From flower to fruit: ecological and evolutionary implications of abiotic and biotic factors affecting Juniperus thurifera L. cone development

ESCUELA DE INGENIARÍA DE LA INDUSTRIA FORESTAL, AGRONÓMICA Y DE LA BIOENERGÍA Dpto. de Ciencias Agroforestales INSTITUTO UNIVERSITARIO DE INVESTIGACIÓN EN GESTIÓN FORESTAL SOSTENIBLE TESIS DOCTORAL: From flower to fruit: Ecological and evolutionary implications of abiotic and biotic factors affecting Juniperus thurifera L. cone development. Presentada por Erik Rodríguez García para optar al grado de doctor con Mención Internacional por la Universidad de Valladolid Dirigida por: Dr. José Miguel Olano Mendoza Dr. Eduardo Tomás Mezquida Diseño de cubiertas y portada: Hugo Rodríguez García Fotografías: Erik Rodríguez García. 2010 - 2016. Portada y capítulo 2: Megastigmus thuriferana ovopositando Capítulo 1: Conos femeninos de sabina albar Capítulo 3: Microscopía óptica de semilla de sabina: nucela con tubos polínicos Capítulo 4: Agujero de salida de M. thuriferana Capítulo 5: Sabinar albar en Cabrejas del Pinar (Soria) Agradecimientos: Aegithalos caudatus Contraportada: Sabinar albar en Arcones (Segovia) “Creo que la naturaleza une a las culturas del mundo. Un árbol tiene raíces en el suelo y ramas que rozan el cielo, y nos recuerda que para prosperar tenemos que saber de dónde venimos.” Wangari Muta Maathai A la memoria de “Takuni” CONTENTS Resumen 11 Capítulo 1 De la flor al gálbulo: Implicaciones biológicas y evolutivas de los factores abióticos y bióticos que controlan la producción de gálbulos en Juniperus thurifera L. Una visión general de la tesis 15 ¿Qué ocurre entre la floración y la dispersión?: el marco general de la tesis 16 ¿Estoicismo o resignación? El caso de la sabina albar 19 Objetivos 22 Metodología 23 Principales resultados 27 Principales conclusiones 29 Referencias 31 Chapter 2 Ripe fruits may not be optimal: green fruits improve the detection of traits associated to predispersal predation risk 39 Abstract 40 Introduction 41 Material and methods 42 Natural history 42 Study area 43 Sampling design and fruit analysis 43 Statistical analysis 44 Results 46 Phenological sequence of events 46 Fruit and full seed production 48 Fruit damage 49 Discussion 52 References 57 Supplementary material 60 Capítulo 3 Deciphering the role of reproductive investment, pollination success and predispersal seed predation on reproductive output in Juniperus thurifera 65 Abstract 66 Introduction 67 Material and methods 69 Natural history 69 Study area 70 Sampling design and sample analyses 70 Statistical analysis 72 Results 73 Discussion 76 References 83 Capítulo 4 Efficiency of pollination and satiation of predators determine reproductive output in Iberian Juniperus thurifera woodlands. 89 Abstract 90 Introduction 91 Material and methods 93 Natural history of the system 93 Population variation in fruit production and seed set 94 Population variation in fruit damage by arthropods 97 Variation in seed output across populations 98 Results 98 Variation in seed number 98 Variation in the proportion of full, empty and aborted seeds 98 Variation in fruit damage by arthropods 100 Seed output across populations 100 Discussion 101 References 107 Supplementary material 113 Capítulo 5 You’d better walk alone: Changes in forest composition affect pollination efficiency and predispersal fruit damage in Iberian Juniperus thurifera forests. 115 Abstract 116 Introduction 117 Material and methods 119 Study area and sampling design 119 Study species 119 Fruit production and seed set 120 Predispersal fruit damage 122 Seed output 122 Statistical analysis 123 Results 124 Discussion 127 References 131 Agradecimientos 137 11 Resumen El desarrollo de frutos y semillas es un factor crítico en la demografía de las plantas. La cantidad de frutos y semillas que se desarrollan correctamente depende de múltiples factores: disponibilidad y uso de los recursos para la reproducción, éxito en la polinización y fertilización y depredación predispersiva. Estos factores actúan a escala individual o poblacional conjuntamente con otros factores como las condiciones climáticas, la densidad de individuos o la composición específica de los bosques. La sabina albar (Juniperus thurifera L.) muestra una alta variabilidad interanual en la producción de gálbulos (denominados indistintamente frutos, aunque desde un punto de vista botánico sean pseudofrutos). Por otra parte, las semillas muestran una viabilidad baja lo que puede estar relacionado tanto con la disponibilidad de recursos, como con la eficiencia de la polinización. Comprender los mecanismos que determinan el éxito reproductivo de la sabina albar y cómo se ven afectados por diferentes motores del cambio global es crítico para conocer el futuro de esta especie. La presente tesis se estructura en cinco capítulos en los que se exploran los diferentes factores que limitan la producción de gálbulos y semillas viables en la sabina albar a diferentes escalas temporales y espaciales. En el capítulo 1 se expone el marco teórico de la tesis, se muestra su diseño, así como los principales resultados y conclusiones obtenidos en la misma. En el capítulo 2 se monitoreó el proceso de maduración de los gálbulos de sabina albar y la incidencia y selección de gálbulos de sus depredadores predispersivos durante tres cohortes de frutos entre los años 2013 y 2016. La mayor pérdida de gálbulos ocurre durante los cinco primeros meses de su desarrollo. Esta pérdida ocurre simultáneamente con la entrada de los principales depredadores predispersivos (ácaros, polillas, cochinillas y avispas). La identificación de los criterios de selección de gálbulos por parte de los artrópodos fue más eficiente si se realizaba en frutos inmaduros, que en maduros tal y como suele realizarse. En el capítulo 3 se evaluó el papel que juegan los niveles de recursos, el éxito en la polinización y la depredación predispersiva en el éxito reproductivo. Se realizó un experimento con un diseño factorial completo que incluía dos tratamientos aporte extra de polen y de pes- Capítulo 1 18 las características de los frutos para hacerlos menos atractivos a los depredadores (Janzen 1969; Espelta et al. 2009; Beckman y Muller-Landau 2011) o la alta variabilidad interanual en la producción de frutos y semillas, que obliga a los depredadores a ajustar sus ciclos demográficos a los ciclos de producción de frutos, aumentando las plantas su éxito reproductivo durante los años de alta producción (Turgeon 1994). El éxito reproductivo es modulado, además, por factores que actúan a mayores escalas espaciales, como las condiciones climáticas (García et al. 2000; Obeso 2002), la fragmentación del hábitat (Sork et al. 2002; Knight et al. 2005), la densidad de individuos y su estructura (Sork et al. 2002; Knight 2003; Sanz and Pulido 2015) o la presencia de otras especies (Mugnaini et al. 2007; Aderkas et al. 2012). De hecho, a una escala geográfica amplia es el clima quien determina la variabilidad en la producción de frutos que existe entre diferentes poblaciones (García et al. 2000; Obeso 2002; Montesinos et al. 2010), mientras que, a una escala local, variaciones temporales del clima determinan la inversión reproductiva (Lee y Bazzaz 1982; Herrera 1991; Crone y Lesica 2006). Para especies con una fuerte variación interanual en su producción de frutos hay una clara correlación entre las condiciones climáticas y una alta producción (Kelly and Sork 2002). El éxito reproductor no es, por tanto, ajeno a la configuración espacial de la vegetación, de modo que los cambios en las características del territorio pueden afectarla. La conversión de bosques en pastos y cultivos es uno de los principales motores del cambio global, especialmente en países en vías de desarrollo (Gibbs et al. 2010, Phelps et al. 2013; Fig. 2). Sin embargo, el patrón opuesto se puede observar en países desarrollados, donde el éxodo rural y la intensificación de las prácticas agrarias ha llevado a un abandono de los usos tradicionales y las zonas menos productivas (Rey Benayas et al. 2007, Valladares et al. 2014). Como resultado, se está produciendo la expansión de los bosques hacia antiguos campos de cultivo (Gimeno et al. 2012), la densificación de las masas forestales (Rey Benayas et al. 2007, Améztegui et al. 2010) o cambiando su composición específica (Hansen et al. 2001, Chauchard et al. 2007, Vayreda et al. 2016). Por un lado, estos cambios en la estructura de los bosques llevan a una mayor competencia por los recursos, a nivel intraespecífico por la densificación (Kenkel 1988, Getzin et al. 2006, Wang et al. 2016) e interespecífico por la entrada de nuevas especies (Costa et al. 1997, Montesinos and Fabado 2015). Por el otro lado, la regeneración del bosque y colonización de nuevas áreas reduce la fragmentación y aumenta la eficiencia en la polinización, tanto por el viento como por animales, así como Una visión general de la tesis 19 de la dispersión (Santos and Tellería 1994, González-Varo et al. 2009). Por el contrario, una densificación de los bosques puede resultar también en una mayor depredación predispersiva de frutos y semillas (Sholes 2008; Guyot et al. 2016) y en el caso de bosques mixtos, la presencia de otras especies puede reducir la eficiencia en la polinización, especialmente por el viento, debido a un efecto barrera o por interferencia polínica entre especies que solapen sus periodos reproductivos (Mugnaini et al. 2007, Aderkas et al. 2012, Millerón et al. 2012). ¿Estoicismo o resignación? El caso de la sabina albar La sabina albar (Juniperus thurifera L.; Fig. 3) es un endemismo mediterráneo occidental con sus principales poblaciones en la Península Ibérica y Marruecos, presentando relictos en Francia, Italia y Argelia (Costa et al. 1997). Se trata de una especie que aparece en zonas de clima mediterráneo de marcada continentalidad y sequía estival, generalmente en suelos pobres predominantemente calcáreos, aunque también en suelos silíceos (p.e. Sierra de Guadarrama en Segovia o Barrios de Luna en León) (Costa et al. 1997). Además, los sabinares albares están incluidos en el anexo IV de la directiva hábitats (código 9560) y si bien la especie está catalogada como preocupación menor (LR/lc) según la Unión Fig. 2 Bosque de araar (Tetraclinis articulata) adehesado en Marruecos. Capítulo 1 20 Internacional para la Conservación de la Naturaleza (UICN) sus poblaciones se encuentran severamente fragmentadas y en continuo decrecimiento (Farjon 2013). La estructura actual de los sabinares albares es consecuencia de un manejo tradicional ligado a la ganadería ovina y caprina (Olano et al. 2008; Fig. 3 A y B), uso que aún persiste en el Atlas marroquí si bien en España no ha desaparecido por completo todavía. Durante la segunda mitad del siglo XX los usos tradicionales de los sabinares han ido cesando su actividad lo que está generando cambios en la estructura de estos bosques (Fig. 3C). Así, están entrando nuevas especies arbóreas como encinas, robles o pinos y se está produciendo una densificación de los sabinares (Olano et al. 2012). La dispersión de la sabina albar se basa en el consumo de sus pseudofrutos carnosos, gálbulos o conos por un amplio grupo de vertebrados, principalmente aves del género Turdus (Tellería et al. 2011) y carnívoros como el zorro (Vulpes vulpes (Linnaeus 1758)) o la garduña (Martes foina (Erxleben 1777)) que son muy eficientes para la dispersión en espacios abiertos (Escribano-Ávila et al. 2012). Sin embargo, la cantidad de gálbulos viables listos para la dispersión está afectada por múltiples factores y puede ser muy baja (Montesinos 2010). Parte de esta limitación reside en la fuerte variabilidad interanual, ligado a la estrategia de reducción de las elevadas tasas de depredación predispersiva por artrópodos (Mezquida y Olano 2013). Mientras que, por otra parte, la sabina albar presenta una reducida viabilidad de las semillas (Montesinos et al. 2010) lo que puede estar relacionado tanto con la eficiencia de la polinización y la disponibilidad de recursos. En general, estos efectos podrían agudizarse en un futuro cercano, como consecuencia de los cambios de uso y del cambio climático. En otras especies de Juniperus se ha predicho un aumento en la tasa de pérdida de gálbulos por factores bióticos y abióticos, reduciendo el número de semillas viables como se ha indicado para otras especies del género (Gruwez et al. 2013), además de una posible disminución de las tasas de dispersión de frutos y semillas. Aunque la sabina albar muestra una gran capacidad de adaptación a variaciones en las condiciones climáticas, haciéndola muy resiliente frente a un futuro escenario de clima más cálido y seco (Camarero et al. 2010), las tasas de depredación predispersiva por artrópodos y aborto de semillas son especialmente sensibles a las condiciones ambientales (Montesinos et al. 2010), pudiendo ser un cuello de botella para el futuro de la especie. Los insectos que se alimentan de frutos y semillas normalmente muestran preferencias por determinados caracteres (Sallabanks y Courtney 1992). En el caso de la sabina albar, existe un diverso grupo de artrópodos que se alimentan de los gálbulos, reduciendo la cantidad Una visión general de la tesis 21 Fig. 3. A: Sabinas albares en Siguero (Segovia). B: Sabinar albar en Calatañazor (Soria). C: Sabinar albar mixto en Megina (Guadalajara) A B C Capítulo 1 22 de ellos que llegan a un estado óptimo para su dispersión (Roques 1984; ver pág. 69-70). En el caso de la península Ibérica este grupo de artrópodos está formado principalmente por: los ácaros Trisetacus quadrisetus (Thomas 1889) (Acari, Phytoptidae), las polillas Pammene juniperana (Millière 1858) (Lepidoptera, Tortricidae) y Mesophleps oxycedrella (Milière 1871) (Lepidoptera, Gelechiidae) y la avispa Megastigmus thuriferana Roques & El Alaoui 2006 (Hymenoptera, Torymidae). Estas especies atacan secuencialmente los gálbulos durante su desarrollo mostrando preferencias por determinadas características de los gálbulos. Esta depredación de gálbulos secuencial lleva a una aborción selectiva de los gálbulos infestados en las fases tempranas de su desarrollo. Además, cuando los gálbulos son atacados en fases de desarrollo similares a la madura, los artrópodos depredadores podrían seleccionar gálbulos con características similares a las preferidas por los dispersores, generando presiones selectivas conflictivas (Siepielski y Benkman 2007). Objetivos El objetivo principal de esta tesis es comprender los diferentes factores que limitan la producción de gálbulos y semillas viables en la sabina albar, así como comprender el impacto y selección de frutos por parte de los depredadores predispersivos, determinando su respuesta a diferentes escalas temporales y ambientales. Los trabajos de investigación llevados a cabo para esta tesis se organizan en cuatro capítulos cuyos objetivos específicos son: Capítulo 2. Realizar un seguimiento del proceso de maduración de los gálbulos de sabina albar, así como del proceso de entrada y biología de los depredadores predispersivos. De esta forma, podremos observar cómo se ajustan los ciclos vitales de los depredadores al desarrollo de los gálbulos y cuáles son los criterios de selección de frutos para cada predador. El seguimiento de los gálbulos y de su tasa de depredación en diferentes momentos permitirá detectar procesos de aborción selectiva de los gálbulos infestados, que pudiera implicar una subestimación de los niveles de depredación predispersiva ejercidos por cada especie. Con ello se contribuirá a mejorar nuestra comprensión de las presiones selectivas que ejercen los depredadores cuando basamos nuestras observaciones e interpretaciones sólo en fases maduras de los frutos. Una visión general de la tesis 23 Capítulo 3. Discernir el papel que juegan los niveles de recursos, el éxito en la polinización y la depredación predispersiva y cómo interactúan entre ellos para determinar el éxito reproductivo final en la sabina albar. Con ello se pretende comprender mejor cómo interactúan entre sí los diferentes filtros que determinan el número final de gálbulos que llegan a madurar. Este trabajo modificó experimentalmente los niveles de polen y de depredadores predispersivos. Capítulo 4. Evaluar a una escala geográfica amplia el impacto combinado de las condiciones climáticas y de la depredación predispersiva sobre la producción de gálbulos y la calidad de las semillas. Con ello se trata de comprender cómo se comportan la sabina y sus depredadores bajo un gradiente de condiciones climáticas y cómo estas condiciones interactúan con los sucesos locales. Capítulo 5. Evaluar cómo la configuración de los sabinares puede afectar a la producción de gálbulos y niveles de depredación predispersiva. Se realizó un diseño muestral comprendiendo bosques puros y mixtos de sabina en diferentes regiones geográficas. En este trabajo se trata de observar de qué modo la tendencia de los sabinares a convertirse en bosques mixtos pueden afectar el éxito reproductor de la especie. Metodología Esta tesis aborda la cuestión con un enfoque observacional, combinado con un diseño experimental en el capítulo 3. Se recolectaron gálbulos en diferentes fases de desarrollo de diversos sabinares de la península Ibérica en los que se ha realizado trabajo de campo entre los años 2012 y 2016 durante la duración de la tesis. Además, en el capítulo 4 se dispuso de muestras recolectadas en años anteriores por el equipo de trabajo. En el capítulo 2 se monitorizó el desarrollo de los frutos de sabina albar. Se seleccionaron cien ramas en veinte árboles, a razón de cinco ramas por árbol, en el sabinar de Villaciervos (Soria) y se muestreó regularmente entre febrero 2013 y octubre de 2016 durante el desarrollo de tres cohortes de frutos. En cada muestreo se contaron los frutos presentes en las ramas seleccionadas. Además, se recolectaron regularmente gálbulos de otras ramas para su disección en laboratorio con el fin de observar el desarrollo de las larvas de los depredadores, estimar la depredación predispersiva y evaluar la selección de frutos por los Capítulo 1 24 artrópodos en diferentes fases fenológicas. Estos datos se compararon con los resultantes de los gálbulos maduros que se recogían en octubre de su segundo año de desarrollo. En el capítulo 3 se trató de discernir experimentalmente el papel que juegan los niveles de recursos, el éxito en la polinización y la depredación predispersiva en el desarrollo de gálbulos. Para ello se realizó un experimento con un diseño factorial completo que incluía diferentes tratamientos: A) control, B) aporte extra de polen, C) aporte de pesticidas y D) aporte extra de polen + pesticidas. Se realizó en el mismo sabinar que el estudio anterior, en Villaciervos (Soria). Para ello se seleccionaron al azar cuarenta árboles. En cada árbol se seleccionaron 8 ramas y se aplicó cada uno de los tratamientos a dos ramas por árbol. El experimento se realizó para la cohorte de gálbulos del año 2014. Se realizó además un seguimiento mensual de la pérdida de frutos. En octubre de 2015 se recolectaron todos los gálbulos maduros de las ramas tratadas para estimar la depredación por artrópodos y analizar la calidad de las semillas. Con el fin de observar el efecto del aporte extra de polen sobre la fertilización en junio de 2015 se recolectaron 100 gálbulos inmaduros (50 control y 50 con aporte extra de polen). Estas muestras fueron procesadas para la observación, en microscopio óptico, de la formación de tubos polínicos (ver metodología en Capítulo 3). La influencia del esfuerzo reproductivo inicial, el aporte extra de polen y la aplicación de pesticidas sobre el desarrollo de gálbulos, la calidad de las semillas y la depredación predispersiva se testó mediante modelos mixtos. En el capítulo 4 se evalúo cómo influyen las condiciones ambientales en el desarrollo de gálbulos y su efecto combinado sobre la calidad de las semillas y la depredación predispersiva a una escala geográfica amplia. Para ello se analizaron muestras provenientes de catorce sabinares a lo largo de su distribución en la península Ibérica recolectadas durante los años 2007 y 2008. En cada sabinar fueron seleccionados al azar treinta sabinas en las que se estimó la producción de gálbulos y se recolectaron treinta gálbulos para la estima de la depredación predispersiva y diez para el análisis de sus características y de la calidad de sus semillas. La influencia combinada de las condiciones ambientales y la producción de gálbulos sobre la calidad de las semillas y la depredación predispersiva se evaluó mediante modelos mixtos. En el capítulo 5 se evaluó cómo pueden afectar los cambios en la estructura y composición específica de los sabinares a la reproducción final. Se seleccionaron bosques puros con una Una visión general de la tesis 25 clara dominancia de la sabina albar (> 90 % de representación en pies) y bosques con baja representación (< 50 %) obviando los estados intermedios (51 – 89 %). Con este esquema, durante otoño de 2014 se realizó un diseño pareado con sabinares puros y mixtos en diferentes localidades del centro de la Península Ibérica: Megina, Pinilla de Molina y Valhermoso (Guadalajara); Arcones, Prádena y Siguero (Segovia); Bayubas de Abajo, Calatañazor y Cabrejas del Pinar (Soria). En cada sabinar fueron seleccionadas al azar treinta sabinas en las que se estimó la producción de gálbulos y se recolectaron treinta gálbulos para la estima de la depredación predispersiva y diez para el análisis de sus características y calidad de sus semillas. El efecto de la composición del bosque sobre la producción de gálbulos y la calidad de sus semillas se evaluó mediante modelos mixtos. Los procedimientos de análisis de los gálbulos recolectados fueron semejantes para todos los capítulos. Se tomaron medidas del tamaño con un calibre digitial y se identificó la actividad de los artrópodos depredadores mediante las señales que estos dejan en los gálbulos maduros o bien por su presencia en el interior del fruto (ver una clave detallada en el capítulo 2). Además, en una porción de los gálbulos fueron extraídas las semillas y tras su secado en estufa a 60 ºC durante 48 horas se pesaron pulpa y semillas por separado. Para las semillas se observó su estado de desarrollo y su calidad. En función de dichos parámetros y siguiendo la clasificación propuesta por Gruwez et al. (2013) fueron asignadas a una de las siguientes tres categorías: Llena (Fig. 4A), si la semilla muestra un desarrollo completo con un embrión y megagametofito bien desarrollados; vacía (Fig. 4B y C) si la semilla no muestra un desarrollo completo del embión o megagametofito o estas estructuras están ausentes pero la semilla se muestra externamente como una semilla viable y abortada (Fig 4D) si la semilla fue abortada antes de la fertilización, siendo reconocible como una semilla deformada, no completamente desarrollada. Los datos obtenidos del análisis de las muestras se han analizado empleando modelos mixtos. Se denominan así porque combinan efectos fijos y aleatorios (Zuur et al. 2009). Este tipo de modelos estadísticos siguen la misma lógica que otros tipos de modelos por la que se trata de describir la relación entre una variable respuesta y una o varias variables (efectos) explicativas fijas. Estos modelos se emplean cuando dada la naturaleza de la toma de datos en los experimentos, las observaciones pueden no ser independientes entre sí. Así, por ejemplo, cuando se disponen de datos provenientes de “n” poblaciones, las observaciones realizadas en una determinada población se parecerán más entre sí que con las de otras poblaciones. El efecto que tienen las diferentes poblaciones sobre la variable respuesta puede no ser de interés para el Capítulo 1 26 experimentador, sino que interesa conocer en qué medida la estimación de los valores de la variable respuesta pueden variar en función de la población (niveles del factor) que se seleccione de un modo aleatorio. De esta forma el factor “población” representaría el efecto aleatorio en el modelo mixto. Estos modelos pueden ser empleados tanto para variables que siguen una distribución normal (modelos mixtos lineales; capítulo 4 y 5) como para variables que siguen otro tipo de distribuciones, como la binomial o la Poisson (modelos mixtos lineales generalizados; capítulos 2 a 5). En cada capítulo de la tesis, los efectos fijos se eligieron en base a los conocimientos biológicos previos. En el capítulo 2 con el fin de simplificar la comparación de modelos 1 mm 1 mm 1 mm 1 mm Fig 4. Diferentes estados identificados en las semillas de sabina albar. A: Semilla llena, mostrando un embrión y megagametofito bien desarrollados. B: Semilla vacía con embrión y megagametofito sin desarrollar por completo. C: Semilla vacía sin embrión ni megagametofito desarrollados. D: Semilla abortada sin desarrollo post-fertilización.. A B C D Una visión general de la tesis 27 obtenidos en las diferentes fases fenológicas sólo se consideraron modelos saturados. En el capítulo 3 se realiza una selección de modelos eliminando secuencialmente las variables menos explicativas hasta obtener un modelo óptimo sencillo. En el capítulo 4 con el fin de obtener información general de los posibles efectos fijos a una escala geográfica amplia sólo se consideraron modelos saturados. Finalmente, en el capítulo 5 se realizó primero una selección de la estructura de efectos aleatorios sobre los modelos saturados y, segundo, una selección de estructura de efectos fijos tal como se indica en Zuur et al. (2009). Esta selección de estructuras se realizó en base al criterio de información bayesiana (BIC) recomendable cuando el objetivo es evaluar hipótesis (Aho et al. 2014). Principales resultados En el capítulo 2 se describió el ciclo vital, asociado al desarrollo de los gálbulos, de los depredadores. Los diferentes artrópodos depredadores atacaron secuencialmente los conos y semillas de la sabina albar. Los ácaros colonizaron las flores a la espera, tras la polinización, que se inicie el desarrollo de semillas para su infestación. Salieron del interior de las semillas durante el invierno en busca de nuevas flores que colonizar. Se identificaron las dos especies de polillas, M. oxycedrella y P. juniperana. Estas realizaron la puesta de huevos sobre los gálbulos, de primer y segundo año, desde finales de mayo a mediados de junio, entrando sus larvas en los gálbulos desde mediados de junio a finales de julio. Las larvas se desarrollaron en el interior del gálbulo y salieron de su interior a partir de finales de octubre. Las cochinillas colonizan nuevos gálbulos en su fase de ninfa durante el mes de junio. Una vez seleccionado el lugar idóneo sufren la metamorfosis a la fase sedentaria y adulta completando su ciclo vital. Finalmente, las avispas depositaron huevos en el interior de las semillas desde finales de julio hasta mediados de septiembre. Las larvas se desarrollaron en el interior de las semillas a lo largo de doce meses, saliendo de su interior desde finales de julio hasta finales de agosto. La identificación de los criterios de selección de gálbulos por parte de los adultos fue diferente si se basaba en gálbulos inmaduros o maduros, siendo en general más robusto cuando se escogían gálbulos inmaduros recogidos poco después de la entrada del depredador. Así durante el primer año de formación del cono: la polilla M. oxycedrella selecciona conos de mayor tamaño y con un mayor número de semillas; mientras la polilla P. juniperana selecciona los mismos árboles que M. oxycedrella pero selecciona gálbulos previamente ocupados por ácaros de los que previsiblemente se alimenta; la avispa M. thuriferana selecciona Capítulo 1 34 Montesinos D., Fabardo J. (2015) Changes in land use and physiological transitions of Juniperus thurifera forest: from decline to recovery. Canadian Journal of Forest Research 45: 764–769 Mugnaini S., Nepi M., Guarnieri M., Piotto B., Pacini E. (2007) Pollination drop in Juniperus communis: response to deposited material. Annals of Botany 100: 1475-1481 Obeso J.R. (2002) The costs of reproduction in plants. New Phytologist 155: 321–348. Olano J.M., Rozas V., Bartolomé D., Sanz D. 2008. Effects of changes in traditional management on height and radial growth patterns in a Juniperus thurifera woodland. Forest Ecology and Management 255: 506-512 Olano J.M., Zavala M.A., Rozas V. 2012. Disruption of Juniperus thurifera woodland structure in its northwestern geographical range: potential drivers and limiting factors. European Journal of Forest Research 131: 563-570. Östergård H., Hambäck P.A., Ehrlén, J. (2007). Pre-dispersal seed predation: The role of fruit abortion and selective oviposition. Ecology 88: 2959–2965. Pessendorfer M.B., Koenig W.D., Pearse I.S., Knops J.M.H., Funk K.A. (2016) Individual resource limitation combined with population-wide pollen availability drives masting in the valley oak (Quercus lobata). Journal of Ecology 104: 637–645. Phelps J., Carrasco L.R., Webb E.L., Koh L.P., Pascual U. (2013) Agricultural intensification escalates future conservation costs. Procedings of the Natural Academy of Sciences 110: 7601-7606. Pías B., Salvande M., Guitián P. (2007) Variation in Predispersal Losses in Reproductive Potential in Rowan (Sorbus aucuparia L. Rosaceae) in the NW Iberian Peninsula. Plant Ecology 188: 191–203. Price M.V., Campbell D.R., Waser N.M., Brody A.K. (2008) Bridging the generation gap in plants: pollination, parental fecundity, and offspring demography. Ecology 89:1596–1604. Rech A.R., Dalsgaard B., Sandel B., Sonne J., Svenning J.C., Holmes N., Ollerton J. (2016) The macroecology of animal versus wind pollination: ecological factors are more important than historical climate stability. Plant Ecology & Diversity 9: 253–262. Rey Benayas J.M., Martins A., Nicolau J.M., Schulz J.J. (2007) Abandonment of agricultural land: an overview of drivers and consequences. CAB reviews: perspectives in agriculture, veterinary sciences, nutrition and natural resources 2. Wallingford (UK): CABI Publishing. 14p. Roques A., Raimbault J.P., Goussard F. (1984). La colonisation des cônes et des galbules des Genévriers Méditerranéens par les insectes et les Acariens et son influence sur la possibilité de régénération Naturelle de ces essences. Ecologia Mediterranea X: 147–169. Rouault G., Turgeon J., Candau J.N., Roques A., von Aderkas P. (2004) Oviposition strategies of conifer seed chalcids in relation to host phenology. Naturwissenschaften 91: 472–480. Sallabanks R., Courtney S.P. 1992. Frugivory, seed predation, and insect-vertebrate interactions. Annual Review of Entomology 37: 377-400. Una visión general de la tesis 35 Santos T., Tellería J.L. (1994) Influence of forest fragmentation on seed consumption and dispersal of Spanish juniper Juniperus thurifera. Biological Conservation 70: 129–134 Sanz R., Pulido F. (2015). Pollen limitation and fruit abortion in a declining rare tree, the Eurasian yew (Taxus baccata L.): A reproductive cost of ecological marginality. Plant Biosystems 149: 818–826. Schoonhoven L.M., van Loon J.J.A., Dicke M. (2005). Insect-Plant Biology. Oxford (U.K.): Oxford University Press. Siepielski A.M., Benkman C.W. (2007). Selection by a predispersal seed predator constrains the evolution of avian seed dispersal in pines. Functional Ecology 21: 611-618. Sholes O.D.V. (2008) Effects of associational resistance and host density on woodland insect herbivores. Journal of Animal Ecology 77: 16-23. Sork V.L., Davis F.W., Smouse P.E., Apsit V.J., Dyer R.J., Fernandez J.F. Kuhn B. (2002). Pollen movement in declining populations of California Valley oak, Quercus lobata: Where have all the fathers gone? Molecular Ecology 11: 1657–1668. Teitel Z., Pickup M., Field D.L., Barrett S.C.H. (2016). The dynamics of resource allocation and costs of reproduction in a sexually dimorphic, wind-pollinated dioecious plant. Plant Biology 18: 98–103. Tellería J.L., de la Hera I., Ramirez A., Santos T. (2011). Conservation opportunities in Spanish juniper Juniperus thurifera woodlands: the case of migratory thrushes Turdus spp. Ardeola 58: 57-70 Turgeon J. (1994) Insect exploiting seed cones of Larix spp., Picea spp., and Pinus spp.: species richness and patterns of exploitation. In: Turgeon J.J., de Groot P. (eds). Biology, damage and management of seed orchard pests. Ontario (Canada): Canadian Forestry Service. pp 15–30 Valladares F., Benavides R., Rabasa S.G., Díaz M., Pausas J.G., Paula S., Simonson W.D. (2014) Global Change and Mediterranean forests: current impacts and potentias responses in Forest and Global Change. In Coomes D.A., Burslem D.F.R.P., Simonson W.D., editors. Forest and Global Change. Cambridge (U.K.): Cambridge University Press. p47-75. Vayreda J., Martinez-Vilalta J., Gracia M., Canadell J., Retana J. (2016) Anthropogenic-driven rapid shifts in tree distribution lead to increased dominance of broadleaf species. Global Change Biology doi: 10.1111/gcb.13394 Walsh R.P., Arnold P.M., Michaels H.J. (2014). Effects of pollination limitation and seed predation on female reproductive success of a deceptive orchid. AoB PLANTS 6: 1–12. Wang Y., Pederson N., Ellison A.M., Buckley H.L., Case B.S., Liang E., Camarero J.J. (2016) Increased stem density and competition may diminish the positive effects of warming at alpine treeline. Ecology 97: 1668-1679. Xia K., Harrower W.L., Turkington R., Tan H.Y., Zhou Z.K. (2016). Pre-dispersal strategies by Quercus schottkyana to mitigate the effects of weevil infestation of acorns. Scientific Reports 6: 37520. https:// doi.org/10.1038/srep37520 Capítulo 1 36 Zuur A., Leno E.N., Walker N., Saveliev A.A., Smith G.M. (2009) Mixed effects models and extensions in ecology with R. Springer. New York (USA): Springer. p549. CHAPTER 2 Ripe fruits may not be optimal: green fruits improve the detection of traits associated to predispersal predation risk “Weeks in the field can save you minutes in the library” Thomas Eisner CHAPTER 2 Capítulo 2 40 Abstract Predispersal seed predation by arthropods has been described as one of the main selective pressures on plants. Arthropods show preferences for certain fruit and seed traits to optimize their fitness. In response, plants have shaped their reproductive parts to reduce losses from seed predators along with enhancing their reproductive output. Fruit and seed preferences by arthropods are usually explored on ripe fruits, but fruit choices usually occur in early phases of fruit formation, and the implications of differences in the timing of analyses are not fully understood. Our aim was to explore whether the analysis of fruit preferences by predispersal seed predators differ when assessed using immature versus ripe fruits. To perform this task, we monitored fruit production and predispersal predation during three cohorts (four years) in a Spanish juniper (Juniperus thurifera) woodland in central Spain. We explored fruit preferences by arthropods using linear mixed models at different phenological stages: fruits collected shortly after the oviposition period of each species and ripe fruits. To perform this task, we assessed the timing of fruit development and predispersal predation. We monitored the fate of 11,480 flowers and analysed predation in 3,740 immature and 2,342 ripe fruits. A complex guild of arthropods predated on juniper fruits. Loss of fruit was very high during the first five months of fruit development, matching the entry of the main predispersal predators. Predation rates for most of the species declined from immature to ripe fruit indicating the selective abortion of infested fruits, although this response was species-specific leading to rank reversals in relative predation impact. Fruits collected just after arthropod entry provided better models for arthropod preferences for fruit traits for most of the species. Evaluation of preferences by different arthropods attacking fruits in the Spanish juniper was improved through multiple fruit sampling matching the timing of arthropods’ entry. This result can probably be extrapolated to other species which have seeds with long maturation periods. and claims for the need of developing a deeper knowledge on natural history to improve our understanding of the biological interactions behind ecological and evolutionary processes. Ripe fruits may not be optimal 41 Introduction Arthropods and plants are among the largest taxa of living organisms in number of species and in terrestrial ecosystem biomass, respectively. Species in both groups interact in different ways with each other, and close interactions has been traditionally considered as the promoters of the current diversity of terrestrial life forms (Ehrlich and Raven 1964). In fact, 90 % of flowering plants are estimated to be pollinated by animals, mainly by insects (Ollerton et al. 2011) and 26 % of the insect species feed on vascular plants (Schoonhoven et al. 2005). Interactions between plants and insects had critical economic implications especially in agricultural systems (Rechcigl and Rechcigl 2000; Schoonhoven et al. 2005, Badii et al. 2015) and are basic for the functioning of natural systems from an ecological and evolutionary point of view (Janzen 1971; Kolb et al. 2007; Suchan and Alvarez 2015). Predispersal seed predators may limit plant reproductive success (Crawley 2000). Indeed, predispersal seed predation has been described as one of the main selective pressures in plants (Kolb et al. 2007). Plants have developed multiple evolutionary strategies to minimize the detrimental effects of seed predators, including the accumulation of secondary metabolites, changes in fruit and seeds traits or inter-annual variation in fruit crop (Janzen 1969, 1971; Crawley 2000; Kolb et al. 2007; Schoonhoven et al. 2005). Natural history is critical to decipher the ecological and evolutionary consequences of plant-seed predators’ interactions. This process becomes more complex when multiple species share the same host plant. Predators must develop life history traits to maximize predation and minimize interspecific competition (Espelta et al 2009; Bonal et al. 2011), whereas plants must deal with contrasting selective pressures (Gagic et al. 2016). The different agents must adjust the phenology of their vital events, including those involving predation pressures, to optimize their fitness (Gillot 2005; Bonal et al. 2010, 2011; Ehrlén 2015). Nevertheless, the inclusion of natural history in ecological research to explain plant-insect relationships and its consequences is scarce (see Östergård et al. 2007; Espelta et al. 2009, Xia et al. 2016). In fact, predispersal seed predation is routinely evaluated on ripe fruits and seed predator preferences assessed based on the characteristics of ripe fruits (Sperens 1997; García et al. 2000; Mezquida and Olano 2013; Mezquida et al. 2016; Moreira et al. 2016), assuming that this relationship reflects the processes occurring at earlier vital stages. Capítulo 2 42 Our aim was to explore whether the common practice of exploring predator-plant selection using ripe fruits provides similar information to evaluating this relationship at the phenological stage corresponding to the timing of fruit use by each seed predator. To perform this task, we selected the conifer Spanish juniper (Juniperus thurifera), which produces pseudo-fruits (hereafter fruits) that mature over a 20 months period when they are attacked by a wide array of specialized arthropods, including mites, moths, chalcid wasps and scale insects (Roques et al. 1984). The ecological and evolutionary implications of seed predation in this species and other junipers has received an intense research effort which has been based on the evaluation of ripe fruits (Mezquida and Olano 2013; Roques et al. 2013; El Alaoui et al. 2013; Mezquida et al. 2016). We hypothesized that plants having long seed maturiation periods, levels of seed predation might be underestimated due to selective fruit losses of infested fruits with different abortion rates determining changes in their relative infestations ranks. Moreover, we predicted that the set of traits associated to predator preferences may differ from the traits evaluated on ripe fruits, and that these differences would be more intense for seed predators showing higher losses of damaged fruits. Material and methods Natural history Spanish juniper (Juniperus thurifera L.) is a dioecious juniper species, endemic to the Western Mediterranean where it inhabits continental summer-dry environments. Wind pollination occurs in late winter-early spring. Pollen germination and megagametophyte formation starts after pollination, but fertilization is delayed until late spring. Embryo development begins after fertilization and seeds are almost fully grown by mid-summer, when they start to harden their seed coat. Megagametophyte and embryo final maturation takes 12-13 additional months (Gruwez et al. 2013). Juniper fleshy cones (hereafter fruits) ripe 20 months after pollination in next year autumn (Montesinos et al. 2012). Thus, individuals display two fruit cohorts from March to November. A variety of arthropods can damage Spanish juniper fruits during this long development period (Roques et al. 1984; El Alaoui et al. 2013). However, in the Iberian Peninsula, three arthropod taxa are the main predispersal seed predators: mites, moths and chalcid wasps. Ripe fruits may not be optimal 43 Mites, Trisetacus quadrisetus (Thomas 1889) (Acari, Phytoptidae), uses seeds as a growth chambers where they feed and reproduce (Roques et al. 1984). The larvae of two moth species: Mesophleps oxycedrella (Millière 1871) (Lepidoptera, Gelechiidae) (yellow juniper moth hereafter) and Pammene juniperana (Millère 1858) (Lepidoptera, Tortricidae) (brown juniper moth hereafter) feed on juniper fruit pulp and seeds (Roques et al. 1984). The larva of a chalcid wasp, Megastigmus thuriferana Roques & El Alaoui 2006 (Hymenoptera, Torymidae) feeds on well-developed megagametophyte and embryo (Rouault et al. 2004). Other frequent arthropods that feed on Spanish juniper fruits include the pulp sucker scale insect Carulaspis juniperi (Bouché 1851) (Hemiptera, Diaspididae) and the pulp eater juniper fly Rhagoletis zerny Hendel 1927 (Diptera, Tephritidae). Study area Sampling site was located in Villaciervos, Soria province, Central Spain (41º 44’ N, 2º 40’ W; 1150 m.a.s.l.). Climate is Mediterranean continental, mean annual precipitation is 533 mm. Mean monthly temperatures range from 2.8 ºC in January to 20 ºC in July. The rock parent material is Cretaceous limestone covered by deep soil with small areas of exposed bedrock. Vegetation is dominated by a J. thurifera woodland with open areas covered by xeric grasslands with shrubs, including Cistus laurifolius L. J. communis L. and Thymus zygis Loefl. Sampling design and fruit analysis Two monitoring schema were preformed simultaneously. First monitoring evaluated fruit success on 20 randomly selected female junipers. We considered fruit success as the number of initial flowers that complete their development to ripe fruits. Five flowering branches from all around the canopy were selected and tagged for each tree. Monitoring was initiated in March 2013 and maintained until October 2016. Thus, monitoring comprised three fruit cohorts (2013, 2014 and 2015). Initial reproductive effort for each branch was estimated by counting the number of female flowers for each cohort in March from 2013 to 2015. Fruit development during the first year was monitored by counting the fruits within the branches monthly from May to September for the three cohorts. Fruit development during the second year was monitored monthly from March to October for the 2013 and 2014 cohorts and in March, April, September and October for the 2015 cohort. In October, when fruits are Capítulo 2 50 3.6 % in August 2015). Estimation of damage rates using ripe fruits was considerably lower for the 2013 cohort (25.4 ± 1.8 %) but of similar magnitude for the 2014 cohort (32.4 ± 1.4 %). Scale insects showed the highest incidence, reaching a maximum of 40.4 ± 3.5 % in September 2013 (Table 3). Yellow juniper moth was the next in importance, reaching 19.2 ± 2.8 % in July 2013. Mites (15.4 ± 2.3 % in June 2013) and chalcid wasps (16.0 ± 2.7 % in September 2014) showed damage rates in the same range. Brown juniper moths had lower, but still significant damage rates with a peak in 2013 August with 7.7 ± 1.7 %. Juniper flies had a moderate impact in 2013 (9.3 ± 2.1 % in September 2013), but was negligible for the other cohorts. Comparison of these values with predispersal damage rates based just on ripe fruits from fruit success monitoring revealed large underestimations of predispersal damage rates for all species but chalcid wasps. Fruit damage monitoring revealed variation in damage rates at seasonal and inter-annual scales. At seasonal scale percentage of fruit damage peaked one or two months after arthropods entry and then declined for all species except for chalcid wasp that maintained similar damage levels along fruit maturation cycle. In species with two entries, a secondary damage peak occurred during second year of fruit development. Differences between maximum damage rates in first year of fruit development and second year September at the end of fruit damage monitoring were rather high especially for both moths first year entry (Yellow juniper moth: from 19.2 ± 2.8 % to 0.6 ± 0.6 %; Brown juniper moth: from 7.7 ± 1.7 % to 0.6 ± 0.6 %) scale insects (from 40.4 ± 3.5 % to 5.6 ± 1.7 %) and for mites (from 15.4 ± 2.3 % to 7.5 ± 2.4 %) and much lower for second entry of arthropods (Yellow juniper moth: from 11.7 ± 2.6 % to 2.8 ± 1.2 %; Brown juniper moth: from 2.6 ± 1.3 % to 3.3 ± 1.3 Fixed terms Estimate SE PRandom SD # Full seeds in ripe fruits Intercept -2.775 8.5·10-4 <0.001 Tree:Branch 0.990 Flowers per branch 0.004 5.8·10-4 <0.001 Branch 3.2·10-4 Fruits per branch -4.2·10-4 8.3·10-4 0.615 Fruit size 0.068 8.5·10-4 <0.001 Table 2. Linear mixed models assessing the effect of flowers and fruit per branch and fruit size on the number of full seeds in ripe fruits in Juniperus thurifera. Significant P-values in bold. SE: Standard error, SD Standard deviation of random factor estimates Ripe fruits may not be optimal 51 %; scale insects: 3.9 ± 1.4 % to 0.8 ± 0.8 %) and chalcid wasps (from 16.0 ± 2.7 % to 12.5 ± 3.0 %). Predispersal damage rate also showed some degree of variability between both monitored years with higher level of infestation for first year moths’ and scale insect entry in 2013 cohort (Table 3). Linear mixed models for fruit preferences for the different taxa differed when the analyses were based on fruit characteristics measured shortly after oviposition compared to models based on ripe fruits. In its first entry, yellow juniper moth preferred larger fruits with more seeds (Table 4). However, the model yellow juniper moth preferences based on ripe fruits did not find any effect of either fruit size or number of seeds per fruit, even with a much larger sample size. In its second entry, yellow juniper moth showed a positive effect of fruit size, but a negative effect of number of seeds per fruit. Interestingly, a positive effect of the incidence of chalcid wasps was also found. In contrast, analysis of yellow juniper moth second entry preferences based on ripe fruits detected no effect of fruit size, seed number or the incidence of chalcid wasps (Table 4). When coexistence at fruit level was evaluated presence of moth yellow juniper moth was marginally positively associated to chalcid wasp 2013 fruit cohort 2014 fruit cohort 2015 fruit cohort Species 1st year 2nd year Ripe 1st year 2nd year Ripe 1st year Ripe Mites 15.4 ± 2.3 11.4 ± 2.5 2.5 ± 0.6 14.0 ± 2.6 7.9 ± 2.1 2.2 ± 0.4 16.0 ± 4.1 4.0 ± 0.7 Yellow Moth 1st 19.2 ± 2.8 7.8 ± 2.5 6.1 ±1.0 5.4 ±1.7 2.4 ±1.2 1.1 ± 0.3 4.8 ± 2.1 4.4 ± 0.8 Yellow moth 2nd 11.7 ± 2.6 6.6 ±1.0 4.4 ±1.5 5.9 ± 0.7 6.9 ± 1.0 Brown moth 1st 7.7 ± 1.7 3.0 ± 1.3 0.7 ±0.3 4.5 ±1.6 1.1 ±0.7 0 3.8 ± 1.7 0.7 ± 0.3 Brown moth 2nd 2.6 ± 1.3 3.5 ±1.7 3.3 ±1.3 1.7 ± 0.4 2.2 ± 0.6 Scale insect 40.4 ±3.5 23.4 ±3.4 22.2 ±1.7 25.3 ±3.2 13.3 ± 2.5 24.3 ± 1.3 23.7 ± 4.0 34.9 ± 1.8 Chalcid wasp 2.1 ± 1.0 14.9 ±2.9 8.3 ±1.1 9.5 ±2.5 16.0 ± 2.7 23.3 ± 1.3 21.5 ± 3.6 14.7 ± 1.4 Juniper fly 9.3 ± 2.1 2.6 ± 1.3 0.3 ±0.2 2.8 ±1.2 0.6 ± 0.6 0.7 ± 0.2 1.5 ± 1.1 0.7 ± 0.3 Total predation 60.6 ± 3.5 54.4 ±4.0 25.4 ±1.8 39.2 ±3.6 37.2 ± 3.6 32.4 ± 1.4 58.5 ± 4.3 56.6 ± 1.9 Table 3. Predation rates (percentage ± standard error) for the main predispersal predators founded within the fruits of Juniperus thurifera. Data represent the maximum observed during the first and second year of fruit development and on ripe fruits for the 2013, 2014 and 2015 fruit cohorts. Data for second year of fruit development for 2015 cohort was not available. Capítulo 2 52 presence (Χ2 = 7.81; P = 0.0501). The linear mixed model for brown juniper moth first entry revealed that their incidence increased in trees with higher incidence of mites and yellow juniper moths. When this analysis was performed on ripe fruits the effects of the incidence of mites or yellow juniper moths were not detected (Table 5). Interestingly, whereas the presence of mites was strongly positively associated to brown juniper moth infestation at fruit level (Χ2 = 29.13; P < 0.001), both moths showed a random pattern (Χ2 = 0.20; P = 0.978). No significant effects were detected for scale insects’ preferences during their first entry (Table S2). However, when this analysis was performed on ripe fruits a positive effect of number of seeds per fruit was found (Table S2). Finally, predation monitoring revealed the preference of chalcid wasps for larger fruits. The same analysis performed on ripe fruits found a positive signal of seed number per fruit and of yellow juniper moth fruit damage rate (Table 6). Discussion Monitoring of Spanish juniper fruit development revealed that fruit set was largely determined by initial reproductive effort. Initial investment in flowers enhanced ripe fruit set levels and number of full seeds per fruit. Several arthropods damaged fruits and depredated the seeds during the developing period, greatly diminishing ripe crop size. The sequential attack of mites, moths and scale insects exerted a strong pressure during the first five months of fruit development. Selective abortion of damaged fruits caused that an important proportion of damaged fruits were not detected in ripe fruits. Moreover, fruit preferences by arthropods differed when analyses were based on developing compared to ripe fruits. Higher resource availability enhances the investment in reproductive structures and supports the demands of fruits and seeds during their development (Obeso 2002; Knight et al. 2005; Pickup and Barret 2012, see chapter 3). Initial flower production, a proxy of reproductive effort, was related to final crop size in absolute terms (Spearman ρ = 0.654; P = < 0.001): branches with more flowers bore more fruits. But also in relative terms: the probability of a flower becoming a ripe fruit increased with number of flowers per branch (Table 2) and in terms of fruit quality, branches with more flowers produced more full seeds per fruit. Ripe fruits may not be optimal 53 Fruits damaged by Yellow juinper moth Fixed terms Estimate SE PRandom SD First entry Intercept -7.389 1.035 <0.001 Tree <0.001 1st year July fruits Fruit size 0.441 0.133 <0.001 Year 0.125 Number of seeds 0.353 0.144 0..014 Mite incidence -1.389 1.054 0.188 Brown moth incidence 5.256 3.875 0.175 First entry Intercept -3.549 0.661 <0.001 Tree 0.878 Ripe fruits Fruit size 2.9·10-5 3.3·10-5 0.379 Year 0.653 Number of seeds -0.026 0.133 0.842 Mite incidence 0.866 2.132 0.685 Brown moth incidence 0.022 3.668 0.995 Secon entry Intercept -7.429 3.166 0.019 Tree 1.031 2nd year July fruits Fruit size 0.646 0.322 0.045 Year 0.006 Number of seeds -0.741 0.323 0.022 Mite incidence -1.370 3.016 0.650 Chalcid wasp incidence 6.474 3.201 0.043 Yellow moth 1st year incidence -0.381 10.649 0.972 Brown moth 1st year incidence 12.127 6.670 0.069 Scale incidence -2.980 3.157 0.345 Second entry Intercept -3.101 0.688 <0.001 Tree 1.258 Ripe fruits Fruit size 2.4·10-5 2.2·10-5 0.277 Year 0.547 Number of seeds 0.067 0.101 0.511 Mite incidence -5.536 4.743 0.243 Chalcid wasp incidence 1.470 1.503 0.328 Yellow moth 1st year incidence -1.359 2.843 0.633 Brown moth 1st year incidence 2.098 3.005 0.485 Scale incidence -0.520 1.226 0.672 Table 4. Linear mixed models evaluating fruit preferences in the yellow juniper moth Mesophleps oxycedrella. Separate models are performed for the first and second year of oviposition as well as for fruit traits measured for fruits gathered during the oviposition period of moths and for ripe fruits. Significant P-values in bold. SE: Standard error, SD Standard deviation of random factor estimates. Capítulo 2 54 Fruit loss concentrated on the first five months of fruit development matching the main period of entry of predispersal fruit predators. The presence of higher predispersal damage rates in developing than in ripe fruits indicates a selective abortion of damaged fruits, with higher investment of resources in high-quality fruits (Stephenson 1981; Ayre and Whelan 1989; Meyer et al. 2014; Riba-Hernández et al. 2016). However, this effect was contingent on the identity of predispersal predators. Mites, moths and scale insects’ damage rates were much higher at early phases of fruit development than in ripe fruits supporting the selective abortion of fruits damaged by these arthropods. In contrast, chalcid wasps and moths second entry fruit damages did not differ from damage rates in ripe fruits. These differences may be related to the stages in which both predators group occur. Chalcid wasps oviposit when seeds are well-developed (Rouault et al. 2004), and probably the cost of fruit abortion at that stage is high compared to previous development phases (Stephenson 1981, Guitian et al. 1992). In addition, during moths’ second year entry, seeds are close to maturity and protected by a thick seed coat, so pulp consumption by moth larvae do not directly damage the seeds, although may negatively affect fruit dispersal (García et al. 1999). Thus, selective abortion did not only lead to a large underestimation of the selective pressure exerted by predispersal predators, but also to changes in their relative importance. Fruits dadmaged by Brown juniper moth Fixed terms Estimate SE PRandom SD First entry Intercept -2.636 1.200 0.028 Tree <0.001 1st year July fruits Fruit size -0.248 0.180 0.167 Year <0.001 Number of seeds -0.001 0.200 0.995 Mites incidence 2.650 0.587 <0.001 Yellow moth incidence 6.732 2.895 0.020 First entry Intercept -7.824 1.768 <0.001 Tree 1.530 Ripe fruits Fruit size 1.3·10-5 1.0·10-4 0.895 Year 0.938 Number of seeds 0.241 0.375 0.521 Mite incidence 3.832 4.527 0.397 Yellow moth incidence 6.286 5.081 0.216 Table 5. Linear mixed models testing for Juniperus thurifera fruit selection by brown juniper moth Pammeme juniperana during oviposition phase based on fruits in first year July and ripe fruits. Significant P-values in bold. SE: Standard error, SD Standard deviation of random factor estimates Ripe fruits may not be optimal 55 The analyses based on ripe fruits showed no preferences by the yellow juniper moth, whereas developing fruits revealed preferences in fruit selection. Yellow juniper moths preferred large fruits and with more seeds in the first year of fruit development. Bigger fruits provide to yellow moths more pulp resources. Indeed, they star breeding directly on seeds that were soft and available for breeding at this phenological stage providing. Thus, pulp and seeds provide the necessary resources, increasing larvae survival chances (Sallabanks and Courtney 1992; Desouhant et al. 2000). Preferences were somewhat different in second year entry. Large fruits were still preferred, but seed number exerted a negative effect. Yellow juniper moths are unable to feed on second-year hardened seeds, so fruits with less seeds have more pulp providing more resources to the larvae. The analyses of fruit preferences by brown juniper moth based on developing fruits revealed an association to yellow moth and mites’ presence. Tree level association with yellow juniper moth would indicate that they use signals emitted either by yellow juniper moth as a cue for favourable oviposition places, a common phenomenon in Lepidoptera (Raitanen et al. 2014). In contrast, tree and fruit level association to mites would indicate an omnivorous behaviour that has been described for other Tortricidae species (see Pierce 1995; Wang and Daane 2014). Fruits damaged by Chalcid wasps Fixed terms Estimate SE PRandom SD Entry 1st year September Intercept -9.006 2.466 <0.001 Tree 1.969 fruits Fruit size 0.652 0.273 0.017 Year 1.047 Number of seeds 0.199 0.197 0.313 Mite incidence -5.465 3.828 0.153 Yellow moth incidence 3.706 3.481 0.287 Brown moth incidence 5.779 6.063 0.341 Scale incidence -1.650 1.979 0.404 Entry Intercept -3.116 0.565 <0.001 Tree 1.314 Ripe fruits Fruit size 7.6·10-5 1.7·10-5 0.654 Year 0.434 Number of seeds 0.211 0.077 0.006 Mite incidence -6.332 4.387 0.149 Yellow moth incidence -10.140 3.667 0.006 Brown incidence 7.404 3.804 0.052 Scale incidence 0.977 1.019 0.338 Table 6. Linear mixed models testing for Juniperus thurifera fruit selection by chalcid wasp Megastigmus thuriferana during oviposition phase based on growing fruits and ripe fruits. Significant P-values in bold. SE: Standard error, SD Standard deviation of random factor estimates. Capítulo 2 56 Scale insects reached the largest fruit damage rate (40%), however their detection in ripe fruits was lower (22 %). The underestimation of the importance of scale insects cannot be attributed to low detectability in ripe fruits (E. Rodríguez pers. obs.), so it could indicate a selective abortion and therefore an impact on final crop size. The inability to detect preferences for the fruit traits we measured may be related to the feeding behaviour these insects as pulp suckers, that probably choose fruits by their water content, as occur in J. communis (García 1998), and not by size or seed related traits. Chalcid wasp was the only taxa where ripe fruit provided more information in traits selection than immature fruits. This discrepancy may be related to the low fruit abortion associated to this species, and the larger statistical power provided from nearly four times larger sample size of ripe fruits. Chalcid wasps selected fruits with more seeds while avoiding trees and fruits already occupied by other arthropods, especially from seed eaters as yellow juniper moth (Mezquida and Olano 2013; Mezquida et al. 2016). Fruit preferences by predispersal seed predators is commonly based on the analysis of ripe fruits (Sperens 1997; García et al. 2000; Mezquida and Olano 2013; Mezquida et al. 2016), however several mechanisms can bias this approach. 1) Selective abortion of damaged fruits produced a high underestimation of damage (from 60.6 ± 3.5 % to 25.4 ± 1.8 %) with species-specific abortion rates altering the relative ranking of predispersal predators. 2) Changes in fruit characteristics from the moment of arthropod entry to ripe fruit that may complicate the identification of preferences. Despite the lower sample size (600 unripe fruits vs. 2342 ripe fruits), fruits collected just after arthropod entry provided better models of arthropod preferences for most species. 3) Finally, species with that can oviposit at different stages of fruit development might differ in their preferences depending on the phenological fruit phase selected, exerting contrasted fruit preferences. Evaluation of preferences by different arthropods attacking fruits in the Spanish juniper was improved through multiple fruit sampling matching the timing of arthropods’ entry. This result can probably be extrapolated to other species which have seeds whit long maturation periods. Overall our results claim for the need of developing a deeper knowledge on natural history to improve our understanding of the biological interactions behind ecological and evolutionary processes. Ripe fruits may not be optimal 57 References Ayre D.J., Whelan R.J. (1989) Factors controlling fruit-set in hermaphroditic plants: Studies with the Australian Proteaceae. Trends in Ecology & Evolution 4: 267–272. Badii, K. B., Billah, M. K., & Nyarko, G. (2015). Review of the pest status, economic impact and management of fruit-infesting flies (Diptera: Tephritidae) in Africa. African Journal of Agricultural Research 10: 1488–1498. Bonal, R., Muñoz, A., María Espelta, J. (2010) Mismatch between the timing of oviposition and the seasonal optimum. The stochastic phenology of Mediterranean acorn weevils. Ecological Entomology 35: 270-278. Bonal R., Espelta J.M., Vogler A.P. (2011). Complex selection on life-history traits and the maintenance of variation in exaggerated rostrum length in acorn weevils. Oecologia 167: 1053–1061. Crawley M.J. (2000) Seed predation and plant population dynamics. In: Fenner M. (ed.) Seeds: the ecology of regeneration in plant communities. 2nd edn. Wallinford (U.K.): GAB International. pp. 167-182. Desouhant E., Debouzie D., Ploye H., Menu F. (2000) Clutch size manipulations in the chestnut weevil, Curculio elephas: fitness of oviposition strategies. Oecologia 122: 493–499. Ehrlich P., Raven P. (1964) Butterflies and plants: a study in coevolution. Evolution 18: 586–608. Ehrlén J. (2015). Selection on flowering time in a life-cycle context. Oikos 124: 92–101. El Alaoui M.A., Yart A., Roques A., Arjouni Y., El Mercht S., Rozenberg M.A., Romane A. (2013) Acariens et insectes ravageurs de deux cupressacées menacées au Maroc: le Genévrier thurifère et le cyprès de l’Atlas. Ecologia Mediterranea 39: 123-128. Espelta J.M., Bonal R., Sánchez-Humanes B. (2009) Pre-dispersal acorn predation in mixed oak forests: Interspecific differences are driven by the interplay among seed phenology, seed size and predator size. Journal of Ecology 97: 1416–1423. García D. (1998). Interaction between juniper Juniperus communis L. and its fruit pest insects: Pest abundance, fruit characteristics and seed viability. Acta Oecologica 19: 517–525. García D., Zamora R., Gomez J.M., Hodar, J.A. (1999). Bird rejection of unhealthy fruits reinforces the mutualism between juniper and its avian dispersers. Oikos 85: 536–544. García D., Zamora R., Gómez J.M., Jordano P., Hodar J.A. (2000) Geographical variation in seed production, predation and abortion in Juniperus communis throughout its range in Europe. Journal of Ecology 88: 435–446. Gagic V., Riggi L.G.A., Ekbom B., Malsher G., Rusch A., Bommarco R. (2016). Interactive effects of pests increase seed yield. Ecology and Evolution 6: 2149-2157 Capítulo 2 58 Gillot C. (2005) Entomology. Dordrecht (The Netherlands): Springer. Gruwez R., Leroux O., De Frenne P., Tack W., Viane R., Verheyen K. (2013) Critical phases in the seed development of common juniper (Juniperus communis). Plant Biology 15: 210–219. Guitián J., Sánchez G., Guitián P. (1992) Niveles de fructificación en Crataegus monogyna Jacq., Prunus mahaleb L. y Prunus spinosa L. (Rosaceae). Anales Jardín Botánico 50: 239–245. Janzen D.H. (1969). Seed-eaters versus seed size, number, toxicity and dispersal. Evolution 23: 1–27. Janzen D.H. (1971) Seed predation by animals. Annual Review of Ecology and Systematics 2:465–492 Knight T., Steets J., Vamosi J., Mazer S., Burd M., Campbell D.R., … Ashman T.L. (2005) Pollen limitation of plant reproduction: pattern and process. Annual Review of Ecology Evolution and Systematics 36: 467–497. Kolb A., Ehrlén J., Eriksson O. (2007). Ecological and evolutionary consequences of spatial and temporal variation in pre-dispersal seed predation. Perspectives in Plant Ecology, Evolution and Systematics 9: 79-100. Meyer K., Soldaat L., Auge H., Thulke H. (2014) Adaptive and Selective Seed Abortion Reveals Complex Conditional Decision Making in Plants. American Naturalist 183: 376–83. Mezquida E.T., Olano J.M. (2013) What makes a good neighborhood? Interaction of spatial scale and fruit density in the predator satiation dynamics of a masting juniper tree. Oecologia 173: 483-492. Mezquida E.T., Rodríguez-García E., Olano J.M. (2016) Efficiency of pollination and satiation of predators determine reproductive output in Iberian Juniperus thurifera woodlands. Plant Biology 18: 147-155. Montesinos D., García-Fayos P., Verdú M. (2012) Masting uncoupling: mast seeding does not follow all mast flowering episodes in a dioecious juniper tree. Oikos 121: 1725–1736. Moreira X., Abdala-Roberts L., Zas R., Merlo E., Lombardero M. J., Sampedro L., Mooney K.A. (2016). Masting behaviour in a Mediterranean pine tree alters seed predator selection on reproductive output. Plant Biology 18: 973-980. Obeso J.R. (2002) The costs of reproduction in plants. New Phytologist 155: 321–348. Ollerton J., Winfree R., Tarrant S. (2011) How many flowering plants are pollinated by animals? Oikos 120: 321–326. Östergård H., Hambäck P.A., Ehrlén J. (2007). Pre-dispersal seed predation: The role of fruit abortion and selective oviposition. Ecology 88 : 2959–2965. Pickup M., Barrett S.C.H. (2012) Reversal of height dimorphism promotes pollen and seed dispersal in a wind-pollinated dioecious plant. Biology Letters 8: 245–248. Pierce, N. E. (1995). Predatory and parasitic Lepidoptera: carnivores living on plants. Journal of the Lepidopterists’ Society 49: 412–453. Ripe fruits may not be optimal 59 Raitanen J., Forsman J.T., Kivelä S.M., Mäenpää M.I., Välimäki P. (2014). Attraction to conspecific eggs may guide oviposition site selection in a solitary insect. Behavioral Ecology 25: 110–116. Rechcigl J.E., Rechcigl N.A. (2000) Insect pest management: Techniques for environmental protection. Boca Ratón (FL. U.S.A.) : CRC Press LLC. Riba-Hernández P., Segura J.L., Muñoz-Valverde J. (2016) Female fruit production depends on female flower production and crown size rather than male density in a continuous population of a tropical dioecious tree (Virola surinamensis). American Journal of Botany 103: 1990-1999. Roques A., Raimbault J.P., Goussard F. (1984) La colonisation des cônes et des galbules des Genévriers Méditerranéens par les insectes et les Acariens et son influence sur la possibilité de régénération Naturelle de ces essences. Ecologia Mediterranea X 147–169. Roques A., Cambecèdes J., Fallour-Rubio D., Raimbault J.P., Lorme P., Auger-Rozenberger, M.A. (2013) Elimination of cone and seed pests by wildfire opened a five-year regeneration window in a non-regenerating incense-juniper (Juniperus thurifera L.) stand. Ecologia Mediterranea 39: 89-98 Roault G., Turgeon J., Candau J.N., Roques A., von Aderkas P. (2004) Oviposition strategies of conifer seed chalcids in relation to host phenology. Naturwissenschaften 91: 472–480. Sallabanks R., Courtney S.P. (1992) Frugivory, seed predation, and insect–vertebrate interactions. Annual Review of Entomology 37: 377–400. Schoonhoven L.M., van Loon J.J.A., Dicke M. (2005). Insect-Plant Biology. Oxford (U.K.): Oxford University Press. Sperens U. (1997). Fruit production in Sorbus aucuparia L. (Rosaceae) and pre-dispersal seed predation by the apple fruit moth (Argyresthia conjugella Zell.). Oecologia 110: 368–373. Stephenson A G. (1981) Flower and fruit abortion: proximate causes and ultimate functions. Annual Review of Ecology and Systematics 12: 253-79. Suchan T., Alvarez N. (2015). Fifty years after Ehrlich and Raven, is there support for plant-insect coevolution as a major driver of species diversification? Entomologia Experimentalis et Applicata 157: 98–112. Wang X.G., Daane, K.M. (2014) Cannibalism of parasitoid-attacked conspecifics in a non-carnivorous caterpillar. Entomologia Experimentalis et Applicata 151: 112-121. Xia K., Harrower W.L., Turkington R., Tan H.Y., Zhou Z.K. (2016). Pre-dispersal strategies by Quercus schottkyana to mitigate the effects of weevil infestation of acorns. Scientific Reports 6: 37520. https:// doi.org/10.1038/srep37520 Capitulo 3 66 Abstract Reproductive output is determined by a complex set of factors including resources available to invest in reproduction, successful pollination and fertilization, and fruit and seed loss from predators during fruit development and ripening. Disentangling the relative contribution of each limiting factor is critical to underpin the factors determining plant fitness. We performed an experimental approach to assess the relative contribution of initial reproductive effort, pollination success and predispersal seed predation on plant reproductive success in the wind-pollinated dioecious conifer Juniperus thurifera. We set a full factorial design with two treatments: pollen addition and pesticide application on 40 female juniper trees, and monitored the full process of fruit development during 20 months. The influence of initial reproductive effort, hand-pollination and pesticide application on fruit set at different stages, seed viability and fruit damage by several specialized arthropods was tested by mixed models. We also set up an additional experiment to test the effect of hand-pollination on pollen tube formation. Most reproductive structures were lost at the early stages of reproduction, particularly during the initial growing and development of the fruits. Resources invested in reproduction had important consequences for fruit and seed set. Production of more flowers reduced the incidence of low-mobile seed predators through a satiation effect, increased the proportion of fruits set and the quality of ripe fruits (having more viable seeds). Pollination was not limiting in the population and year of study and manual addition of pollen diminished levels of fruit set, although fruits that ripen had more viable seeds. Predispersal seed predation by different arthropods was an important limiting factor to fruit set in this juniper species. The initial investment in surplus flowers allowed plants to selectively abort damaged and depredated fruits during the early phase of development and allocate resources to the remaining fruits. Overall, our experiment indicated that levels of fruit set in Spanish junipers were largely determined by initial reproductive effort, predispersal seed predation exerted by arthropods and the interaction between both factors. Deciphering the filters on reproductive output 67 Introduction The number of viable seeds produced by plants is an important determinant of plant recruitment and individual fitness (Herrera 1991; Louda and Potvin 1995). Multiple factors influence seed production and limit reproduction, thus potentially playing important roles in plant population dynamics and evolution (Hainsworth 1984;, Price et al. 2008; Boieiro et al. 2012). Plants invest resources in reproduction by producing flowers that must be pollinated to form fruits, ovules must be fertilized and sustained with resources to develop into viable seeds, and seeds and fruits must escape predation during maturation to produce sound seeds ready for dispersal (Kaye 1999; Pías et al. 2007). From the initial reproductive investment, losses occur sequentially due to factors such as pollination failure, lack of fertilization and seed predation, although their relative influence on reproduction is difficult to disentangle because several components may act in combination (Hainsworth 1984; Herrera 1991; Gruwez et al. 2013; Walsh et al. 2014). Resources allocated to reproduction compete with demands from other plant functions as survival or growth (Obeso; 2002, Hirayama et al. 2008; Teitel et al. 2016), leading to trade-offs between different requirements, with resource allocation priorities determined by life strategy (Pickup and Barret 2012; Martín et al. 2015). Under high availability of resources, plants usually invest more energy in reproductive structures by, for example, producing more ovules (Obeso 2002) or higher pollen loads (Pearse et al. 2015), overall enhancing their reproductive success (Barringer et al. 2013). Nevertheless, the cost of reproduction may extend well beyond the flowering period due to the high energetic demand of fruit development (Obeso 2002; Hirayama et al. 2008; Martín et al. 2015; Teitel et al. 2016). Thus, the investment in energy required for large fruit crops may exhaust current and stored resources compromising resource availability for the next reproductive season (Karlsson et al 1990; Lyles et al. 2015; Pessendorfer et al. 2016). Under these premises, inter-annual changes in fruit production (masting) may respond to inter-annual variation in resource levels as well as inter-cohort resource competition (Knight et al. 2005; Crone et al. 2009). A high initial investment in reproduction may be curtailed if pollination fails. Pollination success depends on the concert of multiple factors and high inter-annual variation is the norm (Koenig and Knops 2014; Koenig et al. 2015). Success at this reproductive stage is Capitulo 3 68 affected by individual plant traits, such as total flower production (Kudo and Harder 2005), flowering synchrony (Albert et al. 2001; Obeso 2002; Koenig et al. 2015; Lyles et al. 2015) or pollen quality and quantity (Knight 2003; Labouche et al. 2016). Floral traits and environmental factors are important determinants of pollination efficiency. For example, abundant precipitations during the pollination period diminishes pollination success in wind-pollinated plants, whereas warm conditions favor the mobility of invertebrate pollinators potentially enhancing pollination in animal-pollinated plants (Knops et al. 2007; Rech et al. 2016). Pollination success is also modulated by factors at higher spatial scales like plant density and structure (Sork et al. 2002; Knight 2003; Sanz and Pulido 2015; see chapter 5), habitat fragmentation (Sork et al. 2002; Knight et al. 2005), interspecific competition for pollinators (Mitchell et al. 2009) or interference from heterospecific pollen (Mugnaini et al. 2007; Aderkas et al. 2012). Thus, fruit set and number of fruits produced may be low even in the presence of high resource levels if conditions for pollination and early fruit development are not satisfactory. Plants accumulate nutritious resources in the embryo and its surrounding tissues during fruit and seed development. The high and concentrated nutritional value of fruits and seeds are targeted by specialized predispersal predators, mainly arthropods. Moreover, arthropods developing inside fruits and seeds are provided with a protective place against desiccation and natural enemies during its development (Sallabanks and Courtney 1992). Thus, predispersal predation may compromise successful plant reproduction and become a strong evolutionary force. In response to predispersal predation, plants develop different strategies to defend their fruit and seeds from attacks. Short-term strategies involve physical and chemical defenses (Janzen 1971; Schoonhoven et al. 2005; Boivin and Auger-Rozenberg 2016), fruit and seed abortion after infestation (Bonal et al. 2010; Meyer et al. 2014; Boivin and Auger-Rozenberg 2016) or delay in fertilization to avoid embryo loss by predator oviposition (Rouault et al. 2004; Aderkas et al. 2005). Long-term responses include higher inter-annual variations in fruit production to constrain the ability of specialist predators to match its demographic cycles with fruit production (Janzen 1971), boosting plant reproductive success during high production years (Turgeon 1994). In the present work, we designed an experimental approach to determine the relative contribution of initial reproductive effort, pollination success and predispersal seed predation on plant reproductive success in the wind-pollinated dioecious conifer Juniperus thurifera. Deciphering the filters on reproductive output 69 We estimated fruit set levels at different stages of fruit development (fruits will be used for the fleshy cones produced by this juniper), and quantified seed predation and seed characteristics in ripe fruits. We address the hypothesis that resources, as represented by the initial individual investment in reproduction, improve reproductive success by counteracting losses at different stages and enabling the selective abscission of damaged or low quality fruits. We also test the hypothesis that, if pollination is limiting, the addition of supplementary pollen would increase fruit or seed set levels. Finally, we assess the relative role of predispersal seed predators in depressing the reproductive potential in this juniper by experimentally reducing their impact and estimating the resulting fruit and seed set levels. Material and methods Natural history Spanish juniper (J. thurifera) is a dioecious juniper species, endemic to the Western Mediterranean where it inhabits continental summer-dry environments. Junipers, as is common in other gymnosperms, exhibit a long interval between pollination and seed ripening. Wind pollination occurs in late winter-early spring. Pollen germinates and megagametophyte formation starts after pollination, but fertilization is delayed until late spring. Embryo development begins after fertilization and seeds are almost fully grown by mid-summer, when seed coat starts to harden. Maturation of the megagametophyte and embryo takes 12-13 additional months (Gruwez et al. 2013). Therefore, the fruits (fleshy cones) produced by this juniper species ripen about 20 months after pollination in next year autumn (Montesinos et al. 2012). A variety of arthropods can damage Spanish juniper fruits during this long development period (Roques et al. 1984; El Alaoui et al. 2013). In the Iberian Peninsula, three arthropod taxa are the main predispersal predators: mites, moths and chalcid wasps. Mites, Trisetacus quadrisetus (Thomas 1889) (Acari, Phytoptidae), colonize the seeds at the very beginning of their development and use them as growth chambers where they feed and reproduce (Roques 1984; see chapter 2). Two moth species: Mesophleps oxycedrella (Millière 1871) (Lepidoptera, Gelechiidae) and Pammene juniperana (Millère 1858) (Lepidoptera, Tortricidae) Capitulo 3 70 feed on juniper fruits. Females oviposit usually one egg on the fruit surface from late-May to early July during the first and second year of fruit development and, by late July, larvae have entered the fruit and seeds. Chalcid wasps Megastigmus thuriferana Roques & El Alaoui 2006 (Hymenoptera, Torymidae) feed on well-developed megagametophyte and embryo (Rouault et al. 2004). Females oviposit one egg inside the seeds of developing fruits from late July-early August to late September. Study area The study took place in Villaciervos, Soria province, Central Spain (41º 44’ N, 2º 40’ W; 1150 m.a.s.l.). Climate is Mediterranean continental with mean annual precipitation of 595 mm, only 110 mm occurring during summer. Mean monthly temperatures ranged from 1.7 ºC in January to 19.7 ºC in July. The rock parent material is Cretaceous limestone covered by deep soil with small areas of exposed bedrock. Vegetation is dominated by the Spanish juniper with open areas among trees covered by xeric grasslands with shrubs, including Cistus laurifolius L., J. communis L. and Thymus zygis Loefl. Sampling design and sample analyses We randomly chose 40 female juniper trees within the study area. For each tree, eight flowering branches from all around the canopy were selected and tagged. We set a full factorial design with two treatments: pollen addition and pesticide application. Two branches per tree were subjected to each of the treatment combinations: pollen, pesticide, pollen plus pesticide and control. Pollen addition consisted on manual addition of pollen collected from 10 different male trees and applied onto female cones with a soft paintbrush shortly after pollen collection. Pollen addition was performed twice during March 2014, coinciding with the pollination period. The pesticide treatment consisted in applying a mixture of acaricide and insecticide that was prepared diluting the acaricide (Fenpyroximate 5 % w/v) and the insecticide (Chlorpyrifos 48 % w/v) in water at 0.2 % v/v. Pesticide was applied twice per month between May and September in 2014 and 2015. Junipers are sensitive to the deposition of particles, other than juniper pollen, during pollination (Mugnaini et al. 2007), so timing of pesticide spraying was chosen to avoid interfering with pollen reception on female flowers during pollination while encompassing the oviposition period of arthro- Deciphering the filters on reproductive output 71 pods. Branches subjected to the pollination plus pesticide treatment were supplemented with pollen and later sprayed with pesticide as explained above. Control branches were manipulated as the experimental branches, but no pollen or pesticide was applied. We counted the number of female flowers produced in each tagged branch in March 2014 as a measure of the initial reproductive effort at branch level. The number of developing fruits in each branch were counted monthly from June to September during the first year of development (2014), when most of the fruits are lost (see chapter 2), and in May, July and October during the second year of development (2015). All fruits remaining in the tagged branches were collected when ripen in late October 2015. We selected five sound fruits (i.e., without signs of damage by arthropods) from each branch (when available) to count and characterize seeds. Fruits were opened to separate and count the number of seeds, and each seed was dissected under a microscope and assigned to one of the following categories: ‘full’, for seeds showing a completely developed embryo and megagametophyte; ‘empty’, for seeds that interrupted the development of the megagametophyte and/or embryo after fertilization, and that are externally similar to fully developed seeds although they do not contain (or only some remnants) megagametophyte or embryo; and ‘aborted’, for seeds that interrupted their development between pollination and fertilization, and that are visually recognizable as small, not completely developed seeds. The rest of the fruits were opened and examined in the laboratory under a dissecting microscope to detect signs of arthropod attack. Damage by arthropods can be easily assigned to each of the three arthropod taxa (Roques et al. 1984; Mezquida and Olano 2013). Mites deform the seeds causing the elongation of their tips that usually stick out of the fruit surface, moths make an irregular hole in different parts of the fruit surface and chalcid wasps make a circular hole in the apical region of the fruit. To further explore the effect of pollen addition on seed development, we evaluated whether the manual addition of pollen increased the likelihood of pollen germination and pollen tube growing. In March 2015, we conducted another pollen addition experiment to a subset of 10 experimental trees. Two branches from each tree were selected and tagged, one was subjected to the pollen addition treatment and the other used as control. In June 2015, we collected five sound developing fruits from each branch (two fruits were damaged or had no viable seeds, so final sample size was 98 developing fruits, 50 and 48 fruits for natural and manual pollination, respectively). The presence of pollen tube was evaluated Capitulo 3 72 under light microscopy. Sampled fruits were stored in 96% v/v ethanol until processing and then embedded in Technovit 7100 resin (Heraeus Kulzer, Wehrheim, Germany) following Leroux et al. (2007). First, we excised outer tissues of the juniper fruits to ensure good resin infiltration and stored them in 99 % v/v ethanol for 2 days to full dehydration. After dehydration was completed, samples were infiltrated with Technovit 7100 infiltration liquid (2-hydroxyethylmethacrylate) diluted to 30%, 50%, and 70% with 99% v/v ethanol. Once the samples were infiltrated with these solutions (1 day each), they were transferred to 100% Technovit 7100 infiltration liquid for 2 days and embedded in polypropylene flat bottom capsules (TAAB laboratories, Berkshire, UK). Transverse sections of 4 µm were cut with a Microm HM360 microtome (Microm International GmbH, Walldorf, Germany), collected on glass slides, stained with a 0.05% w/v solution of toluidine blue O (Merck, Darmstadt, Germany, C.I. No. 52040) in 0.1% w/v Na2B4O7, and mounted in DePeX (Gurr, BDH Laboratory, UK). Sections were observed with a Nikon Eclipse E600 microscope and images were recorded using a Nikon digital camera DXM1200. Statistical analysis We defined four critical stages during fruit development: ‘flowers’, as the number of flowers produced and counted in March; ‘initial fruits’, as the number of initiated fruits that could be identified in May; ‘unripe fruits’, as the number of unripe green fruits that have reached their final size in September of the first year of development, and ‘ripe fruits’, as the number of mature fruits that turn to dark blue in October of the second year. We used linear mixed models with a binomial error structure that accounts for differences in sample size to analyze the influence of experimental treatments on transition success between flowers and ripe fruits. We consider as success the number of fruits that persisted between two stages for each treatment. Firstly, we analyzed the full transition from flowers to ripe fruits using the number of ripe fruits relative to the number of flowers produced. Fixed terms in this model included pollen, pesticide and the interaction between both treatments, and number of flowers per branch was included as a continuous predictor quantifying the initial reproductive effort. Besides, we analyzed the three transitions between the intermediate stages: flower to initial fruits, initial to unripe fruits and unripe to ripe fruits. Model structure was similar to the above model, although number of fruits per branch (initial or unripe fruits; instead of number of flowers) was included as a fixed continuous variable in the models for Deciphering the filters on reproductive output 73 unripe fruit set and ripe fruit set. Tree was included as a random factor in the four models. Linear mixed models with Poisson error structure were used to test whether experimental treatments affected the number of total, full, empty and aborted seeds in ripe fruits. Fixed terms in the models included pollen and pesticide treatments, their interaction and number of flowers per branch. Tree was included as a random factor. Finally, to test whether pollen addition influenced the proportion of fruits showing a grown pollen tube, we fitted a linear mixed model with binomial error structure. Pollen treatment, number of flowers per branch and number of viable seeds per fruit were included in the fixed part of the model, and tree was included as a random factor. To test for the influence of experimental treatments on levels of fruit damage by each arthropod, linear mixed models with a binomial error structure were used. Because fruit abundance may affect damage rates (Mezquida and Olano 2013), we included the number of developing fruits in the branch when each arthropod oviposits or enters the fruit or seed as a covariate. Therefore, we included number of flowers in the mixed model for fruit damage by mites, number of initial fruits (counted in May 2014) in the model for fruit damage by moths and number of developing fruits (July 2014) in the model for fruit damage by chalcid wasps. Because moths and chalcid wasps tend to avoid ovipositing in fruits already infested by other arthropods (Mezquida and Olano 2013), we included the proportion of fruits depredated by other arthropods as explanatory variables in the models for moths and chalcid wasps. Tree was included as a random factor in the models. Count variables used as covariates, such as number of flowers and fruits at different development stages, were standardized to zero mean and unit variance to rescale them and reduce their variance. Mixed models for seed characteristics, pollen tube development and fruit damage by arthropods were simplified by progressively removing the least significant terms. Model calculations were performed using the lme4 package in R environment (R Development Core Team, 2014). Results We monitored 40,532 flowers in 320 branches from 40 female juniper trees in 2014. The number of fruits that started to develop in May were 23,473 (57.9 % of the initial flowers). Major fruit losses occurred during the next two months of development (Fig. 1), with only Capitulo 3 74 10,977 fruits remaining in July (27.1 % of the initial flowers; 46.8 % of the initial fruits). Losses from July to September 2014 were lower with 9,218 fruits remaining (22.7 % of the initial flowers; 39.3 % of initial fruits). Fruit losses continued during the next year until fruit ripening with 5,614 ripe fruits counted in October 2015 (13.9% of the initial flowers; 23.9% of the initial fruits). Trees and branches producing more flowers set a higher proportion of ripe fruits (Table 1). Hand-pollination of flowers decreased, whereas spraying of pesticide increased, the proportion of fruits that ripen, and there was a positive interaction between pollen and pesticide treatment (Table 1; Fig. 2). The analysis of transitions between different developmental stages showed similar results, although branches with more initial fruits set a higher proportion of unripe fruits whereas branches with more unripe fruits set a lower proportion of ripe fruits (Table 1). In addition, the interaction between pollen addition and pesticide treatment was not significant for the transition between initial to unripe fruits, and fruit losses between unripe to ripe fruits did not differ between branches treated with pesticide and control ones (Table 1; Fig. 2). Overall, the proportion of ripe fruits relative to the number of flowers produced were higher in branches treated with pesticide followed by those hand-pollinated plus sprayed with pesticide, control branches and hand-pollinated ones (Fig. 1; Fig. 2). Ripe fruits developed in branches that produced more flowers had more seeds, including more filled and less aborted seeds (Table 2; Fig. 3). Manual addition of pollen and pesticide treatment did not influence the number of seeds per fruit, but both treatments increased the number of full seeds and reduced the number of aborted seeds in ripe fruits (Table 2; Fig. 3). The number of empty seeds was not affected by either the initial number of flowers produced or any treatment (Table 2; Fig. 3). The mixed model to assess the effect of pollen addition on pollen tube development showed that the proportion of seeds with grown pollen tube was higher for fruits in hand-pollinated branches (Mean ± SE: 0.78 ± 0.08) than fruits in control branches (0.61 ± 0.06, Z = 3.1, P = 0.002, n = 98 fruits, 229 seeds). The presence of pollen tube was not influenced by the number of flowers produced by each branch or the number of seeds per fruit (P> 0.15, in both cases). Damage by arthropods affected to 25.4 % of the overall ripe fruits. Chalcid wasps were by far the most abundant seed predator affecting to 17.3 % of the fruits, whereas mites Deciphering the filters on reproductive output 75 (3.9 %) and moths (4.9 %) had low predation rates. The mixed model for the incidence of mites in ripe fruits showed that branches producing more flowers and those treated with pesticide yielded fruits with lower levels of damage by this arthropod (Table 3; Fig. 4). Fruit and seed predation by moths increased with the number of fruits available during the period of oviposition and was reduced in branches treated with pesticide. The incidence of chalcid wasps did also show a negative effect on predation rates by moths (Table 3; Fig. 4). The model for seed predation by chalcid wasps showed that branches with more fruits available during oviposition and those treated with pesticide had higher levels of predation by chalcid wasps. In addition, the incidence of the other two arthropods (mites and moths) negatively influenced damage rates by chalcid wasps (Table 3; Fig. 4). Fig. 1. Percentage of fruits in relation to initial flowers for the 2014-2015 fruit cohort. Lines represent the combination of pollen and pesticide treatments (see inlet legend). Vertical arrows indicate sampling dates. xx: timing of hand-pollination; **: timing of pesticide spraying. The upper panel shows the timing when three arthropod taxa enter (red line), stay (black line) and leave (blue line) the damaged fruits. Capitulo 3 82 investment in surplus flowers allowed plants to selectively abort damaged and depredated fruits during the early phase of development and allocate resources to the remaining fruits. Resources invested in reproduction had important consequences for fruit and seed set (Montesinos et al. 2012, Mezquida et al. 2016). Production of more flowers reduced the incidence of the low-mobile mites through a satiation effect, increased levels of fruit set and the quality of ripe fruits (having more viable seeds). Seed predators with greater dispersal capacity, however, were attracted to trees producing more fruits, which would favor individual as well as population variations in seed production between reproductive episodes to satiate seed predators, as found in this mast-seeding species (Herrera et al. 1998; Montesinos et al. 2012; Mezquida and Olano 2013). Deciphering the filters on reproductive output 83 References Acar I., Eti S. (2008) Effect of pistil receptivity, pollen mixtures, and pollen application distances on fruit set of pistachios (Pistacia vera). New Zealand Journal of Crop and Horticultural Science 36: 295–300. Aderkas P., Rouault G., Wagner R., Rohr R., Roques A. (2005) Seed parasitism redirects ovule development in Douglas fir. Proceedings Biological Sciences The Royal Society 272: 1491–1496. Aderkas P., Nepi M., Rise M., Buffi F., Guarnieri M., Coulter A., Gill K., Lan P., Rzemieniak S., Pacini E. (2012) Post-pollination prefertilization drops affect germination rates of heterospecific pollen in larch and Douglas-fir. Sexual Plant Reproduction 25: 215-225. Albert M.J., Escudero A., Iriondo J.M. (2001) Female reproductive success of narrow endemic Erodium paularense in contrasting microhabitats. Ecology 82: 1734-1747. Ayre D.J., Whelan R.J. (1989) Factors controlling fruit-set in hermaphroditic plants: Studies with the Australian Proteaceae. Trends In Ecology & Evolution 4: 267–272. Barringer B.C., Koenig W.D., Knops J.M.H. (2013) Interrelationships among life-history traits in three California oaks. Oecologia 171: 129–139. Boieiro M., Rego C., Serrano A.R.M., Espadaler X. (2012) Seed production and pre-dispersal reproductive losses in the narrow endemic Euphorbia pedroi (Euphorbiaceae). Plant Ecology 213: 581–590. Boivin T., Auger-Rozenberg M. (2016) Native fruit, cone and seed insects in the Mediterranean Basin. In Paine T.D., Lieutier F. Insects and diseases of mediterranean forest systems. Switzerland: Springer International Publishing. Pp 47-88. Bonal R., Muñoz A., Espelta J.M. (2010) Mismatch between the timing of oviposition and the seasonal optimum. The stochastic phenology of Mediterranean acorn weevils. Ecological Entomology 35: 270-278. Crone E.E. (2009) How do plants know when other plants are flowering? Resource depletion, pollen limitation and mast-seeding in a perennial wildflower. Ecology Letters 12: 1119–1126. Cruzan M.B. (1986) Pollen tube distribution in Nicotiana glauca: evidence for density dependent growth. American Journal of Botany 73: 902-907. El Alaoui M.A., Yart A., Roques A., Arjouni Y., El Mercht S., Rozenberg M.A., Romane A. (2013) Acariens et insectes ravageurs de deux cupressacées menacées au Maroc: le Genévrier thurifère et le cyprès de l’Atlas. Ecologia Mediterranea 39: 123-128. Ehrlén J., Raabova J., Dahlgren J.P. (2015) Flowering schedule in a perennial plant; life-history trade-offs, seed predation, and total offspring fitness. Ecology 96: 2280–2288. Espelta J.M., Bonal R., Sánchez-Humanes B. (2009) Pre-dispersal acorn predation in mixed oak forests: Interspecific differences are driven by the interplay among seed phenology, seed size and predator size. Journal of Ecology 97: 1416–1423. Capitulo 3 84 Gruwez R., Leroux O., De Frenne P., Tack W., Viane R., Verheyen K. (2013) Critical phases in the seed development of common juniper (Juniperus communis). Plant Biology 15: 210–219. Hainsworth F.R., Wolf L.L., Mercier T. (1984) Pollination and Pre-Dispersal Seed Predation: Net Effects on Reproduction and Inflorescence Characteristics in Ipomopsis aggregate. Oecologia 63: 405-409 Herrera C.M. (1991) Dissecting factors responsible for individual variation in plant fecundity. Ecology 72: 1436–1448. Herrera C.M., Jordano P., Guitian J., Traveset A. (1998) Annual variability in seed production by woody plants and the masting concept: reassessment of principles and relationship to pollination and seed dispersal. The American Naturalist 152: 576–594 Hirayama D., Nanami S., Itoh A., Yamakura T. (2008) Individual resource allocation to vegetative growth and reproduction in subgenus Cyclobalanopsis (Quercus, Fagaceae) trees. Ecological Research 23: 451–458. Janzen D.H. (1971) Seed predation by animals. Annual Review in Ecology and Systematics 2: 465–492 Karlsson, P. S., Svensson, B. M., Carlsson, B. Å., Nordell, K. O. (1990) Resource investment in reproduction and its consequences in three Pinguicula species. Oikos 59: 393–398. Kaye T.N. (1999) From flowering to dispersal: Reproductive ecology of an endemic plant, Astragalus australis var. olympicus (Fabaceae). American Journal of Botany 86: 1248–1256. Knight T.M. (2003) Floral density, pollen limitation, and reproductive success in Trillium grandiflorum. Oecologia 137: 557–563. Knight T., Steets J., Vamosi J., Mazer S., Burd M., Campbell D. R., … Ashman T.L. (2005) Pollen limitation of plant reproduction: pattern and process. Annual Review of Ecology Evolution and Systematics 36: 467–497. Knops J.M.H., Koenig W.D., Carmen W.J. (2007) Negative correlation does not imply a tradeoff between growth and reproduction in California oaks. Proceedings of the Nationall Academy of Sciences 104:16982–16985 Koenig W.D., Knops J.M.H. (2014) Environmental correlates of acorn production by four species of Minnesota oaks. Population Ecology 56: 63–71. Koenig W.D., Knops J.M.H., Carmen W.J., Pearse I.S. (2015) What drives masting? The phenological synchrony hypothesis. Ecology 96: 184–192. Kudo G., Harder L.D. (2005) Floral and inflorescence effects on variation in pollen removal and seed production among six legume species. Functional Ecology 19: 245–254. Labouche A.M., Richards S.A., Pannell J.R. (2016) Effects of pollination intensity on offspring number and quality in a wind-pollinated herb. Journal of Ecology 105: 197–208. https://doi.org/10.1111/13652745.12659 Deciphering the filters on reproductive output 85 Leroux O., Van der Kinderen G., Viane R.L.L. (2007) A sandwich-embedding method for oriented sectioning. Journal of Microscopy 227: 79–82. Linhart Y.B., Moreira X., Snyder M.A., Mooney K.A. (2014) Variability in seed cone production and functional response of seed predators to seed cone availability: support for the predator satiation hypothesis. Journal of Ecology 102: 576–583. Louda S.M., Potvin M.A. (1995) Effect of Inflorescence-Feeding Insects on the Demography and Lifetime of a Native Plant. Ecology 76: 229–245. Lyles D., Rosenstock T.S., Hastings A. (2015) Plant reproduction and environmental noise: How do plants do it? Journal of Theoretical Biology 371: 137-144 Martín D., Vázquez-Piqué J., Carevic F. S., Fernández M., Alejano R. (2015) Trade-off between stem growth and acorn production in holm oak. Trees 29: 825–834. Meyer K., Soldaat L., Auge H., Thulke H. (2014) Adaptive and Selective Seed Abortion Reveals Complex Conditional Decision Making in Plants. American Naturalist 183: 376–83. Mezquida E.T., Olano J.M. (2013) What makes a good neighborhood? Interaction of spatial scale and fruit density in the predator satiation dynamics of a masting juniper tree. Oecologia 173: 483-492. Mezquida E.T., Rodríguez-García E., Olano J.M. (2016) Efficiency of pollination and satiation of predators determine reproductive output in Iberian Juniperus thurifera woodlands. Plant Biology 18: 147-155. Mitchell R.J., Flanagan R.J., Brown B.J., Waser N.M., Karron J.D. (2009) New frontiers in competition for pollination. Annals of Botany 103: 1403–1413. Montesinos D., García-Fayos P., Verdú M. (2012) Masting uncoupling: mast seeding does not follow all mast flowering episodes in a dioecious juniper tree. Oikos 121: 1725–1736. Mugnaini S., Nepi M., Guarnieri M., Piotto B., Pacini E. (2007) Pollination drop in Juniperus communis: response to deposited material. Annals of Botany 100: 1475-1481. Obeso J.R. (2002) The costs of reproduction in plants. New Phytologist 155: 321–348. Pearse I.S., Koenig W.D., Funk K.A., Pesendorfer M.B. (2015) Pollen limitation and flower abortion in a wind-pollinated, masting tree. Ecology 96: 587–593. Pessendorfer M.B., Koenig W.D., Pearse I.S., Knops J.M.H., Funk K.A. (2016) Individual resource limitation combined with population-wide pollen availability drives masting in the valley oak (Quercus lobata). Journal of Ecology 104: 637–645. Pías B., Salvande M., Guitián P. (2007) Variation in Predispersal Losses in Reproductive Potential in Rowan (Sorbus aucuparia L. Rosaceae) in the NW Iberian Peninsula. Plant Ecology 188: 191–203. Pickup M., Barrett S.C.H. (2012) Reversal of height dimorphism promotes pollen and seed dispersal in a wind-pollinated dioecious plant. Biology Letters 8: 245–248. Capitulo 3 86 Price M.V., Campbell D.R., Waser N.M., Brody A.K. (2008) Bridging the generation gap in plants: pollination, parental fecundity, and offspring demography. Ecology 89:1596–1604. R Development Core Team (2014) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna Rech A.R., Dalsgaard B., Sandel B., Sonne J., Svenning J.C., Holmes N., Ollerton J. (2016) The macroecology of animal versus wind pollination: ecological factors are more important than historical climate stability. Plant Ecology & Diversity 9: 253–262. Riba-Hernández P., Segura J.L., Muñoz-Valverde J. (2016) Female fruit production depends on female flower production and crown size rather than male density in a continuous population of a tropical dioecious tree (Virola surinamensis). American Journal of Botany 103: 1990-1999. Roques A., Cambecèdes J., Fallour-Rubio D., Raimbault J.P., Lorme P., Auger-Rozenberger, M.A. (2013) Elimination of cone and seed pests by wildfire opened a five-year regeneration window in a non-regenerating incense-juniper (Juniperus thurifera L.) stand. Ecologia Mediterranea 39: 89-98 Rouault G., Turgeon J., Candau J.N., Roques A., von Aderkas P. (2004) Oviposition strategies of conifer seed chalcids in relation to host phenology. Naturwissenschaften 91:472–480. Sallabanks R., Courtney S.P. (1992) Frugivory, seed predation, and insect-vertebrate interactions. Annual Review in Entomology 37:377–400 Sanz R., Pulido F. (2015). Pollen limitation and fruit abortion in a declining rare tree, the Eurasian yew (Taxus baccata L.): A reproductive cost of ecological marginality. Plant Biosystems 149: 818–826. Schoonhoven L.M., van Loon J.J.A., Dicke M. (2005). Insect-Plant Biology. Oxford (U.K.): Oxford University Press. Sork V.L., Davis F.W., Smouse P.E., Apsit V.J., Dyer R.J., Fernandez J.F. Kuhn B. (2002). Pollen movement in declining populations of California Valley oak, Quercus lobata: Where have all the fathers gone? Molecular Ecology 11: 1657–1668. Stephenson A.G. (1981) Flower and fruit abortion: proximate causes and ultimate functions. Annual Reviews of Ecology and Systematics 12: 253-79. Teitel Z., Pickup M., Field D.L., Barrett S.C.H. (2016). The dynamics of resource allocation and costs of reproduction in a sexually dimorphic, wind-pollinated dioecious plant. Plant Biology 18: 98–103. Turgeon J. (1994) Insect exploiting seed cones of Larix spp., Picea spp., and Pinus spp.: species richness and patterns of exploitation. In: Turgeon J.J., de Groot P. (eds). Biology, damage and management of seed orchard pests. Ontario (Canada): Canadian Forestry Service. pp 15–30 Young H.J., Young T.P. (1992) Alternative outcomes of natural and experimental high pollen loads. Ecology 73: 639-647. Walsh R.P., Arnold P.M., Michaels H.J. (2014). Effects of pollination limitation and seed predation on female reproductive success of a deceptive orchid. AoB PLANTS 6 1–12. CHAPTER 4 Efficiency of pollination and satiation of predators determine reproductive output in Iberian Juniperus thurifera woodlands. CHAPTER 4 “I’ve watched them dance To the music that the feeling brings Then dig a hole For the music when the feeling’s gone” Maika Makovski Capitulo 4 90 Abstract Fruit production in animal-dispersed plants has a strong influence on fitness because large crops increase the number of seeds dispersed by frugivores. Large crops are costly, and environmental control of plant resources is likely play a role in shaping temporal and spatial variations in seed production, particularly in fluctuating environments such as the Mediterranean. The number of fruits that start to develop and the proportion of viable seeds produced are also linked to the number of flowers formed and the efficiency of pollination in wind-pollinated plants. Finally, large fruit displays also attract seed predators, having a negative effect on seed output. We assessed the relative impact of environmental conditions on fruit production, and their combined effect on seed production, abortion and seed loss through three predispersal predators in Juniperus thurifera L., sampling 14 populations across the Iberian Peninsula. Wetter than average conditions during flowering and early fruit development led to larger crop sizes; this effect was amplified at tree level, with the most productive trees during more favourable years yielding fruits with more viable seeds and less empty and aborted seeds. In addition, large crops satiated the less mobile seed predator. The other two predispersal predators responded to plant traits, the presence of other seed predators and environmental conditions, but did not show a satiation response to the current-year crop. Our large-scale study on a dioecious, wind-pollinated Mediterranean juniper indicates that pollination efficiency and satiation of seed predators, mediated by environmental conditions, are important determinants of reproductive output in this juniper species. Pollination and predator satiation determine reproduction 91 Introduction The number of fruits produced by individual fleshy-fruited plants frequently has important consequences for plant fitness (Jordano 1995; Martíınez et al. 2007; Ortiz-Pulido et al. 2007; Sobral et al. 2010). Individuals producing larger crops attract more frugivorous animals and, as a result, more seeds can be dispersed from the parent plant (Laska & Stiles 1994; Ortiz-Pulido et al. 2007; Blendinger et al. 2008). However, plants that produce more fruits, also normally attract more predispersal seed predators (Herrera 1986; Sallabanks & Courtney 1992), which may lower plant reproductive success by reducing the number of healthy fruits available for dispersal (Jordano 1987). Fruit production is also dependent on the number of flowers formed, which sets an upper limit to the number of fruits starting to develop. Moreover, flower production largely determines pollination success and embryo development after fertilisation in wind-pollinated species (Nilsson & Wastljung 1987; Kelly & Sork 2002). Reproductive investment during years of large fruit crops is costly in terms of energy devoted to reproduction (Obeso 2002), particularly for female plants in dioecious species that have to allocate resources to both flower and fruit production (Obeso 1997; Montesinos et al. 2012). Environmental conditions and resource levels influence reproductive investment at different spatial and temporal scales (Agren et al. 2008). At a geographic scale, climate largely determines inter-population variability in the resources available to plants for growth and reproduction (García et al. 2000; Obeso 2002; Montesinos et al. 2010; DeSoto et al. 2012). At a local scale, temporal variations in environmental (weather) conditions modulate annual reproductive investment by woody perennials (Lee & Bazzaz 1982; Herrera 1991; Crone & Lesica 2006). For species with high fluctuations in reproductive output, environmental conditions are clearly linked to large crops (Kelly & Sork 2002). However, this connection is not straightforward (Kelly & Sork 2002), and temporal variations in reproduction do not entirely match the availability of resources and may also depend on the allocation of reserves to the current seed crop (Koenig & Knops 2000; Crone et al. 2009; Sala et al. 2012; Hoch et al. 2013; Ida et al. 2013). Therefore, the number of viable seeds produced by individual plants is determined by different factors acting during the reproductive cycle, such as environmental conditions, flower and fruit production, the efficiency of pollination and the amount of seed lost to predators, which typically vary among populations and years in many woody perennials. Capitulo 4 98 Variation in seed output across populations We calculated seed output for each tree as crop size x proportion of fruits not damaged by arthropods x number seeds per fruit x proportion of full seeds, and then used these to estimate an average reproductive output for each population. To assess the relationships between drought index, fruit production, fruit damage by arthropods and seed output at the population level, we used correlations between averaged values of these variables for the 14 populations. Results Variation in seed number Overall, seed numbers per fruit ranged from 1.5 to 6.0 seeds (n = 409 trees), while the average number of seeds per fruit in each of the 14 populations ranged from 3.0 to 4.1. The mixed model indicated that individual fruit production was the main predictor of the number of seeds per fruit (Table 2), so that trees that produced more fruits also had more seeds per fruit. The number of seeds per fruit increased linearly from trees with low to those with large crops (Fig. 2), except for the few trees we sampled with the largest crops of all (n = 4 trees). Drought index and annual precipitation had no effect on seed number (Table 2). Variation in the proportion of full, empty and aborted seeds The proportion of viable to non-viable seeds was very variable among populations. The lowest proportion of full seeds per fruit was recorded in Ciria (0.07) and the highest in the northern populations of Barrios de Luna and Peña Lampa (0.52 and 0.51, respectively). Barrios de Luna and Peña Lampa also had a low proportion of empty seeds (0.38, in both populations), while Ciria had the highest proportion (0.78). The proportion of aborted seeds per fruit ranged from 0.06 in Puerto de Escandón to 0.21 in Nerpio. Pollination and predator satiation determine reproduction 99 Fig. 2. Number of seeds per fruit for Spanish juniper trees (n = 409) producing diverse crop sizes (as estimated by an index of relative fruit abundance) in 14 populations in the Iberian Peninsula. Numbers above bars indicate the sample of trees for each index of fruit production. Fixed effects Estimate SE PRandom SD Number of seeds per fruit Intercept 3.583 0.608 <0.001 Population 0.273 Fruit production per tree 0.128 0.043 0.003 Drought index 0.202 0.121 0.122 Annual precipitation -0.017 0.024 0.502 Proportion of full seeds Intercept -2.917 1.253 0.020 Population 0.625 Fruit production per tree 0.126 0.025 <0.001 Drought index 0.499 0.250 0.046 Annual precipitation 0.074 0.051 0.149 Proportion of empty seeds Intercept 1.564 0.913 0.087 Population 0.453 Fruit production per tree -0.065 0.023 0.005 Drought index -0.157 0.182 0.387 Annual precipitation -0.051 0.037 0.173 Proportion of aborted Intercept -1.120 0.596 0.060 Population 0.278 seeds Fruit production per tree -0.099 0.034 0.003 Drought index -0.484 0.120 <0.001 Annual precipitation -0.024 0.024 0.315 Table 2. Estimates and SE from mixed models for number of seeds per fruit and the proportion of full, empty or aborted seeds. Significant P-values are shown in bold. The SD of the population effect on the estimates of the intercept as random factor is also shown Capitulo 4 100 Mixed models indicated that fruit production per tree was a significant predictor of the proportion of full, empty and aborted seeds (Table 2). Trees that produced more fruits had a higher proportion of full seeds and lower proportion of empty and aborted seeds (Table 2, Fig. 3). In addition, drought index had a significant effect on the proportion of full and aborted seeds. Drier than average conditions during the first year of fruit formation and growth (i.e. low values of the drought index) significantly increased the proportion of aborted seeds and decreased the proportion of full seeds per fruit (Table 2). Variation in fruit damage by arthropods Fruit damage by arthropods was varied highly among populations (see Appendix S1). Levels of fruit damage by mites ranged from 6.4% in Puerto de Escandón to 22.5% in Ciria (Appendix S1). Seed predation by chalcid wasps was lowest in Cabrejas del Pinar (3.3%) and highest in Pina de Ebro (29.7%; Appendix S1). Moths were the main predispersal predators in most populations, with predation rates ranging from 9.8% in Judes to 52.% in Peña Lampa (Appendix S1). Mixed models showed that fruit damage from mites was negatively influenced by individual fruit production and by wetter than average conditions during the first year of fruit formation, and marginally so with higher annual precipitation (Table 3). Seed predation by chalcid wasps was negatively affected by the incidence of the other two arthropods at the same tree, and increased with fruit production per tree (Table 3). Chalcid wasps preferred trees that produced smaller fruits with more aborted and also heavier seeds, but they predated less in sites with higher average annual precipitation (Table 3). Moths preferred trees with larger fruits and heavier seeds, and avoided trees with higher levels of fruit damaged by mites and chalcid wasps (Table 3). Seed output across populations Population-level crop sizes for the surveyed populations were generally medium to low, with large crops occurring in just two populations (e.g. Cobos del Cerrato and Lerma, with an average index of relative fruit abundance >3), as expected for a species with important Pollination and predator satiation determine reproduction 101 inter-annual fluctuations in seed production. Fruit crop size at the population level was positively correlated with drought index (r = 0.70, P = 0.005, n = 14 populations), with wetter than average conditions for each site leading to larger crops. Larger crops, in turn, led to increased population-level seed output (r = 0.84, P < 0.001; Fig. 4). Seed output for the 14 populations showed a significant negative relationship with damage by mites (r = -0.76, P = 0.001), but no relationship with seed predation by chalcid wasps (r = -0.39, P = 0.17) or fruit damage by moths (r = 0.05, P = 0.86). Discussion Production of viable seed in Spanish juniper in the Iberian Peninsula was highly influenced by fruit crop size and environmental conditions during the year of flowering and initial fruit formation. Fruit damage levels by arthropods responded to multiple factors, including crop size, fruit traits, environmental conditions for each population and interactions between arthropods. Overall, wetter than average conditions during early fruit development led to larger crops, which in turn led to increased population-level seed output and lower levels of fruit damage by mites, the least mobile of the three arthropods. These results suggest that female trees that invest more in reproduction during years with less water stress increase both their seed set and seed viability. Fig. 3. Proportion of full (black), empty (grey) and aborted (white) seeds per fruit for Spanish juniper trees (n = 409) producing diverse crop sizes (as estimated by an index of relative fruit abundance) in 14 populations in the Iberian Peninsula. Error bars represent 1 SE. Numbers above columns indicate the sample of trees for each index of fruit production. Capitulo 4 102 Crop size for each population was positively correlated with more favourable local environmental conditions (i.e. wetter than average conditions during the period that included flowering, pollination and initial fruit formation), but not to mean local conditions. This result is not unexpected because good environmental conditions are commonly related to higher reproductive effort in woody perennials (Kelly & Sork 2002), and temporal variations within populations are likely to be dictated by available reserves (Koenig & Knops 2000; Crone et al. 2009). For instance, in the Spanish juniper, experimenta monthly addition of water Fruits damaged by Fixed effects Estimate S.E. PRandom SD Mites Intercept -0.657 0.598 0.272 Population 0.277 Fruit production per tree -0.090 0.036 0.012 Drought index -0.413 0.120 <0.001 Annual precipitation -0.047 0.024 0.051 Chalcid wasps Intercept 2.809 1.219 0.021 Population 0.543 Mites -1.392 0.245 <0.001 Moths -0.734 0.195 <0.001 Fruit production per tree 0.074 0.036 0.040 Fruit diameter -0.125 0.055 0.022 no. full seeds -0.033 0.058 0.570 no. empty seeds -0.011 0.050 0.825 no. aborted seeds 0.178 0.068 0.009 Seed mass 4.006 1.355 0.003 Drought index -0.166 0.221 0.453 Annual precipitation -0.158 0.046 <0.001 Moths Intercept -2.298 1.312 0.080 Population 0.625 Mites -1.021 0.172 <0.001 Chalcid wasps -0.753 0.199 <0.001 Fruit production per tree -0.036 0.026 0.172 Fruit diameter 0.100 0.040 0.012 no. full seeds -0.075 0.043 0.078 no. empty seeds 0.029 0.037 0.432 no. aborted seeds 0.032 0.054 0.553 Seed mass 2.106 1.001 0.035 Drought index 0.165 0.251 0.511 Annual precipitation 0.023 0.051 0.647 Table 3. Estimates and SE from generalised mixed models for the incidence of mites, chalcid wasps and moths. Significant P-values are shown in bold. The SD of the population effect on estimates of the intercept as random factor is also shown Pollination and predator satiation determine reproduction 103 and nutrients to female trees from pollination to fruit ripening resulted in larger tree crops (Montesinos et al. 2012). Larger crops during favourable years were also associated with a higher number of seeds being set (a higher proportion of full seeds and lower abortion rates). This pattern was also observed in individual trees; trees that produced more fruit had fruits with more viable seeds and less empty and aborted seeds. Thus, the combined positive effect of large crops at the population and individual level on seed viability suggests that more efficient pollination combined with the availability of resources and reserves played a role in improving tree fitness. Pollination failure is a significant cause of seed loss in conifers, particularly in dioecious wind-pollinated species, such as Spanish juniper (García et al. 2002; Gruwez et al. 2013), and many junipers and other Cupressaceae regularly produce high proportions of non-viable seeds (Fuentes & Schupp 1998; García et al. 2000; Wesche et al. 2005; Rumeu et al. 2009). Empty seeds are externally similar to full seeds and correspond to seeds aborted during or immediately after fertilisation (Gruwez et al. 2013). Causes for the production of empty seeds include lack of pollination (quantity) or problems with fertilization (quality), so that less empty seeds per fruit in trees with larger crop sizes would suggest a more efficient pollination. Our results suggest that larger crops were also associated with an increase in the number of viable seeds and a reduction in the number of aborted seeds, which is in agreement with the pollination efficiency hypothesis (Nilsson & Wastljung 1987). AB Fig. 4. Seed output estimated for 14 Spanish juniper populations in the Iberian Peninsula in relation to (A) crop size and (B) rate of fruits damaged by mites. Capitulo 4 104 Alternatively, seed abortion after fertilisation may be the result of selective allocation of resources to the developing seeds (Obeso 2004). Seed set may be limited by resources and reserves available to seed development in woody perennials (Koenig & Knops 1998; Crone et al. 2009; Ida et al. 2013). Indeed, overproduction of flowers and subsequent abortion of fruits and seeds is a frequent phenomenon (Obeso 2004). For example, 75% of Spanish juniper fruits that began to develop and grow were lost after the first summer (E. Rodríguez, personal observation) and < 6 % of the initial crop finished the ripening period (Tellería et al. 2011). Our results indicate that good local environmental conditions drove larger crops and that individuals producing more fruits showed increased pollination success and were likely to allocate more resources and reserves to fruits and seeds (leading to more viable seeds per fruit). In short, environmental conditions (resources) modulated crop size that, in turn, determined seed viability (pollen limitation), coupled with maternal resources for fruit development (resource limitation). Damage by arthropods differed by a factor of 2.5 across the 14 populations, with even larger variations when damage by individual arthropod species was assessed. Variability was higher among trees within populations (see also Roques et al. 1984), and there was a strong relationship between individual tree characteristics and levels of fruit and seed damage by arthropods. Mites are important agents of seed damage in Spanish juniper (Roques et al. 1984). We found that populations producing larger crops during wetter than average years showed a lower proportion of fruits damaged by mites, and that the incidence of mites was negatively correlated with the number of viable seeds available for dispersers. Thus, satiation of mites at the population level is presumably a significant determinant of female reproductive output, as predicted by the predation satiation hypothesis (Janzen 1971; Nilsson and Wastljung 1987). The satiation effect is enhanced in large crops that follow years of low fruit production (Turgeon et al. 1994; Poncet et al. 2009; Mezquida & Olano 2013), so predator satiation would be better explored by examining the temporal functional response of the different predators (Linhart et al. 2014). The satiation effect may also occur at the scale of individual trees for specialised seed predators with low dispersal capacity, such as mites (Nilsson & Wastljung 1987; Linhart et al. 2014). In fact, we found that individual trees were able to satiate mites by producing more fruit, which would favour individual variability in seed production (Linhart et al. 2014). In support of satiation at this small scale, we previously demonstrated the influence of crop size on fruit damage by mites at the scale of individual trees, and consistent temporal variations in relative rates of fruit damage within individuals (Mezquida & Olano 2013). Pollination and predator satiation determine reproduction 105 Chalcid wasps did not show a functional response to population-level crops, but responded to crop size, fruit characteristics and the incidence of other arthropods on individual trees. Trees with abundant fruits attracted ovipositing female chalcid wasps, as is common in insect frugivores (Jordano 1987; Sallabanks & Courtney 1992). These small wasps are seed predators whose larvae develop inside juniper seeds so, after finding a tree, females avoid ovipositing in fruits already attacked by mites or moths. The avoidance of fruits occupied by other arthropods is stronger within large crops, when there are more undamaged fruits available to choose from (Mezquida and Olano 2013). Female chalcid wasps insert the ovipositor into the fruit to reach the seeds and lay eggs when fruits are almost fully grown (Roques et al. 1984; see chapter 2), and it follows that seeds inside smaller fruits (particularly those with less pulp) would be easier to reach by ovipositing females. Seed mass per fruit is largely determined by the number and size of full seeds, so the preference of chalcid wasps is likely to be for trees that have fruits with large, viable seeds, as they would logically improve larval development and survival rates (Napela & Grissell 1993; Fidgen et al. 1998). Finally, at a geographic scale the incidence of chalcid wasps was lower for populations with higher average annual precipitation, suggesting a direct impact of climate conditions on these small seed predators (Poncet et al. 2009; Montesinos et al. 2010). As with chalcid wasps, the incidence of moths was not related to local crops nor to population reproductive output, yet showed preferences for certain fruit traits and the absence of other arthropods in individual trees. Fruit damage by moths increased in trees having larger fruits with heavier seeds, which is consistent with our earlier findings in a juniper woodland producing a large crop (Mezquida & Olano 2013). Moth larvae develop by feeding on the pulp and seeds before pupating in the soil, so larger fruits would be preferred in order to provide more resources for the larvae, increasing their chances of survival (Sallabanks & Courtney 1992; Desouhant et al. 2000). Ovipositing female moths seemed to avoid juniper trees with fruits already damaged by mites and chalcid wasps, as previously observed (Mezquida & Olano 2013). The Spanish juniper is a masting species that produces significant seed crops once or twice every 10 years (Montesinos et al. 2012). Therefore, crop size figures recorded for the 14 populations we surveyed were only a 1-year snapshot of the inter-annual variation in fruit production (Montesinos et al. 2012; Mezquida & Olano 2013; Tellería et al. 2014). However, they provide a clear picture of the main factors determining reproductive output in Capitulo 4 106 this juniper. The increase in reproductive output in large crops seemed to be caused by a combination of pollination efficiency (quantity and quality components; Ida et al. 2013) and more resources being devoted to fruit development, resulting in more viable and less empty and aborted seeds. Environmental conditions modulated fruit production in female trees, although reserves accumulated during previous years are probably also important (Kelly & Sork 2002; Montesinos et al. 2012). There was a positive relationship between fruit production, at both tree and population level, and seed set, and large crops produced a satiation effect of the low-mobile predispersal predators (mites), thus leading to a positive feedback on their seed output before dispersal. The other two predispersal seed predator arthropods did not show this satiation effect as a consequence of their ability to cope with variable fruit crops through prolonged diapause (Turgeon et al. 1994), although their incidence can decrease after several years of no or low fruit production (Mezquida & Olano 2013). Our study supports the assertion that large crops during favourable local environmental conditions are associated with increased efficiency of pollination and the satiation of seed predators, which are important determinants of reproductive output in this dioecious, wind-pollinated juniper, inhabiting fluctuating environments (Kelly & Sork 2002; Crone et al. 2009; Montesinos et al. 2012). Pollination and predator satiation determine reproduction 107 References Agren J., Ehrlén J., Solbreck C. (2008) Spatio-temporal variation in fruit production and seed predation in a perennial herb influenced by habitat quality and population size. Journal of Ecology 96: 334–345. Auger-Rozenberg M.A., Kerdelhue C., Magnoux E., Turgeon J., Rasplus J.Y., Roques A. (2006) Molecular phylogeny and evolution of host-plant use in conifer seed chalcids in the genus Megastigmus (Hymenoptera: Torymidae). Systematic Entomology 31: 47–64. Blendinger P.G., Loiselle B.A., Blake J.G. (2008) Crop size, plant aggregation, and microhabitat type affect fruit removal by birds from individual melastome plants in the Upper Amazon. Oecologia 158: 273– 283. Chambers J.C., Vander Wall S.B., Schupp E.W. (1999) Seed and seedling ecology of pinon and juniper species in the pygmy woodlands of western North America. The Botanical Review 65: 1–38. Crone E.E., Lesica P. (2006) Pollen and water limitation in Astragalus scaphoides, a plant that flowers in alternate years. Oecologia 150: 40–49. Crone E.E., Miller E., Sala A. (2009) How do plants know when other plants are flowering? Resource depletion, pollen limitation and mast-seeding in a perennial wildflower. Ecology Letters 12: 1119– 1126. DeSoto L., Camarero J.J., Olano J.M., Rozas V. (2012) Geographically structured and temporally unstable growth responses of Juniperus thurifera to recent climate variability in the Iberian Peninsula. European Journal of Forest Research 131: 905–917. Desouhant E., Debouzie D., Ploye H., Menu F. (2000) Clutch size manipulations in the chestnut weevil, Curculio elephas: fitness of oviposition strategies. Oecologia 122: 493–499. Fidgen L.L., Quiring D.T., Sweeney J.D. (1998) Effect of cone size on adult and larval foraging behavior of Strobilomyia neanthracina and Strobilomyia appalachensis (Diptera: Anthomyiidae). Environmental Entomology 27: 877–884. Fuentes M., Schupp E.W. (1998) Empty seeds reduce seed predation by birds in Juniperus osteosperma. Evolutionary Ecology 12: 823–827. García D., Zamora R., Gómez J.M., Hodar J.A. (2002) Annual variability in reproduction of Juniperus communis L. in a Mediterranean mountain: relationships to seed predation and weather. Ecoscience 9 251– 255. García D., R., Gómez J.M., Jordano P., Hodar J.A. (2000) Geographical variation in seed production, predation and abortion in Juniperus communis throughout its range in Europe. Journal of Ecology 88: 435–446. You’d better walk alone: Changes in forest composition affect pollination efficiency and predispersal fruit damage in Iberian Juniperus thurifera forests. “Déjame a solas con mi sombra que no tengo hueco para nada ni nadie más.” Javier Gallego Crudo CHAPTER 5 Capitulo 5 116 Abstract Changes in land use patterns are a major driver of global environmental change. In developed countries, cessation of traditional management practices led to forest expansion and shifts in forest composition: traditionally managed monospecific forests moved towards mixed forests. However, there is a scarce knowledge on how the presence of other tree species will affect reproduction of formerly dominant species. We explored this question in the wind-pollinated tree Juniperus thurifera. We hypothesized that the presence of heterospecific trees would have a negative effect on pollination due to pollen interference, however they would lead to a reduction in specialized fruit predators, and consequently also have a positive effect on reproduction. We assessed the relative importance of forest composition on fruit production, seed development and predispersal fruit damage on nine paired pure and mixed J. thurifera forests in three regions across the Iberian Peninsula. The effects of forest composition on crop size, fruit and seed characteristics, and damage by predispersal arthropods were tested by mixed models. Fruit production was lower and seed abortion higher in mixed forests, suggesting higher pollination failure. In contrast, fruit damage by arthropods was higher in pure forests, supporting the hypothesis that the presence of non-host plants reduces damage rates. Arthropods response to forest composition was species-specific, and relative damage rates varied depending on individual tree crops. Overall, enhanced crop size in pure forests compensated for increased damage rates, leading to greater net production of sound seeds. This study indicates that ongoing changes in forest composition after land abandonment may impact tree reproduction. Changes in forest composition reduces pollination and predation 117 Introduction Land use change is impacting on major ecosystem services such as climate regulation (Hansen et al. 2001, Foley et al. 2005) and the carbon cycle (Rey Benayas et al. 2007). Forest conversion into croplands and pasture is one of the most significant change in land use across the globe, particularly in non-developed countries (Gibbs et al. 2010, Phelps et al. 2013). However, the opposite pattern is occurring in developed countries, where urbanization and agricultural intensification are leading to the abandonment of low productivity lands and cessation of traditional management practices (Rey Benayas et al. 2007, Valladares et al. 2014). As a result, secondary succession is altering large tracts of land (Rey Benayas et al. 2007, Gimeno et al. 2012a), open areas encroachment (DeSoto et al. 2010, Ewers et al. 2013), the expansion of woodlands into abandoned lands (Gimeno et al. 2012b), forest densification (Rey Benayas et al. 2007, Améztegui et al. 2010), modification of perturbation patterns (Rey Benayas et al. 2007, Valladares et al. 2014) and changes in forest composition (Hansen et al. 2001, Chauchard et al. 2007, Olano et al. 2012, Vayreda et al. 2016). These changes are driving shifts in the abundance of different community components and ecosystem processes (Laiolo et al. 2004, Sirami et al. 2008, Herrando et al 2016). Changes in forest structure and composition after abandonment modify competition levels among adult trees (Gimeno et al. 2012b, Vayreda et al. 2016). On the one hand, forest densification increases competition among conspecifics (Kenkel 1988, Getzin et al. 2006, Wang et al. 2016), while at the same time the colonization of forests by other woody species can intensify interspecific competition (Costa et al. 1997, Montesinos and Fabado 2015). Secondly, forest regeneration and the colonization of abandoned lands can enlarge the area covered by forest, thus reducing fragmentation and improving reproductive success by more effectively attracting animal pollinators and seed dispersers (Santos and Tellería 1994, González-Varo et al. 2009). Furthermore, increased tree density in formerly open forests or woodlands could also enhance reproductive success in wind-pollinated species due to a higher density of conspecifics that both increases pollen load and reduces the necessary pollen dispersal distances (Broadhurst 2015). However, a higher density of trees and fruits may also attract predispersal seed predators that can severely reduce the number of sound seeds before dispersal (Sholes 2008; Guyot et al. 2016). Nevertheless, these potential effects of forest densification could be different if tree density increase is driven Capitulo 5 118 by a higher frequency of heterospecifics, leading to mixed forests. The canopy of other tree species could hinder pollen dispersal in wind-pollinated species, whereas the presence of pollen from other species with overlapping flowering phenology could lead to reduced pollination efficiency and seed set (Mugnaini et al. 2007, Aderkas et al. 2012, Millerón et al. 2012). Conversely, fruit damage by predispersal predators in mixed forests might be lower than in pure, monospecific forests if the presence of heterospecifics prevents specialized fruit predators from finding their host plants due to physical or chemical interference from non-host plants (Sholes 2008, Barbosa et al. 2009). In this study, we assess the consequences of changes in forest structure and composition on the reproduction of Juniperus thurifera L., an evergreen conifer tree endemic to the western Mediterranean. Rural exodus and changes in traditional management practices are driving the colonization of abandoned croplands and livestock pastures by J. thurifera within its distribution range (Rozas et al. 2008, Pías et al. 2014). Concurrently, the reduction in livestock density and grazing pressure is favoring colonization by oak and pine species that were historically scarce due to their lower tolerance of browsing (DeSoto et al. 2010). As a result, J. thurifera forests are rapidly shifting from pure, monospecific open forests to mixed, denser forests, over large geographical scales (Olano et al. 2012). Our aim was to explore the impact of the transition from pure to mixed forests on J. thurifera reproductive success. We hypothesized that lower pollen load and greater interference from heterospecifics in mixed forests would lead to lower production of berry-like cones (fruits hereafter) and higher seed abortion rates. Contrastingly, we expected that mixed forests would experience lower levels of fruit damage, due to the increasing difficulty of predators to find host plants. However, this outcome may depend on the biology and dispersal capacity of the predator species (Barbosa et al. 2009). Finally, we assessed the combined effects of pollination success and predispersal fruit damage in order to evaluate the overall reproductive output of J. thurifera in pure and mixed forests. Changes in forest composition reduces pollination and predation 119 Material and methods Study area and sampling design We sampled J. thurifera forests in three regions in the Iberian Peninsula in Guadalajara, Segovia and Soria provinces (Fig. 1). The climate in these regions is classified as continental Mediterranean with mean annual temperatures of 10-11 ºC and mean annual precipitation ranging from 530 to 740 mm, with a two-month long drought in summer (Table 1). Lithology is calcareous in Guadalajara and Soria and granitic in Segovia. Within each region, we selected three populations and in each population, we sampled two forests differing in composition: monospecific (hereafter pure) and mixed. A forest was considered pure if J. thurifera comprised more than 90 % of the tree stems and mixed if the proportion of juniper stems was below 50 %. To set cleared differences in the effect of composition, we deliberately avoided forests in which the proportion of juniper stems is between 51 and 89%. Three oak (Quercus) and two pine (Pinus) species were the other main tree species present in mixed juniper forests (Table 1). In each case, paired forests were located in the same mountain range and separated by between 500 and 1000 meters. Therefore, sampling design followed a double nested design with forest nested in population and population nested in region (Fig. 1, Table 1). Study species Juniperus thurifera is a long-lived dioecious tree, growing under a continental Mediterranean climate in Spain and Morocco, with smaller populations in France, Italy and Algeria. Female flowers are wind-pollinated; they flower during February-March, and fertilization takes place in March. Fruits grow until reaching their full size in September and then mature for over a year, ripening in October of the second year after pollination. Although a wide array of arthropods may feed on maturing J. thurifera fruits (Roques et al. 1984), predispersal fruit damage in the Iberian Peninsula is mainly driven by three arthropod taxa (Mezquida et al. 2016). A mite species (Trisetacus quadrisetus; Acari, Phytoptidae) uses Juniperus spp. seeds as growth chambers where the colony feeds and reproduces (Roques 1984, El Alaoui et al. 2013). Chalcid wasps (Megastigmus thuriferana; Hymenoptera, Torymidae) are seed Capitulo 5 120 Fig. 1 Distribution of Juniperus thurifera sampled populations and regions. J. thurifera distribution is represented in green. Abbreviations for each population appear in Table 1 predators that only oviposit eggs inside developing seeds of J. thurifera before the seed coat hardens (Rouault et al. 2004, Auger-Rozenberg et al. 2006). Finally, two moth species (Mesophleps oxycedrella; Lepidoptera, Gelechiidae; and Pammene juniperana; Lepidoptera, Tortricidae) are pulp and seed eaters of Juniperus spp. fruits, and are common in J. thurifera (Roques et al. 1984). Fruit production and seed set We estimated fruit production for each forest by randomly selecting 30 female trees selecting the females present along a linear transect in 2014 early autumn just prior to full ripening. Fruit production per tree was estimated after visually inspecting the whole canopy (Koenig et al. 1994) using an index of relative fruit abundance that ranged from 0 (no fruits) to 5 (very high fruiting). This qualitative test has been tested against fruit abundance measurements based on fruit counts showing high correlation (Koenig et al. 1994; Mezquida and Olano Changes in forest composition reduces pollination and predation 121 2013). An advantage of this index allows to compare is its potential to compare tree fruiting effort irrespective of tree size. When available, we collected 40 ripe fruits from each tree by sampling all around the canopy. Fruit production was very low in Bayubas de Abajo (Soria province; Table 1), so this population was excluded from further analyses. Seed and fruit traits were characterized at forest level for each of the 16 remaining forests by analyzing 10 fruits per tree in a subsample of eight randomly selected trees. For each fruit, maximum length and width were measured in the laboratory to the nearest 0.01 mm with a digital caliper and averaged to calculate fruit diameter. Fruits were opened to separate the seeds and pulp mass was measured to the nearest 0.01 mg with a digital scale after it had been oven-dried for 48 hours at 60ºC. Seeds were characterized according to the development stage of the megagametophyte. Every seed was examined under a dissecting microscope and assigned to one of the following categories according to Gruwez et al. (2013): ‘full’, for seeds showing a completely developed embryo and megagametophyte; ‘empty’, for seeds that interrupted the development of the megagametophyte and/or embryo after Region Population Position Elevation (m) Annual mean T (ºC) Precipitation (mm) Accompanying species Guadalajara Megina (MEG) 40° 38’N 1° 54’W 1250 10 740 Q. ilex, Q. faginea, P. nigra Pinilla de Molina (PIM) 40° 40’N 1° 52’W 1370 10 648 Q. ilex Vahermoso (VAL) 40° 47’N 1° 56’W 1150 10 606 P. pinaster Segovia Arcones (ARC) 41° 6’N 3° 42’W 1200 11 653 Q. pyrenaica, F. angustifolia Prádena (PRA) 41° 7’N 3° 40’W 1250 10 739 Q. pyrenaica, I. aquifolium Siguero (SIG) 41° 10’N 3° 37’W 1100 11 694 Q. pyrenaica, F. angustifolia Soria Calatañazor (CAL) 41° 41’N 2° 48’W 1050 10 637 Q. ilex Cabrejas del Pinar (CDP) 41° 46’N 2° 50’W 1100 10 691 Q. ilex, P. pinaster, P. sylvestris Bayubas de Abajo (BDA) 41° 30’N 2° 54’ 950 11 530 P. pinaster Table 1. Characteristics of the nine Juniperus. thurifera populations sampled in the Iberian Peninsula: region, population (population abbreviation), geographic location, elevation, annual mean temperature, average annual precipitation and accompanying species in mixed forests. Capitulo 5 122 fertilization, and that are externally similar to fully developed seeds although they do not contain (or only some remnants of) the megagametophyte or embryo; and ‘aborted’, for seeds that interrupted their development between pollination and fertilization, and that are visually recognizable as small, not completely developed seeds. Parthenocarpic seeds have been described for junipers (Fuentes & Schupp 1998, García et al. 2000), however they can abort megagametophyte an embryo development at different stages along seed maturation process (Gruwez et al. 2013) making difficult to differentiate from those truly parthenocarpic at mature stage. Predispersal fruit damage In order to estimate rates of fruit damage by arthropods, for each tree we used 30 fruit samples (when available). When less than 30 ripe fruits were available, we estimated fruit damage rates whenever there was a minimum of 20 fruits. Forests within Cabrejas del Pinar (Soria) population were excluded from these calculations due to the low fruit availability in the mixed forest; analyses are therefore based on seven paired forests. Fruits were opened and examined in the laboratory under a dissecting microscope to detect signs of damage by arthropods. We counted the number of fruits damaged by mites, chalcid wasps and moths relative to the total number of fruits per tree. Signs of fruit damage can be easily assigned to each of the three arthropod groups (Roques et al. 1984, Mezquida and Olano 2013): mites deform the seeds causing the elongation of their tips that usually stick out of the fruit surface, chalcid wasps make a circular hole in the apical region of the fruit, and moths make an irregular hole in different parts of the fruit surface. Seed output We calculated total seed output for each tree as the result of multiplying fruit production x proportion of fruits not damaged by arthropods x number of seeds per fruit x proportion of full seeds. This calculation included seed traits, and thus seed output could only be estimated for the eight randomly selected trees for each forest (i.e., 128 trees from 16 forests). This index of seed output provides an estimate of reproductive output per individual (Mezquida et al. 2016). Changes in forest composition reduces pollination and predation 123 Statistical analysis We used linear mixed models to assess whether the type of forest (pure or mixed) influenced juniper fruit production and fruit traits. The random components of these and subsequent models consisted in population and population nested within region. We tested for differences in two fruit traits (diameter and pulp mass) that could covary with seed traits and influence fruit damage by predispersal predators (Mezquida and Olano 2013, Mezquida et al. 2016). Moreover, fruit production may affect seed production and seed set by influencing pollination success and resource investment (Mezquida et al. 2016), so we included the index of fruit production as a covariate in the mixed models for total number of seeds and number of full, empty and aborted seeds. We used generalized mixed models with a binomial error structure to test for the effect of forest type on arthropod fruit damage. Fruit production for each tree was included as a covariate (Mezquida et al. 2016) as well as the interaction between fruit production and forest type. Since chalcid wasps and moths tend to avoid ovipositing on fruits already damaged by other arthropods (Mezquida and Olano 2013), we included the proportion of fruits damaged by each of the other arthropods as explanatory variables in the models for chalcid wasps and moths. Difference in total seed output between pure and mixed juniper forests were evaluated with a generalized mixed model with a Poisson error structure. Values were previously rounded to the nearest integer. A similar procedure was followed for all mixed models. We included population nested in region as random factor in all models. The fixed part of the mixed model was then selected by sequentially removing non-significant terms (Zuur et al. 2009). Because our study was focused on hypothesis testing, we used the Bayesian information criterion instead of the Akaike information criterion (Aho et al. 2014). Model calculations were performed using the nlme package in R environment (R Development Core Team, 2014). Capitulo 5 130 production was generally low-medium, any further reduction in fruit production because of heterospecific competition and pollen limitation resulted in even lower reproductive output in mixed forests. This is consistent with findings from experimental alleviations of environmental stress to female trees during reproduction that resulted in larger crop sizes (Montesinos et al. 2012b). Furthermore, differences in reproductive success between pure and mixed forests may also become more pronounced during mast years when trees invest more in reproduction (Mezquida et al. 2016), and predispersal fruit damage rates may become higher in mixed than in pure forests. Our study indicates that rapid changes in the structure of J. thurifera forests, and associated increases in interspecific competition (DeSoto et al. 2010, Gimeno et al. 2012a, Olano et al. 2012) might potentially affect juniper tree reproduction in the long term. This effect may act synergistically with increasing drought intensity (IPCC 2014), amplifying the negative effects of interspecific competition on J.thurifera under more xeric conditions (GómezAparicio et al. 2011). Overall, our work highlights the need to explore the biological impact of current changes in landscape configuration in order to forecast species responses to global change. Changes in forest composition reduces pollination and predation 131 References Aderkas P., Nepi M., Rise M., Buffi F., Guarnieri M., Coulter A., Gill K., Lan P., Rzemieniak S., Pacini E. (2012) Post-pollination prefertilization drops affect germination rates of heterospecific pollen in larch and Douglas-fir. Sex Plant Reproduction 25: 215-225. Aho K., Derryberry D.W., Peterson T. (2014) Model selection for ecologist: the worldviews of AIC and BIC. Ecology 95: 631-636. Améztegui A., Brotons L., Coll L. (2010) Land-use changes as major drivers of mountain pine (Pinus uncinata Ram.) expansion in the Pyrenees. Global Ecology and Biogeography 19: 632-641. Auger-Rozenberg M.A., Kerdelhue C., Magnoux E., Turgeon J., Rasplus J.Y., Roques A. (2006) Molecular phylogeny and evolution of host-plant use in conifer seed chalcids in the genus Megastigmus (Hymenoptera: Torymidae). Systematic Entomology 31: 47–64. Arista M., Talavera P.L. (2001) Reproductive cycles of two allopatric subspecies of Juniperus oxycedrus (Cupressaceae). Flora 196: 114-120. Barbosa P., Hines J., Kaplan I., Martinson H., Szczpaniec A., Szendrei Z. (2009) Associational resistance and associational susceptibility: having right or wrong neigbors. Annual Review of Ecology Evolution and Systematics 40: 1-20. Bernays E.A., Chapman R.F. (1994) Host-plant Selection by Phytophagous Insects. New York (U.S.A.): Chapman and Hall. Broadhurst L. (2015) Pollen dispersal in fragmented populations of the dioecious wind-pollinated tree, Allocasuarina verticillata (Drooping She-Oak: Allocasuarinaceae). PLoS One 10: e0119498 Costa M., Morla C., Sainz H. (1997) Los bosques ibéricos: una interpretación geobotánica. Barcelona (Spain): Planeta. 597p. Chauchard S., Carcaillet C., Guibal F. (2007) Patterns of land-use abandonment control tree-recruitment and forest dynamics in Mediterranean mountains. Ecosystems 10: 936–948. DeSoto L., Olano J.M., Rozas V., De la Cruz M. (2010) Release of Juniperus thurifera woodlands from herbivore-mediated arrested succession in Spain. Applied Vegetation Science 13: 15-25. El Alaoui M.A., Yart A., Roques A., Arjouni Y., El Mercht S., Rozenberg M.A., Romane A. (2013) Acariens et insectes ravageurs de deux cupressacées menacées au Maroc: le Genévrier thurifère et le cyprès de l’Atlas. Ecologia Mediterranea 39: 123-128. Ewers R.M., Didham R.K., Pearse W.D., Lefebvre V., Rosa I.M.D., Carreiras J.M.B., Lucas R.M., Reuman D.C. (2013) Using landscape history to predict biodiversity patterns in fragmented landscapes. Ecology Letters 16: 1221-1233. Capitulo 5 132 Foley J.A., DeFries R., Asner G.P., Barford C., Bonan G., Carpenter S.R., Chapin F.S., Coe M.T., Daily G.C., Gibbs H.K., Helkowski J.H., Holloway T., Howard E.A., Kucharik C.J., Monfreda C., Patz J.A., Prentice I.C., Ramankutty N., Snyder P.K. (2005) Global consequences of Land use. Science 309: 570-574. Friedman J., Barret S.C.H. (2009) Wind of change: new insights on the ecology and evolution of pollination and mating in wind pollinated plants. Annals of Botany 130: 1515-1527. García D., Zamora R., Gómez J.M., Jordano P., Hodar J.A. (2000) Geographical variation in seed production, predation and abortion in Juniperus communis throughout its range in Europe. Journal of Ecology 88: 435–446. Getzin S., Dean C., He F., Trofymow J.A., Wiegand K., Wiegand T. (2006) Spatial patterns and competition of tree species in a Douglas-fir chronosequence on Vancouver Island. Ecography 29: 671-682. Gibbs H.K., Ruesch A.S., Achard F., Clayton M.K., Holmgren P., Ramankutty N., Foley J.A. (2010) Tropical forests were the primary sources of new agricultural land in the 1980s and 1990s. Proceedings of the National Academy of Sciences 107: 16732-16737. Gimeno T.E., Escudero A., Delgado A., Valladares F. (2012a) Previous land uses alters the effect of climate change and facilitation on expanding woodlands of Spanish Juniper. Ecosystems 15: 564579. Gimeno T.E., Pías B., Martínez-Fernández J., Quiroga D.L., Escudero A., Valladares F. (2012b) The decreased competition in expanding versus mature juniper woodlands is counteracted by adverse climatic effects on growth. European Journal of Forest Research 131: 977-987. Gómez-Aparicio L., García-Valdés R., Ruíz-Benito P., Zavala M.A. (2011) Disentangling the relative importance of climate, size and competition on tree growth in Iberian forests: implications for forest management under global change. Global Change Biology 17: 2400-2414. González-Varo J.P., Arroyo J., Aparicio A. (2009) Effects of fragmentation on pollinator assemblage, pollen limitation and seed production of Mediterranean myrtle (Myrtus communis). Biological Conservation 142: 1058–1065. Gruwez R., Leroux O., De Frenne P., Tack W., Viane R., Verheyen K. (2013) Critical phases in the seed development of common juniper (Juniperus communis). Plant Biology 15: 210–219. Guyot V., Castagneyrol B., Vialatte A., Deconchat M., Jactel H. (2016) Tree diversity reduces pest damage in mature forest across Europe. Biology Letters 12: 20151037. http://dx.doi.org/10.1098/ rsbl.2015.1037. Hansen A.J., Neilson R.P., Dale V.H., Flather C.H., Iverson L.R., Currie D.J., Shafer S.C.R., Bartlein P.J. (2001) Global change in forests: Responses of Species, communities and biomes. Bioscience 51: 765-779 Herrando S., Brotons L., Anton M., Páramo F., Villero D., Titeux N., Quesada J., Stefanescu C. (2016) Assessing impacts of land abandonment on Mediterranean biodiversity using indicators based on bird and butterfly monitoring data. Environmental Conservation 43: 69-78. Changes in forest composition reduces pollination and predation 133 IPCC. (2014) Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Geneva (Switzerland): IPCC. 151 pp. Kenkel N.C. (1988) Pattern of self-thinning in Jack pine: testing the random mortality hypothesis. Ecology 69:1017-1024. Koenig W.D., Knops J.M.H., Carmen W.J., Stanback M.T., Mumme R.L. (1994) Estimating acorn crops using visual surveys. Canadian Journal of Forest Research 24: 2105–2112. Laiolo P., Dondero F., Ciliento E., Rolando A. (2004) Consequences of pastoral abandonment for the structure and diversity of the alpine avifauna. Journal of Applied Ecology 41: 294-304. Mezquida E.T., Olano J.M. (2013) What makes a good neighborhood? Interaction of spatial scale and fruit density in the predator satiation dynamics of a masting juniper tree. Oecologia 173: 483-492. Mezquida E.T., Rodríguez-García E., Olano J.M. (2016) Efficiency of pollination and satiation of predators determine reproductive output in Iberian Juniperus thurifera woodlands. Plant Biology 18: 147-155. Millerón M., Heredia U.L., Lorenzo Z., Perea R., Dounavi A., Alonso J., Gil L., Nanos N. (2012) Effect of canopy closure on pollen dispersal in a wind-pollinated species (Fagus sylvatica L.). Plant Ecology 213: 1715-1728 Montesinos D., Fabardo J. (2015) Changes in land use and physiological transitions of Juniperus thurifera forest: from decline to recovery. Canadian Journal of Forest Research 45: 764–769 Montesinos D., García-Fayos P., Verdú M. (2012a) Masting uncoupling: mast seeding does not follow all mast flowering episodes in a dioecious juniper tree. Oikos 121: 1725–1736. Montesinos D., Villar-Salvador P., García-Fayos P., Verdu M. (2012b) Gender in Juniperus thurifera have different functional responses to variations in nutrient availability. New Phytologist 193: 705–712. Mugnaini S., Nepi M., Guarnieri M., Piotto B., Pacini E. (2007) Pollination Drop in Juniperus communis: Response to Deposited Material. Annals of Botany 100: 1475-1481. Olano J.M., Zavala M.A., Rozas V. (2012) Disruption of Juniperus thurifera woodland structure in its northwestern geographical range: potential drivers and limiting factors. European Journal of Forest Research 131: 563–570. Owens J.N. (1995) Constraints to seed production: temperate and tropical forest trees. Tree Physiology 15: 447–484. Pías B., Escribano-Ávila G., Virgós E., Sanz-Pérez V., Escudero A., Valladares F. (2014) The colonization of abandoned land by Spanish juniper: Linking biotic and abiotic factors at different spatial scales. Forest Ecology and Management 329: 186-194. Phelps J., Carrasco L.R., Webb E.L., Koh L.P., Pascual U. (2013) Agricultural intensification escalates future conservation costs. Proceedings of the National Academy of Sciences 110: 7601-7606. Capitulo 5 134 Plath M., Dorn S., Riedel J., Barrios H., Mody K. (2012) Associational resistance and associational susceptibility: specialist herbivores show contrasting responses to tree stand diversification. Oecologia 169: 477-487. R Development Core Team (2014) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. Rey Benayas J.M., Martins A., Nicolau J.M., Schulz J.J. (2007) Abandonment of agricultural land: an overview of drivers and consequences. CAB reviews: perspectives in agriculture, veterinary sciences, nutrition and natural resources 2. Wallingford (UK): CABI Publishing. 14p. Roques A., Raimbault J.P., Goussard F. (1984) La colonization des cônes et galbules des genévriers méditerranéens par les insectes et acariens et son influence sur les possibilités de régénération naturelle de ces essences. Ecologia Mediterranea 10: 147–169. Rouault G., Turgeon J., Candau J.N., Roques A., von Aderkas P. (2004) Oviposition strategies of conifer seed chalcids in relation to host phenology. Naturwissenschaften 91:472–480. Rozas V., Olano J.M. (2013) Environmental heterogeneity and neighbourhood interference modulate the individual response of Juniperus thurifera tree-ring growth to climate. Dendrochronologia 31:105-113. Rozas V., Olano J.M., DeSoto L., Bartolomé D. (2008) Large-scale structural variation and long-term growth dynamics of Juniperus thurifera trees in a managed woodland in Soria, central Spain. Annals of Forest Sciences 65: 809–809. Santos T., Tellería J.L.(1994) Influence of forest fragmentation on seed consumption and dispersal of Spanish juniper Juniperus thurifera. Biological Conservation 70: 129–134. Sholes O.D.V. (2008) Effects of associational resistance and host density on woodland insect herbivores. Journal of Animal Ecology 77: 16-23. Sirami C., Brotons L., Burfield I., Fonderflick J., Martin J. (2008) Is land abandonment having an impact on biodiversity? A meta-analytical approach to bird distribution changes in the north-western Mediterranean. Biological Conservation 141: 450-459. Vayreda J, Martinez-Vilalta J, Gracia M, Canadell J, Retana J (2016) Anthropogenic-driven rapid shifts in tree distribution lead to increased dominance of broadleaf species. Global Change Biology 22: 3984:3995 Turgeon J.J., Roques A., Groot P. (1994) Insect fauna of coniferous seed cones: Diversity, host plant interactions and management. Annual Review of Entomology 39: 179-212. Valladares F., Benavides R., Rabasa S.G., Díaz M., Pausas J.G., Paula S., Simonson W.D. (2014) Global Change and Mediterranean forests: current impacts and potentias responses in Forest and Global Change. Coomes DA, Burslem DFRP, Simonson WD, editors. Cambridge (U.K.): Cambridge University Press. p47-75. Wang Y., Pederson N., Ellison A.M., Buckley H.L., Case B.S., Liang E., Camarero J.J. (2016) Increased stem density and competition may diminish the positive effects of warming at alpine treeline. Ecology 97: 1668-1679. Wesche K., Ronnenberg K., Hensen I. (2005) Lack of sexual reproduction within mountain steppe populations of the clonal shrub Juniperus sabina L. in semi-arid southern Mongolia. Journal Arid Environments 63: 390–405. Zuur A., Leno E.N., Walker N., Saveliev A.A., Smith G.M. (2009) Mixed effects models and extensions in ecology with R. New York (USA): Springer. p549. “Keep truckin’ like the doodah man Together, more or less in line Just keep truckin’ on” Grateful Dead AGRADECIMIENTOS 138 En primer lugar, agradecer a Txemi y Eduardo la posibilidad de realizar este periplo sabinero, que en realidad comenzó en 2011 con una beca de formación de la UVa y el eCyL y se transformó en un proyecto de tesis a finales de 2012. En particular a Eduardo por su meticulosidad y perfeccionismo y a Txemi por su libertad, ímpetu y optimismo sobre todo en los últimos meses. En segundo lugar, a todas aquéllas, personas que de alguna u otra forma acabaron implicadas en el trabajo de la tesis. Txemi y Eduardo que año tras año hemos muestreado juntos en Cabrejas y en Checa. Especialmente estimable la ayuda de Txemi a la hora de recoger los frutos de enebro cada septiembre en Checa. A Ana por dejarme su coche al principio de la tesis durante las primeras visitas en Villaciervos y por su ayuda con R (siempre es un placer trabajar contigo). A Mike, Paula y Fer por su ayuda en campo en el seguimiento fenológico en el sabinar de Villaciervos. A Gonzalo y mi padre por acompañarme a muestrear en Siguero, Prádena y Arcones. A Enrique por preocuparse de preparar el pesticida para la parte experimental y Alfredo por acompañarme y hacer de fotógrafo en Villaciervos. A Vicente, me olvidaba, por acompañarnos a muestrear en Cabrejas. A Helios por sus comentarios y ayuda en la búsqueda de sabinares mixtos. A Olivier Leroux, por sus enseñanzas en el Jardín Botánico de Gante y su implicación en el procesado de las muestras. A Kris Verheyen por aceptarme dos veces en la universidad de Gante, sus interesantes conversaciones y abrir la vía para trabajar con Olivier. A Tomás Carlo por su más que cálida acogida en la Penn State, durante la estancia en Pennsylvania, por aportarme formas distintas de enfocar el trabajo, por ese diseño para un estudio en interacciones tritróficas, sigo pensando que podría funcionar, aunque los zorzales en Europa no son tan amables como en Estados Unidos. Gracias también por las profundas conversaciones sobre la ciencia, su mundo y relación con la situación socio-política actual. Me dijiste que si había llegado hasta aquí era para quedarme. Después de todos estos años sigo sin estar de acuerdo, siempre hay alternativas. En tercer lugar, agradezco a todas aquellas personas que me han acompañado estos últimos años, especialmente duros y en los que me ha costado y mucho seguir adelante. Nada que no se sepa ya, por otra parte. Ya aviso que seguramente me deje a alguien, daos por aludidos igualmente y no seáis rencorosos ni vengativos. O sí, no sé, como veáis las afectadas y afectados. A mis padres les agradezco los esfuerzos realizados estos años, y especialmente por ser los principales actores en aguantarme. A mi hermano, gracias por el apoyo moral y logístico, revisar la maquetación y por su ayuda en los proyectos fotográficos. 139 Sí, ya sé que aún tengo que subir a Vigo, ¿te recuerdo lo de Soria? Ángela gracias por tratar de animarme siempre, por ser tan honesta conmigo, por tu ayuda y consejos, sobre todo en los momentos más difíciles. Gracias por los comentarios sobre la tesis. Y Gracias también por las largas conversaciones sobre el sentido de la vida y Tintín. A Estos, simplemente por estar ahí, por vuestro apoyo durante la tesis y no ponerme el asunto muy difícil al volver tras largo tiempo de desconexión: Gonza, gracias por toda la commedia compartida, los falsos piques políticos y los cafés de los viernes, que no tienen por qué ser en viernes, es sólo una expresión, como patada en el culo, que no tiene porqué ser precisamente en el culo. Hermana, gracias por todos los momentos de enaltecimiento del partirlo, tus visitas a Soria y respetar tanto mi espacio cuando no me ha apetecido compartir nada. Espero que me perdones este cruel crimen. Diego, Fer, gracias también por vuestra comprensión y preocupación y las visitas a Soria. A Diego (otra vez) y a Cris, gracias por acogerme siempre en vuestra casa y dejarme dormir con Coco y Momo, y por llevarme y recogerme del aeropuerto. A Darío, por estar tan presente desde la lejana Dublín. Gracias Tom también por la genial idea del viaje a Irlanda y a Darío por acogernos, enseñarnos Dublín y compartir gran parte del viaje, fueron mis primeras vacaciones verdaderas en cinco años. Carlos Funes o Juan García, que odiosos podían ser los profesores del Laguna, gracias por dejarme colaborar en tu libro “A recoger bombas. Batallones de trabajo forzado en Castilla y León (1937 - 1942)”. Aprendí mucho hablando sobre el tema y discutiendo sobre otros en los viajes Soria – Segovia. A Juankella, gracias por vuestras visitas a Soria cuando nadie más pudo o quiso. Mike y Fran (el profesor de magia de altas energías), gracias también por vuestras visitas y excursiones por Soria. Se nos quedó pendiente el Moncayo (y otros muchos). Dani “Todólogo” e Inés, gracias por darme de comer y de dormir y enseñarme a fondo Utrecht, probablemente la ciudad con la mejor heladería del mundo. Jaime, sólo voy a decirte IECD, ¡qué recuerdos! Joserra, gracias por compartir tantas frikadas varias y me consta que no todas. Ángel, gracias por todos los viajes amenizados con tus llamadas por teléfono y por seguir dejándote la piel en el ayuntamiento como único concejal de izquierdas. Sí, ya sé que no os he nombrado a todos. Los que faltáis, coged cualquiera de los otros nombres y sustituidlos. El sentimiento es el mismo, la commedia y los buenos ratos han sido compartidos con todos, especialmente en las Tradicionales Cenas. P.J., P.D. y P.F. gracias por poner tantísima banda sonora a la tesis. Menos por el electroswing, me niego a dar gracias por ese maquiavélico invento. Elena gracias por tu compañía no sólo estos últimos años también desde la lejana Dublín, sino los