose pa i ions (so called pseudo-F es s). The adonis unc ion also
e u ns a pa ial-R
2
, which is an es ima e o he p opo ion o
a iance explained by he a iable.
We used a mul i a ia e analogue o Le ene’s es o es o
di e ences in egg colo a iances ac oss species and popula ions
(Mul i a ia e homogenei y o g oup dispe sions, MHGD). The
p ocedu e wo ks by i s calcula ing he dis ances be ween egg
colo s and espec i e g oup cen oids ( unc ion be adispe in egan
package). The magni udes o hese dis ances a e hen compa ed
be ween g oups using o dina y ANOVA.
Finally, we used he Euclidean dis ance e sion o a mul i-
esponse pe mu a ion p ocedu e (MRPP [60]) o es o o e all
di e ences in he egg colo dis ibu ions o cha inches and
b amblings. MRPP is closely ela ed o pe mu a ional MANOVA,
since bo h a e pe mu a ion es s based on dis ances among
mul i a ia e obse a ions [61]. In he MRPP p ocedu e, he e ec
size o he di e ence be ween dis ibu ions is gi en by a dispa i y
s a is ic (K), which akes in o accoun di e ences in g oup
cen oids, a iances, skewness, ku osis and shape [53].
Numbe o pe mu a ions was se o 1000 o he Man el and
pa ial Man el es s, and 10000 o he MRPP and pe mu a ional
MANOVAs.
Resul s
The Hos Disc imina ion Func ion
The egg exchange expe imen s e ealed no signi ican in e ac-
ions be ween species and he ou di e en measu es o con as
be ween eggs (DT
C
, b igh ness con as , olume con as , shape
con as ). Fu he mo e, DT
C
was he only e m ha was e ained
a e model simpli ica ion (Table 1). Hence, equal alues o DT
C
gi es equal p obabili ies o egg ejec ion in bo h species.
Mo eo e , since con as in olume o shape does no seem o
a ec ejec ion p obabili y, his indica es ha he hos disc imi-
na ion unc ion (Figu e 2) app oxima es he hos s’ esponses o eal
cuckoo eggs. Figu e 2 shows he ela ionship be ween ejec ion
p obabili y and DT
C
o he pooled da a.
Na u al Pa asi ism and Appea ance o Cuckoo Eggs in
Rela ion o Op imum
The dis ibu ion o pai wise colo con as s (DT
C
) we e simila
in all popula ions s udied (Table 2), meaning ha in e clu ch
a ia ion in egg colo a ion is o simila magni ude. The a e age
mimic y (DT
C
) o he bes achie able mime ic eggs was also
simila . Acco dingly, he a e age accep ance p obabili ies o
op imal cuckoo egg ypes a e o he same magni ude ac oss
popula ions and species (i.e. a ound 50%, Table 2).
The cuckoo eggs measu ed in he ield in Ki ila¨ had colo s ha
we e e y close o he op imum o maximizing accep ance by
b ambling hos s. The same cuckoo eggs would also be close o he
op imum in wo emo e unpa asi ized popula ions (Figu e 3A–D).
When analyzed a he scale o Fennoscandia, museum F ingilla- ype
cuckoo eggs also ended o ha e op imal colo s o pa asi izing
b amblings (Figu e 3E). Fu he mo e, he e was no co ela ion
be ween geog aphical dis ances and colo dis ances o b ambling
eggs (Table 3). This co obo a es he ield da a by sugges ing ha
he op imal egg colo s o cuckoos do no a y no iceably wi hin
Fennoscandia. Acco dingly, he e was no co ela ion be ween
geog aphical dis ances and colo dis ances o cuckoo eggs (Table 3),
indica ing ha F ingilla ype cuckoo eggs do no show any ma ked
geog aphical a ia ion in colo dis ibu ion wi hin Fennoscandia.
Among he museum cuckoo eggs ound in cha inch clu ches,
only 22% (N = 45) could be classi ied as a F ingilla ype cuckoo
Table 1. Gene alized linea models o he p obabili y o
accep ing congene ic pa asi ic eggs in cha inches and
b amblings.
Pa ame e D De iance P
DT
C
1 52.8 ,0.0001
DT
B
1 3.48 0.06
Species 1 0.43 0.51
Shape 1 0.35 0.55
Volume 1 0.33 0.57
DT
B
6Species 1 2.40 0.12
Volume 6Species 1 1.51 0.22
DT
C
6Species 1 0.65 0.42
DT
B
6DT
C
1 0.15 0.70
Shape 6Species 1 0.03 0.87
S epwise backwa d dele ion was used o simpli y he global model including all
pa ame e s, and signi icance o pa ame e s was e alua ed by he change in
de iance be ween models wi h and wi hou he pa ame e in ques ion. DT
C
is
he only e m included in he minimal adequa e model. DT
B
deno es b igh ness
con as , and was calcula ed acco ding o [25]. Volume and shape was
calcula ed acco ding o [86] and [87], espec i ely. Rejec ion a es we e 0.60
(N = 151) o he cha inch and 0.66 (N = 137) o he b ambling. Mean (SD) o
DT
C
and DT
B
, espec i ely, was 0.044 (0.023) and 6.48 (4.20) o he cha inch
and 0.051 (0.031) and 5.52 (4.01) o he b ambling.
doi:10.1371/jou nal.pone.0019288. 001
Egg Colou Coe olu ion
PLoS ONE | www.plosone.o g 5 Ap il 2011 | Volume 6 | Issue 4 | e19288

egg. In compa ison, 78% (N = 55) cuckoo eggs ound in
b ambling clu ches we e F ingilla ype. These numbe s a e in
close ag eemen wi h a p e ious s udy [22] which classi ied 12%
(N = 76) o cuckoo eggs in cha inch clu ches as F ingilla ype
compa ed o only 77% (N = 53) o cuckoo eggs in b ambling
clu ches.
Egg Colo Dis ibu ions o B ambling and Cha inch
The dis ibu ion o clu ches in he colo e ahed on di e ed
signi ican ly be ween he cha inch and b ambling, conside ing
bo h ield da a (Euclidean dis ance MRPP on xyz-coo dina es,
P,0.001, K = 0.12) and museum da a (Euclidean dis ance MRPP
on xyz coo dina es, P,0.001, K = 0.03).
The Robinson p ojec ions o hues illus a e he pa e ns o
a ia ion in hue independen o sa u a ion (Figu e 4). The
dis ibu ion o hues di e ed signi ican ly be ween cha inch and
b ambling clu ches, conside ing bo h ield da a (Euclidean
dis ance MRPP on h,W,P,0.001, K = 0.02, Figu e 4B–C) and
museum da a (Euclidean dis ance MRPP on h,W,P,0.001,
K = 0.01, Figu e 4D–E). The p ojec ions show ha a la ge
p opo ion o he hues a e ound in bo h species, excep om he
mos pu e blue hues which we e no ound among b ambling
clu ches collec ed in he ield (Figu e 4B–C). The p ojec ions also
show ha he hue dis ibu ion o he cha inch is clea ly hicke a
he ails (i.e. mo e bimodal) compa ed o he b ambling
(Figu e 4B–E), bo h o ield and museum clu ches. Acco dingly,
he a iance o he cha inch’s hue dis ibu ion is signi ican ly
g ea e han he a iance o he b ambling’s hue dis ibu ion
(MHGD, ield da a: d = 1, 387, F = 78.2, P,0.0001; museum
da a; d = 1, 966, F = 7.80, P = 0.005). The a iance o he
b ambling’s hue dis ibu ion was simila in he h ee popula ions
s udied (MHGD, d = 2, 229, F = 0.49, P = 0.61).
A bimodal dis ibu ion o hues indica es ha he e a e wo
di e en clus e s in e ach oma ic colo space: i.e. all colo s a e
ound close o wo sepa a e axes unning om he cen e o he
colo space. Fo bo h inches, he sa u a ion o he clu ch colo s
we e app oxima ely no mally dis ibu ed (Figu e S1). In addi ion,
inspec ion o esidual plo s sugges ed ha he e a e no ma ked
Table 2. Summa y s a is ics desc ibing di e en aspec s o he clu ch colo dis ibu ions o he b ambling and cha inch.
Popula ion/Sample N
Colo con as (DT
C
)
max, mean (SD)
Sa u a ion
mean (SD)
A e age colo con as
(DT
C
) o op imal cuckoo
egg ypes
A e age ejec ion a e
o op imal cuckoo egg
ypes (SD)
B ambling
Field da a (2007–2008)
Tydal 47 0.136, 0.047 (0.025) 0.109 (0.019) 0.034 (0.017) 0.53 (0.16)
Ki ila 88 0.177, 0.044 (0.025) 0.100 (0.018) 0.031 (0.018) 0.50 (0.16)
Amma na
¨s 97 0.159, 0.043 (0.025) 0.102 (0.016) 0.031 (0.017) 0.50 (0.16)
Fennoscandia 232 0.180, 0.045 (0.025) 0.103 (0.018) 0.032 (0.018) 0.51 (0.16)
Museum da a (1881–1940)
Fennoscandia 343 0.173, 0.040 (0.024) 0.106 (0.017) 0.028 (0.017) 0.47 (0.16)
Cha inch
Field da a (2007–2008)
S jø dal 157 0.152, 0.045 (0.027) 0.070 (0.016) 0.033 (0.018) 0.52 (0.17)
Museum da a (1881–1940)
G ea B i ain 273 0.151, 0.044 (0.025) 0.095 (0.022) 0.032 (0.017) 0.51 (0.16)
Sweden & Denma k 273 0.143, 0.041 (0.024) 0.089 (0.015) 0.030 (0.016) 0.49 (0.15)
Finland, Ka elen (Russia) & Es onia 79 0.120, 0.042 (0.024) 0.089 (0.017) 0.031 (0.014) 0.50 (0.14)
G ea B i ain and Fennoscandia 625 0.173, 0.043 (0.025) 0.091 (0.019) 0.031 (0.016) 0.50 (0.15)
Op imal cuckoo egg ypes equal he hos egg ype which achie es he lowes DT
C
alue/ ejec ion p obabili y when a e aged o e all possible pai wise combina ions in
which he egg ea u es. Rejec ion a es a e p edic ed om a uni a ia e logis ic eg ession o DT
C
on he p obabili y o ejec ing a pa asi ic egg (Figu e 2).
doi:10.1371/jou nal.pone.0019288. 002
Figu e 2. Hos disc imina ion unc ion. Ba plo showing he
ela ionship be ween ejec ion a e and colo con as (DT
C
) be ween
hos and pa asi ic eggs. Supe imposed is he hos disc imina ion
unc ion ob ained om a uni a ia e logis ic eg ession on he
p obabili y o ejec ion (95% con idence in e al indica ed by do ed
lines). Fo plo ing pu poses he wid h o he ba s o some g oups we e
adjus ed o ob ain simila sample sizes ac oss g oups. Sample sizes o
each g oup depic ed abo e he ba s.
doi:10.1371/jou nal.pone.0019288.g002
Egg Colou Coe olu ion
PLoS ONE | www.plosone.o g 6 Ap il 2011 | Volume 6 | Issue 4 | e19288
non-linea i ies in he ela ionship be ween sa u a ion and hue
(Figu e S2). Toge he , his indica es ha he mul imodal
endencies in he egg colo s a e ound in he hue dimension,
which is cap u ed by he Robinson p ojec ions.
Compa ed o esh clu ches, hue dis ibu ions o museum
clu ches we e skewed owa ds he l ( ed) e ice, and clu ches in
he blue and blue-g een segmen ha e ma kedly educed alues o
he ul a iole componen (W) o hue (Figu e 4B–E). P elimina y
analyses o eggs collec ed in he ield show ha he colo s o hese
eggs ha e changed in he same di ec ion in one yea (JR Vikan,
unpublished esul s). Fo his eason, we will no discuss he
di e ences be ween museum and ield samples any u he in his
pape .
Fo m o Selec ion Imposed on B ambling Clu ches by
Cuckoo Eggs
We calcula ed he a e age DT
C
o each b ambling clu ch in
o de o e alua e he di ec ion o selec ion on hue. We assumed
ha he cuckoo egg dis ibu ion was gi en by he F ingilla cuckoo
egg ypes measu ed in his s udy ( ea ing ield and museum da a
sepa a ely), which a e close o hei op imum colo (Figu e 3). The
calcula ions indica e ha he e should be huge oppo uni ies o
dis up i e selec ion on hue (Figu e 4B, 4D). Fo example, whe eas
b amblings wi h g eenish clu ches would ha e an expec ed
p obabili y o ejec ing cuckoo eggs o abou 50%, he co e-
sponding p obabili y would be 80% o highe o b amblings
laying eggs wi h ex eme hues (i.e. eddish-b own o bluish). An
e olu iona y esponse in he b ambling would he e o e ansien ly
change he hue dis ibu ion in he di ec ion o s onge bimodali y,
which in u n would cause he dis ibu ion o con e ge on he
cha inch dis ibu ion (Figu e 4C, 4E).
Geog aphical Va ia ion in Egg Colo s
The e was a weak bu s a is ically signi ican co ela ion
be ween geog aphic dis ance and egg colo dis ance in he
cha inch, as e ealed by Man el and pa ial Man el es s
(Table 3). The co ela ion emained signi ican when we es ic ed
he analysis o clu ches om Sweden and Denma k (Table 3). The
dis ibu ion o geog aphical dis ances o hese clu ches is
compa able o ha o b amblings in Fennoscandia, o which
he e was no co ela ion be ween colo dis ances and geog aphical
dis ances. In acco dance wi h he Man el es s, pe mu a ional
MANOVA’s e ealed a signi ican e ec o la i ude/longi ude on
cha inch egg colo s (Analyses o dis ance: G ea B i ain and
Fennoscandia: La i ude: d = 1, F = 43.1, R
2
= 0.069, P = 0.0001;
Longi ude: d = 1, F = 32.6, R
2
= 0.053, P = 0.0001; Fennoscandia:
La i ude: d. . = 1, F = 16.5, R
2
= 0.046, P = 0.0001; Longi ude:
d = 1, F = 8.31, R
2
= 0.024, P = 0.0022; Sweden and Denma k:
Figu e 3. Colo s o cuckoo eggs in ela ion o p edic ed
op imum o pa asi ism o b amblings. Black iangles (A–D) and
black ba s (E) indica e he mean colo con as o cuckoo eggs esul ing
om compa ison wi h all b ambling clu ches in a popula ion. The
le mos e ical line in each igu e indica es he lowes a e age colo
con as ha can be achie ed, gi en he colo dis ibu ion o he hos ’s
clu ches. The igh mos e ical line in each igu e indica es he highes
a e age colo con as a cuckoo egg could ha e, and s ill be a pe ec
ma ch o a leas one o he hos clu ches in he popula ion. Cuckoo
eggs in A–D we e all ound in Ki ila
¨(N = 5), whe eas cuckoo eggs in e
we e measu ed in museums (N = 72). A: Ki ila
¨(N = 88 hos clu ches).
B: Amma na
¨s (N = 97 hos clu ches). C: Tydal (N = 47 hos clu ches).
D: Ki ila
¨,Amma na
¨s and Tydal pooled (N=232 hos clu ches). E: museum
clu ches collec ed in he pe iod 1841–1979 (N= 343 hos clu ches). F: a
b ambling clu ch wi h h ee hos eggs and one cuckoo egg ( op).
doi:10.1371/jou nal.pone.0019288.g003
Egg Colou Coe olu ion
PLoS ONE | www.plosone.o g 7 Ap il 2011 | Volume 6 | Issue 4 | e19288
La i ude: d = 1, F = 17.1, R
2
= 0.060, P = 0.0001; Longi ude:
d = 1, F = 10.8, R
2
= 0.039, P = 0.0004), bu no b ambling egg
colo s (Analyses o dis ance: La i ude: d = 1, F = 1.70, R
2
= 0.005,
P = 0.18; Longi ude: d = 1, F = 0.42, R
2
= 0.001, P = 0.59), o
cuckoo egg colo s (Analyses o dis ance: La i ude: d = 1, F = 1.03,
R
2
= 0.015, P = 0.33; Longi ude: d = 1, F = 0.57, R
2
= 0.009,
P = 0.54).
In o de o assess in wha di ec ion he cha inch egg colo
dis ibu ion changes wi h la i ude/longi ude, we compa ed he
b ambling’s egg colo dis ibu ion wi h he cha inch’s dis ibu ion
om G ea B i ain, sou h-wes e n Fennoscandia, and no h-
eas e n Fennoscandia, espec i ely. Clu ches we e g ouped in o
sou h-wes e n o no h-eas e n Fennoscandia acco ding o he
median la i ude. The esul s a e quan i a i ely simila when
Fennoscandian clu ches a e g ouped acco ding o he median
longi ude (because alues o longi ude and la i ude a e s ongly
co ela ed,
s
= 0.75, P,0.0001, Figu e 1C). Ranked by e ec size
(K), b ambling clu ch colo s we e mos simila o cha inch
clu ches om G ea B i ain (Euclidean dis ance MRPP on xyz-
coo dina es, P,0.0001, K = 0.028), less simila o he sou h-
wes e nmos cha inch clu ches (Euclidean dis ance MRPP on xyz-
coo dina es, P,0.0001, K = 0.037), and leas simila o no h-
eas e nmos cha inch clu ches in Fennoscandia (Euclidean
dis ance MRPP on xyz-coo dina es, P,0.0001, K = 0.064). These
esul s we e pa alleled by he magni ude o di e ence in hue
dis ibu ions: he b ambling was mos simila o he cha inch om
G ea B i ain (Euclidean dis ance MRPP on h,W, P = 0.031,
K = 0.004), less simila o he sou h-wes e nmos cha inch
(Euclidean dis ance MRPP on h,W, P = 0.005, K = 0.008), and
clea ly leas simila o he no h-wes e nmos cha inch (Euclidean
dis ance MRPP on h,W, P = 0.0001, K = 0.053, Figu e 4D, Figu e
S3). Finally, he a iance o he hue dis ibu ion inc eased
signi ican ly om sou h-wes o no h-eas : (0.59,0.71,1.05,
MHGD wi h Tukey HSD: P,0.029, Figu e S3).
Selec ion on Cuckoo Egg Colo
To esol e i he e a e any ine de ails in he o m o selec ion on
cuckoo egg colo s, we calcula ed he mean accep ance p obabili y
o a la ge numbe o hypo he ical cuckoo egg colo s andomly
selec ed om wi hin he e ahed al colo spaces occupied by he
wo hos s, espec i ely. Figu e 5 (A–B) shows he o ms o selec ion
imposed on he longi udinal (h) componen o hue by he wo hos
dis ibu ions (selec ion on he la i udinal componen is always
di ec ional, and selec ion on sa u a ion is s abilizing owa ds he
same alue o all h). The esul s show ha selec ion on cuckoo egg
colo s would be mainly s abilizing, bu wi h a wide pla eau in he
cha inch case. To in es iga e i dis inc adap i e peaks could
e ol e, i.e. gi ing ise o wo dis inc cuckoo egg mo phs, we
de ined subse s o he cha inch dis ibu ion ha had s onge
bimodali y and e-calcula ed he cu es. Al hough inc easing
bimodali y had he expec ed e ec o educing he a e age
accep ance p obabili y, i did no gi e ise o dis inc peaks
(Figu e 5C–E).
Discussion
Al hough cuckoo-hos in e ac ions a e amously known as
mimic y sys ems [20,21,62,63,64], mos o he a en ion o da e
has been on he e olu iona y ajec o ies o egg ecogni ion,
whe eas he dynamic aspec s o mimic y ha e been la gely
neglec ed in he majo discussions [65,66,67].
In he p esen s udy, we ha e shown ha cuckoos specializing
on b amblings ha e e ol ed egg colo s ha a e close o
maximizing accep ance a e o hei eggs, and ha his in u n
c ea es ample oppo uni y o dis up i e selec ion on hos egg
colo . One impo an ques ion is whe he his si ua ion could
e lec a ma ching equilib ium (i.e. equal mean pheno ypes).
Theo e ically, coe olu ion leads o a ma ching equilib ium i he
esponse o selec ion, which is p opo ional o he addi i e gene ic
a iance o he ai and he in ensi y o selec ion ac ing on i , is
s onge o he pa asi e han he hos [12,13,18]. Gi en he
b ambling’s pe asi e lack o si e ideli y, he e olu iona y uni is
likely o be e y la ge, possibly comp ising he whole egional
popula ion. In his si ua ion, he a e age pa asi ism a e ac oss he
whole egion would de e mine he selec ion p essu e on b ambling
egg colo (i.e. a panmic ic model would be app op ia e). Vas
numbe s o b amblings b eed in subalpine bi ch o es s in
Fennoscandia, a eas o which he e exis no e en a single epo
o pa asi ism [25,28]. The e o e, al hough i is possible ha
pa asi ism a es can be high locally, he p opo ion o he en i e
Fennoscandian b ambling popula ion being pa asi ized is likely o
be e y small a p esen . By con as , cuckoos a e subjec o s ong
selec ion e e ywhe e because b amblings possess simila egg
disc imina ion capabili ies h oughou Fennoscandia [25]. Fo
Table 3. Man el and pa ial Man el es s.
Da ase N clu ches Mean (SD) Ma ix compa ison P
Yea Geo
Cha inch 569 27 (20) 836 (569) Geo-colo (yea ) 0.061 0.001
Geo-colo 0.062 0.001
Cha inch, Fennoscandia 339 27 (20) 517 (361) Geo-colo (yea ) 0.039 0.019
Geo-colo 0.046 0.008
Cha inch, Sweden and Denma k 320 26 (21) 320 (230) Geo-colo (yea ) 0.055 0.013
Geo-colo 0.068 0.007
B ambling, Fennoscandia 330 29 (23) 393 (266) Geo-colo (yea ) 20.018 0.74
Geo-colo 20.015 0.75
Cuckoo 67 23 (18) 415 (260) Geo-colo (yea ) 0.021 0.36
Geo-colo 0.034 0.29
Man el es s we e pe o med on ma ixes o geog aphical dis ances (geo) and colo space dis ances (colo ), o cha inch, b ambling, and cuckoo eggs. The pa en heses
indica e ha yea o collec ion is con olled o in a pa ial Man el es .
doi:10.1371/jou nal.pone.0019288. 003
Egg Colou Coe olu ion
PLoS ONE | www.plosone.o g 8 Ap il 2011 | Volume 6 | Issue 4 | e19288
his eason, he ecip ocal selec ion p essu es ac ing on egg colo
a e likely o be s ongly asymme ic in he b ambling-cuckoo
in e ac ion. An analogous si ua ion seems o be ound in he
in e ac ion be ween he pa asi ic Maculinea bu e ly and i s an
hos My mica uginodis, whe e high le els o gene low in M. uginodis
p oduces coe olu iona y cold spo s by cancelling ou any selec ion
imposed by he pa chily dis ibu ed Maculinea [4].
The e is some indica ion o consis en egional di e ences in
pa asi ism o b amblings, wi h Finnish Lapland being a po en ial co-
e olu iona y ho spo (Figu e 1C). Al hough pa asi ized b ambling
popula ions should expe ience conside able a iance in ela i e
i ness ela ed o egg colo (Figu e 2B, 2D), cuckoo eggs appea o be
equally well adap ed o each o h ee dis an b ambling popula ions
(Figu e 3). Mo eo e , ega ding he museum da a, we ound no
co ela ion be ween geog aphic dis ances and egg colo dis ances
wi hin Fennoscandia, ei he o he b ambling o he cuckoo. Ou
esul s he e o e join ly ag ee wi h he p edic ed e ec s o lack o si e
ideli y in he b ambling, and a e consis en wi h analyses o
geog aphical a ia ion in o he ai s [25]. The e o e, as long as
pa asi ism o b amblings is pa chily dis ibu ed (i.e. he global
pa asi ism a e is low) nomadic beha iou could p omo e long e m
pe sis ence o he in e ac ion since he op imum o he cuckoo does
no change o e ime (o a mos e y slowly) and does no a y
be ween loca ions in Fennoscandia.
Figu e 4. Colo (hue) dis ibu ions o b ambling, cha inch and cuckoo eggs. (See main ex o explana ion o Robinson p ojec ions). A: he
en i e p ojec ed e ach oma ic hue-space o e e ence. Red iangles indica e he p ojec ions o he ul a iole (u ), sho (s), medium (m) and long
wa eleng h (l) e ices o he e ahed on. Solid lines indica e he p ojec ions o he ou edges connec ing he di e en e ices. B–C: he dis ibu ion
o hues o 232 b ambling (pooled clu ches om Ki ila
¨, Amma na
¨s and Tydal), and 157 cha inch clu ches (S jø dal), espec i ely. D–E: he dis ibu ion
o hues o 343 b ambling (Fennoscandia) and 625 cha inch clu ches (G ea B i ain and Fennoscandia) held in a ious museum collec ions. F: he
dis ibu ion o hues o 72 museum cuckoo eggs classi ied as F ingilla egg mo phs (blue colo ) and i e esh cuckoo eggs ound in b ambling nes s in
Ki ila
¨in he pe iod 2005–2008 ( ed colo ). Fo plo ing pu poses only, h ee clu ches in E loca ed close o he s- e ice (i.e. bluish hues) we e sligh ly
ans o med in o de o make hem appea wi hin he plo ed ame. No e he o e lap be ween he hue-spaces occupied by he wo hos s (B–E). The
ci cles in B and D ha e di e en colo s o illus a e he di ec ion o selec ion on hue unde he assump ion ha cuckoo egg colo dis ibu ions a e
gi en by he i e and 72 cuckoo eggs in Figu e F, espec i ely. Di e en colo s indica e di e ences in he a e age colo dis ance (DT
C
) be ween hos
and cuckoo eggs. Pa asi ism clea ly imposes dis up i e selec ion, and an e en ual e olu iona y esponse is he e o e p edic ed o change he
dis ibu ion owa ds s onge bimodali y, whe eupon i would con e ge wi h he cha inch’s hue dis ibu ion (C, E).
doi:10.1371/jou nal.pone.0019288.g004
Egg Colou Coe olu ion
PLoS ONE | www.plosone.o g 9 Ap il 2011 | Volume 6 | Issue 4 | e19288