Differences in pollination success between local and foreign flower color phenotypes: a translocation experiment with Gentiana lutea (Gentianaceae)

Submi ed 11 July 2016 Accep ed 7 Decembe 2016 Published 7 Feb ua y 2017 Co esponding au ho s Ja ie A. Gui ián, [email p o ec ed] Ma ía Losada, [email p o ec ed] Academic edi o Dezene Hube Addi ional In o ma ion and Decla a ions can be ound on page 12 DOI 10.7717/pee j.2882 Copy igh 2017 Gui ián e al. Dis ibu ed unde C ea i e Commons CC-BY 4.0 OPEN ACCESS Di e ences in pollina ion success be ween local and o eign lowe colo pheno ypes: a ansloca ion expe imen wi h Gen iana lu ea (Gen ianaceae) Ja ie A. Gui ián1, Ma Sob al2, Tania Veiga1, Ma ía Losada2, Pablo Gui ián1 and José M. Gui ián2 1Depa amen o de Bo ánica/Facul ade de Bioloxía, Uni e sidade de San iago de Compos ela, San iago de Compos ela, A Co uña, Spain 2Depa amen o de Bioloxía Celula e Ecoloxía/Facul ade de Bioloxía, Uni e sidade de San iago de Compos ela, San iago de Compos ela, A Co uña, Spain ABSTRACT Backg ound. The adap i e main enance o lowe colo a ia ion is equen ly a - ibu ed o pollina o s pa ly because hey p e e en ially isi ce ain lowe pheno ypes. We es ed whe he Gen iana lu ea—which shows a lowe colo a ia ion ( om o ange o yellow) in he Can ab ian Moun ains ange (no h o Spain)—is locally adap ed o he pollina o communi y. Me hods. We ansplan ed o ange- lowe ing indi iduals o a popula ion wi h yellow- lowe ing indi iduals and ice e sa, in o de o assess whe he he e is a pollina ion ad an age in he local mo ph by compa ing i s isi a ion a e wi h he o eign mo ph. Resul s. Ou ecip ocal ansplan expe imen did no show clea local mo ph ad an age in o e all isi a ion a e: local o ange lowe s ecei ed mo e isi s han o eign yellow lowe s in he o ange popula ion, while bo h local and o eign lowe s ecei ed he same isi s in he yellow popula ion; hus, he e is no e idence o local adap a ion in Gen iana lu ea o he pollina o assemblage. Howe e , some lo al isi o g oups (such as Bombus p a o um,B. so oensis anca icus and B. lapida ius decipiens) consis en ly p e e ed he local mo ph o he o eign mo ph whe eas o he s (such as Bombus e es is) consis en ly p e e ed he o eign mo ph. Discussion. We concluded ha he e is no e idence o local adap a ion o he pollina o communi y in each o he wo G. lu ea popula ions s udied. The consequences o local adap a ion o pollina o on G. lu ea lowe colo would depend on he a ia ion along he Can ab ian Moun ains ange in mo ph equency and pollina o communi y composi ion. Subjec s Conse a ion Biology, Ecology, En omology, E olu iona y S udies, Plan Science Keywo ds Flowe colo a ia ion, Local adap a ion, Pollina o p e e ences, Recip ocal ansplan s, Gen iana lu ea INTRODUCTION Flo al e olu ion is p ima ily d i en by pollina o s (B adshaw & Schemske, 2003;Whi all & Hodges, 2007). The di e si y in lo al ai s ound in angiospe ms, such as lowe colo , could be ela ed wi h ansi ions in hese ai s (Raushe , 2008), as a esul o di e en How o ci e his a icle Gui ián e al. (2017), Di e ences in pollina ion success be ween local and o eign lowe colo pheno ypes: a ansloca ion expe imen wi h Gen iana lu ea (Gen ianaceae). Pee J 5:e2882; DOI 10.7717/pee j.2882 selec i e p essu es exe ed by pollina o s wi h s ong p e e ences o ce ain cha ac e is ics (Schemske & B adshaw, 1999;S eis eld & Kohn, 2007) o by changes in pollina o communi y composi ion (Ellis & Johnson, 2009). Thus, lo al di e si ica ion and specia ion among closely ela ed plan popula ions may esul om he isola ion and/o local adap a ion o he mos e icien pollina o s, which inc ease he numbe o isi s pe plan depending on hei lowe p e e ences (Ha de & Johnson, 2009; and e e ences he ein). The adap i e main enance o lowe colo a ia ion is equen ly a ibu ed o pollina o s, in pa because hey end o p omo e asso a i e ma ing, i.e., p e e en ially isi ing ce ain pheno ypes (Wase & P ice, 1981;S an on, 1987;Jones & Rei hel, 2001). Howe e , many o he easons could explain he e olu ion o lo al di e si ica ion, such as he po en ial uncoupling o he e olu ion o lo al ai s om pollina o -media ed selec ion (S auss & Whi all, 2006). Fo example, pleio opic e ec s on he biosyn he ic pa hways o lo al pigmen a ion may d i e he e olu ion o al e na i e lowe colo s (A mb us e , 1993; Schemske & Bie zychudek, 2007;Cooley, Ca allo & Willis, 2008). Many s udies suppo ha animal pollina o p e e ences may cause selec i e p essu es on lowe colo , which may luc ua e depending on he s uc u e and/o composi ion o he pollina o communi y (Melendez-Acke man & Campbell, 1998;Gómez & Zamo a, 2000). In his con ex , ecip ocal ansplan expe imen s be ween popula ions can be a powe ul ool o assess how changes in lowe colo p e e ences o he pollina o assemblage con ibu e o an adap i e ad an age o he local mo ph o e he o eign mo phs (S eis eld & Kohn, 2007). Gen iana lu ea L. shows lowe colo a ia ion along he Can ab ian Moun ains (No he n Spain). Gen iana lu ea lowe s a e ypically yellow (G. lu ea a . lu ea), hough no hwes Ibe ian popula ions ha e o ange- lowe ing indi iduals, which cons i u es a di e en a ie y o he species (G. lu ea a . au an iaca; Renobales, 2012). The pollina o assemblages a e made up mos ly o bumblebee species (Veiga e al., 2015;Sob al e al., 2015). P e ious s udies conclude ha : (i) Gen iana lu ea lowe colo a ia ion does no esul om adap a ion o en i onmen al ac o s, such as ele a ion, empe a u e, adia ion, and p ecipi a ion (Veiga e al., 2016); (ii) Gen iana lu ea is s ongly dependen on pollina o s o ep oduc ion, and he plan has a posi i e ela ionship be ween he numbe o pollina o isi s and he numbe o seeds i p oduces (Losada e al., 2015); (iii) he e is a pa ial hyb idiza ion ba ie among G. lu ea colo mo phs (Losada e al., 2015); (i ) Gen iana lu ea lowe colo in luences pollina o isi s and i s a ia ion is ela ed o changes in pollina o communi y composi ion ac oss popula ions, since pa o his a ia ion is explained by di e en pollina o selec i e p essu es exe ed on lowe colo among G. lu ea popula ions (Sob al e al., 2015); and ( ) he mos abundan pollina o s o ou s udy species, Bombus e es is and B. p a o um (Veiga e al., 2015), possess pho o ecep o s wi h low sensi i i y o ed colo s; bu o he pollina o s in e ac ing wi h G. lu ea plan s, such as Bombus lapida ius, has high sensi i i y o ed colo s despi e he lack o pho o ecep o s equi ed (Pei sch e al., 1992). In o de o de e mine whe he G. lu ea popula ions a e locally adap ed o he pollina o communi y, we e alua ed he e ec o local and o eign lowe colo mo phs on pollina o isi a ion a e by means o ecip ocal ansplan s in wo Gen iana popula ions. Mo e speci ically, we es ed whe he he e is an adap i e ad an age in he local lowe colo Gui ián e al. (2017), Pee J, DOI 10.7717/pee j.2882 2/15 Figu e 1 Colo mo phs o Gen iana lu ea L. a No he n Spain. (A) a . lu ea, he yellow mo ph in Pon ón loca ion; and (B) a . au an iaca, he o ange mo ph in Anca es loca ion. mo ph ela i e o he o eign mo ph in pollina o isi s by compa ing he (i) o e all pollina o assemblage, and (ii) di e en isi o g oups sepa a ely. We p edic ed ha he local lowe colo mo ph would ecei e mo e pollina o isi s han he o eign mo ph, and ha he wo di e en colo mo phs would ecei e isi s by di e en pollina o g oups in hei local popula ions. MATERIAL AND METHODS Plan species Gen iana lu ea L. is an he baceous pe ennial geophy e (Fig. 1). Du ing summe , he hizome de elops a lowe ing s alk o ca. 80 cm high, wi h a basal ose e o med by 4–8 oblong lea es. Flowe s g ow in wo opposed cymose g oups (each one wi h 15 lowe s) a di e en s alk le els, abo e wo ellip ic lea es; and one lowe c owns he op o he s alk. Flowe s ea u e a o a e co olla wi h 5–7 lobules, 5–7 ee s amens, and a ixed o a y wi h wo nec a ies on he base. The ui is an o oid capsule, holding ellip ic and winged seeds o 3–4 mm (Renobales, 2012). This species needs pollina o s—mainly bumblebees— o p oduce seed (see Veiga e al., 2015;Sob al e al., 2015). Gen iana lu ea lowe colo a ies be ween o ange and yellow (see Fig.1) along he Can ab ian Moun ains (No he n Spain; Fig. 2). The wes e n popula ions (wes o Somiedo Na u al Pa k) ha e o ange- lowe ing indi iduals, whe eas he eas e n popula ions (eas o Pue o de Ven ana) ha e yellow- lowe ing indi iduals; and he in e media e popula ions show indi iduals wi h lowe s o a g ada ion be ween hese wo colo mo phs. S udy si e The s udy was conduc ed in July 2012 a he no hwes o he Ibe ian Peninsula in wo popula ions, one loca ed in Sie a de Anca es (o ange- lowe ed popula ion) and o he in Pue o del Pon ón (yellow- lowe ed popula ion), sepa a ed by app oxima ely 160 km (Fig. 2). Popula ion size was es ima ed by he o al numbe o G. lu ea indi iduals (>3,000 in bo h popula ions). Flowe colo o 75 haphaza dly selec ed lowe s ( h ee lowe s pe plan , 25 plan s) was measu ed he p e ious yea in bo h popula ions by means o a spec ome e (USB2000+; Gui ián e al. (2017), Pee J, DOI 10.7717/pee j.2882 3/15 Figu e 2 Geog aphic loca ion o Gen iana lu ea popula ions a No he n Spain in which ecip ocal ansplan s we e conduc ed. Black do , Anca es, he o ange mo ph popula ion; whi e do , Pon ón, he yellow mo ph popula ion. Ocean Op ics, Inc., Dunedin, FL, USA) and pe al colo spec a we e p ocessed using he Spec aSui e R so wa e (Ocean Op ics, Inc., Dunedin, FL, USA). The CIELab Colo ime ic Sys em (CIE P., 2004) was chosen o desc ibe lowe colo a ia ion in he isible ange o he elec omagne ic spec um, h ough h ee a iables educed by p incipal componen analysis ( o addi ional de ails, see Veiga e al., 2015): L(b igh ness o colo , om black o whi e), a( ed colo a ia ion, om g een o ed) and b(yellow colo a ia ion, om blue o yellow). Indi iduals om he Sie a de Anca es popula ion showed o ange-colo ed lowe s (L=16.42; a=11.77; b=23.61), and om he Pon ón popula ion, yellow ones (L=19.64; a=3.84; b=27.84). Al hough ce ain pollina o species may de ec UV ligh , a p e ious s udy ound no di e ences among G. lu ea popula ions and be ween indi iduals wi h di e en colo mo phs (o ange o yellow, disce nible by human eye) in he UV ligh ange (Veiga e al., 2015). Thus, we p esume ha UV ligh does no d i e local adap a ion among popula ions, and conside lowe colo in bo h, he UV and isible ligh anges in his s udy. Expe imen al design In July 2012, a ecip ocal ansplan expe imen was conduc ed o analyze local adap a ion o colo mo phs o pollina o assemblage o bo h popula ions ( hanks o he ield Gui ián e al. (2017), Pee J, DOI 10.7717/pee j.2882 4/15 pe mission issued by he En i onmen al Te i o ial Se ice ins i u ion om León, Regional Go e nmen o Cas illa and León, Te i o ial Delega ion o Go e nmen o Spain— Iden i ie : 12_LE_325_RNA_PuebladeLilio_INV—Re e ence: 06.01.013.016/ROT/abp— File numbe : AEN/LE/103/12). Gi en he di icul y o ansplan ing plan s in po s due o hei ex ensi e hizomes, lowe ing s ems used o he expe imen we e cu a he heigh o hizome, placed in plas ic con aine s wi h wa e (0.5 L), sealed wi h adhesi e ape, and bu ied a g ound le el o p e en e apo a ion and ensu e longe lowe li e. The e ec i eness o his me hod was p e iously de e mined by checking he du a ion o lowe ing and pollina ion in a con ol popula ion and e i ying ha he lowe s we e well-p ese ed in he i s ou days. Fo he ecip ocal ansplan expe imen , 15 lowe ing s ems we e cu in he o ange popula ion (Anca es) and ans e ed o he yellow popula ion (Pon ón) immedia ely a e collec ion, while simul aneously, 15 yellow s ems we e cu in Pon ón and ans e ed o he Anca es popula ion ( o eign ansplan ea men ). Addi ionally, 15 plan s o each popula ion (o ange in Anca es and yellow in Pon ón) we e subjec ed o he same p ocedu es (cu , ansplan ed, and bu ied) o he local ansplan ea men . The same day, plan con aine s we e placed in each popula ion wi hin 1 m2squa es, se in andom layou by picking he g id poin using a andom numbe able. Addi ionally, 15 plan s we e haphaza dly ma ked o se e as a con ol in bo h popula ions. Thus, each popula ion con ained 30 expe imen al plan s and 15 con ol plan s (45 plan s pe popula ion, 90 plan s o e all). No e ha he s em heigh , he numbe o lo al who ls pe in lo escence, and he numbe o lowe s we e measu ed in all plan indi iduals o co ec o possible plan mo phological e ec s in he expe imen al esul s. Pollina o censuses We conduc ed 15 2-minu e pollina o censuses on each o he 45 plan s in bo h lowe colo mo ph popula ions o e 3 days. Acco ding o p e ious s udies, he obse a ion sampling e o (30 min pe plan , 1,350 min in each popula ion) is app op ia e o ob ain an adequa e ep esen a ion o he pollina o spec a o his species and in his geog aphic a ea (see Ma e ials and Me hods sec ion in Veiga e al., 2015;Sob al e al., 2015). Fo each census, he species ha accessed he lowe s and he numbe o lowe s hey isi ed we e eco ded. E e y day and be o e he census, he numbe o open lowe s on each plan was coun ed o co ec o he e ec o lo al display in pollina o a ac ion. Plan censuses we e andomly conduc ed unde sunny condi ions be ween 0800 and 1800 h. The numbe o isi s pe plan ( isi a ion a e pe 30 min) was used as a measu e o pollina ion success. A he end o each census, a sample o pollina o axa obse ed was collec ed and sen o u he iden i ica ion by en omological specialis s om he Uni e si y o O iedo (Spain). Co ec ing any disc epancies be ween ou isual iden i ica ions o bumblebee species and he lab de e mina ions, species we e g ouped o analyses in o eigh mo phological ca ego ies based on p oboscis leng h (Obeso, 1992). The eigh mo phological g oups we e: g oup 1 (Bombus e es is, B. luco um); g oup 2 (B. ho o um, B. jonellus); g oup 3 (B. p a o um, B. so oensis anca icus, B. lapida ius decipiens); g oup 4 (Bombus wu lenii); Gui ián e al. (2017), Pee J, DOI 10.7717/pee j.2882 5/15 g oup 5 (B. mesomelas); g oup 6 (B. pascuo um); g oup 7 (B. (Psi hy us) upes is); g oup 8 (B. hypno um). S a is ical analyses Se e al gene alized linea models (GLM) we e pe o med, whe e he numbe o isi s pe plan was he dependen a iable. The locali y (Anca es and Pon ón), ea men (con ol, local ansplan , and o eign ansplan ), pollina o g oup (desc ibed ea lie ), and he in e ac ions among hese ac o s we e included as ixed e ec s. The models we e pe o med o all pollina o species pooled oge he and o each o he eigh pollina o g oups independen ly. Residuals o he dependen a iable we e i ed in all models wi h a Poisson dis ibu ion and a loga i hmic link unc ion. Addi ionally, pai swise con as s we e pe o med o analyze he di e ences on he isi a ion a e (numbe o isi s pe plan pe 30 min) by he eigh bumblebee g oups be ween and wi hin he ansplan ea men s and con ol g oup a e aged o bo h des ina ion locali ies. To es o local adap a ion, he mean di e ences on isi a ion a e o g oup 1 and 3 bumblebees ( he mos abundan g oups) we e compa ed be ween each pai o ea men s applied o he ansloca ion expe imen ( o eign ansplan s, local ansplan s, con ols), conside ing bo h locali ies (Anca es and Pon ón). All analyses we e pe o med using SPSS S a is ics 20 (IBM Co p., Some s, NY). RESULTS Pollina o censuses In o al 951 insec s (99% bumblebees belonging o he genus Bombus, and he emaining 1% belongs o he genus Apis o he amily Vespidae) isi ed he plan s (538 in Anca es, and 413 in Pon ón), which ep esen ed 3,268 o al cumula i e lowe isi s (see Da a S1). Bombus e es is (g oup 1) and Bombus p a o um (g oup 3) we e he main lowe isi o o Gen iana lu ea plan s in Pon ón, wi h 447 and 372 o al cumula i e isi s espec i ely o each g oup (excluding con ol ea men , see Fig. 3). Bombus so oensis anca icus +Bombus lapida ius decipiens (g oup 3) and Bombus (Psi hy us) upes is (g oup 7) we e he main lowe isi o s in Anca es (338 and 224 o al cumula i e isi s espec i ely o each g oup, only o ansplan ea men s), being he la es g oup absen in Pon ón, and Bombus hypno um (g oup 8) was only p esen in his loca ion (Fig. 3). Recip ocal ansplan s In Anca es, a popula ion o o ange mo phs, mo e isi s pe 30 min we e obse ed o o ange— han o yellow— lowe ing plan s (32.5% e sus 24.7% plan s isi ed, espec i ely o each colo mo ph); whe eas in Pon ón, a popula ion consis ing o yellow lowe s, he e was no di e ence in isi a ion a e be ween he wo mo phs (39.2% e sus 41% plan s isi ed, espec i ely o yellow and o ange mo phs; see Da a S2). Numbe o isi s pe plan depended on he lo al isi o g oup (P<0.001; Table 1). Addi ionally, he numbe o isi s pe plan by lo al isi o g oups we e a ec ed di e en ly pe ea men : con ol, local ansplan , and o eign ansplan (as sugges ed by he signi ican ea men * lo al isi o s in e ac ion; P<0.05; Table 1), and also by locali y (as sugges ed by he signi ican ea men * lo al isi o s*locali y in e ac ion; P<0.05; Table 1). Gui ián e al. (2017), Pee J, DOI 10.7717/pee j.2882 6/15 Figu e 3 To al cumula i e numbe o isi s o he Gen iana lu ea ansplan ed plan s (black-colo ed ba s, o eign; g ey-colo ed ba s, local) by each g oup o bumblebees in he wo s udy si es a No h- e n Spain (Anca es wi h he o ange mo ph and Pon ón wi h he yellow mo ph). Abb e ia ions: T ans- o eign, o eign ansplan ; T ans-local, local ansplan . Flo al isi o g oups we e he ollowing: g oup 1, Bombus e es is,B. luco um; g oup 2, B. ho o um,B. jonellus; g oup 3, B. p a o um,B. so oensis anca i- cus,B. lapida ius decipiens; g oup 4, B. wu lenii; g oup 5, B. mesomelas; g oup 6, B. pascuo um; g oup 7, B. (Psi hy us) upes is; g oup 8, B. hypno um. Resul s o he GLM o analyze he locali y and ea men e ec s on he numbe o isi s pe plan o each pollina o g oup showed di e en lowe colo p e e ences, bu only in he g oups 1 (B. e es is) and 3 (B. p a o um,B. so oensis anca icus +B. lapida ius decipiens; signi ican ea men ac o ; g oup 1: P<0.05; g oup 3: P<0.01; Table 2). G oup 1 and g oup 3 pollina o s isi ed con ol and local plan s simila ly in bo h colo mo ph popula ions (Table 3 and Fig. 4). G oup 1 isi ed o eign ansplan indi iduals mo e o en han local ansplan indi iduals, conside ing bo h loca ions (signi ican mean di e ences be ween o eign and local ansplan s; P<0.05; Table 3). G oup 1 pollina o s isi ed con ol and local plan s simila ly in bo h colo mo ph popula ions (Table 3 and Fig. 4A). Howe e , he absolu e numbe o isi s by hese pollina o s was highe in Pon ón (yellow mo ph popula ion) han in Anca es (o ange mo ph popula ion). G oup 3 isi ed local ansplan s mo e o en han o eign ansplan s in bo h loca ions (no signi ican ea men *locali y in e ac ion; Table 2). The numbe o isi s by g oup 3 o con ol and o eign plan s was equal conside ing bo h popula ions (Table 3 and Fig. 4B). Once again, he absolu e numbe o isi s by hese pollina o s was highe in Pon ón (yellow- lowe ed popula ion) han in Anca es (o ange- lowe ed popula ion). Gui ián e al. (2017), Pee J, DOI 10.7717/pee j.2882 7/15 Table 1 Resul s o he gene alized linea model o analyze he e ec o colo mo phs o Gen iana lu ea on he isi a ion a e (numbe o isi s pe plan pe 30 min) by eigh bumblebee g oups in wo a eas o No he n Spain ( he o ange colo mo ph was p esen a he Anca es si e, whe eas he yellow mo ph was p esen a he Pon ón si e). We ma ked in bold he s a is ically signi ican ac o s (P<0.05). Fac o s: locali y (Anca es and Pon ón, o ange and yellow mo ph popula ion espec i ely); ea men (con ol, lo- cal ansplan s, and o eign ansplan s); and lo al isi o g oups (g oup 1, Bombus e es is,B. luco um; g oup 2, B. ho o um,B. jonellus; g oup 3, B. p a o um,B. so oensis anca icus,B. lapida ius decipiens; g oup 4, B. wu lenii; g oup 5, B. mesomelas; g oup 6, B. pascuo um; g oup 7, B. (Psi hy us) upes is; g oup 8, B. hypno um). Fac o s Wald Chi-Squa e d. .P Locali y 0.702 1 0.402 T ea men 0.110 2 0.946 T ea men * Locali y 4.336 2 0.114 Flo al isi o s 66.548 3 <0.001 T ea men * Flo al isi o s 16.703 6 0.010 T ea men * Flo al isi o s * Locali y 19.573 9 0.021 Table 2 Resul s o he gene alized linea model o analyze he e ec o locali y, ea men , and he co - esponding in e ac ion be ween bo h, on he isi a ion a e (numbe o isi s pe plan pe 30 min) by wo bumblebee g oups (g oup 1 and 3) in wo a eas o No he n Spain (Anca es and Pon ón, he o - ange and yellow mo ph popula ion espec i ely). We ma ked in bold he s a is ically signi ican ac o s (P<0.05). Fac o s: locali y (Anca es and Pon ón, o ange and yellow mo ph popula ion espec i ely); ea men (con ol, local ansplan s, and o eign ansplan s); and lo al isi o g oups (g oup 1, Bom- bus e es is,B. luco um; g oup 3, B. p a o um,B. so oensis anca icus,B. lapida ius decipiens). No e ha we only included esul s o he isi o g oups ha showed signi ican model e ec s (G oup 1, G oup 3). Flo al isi o s Fac o s Wald Chi-Squa e d. .P Locali y 32.222 1 <0.001 T ea men 6.181 2 0.045 G oup 1 T ea men * Locali y 0.692 2 0.707 Locali y 1.867 1 0.172 T ea men 16.331 2 <0.001 G oup 3 T ea men * Locali y 0.114 2 0.945 DISCUSSION Ou esul s show ha Gen iana lu ea plan s om bo h loca ions ace di e en se s o lowe isi o s wi h di e en colo p e e ences—indica ed by di e ences in isi a ion a e— sugges ing ha hese di e ences may play a ole in lo al colo di e gence pa e ns. We ound ha he o ange- lowe ing plan s may be locally adap ed o he pollina o assemblage since hey ecei ed mo e isi s han he yellow- lowe ing ansplan s in Anca es, he popula ion o o ange mo phs. Howe e , we ound no e idence o local adap a ion o yellow- lowe ing plan s o he o iginal popula ion since he numbe o bumblebee isi s was simila o he o ange- lowe ing ansplan s in Pon ón. This esul di e s om p e ious esea ch, which sugges ed ha lowe colo a ia ion among G. lu ea popula ions is ela ed o selec i e p essu es exe ed by di e en pollina o assemblages (Sob al e al., 2015). Se e al possible explana ions may help o cla i y ou indings. The di e en esponses ob ained in he Anca es and Pon ón popula ions a e likely due o a ia ion in he pollina o Gui ián e al. (2017), Pee J, DOI 10.7717/pee j.2882 8/15 Table 3 Resul s o pai wise con as s o analyze he mean di e ences on he isi a ion a e (numbe o isi s pe plan pe 30 min) by bumblebee g oups 1 and 3, a e aged o wo a eas o No he n Spain (Anca es and Pon ón, he o ange and yellow mo ph popula ion espec i ely) and pe ea men (con- ol, local ansplan , o eign ansplan ). We ma ked in bold he s a is ically signi ican mean di e ences be ween he pai o ea men s (P<0.05). Abb e ia ions: T ans-local, local ansplan s; T ans- o eign, o eign ansplan s; and con ol, unmanipula ed g oups. No e ha we only included esul s om pai wise con as s ha showed signi ican mean di e ences be ween pai o ea men s (i.e., pollina o g oups 1 and 3). Flo al isi o s Pai o ea men s Mean-di e ence S.E. d. .P Con ol T ans- o eign −5.090 2.801 1 0.069 Con ol T ans-local 1.390 2.331 1 0.551 G oup 1 T ans- o eign T ans-local 6.480 2.733 1 0.018 Con ol T ans- o eign 18.720 4.369 1 <0.001 Con ol T ans-local 8.210 4.992 1 0.100 G oup 3 T ans- o eign T ans-local −10.510 3.736 1 0.005 assemblage and/o lo al isi o p e e ences be ween he wo a eas (as expe imen ally demons a ed, Jones & Rei hel, 2001). Anca es and Pon ón ha e di e en pollina o assemblage and he species om he Anca es popula ion may p e e local o ange- lowe ing plan s, while he species om Pon ón popula ion clea ly show no p e e ence o one o he wo colo mo phs. Geog aphic a ia ion in lo al isi o p e e ences can esul om changes in pollina o composi ion, abundance, and di e si y (P ice e al., 2005), e en di e en plan popula ions ha sha e he same pollina o g oup ecei e di e en p opo ions o lowe isi s om each species o unc ional g oup (Fens e e al., 2004; Tas a d e al., 2014). Thus, pollina o -dependen o aging p e e ences may cause shi s in he op imal lo al pheno ypes (see Gómez e al., 2009; and e e ences he ein). Wi hin he pool o G. lu ea lowe isi o s in he Anca es popula ion, he mos abundan species (Bombus e es is) has low sensi i i y o ed colo de ec ion (Pei sch e al., 1992; B iscoe & Chi ka, 2001), which p obably explains he educ ion in his species isi a ion a e when inc easing o ange- lowe ing indi iduals (Fig. 4A). We suspec ha o he pollina o species wi h p obably highe sensi i i y o ed colo may inc ease i ness o he o ange- lowe ing plan s. The e o e, he ac ha hey pe cei e ed colo does no imply ha hey necessa ily would p e e his colo o e o he s. In ac , B. lapida ius, p esen in o he o ange G. lu ea (JA Gui ián, pe s. obs., 2012), shows sensi i i y o ed colo s; al hough i lacks ed colo ecep o s (Kugle , 1943;Pei sch e al., 1992). I pollina o species wi h a highe sensi i i y o inna e p e e ences o ed colo s a e p esen , o ange- lowe ed plan s would inc ease hei i ness h ough he ising isi a ion a e by hese pollina o s. The e o e, a ia ion o he pollina o communi y composi ion migh a ec he selec i e p essu es exe ed on G. lu ea lowe colo , and ul ima ely a ec lowe colo a ia ion among popula ions ia local adap a ion. A second explana ion is ha he p e e ence o ce ain pollina o g oups could be dependen on he equency o each mo ph in he popula ion (i.e., he p e e ences o o ange plan s in Anca es a e dilu ed in a con ex o yellow plan s in Pon ón). Empi ical s udies ha examine plan species polymo phisms in lowe colo , mo phology, o sex Gui ián e al. (2017), Pee J, DOI 10.7717/pee j.2882 9/15