Geographic variation, interactions with animals and sexual dimorphism: glancing at range limits in a gynodioecious plant.

Tesis Doc o al An onio R. Cas illa Ál a ez Se illa, Diciemb e 2011 Supe iso D . Conchi a Alonso Menéndez Resea che Depa men o E olu iona y Ecology Doñana Biological S a ion -CSIC (EBD-CSIC) Se ille-Spain Supe iso D . Ca los M. He e a Maliani P o esso o Resea ch Depa men o E olu iona y Ecology Doñana Biological S a ion -CSIC (EBD-CSIC) Se ille-Spain Tu o D . Sal ado Tala e a Lozano P o esso Depa men o Ecology and Plan Biology Uni e si y o Se ille Se ille-Spain A mis pad es y a Me cedes Según la eo ía del mosaico geog á i co de coe olución, las a iaciones espaciales en la magni ud del l ujo génico y en el signo e in ensidad de la selección eje cida po anima- les sob e los asgos de las plan as, son los dos elemen os esenciales que dan luga a la apa ición de mosaicos de selección y de e minan el g ado de adap ación en e las espe- cies de plan as y animales. Hay cie as si uaciones na u ales donde cabe espe a que se p oduzcan, de mane a p edecible y a lo la go de ex ensiones geog á i cas ela i amen e pequeñas, g adien es en la magni ud del l ujo génico, así como en la o ma e in ensi- dad de selección po p obables a iaciones en la composición y/o abundancia de los polinizado es y he bí o os. Uno de es os luga es do ados de una singula idad ecológica p opia son los má genes de dis ibución de las especies ege ales. Además de sus bien conocidas peculia idades ecológicas, los má genes de dis ibución ambién p esen an peculia idades gené icas que los con ie en en luga es ideales pa a e i i ca las p edic- ciones de la eo ía del mosaico geog á i co de coe olución. El abajo de es a esis doc o al se diseñó pa a e i i ca la siguien e p edicción gene al, basada en los pos ulados de la eo ía del mosaico geog á i co de coe olución: g adien es en la magni ud del l ujo génico que sean consis en es con g adien es de selección po animales (polinizado es y he bí o os) sob e algunos ca ac e es ege ales, da án luga a di e enciación in aespecí i ca en los ca ac e es suje os a selección. Es a p edicción se abo dó median e el es udio de pa ones de selección po animales (polinizado es y he - bí o os), a iación eno ípica y l ujo génico eplicado en g upos de poblaciones cen ales y ma ginales del a bus o ginodioico Daphne lau eola. A con inuación se p esen a un esumen de cada uno de los capí ulos incluidos en es a esis doc o al, señalando los esul ados p incipales de cada uno de ellos. Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 18 - En el Capí ulo 1 se compa an poblaciones cen ales y ma ginales de D. lau eola en las Sie as Bé icas con espec o a su aislamien o espacial, amaño poblacional, di e sidad y di e enciación gené ica. Se comp obó que las poblaciones ma ginales no ienen un amaño meno que las po- blaciones cen ales pe o sí es án más aisladas espacialmen e. Sin emba go, el g ado de aisla- mien o espacial de las poblaciones del ma gen occiden al ue mucho mayo . Es as di e encias en la es uc u a geog á ica de las poblaciones die on luga a un pa ón di e en e de es uc u ación gené ica en los dos má genes de dis ibución es udiados y de inie on dos g upos gené icos den o de la especie en el á ea de es udio. En el Capí ulo 2 aplicamos un en oque genómico pa a de ec a ma cado es (AFLP loci) cuyas ecuencias mos asen un pa ón alejado del pa ón neu al (ou lie s), una e idencia de selección na u al sob e ales loci o loci ligados a los mismos. Encon amos que exis e selección di e encial en poblaciones cen ales y ma ginales de D. lau eola. Una ez de ec ados los loci ou lie s, los e i amos del análisis de la es uc u a gené ica de las poblaciones pa a obse a cuán o de la di e enciación gené ica en e los g upos disjun os de poblaciones se podía a ibui a selección di e- encial. Sin emba go, la ue e di e enciación gené ica en e las poblaciones disjun as ma ginales y el es o de poblaciones pe sis ió. Es e esul ado demos ó que ac o es selec i os y no selec i os son ele an es en la ue e di e enciación gené ica de las poblaciones de D. lau eola. Además, in- di iduos hemb a y he ma odi as mos a on pa ones de selección di e en es, con he ma odi as p esen ando siemp e un mayo núme o de loci bajo selección. Es a di e encia en e sexos ue más ma cada en las poblaciones disjun as ma ginales y sugi ió una mayo p esión selec i a sob e in- di iduos he ma odi as que pod ía elaciona se con el ambien e de polinización. El Capí ulo 3 analiza si los indi iduos de poblaciones disjun as ma ginales y cen ales con inuas se di e encian en el amaño y c ecimien o de los indi iduos, su éxi o ep oduc i o y ambien e de polinización. Los indi iduos de poblaciones cen ales con inuas p esen a on amaños mayo es y p oduje on más lo es que die on luga a cosechas de u os mayo es. Sin emba go, la p opo ción de lo es que die on u os (“ ui se ”) ue mayo en las plan as de las poblaciones ma ginales, Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 19 - sugi iendo un se icio de polinización más e icien e en es as poblaciones. Además no de ec amos di e encias en la densidad de indi iduos ep oduc i os en poblaciones cen ales y ma ginales. Nues os esul ados no apoya on po an o ue es di e encias de es és ecológico en e poblacio- nes cen ales con inuas y ma ginales disjun as, pe o indica on cie os cambios en las in e acciones plan a-polinizado que pod ían ene una especial ele ancia pa a el polimo ismo sexual de es a especie ginodioica. En e ellas cabe des aca que el esca abajo Melige hes elonga us, el p incipal polinizado de la especie en las poblaciones cen ales con inuas, no es á p esen e en las poblacio- nes ma ginales disjun as. En el Capí ulo 4 compa amos el éxi o de polinización en indi iduos de dis in o sexo en pobla- ciones cen ales y ma ginales, di e enciando sus componen es cuan i a i o y cuali a i o. Encon- amos una educción en la di e gencia en e sexos en las poblaciones ma ginales disjun as en la can idad y calidad del polen ecibido. Además, analizamos la a iación eno ípica en algunos asgos lo ales así como los pa ones de selección po polinizado es sob e dichos asgos. Encon- amos una educción en el dimo ismo sexual más e iden e, la longi ud del ubo de la co ola, po educción del amaño de los he ma odi as. Además el único asgo eno ípico es udiado que p e- sen ó un pa ón de selección po polinizado es conco dan e con la a iación geog á ica obse ada ue la longi ud del ubo de la co ola: indi iduos he ma odi as con lo es más la gas u ie on ma- yo éxi o ep oduc i o únicamen e en poblaciones cen ales. Po an o, selección di e encial po polinizado es jun o con mecanismos de selección pos -polinización pa ecen es a con ibuyendo a las di e encias en el ni el de dimo ismo sexual lo al en e poblaciones cen ales con inuas y ma ginales disjun as de D. lau eola. Po úl imo, el Capí ulo 5 se cen a en el papel de los he bí o os sob e la di e enciación in aespe- cí ica de los asgos ege a i os de D. lau eola. Los esul ados de es e capí ulo demos a on que los he bí o os no pa ecen juga un papel decisi o en es e sen ido. No encon amos di e encias en e poblaciones cen ales con inuas y ma ginales disjun as en el ni el medio de he bi o ía, desca - ando una mayo p esión po he bí o os en los má genes dedis ibución. Además las di e encias Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 20 - en los ni eles de he bi o ía en e indi iduos no se co espondie on con las dis ancias gené icas en e ellos. Po an o, es as di e encias pa ecen se u o de la a iación espacial en el consumo po he bí o os debido a a iación (mic o)ambien al den o de las poblaciones. Rasgos ege a i os como el á ea especí i ca olia o el con enido en enoles de las hojas ue on di e en es en poblacio- nes cen ales con inuas y ma ginales disjun as de D. lau eola, pe o es a di e enciación egional no se co espondió con una selección di e encial po he bí o os y pod ía ene elación con cambios egionales en la composición del dosel a bó eo y el ambien e lumínico del so obosque. Lejos de se en es aislados den o de los ecosis emas, las plan as cons i uyen elemen- os dinámicos que con inuamen e in e accionan con o os o ganismos i os. La na u a- leza sésil de las plan as limi a decisi amen e dos ases c í icas en su ciclo ep oduc i o como son la ep oducción sexual y la dispe sión de las semillas (He e a 2002). Es os dos p ocesos equie en el mo imien o en el espacio de es uc u as ep oduc i as. Du an e la ep oducción sexual, los g anos de polen deben se anspo ados desde el indi iduo dis ancias a iables has a alcanza los game o i os emeninos de o as plan as, en la po- blación o ue a de ella. Del mismo modo, la dispe sión de semillas implica el mo imien o de és as lejos de la plan a mad e. Ya que la capacidad de mo imien o es una cualidad ampliamen e ex endida den o del eino animal, no es de ex aña que las plan as hayan desa ollado una eno me a iedad de mecanismos des inados a usa a los animales como ec o es que pe mi an el mo imien o de su polen y semillas. De hecho, ac ual- men e es á ampliamen e acep ado que es a eno me explo ación del mo imien o animal pa a la dispe sión del polen y de las semillas cons i uye uno de los ac o es decisi os en la emenda di e si i cación y éxi o ecológico de las angiospe mas (P ice 2002; Kay & Sa gen 2009). También las in e acciones de ipo an agonis a han con ibuido a la di e si i cación den o de las angiospe mas. La he bi o ía o consumo de las plan as po animales engloba ipos muy di e en es de in e acción que an desde las de ipo dep eda- ción (p.ej. consumo de semillas o plán ulas) a las que no implican la mue e de indi iduo (p.ej. consumo de hojas). Debido a su na u aleza sésil, las plan as no pueden escapa del consumo po he bí o os y po an o han enido que desa olla una eno me a iedad de mecanismos des inados a e i a y/o ole a el daño p oducido po los he bí o os. Po lo an o, in e acciones mu ualis as y an agonis as en e plan as y animales han con ibui- do y con inúan con ibuyendo a la eno me di e si i cación den o de las angiospe mas (Labandei a 2002). Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . Va iación geog á i ca en las in e acciones plan a-animal Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 22 - Una de la ideas cen ales del pensamien o e olucionis a ac ual es que los cambios mac oe o- lu i os son el esul ado acumulado de los p ocesos mic oe olu i os que ienen luga a ni el in- aespecí ico (Simpson 1953, Bock 1970). Po es e mo i o, una de las mane as adicionales de abo da el es udio de la adap ación ha sido in es iga la a iación geog á ica en los ca ac e es de los o ganismos y su elación con los posibles ac o es selec i os que pudie an se esponsables de esa a iación (Endle 1986). Es a idea la exp esan de o ma sin é ica Gould & Johns on (1972, p. 457) cuando a i man: “Since we iew anspeci ic e olu ion as an ex ension o e en s a he species le el, he ounda ion o mos e olu iona y heo y es s upon in e ences d awn om geog aphic a ia ion o upon he e i ica ion o p edic ions made abou i ”. En el caso conc e o de las plan as, se ha p es ado especial a ención a la a iación geog á ica en los asgos ele an es pa a sus in e- acciones con animales (p.ej. polinizado es, he bí o os, dispe san es de semillas, e c) dada la im- po ancia ya comen ada p e iamen e de es e ipo in e acciones en los p ocesos de di e enciación in aespecí ica. En es e sen ido, la comunidad animal con la que una de e minada especie ege al in e ac úa se espe a que a íe geog á icamen e. Po an o, el es udio de la a iación geog á ica in aespecí ica en asgos ele an es pa a las in e acciones plan a-animal y su posible elación con una selección di e gen e e ec uada po animales, es c ucial pa a en ende el ínculo en e mic o y mac oe olución en es os ca ac e es y el papel de las in e acciones plan a-animal en la e olución de los sis emas sexuales en plan as (Ba e e al. 2001; Zange l & Be enbaum 2003; He e a e al. 2006). Sin emba go, la mayo ía de es as in es igaciones sob e las elaciones en e plan as y ani- males se han lle ado a cabo sin a ende a su con ex o geog á ico, a pesa de su impo ancia pa a en ende los p ocesos de e olución adap a i a (He e a e al. 2006). La eo ía del mosaico geog á ico de coe olución (TMGC en adelan e) p opues a po Thomp- son (1994, 2005) y los modelos gené icos asociados a ella (Gomulkiewicz e al. 2000; Alleaume- Benha i a e al. 2006) p opo cionan un modelo eó ico que pe mi e es udia el papel causal de los animales en la apa ición y man enimien o de di e encias in aespecí icas en los asgos eno ípicos impo an es pa a las in e acciones plan a-animal. Según la TMGC, las dinámicas coe olu i as globales de las in e acciones plan a-animal ienen de e minadas po es componen es de la es- Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 23 - uc u a geog á ica de las poblaciones de las especies in e ac uan es: los mosaicos de selección, los pun os calien es coe olu i os y la emezcla de ca ac e es (“ ai emixing”; Thompson 1994; 2005). Es os es concep os se desc iben b e emen e a con inuación: “Mosaicos de selección”: el esul ado de las in e acciones a menudo di ie e en e poblaciones de las especies que in e accionan. Es o es debido p incipalmen e a a iaciones locales en el ambien e abió ico, en la es uc u a demog á ica y gené ica y en el con ex o de la comunidad ecológica donde se desa olla la in e acción. Es as di e encias en e poblaciones en el esul ado de la in e acción c ean un mosaico geog á ico en la in e acción que es el ma e ial básico pa a la dinámica de la coe olución. “Pun os coe olu i os calien es y íos”: consecuencia de la a iación geog á ica en el esul ado de las in e acciones, una in e acción puede coe oluciona en algunas poblaciones (pun os calien es coe olu i os), mien as que en o as puede a ec a a sólo una de las especies pa icipan es o no ene ningún e ec o (pun os íos coe olu i os). Incluso pueden exis i poblaciones de una especie pa ici- pan e en la in e acción que se localicen ue a del á ea de dis ibución geog á ica de la o a especie pa icipan e. “Remezcla de ca ac e es”: el mosaico geog á ico se eajus a con inuamen e debido a los e ec os combinados del lujo génico en e poblaciones, a la de i a gené ica y a las dinámicas de ex inción/ colonización local. Especies sexualmen e polimó icas: impo ancia de las a iaciones geog á icas en las in e acciones plan a-animal Las condiciones ecológicas en las que se desa olla una especie a ían an o espacialmen e como empo almen e. Poblaciones en di e en es egiones den o del ango de una especie a me- nudo se en en an a escena ios ecológicos di e en es que dan luga a cambios en la demog a ía, Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 24 - la es uc u a gené ica y las ca ac e ís icas eno ípicas de las poblaciones (Thompson 1994, 2005; He e a e al. 2006). Tal di e enciación geog á ica puede ene impo an es consecuencias mic o y mac oe olu i as que en úl ima ins ancia pueden lle a al aislamien o ep oduc i o y di e si i- cación de linajes de plan as (Bake 1959; Lloyd 1965; S id 1969; G an 1971; McNeill & Jain 1983; Wya 1988; Olms ead 1990; Ba e 1995). Es a a iación geog á ica an o en las condiciones abió icas como bió icas se piensa que ha con ibuido a di e si ica los sis emas ep oduc i os y sexuales de las Angiospe mas (Ba e e al. 2001; Ba e 2002, 2010). Den o de la imp esionan e di e sidad de sis emas sexuales en plan as podemos encon a des- de especies monomó icas donde cada indi iduo con ibuye ep oduc i amen e a a és de sus unciones sexuales masculina y emenina, has a especies polimó icas donde los indi iduos es án ag upados en dos o más mo o ipos sexuales dis in os que con ibuyen a la siguien e gene ación p incipalmen e a a és de la unción masculina o emenina (Lloyd 1980). Den o de los sis emas monomó icos enemos el he ma odi ismo, donde los indi iduos po an lo es con la unción mas- culina y emenina desa olladas y la monoecia donde las unciones masculinas y emeninas es án sepa adas en lo es dis in as pe o den o del mismo indi iduo. En el caso de las especies dimó icas podemos encon a la dioecia donde las unciones sexuales masculina y emenina apa ecen en lo es dis in as que además son po adas po indi iduos dis in os, la ginodioecia con indi iduos hemb as y he ma odi as coexis iendo en sus poblaciones o la and odioecia con indi iduos ma- chos y he ma odi as. Exis en además sis emas he ma odi as pe o polimó icos, como la dis ilia, a los que no nos amos a e e i . La dioecia no es muy ecuen e en angiospe mas a di e encia del eino animal donde es la con- dición más abundan e (6 % de las especies de angiospe mas; Renne & Rickle s 1995). No obs- an e, la dioecia ha e olucionado desde el he ma odi ismo en mul i ud de ocasiones, al menos hay 100 ansiciones desc i as (Ba e 2002; Cha leswo h 2002). En conc e o, la u a e olu i a desde el he ma odi ismo a la dioecia ía ginodioecia es á apoyada eó icamen e y empí icamen e (Cha leswo h 1999; Webb 1999; Weiblen e al. 2000). Es a u a e olu i a implica dos ases con- Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 25 - secu i as, p ime o la in asión y expansión de mu an es es é iles masculinos (i.e. hemb as) den o de poblaciones he ma odi as y, segundo, la acción de la selección na u al pa a inc emen a la unción masculina en los he ma odi as. Sin emba go, apenas hay in o mación sob e los meca- nismos ecológicos que p omue en es as dos ansiciones, un aspec o cla e pa a de e mina si la ginodioecia puede se un sis ema e olu i amen e es able (Ashman 2006). Da win (1877) ue el p ime o que es ableció una asociación en e dimo ismo sexual y es és abió ico, señalando “a e y d y s a ion appa en ly a ou s he p esence o he emale o m” (p. 301). Es a asociación en e dimo ismo sexual y es és abió ico ha sido documen ado en a ios es udios ilogené icos (p.e. Ha 1985; Welle e al. 1995) y ambién po nume osos es udios de a iación en e poblaciones (Welle e al. 1998; Asikainen & Mu ikainen 2003; Vaugh on & Ram- sey 2004; Case & Ba e 2004). Sin emba go, aunque se piensa que los ac o es bió icos ienen un papel ele an e en la apa ición de sis emas polimó icos sexuales, apenas hay es udios que hayan examinado su papel causal. Po ejemplo se piensa que una educción en la calidad de la polinización p omue e di e sas es a egias ep oduc i as como la ansición desde la ecundación c uzada a la au o ecundación (Schoen e al. 1996; He lihy & Ecke 2002; Moelle & Gebe 2005) o la e olución desde el monomo ismo al dimo ismo sexual (Sakai & Welle 1999; Ba e e al. 2001; Ashman 2006). En el caso conc e o de la ginodioecia, se supone que los polinizado es pue- den a ec a a la e olución del sis ema sexual de dos mane as con apues as (Ashman 2006). Po un lado, educciones o cambios en la composición de la auna de polinizado es pod ían lle a a insu icien e o meno polinización en los indi iduos, espec i amen e ( e Ha de & Ba e 1996). En especies ginodioicas con he ma odi as au ocompa ibles, es e peo ambien e de polinización puede lle a a un aumen o de la au o ecundación en los he ma odi as con el consiguien e inc e- men o de la endogamia en su descendencia. El aumen o de la endogamia en la descendencia de los he ma odi as puede supone una des en aja a la ho a de compe i con la descendencia de los indi iduos hemb a, p oducida exclusi amen e po ecundación c uzada, p omo iendo así la expansión de indi iduos hemb a en las poblaciones (Cha leswo h & Cha leswo h 1978). Po o o lado, la p opagación de los indi iduos hemb as den o de las poblaciones puede se impedida po Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 32 - especie en el á ea de es udio con es poblaciones localizadas en el ma gen disjun o su occiden al (Capí ulos 3, 4 y 5). En el su es e peninsula la especie es ecuen e en el macizo o mado po las sie as Cazo la- Segu a-Cas il-Alca az donde ocupa un amplio g adien e al i udinal (700-1900 m.s.n.m.). En Ca- zo la-Segu a, en la mayo ía de las localidades, D. lau eola o ma pa e del so obosque de pina es de Pinus nig a subsp salzmanii, donde ambién se encuen an o as especies leñosas como Que cus aginea, P unus mahaleb, Taxus bacca a, Ace monspessulanum, Junipe us communis o C a ae- gus monogyna. Mención especial me ecen las poblaciones de D. lau eola localizadas en la sie a de Segu a, en el ex emo no o ien al de las Co dille as Bé icas, donde ecuen emen e apa ece asociada a acebedas. En es as poblaciones la ege ación se ca ac e iza po un dosel a bó eo de Pinus nig a subsp. salzmanii bajo el que se desa ollan abundan es los acebos (Ilex aqui olium), y o as especies ales como Co ylus a ellana, Que cus py enaica o Ace opalus subsp. g ana ense. En el su oes e peninsula D. lau eola p esen a un g upo más educido de poblaciones localiza- das en dis in os macizos mon añosos. En a ias de es as poblaciones la especie que ca ac e iza el es a o a bó eo es Abies pinsapo, que da luga a una de las o maciones boscosas más singula es de la Península Ibé ica (Blanco e al. 1998). Bajo la cobe u a de A. pinsapo c ece una comunidad bas an e pob e en é minos de di e sidad en la que des acan especies como Hede a helix, Ruscus aculea us, Rubia pe eg ina o I is oe idissima (A is a 1995). En zonas donde la cobe u a de A. pin- sapo es meno se es ablecen especies como Que cus aginea, Ulex bae icus, C a aegus monogyna o Rosa canina. Cada una de las poblaciones de es udio ue clasi icada siguiendo el sis ema de clasi icación cli- má ico p opues o po Köppen en su úl ima e sión de 1936, conocida ambién como clasi icación de Köppen-Geige (Köppen 1931, 1936). A pesa de que es a clasi icación se de inió hace unos 100 años, sigue siendo una de las clasi icaciones más u ilizadas en es udios clima ológicos a ni el mundial (Agencia Es a al de Me eo ología & Ins i u o de Me eo ología 2011). La clasi icación de Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 33 - Köppen de ine dis in os ipos de clima a pa i de los alo es medios mensuales de p ecipi ación y empe a u a. Los alo es de p ecipi ación y empe a u as pa a cada una de las poblaciones de es udio ue on es imados a pa i de modelos de in e polación basados en egis os me eo ológicos de 30 años (Bus aman e 2003; Tabla 1). Basándonos en los da os ob enidos, odas las poblaciones de es udio se encuand an en la ca ego ía climá ica Csb p opues a po Köppen. Es deci , las poblaciones de D. lau eola del su de la Península ocupan á eas donde la empe a u a media del mes más ío es á comp endida en e 0 y 18 ºC, obse ándose un pe íodo es i al ma cadamen e seco pe o emplado ( empe a u a media del mes más cálido ≤ 22 ºC y con cua o meses o más con una empe a u a media supe io a 10 ºC). Poblacion X Y Al i ud (m.s.n.m.) P (mm) T(ºC) m(ºC) M(ºC) Ma gen no o ien al Acebeas 537013 4241386 1320 1067 11.5 3.9 22.0 Na alasna 539809 4240677 1202 969 11.5 3.8 21.9 Río Made a 534676 4236195 1220 1009 11.5 3.8 21.9 La Cumb e 528903 4225256 1480 912 11.7 4.1 22.0 Acebeas de los Jabalises 536391 4229352 1430 907 11.9 4.3 22.3 Región cen al Espina ea 503818 4192104 1173 870 11.2 3.8 21.5 Cañada del Espino 510134 4197123 1575 1068 11.0 3.7 21.3 Fuen e Be mejo 514050 4198162 1513 1007 10.2 2.9 20.6 Valdecue as 511627 4196202 1380 1051 10.7 3.4 21.1 Roblehondo 511209 4200234 1235 1036 11.1 3.6 21.3 Ma gen su occiden al Almija a 424124 4079200 1453 655 12.0 4.4 20.5 Algeci as 274886 3996149 506 887 16.9 8.3 22.8 Cañada de las Ánimas 319005 4063911 1333 1030 11.6 4.7 20.9 Fuen e Molina 318354 4063324 1380 1046 11.5 4.6 20.8 G azalema 283308 4072648 1229 1270 13.2 6.2 22.3 Sie a del Reloj 288021 4067469 1126 1665 13.4 6.4 22.5 Tabla 1. Ca ac e ización de las poblaciones de es udio. Se indican las coo denadas geog á icas (X,Y), la al i ud, la p ecipi- ación media anual P, la empe a u a media anual T, la empe a u a media del mes más ío m y la empe a u a media del mes más cálido M. Los alo es de empe a u a y p ecipi ación pa a cada localidad ue on es imados a pa i de modelos de in e polación basados en egis os me eo ológicos de 30 años (Bus aman e 2003). Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 34 - Es uc u a de la Tesis Doc o al Es a Tesis doc o al cons a de cinco capí ulos cuyo con enido y inalidad se de allan a con inua- ción. Todos los capí ulos de es a esis co esponden a abajos o iginales, algunos de ellos publi- cados, o os en ías de publicación. En el Capí ulo 1 (“Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo ”) se examina en p ime luga si en el á ea de es udio hay una educción del amaño poblacional y un inc emen o del aislamien o geog á ico de las poblaciones hacia los má genes de dis ibución de D. lau eola. En un segundo paso, se anali- za la es uc u a gené ica de las poblaciones en el á ea de es udio y conc e amen e se e alúa si hay una educción de la di e sidad gené ica y un aumen o de la di e enciación gené ica en poblaciones ma ginales. Publicado en Plan Biology (en p ensa). En el Capí ulo 2 (“Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genemic scanning”) se adop a un en oque basado en el escaneo genómico (“genomic scanning”) pa a de ec a e idencia de selección na u al a pa i de las ecuencias alélicas de un ele ado núme o de ma cado es gené icos (AFLP loci). Conc e amen e se examina si indi iduos de poblaciones cen ales con inuas y ma ginales disjun as p esen an pa ones de selección di e en- cial, y si den o de cada uno de es os g upos poblacionales, indi iduos con dis in o sexo mues an ambién e idencia de selección di e encial. En e isión. El Capí ulo 3 (“Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s”) es á en ocado en la compa ación en e poblaciones cen ales con inuas y ma ginales disjun as en elación al amaño y éxi o ep oduc i o de sus indi iduos. Tan o la iden i- dad del polinizado p incipal como la abundancia de o as especies en lo ación como po enciales compe ido as po los polinizado es son enidas en cuen a en dicha compa ación. Publicado en 2011, Ac a Oecologica 37: 269-276. Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 35 - En el Capí ulo 4 (“Reduced gende di e gence in lo al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub”) se in es iga cómo un cambio en la iden idad del polinizado p incipal a ec a al éxi o de polinización y a los pa ones de selección lo al en pobla- ciones cen ales con inuas y ma ginales disjun as de D. lau eola. En e isión. Po úl imo, el Capí ulo 5 (“Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non- gene ic basis o adul de olia ion”) in es iga las di e encias en los ni eles de he bi o ía en pobla- ciones cen ales con inuas y ma ginales disjun as de D. lau eola. También se explo an la exis en- cia de di e encias en e es os g upos poblacionales en el e ec o de la he bi o ía sob e la i ness así como los pa ones de selección eno ípica de los he bí o os sob e asgos ege a i os. En e isión. Cas illa, A.R., Alonso, C. and He e a, C.M. In p ess. Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean e ugium and biodi e si y ho spo . Plan Biology. Chap e 1.- Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo Abs ac : Dis ibu ion ma gins cons i u e a eas pa icula ly p one o andom and/ o adap i e in aspeci i c diff e en ia ion in plan s. This end may be pa icula ly ma ked in species discon inuously dis ibu ed ac oss moun ainous anges, whe e sha p geog a- phic isola ion g adien s and habi a bounda ies will enhance gene ic isola ion among popula ions. In his s udy we analysed he le el o neu al gene ic diff e en ia ion among popula ions o he long-li ed sh ub Daphne lau eola (Thymelaeaceae) ac oss he Bae ic Ranges, a glacial e ugium and biodi e si y ho spo in he wes e n Medi e anean Basin. Wi hin a ea, co e and ma ginal popula ions o D. lau eola we e compa ed wi h ega d o hei spa ial isola ion, size, gene ic di e si y and diff e en ia ion. A spa ially explici analy- sis conduc ed on he as majo i y o he species’ known popula ions in he s udy a ea (N = 111) showed ha ma ginal popula ions (wes e n and eas e n) p esen la ge spa- ial isola ion han co e popula ions, bu a e no smalle . We compa ed gene ic di e si y and diff e en ia ion be ween co e and ma ginal popula ions using a subsample o i een popula ions and 225 ampli i ed agmen leng h polymo phism (AFLP) ma ke s. Co e and ma ginal popula ions did no diff e in gene ic di e si y, p obably because o he occu ence o la ge popula ions on he local ma gins. Wes e n popula ions we e s ongly diff e en ia ed om he o he popula ions. In addi ion, spa ial and gene ic diff e en ia ion among popula ions was la ge on he wes e n ma gin. Eas e n popula ions cons i u ed a gene ically homogeneous g oup closely ela ed o co e popula ions, despi e hei g ea- e spa ial isola ion. Resul s sugges ha s udies on pheno ypic diff e en ia ion be ween co e and ma ginal popula ions o D. lau eola, and p esumably o he species likewise ha ing discon inuous dis ibu ions ac oss he Bae ic anges, should ake in o accoun geog aphical diff e ences in le el o gene ic diff e en ia ion be ween he diff e en dis i- bu ion bo de s. Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 38 - Resumen Los má genes en la dis ibución de las especies ege ales cons i uyen á eas donde es pa icula men e ac ible encon a di e enciación in aespecí i ca neu al y/o adap a i a. Es a endencia debe ía se especialmen e ma cada en especies dis ibui- das discon inuamen e en dis in as co dille as mon añosas, donde ue es g adien es de aislamien o geog á i co y he e ogenei- dad de hábi a aumen a án el aislamien o gené ico de las poblaciones. En es e capí ulo analizamos el ni el de di e enciación gené ica neu al en e poblaciones del a bus o Daphne lau eola (Thymelaeaceae) en las Co dille as Bé icas, un e ugio gla- cia y un pun o calien e de biodi e sidad en la cuenca occiden al del Medi e áneo. Poblaciones cen ales y ma ginales de D. lau eola en el á ea de es udio ue on compa adas con espec o a su aislamien o espacial, amaño poblacional, di e sidad y di e enciación gené ica. Un análisis espacialmen e explíci o lle ado a cabo usando la mayo pa e de las poblaciones conoci- das de la especie en el á ea de es udio (N = 111) mos ó que las poblaciones ma ginales (occiden ales y o ien ales) p esen a on un mayo aislamien o espacial que las poblaciones cen ales, pe o no p esen a on un meno amaño poblacional. Usando una submues a de 15 poblaciones y 225 ma cado es de polimo i smo en la longi ud de agmen os ampli i cados (AFLPs) compa amos a las poblaciones cen ales y ma ginales en elación a su di e sidad y di e enciación gené ica. Poblaciones cen- ales y ma ginales p esen a on simila es ni eles de di e sidad gené ica, p obablemen e como consecuencia de la exis encia de poblaciones de g an amaño en los má genes locales de dis ibución. Las poblaciones ma ginales occiden ales es u ie on gené icamen e muy di e enciadas del es o de poblaciones. Además, an o el aislamien o espacial como gené ico ue mayo en el bo de occiden al. Las poblaciones o ien ales cons i uye on un g upo gené icamen e homogéneo muy p óximo a las poblaciones cen ales, a pesa de su mayo aislamien o espacial. Los esul ados de es e es udio sugie en que los es udios so- b e di e enciación eno ípica en e poblaciones cen ales y ma ginales de D. lau eola, y p obablemen e de o as especies con dis ibuciones disjun as en sis emas mon añosos, debe ían ene en cuen a di e encias en el ni el de di e enciación gené ica en e los dis in os bo des de dis ibución de la especie. Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 39 - In oduc ion High moun ains a e con inen al islands om a biogeog aphic pe spec i e, each one su oun- ded by low-al i ude en i onmen s cha ac e ized by unsui able p esen -day clima ic condi ions o cold-adap ed moun ain plan s (al hough see Bi ks & Willis 2008). Consequen ly many moun ain plan s o en show disjunc geog aphical dis ibu ions. Wi hin each dis ibu ion pa ch he species colonises a geog aphical g adien o en i onmen al condi ions and becomes mo e abundan whe- e indi idual su i al, ep oduc ion and hence popula ion g ow h a e highes , and inc easingly less abundan as condi ions depa om his op imum (Henge eld & Haeck 1982; B own 1984). The e o e, wi hin each dis ibu ion pa ch, a species is expec ed o exhibi a local cen e o abundan- ce o co e egion, wi h popula ions becoming p og essi ely smalle and mo e spa ially isola ed owa ds he local ange limi s (B ussa d 1984; Vuce ich & Wai e 2003). Such a combina ion o geog aphic a ia ion in popula ion size and spa ial isola ion is expec ed o ha e impo an con- sequences o he gene ic s uc u e o plan popula ions (Vuce ich & Wai e 2003; Alleaume-Ben- ha i a e al. 2006). Gene ic models o ecological ma gins pe ain o local ange ma gins a he han species global dis ibu ion ma gins because hey conside a con inuous spa ial a ea ha is no much la ge han he dispe sal dis ance (Alleaume-Benha i a e al. 2006; B idle & Vines 2007; B idle e al. 2010). These gene ic models p edic , wi hin an ecological g adien , (i) educed gene- ic di e si y o ma ginal popula ions because o ounde e ec s, bo lenecks, inb eeding, gene ic d i o di ec ional selec ion; and (ii) inc eased gene ic di e en ia ion among ma ginal popula ions h ough educed gene l ow (Young e al. 1996; Lammi e al. 1999; Lowe e al. 2005). Fu he mo e, ma ginal and isola ed popula ions could be gene ically dis inc as a esul o he educed a i al o maladap ed genes om co e popula ions (“gene swamping”; Ga cía-Ramos & Ki kpa ick 1997; Alleaume-Benha i a e al. 2006), a ou ing adap a ion o local en i onmen s and hus becoming pa icula ly aluable o species main enance a he global scale (Lesica & Allendo 1995). The- e o e, de e mining he le el o spa ial and gene ic isola ion o ma ginal popula ions wi h ega d o co e popula ions, as well as gene ic di e si y wi hin popula ions, gene ic di e en ia ion among popula ions and i s ela ionship wi h spa ial isola ion is necessa y o e alua e he dis inc i eness Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 40 - and adap i e po en ial o ma ginal popula ions, pa icula ly in plan s associa ed wi h moun ai- nous en i onmen s. Recen ly Ecke e al. (2008) e iewed 84 s udies in ol ing 67 plan species ha es ed o decli- nes in wi hin-popula ion gene ic di e si y and /o inc eases in among-popula ion di e en ia ion owa ds ange ma gins using nuclea molecula gene ic ma ke s. They ound ha 64.2 % and 70.3 % o s udies de ec ed a decline o gene ic di e si y and an inc ease o gene ic di e en ia ion owa ds ange ma gins, espec i ely, he wo ends being posi i ely associa ed. Howe e , he gene ali y o bo h ends is ques ionable gi en he subs an ial biogeog aphic and axonomic bia- ses in he a ailable s udies. Fi s , mos s udies assume ha ma ginal popula ions a e smalle and mo e spa ially isola ed han co e popula ions, al hough only a ew in es iga ions quan i i ed he spa ial isola ion o popula ions and i s ela ionship wi h gene ic di e en ia ion (bu see Ecks ein e al. 2006; Med ano & He e a 2008). Fu he mo e, mos s udies ha e ocused on no he n bo - de s o a species ye di e en selec i e p essu es could ope a e in con ac ing e sus expanding ma gins (Hampe & Pe i 2005; Cassel-Lundhagen 2010). Fo mos empe a e woody species, he Medi e anean moun ains cons i u e he sou he nmos dis ibu ion limi s. These a eas, in con as o no he nmos expanding dis ibu ion limi s, ep esen long- e m s able ma gins which ha e ac- ed as glacial e ugia du ing he las glacial s age (Bene e al. 1991; Ca ión 2002; Ca ión e al. 2003; Médail & Diadema 2009) . Howe e , he numbe o molecula analyses o plan popula ions in hese a eas is s ill limi ed compa ed wi h mo e no he n Eu opean moun ain anges such as he Alps, Py enees o Ca pa hians (K op e al. 2008; bu see K op e al. 2006; He e a & Bazaga 2008 a,b; Med ano & He e a 2008). In addi ion, Medi e anean moun ains cons i u e one o he Wo ld biodi e si y ho spo s because o hei high complexi y in e ms o geology, clima e and his o y (Thompson 2005; Blondel e al. 2010). The e o e, analysing he cu en gene ic s uc u e o plan popula ions in Medi e anean moun ains may be use ul o unde s and how he o og aphy o he e ogeneous landscapes con ibu es o gene ic isola ion o popula ions p omo ing in aspeci i c di e en ia ion (Médail & Diadema 2009; Thompson 2005; pp: 77-80). Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 41 - In his s udy, we es ed o di e ences in popula ion size, spa ial isola ion and gene ic di e si y and di e en ia ion o ma ginal and co e popula ions o he empe a e e e g een sh ub Daphne lau eola in he Bae ic Ranges, a glacial e ugium and one o he mos biologically di e se egions in he Medi e anean Basin, loca ed in he sou he n Ibe ian Peninsula. This a ea cons i u es he sou he nmos dis ibu ion pa ch o he species in Eu ope, encompassing > 100 popula ions wi h a maximum dis ance among hem ca. 350 km, and la gely isola ed om o he dis ibu ion pa ches (Alonso e al. 2007). Speci i cally, we add essed he ollowing ques ions: (i) A e ma ginal popula- ions smalle and mo e spa ially isola ed han co e popula ions? (ii) Do co e popula ions ha e la ge gene ic di e si y han ma ginal ones? (iii) A e ma ginal popula ions mo e gene ically di e en ia ed han co e popula ions? (i ) Do ma ginal popula ions o m dis inc gene ic g oups o co e popula ions a his sou he n dis ibu ion pa ch? Ma e ial and Me hods S udy Species Daphne lau eola L. (Thymelaeaceae) has a Palaea c ic dis ibu ion, ex ending om he A las Moun ains in Mo occo o he UK and Hunga y no hwa ds (Meusel e al. 1978; Fig. 1a). The spe- cies g ows in he unde s o y o coni e ous and mixed o es s, associa ed wi h he main calca eous anges o Eu ope and No h A ica. In he Ibe ian Peninsula, i shows a disjunc dis ibu ion, being abundan in he no he n Can ab ian Range and he Py enees, and also in he sou he n Bae ic Ranges, bu absen in cen al a eas (Alonso e al. 2007 and e e ences he ein). The species l owe s in win e , and he pollen bee le Melige hes elonga us Rosenhaue and, o a much smalle ex en , small soli a y bees and noc uid mo hs a e i s main pollina o s in sou heas Spanish popula ions (Alonso 2004). Howe e , a ecen s udy has shown ha M. elonga us does no isi D. lau eola l owe s in popula ions on he sou hwes edge (Cas illa e al. 2011). Single-see- ded black d upes ipen in June and a e dispe sed by se e al bi d species (Obeso 1985; Hulme 1992). Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 48 - ma ginal popula ions we e also explo ed wi h Wilcoxon exac es s. The in l uence o popula ion size and connec i i y on gene ic di e si y (PLP and Hj) and di e en ia ion (DW and Fs ) was analy- sed using Spea man co ela ions. A h ee-le el hie a chical analysis o molecula a iance (AMOVA; Exco i e e al. 1992) was cal- cula ed wi h he p og am GenAlEx 6.3. We analyzed he ollowing le els: (i) egion (eas e n, co e and wes e n); (ii) popula ions wi hin egion; and (iii) indi iduals wi hin popula ions. Addi ionally, sepa a e wo le el AMOVAs we e calcula ed o each egion o analyse he pa i ioning o mole- cula a iance among and wi hin popula ions. A s anda d es o gene ic isola ion by geog aphic dis ance was uled ou o wo easons. On he one hand, he low numbe o popula ions pe egion (N = 5) did no allow es ing o gene ic isola ion by dis ance in each egion. On he o he hand, wes e n popula ions we e so dis an om he emaining popula ions (> 138 km) ha i in oduces a la ge discon inui y be e desc ibed by he ca ego ical a iable “ egion” han by he con inuous a iable “geog aphic dis ance”. Rela ionships among popula ions we e isualized h ough neighbou -joining clus e ing o pai - wise FST, and he signi i cance o he b anches was assessed h ough 10,000 boo s ap eplica es (Felsens ein 1985). A second app oach based on s a is ical in e ence wi h Bayesian clus e ing me- hods using STRUCTURE 2.2.3. was also used o u he elucida e he uppe mos le el o gene ic s uc u e in he s udy egion (P i cha d e al. 2000; Falush e al. 2007). This p og am p obabilis- ically assigns geno yped indi iduals in o gene ic g oups in o de o minimise depa u es om Ha dy-Weinbe g equilib ium and linkage equilib ium. The numbe o gene ic g oups was explo ed by pe o ming 20 eplica es o each simula ion om K = 1 o K = 17, wi h a bu n-in o 50,000 and MCMC o 100,000, assuming admix u e and co ela ed allele equencies as ecommended in P i - cha d e al. (2000). We applied E anno e al.’s (2005) modal ΔK pa ame e as he choice c i e ion o de ec he ue numbe o gene ic g oups in he se o N = 533 indi iduals assayed. Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 49 - Resul s A e ma ginal popula ions mo e spa ially isola ed and smalle han co e popula ions? We ound di e ences be ween egions in connec i i y (�2 = 1745.08, d = 2, P < 0.0001; K uskal- Wallis es ), wi h densi y o neighbou popula ions being la ges in he co e egion (Table 3). Co e popula ions had mo e han ou imes he numbe o neighbou popula ions wi hin 25 km han eas e n popula ions (W = 43.29, P < 0.0001; Table 3). The di e ence wi h wes e n popula ions was e en la ge . Co e popula ions had mo e han 80 imes he numbe o popula ions wi hin 25 km han wes e n popula ions (W = 17.20, P < 0.0001; Table 3). The e was a signi i can di e ence in popula ion size among egions (�2 = 8.52, d = 2, P < 0.05; K uskal-Wallis es ). Unexpec edly, he co e egion had he lowes popula ion sizes, wi h only 17 % o popula ions ha ing mo e han 100 ep oduc i e indi iduals, whe eas he mo e isola ed po- pula ions in he wes e n egion had he la ges a e age size, wi h 83 % o popula ions ha ing mo e han 100 ep oduc i e indi iduals (Table 3). Eas e n and co e popula ions did no di e in popula ion size (W = 605.5, P = 0.12) bu wes e n popula ions had la ge a e age popula ion size han co e popula ions (W = 122, P < 0.01). Connec i i y was no co ela ed wi h popula ion size( s = - 0.15, P = 0.11; N = 111). Table 3. A e age alues o connec i i y in eas e n, cen al and wes e n popula ions o D. lau eola in he sou h o Ibe ian Peninsula. N indica es he numbe o popula ions in each egion. Connec i i y e e s o he numbe o neighbou popula ions in a dis ance adius o 25 km. Popula ion size was ca ego ized acco ding o he abundance o ep oduc i e indi iduals (1, ≤30; 2, 31-100; 3, ≥ 100). The able shows he pe cen age o popula ions o each ca ego y pe egion. N Connec i i y Popula ion Size Ca ego y (%) 1 2 3 Eas e n 17 20.8 ± 2.5 12 47 41 Cen al 88 82.6 ± 0.8 8 75 17 Wes e n 6 1.7 ± 0.2 17 0 83 Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 50 - Do Co e Popula ions Ha e La ge Gene ic Di e si y Than Ma ginal Ones? A e age gene di e si y o co e popula ions did no di e signi i can ly om eas e n (W= 27, d = 8, P = 0.5) o wes e n popula ions (W = 33, d = 8, P = 0.1), al hough mean Hj ended o be sligh ly la ge in co e (0.140) han in ma ginal popula ions (0.126 and 0.113 in eas e n and wes e n popu- la ions, espec i ely; Table 1). Wes e n popula ions exhibi ed he b oades a ia ion ange o gene di e si y, p esen ing he minimum and maximum alues (0.068 and 0.195) o all s udied popula- ions. Simila ly, co e popula ions did no show s a is ically signi i can di e ences in he pe cen a- ge o polymo phic loci (PLP) ela i e o eas e n (W = 27, d = 8, P = 0.5) o wes e n popula ions (W = 33.5, d = 8, P = 0.1), al hough mean PLP ended o be la ge in co e han in ma ginal popula ions (33.9 %, 32.0 % and 26.7 % in co e, eas e n and wes e n popula ions, espec i ely; Table 1). Again, he maximum and minimum PLP alues (48.9 % and 13.3 %) we e ound in wes e n popula ions. The FML popula ion had he la ges a e age alues o bo h gene di e si y and PLP in con as o he es o wes e n popula ions wi h low alues o bo h es ima es o gene ic di e si y. In ac , excluding FML om analysis led o an impo an dec ease in gene ic di e si y o wes e n egion (Hj = 0.093 and PLP = 21.1 %). Popula ion size and PLP showed a posi i e ela ionship, al hough i was only ma ginally signi i can ( s = 0.50, P = 0.06, N = 15). No signi i can ela ionship was ound be ween popula ion size and gene di e si y ( s= 0.29, P = 0.30, N = 15). PLP and gene di e si y we e bo h posi i ely ela ed o connec i i y, al hough he ela ionships we e only ma ginally signi i - can ( s = 0.47, P = 0.07, N = 15 in bo h cases). A e Ma ginal Popula ions Mo e Gene ically Di e en ia ed Than Co e Ones? On a e age, wes e n popula ions had mo e p i a e agmen s (2.0) han ei he eas e n o co e popula ions (0.6 and 0.2 p i a e agmen s espec i ely; Table 1). The wes e nmos popula ion (ALG) had he highes numbe o p i a e agmen s despi e i s low alues o mean gene di e si y and pe cen age o polymo phic loci (Table 1). Mean pai wise FST was la ge in wes e n han in co e popula ions (W = 55, d = 8, P < 0.0001; Fig. 2), which we e in u n mo e gene ically di e en- ia ed han eas e n ones (W = 64, d = 8, P < 0.001; Fig. 2). Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 51 - Figu e 2. Gene ic di e en ia ion o he popula ions in co e and ma ginal egions. Popula ion pai wise FST and gene ic a i y (DW) pe egion (eas e n, co e and wes e n) a e ep esen ed by means (±SE). S a is ically signi i can a ia ion be ween ma ginal and co e popula ions is indica ed wi h as e isks (* P <0.01; ** P <0.001; *** P <0.0001 ). Fu he mo e, he gene ic a i y index DW was signi i can ly la ge in wes e n popula ions han in co e popula ions (W = 16, d = 8, P < 0.01; Fig. 2). In con as , he e we e no di e ences in DW be ween co e and eas e n popula ions (W = 26.5, d = 8, P = 0.4; Fig. 2). In e es ingly, gene ic a i y we e nega i ely ela ed o connec i i y ( s = -0.59, P < 0.05, N = 15). Popula ion pai -wise FST also was nega i ely ela ed o connec i i y o popula ions, al hough he ela ionship was only ma gi- nally signi i can ( s = -0.47, P = 0.08, N = 15). Popula ion size was un ela ed o gene ic a i y and popula ion pai -wise FST ( s = 0.03, P = 0.90 and s = 0.08, P = 0.79 espec i ely; N = 15). Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 52 - The h ee-le el AMOVA e ealed a s ong gene ic di e en ia ion among egions (�ST = 0.45). Mos molecula a iance was accoun ed o di e ences among egions and wi hin popula ions (46 and 40 % espec i ely), whe eas di e ences among popula ions wi hin a egion accoun ed o a educed amoun o molecula di e si y (16 %). Resul s o he AMOVA analysis also showed ha wes e n popula ions we e mo e gene ically di e en ia ed among hem, wi h la ge alues o mo- lecula a iance explained among popula ions, and an a e age �ST highe han eas e n and co e popula ions (Table 4). Eas e n popula ions p esen ed he lowes alues o molecula a iance explained among popula ions and �ST (Table 4). Sou ce o a ia ion d. . % Va iance ФST P- alue 1. Th ee-le el AMOVA Among egions 2 45 0.455 0.0001 Among popula ions 12 14 Wi hin popula ions 518 40 2. Eas e n Among popula ions 4 9 0.090 0.0001 EWi hin popula ion 170 91 3. Co e Among popula ions 4 13 0.125 0.0001 Wi hin popula ion 169 87 4. Wes e n Among popula ions 4 47 0.472 0.0001 Wi hin popula ion 179 53 Table 4. Gene ic di e en ia ion (�ST) and pe cen age o molecula a iance dis ibu ed among egions, among popula ions and wi hin popula ions in he h ee egions de i ned in his s udy. 9999 pe mu a ions we e used in he analysis. Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 53 - Do Ma ginal Popula ions Fo m Gene ically Dis inc G oups? Two dis inc gene ic clus e s, one ep esen ed by he wes e n popula ions and he o he g ou- ping oge he co e and eas e n popula ions, we e suppo ed by bo h neighbou -joining and Baye- sian app oaches (Fig. 3 and 4). The dend og am based on neighbou -joining app oach showed a spli be ween wes e n and he o he popula ions wi h a suppo o 100 % (Fig. 3). Wi hin he la ges g oup, wo popula ions, VCV and FBJ, bo h loca ed in a pa icula wa e shed (Guadalen ín Valley; Fig. 4b), appea ed di e en ia ed om each o he and om he es o he co e and eas e n popula- ions, wi h a suppo o 97 % in bo h cases (Fig. 3). In addi ion, he Bayesian analysis showed a dis- inc modal maximum o ΔK a K = 2 gene ic g oups. Membe ship assignmen s o he wo gene ic g oups exhibi ed a dis inc geog aphic pa e n coinciden wi h he esul s o he neighbou -joining app oach (Fig. 4a), suppo ing ha wes e n popula ions o med a gene ic g oup di e en ia ed om he es o popula ions analysed. Figu e 3. Neighbou -joining clus e ing based on FST, showing he ela ionship among he 15 sampled popula ions o D. lau eola. Boo s ap pe cen age alues a e indica ed abo e he b anches. Popula ions coded as in Table 2 and egions as in Fig. 1. Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 54 - Figu e 4. (a) Gene ic popula ion s uc u e o Daphne lau eola in he Bae ic Ranges in e ed using he model-based Bayesian clus e ing implemen ed in STRUCTURE which does no equi e a p io i ca ego ies. The indi idual pie cha s indica e he mean p opo ion o membe ship o each local popula ion o he in e ed numbe o K = 2 gene ic g oups. (b) De ailed popula ion s uc u e wi hin he la ges gene ic g oup ha combines co e and eas e n popula ions. Dashed line indica es he highes ele a ion o he moun ainous ange. Popula ions coded as in Table. Discussion Gene ic cha ac e is ics o popula ions may be s ongly in l uenced by he size and spa ial dis- ibu ion o popula ions h ough hei ela ionships wi h gene ic d i and gene l ow. Thus, hese cha ac e is ics a e expec ed o a y ac oss species’ geog aphic anges (Ecke e al. 2008). In he o- llowing pa ag aphs, we will discuss ou esul s suppo ing, i s , he inc eased spa ial isola ion o ma ginal popula ions and, second, he gene ic di e en ia ion o he mos spa ially isola ed local ma gin o he sh ub D. lau eola in he Medi e anean Bae ic Ranges, an a ea close o he species sou he n dis ibu ion ange. Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 55 - Spa ial Isola ion and Size o D. lau eola Popula ions in a Sou he n Dis ibu ion Pa ch Ou esul s, based on an ex ensi e sampling ac oss he Bae ic Ranges, showed ha popula ions o his empe a e sh ub we e ewe and mo e spa ially isola ed owa ds local eas e n and wes e n ma gins, suppo ing he hypo hesis o inc eased isola ion a local ange ma gins (B own 1984; Vuce ich & Wai e 2003; B idle e al. 2010), and highligh ing he sui abili y o he spa ially explici analysis adop ed he e o iden i y meaning ul geog aphic discon inui ies (Table 2). The wes e n egion comp ised a ew highly isola ed popula ions loca ed in di e en moun ain anges (Fig. 1b), he co e egion wi h abundan and e y connec ed popula ions a all dis ances comp ised popu- la ions mainly loca ed in he Sie as de Cazo la and Las Villas, and he eas e n egion associa ed wi h hose popula ions in Sie as de Segu a and Alca az, whe e he species is s ill common bu popula ions exhibi a educ ion in connec i i y among hem (Fig. 1c; Table 2). The p edic ed decline in popula ion size owa ds local bo de s o dis ibu ion was no con i med in D. lau eola popula ions ac oss he Bae ic Ranges (see also Jump & Woodwa d 2003; He lihy & Ecke 2005). In ac , co e popula ions we e on a e age smalle han ma ginal popula ions, al hough he end was only signi i can in he compa ison be ween co e and wes e n popula ions. The wide al i udinal ange o co e popula ions could likely con ibu e o a iabili y in popula- ion size in his egion h ough al i udinal g adien s in ecological condi ions (He e a and Bazaga 2008b). The p esence o la ge ma ginal popula ions wi h mo e han 100 ep oduc i e indi iduals sugges s he exis ence o some sui able loca ions wi hou e idence o ecological s ess o D. lau- eola owa ds he local bo de s o dis ibu ion (Cas illa e al. 2011). Gene ic Di e si y, Di e en ia ion and S uc u e, and hei Rela ionships wi h he Spa ial Isola ion o Popula ions Co e and ma ginal popula ions o D. lau eola in he Bae ic Ranges exhibi ed simila le els o gene ic di e si y, and egions did no di e signi i can ly in ei he gene di e si y o pe cen age o polymo phic loci wi hin a popula ion. These esul s con as wi h hose ound o o he plan s (e.g., Lammi e al. 1999; Lönn & P en ice 2002; A naud-Haond e al. 2006; Ecke e al. 2008). The absence o clea di e ences be ween co e and ma ginal popula ions in gene ic di e si y could be Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 56 - ela ed o he in l uence o popula ion size (Lammi e al. 1999; Leimu e al. 2006) because ma ginal popula ions, con a y o expec a ions, we e no smalle han co e popula ions. Fu he , ou esul s suppo ed ha PLP was ma ginally ela ed o popula ion size. The e o e, he exis ence o la ge popula ions on D. lau eola’s local bo de s o dis ibu ion could p e en gene ic e osion h ough mechanisms o gene ic escue (Young e al. 1996; Ga an e al. 2007). In e es ingly, he wes e n ma gin was in e nally e y he e ogeneous wi h ega d o he gene ic di e si y o he popula ions. Wi h he excep ion o FML popula ion ha exhibi ed he highes gene ic di e si y among all s udy popula ions, he o he ou ended o ha e a much mo e educed gene ic di e si y han co e and eas e n popula ions, mos ly suppo ing he expec a ion o a educed gene ic di e si y in he mos spa ially isola ed popula ions o he wes e n ma gin, wi h he single excep ion al eady men ioned. S ong gene ic di e en ia ion among D. lau eola popula ions ac oss he Bae ic Ranges was in- dica ed in ou esul s using AFLP ma ke s (�ST = 0.45). These esul s a e consis en wi h obse - ed gene ic di e en ia ion in DNA sequence in a much educed numbe o sampled popula ions and indi iduals (Alonso and He e a 2011). In addi ion, he wo es ima es o gene ic di e en ia ion (FST and DW) we e posi i ely ela ed o he spa ial isola ion o popula ions. In pa icula , esul s showed ha he highly isola ed wes e n popula ions cons i u e a gene ic g oup s ongly di e en- ia ed om co e and eas e n popula ions, suppo ing a posi i e ela ionship be ween geog aphic isola ion and gene ic di e en ia ion in ou s udy sys em (Ecks ein e al. 2006; Med ano & He e a 2008). Also, ou esul s sugges he exis ence o wo ypes o local bo de s o dis ibu ion in he s udy a ea om a gene ic iewpoin : a wes e n ma gin wi h popula ions being bo h mo e spa ially and gene ically isola ed among hem, as usually happens in he ea edge o species dis ibu ions (Hampe & Pe i 2005); and an eas e n ma gin comp ising a g oup o mo e spa ially isola ed bu gene ically homogeneous popula ions, closely ela ed o co e popula ions and he e o e mo e y- pical o a expanding edge (Schönswe e e al. 2002; Csë go e al. 2009). The gene ic simila i y among co e and eas e n popula ions sugges s ex ensi e gene l ow be ween he wo egions. Eas- e n popula ions a e dis an by only abou 20-30 km om co e popula ions, all loca ed along an ex ensi e moun ainous ange, which could acili a e local gene l ow h ough na u al co ido s Gene ic s uc u e o he sh ub Daphne lau eola ac oss he Bae ic Ranges, a Medi e anean glacial e ugium and biodi e si y ho spo . - 57 - o he dispe sal ec o s o he species (see e.g. Gaudeul e al. 2000; Ba e e al. 2004). Howe e , wes e n popula ions a e loca ed a away om he co e popula ions (130 km app oxima ely), limi ing he gene l ow among wes e n and co e popula ions o long-dis ance in equen dispe sal e en s (K op e al. 2006; K op e al. 2008). In addi ion, wes e n popula ions a e loca ed in di e- en moun ainous anges sepa a ed by dozens o hund eds o kilome es o unsui able habi a s o he species, which could also educe gene l ow among popula ions on he wes e n edge, hus p omo ing hei gene ic di e en ia ion. Suppo ing his no ion, wes e n popula ions also had he highes alues o gene ic a i y es ima es and ecen s udies ha e also e ealed some ecological pa icula i ies o hese popula ions. Fo ins ance, he emale equency dec eases in sou hwes po- pula ions, he ALG popula ion being cons i u ed exclusi ely by he maph odi e plan s in he mos ex eme case (Alonso e al. 2007). In addi ion, indi iduals o wes e n popula ions p esen lowe l o al displays han hose o co e popula ions, coinciden wi h a change in he pollina o auna (Cas illa e al. 2011). The ecological pa icula i ies o wes e n popula ions and hei s ong gene ic isola ion highligh he po en ial ole ha local adap a ion e en s could ha e in hese ma ginal disjunc popula ions (Lesica & Allendo 1995; Ga cía-Ramos & Ki kpa ick 1997; A naud-Haond e al. 2006; Cassel-Lundhagen 2010). Concluding Rema ks Among-popula ion pheno ypic di e gence is ela i ely common ea u e in plan species ac oss Medi e anean moun ain anges, geog aphic ma gins o species’ anges being pa icula ly p one o he e olu ion o his andom o adap i e di e en ia ion. Backg ound in o ma ion on pa e ns o neu al di e en ia ion is pa icula ly in e es ing o disen angling he ole o na u al selec ion on his pheno ypic di e gence. This app oach is especially use ul in plan species o Medi e anean moun ain anges because o he s ong ele ance o his o ical p ocesses on he gene ic s uc u e o hei popula ions. In his sense, gene ic s uc u e analyses compa ing co e and ma ginal po- pula ions based on neu al ma ke s eme ge as a use ul ool o es ima e he le el o neu al di e- en ia ion in popula ions whe e among-popula ion pheno ypic di e en ia ion is expec ed. Resul s o ou s udy show ha neu al di e en ia ion a ies on di e en local ma gins o dis ibu ion. In Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 64 - phism in ep oduc i e ai s con as ing wi h he modes sexual dimo phism in ege a i e ai s (Ashman, 2005). The e o e, di e en ial selec i e egimes on ep oduc i e ai s (e.g. l o al ai s) o he maph odi e and emale indi iduals a e expec ed o occu . Pollina o s cons i u e impo an selec i e agen s on ep oduc i e ai s and hus, hey may ha e a ele an ole on de e mining he ex en o sexual dimo phism in hese ai s, pa icula ly because pollina o auna o en a y geog aphically (He e a e al., 2006; Ashman, 2000; Ashman & Die ende e , 2001; Ashman, 2006 and e e ences he ein). Daphne lau eola L. (Thymelaeaceae) is a sexually polymo phic sh ub ha p esen s gyno- dioecious popula ions wi h he maph odi e and emale indi iduals, and monomo phic popula ions made up exclusi ely o he maph odi es (Alonso e al., 2007). In gynodioecious popula ions, D. lau eola shows sexual dimo phism in ai s ela ed o i s l o al mo phology and chemical de en- ces. He maph odi es ha e la ge l owe s han emales (Alonso, 2005), a widesp ead pa e n in gynodioecious species (Eckha , 1999; Shyko e al., 2003). Fu he mo e, emales p esen la ge concen a ion o speci i c chemical de ences (couma ins) han he maph odi es (Alonso e al., 2005). We s udied pa e ns o selec ion in popula ions o D. lau eola loca ed ac oss he Bae ic Ranges in sou he n Ibe ian Peninsula. Gene ic s uc u e analyses ha e shown ha D. lau eola p esen s wo gene ically di e en ia ed popula ion g oups (SE and SW popula ions he ea e ; Alonso & He e a, 2011; Cas illa e al., unpublished da a). Howe e , he inclusion o only a ew ou lie loci unde selec ion could g ea ly bias he es ima es o popula ion e olu iona y his o y as de e minan o gene ic popula ion s uc u e (Luika e al., 2003; He e a & Bazaga, 2008). The e o e, he analysis o he same polymo phic ma ke s disca ding he ou lie loci may e eal how much o he gene ic di e gence be ween SE and SW popula ions o D. lau eola ac oss he Bae ic Ranges is accoun ed by di e ences in he selec i e egimes. Fu he mo e, SE and SW popula ions di e also in he po- llina o auna and hei o e all emale equency (Alonso e al., 2007; Cas illa e al., 2011). Di e- ences in ep oduc i e success and pheno ypic ai s (e.g. l o al display) sugges he exis ence o local adap a ion in he mos isola ed popula ions o he SW egion (Cas illa e al., 2011). In his s udy, we adop ed a genomic scan app oach o de ec signa u es o di e en ial selec ion be ween Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 65 - emale and he maph odi e indi iduals o D. lau eola, and o asses whe he pa e ns o sex-speci i c selec ion emained consis en be ween he SE and SW popula ions. Speci i cally, we add essed he ollowing ques ions: (1) a e he e gene ic signa u es o di e en ial selec ion be ween he wo s u- dy egions? (2) does he gene ic di e gence be ween SE and SW popula ions o D. lau eola emain a e excluding loci unde selec ion? (3) a e he e con as ing selec i e egimes on indi iduals o di e en sex? (4) do di e ences be ween sexes in he selec i e pa e ns emain consis en ac oss egions? Ma e ial and me hods S udy Species and Popula ions Daphne lau eola L. (Thymelaeaceae) is an e e g een sh ub wi h a disjunc Palaea c ic dis i- bu ion, wi h dis ibu ion pa ches on he main calca eous moun ainous anges o Eu ope and no - he n A ica, including he Bae ic Ranges in Sou he n Spain whe e his s udy was conduc ed (Fig. 1). Wi hin he Bae ic Ranges, D. lau eola is less equen in he SW egion, whe e popula ions a e cha ac e is ically mo e isola ed and gene ically di e en ia ed om hose in he SE egion, which ep esen s he co e a ea o dis ibu ion (Alonso & He e a, 2011; Cas illa e al., unpublished da a). The wo egions also di e in he composi ion o pollina o auna, wi h SE popula ions p esen ing as main pollina o he pollen bee le Melige hes elonga us Rosenhaue , which is absen in SW po- pula ions (Alonso, 2004; Cas illa e al., 2011). Fu he mo e, o e all emale equency is lowe in SW popula ions (Alonso e al., 2007). Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 66 - Figu e 1. Dis ibu ion o Daphne lau eola popula ions in he Bae ic Ranges (Sou he n Spain), and loca ion o he SE and SW popula ions de i ned in his s udy, he e deno ed by i lled ci cles and black iangles, espec i ely. The g ey scale co esponds o diff e en in e als o al i ude (m.a.s.l.). AFLP analyses We e sampled 15 popula ions o D. lau eola, en in he mos densely popula ed SE egion and i e in he ma ginal SW egion (Fig. 1). In Ma ch 2007, esh lea ma e ial was collec ed om 25-40 andomly chosen ep oduc i e indi iduals pe popula ion, placed in small pape en elopes and d ied immedia ely a ambien empe a u e in sealed con aine s wi h abundan silica gel. D ied lea es we e homogenized o a i ne powde using a Re sch MM 200 mill. To al genomic DNA was ex ac ed om app oxima ely 15 mg o g ound lea ma e ial using DNeasy Plan Mini Ki (Qiagen) and ollowing he manu ac u e p o ocol. DNA concen a ion o ex ac s was es ima ed by un- ning elec opho eses o 5 μl aliquo s on 0.8 % aga ose gels. The AFLP analysis was pe o med es- sen ially as o iginally desc ibed by Vos e al. (1995), wi h modi i ca ions in ol ing he use o l uo- Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 67 - escen dye-labeled selec i e p ime s ollowing Applied Biosys ems (2005). Res ic ion-liga ion was conduc ed using EcoRI / MseI and Ps I / MseI endonuclease mix u es and doubled-s anded adap o s. A o al o 32 EcoRI + 3 MseI + 3 and 32 Ps I + 3 MseI + 3 p ime pai s we e i s sc eened o selec i e ampli i ca ion in a pilo s udy conduc ed wi h a andom subsample o 8 indi iduals om ou widely sca e ed popula ions. We i nally selec ed ou EcoRI + 3 / MseI + 3 and ou Ps I / MseI + 3 p ime combina ions on he basis o hei highe polymo phism and ease o in e p e a ion o band pa e n (see Table 2 in Chap e 1). Each plan was i nge p in ed using he eigh combina ions chosen. F agmen sepa a ion and de ec ion was made using an ABI PRISM 3100 DNA sequence . The p esence o absence o each ma ke in each indi idual plan was sco ed manually by isualizing elec opho eg ams wi h Ge- neMappe 3.7 so wa e. All sco ing was done by he same pe son (A.R.C.). P io o s a is ical analy- ses, we excluded ma ke s p esen in < 1 o > 99 % o indi iduals. Fu he mo e, only agmen s ≥ 150 base pai s in size we e conside ed, as a way o educing he po en ial impac o size homoplasy (Vekemans e al., 2002; Caballe o e al., 2008). In addi ion, i e plan s ha p oduced noisy sequen- ce elec opho eg ams o some p ime combina ions a e unning he analyses se e al imes, we e also excluded om he sample. Ou i nal da a se consis ed o a o al o 493 indi iduals, each sco ed o p esence / absence o 225 polymo phic loci. Geno yping e o a es we e de e mined o each p ime combina ion by unning epea ed, in- dependen analyses o a o al o 32 andomly chosen indi idual plan s, and es ima ed as he a io o he o al numbe o loci wi h disco dan sco es (all indi iduals combined) o a p oduc o he numbe o indi iduals by he o al numbe o sco ed loci (Bonin e al., 2004; Pompanon e al., 2005). E o a es a ied among p ime s combina ions, being la ge in EcoRI combina ions han in Ps I combina ions (see Table 2 in Chap e 1). The mean e o a e (± SE) o he whole se o eigh combina ions was 0.9 ± 0.4%. Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 68 - De ec ion o Loci Unde Selec ion Analyses o iden i y loci unde selec ion we e conduc ed sepa a ely o SE and SW popula ions o D. lau eola. In addi ion, we analysed sepa a ely emale and he maph odi e indi iduals wi hin each popula ion g oup (57 emales and 141 he maph odi es in SW popula ions; 102 emales and 193 he maph odi es in SE popula ions). The so wa e package BayeScan e sion 2.01 (Foll & Gagio i, 2008; Foll e al., 2010) was used o es o he neu ali y o he AFLP ma ke s. BayeScan ex ends Beaumon & Balding’s (2004) Ba- yesian me hod ha implemen s he mul inomial-Di ichle likelihood, which a ises in a wide ange o neu al popula ion gene ic models. The me hodology consis s o iden i ying loci ha p esen po- pula ion di e en ia ion coe i cien s (FST) ha a e dis inc om hose unde neu al expec a ions. A locus is suspec ed o be unde selec ion i he locus-speci i c e ec is signi i can ly di e en om ze o. The me hod de eloped by Foll & Gagio i (2008) uses a hie a chical Bayesian app oach o es ima e he pos e io p obabili ies o wo al e na i e models, one including he e ec s o he se- lec ion and one excluding hem. In he con ex o he mul iple es ing, we also need o inco po a e ou scep icism ela ed o s a is ical e o ype I abou he chance ha each locus is unde selec ion. This is done by BayeScan se ing he p io odds o he neu al model and using he Pos e io Odds ins ead o Bayes ac o s o make decisions (Foll e al., 2010). Pos e io odds a e simply he a io o pos e io p obabili ies, and indica e how mo e likely he model wi h selec ion is compa ed o he neu al model. Pos e io p obabili ies di ec ly allow he con ol o he False Disco e y Ra e (FDR). In ou analyses, we calcula ed he pos e io p obabili ies h eshold o de ec loci unde selec ion leading o a False Disco e y Ra e ≤ 0.001 and we used a alue o p io odds o 10, which is easo- nable o he iden i i ca ion o candida e loci wi hin a ew hund eds o ma ke s (Foll e al., 2010). We an BayeScan wi h 10 pilo uns, bu n-in o 50000, o al o 550000 i e a ions, sample size o 10000 and hinning in e al o 50. Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 69 - Gene ic S uc u e o Popula ions We emo ed ou lie loci om he analysis o he gene ic s uc u e o de ec how much o he gene ic di e gence be ween SE and SW popula ions o D. lau eola ac oss he Bae ic Ranges was accoun ed by di e ences in he selec i e egimes. The inclusion o only a ew ou lie loci among many neu al loci can g ea ly bias es ima es o he e ec s o popula ion e olu iona y his o y (Luika e al., 2003; He e a & Bazaga, 2008). The e o e, o a oid making e oneous conclusions om using ou lie loci, popula ion pa ame e s should be es ima ed wi hou ou lie s (Luika e al., 2003; He e a, 2011). The loci depa ing signi i can ly om neu al expec a ion we e emo ed om u he analyses o gene ic s uc u e and gene l ow, al oge he 13 loci (Supplemen a y ma e ial). An app oach based on s a is ical in e ence wi h Bayesian clus e ing me hods using STRUCTURE 2.2.3. was hen used o u he elucida e he uppe mos le el o gene ic s uc u e in he s udy e- gion (P i cha d e al., 2000; Falush e al., 2007). This p og am p obabilis ically assigns geno yped indi iduals in o gene ic g oups in o de o minimize depa u es om Ha dy-Weinbe g equilib ium and linkage disequilib ium. The numbe o gene ic g oups was explo ed by pe o ming 20 epli- ca es o each simula ion om K=1 o K=17, wi h a bu ning o 50000 and MCMC o 100000, as- suming admix u e and co ela ed allele equencies as ecommended by P i cha d e al. (2000). We applied E anno e al.’s (2005) modal ΔK pa ame e as he choice c i e ion o de ec he ue numbe o gene ic g oups in he se o indi iduals assayed. Resul s Loci Unde Selec ion and Neu al Gene ic S uc u e o Popula ions BayeScan iden i i ed six and se en AFLP ma ke s wi h e idence o pu i ying selec ion in he SE and SW popula ions, espec i ely (3 % o all ma ke s app oxima ely in bo h cases; Fig. 2a). No e ha loci unde pu i ying selec ion in one g oup o popula ions we e no unde de ec able selec ion in he o he g oup (Fig. 2b). Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 70 - Figu e 2. De ec ion o selec ion using a Bayesian genomic scan app oach in SE and SW popula ions o Daphne lau eola (black and g ey do s espec i ely). Each do co esponds o an AFLP ma ke (N = 225). Dashed lines indica e he h eshold alue o he log (PO) beyond which he e is e idence o a locus o be subjec o pu i ying selec ion a a False Disco e y Ra e o 0.001. No e ha a h eshold alue was calcula ed independen ly o SE and SW popula ions. (a) Plo o he es ima ed FST alues agains he (log PO) o each locus wi h labels o he ou lie loci. (b) Compa ison o loci-speci i c pa e ns o pu i ying selec ion (log PO) in SE and SW popula ions. Solid lines join alues o log (PO) o he same locus in SE and SW popula ions. Only loci unde pu i ying selec ion in a leas one g oup o s udied popula ions we e ep esen ed. Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 71 - Bayesian analysis o pu a i ely neu al gene ic a ia ion (i.e. a e emo ing selec ed loci om he da a), which did no ely on p ede e mined assignmen o samples o egions o popula ions, showed a dis inc modal maximum o ΔK a K = 2 gene ic g oups. No e ha SW popula ions o - med a gene ic g oup di e en ia ed om he es o popula ions wi h nea ly all popula ions being assigned exclusi ely o he SW gene ic g oup (Fig. 3). Figu e 3. Gene ic popula ion s uc u e o Daphne lau eola in he Bae ic Ranges in e ed wi h he model-based Bayesian clus e ing implemen ed in STRUCTURE om da a on pu a i ely neu al loci alone. The indi idual pie cha s indica e he mean p opo ion o membe ship o each local popula ion o he in e ed numbe o K = 2 gene ic g oups. Di e en ial Selec ion on Sexes Ve y ew loci we e subjec o pu i ying selec ion in emale plan s. In SE popula ions, he maph o- di e indi iduals p esen ed h ee loci unde pu i ying selec ion in compa ison o only one in emales (Fig. 4 a-c). The di e ence be ween sexes was e en sha pe in SW popula ions, whe e he maph o- di e and emale indi iduals exhibi ed se en and one only locus unde pu i ying selec ion, espec i- ely (Fig. 4d- ). The sepa a e analyses by sex con i med ha he iden i y o he loci unde selec ion on each sex was di e en in SE and SW popula ions (Fig. 4). Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 72 - Figu e 4. Sex- ela ed selec ion pa e ns e ealed by Bayesian genomic scan in he SE (uppe ow) and SW (lowe ow) popu- la ions o he gynodioecious sh ub Daphne lau eola. Wi hin each ow, om le o igh , he i s wo g aphs plo es ima ed FST o each locus agains i s co esponding alue o log (PO) wi h labels o he loci ou lie s, sepa a ely o he maph odi es (a, d) and emales (b, e), espec i ely. The hi d g aph illus a es diff e ences be ween sexes in he log (PO) alues o loci ound subjec o pu i ying selec ion in a leas one sex (c, ). Solid lines join he alues o log (PO) o he same locus in he maph o- di es and emales. Each do co esponds o an AFLP ma ke (N = 225; whi e and black do s, he maph odi es and emales, espec i ely). On each panel, dashed lines indica e he h eshold alue o he log (PO) beyond which he e is e idence o a locus o be subjec o pu i ying selec ion a a False Disco e y Ra e o 0.001. Independen h eshold alues we e calcula ed o each da ase (SE he maph odi es, SE emales, SW he maph odi es, SW emales). Discussion Di e en ial Selec ion Be ween Regions and Neu al Gene ic S uc u e S ong gene ic di e en ia ion among disjunc popula ions is a common ea u e in plan species o Medi e anean moun ainous anges (Schönswe e e al., 2006; Ronikie e al., 2008; Cse gö e al., 2009). This gene ic di e gence a in aspeci i c le el could be he esul o non-selec i e me- chanisms such as andom i xa ion o di e en alleles in disjunc dis ibu ion pa ches because o ounde e en s du ing coloniza ion o gene ic d i a ising om educ ion o popula ion sizes du ing wa me s ages in he Holocene (Young e al., 1996; Thompson, 2005). Con as ing selec i e egimes in disjunc pa s o he species dis ibu ion ange can also lead o s ong gene ic di e en ia ion among disjunc dis ibu ion pa ches (Kawecki, 2008; Sex on e al., 2009). Di e gen selec i e e- gimes in disjunc g oups o popula ions because o a ia ion in abio ic and/o bio ic ac o s ha e Di e en ial selec ion be ween sexes and egions in a gynodioecious sh ub e ealed by genomic scanning. - 73 - been documen ed in some plan species (Allan & Pannell, 2009; He e o d & Winn, 2008; Pa isod & Joos , 2010). Popula ions o D. lau eola in sou he n Ibe ian Peninsula p esen a s ong gene ic di e en ia ion, exhibi ing a clea spli be ween SW and SE popula ions (Alonso & He e a, 2011; Cas illa e al., unpublished da a). Resul s o he p esen s udy sugges ha bo h non-selec i e and selec i e mechanisms ha e mos likely con ibu ed o he s ong gene ic di e gence o D. lau- eola popula ions ac oss he Bae ic Ranges. In pa icula , we ound ha SE and SW popula ions o D. lau eola p esen a simila numbe o loci unde pu i ying selec ion, which may en a i ely be in e p e ed as an indica ion o equi alen le els o selec i e p essu e. This esul con adic s o he expec a ion o ma ginal popula ions ep esen ing a eas wi h s onge selec i e p essu es han co e popula ions (Kawecki, 2008; Sex on e al., 2009). Howe e , he non-o e lapping iden i y o he loci unde selec ion in bo h g oups o popula ions sugges s di e en selec i e egimes in SE and SW popula ions. Reduced l o al display sizes and inc eased ela i e ui se associa ed o a change in he pollina o auna has been epo ed in SW popula ions (Cas illa e al., 2011). S udies o pheno ypic selec ion by pollina o s on l o al ai s in hese wo disjunc g oups o popula ions a e necessa y o unde s and he ole o geog aphic a ia ion in plan -pollina o in e ac ions on he in aspeci i c di e en ia ion o D. lau eola popula ions. Ou esul s also suppo ed a ole o non-selec i e mechanisms in con ibu ing o he s ong ge- ne ic di e en ia ion o he D. lau eola ac oss he Bae ic Ranges. The clea gene ic spli be ween SW and SE popula ions o species p e iously epo ed (Alonso & He e a, 2011; Cas illa e al., unpubl. da a) emained a e emo ing all he loci unde pu i ying selec ion. Range shi s associa ed wi h clima ic oscilla ions du ing he Qua e na y ha e led o andom i xa ion o di e en alleles in dis- junc g oups o popula ions con ibu ing o gene ic in aspeci i c di e en ia ion o plan species (Bi kau & Comes, 2005; Schönswe e e al., 2006). In his sense, Bae ic Ranges eme ge as a moun- ainous ange less a ec ed by he glacia ions du ing he Qua e na y han mo e no he n moun ain sys ems (Bene e al., 1991; Médail & Diadema, 2009) and hus he s ong gene ic di e gence o D. lau eola popula ions could e l ec a long- e m his o y o gene ic isola ion be ween he wo cu- en ly geog aphically disjunc g oups o popula ions he e s udied. Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 80 - Plan -pollina o in e ac ions a e an example o selec ion p essu es expec ed o a y geog aphi- cally ollowing spa ial changes in he abundance and iden i y o he l o al isi o s (Moelle , 2005, 2006; Thompson, 2005). In gene al, small and/o isola ed popula ions a ac ewe and less di- e se pollina o s (G oom, 1998; Sih and Bal us, 1987). Also, ma ginal popula ions ha e been ound o ha e poo e pollina o aunas han cen al popula ions (Sil a-Mon ellano and Eguia e, 2003; S one and Jenkins, 2008). Pollina ion egimes di e ing in composi ion and abundance o pollina- o s can in u n lead o di e en ial ep oduc i e success (He e a e al., 2006; Pé ez-Ba ales e al., 2007), and in he speci i c con ex o he compa ison be ween cen al and ma ginal popula ions, could lead o a highe ep oduc i e success in cen al popula ions. In he p esen s udy, we compa ed cen al con inuous and ma ginal disjunc popula ions o D. lau eola wi hin i s dis ibu ion ange in he sou h o he Ibe ian Peninsula. We used a b oad a ay o popula ion ea u es and plan i ness co ela es, in addi ion o conside a ion o he main pollina- o iden i y and abundance o co- l owe ing species. Daphne lau eola is an e e g een sh ub wi h a wide dis ibu ion ange wi hin Eu ope, g owing in he unde s o y o moun ain o es s and p esen- ing dis ibu ion pa ches on he main calca eous moun ain anges o sou he n Eu ope and no h o A ica (Meusel e al., 1978). In ou s udy a ea, mos popula ions a e gynodioecious wi h he ma- ph odi e and emale plan s coexis ing wi hin popula ions (Alonso e al., 2007). The main pollina o o D. lau eola in he cen al a ea is a small pollen bee le, Melige hes elonga us Rosenhaue (Alon- so, 2004), bu i s ole as pollina o in ma ginal popula ions was unce ain. We analyzed popula- ion ea u es such as popula ion size, densi y o ep oduc i e and non- ep oduc i e indi iduals wi hin popula ion, iden i y o main pollina o s and al e na i e l o al esou ces o he pollina o s, and indi idual li e-his o y ai s including plan size, shoo g ow h, l owe and ui p oduc ion, ui -se , and mo ali y. We u he s udied di e ences be ween emales and he maph odi es in size and ep oduc i e ai s p esuming he exis ence o mo e s ess ul condi ions in ma ginal po- pula ions ha would allow o es he “sex-di e en ial plas ici y hypo hesis” (Case and Ashman, 2007; Do ken and Mi cha d, 2008). A conside able numbe o s udies suppo ha emales a e mo e abundan unde s ess ul condi ions because by in es ing in one only sex unc ion, emales Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 81 - ha e mo e a ailable esou ces o o he unc ions hus inc easing hei ela i e i ness compa ed o he maph odi es ( e iewed in Ashman, 2006). Howe e , acco ding o he sex-di e en ial plas i- ci y hypo hesis (Do ken and Mi cha d, 2008), i he maph odi es a e plas ic in hei in es men in emale ep oduc ion, hey will alloca e ewe esou ces o seed p oduc ion unde s ess ul condi- ions coun e ac ing he emales’ ad an ages. Speci i cally, we add essed he ollowing ques ions: (1) do popula ion size and densi y o indi iduals di e be ween cen al con inuous and ma ginal disjunc popula ions? (2) do plan s in cen al con inuous and ma ginal disjunc popula ions di e in size, g ow h, l owe and ui p oduc ion? (3) do he iden i y o main pollina o and he abun- dance and di e si y o o he co- l owe ing species a y in cen al con inuous and ma ginal disjunc popula ions? and (4) do he maph odi e plan s p esen a lowe seed ou pu han emales in he ma ginal disjunc popula ions? Ma e ial and me hods S udy Species D. lau eola L. (Thymelaeaceae) is an e e g een sh ub wi h a disjunc Palaea c ic dis ibu ion associa ed wi h he main calca eous moun ain anges o Eu ope, and no h o A ica. In he Ibe ian Peninsula i shows a disjunc dis ibu ion, being abundan in he no he n Can ab ian Range and he Py enees, and also in he sou he n Bae ic Ranges (Alonso e al., 2007; Fig. 1a). This s udy was conduc ed in 2007-2009 in he Bae ic Ranges. Plan s o D. lau eola p esen a a iable numbe o b anches ha elonga e o e se e al conse- cu i e g owing seasons wi hou any seconda y b anching. Lea es occu only a he dis al end o b anches, whe e hey o m a well-de i ned ose e, e med “lea who l” he ea e . I has a b oad win e - l owe ing pe iod (Janua y-Ap il) and p oduces a la ge numbe o small, ubula , g een- yellowish l owe s agg ega ed in o se e al compac axilla y in l o escences pe s em. D. lau eo- la p esen s gynodioecious popula ions whe e emale and he maph odi e indi iduals coexis , and popula ions ha ing exclusi ely he maph odi e plan s, he las ones being a e in ou s udy a ea (Alonso e al., 2007). F uc i i ca ion begins in June, p esen ing one only seed pe ui . Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 82 - Figu e 1. (a) Dis ibu ion ange o Daphne lau eola in he Ibe ian Peninsula. (b)Mo e de ailed dis ibu ion o popula ions in he s udy a ea and loca ion o cen al and ma ginal egions de i ned in his s udy. In e als o al i ude (m.a.s.l.) in he s udy a ea a e ep esen ed wi h g ey scale (see legend). S udy a ea Ac oss he Bae ic Ranges, D. lau eola is especially abundan in he Sie a de Cazo la, wi h a leas 88 popula ions highly connec ed among hem loca ed along a b oad al i udinal ange. In ac , conside ing popula ion as a disc e e g oup o indi iduals dis an om o he g oups o indi i- duals by a leas 1 km, mos popula ions o Cazo la ha e > 60 neighbou ing popula ions in a a- dius o 25 km (Cas illa e al., 2011). The geog aphic isola ion among popula ion inc eases owa ds eas e n and wes e n egions o he Bae ic Ranges, being isola ion mo e ma ked in he wes e n popula ions he e analysed (Cas illa e al., 2011). In pa icula , wes e n popula ions o D. lau eola ha e on a e age < 3 neighbou ing popula ions in a adius o 25 km. The e o e, in his s udy we The main pollina o o D. lau eola in popula ions o he Sie a de Cazo la, which cons i u e he cen e o abundance o he species wi hin he s udied sou he n Ibe ian egion (Fig. 1b), is M. elonga- us Rosenhaue , a small pollen bee le ha mo es equen ly among l owe s o he same indi idual (Alonso, 2004). Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 83 - conside ed he Sie a de Cazo la as he local con inuous cen e (“cen al egion”, he ea e ) and popula ions a he wes e n edge as a local disjunc ma gin (“ma ginal egion”, he ea e ) o he species’ dis ibu ion pa ch in sou he n Ibe ian Peninsula (Fig. 1b). We s udied h ee accessible popula ions pe egion. In he cen al egion we used Valdecue as, Cañada del Espino and Fuen e Be mejo, he dis ance among hem being 2-5 km. In he ma ginal egion we used wo di e en moun ain anges. Two popula ions (Fuen e Molina and Cañada de las Animas) we e loca ed in he Sie a de las Nie es, wi h a dis ance be ween hem < 3 km. The hi d ma ginal popula ion was loca ed in he Sie a de G azalema, dis an om he o he wo popula ions abou 35 km. All s udy popula ions we e gynodioecious, and he p opo ion o emale and he maph odi e indi iduals de e mined du ing he l owe ing pe iod o 2007 by examining 100 andomly chosen indi iduals was ca. 20 % in all o hem (Table 1). A he beginning o he s udy we haphaza dly ma ked 25 he maph odi e and 15 emale plan s in each s udy popula ion, o a oid po en ial a e ac s due o uncon olled a ia ion on sample sex a io. Popula ion Fea u es: Size, Plan Densi y and Mo ali y o Indi iduals Popula ions we e classed in o one o he ollowing size ca ego ies: <500, 500-1000, 1000-1500, 1500-2000, >2000 ep oduc i e indi iduals. Densi ies o ep oduc i e and non- ep oduc i e indi- iduals wi hin each popula ion we e es ima ed using 20 andomly loca ed ci cula a eas o 10 m diame e in which he numbe o ep oduc i e and non- ep oduc i e indi iduals we e coun ed. In 2009 summe , he mo ali y o ma ked indi iduals was eco ded. Indi idual Fea u es: Size, Annual Shoo G ow h, Flo al Display and Seed P oduc ion Fou measu es ela ed o indi idual size we e ob ained o e e y ma ked plan : heigh o he alles b anch (“plan heigh ” he ea e ), o al numbe o lea who ls (“numbe o lea who ls” he ea e ) and he majo and he mino diame e o he plan e ical p ojec ion, used o es ima e he a ea o he plan assimila ing i o an ellipse (“plan a ea” he ea e ). Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 84 - Th ee measu es ela ed o l o al display we e also ob ained o e e y ma ked plan : p opo - ion o s ems bea ing l owe s; numbe o in l o escences/s em (based on coun s o i e s ems pe plan ) and numbe o l owe s/in l o escence (coun ed in one in l o escence pe sampled s em). To al l owe p oduc ion pe plan was hen es ima ed as he numbe o l owe ing s ems x a e age num- be o in l o escences/s em x a e age numbe o l owe s/in l o escence. The annual shoo g ow h was es ima ed by measu ing he shoo elonga ion a he end o he g owing season in he i e s ems pe plan used o he l o al display (“shoo g ow h” he ea e ). In addi ion, he o al numbe o ui s p oduced pe indi idual was coun ed (“ ui p oduc ion” he ea e ) and he p opo ion o ui s pe l owe (“ ui se ” he ea e ) was es ima ed using he i e ma ked in l o escences. Main Pollina o and Al e na i e Flo al Resou ces P e ious s udies showed ha he pollen bee le M. elonga us (Ni idulidae) is he main pollina o o D. lau eola in Sie a de Cazo la, de e mining impo an aspec s o species’ ep oduc i e biology (Alonso, 2004; Med ano e al., 2005). Howe e , occasional obse a ions on D. lau eola l owe s du- ing he l owe ing pe iod o 2007 (A.R.C. pe sonal obse a ion) indica ed ha pollen bee les we e appa en ly absen om ma ginal popula ions, whe e se e al species o Hymenop e a, mainly Bombus e es is L, Xylocopa iolacea L. and Apis melli e a L., and ano he bee le Tachypo us ni idulus Fab icius (S aphylinidae) we e obse ed (Figu e 2). The e o e, in 2009 censuses we e conduc ed o de e mine he abundance o M. elonga us in ma ginal popula ions o D. lau eola. The censuses we e done in sunny days. In each census, he obse e was close o he plan and coun ed he numbe o M. elonga us indi iduals sys ema ically by examining all he l owe ing s ems o he plan du ing a maximum in e al o h ee minu es. The numbe o l owe ing s ems, he ime spen and whe he he plan was in sun o shade du ing he obse a ion we e eco ded. Fo y indi iduals pe popula ion we e censused in all cases. In ma ginal popula ions, censuses we e conduc ed in h ee di e en da es co e ing all he l owe ing pe iod (ea ly, peak and la e l owe ing) o exclude he possibili y o a misma ch be ween he p esence o M. elonga us and he D. lau eola’s l owe ing peak. Fo compa ison wi hin he same season, plan s in cen al popula ions we e simila ly censused al hough only in a single da e a l owe ing peak. We censused he same Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 85 - numbe o indi iduals in ma ginal and cen al popula ions, in es ing 168 minu es in ma ginal po- pula ions du ing he l owe ing peak and al oge he 539 and 304 minu es in ma ginal and cen al popula ions, espec i ely. In addi ion, he abundance o o he l owe ing species which could po en ially sha e pollina o s wi h D. lau eola was es ima ed by coun ing he numbe o l owe ing indi iduals o each species wi hin i e me e s a ound e e y censused plan . Figu e 2. Pollina o assemblage in ma ginal popula ions o Daphne lau eola in sou he n Ibe ian Peninsula. (a) Bombus e es is; (b) Xylocopa iolacea; (c) Apis melli e a. a b c Da a Analyses All s a is ical analyses we e pe o med using he SAS s a is ical package (SAS, Ins i u e 2002). In o de o compa e ma ginal and cen al popula ions, es s conside ing he e ec s o egion (cen al s. ma ginal) and sex ( emale s. he maph odi e) as i xed ac o s and he e ec o popula ion as andom, we e conduc ed. Densi y o ep oduc i e and non- ep oduc i e indi iduals, plan a ea, plan heigh , numbe o lea who ls, l owe p oduc ion, shoo g ow h, ui p oduc ion and ui se we e he esponse a iables analysed sepa a ely. F ui se was no mally dis ibu ed and plan heigh was no malized a e being log ans o med. Bo h esponses we e analyzed by mixed mo- dels (P ocedu e MIXED). We analyzed all he o he a iables wi h gene alized linea mixed models (P ocedu e GLIMMIX), using he nega i e binomial unc ion. Di e ences among egions in he e- Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 86 - la ionship be ween l owe p oduc ion and indi idual size (plan a ea and numbe o lea who ls pe plan ) we e s udied wi h an analysis o co a iance (ANCOVA). Flowe p oduc ion was he es- ponse a iable, plan a ea, and numbe o lea who ls we e he explana o y a iables and egion was he ca ego ical e ec . In addi ion, we explo ed po en ial ade-o s be ween ep oduc ion and shoo g ow h wi h an analysis o co a iance, dis inguishing shoo g ow h as he esponse a ia- ble, l owe and ui p oduc ion as he explana o y a iables and egion as he ca ego ical e ec . Finally we used Fishe ’s exac es o analyze di e ences in he equency o dead indi iduals be - ween cen al and ma ginal popula ions. Means a e gi en ± 1 SE. Resul s Popula ion Fea u es: Size, Plan Densi y and Mo ali y o Indi iduals All he popula ions s udied had mo e han 500 ep oduc i e indi iduals (Table 1). The wo la - ges popula ions occu ed in he cen al egion, wi h > 2000 ep oduc i e plan s. The wo smalles popula ions, wi h 500-1000 ep oduc i e indi iduals we e ound one in he cen al egion and he o he in he ma ginal egion. Table 1. Popula ion ea u es o he six Daphne lau eola popula ions s udied. Popula ion Coo dina es Region Al i ude (m.a.s.l.) Size (No. ep oduc i e indi iduals) Female equency (%) Valdecue as 37º 54’ N 2º 52’ W Cen al 1575 >2000 25.5 Cañada del Espino 37º 55’ N 2º 53’ W Cen al 1513 500-1000 20 Fuen e Be mejo 37º 55’ N 2º 50’ W Cen al 1380 >2000 21 G azalema 36º 46’ N 5º 25’ W Ma ginal 1229 1500-2000 10 Fuen e Molina 36º 41’ N 5º 1’ W Ma ginal 1380 1000-1500 20 Cañada de las Animas 36º 42’ N 5º 1’ W Ma ginal 1333 500-1000 18.5 Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 87 - The e was ma ginally signi i can a ia ion among egions o plan densi y wi hin popula ion (F1,106 = 3.86, P = 0.05). Ma ginal popula ions we e dense han cen al popula ions (0.09 ± 0.01 and 0.06 ± 0.01 indi iduals / 100 m2 espec i ely). Densi y o ep oduc i e indi iduals did no di e signi i can ly be ween egions (F1,118 = 2.48, P > 0.1; Fig. 2). Howe e , densi y o non- ep oduc i e indi iduals in ma ginal popula ions was la ge han in cen al popula ions (F1,118 = 13.84, P < 0.001; Fig. 2). The pe cen age o dead indi iduals a e wo yea s was also la ge in ma ginal han in cen al popula ions (17 % and 2 % espec i ely; Fishe ’s exac es : P < 0.0001). Figu e 3. Regional di e ences in he densi y o ep oduc i e (a) and non- ep oduc i e indi iduals o Daphne lau eola ba- sed on da a o h ee popula ions pe egion and 20 plo s pe popula ions. In each plo , ep oduc i e and non- ep oduc i e indi iduals we e coun ed in a ci cula a ea wi h a diame e o i e me e s. Ba s ep esen he leas -squa e means (± s.e.) Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 88 - Indi idual Fea u es: Plan Size, Shoo G ow h, Flo al Display and F ui P oduc ion Plan s o cen al popula ions we e la ge han hose g owing in ma ginal popula ions. The a e- age a ea o indi iduals in cen al popula ions was mo e han wice he a ea o plan s in ma ginal popula ions (F1,232 = 17.65, P < 0.0001; Fig. 3a). Also he numbe o lea who ls was mo e han h ee imes la ge in cen al popula ions (F1,232 = 16.92, P < 0.0001; Fig. 3b). Plan s in ma ginal popula ions ended o be alle han hose in cen al popula ions bu hei heigh a e ages did no di e signi i can ly (84.72 ± 2.44 cm s. 80.22 ± 2.28 cm; F1,232 = 0.71, P = 0.40). Shoo g ow h was signi i can ly la ge in ma ginal popula ions (5.9 ± 0.2 cm s. 3.9 ± 0.1 cm; F1,229 = 24.07, P < 0.0001). Also, shoo g ow h was nega i ely ela ed o l owe and ui p oduc ion (F1,235 = 36.21, P < 0.0001 and F1,235 = 9.01, P < 0.01, espec i ely). The ANCOVA e ealed ha nega i e ela ionships be ween g ow h and l owe and ui p oduc ion we e consis en in he wo s udy egions (P = 0.28 and P = 0.29 o he in e ac ion be ween egion and l owe and ui p oduc ion, espec i ely). Fu - he mo e, he e we e no signi i can di e ences be ween he maph odi e and emale plan s in any size- ela ed a iable (P > 0.5 in all cases). Absence o di e en ial size be ween sexes was consis en be ween he wo s udy egions (P > 0.5 in he in e ac ion egion × sex o all size a iables). Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 89 - Figu e 4. Di e ences in a e age plan a ea (a), numbe o lea who ls (b) and l owe p oduc ion pe plan (c) in cen al and ma ginal popula ions o Daphne lau eola, based on da a o h ee popula ions pe egion and 40 indi iduals pe popula ion. Ba s ep esen he leas -squa e (+ s.e.) a e accoun ing o sex ( i xed e ec ) and popula ion ( andom e ec ) a ia ion. Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 96 - adul s and highe densi y o non- ep oduc i e indi iduals in ma ginal popula ions would sugges a highe u no e in hese popula ions. A highe demog aphic a iabili y in ma ginal popula ions o plan species has been ound by se e al s udies (Lönn and P en ice, 2002; Moo e, 2009; Nan el and Gagnon, 1999). Howe e , mo e demog aphic analyses a e equi ed o de e mine i his is a common pa e n in plan s and o in es iga e he possible unde lying ac o s. Indi idual Fea u es: Plan Size, Annual G ow h, Flo al Display and Rep oduc i e Ou pu Adul plan s o D. lau eola a y widely in size and a chi ec u e h ough he Ibe ian Peninsula (Alonso e al., 2007). The p esen s udy ound ha in i s sou he n dis ibu ion a ea, indi iduals o cen al popula ions we e la ge han hose in ma ginal popula ions, which had as a di ec conse- quence a huge a ia ion in l owe p oduc ion. Plan s in cen al popula ions p oduced one o de o magni ude mo e l owe s pe indi idual han in ma ginal popula ions, because hey consis en ly p oduced mo e ep oduc i e s ems, mo e in l o escences pe s em and mo e l owe s pe in l o es- cence. These esul s sugges ed mo e op imal condi ions o he species in he cen al loca ions, u - he suppo ed by he s eepe slope o he ela ionship be ween l owe p oduc ion and indi idual size in cen al popula ions. We mus poin ou , howe e , ha shoo g ow h was lowe in cen al popula ions, likely showing a ade-o be ween ep oduc ion and g ow h wi hin he season wi h di e en ou comes in cen al and ma ginal popula ions (Obeso, 2002 and e e ences he ein; bu see also Knops e al., 2007). Shoo g ow h was nega i ely ela ed o l owe and ui p oduc ions. Plan s o cen al popula ions p oduced mo e l owe s and ui s han plan s o ma ginal popula- ions, which could ha e cons ained mo e he shoo g ow h in indi iduals o cen al popula ions. Ecological s ess in dis ibu ion bo de s can also esul in di e en ial ep oduc ion be ween cen al and ma ginal popula ions (Henge eld and Haeck, 1982; B own, 1984; Law on, 1993). We ound ha D. lau eola indi iduals o cen al popula ions p oduced mo e ui s han hose on ma - ginal popula ions, associa ed wi h he di e ences in size and l owe p oduc ion men ioned abo e. The connec ion be ween he numbe and densi y o adul indi iduals in popula ions and hei e- Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 97 - p oduc i e ou pu was no , howe e , suppo ed because cen al and ma ginal popula ions did no di e in size o densi y o adul plan s (see also Samis and Ecke , 2007). The expec a ion ha popula ion size and densi y o indi iduals a e mainly de e mined by indi idual seed p oduc ion, howe e , could be un ealis ic in some ci cums ances. Likely he e a e o he ac o s cons aining seedling es ablishmen , such as abundance and quali y o a ailable si es o es ablishmen and u he pos dispe sal ac o s ha a ec seed su i al, ge mina ion and seedling su i al (Ha pe , 1977; Cla k e al., 2007) ha could be mo e de e minan o popula ion size and densi y o indi i- duals. Finally, acco ding o he sex-di e en ial plas ici y hypo hesis (Do ken and Mi cha d, 2008), i he he maph odi es a e plas ic in hei in es men on emale ep oduc ion hey will alloca e ewe esou ces o he seed p oduc ion unde mo e s ess ul condi ions, and subsequen ly hei ela i e emale i ness should change. Ou esul s, howe e , indica ed ha he maph odi e indi iduals did no p oduce less ui s han emale indi iduals in D. lau eola cen al o ma ginal popula ions, sugges ing ei he he absence o ecological s ess in bo h s udy egions o a simila esponse o he wo sexes. Pollina ion En i onmen Gi en he synch onic l owe ing o D. lau eola indi iduals (Alonso, 2004), he much highe l owe p oduc ion eco ded in cen al popula ions may ha e consequences o he in e ac ion wi h polli- na o s, since l o al display can in l uence bo h he a ac ion and he beha iou o he pollina o s (Eckha , 1991; Ashman and King, 2005). In addi ion, poo e pollina o aunas a e expec ed in ma - ginal popula ions (Moelle , 2006; Sil a-Mon ellano and Eguia e, 2003; S one and Jenkins, 2008). P e ious s udies conduc ed in he cen al egion documen ed bo h high sel i ng a es in he ma- ph odi es and s ong inb eeding dep ession om seedling o ep oduc i e s age in D. lau eola po- pula ions cha ac e ized by indi iduals wi h la ge l o al displays (Med ano e al., 2005) and he po- llen bee le M. elonga us as he main pollina o (Alonso, 2004). The educ ion o indi idual size and l o al display in ma ginal popula ions could he e o e ha e some consequences o he maph odi e Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 98 - ep oduc ion h ough seed p oduc ion, associa ed wi h a dec ease in gei onogamy (Klinkhame and de Jong, 1993). The inc eased ui -se eco ded in ma ginal popula ions sugges ed ha pollina o se ices migh be imp o ed in hese si es. A possible explana ion could be ela ed o di e ences in he iden i y o he main pollina o in cen al and ma ginal popula ions. Pollina o censuses in 2009 con i med ha M. elonga us was he main pollina o in he cen al popula ions s udied, de ec ing i s p esence in all popula ions, wi h an a e age o 2 bee les/plan . In con as , M. elonga us was no ound in any census a he ma ginal popula ions s udied, despi e conside ably highe and pheno- logically wide sampling e o han in cen al popula ions. Addi ionally, se e al species o Hyme- nop e a as Bombus e es is, Xylocopa iolacea and Apis melli e a we e obse ed isi ing l owe s o D. lau eola a he ma ginal popula ions. Thus, he absence o a key pollina o ha p omo es gei onogamy as M. elonga us and he exis ence o o he mo e dynamic isi o s did no suppo he p edic ion o poo e pollina o aunas a ma ginal popula ions and sugges ed ha he la ge ui se in ma ginal popula ions could be ela ed o changes in he pollina o auna po en ially leading o educed gei onogamy. A mo e exhaus i e sampling e o is equi ed o quan i y he abundance o bees and bumblebees as pollina o s o ma ginal popula ions o D. lau eola and hei consequen- ces o ealized ma ing sys em. Fu he mo e, he di e si y o al e na i e l o al esou ces o insec s was highe in cen al popula ions, wi h up o six di e en co- l owe ing species in he same popu- la ion. In his con ex , D. lau eola l owe s appea o be a low- alued esou ce o mos day- l ying l o al isi o s in cen al popula ions, whe e p e ious s udies and anecdo al obse a ions showed ha bumblebees and bees only excep ionally isi ed D. lau eola l owe s and p e e ed isi ing o he coexis ing l owe ing species (Alonso, 2004). In con as , pollina o s in ma ginal popula ions sca - cely ha e al e na i e l o al esou ces, mainly Hellebo us oe idus in low abundance, which could d i e hem o isi a less a ac i e species such as D. lau eola. This shi o pollina o s, oge he wi h he inc eased ui se in ma ginal popula ions despi e a lowe l o al display, sugges a mo e e ec i e pollina ion en i onmen in ma ginal popula ions. In e es ingly, one o he cen al popula- ions p esen ed a e y poo l owe ing communi y, simila o ma ginal popula ions. Fu he s udies in popula ions wi h poo win e co- l owe ing communi ies could be use ul o es i indi iduals o D. lau eola a e isi ed mo e equen ly by bees and bumblebees in hem. Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 99 - To summa ize, ma ginal popula ions o D. lau eola in he sou he n Ibe ian Peninsula did no di e signi i can ly om cen al ones in size and densi y o indi iduals. In addi ion, he lowe ep o- duc i e in es men o ma ginal popula ions and hei highe ui se could be ela ed o obse ed a ia ion in plan -pollina o in e ac ions, which could be ele an o he sexual polymo phism in his gynodioecious species i le els o gei onogamy we e educed in ma ginal popula ions. Di e- ences in ep oduc i e success be ween cen al and ma ginal popula ions o D. lau eola seemed he e o e mo e closely ela ed o a ia ion in bio ic selec i e p essu es han o changes in he spe- cies’ abio ic niche. Addi ional wo k is cu en ly unde way ocusing on he e ec s o changes in he pollina o communi y on l o al ai s and he gene ic s uc u e o ma ginal and cen al popula- ions, which may in o m abou he impo ance o o he mechanisms po en ially in ol ed in he delimi a ion o dis ibu ion anges (e.g. limi ed gene l ow, uns able me apopula ion dynamics). These s udies should con ibu e o a be e unde s anding o he e olu iona y po en ial o ma gi- nal disjunc popula ions. Acknowledgmen s We a e g a e ul o Me cedes Pé ez o he en husias ic suppo in he i eld wo k. Mónica Med a- no, José L. Ga ido, Cla a de Vega and wo anonymous e iewe s p o ided use ul commen s and c i icisms on he manusc ip . Raimundo Ou e elo iden i i ed Tachypo us ni idulus. Also we a e g a- e ul o Conseje ía de Medio Ambien e, Jun a de Andalucía o pe mission o wo k wi hin Na u al Pa ks and o he en i onmen agen s, especially o Manolo Cal illo, Paco and Da id, o acili ies p o ided du ing i eld wo k. This s udy was unded by he Spanish Minis e io de Educación y Cien- cia h ough esea ch p ojec CGL2006-01355, he Consejo Supe io de In es igaciones Cien í i cas (CSIC) h ough an I3P ellowship o A.C., and he Conseje ía de Inno ación Ciencia y Emp esa, Jun a de Andalucía, h ough he esea ch p ojec RNM156-2005. This pape is dedica ed o he memo y o Miguel Blázquez. Explo ing local bo de s o dis ibu ion in he sh ub Daphne lau eola: indi idual and popula ion ai s. - 100 - Cas illa, A.R., Alonso, C. and He e a, C.M. In e iew. Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub. Chap e 4.- Reduced gende di e gence in fl o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub Abs ac : In sexually polymo phic plan species he ex en o gende di e gence in l o al mo phology and phenology a e expec ed o be in l uenced by gende diff e ences in pollina ion success and pheno ypic selec ion by po- llina o s. Ma gins o dis ibu ion cons i u e na u al a eas whe e changes in he pollina o auna and hus a ia ion in gende di e gence o l o al ai s a e expec ed. He e we indi ec ly assessed he eff ec s o a change in he iden i y o main pollina o occu ing in a disjunc g oup o ma ginal popula ions o he gynodioecious sh ub Daphne lau eola by analyzing gende ela i e pollina ion success and he gende -speci i c selec ion pa e ns o pollina o s on l o al ai s and l owe ing phenology. We ound ha pollina o s deposi ed la ge pollen loads on indi iduals o ma ginal popula ions, con a y o he expec a ion o less effi cien pollina o se ice on dis ibu ion ma gins, and diff e ences be ween gende s in he e- quency o isi ed l owe s also anished. Fu he mo e, he maph odi es and emales ecei ed simila quali y pollen in ma ginal popula ions con as ing wi h co e popula ions, whe e emales ecei ed highe quali y pollen han he - maph odi es. Pollina o s selec ed longe l owe s only on he maph odi e indi iduals o co e popula ions. In addi ion, he maph odi es o co e popula ions wi h longe l owe s p esen ed g ea e ui se s and a ma ginally signi i can nega i e associa ion wi h a locus unde pu i ying selec ion. Consis en wi h hese selec ion pa e ns, he gende di e gence in he longi ude o l owe ube was g ea e in co e popula ions. On he o he hand, pollina o s selec ed also owa ds la e l owe ing on he maph odi e indi iduals o ma ginal popula ions, being his phenological ai nega i ely ela ed o ano he locus unde pu i ying selec ion. Howe e , a geog aphic a ia ion in his phenological ai cong uen wi h his pheno ypic selec ion pa e n was no obse ed. Ou esul s suppo ed ha geog aphic a ia ion in he composi ion o he pollina o aunas could lead o changes in bo h quali a i e and quan i a i e componen s o pollina ion success and o diff e ences in he gende -speci i c se- lec ion pa e ns by pollina o s. These changes can ha e a ele an ole in he geog aphic a ia ion o he gende di- e gence o l o al ai s. The e o e, he esul s o he p esen s udy suppo ed ha ma ginal disjunc popula ions. Re- duced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub cons i u e na u al a eas whe e changes in he e olu iona y dynamics o plan pollina o in e ac ions can be expec ed, especially in sexually polymo phic plan species. Resumen Las di e encias en e plan as de dis in o sexo en el éxi o de la polinización y en los pa ones de selección eno ípica eje cidos po los polinizado es pueden de e mina el g ado de di e gencia en e sexos en los asgos lo ales y en la enología de la lo ación en especies ege ales con sis emas sexuales polimó icos. Los má genes de dis ibución cons i uyen á eas na u ales donde se espe an cambios en la auna de polinizado es y po an o, donde se espe a di e gencia en e plan as de dis in o sexo en los asgos lo ales. En es e es udio, noso os hemos e aluado indi ec amen e el e ec o del cambio en la iden idad del polinizado p incipal en poblaciones ma ginales del a bus o ginodioico Daphne lau eola. Pa a ello, hemos analizado las di e- encias en e plan as de dis in o sexo en el éxi o de polinización y su elación con los asgos lo ales y la enología de lo ación. En es e abajo hemos encon ado que los polinizado es en las poblaciones ma ginales deposi a on ca gas polínicas ma- yo es en los indi iduos, con a io a la expec a i a de un se icio de polinización menos e icien e en los má genes de dis i- bución. Los polinizado es isi a on con simila ecuencia las lo es de indi iduos hemb as y he ma odi as. Los indi iduos he ma odi as y hemb a ecibie on polen de simila calidad en las poblaciones ma ginales en con as e con las poblaciones cen ales, donde los indi iduos hemb as ecibie on polen de mejo calidad que los he ma odi as. Po o o lado, los poliniza- do es selecciona on posi i amen e lo es más la gas sólo en los he ma odi as de las poblaciones cen ales. Además, los he - ma odi as de las poblaciones cen ales con lo es más la gas ambién p esen a on mayo p opo ción de lo es p oduciendo u os. La a iación en la longi ud del ubo lo al es u o elacionada nega i amen e, de mane a ma ginal, con la p esencia de un locus con signos de es a bajo selección pu i icado a. Consis en e con es os pa ones de selección, encon amos que la di e gencia en e plan as de dis in o sexo con espec o a la longi ud del ubo lo al ue mayo en las poblaciones cen ales. En las poblaciones ma ginales, los polinizado es selecciona on posi i amen e enologías de lo ación más emp anas sólo en los indi iduos he ma odi as. Además la a iación en es e asgo enológico es u o elacionada nega i amen e con o o locus con signos de selección pu i icado a. Sin emba go, la a iación geog á ica en es e asgo no ue cong uen e con el pa ón de selección eno ípica obse ado. Nues os esul ados apoya on la idea de que los cambios en la composición de la auna de polinizado es dan luga a a- iación en los componen es cuan i a i o y cuali a i o del éxi o de la polinización; y a di e encias en los pa ones de selección eje cidos po los polinizado es en e plan as de dis in o sexo. Es os cambios pod ían ene una especial ele ancia sob e la a iación geog á ica de la di e gencia de los asgos lo ales en e plan as de dis in o sexo. Po an o, los esul ados de es e es udio señalan las poblaciones ma ginales disjun as de las especies ege ales como á eas na u ales impo an es pa a el es udio de los cambios en las dinámicas coe olu i as de las in e acciones plan a-animal, especialmen e en plan as con sis emas polimó icos sexuales. In oduc ion Geog aphic a ia ion in he composi ion o he pollina o auna in plan popula ions may lead o impo an changes in he pollina ion success o indi iduals h ough a ia ion in quan i y and/ o quali y o pollen ecei ed (He e a 1987; He e a 2004; Gómez e al. 2010; Alonso e al. 2011). Va ia ion in he pollina ion success may ha e a ele an ole in he e olu ion o sexually polymo - phic sys ems in plan species (Ba e 2002; Ashman 2006). Fo ins ance, di e gence in pollina ion success be ween sex mo phs in gynodioecious species can acili a e he sp eading o emale indi- iduals h ough inc eased numbe and/o be e quali y o hei o sp ing (Cha leswo h & Cha - leswo h 1978). Many gynodioecious species p esen sel -compa ible he maph odi es (Meaghe 2007; Ehle s & Schie up 2008). As a consequence, highe quali y seed p oduc ion by emales can esul om he inabili y o ha mo ph o sel - e ilize, he eby limi ing inb eeding in hei o sp ing (e.g. Thompson & Ta ay e 2000; Delph 2004; Chang 2007), ha linked o s ong inb eeding dep es- sion may con ibu e o emale pe sis ence wi hin popula ions (Cha leswo h & Cha leswo h 1978; Med ano e al. 2005; Ramsey e al. 2006). Habi a agmen a ion simul aneously a ec s a a ie y o abio ic and bio ic ac o s, which di- ec ly o indi ec ly can change he abundance and composi ion o pollina o s (S e an-Dewen e & Tscha n ke 1999; Knigh e al. 2005; González-Va o e al. 2009). In his sense, ecological ma - gins in plan dis ibu ions cons i u e a eas whe e popula ions a e geog aphically mo e isola ed and, hus, whe e shi s owa ds poo e pollina o auna may be expec ed (Sil a-Mon ellano & Eguia e 2003; S one & Jenkins 2008). Di e en pollina o aunas may exe di e en pa e ns o pheno ypic selec ion on lo al o phenological ai s (He e a e al. 2006; Ha de & Johnson 2009). This di e en ial selec ion could lead o in aspeci ic pheno ypic di e en ia ion and ul ima ely o specia ion e en s (Johnson 2006). Howe e , ew s udies ha e analyzed speci ically he causal ole o he spa ially a iable selec ion by pollina o s on he geog aphical di e en ia ion in lo al ai s o plan species (He e a e al. 2006), despi e mos o hem ha e ound a ele an ole o di e gen selec ion by pollina o s on in aspeci ic lo al di e en ia ion o plan species (e.g. To land 2001; Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 104 - He e a e al. 2006; Gómez & Pe ec i 2010). Cla i i ca ion o he ole o pollina o s in he e olu- ion o sexually polymo phic sys ems awai s s udies ha documen i changes in pollina o auna composi ion a ec di e en ly o indi iduals wi h di e en sex in hei pollina ion success and hei pheno ypic ai s (Ashman 2006). Flowe s o angiospe ms wi h unisexual l owe s equen ly exhibi dimo phism in seconda y sex cha ac e s (Eckha 1999), including gende di e ences in l owe size, l owe shape, nec a p oduc- ion, l owe longe i y, l o al scen o l owe ing phenology (Eckha 1999; Asikainen & Mu ikainen 2005; Collin & Shyko 2010). In gynodioecious species, whe e he maph odi e and emale plan s occu in popula ions, l owe size is almos uni e sally la ge in he maph odi es han emales (Ec- kha 1999; Shyko e al. 2003; bu see Delph & Li ely 1992). Gende di e gence has been also con i med in l owe ing phenology o gynodieocious species. Female indi iduals equen ly l owe somewha ea lie han he maph odi es (Uno 1982; Ashman & S an on 1991; Asikainen & Mu ikai- nen 2005). Bo h l owe size and l owe ing phenology may be la gely in l uenced by pollina o ’s selec ion pa e ns (Campbell 1989; Galen 1989; Johns on 1991; Conne 1996). The e o e, he le els o gende di e gence in l owe size and l owe ing phenology a e expec ed o a y in popula ions wi h di e en pollina o aunas. In his s udy, we analyzed he geog aphic a ia ion in pollina ion success and gende di e gen- ce in l o al mo phology and l owe ing phenology o he long-li ed gynodioecious sh ub Daphne lau eola. We compa ed co e con inuous and ma ginal disjunc popula ions in sou he n Ibe ian Peninsula which di e in he iden i y o hei main pollina o (Alonso 2004; Chap e 3). The pe sis- ence and equency o emale indi iduals in co e con inuous popula ions is accoun ed o a com- bina ion o ex ensi e sel i ng in he maph odi es and s ong inb eeding dep ession in hei o sp ing (Med ano e al. 2005). In addi ion, he maph odi es o D. lau eola p esen longe l owe s and s a hei l owe ing ea lie han emales in co e popula ions (Alonso 2004; Alonso 2005), bu he le el o spa ial a ia ion o gende di e gence has no been analysed in a b oade geog aphic con ex . In addi ion, a p e ious wo k has shown he exis ence o di e en loci unde pu i ying selec ion Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 105 - be ween ma ginal disjunc and he emaining popula ions o sou he n Ibe ian Peninsula (Chap e 2). In he p esen s udy, we add essed speci i cally he ollowing ques ions: (1) is he pollina ion success in i s quan i a i e and quali a i e componen s lowe in ma ginal disjunc popula ions o D. lau eola in sou he n Ibe ian Peninsula?, (2) is he le el o gende di e gence in fl o al and phenological ai s di e en in co e and ma ginal disjunc popula ions?, (3) is he e di e en ial pheno ypic selec ion on fl o al and phenological ai s in con inuous co e and ma ginal disjunc popula ions o D. lau eola?, (4) do pa e ns o pheno ypic selec ion by pollina o s ma ch o geog aphic a ia ion in he le el o gende di e gence in fl o al ai s?, and (5) a e fl o al and phenological ai s associa ed o loci unde pu i ying selec ion? Ma e ial and Me hods S udy Species Daphne lau eola L. (Thymelaeaceae) is an e e g een sh ub wi h a disjunc Palaea c ic dis ibu- ion associa ed wi h he main calca eous moun ainous anges o Eu ope and no he n A ica. In he Ibe ian Peninsula, i shows a disjunc dis ibu ion, being abundan in he no he n Can ab ian Range and he Py enees and also in he sou he n Bae ic Ranges (Fig. 1a; Alonso e al. 2007). This s udy was conduc ed in he Bae ic Ranges (Fig. 1b), whe e he species l owe s om Janua y o Ap il, a pe iod cha ac e ized by low empe a u es, equen ains and i egula snow alls. Each plan p oduces a la ge numbe o small, ubula , g een-yellowish l owe s agg ega ed in o se e al compac in l o escences pe s em. The wo sexes a e simila in size, l owe and ui p oduc ion (Alonso & He e a 2001). Indi idual l owe s o bo h gende s ha e a single o ule and emain open con inuously du ing app oxima ely one mon h. Flowe s consis o a co olla ube (4-12 mm) wi h ou lobes and bea eigh s amens a anged in wo who ls o ou s amens each. He maph odi e plan s a e ully sel -compa ible, bu ui p oduc ion equi es l owe isi a ion by pollina o s and an excess o pollen may clog-up he s igma o he maph odi es, educing hei ui se (Alonso & He e a 2001). The small bee le Melige hes elonga us Rosenhaue cons i u e he main pollina o in Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 112 - Resul s Quan i y and Quali y Componen s o Pollina ion Success He maph odi e l owe s we e quan i a i ely mo e success ul han emales bu he magni ude o di e gence be ween he wo sexes was highly di e en in he wo egions (Fig. 2). Flowe s o he - maph odi e indi iduals ecei ed la ge mean pollen loads han emales and he magni ude o he di e ence was highe in ma ginal popula ions (F1,1865 = 16.14, P < 0.0001 egion x sex in e ac ion; Figu e 2a). Fu he mo e he di e ence be ween sexes in he mean pe cen age o isi ed l owe s was widely di e en among egions (F1,227 = 89.20, P < 0.0001 egion x sex in e ac ion), because he - maph odi es had la ge mean pe cen age o isi ed l owe s han emales only in co e popula ions (Fig. 2b). Figu e 2. Va ia ion in a e age pollina ion success (± SD/SE) be ween gende s (he maph odi es s. emales) and egions (co e s. ma ginal) o Daphne lau eola in sou he n Ibe ian Peninsula. (a) (a) Pollen load e e s o he a e age numbe o pollen g ains deposi ed on he s igma. (b) Visi ed l owe s e e o he pe cen age o l owe s wi h a leas one pollen g ain eco ded om a sample o 12 l owe s pe plan . In bo h panels N = X emales and XX he maph odi es pe egion. Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 113 - Piecewise analyses u he indica ed ha isi ed emale l owe s we e mo e equen ly limi ed by quan i y han he maph odi e l owe s (�22 = 114.39, P < 0.0001 and �22 = 1226.25, P < 0.0001, o co e and ma ginal popula ions), al hough wi h some ele an di e ences be ween he wo egions. In co e popula ions quan i a i e limi a ion was mo e equen in emales bu ela i ely common in he wo sexes, wi h 64.6 and 36.8 % samples alling below he lowe BCa. In ma ginal popula ions emale l owe s we e simila ly a ec ed by quan i y limi a ion (66.75 %) bu he e- quency o quan i y limi a ion in he maph odi e l owe s was much lowe (6.43 %). As ega ds limi- a ion by quali y, only 11.8 and 11.2 % o isi ed emale l owe s all abo e he highe BCa in co e and ma ginal popula ions, espec i ely. Howe e , 85.4 % o he maph odi e l owe s in ma ginal popula ions we e limi ed by quali y, in con as o 50.6 % eco ded in co e popula ions. Fu he mo e, in co e popula ions he slope o he ini ial ela ionship (b1) was nea ly wice as high in emales as in he maph odi es whe eas in ma ginal popula ions i was simila be ween sexes and subs an ially lowe in he emales, hus, indica ing a educ ion in he quali y o po- llen ecei ed by emales o ma ginal popula ions (Table 1 and Figu e 3). Finally, he slope o he second ela ionship (b2) did no di e om 0 in any case (Table 1 and Figu e 3). This las esul sugges s ha he s yles we e ully sa u a ed wi h pollen ubes, al hough he numbe o ubes co esponding o he b eakpoin was 16 in he maph odi es and 6 in emales o co e popula ions, and 8 in he maph odi es and 13 in emales o ma ginal popula ions; always highe han he single o ule pe l owe . Region Sex b1 B eakpoin (BCa) b2 Co e Female 0.55 11 (4.51-19.34) 0.12 He maph odi e 0.34 45.96 (37.06-55.37) -0.001 Ma ginal Female 0.26 50.52 (29.27-71.26) -0.058 He maph odi e 0.31 24.80 (19.74-43.34) 0.002 Table 3. Pa ame e s o piecewise eg ession analysis o he ela ionship be ween numbe s o pollen g ains and pollen ubes o he maph odi e and emale indi iduals in co e and ma ginal popula ions o Daphne lau eola. The b eakpoin indica es he numbe o pollen g ains a which he slope o he ela ionship changes. Boo s apping Con i dence in e als (BCa) we e es ima ed wi h n = 1000 boo s aps. The slopes o he ela ionship be o e and a e he b eakpoin a e ep esen ed by b1 and b2, espec i ely. Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 114 - Figu e 3. Piecewise eg ession o he ela ionship be ween he numbe o pollen g ains and pollen ubes o emales and he - maph odi es o D. lau eola in co e and ma ginal popula ions (solid and dashed lines espec i ely). Fo compa ison, he ou piecewise models a e p esen ed on a common scale. Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 115 - In aspeci i c Flo al T ai s Di e en ia ion He maph odi e l owe s we e longe and wide han emale l owe s. Flo al ubes we e longe in he maph odi es han emales, al hough he magni ude o he gende di e gence was lowe a ma ginal popula ions (F1,2392 = 30.63, P < 0.0001; egion x sex in e ac ion; Fig. 4a). In pa icula , he maph odi e l owe s o ma ginal popula ions had sho e ubes han hose o co e popula ions (Fig. 4a), whe eas emale l owe s we e simila be ween egions (Fig. 4a). Flowe s o he maph odi- es had g ea e mean ou e diame e han hose o emales (F1,2321 = 174.08, P < 0.0001; Fig. 4b). The magni ude o he di e ence be ween bo h sexes was simila in co e and ma ginal popula ions (P = 0.8; egion x sex in e ac ion; Fig. 4b). Flowe s o he maph odi es showed la ge mean ape u e o ube han hose o emales in bo h co e and ma ginal popula ions (F1,2377 = 491.46, P < 0.0001; Fig. 4c). The a e age alues o ape u e o ube and he magni ude o he di e ence be ween bo h sexes was simila be ween co e and ma ginal popula ions (F1,2377 = 2.22, P = 0.1; egion x sex in e ac ion). In addi ion, l owe s o he wo sexes had g ea e mean ape u e o ube in ma ginal han in co e popula ions (F1,2377 = 52.95, P < 0.0001; Fig. 4c). Gende di e gence in phenological ai s was di e en in co e and ma ginal popula ions o D. lau eola. Sexes di e ed in he ime o l owe ing in ma ginal popula ions, whe e emales we e delayed, bu no in co e popula ions (F1,232 = 25.66, P < 0.0001, in e ac ion egion x sex; Fig. 4d). Also, indi iduals in co e popula ions had mo e open l owe s pe in l o escence a l owe ing peak han hose o ma ginal popula ions. Wi hin-plan asynch ony was highe in co e popula ions and he sign o gende di e gence changed be ween egions (F1,232 = 22.06, P < 0.0001; in e ac ion e- gion x sex). He maph odi es p esen ed g ea e a ia ion in he numbe o open l owe s du ing he l owe ing peak in co e popula ions, in con as o he ma ginal popula ions we e emales we e mo e a iable (Fig. 4e). Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 116 - Fig 4. Va ia ion in a e age o l o al pheno ypic ai s be ween gende s (he maph odi es s. emales) and egions (co e s. ma ginal) o Daphne lau eola in sou he n Ibe ian Peninsula. Th ee mo phological ai s ela ed o l owe size we e s udied: (a) longi ude o ube; (b) ou e co olla diame e ; and (c) ape u e o ube. Addi ionally, wo phenological ai s we e s udied: pe cen age o open l owe s du ing l owe ing peak, l owe ing ime; and coeffi cien o a ia ion o he numbe o open l owe s du ing l owe ing peak, l owe ing asynch ony. Values plo ed a e he model-adjus ed leas -squa es means (± SE). Pheno ypic Selec ion Es ima ed F om Pollina ion Success Only some l o al ai s o he maph odi e indi iduals we e in l uen ial o he mean numbe o pollen ubes eco ded pe plan , ou es ima e o indi idual ela i e pollina ion success. In co e popula ions, mean numbe o pollen ubes was signi i can ly ela ed o longi ude o ube and ape - u e o ube in he maph odi es, being he le el o s a is ic signi i cance a iable among popula ions (Table 2). In pa icula , he mean numbe o pollen ubes was posi i ely ela ed o longi ude o ube in he maph odi es o wo co e popula ions (CDE and VCV; Fig. 5). Fu he mo e, mean num- be o pollen ubes showed a ma ginally signi i can posi i e ela ionship wi h ape u e o ube in he maph odi es o VCV popula ion ( = 1.85, P = 0.07; b = 0.14 ± 0.08). Howe e , he e was no any signi i can ela ionship in he o he wo co e popula ions (P = 0.12 and P = 0.43 o he maph odi es o CDE and FBJ espec i ely). Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 117 - Co e He maph odi es Co e Females Ma ginal He maph odi es Ma ginal Females E ec d. . F(62) Pb (SE) F(33) Pb (SE) F(62) Pb (SE) F(36) Pb (SE) LT 1 6.16 < 0.05 0.20 (0.08) 0.41 0.52 0.09 (0.14) 0.00 0.96 0.01 (0.09) 0.12 0.74 -0.04 (0.11) OD 1 0.98 0.33 0.05 (0.05) 0.07 0.79 -0.04 (0.14) 0.28 0.60 0.05 (0.10) 0.21 0.65 0.05 (0.12) AT 1 0.50 0.48 0.14 (0.08) 0.03 0.87 0.03 (0.15) 0.44 0.51 0.07 (0.10) 0.47 0.50 0.08 (0.11) FT 1 0.48 0.49 0.07 (0.11) 0.93 0.34 -0.25 (0.26) 5.30 < 0.05 -0.25 (0.11) 1.30 0.26 -0.14 (0.12) FA 1 0.02 0.90 0.01 (0.10) 1.12 0.30 -0.22 (0.21) 0.05 0.82 -0.03 (0.11) 2.16 0.15 0.16 (0.11) Pop 2 0.79 0.46 0.41 0.66 0.59 0.56 0.05 0.95 LT x Pop 2 3.31 < 0.05 AT x Pop 2 3.33 < 0.05 In addi ion, in ma ginal popula ions he pollina ion success was no ela ed o l o al mo pho- logy bu i was la ge in la e l owe ing he maph odi e indi iduals (i.e. wi h lowe l owe ing ime; Table 2). Nei he l o al mo phological no phenological ai s we e ela ed o he mean numbe o pollen ubes in emales o he wo s udy egions (Table 2). Table 2. ANCOVA analyses compa ing pa e ns o selec ion by pollina o s on l o al ai s in emales and he maph odi es o co e and ma ginal popula ions o Daphne lau eola. Dependen a iable was he ela i e mean numbe o pollen ubes. Independen a iables we e s anda dized o mean 0 and a iance 1. Signi i can ela ionships (P ≤ 0.05) a e in bold ace. LT, OD, AT, FT and FA e e o longi ude o ube, ou e diame e , ape u e ube, l owe ing ime and l owe ing asynch ony espec i ely. We i s an he model wi h all o he in e ac ions and hen excluded he nonsigni i can ones. Only linea selec ion g adien s we e conside ed (see Ma e ial and Me hods). Resul s show he s anda dized selec ion coeffi cien (b) and i s s anda d e o mean (SE). Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 118 - Fig 5. S anda dized linea pheno ypic selec ion g adien s o longi ude o l owe ube o D. lau eola, based on numbe o pollen ubes pe s yle as an es ima e o pollina ion success (uppe panels) and ui se as an es ima e o he ma e nal com- ponen o emale i ness (bo om panels). Rela i e alues o dependen a iable, es ima ed by di iding he mean alues pe indi idual by he popula ion means, we e used. Independen a iable was s anda dized o mean 1 and a iance 0. Solid and dashed lines indica e signi i can and non-signi i can ela ionships espec i ely. Pheno ypic Selec ion Es ima ed F om Ma e nal Componen o Fi ness Again, pheno ypic selec ion pa e ns we e a iable among sexes, popula ions and egions. In co e popula ions, posi i e selec ion on longi ude o ube a iable among popula ions was de ec ed in he maph odi es (Table 3). As no ed om pollina ion success (see p e ious sec ion), he maph o- di es wi h g ea e ui se p esen ed also longe co olla ubes in CDE and VCV popula ions bu no in FBJ (Fig. 5). Females o all co e popula ions exhibi ed posi i e selec ion on ou e diame e (Table 3). Also in emales we de ec ed selec ion on l owe ing asynch ony ha a ied among popula ions (Table 3). Highe ui se was posi i ely ela ed o l owe ing asynch ony only in CDE (Fig. 6). In ma ginal popula ions, we did no de ec any selec ion on pheno ypic ai s o he maph odi- e plan s (Table 3). Howe e , pheno ypic selec ion on l owe ing ime and l owe ing asynch ony Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 119 - a iable among popula ions in emales o ma ginal popula ions was de ec ed (Table 3). F ui se was nega i ely ela ed o l owe ing ime and l owe ing asynch ony in emales o GRZ popula ion (Fig. 7). Any pheno ypic selec ion on l owe ing ime was de ec ed nei he CDA no FML ( = 0.19, P = 0.85 and = -0.71, P = 0.48). Flowe ing asynch ony was no ela ed o ui se nei he CDA no FML ( = -0.23, P = 0.82 and = -0.63, P = 0.53). Table 3. ANCOVA analyses compa ing pa e ns o selec ion on l o al ai s in emales and he maph odi es o co e and ma - ginal popula ions o Daphne lau eola. Dependen a iable was ela i e ui se , as an es ima e o he ma e nal componen o i ness. Independen a iables we e s anda dized o mean 0 and a iance 1. Signi i can ela ionships (P ≤ 0.05 o be e ) a e in bold ace. LT, OD, AT, FT and FA e e o longi ude o ube, ou e diame e , ape u e ube, l owe ing ime and l owe ing asynch ony espec i ely. We i s an he model wi h all o he in e ac ions and hen excluded he nonsigni i can ones. Only linea selec ion g adien s we e conside ed (see Me hods). Resul s show he s anda dized selec ion coeffi cien (b) and i s s an- da d e o mean (SE). Co e He maph odi es Co e Females Ma ginal He maph odi es Ma ginal Females E ec d. . F(64) Pb (SE) F(34) Pb (SE) F(62) Pb (SE) F(32) Pb (SE) LT 19.59 < 0.01 0.23 (0.11) 0.34 0.56 0.04 (0.08) 0.00 0.96 0.002 (0.05) 3.50 0.07 0.15 (0.08) OD 1 0.13 0.72 -0.02 (0.06) 8.49 < 0.01 0.24 (0.08) 0.11 0.74 -0.02 (0.05) 0.22 0.64 0.04 (0.09) AT 1 0.05 0.83 -0.01 (0.06) 2.58 0.12 0.17 (0.10) 0.32 0.57 0.03 (0.05) 0.65 0.42 -0.06 (0.08) FT 1 0.19 0.66 -0.06 (0.14) 3.03 0.09 0.41(0.24) 0.39 0.53 0.04 (0.06) 5.13 < 0.05 -0.56 (0.18) FA 1 0.03 0.87 -0.02 (0.14) 1.72 0.20 0.09 (0.21) 0.03 0.86 -0.01 (0.06) 12.84 < 0.01 -0.91 (0.20) Pop 2 1.39 0.26 0.25 0.78 0.34 0.71 2.59 0.09 LT x Pop 24.02 < 0.05 FT x Pop 23.34 < 0.05 FA x Pop 25.15 < 0.05 7.13 < 0.01 Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 120 - Fig 6. S anda dized linea pheno ypic selec ion g adien s o l owe ing asynch ony (FA) in emale indi iduals o co e popula- ions o Daphne lau eola. Flowe ing asynch ony be ween in l o escences o he same indi idual was es ima ed h ough he coeffi cien o a ia ion o he numbe o open l owe s pe in l o escence (N = 5 pe plan ) du ing he popula ion l owe ing peak. Dependen a iable was he ela i e ui se , es ima ed by di iding he mean alues pe indi idual by he popula ion means. Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 121 - Fig 7. S anda dized linea pheno ypic selec ion g adien s o l owe ing ime and l owe ing asynch ony (FT and FA, espec i- ely) in emale indi iduals o he ma ginal popula ion GRZ. Signi i can ela ionships be ween hese phenological a iables and ela i e ui se we e only ound in emales o his ma ginal popula ion (see esul s). Flowe ing ime was es ima ed as he mean pe cen age o opened l owe s pe in l o escence in each indi idual du ing he popula ion l owe ing peak. Flowe ing asynch ony was es ima ed h ough he a ia ion coeffi cien o he numbe o opened l owe s pe in l o escence in each indi- idual du ing he popula ion l owe ing peak. Dependen a iable was he ela i e ui se , es ima ed by di iding he mean alues pe indi idual by he popula ion means. Reduced gende di e gence in l o al ai s and pollina ion success a disjunc ma ginal popula ions o a gynodioecious sh ub - 128 - Acknowledgemen s We hank Ma ía del Ma Alonso and Rocío Reque ey o hei help in he measu ing o l owe s and pollina ion success. Pila Bazaga o assis ance wi h AFLP analysis. L.F. Delph and T.-L. As- hman o help ul commen s. This s udy was unded by he Spanish Minis e io de Educación y Ciencia h ough esea ch p ojec CGL 2006-01355/BOS, he Consejo Supe io de In es igaciones Cien í i cas (CSIC) h ough an I3P ellowship o A.R.C. and he Conseje ía de Inno ación Ciencia y Emp esa, Jun a de Andalucía, h ough he esea ch p ojec RNM156-2005. Cas illa, A.R., Alonso, C. and He e a, C.M. In e iew. Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion. Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion Chap e 5.- Abs ac : An agonis in e ac ions a e conside ed an impo an selec i e ac o on plan species. Geog aphic a ia ion in he pheno ypic selec ion pa e ns by he bi o es on plan ai s could lead o in aspeci i c diff e en ia ion in plan ai s ele an o plan -he - bi o e in e ac ion. In he p esen s udy, we analyzed diff e ences in he a e age he bi o y le el and i s eff ec on ma e nal i ness in co e con inuous and ma ginal disjunc popu- la ions o Daphne lau eola. Fu he mo e, we in es iga ed he exis ence o in aspeci i c diff e en ia ion in ege a i e ai s and i s po en ial connec ion o diff e en ial selec ion by he bi o es in bo h g oups o popula ions. Gene ic dis ance among indi iduals calcula ed om 225 polymo phic AFLP loci was ela ed o diff e ences in he bi o y among indi i- duals o in es iga e he po en ial gene ic basis o he esis ance o ole ance o he bi o y. Ou esul s did no suppo diff e en ial he bi o y incidence bu did a diff e en ial eff ec o he bi o y on ma e nal i ness be ween co e con inuous and ma ginal disjunc popula- ions o D. lau eola. Howe e , diff e ences in he he bi o y le el among indi iduals did no ela e o gene ic dis ance among hem uling ou a gene ic basis o esis ance o ole- ance o he bi o y in D. lau eola. Changes in he he bi o e abundance because o spa ial a ia ion o abio ic condi ions could explain diff e ences in he le el o he bi o y among indi iduals. In addi ion, he bi o es did no exe pheno ypic selec ion consis en wi h he geog aphic a ia ion in s udied plan ai s. Regional diff e ences in he ligh en i onmen o he unde s o y because o changes in he composi ion o ee canopy could be ela ed o egional diff e en ia ion o ege a i e ai s o D. lau eola. The e o e, bo h among indi- idual diff e ences in he he bi o y le el wi hin popula ions and geog aphic a ia ion in ege a i e ai s o D. lau eola seem o be consequence o en i onmen al he e ogenei y mo e han ou pu o geog aphically diff e en ial selec ion by he bi o es. Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 130 - Resumen Las in e acciones an agonis as es án conside adas un agen e selec i o impo an e pa a las especies ege ales. La a iación geog á i ca en los pa ones de selección eno ípica de los he bí o os sob e asgos eno ípicos de las plan as puede o igina di e enciación in aespecí i ca en ales asgos. En es e es udio compa amos poblaciones cen ales con inuas y ma ginales disjun as de Daphne lau eola en elación a sus ni eles medios de he bi o ía y al e ec o de és a sob e la p opo ción de l o es que p oducen u os. Además in es igamos la exis encia de di e enciación in aespecí i ca en asgos ege a i os y su po encial conexión con selección di e encial lle ada a cabo po he bí o os en ambos g upos de poblaciones. Las dis ancias gené icas en e indi iduos ue on es imadas median e el análisis de 225 AFLP loci polimó i cos y elacionadas con las di e encias en el ni el de he bi o ía de los indi iduos pa a in es iga la posible base gené ica de la esis encia o ole ancia a la he bi o ía. Nues os esul ados no apoya on una incidencia di e encial pe o sí un e ec o di e encial sob e la p opo ción de l o es que p oducen u os en poblaciones cen ales con inuas y ma ginales disjun as de D. lau eola. Sin emba go, las di e encias en el ni el de he bi o ía expe imen ado po los indi iduos no se elacionó con la dis ancia gené ica en e ellos desca ando una base gené ica de la esis encia o ole ancia a la he bi o ía en D. lau eola. Cambios en la abundancia de los he bí o os debido a a iación espacial de las condiciones abió icas den o de las poblaciones de la plan a pod ían explica las di e encias en los ni eles de he bi o ía en e indi iduos de una misma población. Además, los he bí o os no eje cie on selección eno ípica consis en e con el pa ón de di e enciación geog á i ca mos ado po los asgos ege a i os es udiados. Di e encias egionales en el ambien e lumínico del so obosque debido a cambios en la composición del dosel a bó eo pod ían es a elacionadas con la di e enciación egional en los asgos ege a i os de D. lau eola. Po an o, an o las di e encias en e indi iduos den o de las poblaciones en los ni eles de he bi o ía como la a iación geog á i ca en los asgos ege a i os de D. lau eola pa ecen se una consecuencia de la he e ogeneidad ambien al más que un esul ado de selección di e encial lle ada a cabo po los he bí o os. Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 131 - In oduc ion He bi o y can educe g ow h, ecundi y and su i al o plan s (C awley 1989; S auss 1991). Geog aphic a ia ion in plan -he bi o e in e ac ions is also commonly obse ed and hus bo h he magni ude and he e ec s o he bi o y on plan popula ions should be geog aphically a iable (S auss and Zange l 2002; Pennings and Silliman 2005; Adams and Zhang 2009). Spa ial a ia- ion in he in ensi y o in e ac ions is impo an because i es ablishes geog aphic selec ion mosaics whe e in aspeci i c di e en ia ion p ocesses could occu (Thompson 1994). He bi o e abundance can a y spa ially and empo ally because o abio ic (e.g. wea he con- di ions) and bio ic (e.g. abundance o hei hos plan ) ac o s (Schoonho en e al. 1998). Such a ia ions can be spa ially andom bu we can p edic ha se e al ac o s can lead o changes in plan -he bi o e in e ac ions associa ed wi h ma ginal popula ions (Ma on and Ch one 2006; Gas on 2009). Ma ginal popula ions o plan species equen ly p esen ecological and gene ic peculia i ies ha could con ibu e o modi y he ela ionships be ween plan s and hei he bi o es (e.g. educed gene ic di e si y could dec ease he adap a i e po en ial o ma ginal popula ions o he bi o es). Fi s , changes in he bi o e auna ha e been equen ly epo ed in ma ginal popula- ions o he hos plan (Lewinsohn e al. 2005). Fu he mo e, he nega i e impac o he bi o y on i ness o indi iduals could be la ge in ma ginal popula ions con ibu ing o he main enance o s able dis ibu ional limi s (B uelheide and Scheidel 1999; Ma on and Ch one 2006). The e o e, ma gins o dis ibu ion eme ge as na u al loca ions whe e plan -he bi o e in e ac ions a e expec- ed o a y (Ma on and Ch one 2006). Howe e , ew empi ical s udies ha e examined he in l uence o na u al enemies a ange limi s (Gas on 2009). Fu he mo e, plan ai s ela ed o esis ance o he bi o y should be unde di e en ial selec i e p essu e in a landscape whe e he he bi o e-plan in e ac ions a y spa ially and empo a ily (Zange l and Be enbaum 2003; Muola e al. 2010a; Ve gee and Kunin 2011). Plan quali y om he he bi o e’s iewpoin depends on a wide a ie y o pheno ypic ai s encompassing om nu i io- Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 132 - nal and seconda y chemicals o plan a chi ec u al ai s. Va ia ion in hese ai s can e l ec no only gene ic a ia ion among indi iduals, bu also une enly dis ibu ed esou ces (Ka ban 1992). The e o e, he analysis o he geog aphic a ia ion in pheno ypic ai s ela ed o he bi o y, hei connec ion o di e en ial selec ion egimes by he bi o es and he de e mina ion o he gene ic o en i onmen al con ol o such pheno ypic a ia ion is a ma e o maximal ele ance in e olu io- na y ecology. In he p esen s udy, we compa ed co e con inuous and ma ginal disjunc popula ions o he sh ub Daphne lau eola wi hin i s dis ibu ion ange in he Sou h o Ibe ian Peninsula. In his e- gion, mos o popula ions o he species a e gynodioecious wi h emale and he maph odi e indi- iduals coexis ing wi hin popula ions (Alonso e al. 2007). Fou noc uid ca e pilla s cons i u e he main he bi o es o he species in he s udy a ea (Alonso and He e a 1996). Gende di e ences in esis ance and ole ance o damage by he bi o es has been epo ed in gynodioecious species. P e ious s udy did no i nd ha D. lau eola he bi o es disc imina e be ween lea es o emale and he maph odi e plan s (Alonso 2003). Howe e , ma ginal popula ions equen ly p esen ha she condi ions ha could p omo e sex-di e en ial esis ance and ole ance o damage by he bi o es (Ashman 2006; Kawecki 2008). The e o e, in he p esen s udy we analyze he e ec o he bi o y on plan s wi h di e en sex in bo h co e con inuous and ma ginal disjunc popula ions. Fu he mo- e, in aspeci i c a ia ion in de olia ion le els has been ela ed o plan a chi ec u e (Alonso and He e a 1996) and lea nu ien composi ion (Alonso and He e a 2003). Howe e , he in l uence o seconda y me aboli es on his plan -he bi o e sys em is s ill no ully unde s ood. Speci i cally we add essed he ollowing ques ions: (1) do co e con inuous and ma ginal disjunc popula ions o D. lau eola di e in hei le el o he bi o y damage? (2) does he bi o y damage ha e s on- ge impac on i ness in indi iduals o ma ginal disjunc popula ions? (3) is he e di e en ial pheno ypic selec ion by he bi o es in co e con inuous and ma ginal disjunc popula ions? and (4) do indi iduals gene ically di e en ia ed showed signi i can ly di e en le el o he bi o y damage? Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 133 - Ma e ial and me hods S udy Species Daphne lau eola L. (Thymelaeaceae) is a long-li ed e e g een sh ub wi h a disjunc Palaea c ic dis ibu ion associa ed o he main calca eous moun ainous anges o Eu ope and no h o A ica (Meusel e al. 1978). This s udy was conduc ed in 2007-2009 in he Bae ic Ranges in sou he n Ibe- ian Peninsula. The l owe ing occu s in Janua y-Ap il sho ly be o e new lea es a e p oduced. D. lau eola plan s consis o a a iable numbe o e ec s ems ha ise a g ound le el om a com- mon unk. Lea es a e ound only a he dis al end o each b anch, o ming a single, well-de i ned lea who l. Fou polyphagus noc uid species a e he main he bi o es o he species in he s udy egion (T igonopho a l ammea Espe ., Noc ua jan he Bkh., Noc ua i mb ia a Sch ebe , Pseudena - gia ulicis S aud.). The pe iod o ac i i y o he bi o es encompasses om Ap il o mid-June. We only e alua ed he impo ance o ca e pilla s as consume s o oliage, al hough hey also consume l owe s and un ipe ui s (Alonso and He e a 1996). S udy A ea Ac oss he Bae ic Ranges, D. lau eola p esen s a local con inuous cen e in he Sie a de Cazo la wi h nume ous and la gely connec ed popula ions (Cas illa e al. in p ess). The spa ial isola ion inc eases owa ds eas e n and wes e n ma gins, al hough he inc ease is sha pe in wes e n edge (Cas illa e al. in p ess). The e o e, in his s udy we conside ed he Sie a de Cazo la as he local con inuous cen e (“co e egion” he ea e ) and popula ions a he wes e n edge as a local disjunc ma gin (“ma ginal egion” he ea e ) o he species’ dis ibu ion pa ch in sou he n Ibe ian Penin- sula. We s udied h ee accessible popula ions pe egion (see Chap e 3). All s udy popula ions we e gynodioecious, and he p opo ion o emale and he maph odi e indi iduals was de e mined du ing l owe ing pe iod o 2007 by examining 100 andomly chosen indi iduals. The equency o emales was ca. 20 % in all o hem (see Chap e 3). A he beginning o he s udy, we haphaza dly ma ked 25 he maph odi e and 15 emale plan s in each s udy popula ion. Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 134 - Plan Pheno ypic and Gene ic T ai s In 2007, we measu ed he basal diame e , a g ound le el, o he suppo ing s em o each ma - ked plan using a Mi u oyo digi al callipe (basal diame e , he ea e ). A p e ious s udy showed basal diame e is unde nega i e pheno ypic selec ion by he bi o es in a leas some popula ions o D. lau eola in he co e egion (Alonso and He e a 1996). In addi ion, we es ima ed p opo ion o ui s pe l owe using andomly ma ked i e in l o escences pe indi idual (“ ui se ” he ea e ). In 2009 July, i e in ac lea s (i.e. wi hou signs o he bi o y) we e haphaza dly emo ed wi h scisso s o each ma ked plan , placed in o sealed plas ic bags and kep in a po able coole o mi- nimize wa e loss du ing hei anspo o he i eld s a ion. A e wa ds, lea es we e indi idually placed in small pape en elopes and d ied immedia ely a ambien empe a u e in sealed con ai- ne s wi h abundan silica gel. D ied lea es we e indi idually weigh ed on an analy ical balance Me le Toledo PL 203.S and hei a ea measu ed using a lea a ea me e LI-3000C (LI-COR, Lin- coln, USA). Then, d y weigh and lea a ea we e used o calcula e he speci i c lea a ea mm2 g-1 o each lea (SLA, he ea e ). D y lea es om he same plan (N = 5), we e pooled in o a single sample o chemical analyses and homogenised o a i ne powde using a Re sch MM 301 mill. This i ne powde was he s a ing ma e ial o he ex ac ion p ocedu e. 250 mg o i ne powde o each indi idual we e ex ac ed wi h 100 ml o me hanol/wa e (70:30, / ) by agi a ion a oom empe a u e o 24 hou s. The ex ac ion p ocedu e was epea ed wice o each indi idual. The mix u es we e cen i uga ed a 4000 pm, i l e ed h ough a Wha man i l e pape and he i l a e was concen a ed by e apo a- ion o he sol en , using a o a o y e apo a o (BÜCHI Ro a apo R-215, BÜCHI, Lausanne, Swi - ze land). The ex ac was hen e-dissol ed in 10 ml o me hanol/wa e (80:20, / ) To al phenolic concen a ion in plan ex ac s (Phenolic con en , he ea e ) we e de e mined spec opho ome i- cally by he Folin-Ciocal eau assay (Single on e al. 1999) using gallic acid as a s anda d. An ali- quo o 20 µl dilu ed wi h 1580 µl wa e was mixed wi h 100 µl o Folin-Ciocal eau phenol eagen Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 135 - and allowed o eac o 5 min. A e wa ds, we added 300 µl o sa u a ed Na2CO3 solu ion and allowed o s and o 120 min a oom empe a u e be o e he abso bance o he eac ion mix u e was measu ed in iplica e a 765 nm. Phenolic con en o he plan ex ac s was exp essed as mg gallic acid equi alen s pe g o plan ma e ial. Mo eo e , we quan i i ed he concen a ion in lea es o he h ee mos abundan couma ins in D. lau eola using a subsample o 15 indi iduals (Alonso e al. 2005; Alonso e al. 2009). Phenolic con en showed a s ong posi i e co ela ion wi h concen- a ion o he h ee couma in compounds analyzed (Fig. 1). Figu e 1. Rela ionship be ween o al phenolic con en and he h ee mos abundan couma ins in lea es o D. lau eola (N = 15). Pea son co ela ion was used. Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 136 - Ampli i ed agmen leng h polymo phism (AFLP) ma ke s we e used o cha ac e ize 240 s udy plan s gene ically. The AFLP analysis was pe o med essen ially as o iginally desc ibed by Vos e al. (1995), wi h modi i ca ions in ol ing he use o l uo escen dye-labelled selec i e p ime s ollowing Applied Biosys ems (2005). Each plan was i nge p in ed using ou EcoRI + 3/MseI + 3 and ou Ps I + 2/MseI + 3 p ime combina ions. F agmen sepa a ion and de ec ion we e made using an ABI PRISM 3130xl DNA sequence , and he p esence/absence o each ma ke in each in- di idual plan was sco ed manually by isualizing elec opho eg ams wi h GeneMappe 3.7 so - wa e. Only agmen s ≥ 150 base pai s in size we e conside ed, as a way o educing he po en ial impac o size homoplasy (Vekemans e al. 2002)., De ails on p ime combina ions, numbe o ma ke s, sco ing e o a es and le els o polymo phism can be ound in Cas illa e al. in p ess. He bi o y Incidence The p opo ion o lea a ea emo ed by he bi o es was quan i i ed in he lea who ls o 20 % o o al s ems pe plan , excep o indi iduals wi h ≤ 12 s ems, in which he p opo ion o lea a ea emo ed was quan i i ed in all s ems (He bi o y le el, he ea e ). He bi o y le el o indi iduals was quan i i ed in wo di e en yea s (2007 and 2009). In each lea who l, each lea e was classi- i ed in o one o 6 he bi o y classes acco ding o pe cen age o lea a ea emo ed: 0, no signs o he bi o y; 1, 1-5% a ea emo ed; 2, 6-25%; 3, 26-50%; 4, 51-75%; and 5, >75%. An o e all es ima e o he bi o y le el o each lea who l was ob ained ollowing he me hod p oposed by Alonso and He e a (1996). Finally we calcula ed he mean alue o he bi oy le el pe indi idual a e a- ging he he bi o y alues o all i s lea who ls. Da a Analysis All s a is ical analyses we e pe o med using SAS s a is ical package (SAS Ins i u e 2002). Di e ences in he le el o he bi o y be ween egions and sexes we e analyzed by gene alized linea mixed models using he nega i e binomial unc ion (P ocedu e GLIMMIX). E ec s o egion Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 137 - (co e s ma ginal), popula ion and sex (he maph odi e s emale) we e ea ed as i xed ac o s. Di e ences in he le el o he bi o y in 2007 and 2009 we e analyzed wi h sepa a ed analyses. In addi ion, empo al a iabili y in he le el o he bi o y o indi iduals was explo ed wi h spea man ank co ela ion. Di e ences be ween sexes and egions in basal diame e , phenolic con en and SLA o indi i- duals we e analyzed by gene alized linea mixed models (P ocedu e GLIMMIX) wi h no mal e o dis ibu ion o phenolic con en and nega i e binomial dis ibu ion o bo h basal diame e and SLA. Sex, popula ion and egion e ec s we e ea ed as i xed. We s udied he ela ionship be ween he bi o y and ui se in 2007, analyzing sepa a ely he - maph odi e and emale indi iduals wi hin each egion, because i is expec ed ha he e ec o he bi o y on i ness a ies be ween he maph odi e and emale indi iduals in gynodioecious spe- cies (Ashman 2006, and e e ences he ein). Fo each sex, we explo ed po en ial among popula ion di e ences in he ela ionship be ween he bi o y le el and ui se using an analysis o co a iance (P ocedu e MIXED). F ui se was he esponse a iable, showing a no mal dis ibu ion. He bi o y le el was he con inuous explana o y a iable and popula ion was he ca ego ical e ec . Pheno ypic selec ion a each egion was examined by Lande-A nold selec ion g adien models (Lande and A nold 1983). Wi hin each egion, we p e iously analyzed pheno ypic selec ion pa - e ns by he bi o es on indi iduals o di e en sex sepa a ely. Howe e , he simila i y be ween sexes in pheno ypic selec ion analyses oge he wi h he absence o di e ences in he he bi o y le el and pheno ypic ai s lead us o analyzed he maph odi e and emale indi iduals o each egion join ly o ge la ge s a is ical powe in he analyses. The e o e, we conduc ed sepa a ed pheno ypic selec ion analyses o indi iduals o co e and ma ginal popula ions wi hou acking in o accoun sex o indi iduals. We quan i i ed pheno ypic selec ion only in 2009 because we only measu ed olia ai s (SLA and phenolic con en ) in ha yea and olia ai s equen ly p esen among-yea a ia ion. We used ela i e alues o he esponse a iable, he bi o y le el (es ima- Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 144 - Selec ion by He bi o es on Plan T ai s Acco ding o he p e ious esul s, we could expec s onge pheno ypic selec ion on indi iduals o ma ginal popula ions. In co e popula ions plan s wi h lowe phenolic con en p esen ed ma gi- nally la ge he bi o y le el, in con as o ma ginal popula ions whe e he phenolic con en had no any e ec (Table 2). In ma ginal popula ions, he bi o es also exe ed pheno ypic selec ion on basal diame e bu his e ec was a iable among popula ions (Table 2). Indi iduals wi h la ge basal diame e p esen- ed also la ge he bi o y le el in FML popula ion, bu he e we e no any e ec in he o he wo popula ions (Fig. 6). In co e popula ions, he e was no any ela ionship be ween he bi o y le el and basal diame e (Table 2). Finally, he bi o es did no exe any pheno ypic selec ion on SLA nei he co e no ma ginal popula ions (Table 2). Co e Ma ginal E ec d. .F(108) Pb (SE) F(101) Pb (SE) BD 1 2.64 0.11 -0.18 (0.11) 0.71 0.40 -0.19 (0.12) PC 1 3.90 0.05 -0.21 (0.11) 2.71 0.10 -0.13 (0.08) SLA 1 0.11 0.74 -0.04 (0.11) 1.26 0.27 -0.09 (0.08) Pop 2 0.03 0.97 0.11 0.89 BD x Pop 27.31 < 0.01 Table 2. ANCOVA analyses compa ing pa e ns o selec ion by he bi o es on ege a i e and olia ai s in co e and ma ginal popula ions o D. lau eola. Dependen a iable was he le el o he bi o y. BD, PC and SLA ep esen basal diame e , phenolic con en and speci i c lea a ea espec i ely. Independen a iables we e s anda dized o mean 0 and a iance 1. Signi i can ela ionships (P < 0.05 o be e ) a e in bold ace. Ma ginally signi i can ela ionships (0.05 ≤ P < 0.09) a e in i alics. Only linea selec ion g adien s we e conside ed (see Ma e ial and Me hods). Resul s show he s anda dized selec ion coeffi cien (b) and i s s anda d e o mean (SE). Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 145 - Figu e 6. S anda dized linea pheno ypic selec ion g adien s o he basal diame e (BD) in indi iduals o ma ginal popula- ions o D. lau eola (FML, CDA, GRZ). Dependen a iable was he ela i e ui se , es ima ed by di iding he mean alues pe indi idual by he popula ion means. Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 146 - Discussion He bi o y and I s E ec on Fi ness in Co e and Ma ginal Popula ions An agonis ic in e ac ions a e conside ed impo an agen s limi ing geog aphic dis ibu ion o species (Gomulkiewicz e al. 2000; Nuisme e al. 2000; Pa e son e al. 2010). Fo ins ance, escape om p eda o s o pa hogens has been epo ed as a causal agen o geog aphic expansion in nu- me ous in asi e species (Ma on and Vilá 2001; To chin e al. 2003). In addi ion, a g owing body o expe imen al li e a u e demons a ing a key ole o he bi o y in he spa ial s uc u e o plan popula ions is eme ging (B uelheide and Scheidel 1999; Fine e al. 2004; Gómez e al. 2005; Mille e al. 2009). Howe e , he bi o y has no been adi ionally conside ed an impo an ac o limi ing geog aphic dis ibu ion o plan species (Gas on 2009). Co e popula ions o D. lau eola p esen ed la ge he bi o y le el in 2009, whe eas co e and ma - ginal popula ions did no di e in he a e age he bi o y le el in 2007. Dec ease o he bi o y da- mage in ma ginal popula ions has been epo ed in o he s udies (Moo e 2009; Ve gee and Kunin 2011). Howe e , he bi o y a ec ed nega i ely o ui se o indi iduals o bo h sexes in ma ginal popula ions, bu did no in co e popula ions. In e es ingly, indi iduals o ma ginal popula ions had g ea e ui se han hose o co e popula ions (Cas illa e al. 2011a). The e o e, esul s o he p esen s udy gi e s ill mo e suppo o ou hypo hesis o a mo e e i cien pollina ion en i onmen in ma ginal popula ions o D. lau eola (Cas illa e al. 2011a). Fu he mo e, we ound low a e a- ge le els o he bi o y in popula ions o D. lau eola, bu he di e ences among indi iduals we e empo a ily consis en in line wi h p e ious in es iga ions o he s udy species a o he co e popu- la ions (Alonso and He e a 1996). Ne e heless, his esul does no ule ou a ele an ole o D. lau eola he bi o es as selec i e agen s because ch onic le els o ela i ely insubs an ial he bi o y ( ha is, damage ha esul s in ela i ely li le issue los ) can ha e signi i can impac s on long- li ed plan s o e hei li e imes (Doak 1992; Mille e al. 2009). Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 147 - Selec ion by He bi o es in Co e and Ma ginal Popula ions o D. lau eola He bi o es equen ly exe pheno ypic selec ion on ege a i e ai s o plan species, al hough he in ensi y o such pheno ypic selec ion equen ly a ies geog aphically (Gómez e al. 2005; Pennings e al. 2009; Muola e al. 2010a; Ve gee and Kunin 2011). This geog aphic a ia ion in pheno ypic selec ion could lead o in aspeci i c di e en ia ion among plan popula ions in pheno- ypic ai s impo an o such in e ac ions (e.g. He e a e al. 2006 o selec ion by pollina o s). A igo ous demons a ion o a ele an ole o D. lau eola’s he bi o es as selec i e agen s equi es demons a ing ha obse ed di e ences in he he bi o y le el among indi iduals ha e a gene- ic basis. Resis ance and ole ance o he bi o y a e commonly gene ically de e mined ai s ha show he i able gene ic a ia ion wi hin species (e.g. Muola e al. 2010b; Kuzina e al. 2011; E ans e al. 2011). In model o ganisms and c op species, g ea s ides ha e been made in unde s anding he e olu ion o plan de ence, as well as exp ession changes in esponse o he bi o y con as ing o non-model o ganisms whe e ew s udies ha e examined his issue (Ande son and Mi chell-Olds 2011). Ou esul s did no suppo a p edic able, and empo a ily cons an ela ionship be ween di- e ences in he he bi o y le el among indi iduals and hei gene ic dis ance in wild popula ions o D. lau eola. Fu he mo e, al hough all s udied ege a i e ai s did di e be ween co e and ma gi- nal popula ions o D. lau eola, he bi o es did no p omo e such pheno ypic di e en ia ion h ough pheno ypic selec ion. He bi o es exe ed nega i e selec ion on he phenolic con en o lea es in indi iduals o co e popula ions, whe eas he bi o es did no exe any selec ion on his ai a ma ginal popula ions. This esul seems o ma ch wi h he geog aphic a ia ion in he chemical composi ion o D. lau eola popula ions, wi h an inc ease o he le el o phenolic compounds o de- ence in popula ions whe e he bi o es exe nega i e pheno ypic selec ion on his ai . Howe e , we mus be cau ious he e because selec ion in co e popula ions was only ma ginally signi i can . On he o he hand, nei he SLA no basal diame e geog aphic a ia ions ma ched wi h he pheno- ypic selec ion pa e ns by he bi o es in co e and ma ginal popula ions. The e o e bo h di e ences in he le el o he bi o y among indi iduals wi hin popula ions and egional di e ences in ege a i e ai s seem o be ela ed o abio ic a ia ion. On he one hand, Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 148 - abundance o e es ial insec he bi o es can be limi ed by abio ic condi ions, pa icula ly em- pe a u e and mois u e (Kingsol e 1989; Hun e and P ince 1998; Ri chie 2000). The e o e, ou e- sul s sugges ha en i onmen al he e ogenei y among and wi hin D. lau eola popula ions lead o spa ial a ia ion in he abundance o he bi o es, p omo ing di e ences in he he bi o y le el among indi iduals and popula ions. On he o he hand, indi iduals o co e popula ions p esen ed lowe a e age basal diame e and lea s wi h lowe SLA bu la ge phenolic con en han hose indi iduals o ma ginal popula ions. Di e ences in SLA and phenolic con en could be ela ed o con as ing ligh en i onmen s in co e and ma ginal popula ions o D. lau eola. Plan s g owing in shaded en i onmen s equen ly p esen lea es wi h la ge SLA which a e mo e e i cien in cap- u ing a limi ing ligh esou ce (Chen e al. 1996; Meziane and Shipley 1999; Ba be and Ma quis 2011). On he o he hand, ligh exposu e can al e lea quali y, especially in plan s wi h ca bon- based de ences, esul ing in plan s wi h lowe phenolic con en in lea es in shaded en i onmen s (Nichols-O ions 1991; Dud and Shu e 1994; Ba be and Ma quis 2011). In co e egion, D. lau eola g ows in he unde s o y o mixed o es s cha ac e ized by shading he e ogenei y (Alonso and He- e a 2008). In ma ginal a eas, indi iduals o D. lau eola a e es ablished in he sh ub s a um o closed i o es s, Abies pinsapo Boiss, cha ac e ized by mo e homogeneous and closed canopy (A is a 1995). The e o e, he la ge SLA and lowe phenolic con en in lea es o indi iduals o ma - ginal popula ions could be a consequence o li ing in mo e shaded en i onmen s. The e o e, esul s o he p esen s udy sugges ed ha he bi o es did no ha e a ele an ole p omo ing in aspeci i c di e en ia ion in ege a i e ai s o adul plan s among D. lau eola popula ions. Howe e , we only ha e s udied he e ec s o olia he bi o y on adul plan s and some s udies poin ou he ele ance o he bi o y di e s among di e en li e s ages in plan spe- cies. Fo ins ance, slugs also seem o ha e a ele an ole in he su i al o D. lau eola seedlings (Alonso pe sonal comm.), bu we did no analyze hei e ec in he p esen s udy. In his sense, B uelheide and Scheidel (1999) ound ha slug he bi o y was an impo an ac o limi ing he geog aphical dis ibu ion o A nica mon ana. On he o he hand, olia he bi o es o D. lau eola also consume l owe s and ui s bu hei e ec s a e s ill wai ing o be e alua ed. Loose he bi o y in e ac ions in a long-li ed unde s o y plan : non-gene ic basis o adul de olia ion - 149 - Concluding Rema ks Resul s o he p esen s udy showed impo an di e ences in he he bi o y le el among nea by popula ions mo e han clea egional di e gence. Among popula ion and among indi idual di- e ences seem o be ela ed o en i onmen al he e ogenei y de e mining spa ial a ia ion in he abundance o he bi o es mo e han di e ences in he esis ance o indi idual plan s wi h a gene ic basis. In addi ion, ou esul s uled ou a egional di e en ia ion in ege a i e ai s conduc ed by di e gen pheno ypic selec ion by he bi o es in co e and ma ginal popula ions o D. lau eola. Ra he con as ing ligh en i onmen s because o di e ences in he species composi ion o he ee canopy in co e and ma ginal popula ions o D. lau eola could be ela ed o egional di e ences in olia ai s. In addi ion, sexes did di e nei he ege a i e ai s no pheno ypic selec ion by he bi- o es suppo ing he widely accep ed issue o sca ce gende di e en ia ion o ege a i e ai s in con as o ep oduc i e ai s in gynodioecious species. Acknowledgemen s We hank Ma ía del Ma Alonso o he in aluable help in he i eld wo k. Isabel Ma ía Ga cía ca ied ou he chemical analyses. This s udy was unded by he Spanish Minis e io de Educación y Ciencia h ough esea ch p ojec CGL 2006-01355/BOS, he Consejo Supe io de In es igaciones Cien í i cas (CSIC) h ough an I3P ellowship o A.R.C. and he Conseje ía de Inno ación Ciencia y Emp esa, Jun a de Andalucía, h ough he esea ch p ojec RNM156-2005. Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 151 - El es udio de la a iación geog á i ca in aespecí i ca en asgos ele an es pa a las in e acciones plan a-animal y su posible elación con una selección di e gen e e ec uada po animales, es c u- cial pa a en ende el ínculo en e mic o y mac oe olución en es os ca ac e es y el papel de las in- e acciones plan a-animal en la e olución de los sis emas sexuales en plan as (Ba e e al. 2001; Zange l & Be enbaum 2003; He e a e al. 2006). La eo ía del mosaico geog á i co de coe olución p opues a po Thompson (1994, 2005) y los modelos gené icos asociados a ella (Gomulkiewicz e al. 2000; Alleaume-Benha i a e al. 2006) p opo cionan un modelo eó ico que pe mi e el ace ca- mien o al es udio del papel causal de los animales en la apa ición y man enimien o de di e encias in aespecí i cas en los asgos eno ípicos impo an es pa a las in e acciones plan a-animal (Ve - mee e al. 2011). Sin emba go, es e ma co eó ico no p edice las á eas conc e as donde el cambio de las dinámicas coe olu i as en las in e acciones plan a-animal sea espe able a p io i. El obje i o gene al de es a Tesis Doc o al ha sido a a de mos a cómo los má genes de dis ibución de las especies ege ales, debido a sus peculia idades ecológicas y gené icas, cons i uyen á eas donde cabe espe a cambios en las in e acciones plan a-animal. La in eg ación de in o mación ob enida median e mues eos de campo (cuan i i cación del éxi o de polinización y ep oducción, censos de polinizado es, es imas de he bi o ía), écnicas molecula es (di e sidad y di e enciación gené ica, de ección de loci bajo selección na u al) y medición de asgos eno ípicos en labo a o io (mo olo- gía l o al y olia , enología, análisis de la concen ación de enoles en hojas) han pe mi ido in e i la impo ancia de polinizado es y he bí o os como agen es del cambio en los asgos eno ípicos de D. lau eola en má genes de su dis ibución. La Es uc u ación Geog á i ca de las Poblaciones de D. lau eola El análisis espacialmen e explíci o lle ado a cabo con la mayo pa e de las poblaciones de D. lau eola exis en es en las Co dille as Bé icas e eló que la es uc u ación geog á i ca de las po- Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 152 - blaciones a ió hacia los má genes de la dis ibución. Los esul ados de es e es udio mos a on un inc emen o del aislamien o espacial de las poblaciones hacia los má genes de dis ibución de la especie en el á ea de es udio (Capí ulo 1). Sin emba go, no hubo e idencias que apoyasen una educción del amaño poblacional hacia los má genes de dis ibución (Capí ulo 1). De hecho, las poblaciones ma ginales de D. lau eola en la egión de es udio se ca ac e iza on po p esen a en muchos casos más de 100 indi iduos ep oduc i os. Es os esul ados sugi ie on la exis encia de lo- calidades ecológicamen e óp imas pa a la especie en odo el á ea de es udio, aunque la abundan- cia de las mismas disminuyó hacia los má genes de dis ibución (Ecke e al. 2008; Yakimowski and Ecke 2008; Sex on e al. 2009). Po o o lado, aunque el aislamien o espacial de las pobla- ciones aumen ó an o hacia el ma gen o ien al como hacia el ma gen occiden al, la magni ud del aislamien o no ue equi alen e. Las poblaciones del ma gen occiden al p esen a on una ni el de aislamien o espacial mucho mayo que las poblaciones del ma gen o ien al. El análisis de la es- uc u a gené ica de las poblaciones demos ó que es as di e encias en el aislamien o espacial de las poblaciones ienen impo an es consecuencias sob e el pa ón de di e enciación gené ica de las poblaciones (Capí ulo 1). En es a Tesis Doc o al demos amos que la es uc u ación gené ica de las poblaciones puede se di e en e en dis in os má genes de dis ibución (Hampe and Pe i 2005; Sex on e al. 2009). Las poblaciones bé icas del ma gen o ien al ue on gené icamen e muy simi- la es a las poblaciones cen ales y además, cons i uye on un g upo gené icamen e homogéneo, cons i uyendo un lími e en expansión (“expanding edge”). En cambio, las poblaciones disjun as del ma gen occiden al es u ie on gené icamen e muy di e enciadas del es o de las poblaciones bé icas y p esen a on además una mayo he e ogeneidad gené ica en e ellas, ca ac e ís icas ípi- cas de un ma gen de dis ibución de e agua dia (“ ea o ailing edge”). Es a dis inción gené ica en e di e en es bo des de dis ibución de una especie ege al ha sido p opues a a escala de la dis ibución geog á ica global de las especies, donde la dis ancia geog á ica en e bo des ocu e a escala con inen al (Hampe and Pe i 2005). Sin emba go, noso os hemos demos ado que la es uc u ación gené ica di e encial en dis in os bo des de dis ibución puede p oduci se igualmen- e a escalas geog á icas más educidas, como son cada uno de los pa ches de dis ibución de las especies ege ales con dis ibuciones disjun as. Es e ipo de si uaciones pueden se especialmen e Geog aphic a ia ion, in e ac ions wi h animals and sexual dimo phism: glancing a ange limi s in a gynodioecious plan . - 153 - ecuen es en sis emas mon añosos que no se ie on a ec ados d ás icamen e po los pe íodos glacia es, como es el caso de las Co dille as Bé icas (Bene e al. 1991; Ca ión 2002; Ca ión e al. 2003; Médail & Diadema 2009). Es e pa ón di e encial de es uc u ación gené ica puede in lui de mane a di e en e en las dinámicas plan a-animal a a és de la emezcla de ca ac e es ía lujo génico, de i a gené ica y dinámicas de ex inción/colonización (Thompson 1994, 2005). La exis encia de una in ensa selección na u al di e encial pod ía explica la ue e di e enciación gené ica en e las poblaciones disjun as del ma gen occiden al en e al es o de poblaciones. Po ello, analizamos los pe iles gené icos de los indi iduos siguiendo un en oque genómico pobla- cional des inado a de ec a la p esencia de ma cado es suje os a selección en e los ma cado es polimó icos usados (Capí ulo 2). Nues os esul ados con i ma on la exis encia de selección na- u al di e encial en e ambas egiones. Los loci que mos a on signos de selección na u al en las poblaciones disjun as del ma gen occiden al no mos a on ales signos en el es o de poblaciones, y ice e sa. Una ez de ec ados los loci ou lie s, se e i a on del análisis de la es uc u a gené ica de las poblaciones pa a disce ni cuán o de la di e enciación gené ica en e los g upos disjun os de poblaciones se podía a ibui a selección di e encial y cuán o a a iación neu al a ec ada po e ec os his ó icos de de i a y lujo génicos. La ue e di e enciación gené ica en e las poblaciones disjun as del ma gen occiden al y el es o de poblaciones pe sis ió después de e i a los loci ou - lie s del análisis. Po an o, nues os esul ados con i ma on que an o la selección na u al di e en- cial como el ue e aislamien o gené ico pa ecen es a con ibuyendo a la ma cada di e enciación gené ica en e las poblaciones disjun as del ma gen occiden al y el es o de poblaciones bé icas de D. lau eola. El papel de la selección di e encial sob e plan as de dis in o sexo en el man enimien o del po- limo ismo sexual en plan as ha sido demos ado en o as especies ginodioicas (Ashman 1999, 2003; Sakai e al. 2007). Nues os esul ados con i ma on un pa ón di e encial en e indi iduos de dis in o sexo en ambas egiones (Capí ulo 2). Los indi iduos he ma odi as siemp e p esen a on un mayo núme o de loci con ecuencias alejadas del pa ón neu al que los indi iduos hemb a.