Higher mobility of butterflies than moths connected to habitat suitability and body size in a release experiment

Highe mobili y o bu e lies han mo hs connec ed o habi a sui abili y and body size in a elease expe imen Mikko Kuussaa i 1 , Ma ias Saa inen 2 , Ee a-Liisa Ko pela 1 , Juha P€ oy y 1 & Te ho Hy € onen 3 1 Finnish En i onmen Ins i u e, Na u al En i onmen Cen e, P.O. Box 140, FI-00251 Helsinki, Finland 2 Cas  eninka u 10 b 17, FI-00530 Helsinki, Finland 3 MTT Ag i ood Resea ch Finland, Plan P oduc ion Resea ch, FI-31600 Jokioinen, Finland Keywo ds Animal mo emen , dispe sal p opensi y, expe imen al s udy on mig a ion, habi a p e e ence, in e speci ic di e ences in mobili y, ma k- elease- ecap u e s udy, elease habi a sui abili y, species ai s, a ia ion in dispe sal abili y, wingspan. Co espondence Mikko Kuussaa i, Finnish En i onmen Ins i u e, Na u al En i onmen Cen e, P.O. Box 140, FI-00251 Helsinki, Finland. Tel: +358 40 5256 249, Fax: +358 20 490 2290; E-mail: [email p o ec ed] Funding In o ma ion This s udy was unded by he Finnish Minis y o Ag icul u e and Fo es y ia he ECONTOOLS p ojec and by he Maj and To Nessling Founda ion. Recei ed: 13 Ma ch 2014; Re ised: 1 July 2014; Accep ed: 13 July 2014 Ecology and E olu ion 2014; 4(19): 3800– 3811 doi: 10.1002/ece3.1187 Abs ac Mobili y is a key ac o de e mining lepidop e an species esponses o en i on- men al change. Howe e , di ec mul ispecies compa isons o mobili y a e a e and empi ical compa isons be ween bu e lies and mo hs ha e no been p e i- ously conduc ed. He e, we compa ed mobili y be ween bu e lies and diu nal mo hs and s udied species ai s a ec ing bu e ly mobili y. We expe imen ally ma ked and eleased 2011 bu e ly and 2367 mo h indi iduals belonging o 32 and 28 species, espec i ely, in a 25 m 925 m elease a ea wi hin an 11-ha, 8- yea -old se -aside ield. Dis ance mo ed and emig a ion a e om he elease habi a we e eco ded by species. The elease expe imen p oduced di ec ly compa able mobili y da a in 18 bu e ly and 9 mo h species wi h almos 500 indi iduals ecap u ed. Bu e lies we e ound mo e mobile han geome oid mo hs in e ms o bo h dis ance mo ed (mean 315 m s. 63 m, espec i ely) and emig a ion a e (mean 54% s. 17%, espec i ely). Release habi a sui abil- i y had a s ong e ec on emig a ion a e and dis ance mo ed, because bu e - lies ended o lea e he se -aside, i i was no sui able o b eeding. In addi ion, emig a ion a e and dis ance mo ed inc eased signi ican ly wi h inc easing body size. When phylogene ic ela edness among species was included in he analyses, he signi ican e ec o body size disappea ed, bu habi a sui abili y emained signi ican o dis ance mo ed. The highe mobili y o bu e lies han geome oid mo hs can la gely be explained by mo phological di e ences, as bu e lies a e mo e obus lie s. The impo an ole o elease habi a sui abili y in bu e ly mobili y was expec ed, bu seems no o ha e been empi ically documen ed be o e. The obse ed posi i e co ela ion be ween bu e ly size and mobili y is in ag eemen wi h ou p e ious indings on bu - e ly coloniza ion speed in a long- e m se -aside expe imen and ecen me a- analyses on bu e ly mobili y. In oduc ion Dispe sal abili y is a key ac o a ec ing occu ence pa - e ns and popula ion ends in animals (Ewe s and Did- ham 2006). Ongoing changes in land use and clima e also pose s ong selec i e p essu es on species ai s ha a e connec ed o animal mobili y (Bon e e al. 2012; Bague e e al. 2013). An inc eased need o unde s and he impac s o en i onmen al change a popula ion and communi y le els has ecen ly a ac ed much in e es in he measu emen o mobili y di e ences ac oss indi iduals, popula ions, and species (Bowle and Ben on 2005; Clo- be e al. 2012). Howe e , despi e he accumula ing expe ience in es ima ing mobili y (Na han e al. 2008), p oducing eliable mul ispecies compa isons has emained a challenging ask. He e, we used bu e lies and mo hs o a mul ispecies mobili y compa ison o examine di e - ences in dispe sal abili y among species and be ween spe- cies g oups. Bu e lies a e one o he mos popula g oups in animal mobili y esea ch (S e ens e al. 2010), whe eas knowledge on o he insec g oups, e en among Lepidop e a, has emained scan y. 3800 ª2014 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. Se e al p e ious s udies on bu e lies ha e demon- s a ed he impo an ole o in e speci ic mobili y di e - ences in species dis ibu ions and species esponses o habi a and clima e change. Fo example, he e ec s o habi a agmen a ion ha e been shown o di e be ween bu e ly species wi h a ying mobili y ( € Ockinge e al. 2009, 2010). € Ockinge e al. (2010), using body size as a p oxy o mobili y, showed ha bu e ly species wi h low mobili y ha e been mos s ongly a ec ed by habi a loss and o he s udies ha e epo ed simila esul s. Ko iaho e al. (2005) ound ha h ea ened bu e ly species a e cha ac e ized by low mobili y, and he me a-analysis by Thomas e al. (2011) showed ha dispe sal abili y is one o he main d i e s o long- e m bu e ly popula ion ends. These esul s indica e ha dispe sal abili y may c ucially a ec how species can cope wi h global h ea s such as clima e change and habi a loss and agmen a ion. Mo eo e , ecen s udies ha e highligh ed he impo - ance o in aspeci ic a ia ion in mobili y and ha ela- i ely as mic oe olu iona y changes in dispe sal abili y and emig a ion p opensi y may play a signi ican ole when species a e adap ing o changing en i onmen s (Me ckx e al. 2003; Sch ickzelle e al. 2006; Duplouy e al. 2013). Fas e olu iona y changes may in luence eco- logical popula ion dynamics and ice e sa, po en ially causing complex eco-e olu iona y dynamics in dispe sal (Hanski and Mononen 2011). Howe e , he la ge numbe o ac o s in luencing e olu ion o dispe sal complica es p edic ions on wha would be he op imal dispe sal s a - egy in di e en landscapes and in case o di e en popu- la ion s uc u es (Clobe e al. 2012). Bu e ly mobili y has been empi ically s udied using a numbe o di e en app oaches (S e ens e al. 2010; Seka 2012). The mos popula app oach has been o conduc ma k- elease- ecap u e (MRR) s udies in na u al bu e ly (me a)popula ions (Ho es ad and Nieminen 2009). How- e e , mobili y es ima es om di e en single-species MRR s udies a e no di ec ly compa able, because he esul s a e s ongly dependen on he spa ial scale (Schneide 2003; F anz en and Nilsson 2007) and landscape s uc u e (Mennechez e al. 2003; Do e and Se ele 2009) o di e - en s udies. Manipula i e expe imen al app oaches ha e enabled o answe mo e speci ied ques ions conce ning di e en componen s o bu e ly mobili y and o ca y ou in a- and in e speci ic compa isons. Howe e , expe i- men al eleases o bu e lies in he ield (S€ ode s € om and Hedblom 2007; Kallioniemi e al. 2014) and s udies con- duc ed in la ge habi a cages (No be g e al. 2002; Hanski e al. 2006) ha e been ela i ely es ic ed in spa ial scale and ha e a ely in ol ed mo e han wo species. Because o he g ea demand o compa able mobili y es ima es in communi y ecological s udies, he e is an ob ious need o empi ical s udies p oducing compa able mobili y es ima es o a la ge numbe o species simul a- neously and in s anda dized condi ions. We p oduced such es ima es by expe imen ally eleasing a la ge numbe o ma ked indi iduals o 60 bu e ly and diu nal mo h species in a la ge se -aside ield and hen collec ing ecap- u es wi hin he s udy landscape. Ou aim was o collec a su icien amoun o compa able da a in o de o ana- lyse in e speci ic di e ences in mobili y and es ou hypo heses on he e ec s o speci ic species ai s on bu - e ly mobili y based on ea lie s udies. Mo e speci ically, we aimed o answe he ollowing s udy ques ions: (1) Do bu e lies di e signi ican ly om geome oid and noc u- oid mo hs in mobili y? (2) Does body size (wingspan) explain mobili y di e ences be ween bu e ly species? (3) Which o he species ai s a ec mobili y di e ences be ween bu e ly species? Based on p e ious s udies on mo hs (Nieminen 1996; Nieminen e al. 1999), we hypo hesized geome oids o be less mobile han noc uoids. Ou expec a ion o he ela ionship be ween bu e ly and mo h mobili y was less clea , because much a ia ion has been epo ed in bo h species g oups and di ec mul ispecies compa isons be ween bu e lies and mo hs ha e been lacking. How- e e , ou ea lie esul s o a six-yea se -aside expe imen showed ha bu e lies colonized he se -aside as e han diu nal mo hs (Alanen e al. 2011), sugges ing highe mobili y in bu e lies han mo hs. Based on ecen me a-analyses on bu e ly mobili y (S e ens e al. 2010, 2012; Seka 2012) and ou own esul s on coloniza ion speed in bu e lies (Alanen e al. 2011), we hypo hesized mobili y o inc ease wi h inc easing body size (wingspan). The mo i a ion o es he ole o a se o o he species ai s s ems om ecen s udies epo ing sig- ni ican e ec s o a ious ai s on bu e ly mobili y (S e- ens e al. 2010, 2012; Seka 2012). Fu he mo e, we used he oppo uni y o e ed by ou expe imen al se -up o es also he e ec o elease habi a sui abili y on mobili y o species o igina ing om di e en habi a ypes, hypo he- sizing ha dec easing habi a sui abili y would inc ease emig a ion a e (Bowle and Ben on 2005). Finally, we also conside ed he po en ial e ec s o phylogene ic ela edness on bu e ly mobili y. Cha ac e is ics o closely ela ed spe- cies a e o en mo e simila compa ed wi h dis an ly ela ed species, and hus he assump ion o independen da a poin s may be iola ed in compa a i e analyses including mul iple species (I es and Zhu 2006). Ma e ials and Me hods Expe imen al design and s udy a ea The expe imen had a simple design in which ma ked lepidop e an indi iduals we e eleased daily in a ª2014 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 3801 M. Kuussaa i e al.Highe Mobili y o Bu e lies han Mo hs 25 m 925 m elease a ea wi hin a 11-ha se -aside ield, which was es ablished eigh yea s ea lie (Fig. 1; o he o me six-yea se -aside expe imen , see Alanen e al. 2011). Mo emen dis ances o he ma ked indi iduals we e hen sys ema ically eco ded by ecap u ing hem a di e en dis ances om he elease a ea bo h wi hin and ou side he se -aside ield (Fig. 1). This design enabled us o eco d dis ance mo ed and emig a ion a e in a compa able manne o a la ge se o bu e ly and mo h species han o ou knowledge in any p e ious s udy. The elease se -aside ield was loca ed in Yp€ aj€ a, sou h- wes e n Finland (ETRS-TM35FIN N 6745551 E 299807), in an ag icul u al landscape domina ed by sp ing ce eal p oduc ion. The landscape su ounding he se -aside ield was la and open ag icul u al land in all di ec ions excep owa d he no hwes , whe e he e was a mosaic a ea o o es s, species- ich semi-na u al g asslands, and buil -up a eas s a ing om c. 600 m om he se -aside (Fig. 1). The elease se -aside was occupied by a ela i ely di e se communi y o g assland bu e lies and diu nal mo hs, wi h many species e en mo e abundan a he ime o ou elease expe imen han in yea 2008, when he six- yea se -aside expe imen ended (see Table S1). Fo ins ance, Lycaena hippo hoe had clea ly es ablished a local popula ion on he se -aside a e yea 2008. Bu e ly and mo h eleases A o al o 2011 bu e ly and 2367 mo h indi iduals belonging o 32 bu e ly and 28 mo h species we e ma ked and eleased in he 25 m 925 m elease a ea wi hin he se -aside ield ( o a de ailed lis o eleased species, see Table S1; nomencla u e acco ding o Kull- be g e al. 2002). Indi iduals o he eleases we e collec ed om he se -aside ield (40% o eleased indi- iduals) as well as om he su ounding landscape (nine si es, 60% o indi iduals). The nine si es we e loca ed 50–3800 m om he elease se -aside ield and we e good bu e ly habi a s, mos ly pa ches o semi-na u al g asslands and shel e ed, sunny o es edges wi h some semi-na u al ege a ion. These si es we e selec ed in o de o maximize bo h he numbe o indi iduals and species eleased in he expe imen . Collec ing (unma ked) indi iduals om hese si es o he eleases also e ec i ely se ed in collec ing ecap u es o ma ked indi iduals ha had al eady emig a ed om he elease se -aside ield (see below). Bu e lies and diu nal mo hs we e ma ked, eleased, and ecap u ed daily du ing wo s udy pe iods: om 30 May o 11 June and om 28 June o 14 July 2011. The i s pe iod co e ed he ligh season o ea ly summe spe- cies in sou hwes e n Finland, whe eas he second pe iod co e ed he ligh season o mid-summe species. This p ocedu e enabled us o co e a la ge p opo ion o bu - e ly and diu nal mo h species’ occu ence du ing he summe season. The wea he was mos ly wa m and sunny (i.e., a o able o lepidop e an ac i i y) du ing he wo s udy pe iods. Usually, indi iduals we e collec ed o he eleases om he elease se -aside ield du ing he mo ning and om he su ounding landscape du ing he a e noon. Figu e 1. Ae ial pho og aph o he s udy a ea. Le e A indica es he elease a ea wi hin he ocal se -aside ield. The black line wi h a ows indica es he 2500-m-long ansec in which ma ked indi iduals we e sys ema ically sea ched. Solid whi e lines show he sea ching ou es ou side he elease se -aside, and dashed whi e lines show he ou es which we e walked less equen ly. Numbe s 1–4 indica e a o able bu e ly and mo h habi a s, which we e used bo h o collec ing indi iduals o he eleases and o sea ching ecap u es o emig a ed indi iduals; especially si es 1 (abandoned a mya d and a shel e ed o es edge) and 2 (semi-na u al g assland pa ch) a ac ed many emig an s. 3802 ª2014 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Highe Mobili y o Bu e lies han Mo hs M. Kuussaa i e al. Bu e lies we e always ma ked wi h an indi idual numbe on he wing using a ine-poin pen, whe eas o he lepi- dop e an species we e ma ked wi h a colo spo made on he wing wi h a hicke ma ke pen. The la e was pe - o med by gen ly p essing he pen h ough he bu e ly ne wi hou aking he mo h indi idual in hand, in o de o a oid damaging i s agile wings. Immedia ely a e ma king, each indi idual was placed indi idually wi hin a 120-ml plas ic con aine which was hen s o ed in a cool box in o de o keep he ma ked indi iduals inac i e be o e he elease. Indi iduals ma ked wi hin he elease se -aside du ing he mo ning session we e eleased close o he cen e o he 25 m 925 m elease a ea daily app oxima ely a 12 o’clock, whe eas he ma ked indi iduals collec ed om he su ounding landscape we e ypically eleased be ween 16 and 18 o’clock. In he elease a ea, he bu e lies and mo hs we e gen ly placed indi idually on plan lea es and lowe s. Recap u es we e ne e collec ed wi hin he 25 m 925 m elease a ea. P o ocol o ecap u es In collec ing da a on mo emen s o eleased bu e lies and mo hs, he ocus was on bo h wi hin se -aside mo e- men s and mo emen s o he su ounding landscape. The e o e, ecap u es we e sea ched daily in a sys ema ic way a di e en dis ances om he elease a ea, bo h in he elease se -aside and in i s su oundings. App oxima ely one hou was spen on collec ing ecap- u es wi hin less han 100 me e s om he elease a ea e e y mo ning. Such a high e o was di ec ed on he el- a i ely close icini y o he elease a ea in o de o asce - ain a leas some ecap u es om as many eleased species as possible, including he leas mobile species. In addi ion, he whole elease se -aside ield was sys ema i- cally sea ched h ough by walking a 2500-m-long cons an ansec (Fig. 1) e e y day. App oxima ely simila e o was di ec ed on ga he ing ecap u es o emig a ed indi- iduals in he su ounding landscape. Fig. 1 shows he ou es along ield ma gins and oad e ges in he icini y o he elease se -aside ield in which ecap u es we e sea ched o as o en as ime allowed (almos daily). In addi ion, ou a o able bu e ly and mo h habi a s (numbe s 1–4 in Fig. 1) u ned ou o a ac many emi- g a ed indi iduals, and he e o e, hese a eas we e isi ed almos daily. In summa y, an a ea o c. 1km 2 was well- su eyed daily, whe eas in o al ecap u es we e collec ed om an a ea o c. 4km 2 in size. Fo each ecap u ed indi idual, he ollowing in o ma- ion was eco ded: da e, ime, species, sex, indi idual numbe ( o bu e lies), and he exac loca ion o he ecap u e, ma ked on an ae ial pho og aph o he a ea. Measu emen o mo emen pa ame e s Two main measu es o mobili y we e eco ded o each species wi h ecap u es: a e age dis ance mo ed and emig a ion a e. Dis ance mo ed was measu ed o each ecap u ed bu - e ly indi idual as he dis ance be ween he elease poin and he loca ion o he las ecap u e, he eby each indi- idual con ibu ed o he esul s only once. Fo diu nal mo hs, which we e no ma ked indi idually, he dis ance om he elease poin was eco ded o e e y ecap u e poin . Dis ances mo ed we e measu ed om he ae ial pho og aphs in which he ecap u e poin s we e ma ked in he ield. Fo he s a is ical analyses, dis ances mo ed we e ln- ans o med a e which hey ollowed a no mal dis ibu ion. An indi idual was conside ed as emig a ed, i i was ecap u ed ou side he elease se -aside ield. Based on he same logic as wi h dis ance mo ed, only he las ecap u e o a bu e ly indi idual was used o indi- ca ing emig a ion, i espec i e o i s p e ious ecap u e eco ds. In con as , all ecap u es o mo hs we e consid- e ed as independen obse a ions. As a hi d measu e po en ially ela ed o mobili y, he p opo ion o ecap u ed indi iduals was eco ded o all s udied species. In p e ious s udies on lepidop e an mobili y, inc easing ac ion o disappea ed (i.e., no ecap u ed) indi iduals has some imes been conside ed as an indica ion o inc easing mobili y o emig a ion (Kuussaa i e al. 1996; Me ckx e al. 2009). In con as o he o he wo mobili y measu es which a e solely based on ecap u es, all eleased lepidop e an indi iduals con- ibu ed o his measu e and hus he ac ion o ecap- u ed indi iduals could po en ially gi e some addi ional in o ma ion on mobili y. In o de o acili a e an unbiased compa ison o mobil- i y be ween bu e lies and mo hs in s a is ical analyses, we also calcula ed all h ee mobili y measu es o bu e - lies using he same logic as in mo hs, ha is, ea ing each bu e ly ecap u e as a sepa a e da a poin in he da a se . Species ai s The analyses on he ole o species ai s ocused only on bu e lies as published species ai da a a e scan y o mo hs. The ollowing six species ai s we e examined in o de o explain obse ed mobili y di e ences in bu e - lies: body size, adul habi a speci ici y and p e e ence, la al hos plan speci ici y and hos plan ype, and elease habi a sui abili y. Body size was measu ed as a con inuous a iable, whe eas all he o he species ai s we e measu ed as ca ego ical a iables. The ai classi ica- ions o each s udied species a e shown in Table S2. ª2014 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 3803 M. Kuussaa i e al.Highe Mobili y o Bu e lies han Mo hs Body size o each species was measu ed as he a e age emale wingspan (in mm), based on he Finnish bu e ly handbook by Ma ila e al. (1990). Adul habi a speci ic- i y was classi ied as a bina y a iable: habi a specialis s occupying one o wo and gene alis s occupying mo e han wo habi a ypes ollowing Ek oos e al. (2010) and o igi- nally based on Komonen e al. (2004). Habi a p e e ence had h ee classes: o es edges and clea ings, semi-na u al g asslands, and ield ma gins in open a mland, ollowing Kuussaa i e al. (2007). The speci ici y o la al hos plan use was measu ed as a bina y a iable: mono- and oligoph- agous species eeding only on one hos plan genus and polyphagous species eeding on mo e han one plan genus, based on Komonen e al. (2004). La al hos plan ype was classi ied o he ollowing ou ca ego ies: woody plan s (i.e., ees and sh ubs as well as species in he amily E icaeae), g asses (Poaceae), leguminous plan s (Fabaceae), and o he he bs, based on Alanen e al. (2011). Habi a sui abili y o he elease se -aside ield was a a iable cons uc ed speci ically o ou cu en analyses. I was based on ex ensi e quan i a i e obse a ions on he na u al occu ence o he s udied bu e ly species in he elease se -aside ield, as explained in Table S1. All he species eleased in ou mobili y expe imen we e classi ied in o h ee g oups: 1 =species ne e eco ded, 2 =species wi h 1–5 eco ds, and 3 =species wi h >5 eco ds du ing yea s 2003–2011. Class 3 ep esen s species o which he se -aside ield was mos sui able as a b eeding habi a . This measu e o habi a sui abili y was conside ed as an empi ically well-jus i ied and o ou pu poses mo e accu- a e measu e o species habi a p e e ence han he p e i- ously published classi ica ion, p esen ed abo e. S a is ical analyses The i s se o s a is ical analyses ocused on mobili y di - e ences be ween wo phylogene ically delinea ed species g oups, bu e lies (Papilionoidea) and geome oid (Geo- me oidea) mo hs ( an Nieuke ken e al. 2011), using compa ably calcula ed mobili y a iables as explained abo e. Noc uoid (Noc uoidea) mo hs we e excluded om hese analyses, as he e we e only a ew ecap u es (mo e han one indi idual ecap u ed only in one species; Table 1). Di e ences in mean dis ance mo ed be ween he spe- cies g oups we e es ed using linea mixed models (LMM) using species g oup as a ca ego ical ixed ac o . Species was included in he model as a andom ac o in o de o ake in o accoun he nonindependence o obse - a ions om di e en indi iduals o he same species. Model i ing was conduc ed using es ic ed maximum- likelihood (REML) es ima ion wi h he deg ees o ee- dom calcula ed acco ding o he Kenwa d–Roge me hod (Bolke e al. 2009). Di e ences in emig a ion a e and ecap u e p obabili y be ween he h ee species g oups we e es ed using he same logic, bu by i ing gene al- ized mixed models (GLMM) wi h logis ic link unc ion and binomial e o dis ibu ion (due o bina y esponse a iables). GLMM i ing was conduc ed using adap i e Gauss–He mi e quad a u e es ima ion (Bolke e al. 2009) wi h he deg ees o eedom calcula ed wi h he be ween– wi hin deg ees o eedom app oxima ion. Fo all h ee esponse a iables, he pai wise di e ences be ween he h ee species g oups we e es ed using Tukey’s es . As he second s ep o analyses, mul i a ia e models we e buil o examine which combina ions o species ai s bes explained mobili y di e ences be ween bu e ly species. He e, only he las ecap u e o each bu e ly indi idual was aken in o accoun . Also, he sex o each indi idual was included in hese models, because he mo i a ion o he wo sexes o mo e and emig a e may be qui e di e en . Howe e , be o e mul i a ia e model building, he uni a ia e ela ionships be ween each species ai , sex and he h ee mobili y measu es we e examined by building a sepa a e s a is ical model o each species ai and mobili y measu e (Appendix S1). Pai wise ela- ionships be ween he explana o y species ai s we e examined be o e model building in o de o a oid inclu- sion o collinea explana o y a iables. Consequen ly, wo species ai s (la al hos plan ype and habi a p e e - ence) we e omi ed om mul i a ia e model building, due o signi ican ela ionships wi h o he ai s (Appen- dix S1). Mo eo e , he po en ial e ec o he o iginal col- lec ion a ea ( om he se -aside ield o om su ounding landscape) o he eleased bu e ly indi iduals on mobil- i y was es ed, and i did no a ec emig a ion a e o dis- ances mo ed (Appendix S1). Thus, he ole o he sou ce a ea could be igno ed in he analyses. Fo wa d selec ion was used in building he LMM and GLMM wi h mul iple a iables, ha is, he s a is ically signi ican a iables (P<0.05) we e en e ed in o he model in he o de o hei explana o y powe . Fo he only con inuous a iable, body size, bo h linea and qua- d a ic e ec s we e es ed. S a is ical signi icances we e cal- cula ed using an F- es . No o e dispe sion was obse ed in he analyses. Pai wise di e ences be ween he ca ego ies o he ca ego ical species ai s we e es ed using Tukey’s es . All LMM and GLMM models desc ibed abo e we e buil using he s a is ical package SAS/STAT  9.2 (SAS ins i u e Inc., Ca y, NC). In o de o ake in o accoun he po en ial e ec s o phylogene ic ela edness on bu e ly mobili y in ou s udy, he inal mul i a ia e models o dis ance mo ed and emig a ion a e we e e i ed using gene alized es ima ion equa ions (GEE) as implemen ed in he ape lib a y, e sion 3.0.11 (Pa adis e al. 2004) in he R 3804 ª2014 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Highe Mobili y o Bu e lies han Mo hs M. Kuussaa i e al. s a is ical en i onmen (R Co e Team 2013). GEE a e ex ensions o gene alized linea models (GLMs) o be applied when he s a is ical nonindependence o he da a can be de e mined wi h a co ela ion ma ix (Pa adis and Claude 2002). Pa adis and Claude (2002) ha e demon- s a ed he applicabili y o GEE in compa a i e s udies using a be ween-species co ela ion ma ix de i ed om a phylogene ic ee, and P€ oy y e al. (2009) p o ide a p e i- ous example on bu e lies. GEE a e especially sui able o da a ha include ca ego ical a iables (Pa adis and Claude 2002), as was he case in ou s udy. To calcula e a co ela ion ma ix o ela edness in GEE, a phylogene ic hypo hesis was de i ed o he 32 bu e ly species included in ou s udy (Appendix S2). The b anching sequences o bu e ly amilies we e de i ed om ecen amily-le el phylogene ic s udies co - e ing all highe axa o bu e lies (e.g., Heikkil€ a e al. 2011). Placemen o lowe axa down o indi idual species was deduced om he phylogene ic s udies ocusing spe- ci ically on each g oup (Appendix S2). B anches wi h weak suppo o un esol ed b anches in he o iginal s ud- ies we e ea ed as poly omies. Fo simplici y, all ee b anches we e assumed o be o equal leng h. In o de o include indi iduals in he analysis, we placed hem on species b anches so ha be ween-indi idual dis ances we e assumed o be 0.01 x species b anch leng h. S a is ical sig- ni icances we e calcula ed using an F- es , and he phylo- gene ic hypo hesis was used o calcula e he co ec ed Table 1. Mobili y esul s o all ecap u ed species: Numbe o eleased indi iduals (n), numbe o ecap u ed indi iduals (RC ind ), ecap u e p oba- bili y (%; RC % ), emig a ion p obabili y (%; Emig % ), mean dis ance mo ed s anda d e o (m; D mean SE), and maximum dis ance mo ed (m; D max ). Fo each bu e ly species, he alues in pa en heses indica e he o al numbe o ecap u es and es ima es o emig a ion a e and mean dis ance mo ed, based on all ecap u es and calcula ed simila ly as in diu nal mo hs. Species nRC ind RC % Emig % D mean SE D max Bu e lies An hocha is ca damines 22 2 (3) 9.1 100 (100) 985 565 (779 386) 1550 Aphan opus hype an us 188 79 (110) 42.0 14 (10) 113 11 (105 8) 510 A icia a axe xes 24 5 (6) 20.8 40 (33) 347 119 (310103) 730 Bolo ia euph osyne 21 4 (5) 19.0 100 (100) 619 96 (586 81) 893 Bolo ia selene 236 40 (45) 16.9 45 (33) 250 43 (237 38) 885 B en his ino 92 35 (56) 38.0 51 (41) 147 19 (138 13) 520 Coenonympha glyce ion 161 25 (34) 15.5 16 (15) 138 21 (132 16) 539 Gonep e yx hamni 44 1 2.3 100 878 878 Lep idea sinapis 56 9 (15) 16.1 100 (100) 488 18 (491 12) 548 Lycaena hippo hoe 33 13 (17) 39.4 8 (12) 84 13 (83 10) 196 Lycaena i gau eae 14 6 (8) 42.9 83 (75) 460 71 (430 67) 550 Meli aea a halia 21 3 14.3 67 339 241 817 Nymphalis io 26 2 (3) 7.7 100 (100) 290 129 (23592) 419 Pie is napi 480 47 (55) 9.8 77 (78) 396 40 (39335) 1720 Polyomma us amandus 253 79 (127) 31.2 19 (20) 119 12 (12010) 520 Polyomma us ica us 25 5 (6) 20.0 40 (33) 191 84 (170 72) 510 Polyomma us semia gus 72 15 (17) 20.8 20 (24) 142 34 (13730) 520 Thymelicus lineola 97 15 (17) 15.5 27 (29) 106 13 (112 13) 214 To al 2011 1 385 (528) 22.2 2 55.9 (53.9) 2 338 (315) 2 Noc uoid mo hs Callis ege mi 19 1 5.3 0 34 34 C yp ocala cha dinyi 16 1 6.3 100 128 128 Euclidia glyphica 594 40 6.7 10 122 23 913 To al 673 1 42 6.1 2 36.7 2 95 2 Geome oid mo hs Chiasmia cla h a a 930 41 4.4 0 47 4 130 Ema u ga a oma ia 192 8 4.2 0 64 10 115 Odezia a a a 39 1 2.6 100 120 120 Sco op e yx chenopodia a 348 11 3.2 0 49 9 107 Scopula immo a a 38 2 5.3 0 51 151 Siona linea a 13 2 15.4 0 44 17 61 To al 1694 1 65 5.9 2 16.7 2 63 2 To al all 4378 1 492 16.1 2 45.1 2 250 2 1 Including also species wi h no ecap u es in he da a. 2 Unweigh ed mean o he ecap u ed species. ª2014 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 3805 M. Kuussaa i e al.Highe Mobili y o Bu e lies han Mo hs deg ees o eedom o he da a. Fo ecap u e p obabili y, he models did no con e ge using he GEE app oach. Resul s A o al o 385 indi iduals o 18 species o bu e lies and 107 indi iduals o 9 species o mo hs (6 geome oids and 3 noc uoids) we e ecap u ed wi hin he elease se -aside ield (328 indi iduals) and in i s su oundings (164 indi- iduals). Table 1 summa izes in o ma ion on he eleased and ecap u ed indi iduals and hei mobili y o all spe- cies wi h a leas one ecap u e ( o in o ma ion on all eleased species, see Table S1). Di e ences be ween bu e lies and mo hs The wo compa ed species g oups, bu e lies and geome - oid mo hs, di e ed signi ican ly in all h ee examined measu es o mobili y (Table 2, Fig. 2). Bu e lies we e mo e mobile han geome oids as indica ed by hei longe mean dis ances mo ed (315 m s. 63 m) and highe emig a ion a e (54% s. 17%). The highe ecap- u e a e o bu e lies han geome oids (22% s. 6%) mos p obably e lec ed he be e de ec abili y in bu e - lies han geome oids. The mobili y o noc uoid mo hs seemed o be somewhe e be ween bu e lies and geome - oids (Table 1), bu he noc uoid da a we e oo limi ed o allow meaning ul s a is ical analyses. Bu e ly mo emen s in ela ion o species ai s The s udied bu e ly species showed a lo o in e speci ic a ia ion in mobili y. A e age dis ance mo ed a ied om 84 m and 106 m in he wo mos seden a y species (Lycaena hippo hoe and Thymelicus lineola, espec i ely) o 619 m and 985 m in he wo mos mobile species (Bolo- ia euph osyne and An hocha is ca damines, espec i ely). Emig a ion om he elease se -aside ield a ied om 8% (L. hippo hoe) and 14% (Aphan opus hype an us) o 100% in i e o he s udied species (Table 1). Two species ai s, elease habi a sui abili y and body size, became included oge he in he mul i a ia e models bes explaining he wo main mobili y a iables, dis ance mo ed (LMM) and emig a ion a e (GLMM) (Table 3A). The e ec s o he wo ai s we e e y simila in bo h models. Dis ance mo ed and emig a ion a e we e lowe in bu e ly species o which habi a sui abili y was he highes . Fu he mo e, bo h dis ance mo ed and emig a- ion a e ended o inc ease wi h inc easing body size, when he e ec o habi a sui abili y was aken in o accoun (see also Fig. 3A and C). Fo he hi d mobili y a iable, ecap u e a e, body size, and sex we e he wo a iables included oge he in he mul i a ia e GLMM (Table 3A). The e ec o body size became signi ican only when i s nonlinea compo- nen was included in he model. Recap u e a e was high- es in bu e ly species o in e media e size and pa icula ly low in he la ges species eleased in he Table 2. LMM and GLMM esul s on he di e ences in he h ee mobili y a iables be ween bu e lies and geome oid mo hs. The di - e ences be ween he species g oups emained signi ican in all h ee a iables when he models we e e i ed o he subse o species o which he elease se -aside p o ided sui able habi a ( elease habi a sui abili y class =3). Response a iable Model n d (nume a o : denomina o ) FP Dis ance mo ed LMM 593 1:23.2 14.71 0.0008 Emig a ion a e GLMM 593 2:22 8.22 0.0090 Recap u e a e GLMM 3848 2:52 16.69 0.0002 (A) (B) (C) Figu e 2. Di e ences in (A) mean dis ance mo ed, (B) emig a ion a e, and (C) ecap u e a e be ween bu e lies and geome oid mo hs. Means a e leas squa es means (LSM) wi h 95% con idence in e als based on he s a is ical models i ed o collec ed da a (Table 2). The as e isks indica e he s a is ical di e ence be ween he species g oups (**P<0.01, ***P<0.001). 3806 ª2014 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Highe Mobili y o Bu e lies han Mo hs M. Kuussaa i e al. expe imen . The signi ican e ec o sex was due o he highe ecap u e a e o males han emales. When he inal LMM and GLMM models we e e i ed using gene alized es ima ion equa ions (GEE) in o de o ake in o accoun he po en ial e ec s o phylogene ic ela edness, he esul s changed sligh ly (Table 3B). In he GEE model o dis ance mo ed, he e ec o body size did no emain signi ican (P=0.14), bu habi a sui abil- i y s ill had a signi ican e ec . In he GEE model o emig a ion a e, bo h habi a sui abili y and body size had a signi ican e ec . Discussion The elease expe imen success ully p oduced di ec ly com- pa able mobili y da a o bu e lies and mo hs. Recap u es we e collec ed om almos 500 indi iduals belonging o 27 species. The da a se enabled us bo h o de ec di e ences in mobili y be ween wo lepidop e an supe amilies and o iden i y signi ican e ec s o species ai s on dis ance mo ed and emig a ion a e in 18 species o bu e lies. Di e ences be ween bu e lies and mo hs As expec ed, expe imen ally eleased bu e lies we e mo e mobile han hin-bodied, weakly lying geome oid mo hs in e ms o bo h dis ance mo ed and emig a ion a e. Bu e ly mo emen dis ances we e on he a e age i e imes longe and emig a ion a e h ee imes highe han in geome oid mo hs. Da a o noc uoid mo hs emained oo spa se o in e any gene al esul s. Ou indings a e in ag eemen wi h ou p e ious esul s on coloniza ion o se -asides by bu e lies and diu nal mo hs (Alanen e al. 2011) and an expe imen compa ing mobili y o lepidop e an species g oups (Nieminen 1996) in a ne wo k o small islands. Like ou esul s, he esul s o Nieminen also sugges ed ha bu e lies a e mos and hin-bodied geome oids leas mobile, whe eas noc uoids show in e media e mobili y. I should be no ed, howe e , ha Nieminen s udied only wo bu e ly species, Vanessa a alan a and Hippa chia semele, o which V. a alan a is known as a egula long-dis ance mig an , ep esen ing one o he mos mobile bu e ly species occu ing in Eu ope (S e anescu 2001). Recap u e a e was gene ally much lowe in diu nal mo hs han in bu e lies. We a gue ha he e a e wo likely easons o his: he lowe ligh ac i i y and hus he lowe de ec abili y o mo hs and he highe popula- ion densi ies o he mos abundan mo hs compa ed o he mos abundan bu e lies. Based on he obse ed el- a i e abundances o ma ked s. unma ked indi iduals in he elease se -aside ield, we es ima ed ha he geome - oid Semio hisa cla h a a and he noc uoid Euclidia glyph- ica, o ins ance, we e an o de o magni ude mo e abundan han he mos abundan bu e lies, such as Aphan opus hype an us and Polyomma us amandus. Ne - e heless, due o ou sys ema ic sampling p o ocol, he ela i e ecap u e p obabili ies o he s udied axonomic g oups did no di e a di e en dis ances om he elease poin , and hus he mobili y esul s can be eliably compa ed be ween di e en species and species g oups. Ou esul s indica e ha i is mo e di icul o ob ain eli- able mobili y da a om diu nal mo hs han bu e lies by ma k– elease– ecap u e me hod. In ligh o he heo e ical model by T a is and Dy- ham (1999), he obse ed pa e n o mobili y a ia ion ac oss mo h and bu e ly species in ou expe imen has po en ial consequences o species pe sis ence. Acco ding o hei p edic ions, species wi h ei he low o high Table 3. Resul s o he inal mul i a ia e models (LMM, GLMM, and GEE) o he h ee s udied mobili y a iables. (A) LMM and GLMM models wi hou accoun ing o he phylogene ic ela edness. (B) GEE models accoun ing o he phylogene ic ela edness among species Model 1 Es ima e SE d FP (A) Dis ance mo ed (LMM) Cons an 3.354 0.589 Hab sui 2:30.5 15.57 <0.001 Class 1 1.168 0.278 Class 2 1.316 0.325 Body size 0.042 0.017 1:27.8 6.21 0.019 Emig a ion a e (GLMM) Cons an 5.517 1.850 Hab sui 2:14 6.76 0.009 Class 1 3.334 1.195 Class 2 2.384 0.967 Body size 0.139 0.054 1:14 6.70 0.022 Recap u e a e (GLMM) Cons an 12.468 4.301 Body size 0.605 0.238 1:29 6.48 0.017 Body size*Body size 0.009 0.003 1:29 7.08 0.013 Sex 1:22 6.32 0.020 Male 0.372 0.148 (B) Dis ance mo ed (GEE) 2 Hab sui 3 0.729 0.186 2 29.43 0.011 Body size 0.039 0.020 1 3.95 0.141 Emig a ion a e (GEE) 2 Hab sui 3 1.315 0.739 2 13.99 0.030 Body size 0.141 0.028 1 25.19 0.015 Hab sui = elease habi a sui abili y (Class 1 =unsui able, Class 2= ai ly unsui able, Class 3 =sui able habi a o b eeding), Body size =wingspan (mm). 1 GEE Model o ecap u e p obabili y did no con e ge. 2 Phylogene ic deg ees o eedom: 7.00. 3 Habi a sui abili y (Hab sui ) was ea ed as an o de ed ac o in bo h models, and model es ima es o linea con as s a e p esen ed in he able. ª2014 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 3807 M. Kuussaa i e al.Highe Mobili y o Bu e lies han Mo hs dispe sal a e should pe o m bes in highly agmen ed landscapes, whe eas species wi h in e media e mobili y a e p edic ed o pe o m wo s . Ou indings seem o i hese p edic ions because he geome oid mo hs, ha we e ound o be he leas mobile lepidop e ans, ha e no declined in Finland (Hulden e al. 2000) and a e ypically common and abundan in many kinds o uncul i a ed g assland. Simila ly, la ge bu e ly species wi h high mobili y ha e no su e ed om habi a ag- men a ion, whe eas some g assland specialis bu e lies wi h in e media e mobili y, such as L. hippo hoe, ha e disappea ed om many in ensi ely cul i a ed landscapes (Ek oos and Kuussaa i 2012). This model p edic ion has p e iously ecei ed empi ical suppo om B i ish bu - e lies (Thomas 2000). Bu e ly mo emen s in ela ion o species ai s Bu e ly mobili y was s ongly a ec ed by habi a sui - abili y. Bu e lies ended o quickly emig a e om he elease se -aside ield, i i did no o e sui able b eeding habi a o he species in ques ion. Body size explained addi ional a ia ion in mobili y a e he e ec o habi a sui abili y had been aken in o accoun in he s a is ical models. Bo h dis ance mo ed and emig a ion a e inc eased wi h body size, as expec ed based on ou ea lie esul s on bu e ly coloniza ion speed (Alanen e al. 2011) and me a-analyses on bu e ly mobili y (S e ens e al. 2010, 2012; Seka 2012). When phylogene ic ela edness among species was included in he analyses, (A) (B) (C) (D) (E) (F) Figu e 3. S a is ically signi ican ela ionships be ween species ai s and he h ee mobili y a iables: (A–B) dis ance mo ed, (C–D) emig a ion a e, and (E–F) ecap u e a e in bu e lies. Means a e leas squa es means (LSM) wi h 95% con idence in e als based on he mul i a ia e models i ed o collec ed da a (Table 3). In he panels A, C, and E, he do s ep esen means o indi idual species. The le e s a and b wi hin he panels B, D, and F indica e homogeneous g oups and hus he ea men s which di e ed signi ican ly in pai wise compa isons. 3808 ª2014 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Highe Mobili y o Bu e lies han Mo hs M. Kuussaa i e al.