(Lamiaceae) Floral Mixture in Conservation of Pollinators in Central Spain

UNIVERSIDAD POLITÉCNICA DE MADRID

ESCUELA TÉCNICA SUPERIOR DE INGENIEROS AGRÓNOMOS

HABITAT MANAGEMENT FOR CONSERVATION OF

POLLINATORS IN AGRO-ECOSYSTEMS OF CENTRAL SPAIN

TESIS DOCTORAL

JELENA BARBIR

BSc in Biology Madrid, 2014

UNIVERSIDAD POLITÉCNICA DE MADRID

ESCUELA TÉCNICA SUPERIOR DE INGENIEROS AGRÓNOMOS

Habitat management for conservation of pollinators in

agro-ecosystems of Central Spain

TESIS DOCTORAL

Jelena Barbir

BSc in Biology

Madrid, 2014

Departamento de Producción Vegetal: Botánica y Protección Vegetal

Escuela Técnica Superior de Ingenieros Agrónomos

Habitat management for conservation of pollinators in

agro-ecosystems of Central Spain

PhD candidate: Jelena Barbir

BSc in Biology

Supervisors: José Dorado

PhD in Agronomy (ICA, CSIC, Madrid)

Francisco Rubén Badenes Pérez

PhD in Entomology (ICA, CSIC, Madrid)

Madrid, 2014

Tribunal nombrado por el Magfco. Y Excmo. Sr. Rector de la Universidad Politécnica de Madrid, el día de de 2014.

Presidente D./Da

Vocal D./D.a

Secretario D./D.a

Suplente D./D.a

Realizada la lectura y defensa de la tesis el día de de 2014 en Madrid, en la Escuela Técnica Superior de Ingenieros Agrónomos.

Calificación:

El Presidente Los Vocales

El Secretario

To the planet Earth ...

... with hope that this work will contribute its beauty and magnificence;

... and to my family.

Acknowledgments

First of all, I would like to use this opportunity to express gratitude to my supervisors, José Dorado and Francisco Rubén Badenes Pérez.

José, thank you for being like a father to me, for helping me to adapt easily to the new environment and group. Thank you for sharing with me your knowledge in the field and your passion for science and agriculture. Thank you for your patience, for all your time spent helping me out with my decisions, experimental designs and doubts. Also, thank you for being always available when I needed someone to talk about anything (work, stress, happy moments, vacations, health or even about missing my family). It was a real pleasure to work with you during those 4 years.

Ruben, thank you for sharing with me all your knowledge and passion for science and entomology. It was a unique experience to work with you and I am happy that I had an opportunity to learn so much from you about science and life in general. I would like to thank you for encouraging me to participate in several international conferences and for financing all the inscriptions and trips. Thank you for being always there for me when I needed an advice regarding my work, and also thank you for always being available to prioritize my deadlines.

Although not my supervisor, I feel like he was one of them. César, thank you for recognizing my passion for science and for giving me an opportunity to be part of your research team where I felt as another member of the family. Also, I want to thank you for being always so nice to me, and always finding time for me with a big smile on your face. Thank you for your positive energy!

In addition, I would like to thank to all members of the group “Weed Ecology”: to David and José Manuel, who were always there for me whenever I needed help in the field, lab or to share a tortilla in a bar; to Carolina, for her smiles and amazing trip to Japan; to Bea, for her positive energy, empathy, smiles and help in the field where we shared some great moments counting Plutellas; also special thanks to my friend, colleague and flatmate Dioni, for motivating me to finish my thesis as soon as possible and have life again:).

I would also like to use this chance to thank Elisa Viñuela, my tutor at Polytechnic University of Madrid, who helped me to identify some of the first pollinators in my study area and for always being so positive and energetic. Also, I would like to use this chance to thank all members of her group for being supportive and nice to me. Especially, I would like to thank to my friend Agustin for sharing with me all those happy and frustrating moments during the PhD thesis “process”, and to Celeste, who helped me in the field during that hot summer of 2012, patiently counting bees and keeping aromatic plants alive.

My thanks go to Cedo Maksimovic (Imperial College London, UK) and Tijana Blanusa (University of Reading, UK) who warmly received me during my 3-month research visit and helped me with the experiments I was conducting there. Also, thanks to Rebeka Siling who was helping me with my experiments in Reading and to Simon Potts and Robert Stuart for finding time to receive me in Reading and help me sort out some of my major doubts regarding experimental designs.

I would like to express my gratitude to Luis Oscar Aguado Martín and Mike Edwards. Those two entomologists made me understand bees. Their passion reflects through my work, my experiments and love for what I was doing those 4 years. Without this passion, all this work would be just boring. So, Oscar and Mike, thank you for enabling me to have fun during those 4 years of my PhD, and thanks to your wives for being so friendly and cooking such a lovely dishes while I was visiting you.

Probably none of my field experiments would turn out right without help of Pedro Hernaiz and his “crew”, who solved all my problems in the experimental farm “La Poveda”. Pedro, gracias por tu paciencia y todo el esfuerzo que hiciste para que mis experimentos salgan bien. Muchas gracias a José, Tasio, Jesús, Paulino y especialmente quiero dar las gracias a Carlos que venía incluso durante sus vacaciones a regar mis plantas en el campo.

I would like to thank to the Spanish National Research Council (CSIC) that awarded me with the JAEpredoc scholarship, to the head of the Institute of Agricultural Sciences - María Teresa García González, and to the CSIC staff: Cristina, Maria Cruz, Belen, Raimundo and Chelo for helping me out with all the paperwork and documentation I needed to deal with during those 4 years; to Clara for assuring me to get any weird or super old article I needed to include in my thesis; to Laura Barrios for her help with statistic analysis and experimental designs; to Antonio for having the cars always ready for me; to David and Victor for constantly repairing my office light; and last but not least, I would like to thank to my “polish” friend Casimiro for keeping my plants happy and in a good condition and to Elisa and Antonio who were always so helpful and nice to me, maintaining my greenhouse experiments running. Also, I would like to thank to all techs, master, PhD and postdoc students I have met in CSIC: to Clara, Sergio, Raul, Lilia, Mikel, Pablo, MJ, Vanesa, Irene, Francesca, Leti, Carlos, Carlos, Dani, Mario, Miguel, Celia, Anita, Rocco, Mariana... thanks for your friendship, for sharing with me Spanish food, Madrid and “out of Madrid” excursions, parties ...

Thanks to all those amazing people I have met in Madrid: José, Paula, Fahmi, Ruth (for being so patient correcting my English), Gorjana, Alicia, Ana Cobo ..., and of course to my “precious” friend Clemence, for being always there for me whenever I needed a nice word or a hug. To all my great international friends for being there for me: to Sergio, Paul, Andy, Marta,

Manal, Jojo, Jasna, Mike, Rob, Quentin, Hara, Jovanka, Meera, Gabi, Catalan group ... Many thanks to my fantastic Serbian friends: Jelena, Duska, Ivana, Marija, Sanja NS, Elvira, Vesna, Sanela, Ranko, Alpi, Zeljko... guys thank you for all those years of friendship and time we have spent together. Especially, I would like to thank to Sanja and Jovanka for always, always, always being there for me weather I am feeling good or bad, happy or sad, being annoying or super-fun. Thanks girls for always holding my back!

The most important people in my life, I saved for the end. To my parents, Ljubisa and Dragana, who passed through the war and fights, but they never stopped believing in life and people. I want to thank them, for transmitting that kind of thinking to me and for making me the person I am. Thank you for all fantastic moments, endless chats, trips, advices, hugs and positive energy... but above all thank you for believing in me. Mama, tata, ne zaboravite, vi ste najbolji roditelji koje jedno dete moze pozeleti. Ponosna sam na vas i volim vas do neba! I would like to thank to my sister, Natasa, for her endless love, for her happy and funny spirit that always cheers me up, for her “lousy” karaoke singing, for helping me out in the field and for never stopping to believe in me. Also, I would like to thank to my super-grandmas, Milica and Zlata, for being my inspiration and support. Hvala vam na vasoj neizmernoj ljubavi, na svakom telefonskom razgovoru, savetu, poljupcu i zagrljaju. Takodjer, hvala vam na svakom obroku, torti i soku pripremljenom sa ogromnom kolicinom ljubavi. Also, I would like to thank to the rest of my family for always being there for me: Vedo, Tiki, Ivana, Bojana, tetka and Saco, teco, strikan and Vesna, tetka Anica, Bojan, Bojan, Bilja, cika Mijo, and of course thanks to the youngest crew of our family: Stefku, Daci and Dori! Also, I thank to the newest members of our family, Alonso and Maite, for their love, support and for being part of my life.

Finally, I would like to thank to the love of my life and the best life partner one can imagine - Xavieru. Your love, support, indefinite patience and care for me keep me alive and joyful. I am grateful for every moment I have spent with you and I am looking forward to spend with you the rest of my life.

Table of Content

Abstract ...... i

Resumen ...... iv

I. General introduction ...... 1

1. Agriculture and pollination ...... 1

2. Plant-pollinator interactions ...... 3

3. Important groups of pollinators ...... 4

4. Managing semi-natural habitats of pollinators within agro-ecosystems10

II. Research question and objectives ...... 15

III. Functionality of aromatic (Lamiaceae) floral mixture in conservation of pollinators in Central Spain ...... 18

1. Introduction ...... 18

2. Materials and methods ...... 20

3. Results ...... 24

4. Discussion ...... 33

IV. The attractiveness of flowering herbaceous plants to bees (Hymenoptera: Apoidea) and hoverflies (Diptera: Syrphidae) in agro-ecosystems of Central Spain ...... 38

1. Introduction ...... 38

2. Materials and methods ...... 39

3. Results ...... 44

4. Discussion ...... 52

V. Wild rocket - effect of water deficit on growth, flowering and attractiveness to pollinators ...... 57

1. Introduction ...... 57

2. Material and methods ...... 58

3. Results ...... 63

4. Discussion ...... 77

VI. Can proximity of floral field margins improve pollination and seed production of coriander crop, Coriandrum sativum L. (Apiaceae)? ...... 81

1. Introduction ...... 81

2. Material and methods ...... 83

3. Results ...... 87

4. Discussion ...... 93

VII. General discussion ...... 96

VIII. Conclusions ...... 102

IX. Conclusiones ...... 104

X. References ...... 106

XI. Appendices ...... 121

1. List of abbreviations ...... 121

2. List of figures ...... 123

3. List of tables ...... 126

Abstract

Habitat management, aimed to conserve pollinators in agro-ecosystems, requires selection of the most suitable plant species in terms of their attractiveness to pollinators and simplicity of agronomic management. However, since all flowers are not equally attractive to pollinators and many plant species can be weedy or invasive in the particular habitat, it is important to test which plant species are the most appropriate to be implemented in specific agro-ecosystems. For that reason, this PhD dissertation has been focused on determination of the most appropriate aromatic and herbaceous plants for conservation of pollinators in agro-ecosystems of Central Spain.

Therefore, in a first approximation, spatial expansion (i.e. potential weediness) and attractiveness to pollinators of six aromatic perennial plants from the Lamiaceae family, native and frequent in the Mediterranean region, were evaluated. Preliminary plant selection was based on designing a functional mixed margins consisting of plants attractive to pollinators and with different blooming periods, in order to extend the availability of floral resources in the field. After a year of vegetative growth, the next two years the plant species were studied in a randomized block design experiment in order to estimate their attractiveness to pollinators in Central Spain and to investigate whether floral morphology and density affect attractiveness to pollinators. The final aim of the study was to evaluate how their phenology and attractiveness to pollinators can affect the functionality of a flowering mixture of these plants. In addition, the spatial expansion, i.e. potential weediness, of the selected plant species was estimated under field conditions, as the final purpose of the studied plants is to be implemented within agro-ecosystems. The results of the experiment showed that floral morphology did not affect the attractiveness of plants to pollinators, but floral density did, as plant species with higher floral density (i.e. Nepeta tuberosa and Hyssopus officinalis) showed significantly higher attractiveness to pollinators. In addition, of six plant species, two summer species (Melissa officinalis and Thymbra capitata) did not efficiently contribute to the attractiveness of the mixture to pollinators, which reduced its attractiveness during the summer period. Finally, as none of the plants showed weedy behaviour under field conditions, the attractive plant species, i.e. N. tuberosa, H. officinalis and the early spring flowering Salvia verbenaca, showed good potential to conserve the pollinators.

Abstract

Similarly, in a second approximation, the attractiveness to pollinators and agronomic behaviour of twelve herbaceous plants blooming in spring were studied. This experiment was conducted over two years in a randomized block design in order to evaluate attractiveness of preselected plant species to pollinators, as well as their attractiveness efficiency (a combination of duration of flowering and insect visitation), their response to two different agronomic management practices (growing in mixed vs. mono-specific plots; tillage vs. no-tillage), and their potential weediness. The results of this experiment showed that the flowers of Borago officinalis, Echium plantagineum, Phacelia tanacetifolia and Diplotaxis tenuifolia were attractive to bees, while Calendula arvensis, Coriandrum sativum, D. tenuifolia and Lobularia maritima were attractive to hoverflies. In addition, floral mixture resulted in lower attractiveness efficiency to pollinators than mono-specific D. tenuifolia, but higher than most of the mono-specific stands. On the other hand, although some of the most attractive plant species (e.g. P. tanacetifolia and C. arvensis) showed potential weediness, their self-seeding was reduced by tillage. After comparing attractiveness efficiency of various herbaceous species to pollinators and their potential weediness, the results indicated that D. tenuifolia showed the highest attractiveness efficiency to pollinators and efficient self- reproduction, making it highly recommended to attract bees and hoverflies in agro- ecosystems of Central Spain. In addition, this plant, commonly known as wild rocket, has a supplementary economic value as a commercialized crop.

The implementation of a new floral margin in agro-ecosystems means increased production costs, especially in regions with frequent and long droughts (as it is Central and South-East area of Iberian Peninsula), where the principal agricultural cost is irrigation. Therefore, before recommending D. tenuifolia for sustainable habitat management within agro-ecosystems, it is necessary to study the effect of drought stress and moderate and severe deficit irrigation on its growth, flower development and attractiveness to pollinators. The results of this experiment showed that in greenhouse conditions, potted D. tenuifolia could be without irrigation for 4 days without affecting its growth, flowering and attractiveness to pollinators. However, lack of irrigation for 8 days or longer significantly reduced the vegetative growth, number of open flowers, total floral area, flower diameter, corolla tube diameter and corolla tube length of D. tenuifolia. This study showed that regulated deficit irrigation can improve water use efficiency, and depending on the purpose of growing D. tenuifolia, as a crop

Abstract or as a beneficial plant to attract pollinators, it can reduce water consumption by 40% to 70% without affecting its vegetative and floral development and without reducing its attractiveness to pollinators.

Finally, the following experiment was developed in order to understand how habitat management can influence on the agricultural production. For this purpose, it was evaluated if the vicinity of mixed and mono-specific field margins, preselected to conserve pollinators within agro-ecosystems, can improve seed production in coriander (C. sativum). The selection of this plant species for the experiment was based on its necessity for insect pollination for production of seeds (even though some pollen can be transmitted from one flower to another by wind) and the fact that under semi-controlled field conditions established in the field it is possible to estimate its total seed production. Since D. tenuifolia is attractive for both bees and hoverflies in Central Spain, the main objective of this experiment was to estimate the impact of two different types of field margins, i.e. mono-specific margin with D. tenuifolia and mixed margin with six herbaceous species, on the seed production of potted coriander. For that reason, it was tested: i) if open pollination (control without proximate field margin and treatments with nearby mono-specific and mixed margin) increases the seed production of coriander when compared with no-pollination (covered inflorescences of coriander) under field conditions; ii) if frequency of pollinator visitation to the flowers of coriander was higher in the presence of field margins than in the control without field margin; and iii) if seed production was higher in the presence of field margins than in control plants of coriander without field margin. The results showed that the proximity of both types of floral margins (mixed and mono-specific) improved the seed quality of coriander plants, as seed weight and germination rate were higher than in control plants without field margin. Furthermore, the number of seeds produced was significantly higher in coriander plants grown near mixed margins than near mono-specific margin, probably due to an increase in pollinator visits. Since the experiment was conducted under semi- controlled field conditions, it can be concluded that pollinator visits was the main factor that biased the results, and that presence of both mixed or mono-specific (D. tenuifolia) margins can improve the production of coriander for more than 200% in small-scale gardens and, in addition, conserve the local pollinators.

iii

Resumen

La gestión de hábitat orientada a la conservación de polinizadores en los agro- ecosistemas requiere una selección de especies vegetales atendiendo fundamentalmente a dos criterios: i) el potencial atractivo de sus flores a los polinizadores; y ii) la simplicidad en su manejo agronómico. Además de estas premisas, es necesario considerar la capacidad invasora de estas especies vegetales, debido a que algunas de las más atractivas pueden resultar invasoras en determinados agro-ecosistemas. Por lo tanto, es preciso determinar qué especies vegetales son las más indicadas para ser implementadas en cada agro-ecosistema. En la presente tesis doctoral se plantea la búsqueda de las especies vegetales adecuadas para atraer polinizadores en los agro- ecosistemas del centro de España.

En una primera aproximación, se ha evaluado la atracción y expansión espacial (potencial invasivo) de seis plantas perennes de la familia Lamiaceae (aromáticas), elegidas por ser nativas de la región mediterránea. La elección de las especies vegetales se ha llevado a cabo con el fin de crear márgenes funcionales basados en la mezcla de especies vegetales con distintos periodos de floración, de modo que prolonguen la disponibilidad de recursos florales en el tiempo. Tras un primer año dedicado al establecimiento de las especies aromáticas, en los dos años siguientes se ha estudiado la atracción individual y combinada de las especies vegetales sobre los polinizadores, y como ésta se ve afectada por la densidad y la morfología floral, utilizando para ello un diseño experimental en bloques al azar. Los resultados de este estudio han puesto de manifiesto que la morfología floral no tuvo influencia sobre la atracción de las especies vegetales, pero si la densidad floral, puesto que las especies vegetales con mayor densidad de flores (Nepeta tuberosa e Hyssopus officinalis) han mostrado mayor atracción a polinizadores. Cabe destacar que de las seis especies consideradas, dos especies de verano (Melissa officinalis y Thymbra capitata) no han contribuido de forma efectiva a la atracción de la mezcla hacia los polinizadores, mostrando una reducción significativa de este parámetro respecto a las otras especies aromáticas a lo largo del verano. Se ha observado que ninguna de las especies aromáticas evaluadas ha mostrado tendencia invasora a lo largo del estudio. En base a estos resultados, se puede concluir que entre las especies aromáticas estudiadas, N. tuberosa, H. officinalis y Salvia verbenaca son las que ofrecen mayor potencial para ser utilizadas en la conservación de polinizadores. iv

Resumen

De forma similar al caso de las plantas aromáticas, se ha llevado a cabo una segunda experimentación que incluía doce plantas anuales con floración de primavera, en la que se evaluó la atracción a polinizadores y su comportamiento agronómico. Este estudio con especies herbáceas se ha prolongado durante dos años, utilizando un diseño experimental de bloques aleatorios. Las variables analizadas fueron: el atractivo de las distintas especies vegetales a los polinizadores, su eficiencia de atracción (calculada como una combinación de la duración de la floración y las visitas de insectos), su respuesta a dos tipos de manejo agronómico (cultivo en mezcla frente a monocultivo; laboreo frente a no-laboreo) y su potencial invasivo. Los resultados de esta segunda experimentación han mostrado que las flores de Borago officinalis, Echium plantagineum, Phacelia tanacetifolia y Diplotaxis tenuifolia son atractivas a las abejas, mientras que las flores de Calendula arvensis, Coriandrum sativum, D. tenuifolia y Lobularia maritima son atractivas a los sírfidos. Con independencia del tipo de polinizadores atraídos por cada especie vegetal, se ha observado una mayor eficiencia de atracción en parcelas con monocultivo de D. tenuifolia respecto a las parcelas donde se cultivó una mezcla de especies herbáceas, si bien en estas últimas se observó mayor eficiencia de atracción que en la mayoría de parcelas mono-específicas. Respecto al potencial invasivo de las especies herbáceas, a pesar de que algunas de las más atractivas a polinizadores (P. tanacetifolia and C. arvensis) mostraron tendencia a un comportamiento invasor, su capacidad de auto- reproducción se vio reducida con el laboreo. En resumen, D. tenuifolia es la única especie que presentó una alta eficiencia de atracción a distintos tipos de polinizadores, conjuntamente con una alta capacidad de auto-reproducción pero sin mostrar carácter invasor. Comparando el atractivo de las especies vegetales utilizadas en este estudio sobre los polinizadores, D. tenuifolia es la especie más recomendable para su cultivo orientado a la atracción de polinizadores en agro-ecosistemas en el centro de España. Esta especie herbácea, conocida como rúcula, tiene la ventaja añadida de ser una especie comercializada para el consumo humano.

Además de su atractivo a polinizadores, deben considerarse otros aspectos relacionados con la fisiología y el comportamiento de esta especie vegetal en los agro- ecosistemas antes de recomendar su cultivo. Dado que el cultivo en un campo agrícola de una nueva especie vegetal implica unos costes de producción, por ejemplo debidos a la utilización de agua de riego, es necesario evaluar el incremento en dichos costes en

Resumen función de demanda hídrica específica de esa especie vegetal. Esta variable es especialmente importante en zonas dónde se presentan sequías recurrentes como es el caso del centro y sur-este de la península Ibérica. Este razonamiento ha motivado un estudio sobre los efectos del estrés hídrico por sequía y el estrés por déficit moderado y severo de riego sobre el crecimiento y floración de la especie D. tenuifolia, así como sobre la atracción a polinizadores. Los resultados muestran que tanto el crecimiento y floración de D. tenuifolia como su atracción a polinizadores no se ven afectados si la falta de riego se produce durante un máximo de 4 días. Sin embargo, si la falta de riego se extiende a lo largo de 8 días o más, se observa una reducción significativa en el crecimiento vegetativo, el número de flores abiertas, el área total y el diámetro de dichas flores, así como en el diámetro y longitud del tubo de la corola. Por otro lado, el estudio pone de manifiesto que un déficit hídrico regulado permite una gestión eficiente del agua, la cual, dependiendo del objetivo final del cultivo de D. tenuifolia (para consumo o solo para atracción de polinizadores), puede reducir su consumo entre un 40 y un 70% sin afectar al crecimiento vegetativo y desarrollo floral, y sin reducir significativamente el atractivo a los polinizadores.

Finalmente, esta tesis aborda un estudio para determinar cómo afecta el manejo de hábitat a la producción de los cultivos. En concreto, se ha planteado una experimentación que incluye márgenes mono-específicos y márgenes con una mezcla de especies atractivas a polinizadores, con el fin de determinar su efecto sobre la producción del cultivo de cilantro (C. sativum). La elección del cultivo de cilantro se debe a que requiere la polinización de insectos para su reproducción (aunque, en menor medida, puede polinizarse también por el viento), además de la facilidad para estimar su producción en condiciones semi-controladas de campo. El diseño experimental consistía en la siembra de márgenes mono-específicos de D. tenuifolia y márgenes con mezcla de seis especies anuales situados junto al cultivo de cilantro. Estos cultivos con márgenes florales fueron comparados con controles sin margen floral. Además, un segundo grupo de plantas de cilantro situadas junto a todos los tratamientos, cuyas flores fueron cubiertas para evitar su polinización, sirvió como control para evaluar la influencia de los polinizadores en la producción del cultivo. Los resultados muestran que la presencia de cualquiera de los dos tipos de margen floral mejora el peso y el porcentaje de germinación de las semillas de cilantro frente al control sin margen. Si se comparan los dos tipos de margen, se ha observado un mayor número de semillas de cilantro junto al

Resumen margen con mezcla de especies florales respecto al margen mono-específico, probablemente debido al mayor número visitas de polinizadores. Puesto que el experimento se realizó en condiciones de campo semi-controladas, esto sugiere que las visitas de polinizadores fueron el factor determinante en los resultados. Por otro lado, los resultados apuntan a que la presencia de un margen floral (ya sea mono-especifico o de mezcla) en cultivos de pequeña escala puede aumentar la producción de cilantro en más de un 200%, al tiempo que contribuyen a la conservación de los polinizadores.

vii

Chapter I

I. General introduction

1. Agriculture and pollination

Agriculture, defined as cultivation of plants, animals and other forms of life for food, biofuel, medicinal and other products used to sustain and enhance human life (ILC, 2000), is one of the oldest and greatest modifications of the ecosystems. Through the history, a constant increase in human population have caused a gradual switching from small-scale farming to an application of intensive agricultural practices where usage of modern technologies, i.e. tillage, pesticides and fertilizers, have resulted in massive biodiversity simplification and a reduction of ecosystem services (Altieri, 1999; Benton et al., 2003; Watling and Donnelly, 2006; Gill et al., 2012). Nowadays, agricultural land occupies around 45% of total Earth land area (New, 2005), irregularly distributed worldwide (Fig. I-1) and therefore it should be managed in the most sustainable way.

Fig. I-1. Percentage of agricultural land relative to the total land area across the Earth Source: World Bank, 2013

Chapter I

Pollination, i.e. transfer of pollen from anther to stigma, is one of the most important ecosystem services in agriculture as it enables reproduction of plants (Bastian, 2013). There are two types of pollination, abiotic and biotic. Abiotic pollination takes place without the involvement of living organisms, for example, where pollen is transported by wind (anemophily), while biotic pollination is the result of the movement of pollen by living organisms and it is the most common form of pollination. The most diverse and abundant group of pollinators are insects (Steffan-Dewenter et al., 2005), hence in the following text the term “pollinators” will be used for insect pollinators restrictively.

Focusing on insects, in one natural and balanced ecosystem, it is assumed that all insects contribute to maintaining the system and therefore no species can be considered to have positive or negative impact on the ecosystem. On the other side, in agriculture, where the goal is to raise selected crops, insects that hinder production are classified as pests, while insects that assist production are considered beneficial. Along with natural enemies of pests, pollinators are the most important beneficial insects for agricultural production (Carvalheiro et al., 2012) and wild plant reproduction (Biesmeijer et al., 2006), as approximately 75 % of crop species of global significance for food production and 65 % of plant species rely on pollinators (Klein et al., 2007).

Recently, interactions between pollinators and agriculture have generated considerable research interest, stimulated by rising recognition of the significance of insect pollination in maintaining yields of some crops (Santos et al., 2008; Carvalheiro et al., 2012; Bartomeus et al., 2014) and coupled with the concern that many species of pollinating insects have declined substantially (Biesmeijer et al., 2006). In the face of global decline in biodiversity, the species interactions deserve particular attention. As pollinators are strongly dependent on floral diversity due to mutual specialization, declines in plant diversity (caused by land use intensification), may be associated with extinction of pollinators (Frund et al., 2010). Although intensive agriculture is mainly responsible for pollinator decline (Tscharntke et al., 2005; Rundlof et al., 2008; Le Feon et al., 2010), there are other factors that negatively affect diversity and abundance of pollinators, as the potential spread of diseases (Vanbergen and Insect Pollinators Initiative, 2013) and changes in climate (Memmott et al., 2010; Rader et al., 2013). Furthermore, various scientific proofs have been published recently showing the

Chapter I negative effect of neonicotinoid insecticides on development and health of pollinators (Gill et al., 2012; Stokstad, 2013; van der Sluijs et al., 2013).

Consequently, the decline in abundance and diversity of pollinators has become a subject of public and political interest. Therefore, to promote farming in a way that it supports biodiversity, many regions have introduced agri-environment schemes (AES), including Australia, the European Union, USA and Switzerland (Morris and Potter, 1995; Albrecht et al., 2007; Batary et al., 2011; Elts and Lohmus, 2012). In that way, conservation of semi-natural habitats has been supported economically by AES, compensating the farmers’ income losses associated with employing environmentally sustainable land (Dobbs and Pretty, 2004). In addition, in 2013 the European Union decided to suspend the use of a number of neonicotinoids, which has boosted further research on pollinator exposure to insecticides and the potential for sublethal effects (Sandrock et al., 2014).

2. Plant-pollinator interactions

Recently, parallel diversity decline of bees and insect-pollinated plants has been reported in Europe (Biesmeijer et al., 2006), suggesting a functional coupling and mutual dependence between both guilds. This diversity correlation may be driven by an influence of pollinator diversity on the seed set of plants (Klein et al., 2003a; Fontaine et al., 2006) and in turn, as pollinators are strongly dependent on floral diversity due to mutual specialization, declines in plant diversity, e.g. caused by land use intensification, may be associated with linked extinctions of pollinators (Frund et al., 2010).

Although the general extent of pollinator specialization is still poorly known many findings highlight the complexity of plant-pollinator interactions and confirm the importance of flower diversity for bee and hoverfly communities (Ambrosino et al., 2006; Bosch et al., 2009). Since floral density and diversity are one of the most important factors for conservation of pollinators (Scriven et al., 2013), importance of a positive correlation between richness of flower species and pollinators has been reported from several countries and habitats (Steffan-Dewenter and Tscharntke, 2000; Frund et al., 2010; Nicholls and Altieri, 2013). For that reason, exposure of diverse flowers to pollinators is a way of maintaining and promoting genetic diversity in crops and other plants, and studies focusing on the behaviour of floral visitors are an 3

Chapter I important tool to identify patterns of ecological interactions between visitors and flowers. The majority of such ecological interactions are mutualistic, whereby plant species offer floral resources and obtain pollination benefits. However, some relations are unilateral, where plants can be used for other purposes such as for prey capture, mating, and resting (Comba et al., 1999). One of interesting ecological relations are those that involve the pillage of plant resources by floral visitors, i.e. the use of resources by illegitimate visitation incapable of promote pollination (Inouye, 1980). For those reasons it is important to know which insects are attracted to which floral species, and also which insects are efficient in pollination.

3. Important groups of pollinators

The most recognized pollinators are various species of bees (Hymenoptera: Apoidea), with their body structure plainly adapted to pollinate (Michener, 2007). Other groups of insects important in pollination are: hoverflies (Diptera: Syrphidae); ants and wasps (Hymenoptera); moths and butterflies (Lepidoptera) and some beetles (Coleoptera).

This study has been mainly focused only on bees and hoverflies as those two groups of pollinators were the most abundant and diverse in the study area. In addition, several species of beetles were included in the study as they were highly abundant visitors to some of the studied plant species. For these reasons, taxonomy and biology of these three groups will be described, while other groups of pollinators will be briefly introduced.

3.1. Bees (Hymenoptera: Apoidea)

Bees (Hymenoptera: Apoidea) are the most important pollinating insects as they depend on flower nectar and pollen. Nectar is the main energy source for adults, while pollen is a protein source, collected by adult females primarily to feed their larvae (Michener, 2007). However, they are important pollinators because some of the collected pollen that they inevitably lose in going from flower to flower falls on the pistils (reproductive structures) of other flowers of the same species, resulting in cross- pollination.

Chapter I

The bees have been classified by different authors following different criteria, but in this research the classification proposed by Michener (2007) has been used, where seven families of bees were set: Andrenidae, Colletidae, Halictidae, Melittidae, Stenotritidae, Megachilidae and Apidae. Two of those families (Melittidae, Stenotritidae) are small families, mainly distributed in Africa and Australia, while other five families (Apidae, Halictidae, Colletidae, Andrenidae and Megachilidae) are cosmopolitan families and numerous species of those families are present in Spain (Fig. I-2).

Fig. I-2. Different groups of pollinators [1. Honey bee (Apidae: Apis sp.); 2. Bumblebee (Apidae: Bombus sp.); 3. Carpenter bee (Apidae: Xylocopa sp.); 4. Digger bee (Apidae: Amegilla sp.); 5. Leafcutter bee (Megachilidae: Megachile sp.); 6. Mining bee (Andrenidae: Andrena sp.; 7. Sweat bee (Halictidae: Halictus sp.; 8. Hoverfly (Syrphidae: Sphaerophoria scripta.); 9. Beetle (Coleoptera: Heliotaurus ruficollis)]

Nevertheless, in pollination ecology it is frequent to group bees in accordance with their morpho-physiological features to function as pollinators, such as body size

Chapter I

and method of pollen collection. Therefore, honey bees, bumblebees and wild bees are in general analyzed separately (Carrie et al., 2012; Miñarro and Prida, 2013; Bartomeus et al., 2014) and sometimes, depending on the scope of the study, wild bees can also be divided into different groups (Barbir et al., 2014a) easily distinguished in the field. In continuation, bee groups that are frequent in Central Spain will be briefly described.

Honey bees

Honey bees (Family: Apidae) represent only a small fraction of the approximately 20,000 known species of bees and are a subset in the genus Apis, primarily distinguished from other species by the production of honey and the construction of nests out of wax. Currently, there are only seven recognized species of honey bee with a total of 44 subspecies (Engel, 1999).

Honey bees (as well as bumble bees) are social insects, with a firmly specialized division of labour (Johnson, 2010). Besides the important role in producing honey, honey bees have the important ecological and economical role of pollination, for which their hind legs are modified into a structure called corbicula, widely known as the "pollen basket" (Fig. I-2.1). Economically, for all USA agriculture, the marginal increase in the value attributable to honey bees—that is, the value of the increased yield and quality achieved through pollination by honey bees alone—was $9.3 billion in 1989 © P and $14.6 billion in 2000 (Morse and Calderone, 2000).

Bumblebees

Another important genus from the family Apidae in pollination of plants is Bombus (bumblebees). There are over 250 known species, existing primarily in the Northern Hemisphere. Bumblebees are, like honey bees, social insects and are mostly characterized by black and yellow body (Fig. I-2.2) while in some species it might be orange, red or entirely black (Michener, 2007).

In Spain there are just twelve species of bumblebees of which the most abundant is Bombus terrestris L. (Rasmont and Iserbyt, 2010). However, in other regions, e.g. UK or USA, bumblebees are rather diverse and abundant, but this diversity is declining rapidly and therefore conservation of bumblebees have received a lot of attention © 6

Chapter I recently (Carvell et al., 2007; Heard et al., 2007). In general, bumblebees are good pollinators due to their hairy body, but some species of bumblebee also exhibit what is known as "nectar robbing", i.e. instead of inserting the mouthparts into the flower, they bite directly through the base of the corolla to extract nectar, avoiding pollen transfer (Maloof, 2001).

Solitary bees

Solitary bees (frequently called wild bees) are important pollinators, because of their advanced types of pollen-carrying structures, i.e. the scopa (made up of thick, plumose setae), on the lower abdomen (Fam. Megachilidae) or on the hind legs (other families of wild bees).

Solitary bees are often oligoleges (or simply specialist pollinators), gathering pollen from one or a few species/genera of plants (unlike honey bees and bumblebees which are generalists). Due to their high specialization, some solitary bee species are endangered, simultaneously with the plant species they are associated with (Biesmeijer et al., 2006). In addition, solitary bees create nests in hollow reeds, holes in wood, or most commonly, in tunnels in the soil, which are frequently destroyed by agricultural tillage (Nicholls and Altieri, 2013).

Although solitary bees are important pollinators, a very few species of solitary bees are reared for commercial pollination, e.g. Osmia ribifloris Cockerell, because of their difficult management (providing nest sites). Instead, especially in organic gardening, it became widely popular to provide nest boxes for solitary bees and in that way improve the conservation of the native solitary bee species. In the landscape level, conservation of pollinators mainly takes place within agro-ecosystems by managing the habitat and offering additional floral sources to forage, with final aim to maintain pollinators close to agricultural fields (New, 2005).

Among the solitary bees different groups can be found, such as:

Carpenter bees (Fam. Apidae: Xylocopini) are large bees which make their nests by tunnelling into wood. These bees are important pollinators of open-faced flowers, although some species are known to rob nectar by slitting the sides of flowers with deep 7

Chapter I

corollas (Zhang et al., 2007). One of the most abundant species of this group in Central Spain is Xylocopa violacea L. (see Xylocopa sp. In Fig. I-2.3).

Digger bees (Fam. Apidae: Anthophorini) are very buzzy solitary bees which dig their nests in the soil. This group consists of over 750 species with different shapes and sizes, but the majority of bees are large (up to 3 cm), hairy, fast-flying and robust (Michener, 2007). The most abundant genera in Spain are: Amegilla spp. (Fig. I-2.4), Antophora spp. and Habropoda spp.

Leafcutter bees (Fam. Megachilidae) (Fig. I-2.5) are a cosmopolitan family of solitary bees whose pollen-carrying structure is restricted to the ventral surface of the abdomen and because of this they are easy to recognize in the field (Michener, 2007). They also produce characteristic buzzy sound when flying. As they nest in the wood, it is easier to commercially produce some of the species and to use them for pollination in greenhouses (O. ribifloris). Some of the frequent species in Central Spain are: Chelostoma florisomnis, Chalicodoma pyrenaica, Osmia spp., Anthidium spp.

Mining bees (Fam. Andrenidae) is a large non-parasitic bee family, with most of the diversity in temperate and arid areas (warm temperate xeric). Therefore, this group of bees is very diverse and abundant in the Mediterranean region. They are typically small to moderate-sized bees and often have scopa on the basal segments of their legs in addition to the tibiae. They can be separated from other bee families by the presence of two subantennal sutures on the face (Michener, 2007). The most frequent genus in Central Spain is Andrena with many different species (Fig. I-2.6). © J l Sweat bees (Fam. Halictidae) is a cosmopolitan family consisting mainly of small to midsize bees (up to 10 mm), although few species can be bigger. Their hives are usually found inside the hollow trunk or wood or underground. Several species are completely or partly green and a few are red; a number of them have yellow markings, especially the males, which commonly possess yellow faces, a pattern widespread among the various families of bees (Engel, 2000). Two genera are the most abundant in Spain: Halictus spp. (Fig. I-2.7) and Lasioglossum spp.

Chapter I

3.2. Hoverflies (Diptera: Syrphidae)

Hoverflies, also known as syrphid flies, are classified in the family Syrphidae (Diptera), which consists of about 6,000 species. Hoverflies are an extremely variable family of flies that range from the large, bulky and hairy to the small, slender, and shiny adults which are easily recognized by their ability to hold a seemingly motionless position in the air (Stubbs and Falk, 2002). The genera Sphaerophoria (Fig. I-2.8), Eristalis, Episyrphus, Syrphus and Eupeodes are some of the most abundant genus of the family Syrphidae in Central Spain (Pineda and Marcos-Garcia, 2008).

Many adult hoverflies survive on nectar and pollen, but the larvae eat a much more varied diet that may include other insects or decaying plants and animals. For that reason, some species of hoverfly are highly prized as biological control agents because the larval forms of those flies feed on aphids (Chambers and Adams, 1986; Ambrosino et al., 2006; Smith and Chaney, 2007). Therefore, it has been recommended to plant flowers that produce pollen and nectar that adult hoverflies use as the energy to produce ©Jelena large numbers of eggs all of which turn into aphid-munching larvae (Stubbs and Falk,

2002).

3.3. Beetles

Beetles (Coleoptera) are generally viewed as poor pollinators (Bartomeus et al., 2008), although in Mediterranean ecosystems they are very abundant floral visitors (Bernhardt, 2000) (Fig. I-2.9). On the other side, beetles spend long periods of time on each flower they visit, and therefore they visit fewer flowers than bees (Bosch, 1992). In addition, some beetles can occupy or destroy flowers, not allowing other pollinators to feed on the flowers (Barbir et al., 2014a). In this study, only the most abundant flower visitors have been included.

3.4. Other pollinator groups

Butterflies and moths are the second largest order of insects (Lepidoptera) however, they are not major pollinators. Butterflies and moths tend to sit at the edges of flowers, and extend their long tongues (probosces) to reach nectar (Subba Reddi and Meera Bai, 1984). They rarely come in contact with the pollen at the centres of the 9

Chapter I flowers. Nevertheless, some moths are effective pollinators of deep-throated flowers that require the insects to crawl inside to reach the nectar.

In addition, ants and wasps (Hymenoptera) are also insects that frequently visit flowers and contribute to their pollination, but they are not as efficient pollinators as bees (Philpott et al., 2006).

4. Managing semi-natural habitats of pollinators within agro- ecosystems

Since agricultural practices can damage biodiversity through various pathways, agriculture is frequently considered anathema to conservation. However, appropriate management can ameliorate many of the negative impacts of agriculture, while largely maintaining provisioning ecosystem services (Power, 2010). Although honey bee has high economic importance in honey production and it is the most abundant cosmopolitan pollinator, it has been discovered that there are various crops for which honey bees are poor pollinators compared to wild bees (Michener, 2007). For that reason, national and international organizations are publicizing the need to conserve native pollinators (Michener, 2007).

Although wild pollinators are better conserved within protected areas and natural parks, where floral diversity and shelters are available, for agricultural production it is more convenient to conserve pollinators within agricultural fields. Conventional farmers routinely manage their fields for greater provisioning services by using inputs and practices to increase yields, but management practices can also enhance other ecosystem services, such as pollination, biological control, soil fertility and structure, and support biodiversity (Power, 2010). Therefore, if introduced properly, habitat management within the agro-ecosystem can provide resources needed for conservation of pollinators and natural enemies (Tscharntke et al., 2005).

Large monocultures of bee-pollinated crops, i.e. almond, melon, or canola, normally provide abundant food sources for pollinators, but only for a few weeks period. Therefore, lack of within field or adjacent wild plants, blooming before and after the main crops bloom, can result in a decline of healthy pollinators abundance (Goulson, 2003). Recently, implementation of floral margins in agricultural landscapes

Chapter I has been shown to increase the abundance of pollinators in agro-ecosystems (Kells et al., 2001; Marshall, 2005). Naturally regenerated uncropped margins were primarily designed to conserve rare local arable flora and their associated fauna (Kells et al., 2001). However, sowing an uncropped margin with wildflower seed mixture can offer better food sources to pollinators and improve their conservation (Pywell et al., 2011).

Therefore, in order to maintain native plant species in each region for pollinator conservation, there is a need for pollination biologists to integrate their findings with those of plant restoration ecologists and to ensure sustainable restoration of pollinator habitats within agro-ecosystems by using native plants (Menz et al., 2011). For that reason, special attention should be focused on the selection of beneficial insectary plants and their management in the agro-ecosystems (Nicholls and Altieri, 2013).

4. 1. Beneficial insectary plants

Beneficial insectary plants are plants that attract beneficial insects (pollinators and natural enemies of pests) by producing nectar and pollen which are their main food sources (Colley and Luna, 2000). Although many pollinators are generalist, e.g. honey bee and the majority of hoverfly species (Branquart and Hemptinne, 2000; Frund et al., 2010), some pollinators have preferences when choosing the plants to visit (Frund et al., 2010).

Pigmentation in flowers seems to play a major role in pollination success (Miller et al., 2011) and can significantly influence the insects’ preference (Campbell et al., 2012b). In addition, colour can help beneficial insects with plant species selection, flower location and ripeness, and nectar and pollen location within the flower. Besides, shape and size of the flower (Gomez et al., 2008; Schiestl and Johnson, 2013), nectar availability and chemical composition of the plant species (Baker and Baker, 1983; Stout and Goulson, 2002) also contribute to the preferences of pollinators.

Some plant families have been widely known as highly attractive to beneficial insects, for instance Boraginaceae, Brassicaceae and Apiaceae, and many plant species from those families have been introduced within agricultural fields worldwide. Plants as buckwheat (Fagopyrum esculentum Moench) (Fig. I-3.1), phacelia (Phacelia tanacetifolia Benth.) (Fig. I-3.2), alyssum (Lobularia maritima L. Desv.) (Fig. I-3.3)

Chapter I and coriander (Coriandrum sativum L.) (Fig. I-3.4), have been used in beneficial insect conservation practices in the last 10 years (Landis et al., 2000; Pfiffner and Wyss, 2004), where their attractiveness to beneficial insects have been highly ranked on the basis of feeding visit frequencies (Ambrosino et al., 2006).

1 2

3 4

Fig. I-3. Beneficial insectary plants [1. Buckwheat (Fagopyrum esculentum Moench); 2. Phacelia (Phacelia tanacetifolia Benth.); 3. Alyssum (Lobularia maritima L. Desv.); 4. Coriander (Coriandrum sativum L.)]

While it is important to consider all the factors mentioned above when choosing plants to attract beneficial insects, it is also necessary to consider the ecology, physiology, floral phenology, and potential weediness of the plants prior introducing them in agro-ecosystems (Nicholls and Altieri, 2013), with the final aim to apply the most appropriate and sustainable habitat management for conservation of pollinators.

Chapter I

4. 2. Management of beneficial insectary plants within agro-ecosystems

Although some plants can be widely attractive to cosmopolitan pollinators, other plants have flowers that are more attractive to specific groups of pollinators, e.g. open flowers are attractive to hoverflies and plants from Boraginaceae family are very attractive to bumblebees. Therefore, it is important to have fundamental knowledge about the abundance and diversity of groups of pollinators in the region before choosing plant species for efficient management of pollinator habitats within the landscapes.

In addition, to manage beneficial insectary plants within agro-ecosystems it is essential to consider the climate conditions of the region (Memmott et al., 2010). In the Mediterranean region, where precipitation and irrigation water are scarce, usage of plant species that can cope with water stress are preferred, while in other regions, i.e. tropical, Central and Northern Europe, irrigation is not a limiting factor. Under different environmental conditions, beneficial insectary plants easily managed in one region can show weediness or invasiveness in another region. Therefore, it is recommendable to test under field conditions the agricultural behaviour and the attractiveness to pollinators of introduced or native beneficial plants (Comba et al., 1999; Corbet et al., 2001) before recommending its usage in specific agro-ecosystems.

Furthermore, beneficial insectary plants can be managed in different ways within agro-ecosystems depending on the scale of the field, climate and agricultural practices of specific region (Altieri, 1991). One of the most common approaches to implement beneficial plants within large agricultural fields is to sow field margins. However, some plant species are more efficient if sown as mono-specific margins (Fig. I-4.1), while others are more efficient if mixed with other plant species (Fig. I-4.2) (Pontin et al., 2006; Barbir et al., 2014a).

Chapter I

Fig. I-4. Sowed field margins: 1. Mono-specific field margin (Phacelia tanacetifolia Benth.), 2. Mixed field margin

In addition, sustainable and organic agriculture have recently received a lot of attention as strategies that will stop or at least reduce the damage to ecosystems (Andersson et al., 2012; Crowder et al., 2012) and improve conservation of beneficial insects (Power and Stout, 2011; Gill et al., 2012). In organic farming, beneficial insectary plants are widely used for their positive impact in conservation of natural enemies of pests and pollinators (Power and Stout, 2011). Since the commercialization of organic products is becoming increasingly widespread, this implies an unquestionable opportunity for organic production on a small-scale, where the use of beneficial insectary plants should be a key part of the production system (Fig. I-5) (Markussen et al., 2014).

©OrganicGarden ©The ClimateCompliannce Conference

Fig. I-5: Use of beneficial insectary plants in small-scale gardening 14

Chapter II

II. Research question and objectives

There is a need to improve habitat management to promote the conservation of pollinators in agro-ecosystems from Central Spain. Considering the specific characteristics in term of climate, native flora, agricultural practices, pollinator species and their habitats within different regions worldwide, it is important to find the most appropriate beneficial plants in a specific region and to manage them in the most suitable way. For this, the following research question should be posed:

How to manage the habitat of pollinators in order to conserve them within agro-ecosystems of Central Spain?

In order to answer the research question and to conduct the experiments systematically, several specific objectives have been proposed:

Objective 1

to improve the functionality of a mixture of aromatic plants (Fam. Lamiaceae) in order to conserve pollinators (Chapter III)

Objective 2

to assess the preferences of pollinators to different species of annual herbaceous plants in order to use them for their conservation in agro-ecosystems of Central Spain (Chapter IV)

Objective 3

to evaluate the agronomic behaviour, i.e. potential weediness and self-seeding, of different aromatic and herbaceous plant species based on their biological, physiological and morphological characteristics (Chapter III and IV)

Chapter II

Objective 4

to study the effects of drought and regulated deficit irrigation (RDI) on growth and potential attractiveness of the plant selected as the most appropriate to conserve pollinators in Central Spain (Chapter V)

Objective 5

to test if the proximity of mono-specific or mixed field margins can improve pollination in the neighbouring crop (Chapter VI)

CHAPTER III

Submitted to: Insect Conservation and Diversity Journal (Submitted: 06/06/2014)

Barbir, J., Azpiazu, C., Badenes-Peréz, F.R., Fernández-Quintanilla, C., Dorado, J. (submitted). Functionality of aromatic (Lamiaceae) floral mixture in conservation of pollinators in Central Spain, Insect Conservation and Diversity

Chapter III

III. Functionality of aromatic (Lamiaceae) floral mixture in conservation of pollinators in Central Spain

1. Introduction

In the Mediterranean region, medicinal and aromatic plants (in further text referred only to family Lamiaceae) are broadly distributed in agro-ecosystems as both cultivated and spontaneous (Carrubba and Scalenghe, 2012). Besides being commercially used by food, perfume and pharmaceutical industries (Hussain et al., 2011; Zihlif et al., 2013), those plants have good potential as a forage source for pollinators and other beneficial insects (Song et al., 2010; Macukanovic-Jocic et al., 2011). As such, native aromatic species could play a significant role in the enhancement of multi-functionality of agriculture (Carrubba and Scalenghe, 2012), since environmentally friendly farming practices are mandatory in the European Union. Furthermore, use of aromatic plants could improve the conservation of beneficial insects and their natural habitat within agro-ecosystems.

Agricultural landscapes, dominated by large monocultures, have caused alarming declines in farmland biodiversity, threatening the optimal functionality of agro- ecosystems (Krebs et al., 1999; Benton et al., 2003). As reported by Landis et al. (2000) the presence of wild plants in crop ecosystems enhances the survivorship of beneficial insects responsible for ecosystem services such as pollination and biological control of pests in both agricultural and natural habitats (Klein et al., 2007). Thus, the recent report of parallel decline of pollinators and insect-pollinated plants (Biesmeijer et al., 2006) encouraged further investigations in plant-pollinator interactions (Campbell et al., 2012a; Carrie et al., 2012; Wratten et al., 2012) and habitat management (Cranmer et al., 2012; Holzschuh et al., 2012; Wratten et al., 2012) to enhance the presence of beneficial insects within agro-ecosystems.

In the intensively managed agro-ecosystems, insect species tend to suffer from deficit of nectar and pollen sources, shelter and nesting sites (Cane, 2008; Pywell et al., 2011; Carvalheiro et al., 2012). Therefore, some recent studies have focused on improving the habitat of beneficial insects within agricultural landscapes by sowing field margins with a mixture of wild flowers (Marshall, 2005; Carvell et al., 2007;

Chapter III

Olson and Wackers, 2007). For that purpose, some studies recommend plant species that are attractive to pollinators (Pontin et al., 2006), while others suggest the usage of native or naturalized plants in the area (Tuell et al., 2008; Morandin and Kremen, 2013). In addition, the importance of adjusting the floral phenologies of the plants used in the field margins has been emphasized, by combining their different bloom periods and providing pollinators with constant floral resources available from early spring to early autumn (Memmott et al., 2010). Although the attractiveness to pollinators of many aromatic species has been combined with their floral phenology and morphology (Herrera, 2001; Keasar et al., 2006), only few studies have simultaneously compared the attractiveness of various aromatic plant species under the same field conditions (Petanidou and Vokou, 1993; Macukanovic-Jocic et al., 2011). Some studies have emphasized that the beneficial plants should not be invasive towards neighbouring crop fields (Stout and Morales, 2009), but in the Mediterranean region little is known about the behaviour of beneficial plants in floral mixtures as well as their competitiveness towards neighbouring crops.

Considering that many native aromatic species are naturally present in the landscapes, hedgerows and margins of crop fields, there is a need to compare their attractiveness to pollinators and to study their role in floral mixtures in order to choose the most efficient plants for functional field margins. Therefore, it was hypothesized that mixture of plants attractive to pollinators and with different floral phonologies, can create a functional, long-flowering floral margin which will offer food for pollinators during their active season. Consequently, six plant species from the Lamiaceae family, with different floral phenologies, frequent and native in the Mediterranean region, were selected for this study by following three main objectives: i) to compare their attractiveness to pollinators, ii) to investigate whether floral morphology affects the attractiveness; and iii) to evaluate the functionality of the aromatic plants grown in mixture by taking into account different phenologies and attractiveness to pollinator. Finally, as the purpose of the studied plants is to be implemented within agro- ecosystems, their spatial expansion, i.e. potential weediness, was estimated under field conditions.

Chapter III

2. Materials and methods

2.1. Study area and experimental design

This study has been conducted over three years at the experimental farm La Poveda (40º 19' N and 3º 29' W, elevation 536 MASL), Arganda del Rey, Madrid, Spain. The study area is characterized by a continental Mediterranean climate with cold winters and hot summers and scant precipitation (about 400 mm per year). The meteorological data for both years of study was obtained from the meteorological station located at a distance of 1 km from the experiment, available online at Agro- climatic Information System for Irrigation (SIAR, 2014).

In February 2011, three-month-old seedlings of six aromatic species (Lamiaceae) were organised in randomized block design experiment (24.0 m × 15.2 m) within a barley field, with the longer side orientated NE-SW. The experiment was divided into three blocks (repetitions) each consisting of eight plots (2.4 m × 2.4 m) including six mono-specific plots and two plots (M1 and M2) with a mixture of these six species equally proportioned and randomly distributed within a plot (Fig. III-1).

1 6 M2 4 1 5

5 4 6 1 3 M1

3 M1 M1 5 M2 4

M2 2 3 2 2 6

Fig. III-1. Experimental design (randomized block design - each number represents one Lamiaceae plant species (mono-specific plot), while M1 and M2 represent mixed plots)

Chapter III

Each plot consisted of 64 plants organized in 8 lines and 8 columns, with a separation of 0.3 m between them. Finally, in order to avoid the influence of the neighbouring species, a separation between plots was 1.2 m and between blocks 2.0 m.

2.2. Groups of pollinators

Insects were grouped in accordance with their morpho-physiological features to function as pollinators, such as body size and method of pollen collection (Barbir et al., 2014a). The study was focused on seven groups of insects easily distinguished in the field: i) honey bees (Apis mellifera L.), ii) leafcutter bees (Fam. Megachilidae), iii) digger bees (Fam. Apidae, tribe Anthophorini), iv) small solitary bees (Fam. Halictidae and Andrenidae, with body size≤ 1cm), v) big solitary bees (Fam. Halictida e and Andrenidae, with body size > 1cm), vi) hoverflies, and vii) beetles. Bumblebees, Bombus spp. (Apidae) and carpenter bees, Xylocopa violacea L. (Apidae) were not included in the final analysis as fewer than ten visits were recorded during the whole study.

2.3. Attractiveness of the plants to pollinators

In order to estimate the attractiveness of aromatic flowers to local pollinators, their visitation rate, i.e. the mean number of insect visits per minute (Campbell et al., 2012a), was compared between plant species and their mixtures during 2012 and 2013. Insect visits to the flowers (direct contact with the reproductive structures of the flower, i.e. stamens and pistils) were counted once per week (9 min/plot) within three marked areas (0.25 m × 0.25 m) randomly scattered within each plot. The observations were conducted by three observers throughout the flowering period of each species, always in the morning (9.00 h – 13.00 h local time) when the weather conditions were preferable for the insects to fly and visit flowers (sunny days without wind). Switching plots, each observer counted visits of insects for three minutes in each plot.

2.4. Floral assessment

As one of the objectives of the study was to create a functional long flowering mixture of aromatic species, floral phenology was one of the main criteria for plant selection. For that purpose, six aromatic species (Fam. Lamiaceae), native and frequent

Chapter III in ecosystems of Central Spain, were selected for this study (Table III-1) in order to assess their potential contribution in floral mixtures.

Table III-1. Aromatic plant species used (taken from Castroviejo (1993))

Plant height Flowering Plant species Habitat Use (cm) period grasslands, cultivated fields, Salvia verbenaca L. 5 – 60 april -may medical roadsides, wastelands cultivated or sub- Salvia officinalis L. < 60 may -june food and medical spontaneous juniper and oak woods, thickets, ornamental and Nepeta tuberosa L. 30 – 100 may - june roadsides, meadows, medical grasslands nitrified, shady and food, medical and Melissa officinalis L. 45 – 83 june - july cool places perfume industry rocky slopes, food, medical and Thymbra capitata (L.) Cav. 10 – 40 june - july limestone, marl or perfume industry clay soils july - Hyssopus officinalis L. 15 – 52 thickets, roadsides medical octobre

In order to understand why some plant species can be more attractive to pollinators than others, morphological characteristics of flowers were assessed. Morphological characteristics of flowers were measured at the full bloom phenological stage in ten randomly chosen flowers, using a Vernier caliper. Immediately after cutting off the flower, certain morphological characteristics (flower length, flower height, corolla tube length, corolla height and calyx length) were measured, and the mean value of those measurements was used in further analysis.

The capacity of plants to produce flowers was assessed. For this, the number of flowers was counted in three frames of 0.25 m × 0.25 m within each plot once per week. In addition, each observer estimated the percentage of flower coverage in each plot, and the mean value of the three readings was calculated. Both parameters were used to estimate floral density as well as floral phenology of each plant species under the field conditions, with the final aim to estimate their influence on the attractiveness of the plants (Stout et al., 1998). Any plot was considered in bloom when it had more than five open flowers per observed area.

Chapter III

2.5. Spatial expansion of the plant species

In the early spring of 2014, the spatial expansion from the original area (i.e. the percentage of increased area) of each plant species, i.e. the three replicates of each mono-specific plots (2.40 m × 2.40 m) and the proportional part (i.e. one-sixth) of all the mixed plots, was calculated considering the area occupied the following year by using real time kinematic differential GPS (RTK-GPS). The percentage increase in area occupied by each plant species relative to their original area (i.e. potential weediness) was calculated using the software ArcGIS® 9.3 (ESRI, Redlands, California, USA).

2.6. Data analysis

The data from 2012 and 2013 were independently analysed with SPSS® Statistics19 software (IBM, Chicago, IL, USA). A simple general linear model (GLM) was used in order to assess the ability of different insects to discriminate between flower species. The mean value of insect visits per minute within the observation area (0.25 m × 0.25 m) was used as response variable for these analyses. Prior to ANOVA analysis, the assumptions of normal distribution (Kolmogorov Smirnov test) and homogeneity of variances (Levene’s test) were tested. Differences among means were explored by Tukey’s B post hoc test (P ≤ 0.05). In addition, univariate ANOVA was carried out to assess if there were significant differences between measured morphological characteristics of the flowers.

Discriminate Canonical Analysis (DCA) was performed in order to test if any of morphological characteristics influence on the discrimination between selected plants, and number of dimensions necessary to express this relationship. Finally, Categorical Partial Component Analysis (CATPCA) was used to assess the relationship between morphological characteristics and number of insect visits per minute. A bivariate two- tailed Pearson analysis was used to assess the correlation between floral density and visits of pollinators per minute.

Chapter III

3. Results

3.1. Attractiveness of the selected aromatic plants to pollinators

The attractiveness of the plants to pollinators was estimated by insect visitation rate (Table III-2). Nepeta tuberosa and Hyssopus officinalis, which had the highest visitation rate of pollinators in both years of the study, were characterized as the most attractive plant species. On the other hand, Salvia verbenaca, with no insect visits in 2012, and few beetles, digger bees and hoverflies visits in 2013, was characterized as the least attractive plant species (Table III-2).

Chapter III a abc b c b b b bc bcd bc ab bc b b b b 01 01 01 01 01 01 02 01 03 001 01 01 01 4 03 M2 ...... 01 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 04 01 01 02 04 03 08 06 12 02 01 07 01 03 31 23 ......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 bc b ab b b c d bc a ab bc b b b b 01 01 01 02 01 00 01 01 03 03 01 01 01 01 ...... 01 . . . M1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ± ± c ± ± ± ± ± ± ± ± ± ± ± ± ± 0 02 01 ± 03 03 02 05 05 09 03 01 05 01 03 20 23 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a b b ab a a a bc bc b a a a a ab 01 01 02 03 04 09 03 02 02 02 03 01 12 08 05 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ± ± ± ± ± ± ± ± ± ± ± ± b ± ± ± 0 H. officinalis 14 03 04 05 12 20 37 18 04 08 04 10 ± 01 74 64 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c b ab b a bc a ab bc b a a 01 01 02 10 02 02 00 03 09 14 01 . . . . 07 ...... 0 0 0 0 0 0 0 0 0 0 0 0 ± ± c b ± ± ± ± ± ± ± b b ± ± ± 0 0 0 0 T. capitata 03 01 29 55 06 05 01 08 ± ± 39 69 01 01 ± ± ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b b c b ab b abc b c ab 01 01 01 02 04 07 01 01 01 05 07 ...... 0 0 0 0 0 0 0 0 0 0 0 c b ± ± ± ± ± ± ± a ± ± b b ± ± 0 0 0 0 0 M. officinalis ± ± 02 01 04 03 15 22 02 ± 01 01 ± ± 24 27 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ≤ 0.05 as determined by Tukey’s B multiple comparison test. comparison multiple B Tukey’s by determined as 0.05 ≤ P P ab a a a a ab ab bcd a a a a a a a 02 02 02 03 03 03 02 02 06 02 01 04 11 11 14 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± N. tuberosa 0 07 04 09 09 12 09 14 06 44 06 03 17 ± 29 93 83 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 nificantly different at a abc c b b a d b a ab a b 01 01 11 04 02 01 05 01 03 03 03 05 ...... 0 0 0 0 0 0 0 0 0 0 0 0 ± b ± ± ± b ± ± ± ± b ± ± b ± ± 0 0 0 0 S. officinalis ± 01 02 10 14 ± 05 02 ± 28 01 12 03 ± 61 15 ...... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ab b b c 04 02 04 04 . . . . 0 0 0 0 b c b c ± b b c d c a b ± ± c ± 0 0 0 0 0 0 0 0 0 0 0 0 ± ± ± ± ± ± ± 02 ± ± ± 07 ± 06 ± 15 . . . . S. verbenaca S. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 SE) insect visits per minute to flowers of selected aromatic selected in of 2012 visits minute SE) insect toplant species 2013 flowers and per ± 2012 2013 2012 2013 2012 2013 2012 2013 2012 2013 2012 2013 2012 2013 2012 2013 Year 2. Mean (

Beetles Hoverflies Honey bees Honey Digger bees Digger Leafcutter bees Leafcutter Big solitary beesBig Pollinators (total)Pollinators Small solitary bees Table III - noteach rowsig are in letter the same with Means

Chapter III

In 2012, Salvia officinalis was highly attractive to small solitary bees (F6,363 =

8.98, P ≤ 0.01) and beetles (F6,363 = 11.56, P ≤ 0.01), while in 2013 was attractive only to hoverflies (F7,231 = 7.40, P ≤ 0.01). The flowers of N. tuberosa showed high attractiveness to digger bees (2012: F6,363 = 15.73, P ≤ 0.01; 2013: F7,231 = 3.79, P ≤

0.01), leafcutter bees (2012: F6,363 = 5.75, P ≤ 0.01; 2013: F7,231 = 10.17, P ≤ 0.01) and hoverflies (2012: F6,363 = 3.25, P ≤ 0.05; 2013: F7,231 = 7.40, P ≤ 0.01) in both years of the study. In 2012, N. tuberosa was attractive to small solitary bees (F6,363 = 8.98, P ≤

0.01) and big solitary bees (F6,363 = 18.23, P ≤ 0.01), and in 2013 was the most attractive plant species to honey bees (F7,231 = 3.06, P ≤ 0.05) and beetles (F7,231 = 3.79, P ≤ 0.01). In general, Melissa officinalis was not attractive to pollinators in both years of the study. The most frequent visitors to M. officinalis flowers were leafcutter bees

(F6,363 = 5.75, P ≤ 0.01) and small solitary bees (F6,363 = 8.98, P ≤ 0.01), but only in 2012. Small solitary bees were the most frequent visitors in the plots of Thymbra capitata in both years, 2012 (F6,363 = 8.98, P ≤ 0.01) and 2013 (F7,231 = 16.75, P ≤ 0.01).

In addition, H. officinalis was highly attractive to leafcutter bees (2012: F6,363 = 5.75, P

≤ 0.01; 2013: F7,231 = 10.17, P ≤ 0.01) in both years of the study, while in 2012 was frequently visited by honey bees (F6,363 = 5.22, P ≤ 0.01), small solitary bees (F6,363 =

8.98, P ≤ 0.01) and hoverflies (F6,363 = 3.25, P ≤ 0.05), and in 2013 by digger bees

(F7,231 = 3.79, P ≤ 0.01). On the other hand, when compared with the mono-specific aromatic plots, both mixed plots (M1 and M2) showed low attractiveness to all groups of pollinators, although the diversity of insects was high (Table III-2).

3.2. Relationship between morphological characteristics of the flowers and insect visitation

Although the results showed statistically significant differences among plant species for each of the measured morphological flower characteristics (Table III-3), the results of DCA showed that only two of the studied morphological characteristics were significant for plant discrimination (Table III-4).

Chapter III

Table III-3. Morphological characteristics of flowers of the six Lamiaceae species studied

Corolla tube Plants Flower length Flower height Calyx length Corolla height length S. officinalis 1.55 ± 0.04a 0.87 ± 0.03c 0.74 ± 0.02a 1.04 ± 0.04b 0.82 ± 0.04b N. tuberosa 1.02 ± 0.02c 0.91 ± 0.01c 0.74 ± 0.02a 1.02 ± 0.02b 0.72 ± 0.02bc T. capitata 0.68 ± 0.02e 1.17 ± 0.04a 0.42 ± 0.01c 0.62 ± 0.01d 0.58 ± 0.01d M. officinalis 1.28 ± 0.04b 1.04 ± 0.03b 0.67 ± 0.01a 1.33 ± 0.04a 0.66 ± 0.03cd H. officinalis 0.89 ± 0.03d 0.86 ± 0.02c 0.57 ± 0.01b 0.78 ± 0.02c 0.71 ± 0.03bc S. verbenaca 1.24 ± 0.02b 1.15 ± 0.04a 0.69 ± 0.02a 0.87 ± 0.02c 0.99 ± 0.05a *Means (± SE) with the same letter in each column are not significantly different at P ≤ 0.01 as determined by Tukey’s B multiple comparison test. All values are represented in cm.

Flower length contributed to discrimination of the plant species by the first canonical axis, while corolla tube length contributed to discrimination of the plant species by the second canonical axis (Table III-4).

Table III-4. Correlation between discriminating variables and 1st and 2nd standardized canonical discriminant functions (80.04%)

Morphological Function characteristics 1 2 3 4 5 Flower length 0.824* 0.246 0.503 0.032 -0.080 Corolla tube length 0.623 -0.634* 0.239 0.388 0.049 Flower height -0.208 0.020 0.478 0.752 0.404 Calyx length 0.548 0.061 -0.265 0.600 -0.516 Corolla height 0.179 0.378 -0.035 0.514 0.749 *Morphological characteristic that discriminate between selected plants at P ≤ 0.01

In addition, the graphical representation of DCA showed that T. capitata was separated from other five species upon the first canonical axis, while the second axis discriminated M. officinalis from other species (Fig. III-2). As shown in Table III-4, T. capitata had significantly the shortest flower length (F5,59 = 115.05, P ≤ 0.01), while M. officinalis had the longest corolla tube among selected plant species (F5,59 = 88.99, P ≤ 0.01).

Chapter III

Fig. III-2. Discrimination among plant species upon the first and second canonical axis

Figure III-3 shows the two-dimensional scatterplot of CATPCA analysis, relating the morphological features of flowers and the visits of the different groups of pollinators.

Chapter III

Fig. III-3. Relationship between morphological characteristics and number of insect visits in 2012 (A) and in 2013 (B)

Chapter III

The figure shows that beetles and hoverflies are closest to the centroid and distant from the two principal components, meaning that their contribution to the analysis is low. Although DCA showed that some floral characteristics discriminate between plant species, CATPCA showed that there is no correlation between insect visitation and morphological characteristics of the flowers. Therefore, the fact of no clustering between floral characteristics and insect visits emphasized that those two traits are not related (Fig. III-3).

3.3. Relationship between floral phenology and insect visitation

In early spring 2012 mean daily temperatures were below 15 ºC and no pollinators visited the flowers of S. verbenaca (Fig. III-4).

Chapter III

Fig. III-4. Relationship between floral phenology and insect visitation

Chapter III

In May, when the mean temperatures increased, S. officinalis and N. tuberosa started to bloom, and the peaks of floral density and insect visits overlapped at the end of May (S. officinalis) and beginning of June (N. tuberosa), respectively. The peaks of floral density and the insect visits overlapped in mid-June for M. officinalis, and at the beginning of July 2012 for T. capitata. Finally, the peak of floral density in H. officinalis occurred at the end of September, while the insect visits had two peaks, the first at the end of August and the second at the end of September. Both mixtures (M1 and M2) started to bloom at the beginning of May, reaching their peaks of floral density and insect visits at the beginning of June. In the summer period (July-August) both traits decreased until September, when they reached their second peaks (Fig. III-4).

In early spring 2013 mean daily temperatures were higher than the previous year and pollinators visited the flowers of S. verbenaca. However, May 2013 was rainy and cool (with the mean temperatures around 15 oC). This cool spring favoured the high floral density of S. officinalis and N. tuberosa, but it led to low frequency of insect visits (Fig. III-4). At the beginning of June the mean temperature started to increase and stayed between 20 and 30 oC until September, when it started to decrease again. The curves of floral density and insect visits for M. officinalis and T. capitata in 2013 showed similar tendencies as in 2012. However, in 2013 both the insect visits and the floral density of H. officinalis reached one peak in mid-September, but the abundance of insects was lower than the previous year. Similarly to 2012, both mixtures (M1 and M2) started to bloom in May 2013, reaching the peaks of floral density and insect visits at the beginning of June. In the summer period (July-August), both traits reduced their values approximately to zero, and at the end of September reached their second peak (Fig. III-4). In both years of study, the results showed a strong positive correlation between floral density and number of insect visits (year 2012: r = 0.748, n = 608, P ≤ 0.01; year 2013: r = 0.661, n = 648, P ≤ 0.01).

3.4. Spatial expansion of the plant species

The results of spatial expansion from 2014 showed that of the six plant species included in the experiment, only two, S. verbenaca and N. tuberosa, expanded from their original plots planted in 2011 (Table III-5).

Chapter III

Table III-5. Spatial increase (%) of the area occupied by plant species one year after planting

Initial area in Area in 2014 Spatial increase of Plant species 2011 (m2) (m2) initial area (%) S. verbenaca 23.04 32.17 40 S. officinalis 23.04 23.04 0 N. tuberosa 23.04 31.39 36 M. officinalis 23.04 17.27 -25 T. capitata 23.04 23.04 0 H. officinalis 23.04 23.04 0

In contrast, the area occupied by M. officinalis was reduced by 25% in comparison with its initial area. Other three plant species (S. officinalis, T. capitata and H. officinalis) maintained their area within the initial plots during 3 years of the study (Table III-5). None of the plant species intruded in the neighbouring barley fields, thus showing no weediness or danger to invade the crop.

4. Discussion

The selected aromatic plant species attracted a great diversity of bees, while hoverflies were in general less frequent visitors. Nevertheless, hoverflies are described as less effective pollinators of Lamiaceae species than bees (Garcia, 2000). However, results of both years of study showed clear differences in the floral attractiveness to pollinators among aromatic flower species, where N. tuberosa and H. officinalis were more attractive to local pollinators than S. verbenaca, M. officinalis and the mixtures (M1 and M2). In addition, similarly to previous studies (Pontin et al., 2006), preferences of the different groups of pollinators towards specific plants were detected. In both years of study, T. capitata was mostly visited by small solitary bees. Nepeta tuberosa and H. officinalis were mainly visited by leafcutter bees and hoverflies, while digger bees preferred the flowers of N. tuberosa. Although honey bees are polylectic and have been reported as the most frequent visitors and the most efficient pollinators of H. officinalis, S. officinalis and M. officinalis (Macukanovic-Jocic et al., 2011), in this study, honey bees were in general scarce, visiting only the flowers of H. officinalis and N. tuberosa. Nevertheless, since the main purpose of field margins is the conservation

Chapter III of beneficial insects, plants that attract mayor diversity of pollinators, like N. tuberosa are preferable.

Although many studies have stated that the morphological characteristics of flowers influence the plant’s attractiveness to insect visitors (Buide, 2006; Reith et al., 2006; Reith et al., 2007), few studies have compared the morphology and the attractiveness to pollinators of various aromatic species under the same conditions (Petanidou and Vokou, 1993; Macukanovic-Jocic et al., 2011). This study shows that only two morphological characteristics discriminated among plant species in terms of their attractiveness to pollinators, i.e. flower length and corolla tube length. Thus, T. capitata, visited mainly by short-tongued small solitary bees, was discriminated from other plant species due to its short flowers. Conversely, despite the longest corolla tube of M. officinalis, this species was rarely visited by the long-tongued bees and was mainly visited by small solitary bees. Although this finding is in accordance with Engel and Dingemansbakels (1980), the CATPCA analysis showed that morphological characteristics and visits of pollinators were unrelated, suggesting that flower size does not have a crucial influence on the attraction to pollinators. However, although morphological characteristics of flowers did not influence the insect preference toward different plant species, the results revealed a strong positive correlation between insect visits and floral density, showing its important role in plant attractiveness (Dauber et al., 2010; Scriven et al., 2013).

Low mean temperatures limited the presence of pollinators in the field and consequently their visits to the flowers of S. verbenaca in April 2012, while April 2013 was warm and pollinator’s visits were observed. Although S. verbenaca was not frequently visited by pollinators, it blooms in early spring when flowers in landscapes are scarce, offering food source for pollinators before the crops flower (Nicholls and Altieri, 2013). As such, the contribution of S. verbenaca may be important in a floral mixture (Petanidou and Vokou, 1993). The flowering period of N. tuberosa and S. officinalis coincided with high temperatures and high abundance of pollinators in the field in 2012. Nevertheless, in 2013, May and June were cool and rainy, and the plants in full bloom with high floral density were rarely visited by pollinators. Therefore, this study emphasizes that weather conditions can limit the presence of pollinators in the field, overriding the influence of floral phenology (Martin Gonzalez et al., 2009; Abrahamczyk et al., 2011). However, as weather conditions can change from one year 34

Chapter III to another, it is important to maintain floral sources available to pollinators from early spring until late autumn.

Field margins containing a large floral diversity have become the focus of numerous conservation efforts (Kells et al., 2001). In this study, although plant mixtures allowed a longer blooming period than any of the individual stands, maximum flower densities and insect visits were attained with some specific plant species. Several reasons may explain this fact. Although mono-specific plots of S. officinalis and N. tuberosa flower profusely and attract numerous pollinators in early spring, N. tuberosa was significantly more attractive to all functional groups of pollinators than S. officinalis. This probably was reflected on the lower relative attractiveness of the mixture in the period of their flowering (May-June). During early summer (June-July), flowering of M. officinalis and T. capitata in both mixed plots was considerably reduced compared with their respective mono-specific plots. Although relatively high floral density and insect visitation was observed in mono-specific plots of T. capitata, due to its low height (20-30 cm), flowers were not available to pollinators when mixed with taller plants, limiting its contribution to the mixture. During late-summer and early- autumn, H. officinalis was the only plant species in the mixture with high floral density, which contributed to the high attractiveness of the mixed plots. Even so, it is likely that the growth and flowering capacity of this late growing plant was somewhat reduced by competition from the other early growing species. In addition, it has to be considered that floral density in the mixed plots has been lower than in mono-specific plots, due to the mismatch in flowering periods of different plant species.

The plants assessed in this study are native and frequent in the study area, and none of them showed a weedy behaviour nor intruded in the neighbouring crop. Most of them maintained their original stands during three years, with the exception of the reduction in the case of M. officinalis. Considering that perennial plants are recommended as floral sources to pollinators in agro-ecosystems due to their low maintenance (Tuell et al., 2008), the fact that the initial area of M. officinalis was reduced may represent a limitation for its usage in floral margins. Therefore, a correct plant selection is essential for the efficient functioning of plant mixtures (Nicholls and Altieri, 2013).

Chapter III

The plants in this study were pre-selected to obtain continuous flowering of their mixture. However, the results showed that under field conditions, other factors such as weather and competition between plants can influence their availability to pollinators in the mixtures. Therefore, the results provide some interesting hints for the design of improved species mixtures, where mixing individual species with different flowering periods and avoiding competition between them may be critical. Thus, other possible option could be planting a mosaic of mono-specific stands and trying to optimize their management. The explanation for lower number of pollinator visits in mixed plots was probably due to the lower floral density in these plots (Essenberg, 2012; Hudewenz et al., 2012), as within the mixed plots it never occurred that all plants were flowering at the same time. Therefore, there is a need to exclude the plants that were not efficient in the floral mixture (i.e. M. officinalis and T.capitata) in order to increase the floral density of the mixture, and offer constant food sources to pollinators.

CHAPTER IV

Published online: Agricultural and Forest Entomology (Accepted: 15/05/2014)

Barbir, J., Badenes-Peréz, F.R., Fernández-Quintanilla, C., Dorado, J. (2014). The attractiveness of flowering herbaceous plants to bees (Hymenoptera: Apoidea) and hoverflies (Diptera: Syrphidae) in agro-ecosystems of Central Spain. Agricultural and Forest Entomology (DOI: 10.1111/afe.12076)

Chapter IV IV. The attractiveness of flowering herbaceous plants to bees (Hymenoptera: Apoidea) and hoverflies (Diptera: Syrphidae) in agro-ecosystems of Central Spain

1. Introduction

The global biodiversity decline is affecting the functioning of ecosystems given that it is threatening the stability of the ecosystem services (Dobson et al., 2006; Wratten et al., 2012; Vanbergen and Insect Pollinators Initiative, 2013). Under these circumstances, agro-biodiversity can play a fundamental role in providing ecosystem services such as pollination and pest control in agro-ecosystems (Altieri, 1999; Gurr et al., 2003; Scheper et al., 2013). Bees are generally the most efficient pollinators of wild flowers and commercial crops (LaSalle and Gauld, 1993; Jauker et al., 2012; Woodcock et al., 2013), while hoverflies are effective pollinators of open flowers and many species are also natural enemies of aphid pests (Ambrosino et al., 2006; Fontaine et al., 2006).

Many studies have reported an overall decline in the abundance and diversity of wild pollinators (Altieri, 1991; Potts et al., 2010; Carvalheiro et al., 2013). Some studies have linked this decline to the potential spread of diseases (Vanbergen and Insect Pollinators Initiative, 2013) or competition (Goulson, 2003; Walther-Hellwig et al., 2006) from managed honey bees. On the other hand, changes in climate (Memmott et al., 2010; Rader et al., 2013), a reduction and a fragmentation of habitats (Pfiffner and Wyss, 2004), an expansion of intensive agriculture (Benton et al., 2003; Watling and Donnelly, 2006) and high inputs of pesticides (Brittain and Potts, 2011; Gill et al., 2012) are also important factors that are associated with the decline in pollinators.

Nevertheless, semi-natural habitats can offer diverse forage plants and shelter for beneficial insects within the agricultural fields, which has been recognized as essential for both biodiversity conservation (Billeter et al., 2008) and for the maintenance of ecosystem services in agro-systems (Klein et al., 2003b; Carvalheiro et al., 2012). Therefore, as the parallel diversity decline of pollinators and insect-pollinated plants has been reported (Biesmeijer et al., 2006), suggesting a functional coupling and mutual dependence between both guilds (Frund et al., 2010), several studies have focused on

Chapter IV improving the habitat for beneficial insects within agro-ecosystems in order to conserve them. Some of those studies have suggested using diverse field margins to provide nectar and pollen resources (Carvell et al., 2007; Heard et al., 2007) or nesting habitat (Goulson et al., 2010) to pollinators in agro-ecosystems, while others have investigated the differences in attractiveness of flowers to pollinators (Pontin et al., 2006; Hogg et al., 2011). Although plant attractiveness is considered an important driver when choosing the plants for beneficial insects’ conservation practices (Campbell et al., 2012a; Carrie et al., 2012), it is essential to combine it with the factor of flowering duration of the different plant species (Nicholls and Altieri, 2013).

Due to the lack of practical information on attractiveness and management of beneficial plants in Spain (Miñarro and Prida, 2013), the implementation of known beneficial plants within Spanish agro-ecosystems is still limited. Therefore, the main objective of this study was to find the most appropriate plant species to improve the habitat for pollinators in agro-ecosystems of Central Spain, where a preliminary selection of plants prioritized native or naturalized (Tuell et al., 2008; Morandin and Kremen, 2013) and not invasive (Stout and Morales, 2009) plants. Consequently, this study addressed several specific objectives: i) to evaluate both attractiveness (insect visitation) and attractiveness efficiency (a combination of duration of flowering and insect visitation) of various plant species from diverse botanical families; ii) to evaluate how the selected plants respond to two different agronomic managements (growing in mixture vs. mono-specific stands; tillage vs. no-tillage at the end of the season to check their ability of self-seeding); and iii) to estimate the spatial expansion, i.e. potential weediness, of the different plant species, because it may threaten the functionality of agro-ecosystems (Marshall and Moonen, 2002).

2. Materials and methods

2.1. Study area and experimental design

This study has been conducted over two years at the experimental farm La Poveda (40º 19' N and 3º 29' W, elevation 536 m), Arganda del Rey, Madrid, Spain. The study

Chapter IV area is characterized by a continental Mediterranean climate with cold winters and hot summers and very scant precipitation (about 400 mm per year).

Table IV-1 shows the twelve herbaceous plant species used in 2011, which were organised in two experiments (16.5 m × 11.4 m) within a barley field and separated by 50 m, with the longer side orientated NE-SW. Each experiment was divided into three blocks (repetitions) that consisted of eight plots (1.30 m × 1.30 m) including six plant species in monoculture and two plots with a mixture of these six species equally proportioned and randomly distributed within a plot (M1 and M2). The plots within each block were randomly organized in order to avoid any systematic influence of the neighbouring species (separation between plots was 1.2 m and between blocks 2.0 m).

Table IV-1. Plant species and their flowering periods in 2011 and 2012

Common name Latin name Family Flowering period 2011 2012 Chives Allium schoenoprasum L.1 Liliaceae 14 June - 27 June Excluded Snapdragon Antirrhinum majus L.2 Scrophulariaceae 8 June - 5 July Excluded Borage Borago officinalis L.1 Boraginaceae 3 May - 21 June 11 May - 26 June Field marigold Calendula arvensis L.1 Asteraceae 18 April - 8 June 9 May - 4 June Cornflower Centaurea cyanus L.1 Asteraceae 18 May - 8 June 11 May - 5 July Coriander Coriandrum sativum L.2 Apiaceae 27 April - 26 May 9 May - 24 May Wild-rocket Diplotaxis tenuifolia (L.) DC.1 Brassicaceae 18 April - 27 June 11 May - 10 July Paterson's Curse Echium plantagineum L.2 Boraginaceae 10 May - 21 June 18 May - 21 June Sweet alyssum Lobularia maritima (L.) Desv.2 Brassicaceae 18 April - 27 June Excluded Common melilot Melilotus officinalis (L.) Pall.2 Fabaceae did not flower* Excluded Phacelia Phacelia tanacetifolia Benth. 1 Hydrophyllaceae 27 April - 21 June 11 May - 26 June French marigold Tagetes patula L.2 Asteraceae 18 Abril - 5 July Excluded 1 plants in experiment 1 in 2011, 2 plants in experiment 2 in 2011

*Melilotus officinalis was not included in analysis in 2011, since the plants were eaten by rabbits in its early stage and did not flower.

After comparing the attractiveness among the species, the duration of flowering and their contribution in the mixtures in 2011, the least appropriate plants were excluded (Table IV-1) and the study was continued in 2012 with the seven selected plant species and their mixture.

Chapter IV

2.2. Insect visitation assessment

Grouping of pollinators. The local insects were grouped in accordance with their morpho-physiological features to function as pollinators, such as body size and method of pollen collection. The analysed groups were: honey bees, Apis mellifera L. (Apidae), leafcutter bees (Megachilidae), small solitary bees (Halictidae and Andrenidae, with body size ≤ 1cm), big solitary bees (Apidae, Halictidae and Andrenidae, with body size > 1cm) and hoverflies. Bumblebees, Bombus spp. (Apidae) and carpenter bees, Xylocopa violacea L. (Apidae) were not included in the final analysis as fewer than ten visits were recorded during the whole experiment.

Although the main focus of this study was on the flower visitation of bees and hoverflies, some beetles were studied as they occupied the same feeding niche as bees and hoverflies and in that way indirectly influenced the visits of the main pollinators.

Attractiveness of the plants. In order to estimate the attractiveness of flowers to local pollinators, the mean number of visits of bees and hoverflies per minute (Campbell et al., 2012) was compared between plant species and their mixtures during the spring of 2011 and 2012. Insect visits to the flowers were counted twice per week (9 min/plot) in a marked area of 0.5 m × 0.5 m, located in the centre of each plot. Visits were counted only when the insect had a direct contact with the reproductive structures of the flower (i.e., stamens and pistils). The observations were conducted by three observers, always in the morning (9.00 h – 13.00 h local time, Greenwich Mean Time + 2:00) when the weather conditions were preferable for the insects to fly and visit flowers (sunny days without wind and mean temperatures between 22oC and 32oC). Switching plots, each observer counted visits of insects for three minutes in each plot. In order to identify the different insect species, insect captures were made twice per month using a butterfly net and an aspirator. Insect were always caught in the morning, a day after the experimental observations thus not biasing the results.

Attractiveness efficiency of the plants. Considering that the attractiveness of a plant species to pollinators cannot be efficient in agro-ecosystems if its flowering duration is short, the term attractiveness efficiency has been introduced in this study, visualizing a relation between the total pollinator visits to the flower/minute and the duration of flowering (the total number of observation days).

Chapter IV

Fig. IV-1. Attractiveness efficiency of flower species. Schematic diagram depicting the relationship between the total number of insect visits/minute (y-axis) and flowering duration (number of observation days) of plant species (x-axis): A (upper-right corner): high total number of insect visits/minute in a long observation period (high attractiveness efficiency) ; B (upper-left corner): high total number of insect visits/minute in a short observation period; C (bottom-right corner): low total number of insect visits/minute in a long observation period; and D (bottom-left corner): low total number of insect visits/minute in a short observation period

Therefore, when dividing the scatter plot into four quadrants, the highest attractiveness efficiency of a plant species (i.e., a high total number of insect visits/minute over a long period of observation) will be located in the upper right quadrant (Fig. IV-1). The flowering period was presented as the number of observation days since those two parameters were proportional (observation of insect visits was conducted every 3-4 days).

Chapter IV

2.3. Floral density

The capacity of plants to produce flowers was assessed. For this, the number of flowers was counted on each observation day within the marked area (0.5 m × 0.5 m), while the percentage of floral coverage per plot was assessed once per week. Each observer estimated the percentage of flower coverage per m2 in each plot, and the mean value of the three readings was calculated. Those parameters were used in order to estimate the floral density within the studied plots and to investigate its influence on the attractiveness of the plants (Stout et al., 1998). Each capitulum of species belonging to the Asteraceae family (i.e., T. patula, C. cyanus and C. arvensis) was considered as an individual flower. Any plot was considered in bloom when it had more than five open flowers per observed area.

2.4. Effect of tillage and no-tillage on the self-seeding of plant species

At the end of each floral season (July), how the two different soil management practices (tillage and no-tillage) affected self-seeding (the ability of the plant to reproduce itself) in the studied plant species was compared. In order for this to take place, half of each plot was shallow tillage while in the other half no-tillage was applied. At the beginning of the following season (January), when the majority of seedlings emerged, the number of seedlings under each treatment was counted three times using a 25 cm × 25 cm frame. At this time, the density of seedlings was used to estimate which agronomical treatment was more appropriate for the successful maintenance of the plant species in successive years.

2.5. Potential weediness of plant species

In spring (April-May), the spatial expansion from the original area [Δ Area (%)] of each plant species (the sum of the three mono-specific plots (1.30 m × 1.30 m) and one sixth of all the mixed plots), was calculated considering the area occupied the following year by using real time kinematic differential GPS (RTK-GPS). The percentage increase in area occupied by each plant species relative to their original area (i.e. potential weediness) was calculated using the software ArcGIS® 9.3 (ESRI, Redlands, California, USA). 43

Chapter IV

2.6. Data analysis

The data from 2011 and 2012 were independently analysed with SPSS® Statistics19 software (IBM, Chicago, IL, USA). A simple general linear model (GLM) was used in order to assess the ability of different insects to discriminate between flower species. The mean value of insect visits per minute within the observation area (0.5 m × 0.5 m) was used as response variable for these analyses. The data of hoverflies visitation in 2012 was square root transformed prior to analysis in order to meet the assumption of normally distributed residuals and homogeneity of variance. Differences among means were explored by Tukey’s B post hoc test (P ≤ 0.05). Univariate ANOVA was also used to test whether there were significant differences between tillage and no- tillage treatments. Finally, a bivariate two-tailed Pearson analysis was used to assess the correlation between floral density and visits of pollinators/minute.

3. Results

3.1. Diversity of pollinators in the studied area

Bees were the most diverse group of pollinators (Table IV-2) with 41 different species from five families (Apidae, Andrenidae, Colletidae, Halictidae and Megachilidae). Due to the semiarid conditions of the experimental area, only 7 species of hoverflies were observed (Table IV-2). Even though weather conditions in the spring 2011 were favourable for hoverflies (mean temperature 17oC and mean daily precipitation 3 mm), a drier and colder spring in 2012 (mean temperature 15oC and mean daily precipitation 1 mm) reduced their abundance and diversity in 2012 (SIAR, 2014). In addition, four species of beetles (Table IV-2) were found occupying the same feeding niche as the insects of interest.

Chapter IV Table IV-2. List of insects visiting the studied plant species Orden Family Species Size (mm) Groups of insects *

Coleoptera Dasytidae Psilothrix viridicoerulea (Geoffroy, 1785) 7 beetles Coccinelidae Coccinela septempunctata (Linnaeus, 1758) 7 beetles Tenebrioidae Heliotaurus ruficollis (Fabricius, 1781) 10 beetles Scarabaeidae Tropinota squalida (Scopoli, 1783) 14 beetles

Hymenoptera Apidae Apis mellifera (Linnaeus, 1758) 11 -16 honey bee Bombus terrestris (Linnaeus, 1758) 12-21 bumblebee Bombus muscorum (Linnaeus, 1758) 10-20 bumblebee Xylocopa violacea (Linnaeus, 1758) 22-23 carpenter bee Amegilla quatrifasciata (Villers, 1789) 12-15 solitary big Anthophora agama (Radoszkowski, 1869) 15-20 solitary big Anthophora atroalba (Lepeletier, 1841) 12-16 solitary big Anthophora fulvitarsis (Brullé, 1832) 16-18 solitary big Habropoda zonatula (Smith 1854) 14-17 solitary big Eucera caspica (Morawitz, 1873) 11-12 solitary big Eucera clypeata (Erichson, 1835) 11-12 solitary big Eucera elongatula (Vachal, 1907) 9-11 solitary big

Andrenidae Andrena albopunctata ssp. melona (Rossi, 1792) 13 -16 solitary big Andrena bicolorata (Rossi, 1792) 11-15 solitary big Andrena bimaculata (Kirby 1802) 10-13 solitary big Andrena florea (Fabricius, 1781) 10-15 solitary big Andrena gelriae ssp. gredana (van der Vecht, 1927) 9 - 13 solitary big Andrena haemorrhoa (Fabricius, 1781) 9-11,5 solitary big Andrena hispania (Warncke, 1967) 13-16 solitary big Andrena morio (Brullé, 1832) 12-19 solitary big Andrena nigroaenea (Kirby 1802) 10-14 solitary big Andrena ovatula (Kirby 1802) 10-13 solitary big Andrena similis (Smith, 1849) 7-12 solitary big Andrena thoracica (Fabricius, 1775) 11-15 solitary big Andrena truncatilabris ssp. espanola (Morawitz, 1877) 9-12 solitary big Andrena variabilis (Smith 1853) 10-16 solitary big Andrena vetula (Lepeletier, 1841) 10-12,5 solitary big Andrena sp. 2-7 solitary small Andrena sp. 1-5 solitary small Panurgus canescens (Latreille, 1811) 4 -6 solitary small

Colletidae Colletes cunicularius (Linnaeus, 1761) 12-13,5 solitary big

Halictidae Halictus scabiosae (Rossi, 1790) 14 -17 solitary big Lasioglossum callizonius (Perez, 1895) 6-8 solitary small Lasioglossum sp. 2-5 solitary small Lasioglossum sp. 9 solitary small Lasioglossum sp. 2-8 solitary small

Chapter IV

Family Species Size (mm) Group of insect Megachilidae Anthidium oblongatum (Illiger, 1806) 7-8,5 leafcutter bee Chelostoma emarginatum (Nylander, 1856) 9 leafcutter bee Megachile sp. 1 9 leafcutter bee Megachile sp. 2 11 leafcutter bee Osmia tricornis (Latreille, 1811) 10-14 leafcutter bee

Diptera Syrphidae Episyrphus balteatus (De Geer, 1776) 9 -12 hoverfly Eupeodes corollae (Fabricius, 1794) 6-11 hoverfly Sphaerophoria scripta (Linnaeus, 1758) 8 -11 hoverfly Sphaerophoria rueppellii (Weidemann, 1820) 8-11 hoverfly Eristalis tenax (Linnaeus, 1758) 10-12 hoverfly Melanostoma sp. 9-11 hoverfly Scaeva sp. 8-10 hoverfly

* Insects were classified in eight groups (beetles, honey bees, bumblebees, carpenter bees, solitary big, solitary small, leafcutter bees and hoverflies). All solitary bees, except leafcutter bees, were classified in two groups using the body length [solitary small (≤ 10 mm) and solitary big (> 10mm)] as main criteria

3.2. Attractiveness of different plants and their mixtures to pollinators

In general, the visitation frequency of bees was significantly higher in 2012 than in 2011, while the mean number of visits per minute of hoverflies and beetles was much higher in 2011 than in 2012 (Fig. IV-2 and IV-3).

In 2011, bees were most attracted to the flowers of B. officinalis, P. tanacetifolia and D. tenuifolia in experiment 1 (F6,17 = 6.78, P≤ 0.01), while in experiment 2 the flowers of E. plantagineum and M2 were most attractive to the bees (F5,15 = 24.08, P≤ 0.01). In 2012, flowers of C. sativum, E. plantagineum and P. tanacetifolia were significantly more visited by the bees than the flowers of B. officinalis and C. arvensis

(F7,16 = 7.20, P≤ 0.01) (Fig. IV-2). In general, the most frequent visitors were solitary bees, although the results of both years point to the preference of honey bees for B. officinalis (year 2011: F6,17 = 16.80, P≤ 0.01; year 2012: F7,16 = 6.23, P≤ 0.01) as well as leafcutter bees for the flowers of C. cyanus (year 2011: F6,17 = 2.35, P= 0.07; year

2012: F7,16 = 9.79, P≤ 0.01).

In 2011, hoverflies visited significantly more frequently the flowers of C. arvensis, D. tenuifolia and M1 (F6,17 = 16.69, P≤ 0.01) in experiment 1, while in experiment 2 the flowers of L. maritima were the most attractive followed by C. sativum (F5,15 = 49.08, P ≤ 0.01). The most regular hoverflies to visit the plants in 2011 were Sphaerophoria scripta and Sphaerophoria rueppellii, two species grouped as 46

Chapter IV Sphaerophoria spp. (Fig. IV-2). In 2012, few hoverflies were observed in the study area, preferably visiting the same plant species as in 2011, i.e. D. tenuifolia, C. sativum and C. arvensis (F7,16 = 4.48, P≤ 0.01), except L. maritima that was not used in 2012.

Abbreviations: 2011 (1): Bor= Borago officinalis; Cal= Calendula arvensis; Cen= Centaurea cyanus; Dpl= Diplotaxis tenuifolia; Pha= Phacelia tanacetifolia; Ali= Alium schoenoprasum; M1= mixture of those species. 2011 (2): Cor= Coriandrum sativum; Ech= Echium plantagineum; Lob= Lobularia maritima; Tag= Tagetes patula; Ant= Antirrhinum majus; M2= mixture of those species. 2012: M= mixture of the species used in 2012.

Chapter IV

According to the results, the flowers of E. plantagineum (F5,15 = 9.62, P≤ 0.01) and P. tanacetifolia (F6,17 = 32.61, P≤ 0.01) were the most attractive to beetles in 2011, while in 2012 beetles were more abundant on the flowers of C. sativum (F7,16 = 9.01, P≤ 0.01) (Fig. IV-2). The most frequent visitor to all flowers was Heliotaurus ruficolis, except C. arvensis and C. cyanus which mainly attracted Psilothrix viridicoerulea and Coccinella septempunctata.

Furthermore, the results showed a strong positive correlation between floral density and the number of visits by bees in both years of the study (year 2011: r= 0.119, n= 810, P≤ 0.01; year 2012: r= 0.413, n= 407, P≤ 0.01), while visits of hoverflies were significantly positively correlated with the floral density in 2011 (r= 0.634, n= 810, P≤ 0.01), but not in 2012, probably because of the low abundance of hoverflies in the field area that year.

3.3. Attractiveness efficiency of different plants and their mixtures to pollinators

Figure IV-3 illustrates the scatter plot showing the relationship between the total number of beneficial insect visits and the duration of the flowering period of each plant species (i.e., the attractiveness efficiency). The plant species with the highest attractiveness efficiency (i.e., the upper-right square of the scatter plot) for bees were B. officinalis, D. tenuifolia, P. tanacetifolia and the mixtures in 2011, and P. tanacetifolia, E. plantagineum, C. cyanus, D. tenuifolia and the mixture in 2012. Regarding hoverflies, the highest attractiveness efficiency was observed in L. maritima in 2011 and D. tenuifolia in 2012. The lowest attractiveness efficiency (i.e., bottom-left square of the scatter plot) for both bees and hoverflies was found in A. majus, A. schoenoprasum and C. cyanus in 2011 and in C. sativum and C. arvensis in 2012 (Fig. IV-3).

Chapter IV

Fig. IV - 3. Attractiveness efficiency of the studied plant species for bees and hoverflies in 2011 and 2012 (abbreviations of plant species as in Fig. IV-2.).

Scatter-plots depict the relationship between the total number of insect visits/minute (y-axis) and flowering duration (number of observation days) of plant species (x-axis). Bars represent the standard deviation from 3 replicated plots.

After consulting the results of attractiveness efficiency for 2011 it was decided to exclude some plant species which were not attractive to bees (L. maritima), not attractive to hoverflies (T. patula and A. schoenoprasum) and that which was not attractive to neither bees nor hoverflies (A. majus). Additionally, A. schoenoprasum, L. maritima and T. patula were also excluded from the experiment as their height (maximum 20 cm) was limiting their contribution when mixed with taller plant species (60-100 cm) used in the experiment (J. Barbir, personal observation). Considering that the flowering period of C. cyanus was significantly reduced in 2011 due to aphid attack, it was decided to test again its attractiveness to pollinators in 2012.

Chapter IV

3.4. Effect of tillage and no-tillage on the self-seeding of plant species

The results of the first year (2011) showed that seedling emergence was significantly reduced by tillage in C. arvensis (F1,16 = 6.57, P= 0.02) and P. tanacetifolia

(F1,16 = 4.79, P= 0.04) with respect to no-tillage treatment. In contrast, C. sativum showed higher seedling emergence after tillage (F1,16 = 12.99, P= 0.02). Some plant species such as C. cyanus and D. tenuifolia did not emerge in 2011 (Table IV-3).

Table IV-3. Differences in seedling emergence (Nº plants/m2) after applying two different managements at the end of plants cycle: shallow tillage and no-tillage

Plant species Treatment No. plants m–2 ± SD 2011 2012

tillage 26.67 ± 30.98 1.78 ± 5.33 B. officinalis no-tillage 26.67 ± 17.88 19.56 ± 31.78 tillage 2816.89 ± 1103.66 231.11 ± 119.23 C. arvensis no-tillage 4522.67 ± 1664.15 * 1223.11 ± 78.02 * tillage 0 58.67 ± 32.98 C. cyanus no-tillage 0 190.22 ± 97.47 * tillage 313.56 ± 246.96 * 1.78 ± 5.33 C. sativum no-tillage 16.19 ± 16.54 0 tillage 0 5.33 ± 8.00 D. tenuifolia no-tillage 0 152.89 ± 78.43 * tillage 234.61 ± 60.1 1.78 ± 5.33 E. plantagineum no-tillage 230.26 ± 135.35 236.44 ± 127.94 * tillage 226.12 ± 96.79 1.78 ± 5.33 P. tanacetifolia no-tillage 318.13 ± 80.83 * 115.56 ± 160.75 * *significant difference between the treatments at P≤ 0.05

Similar results to those of 2011, only with a lower seedling emergence after tillage, were obtained in 2012 (Table IV-3) for C. arvensis (F1,16 = 436.24, P≤ 0.01) and

P. tanacetifolia (F1,16 = 4.50, P= 0.05). Although there was no significant difference between tillage and no-tillage treatments in C. sativum in 2012, there was a tendency that tillage positively affected the seedling emergence of this plant. Both, C. cyanus and D. tenuifolia emerged in 2012, and the number of seedlings was significantly reduced by tillage for both species (F1,16 = 14.71, P≤ 0.01 and F1,16 = 31.53, P≤ 0.01, respectively).

Chapter IV

3.5. Potential weediness of plant species

The results of both years demonstrated the potential weediness of C. arvensis since its initial area increased by 1632% (2011) and 1300% (2012) within one year (Table IV-4). Furthermore, in 2011 the initial area of P. tanacetifolia increased by 611% and in 2012 by 124%. Although the initial area of E. plantagineum increased by 275% in 2011, in 2012 it increased by only 12% (Table IV-4).

Table IV-4. Spatial increase (%) of the area occupied by plant species one year after planting

Initial Area in the year after Plant species Δ Area (%) area (m2) planting (m2) 2011 2012 2011 2012 2011 2012

B. officinalis 7.29 6.38 0.17 4.19 -98 -34 C. arvensis 7.29 6.38 126.26 89.35 1632 1300 D. tenuifolia 7.29 6.38 3.58 8.05 -51 26 C. cyanus 7.29 6.38 1.25 7.84 -83 23 P. tanacetifolia 7.29 6.38 51.8 14.26 611 124 A. schoenoprasum 7.29 - 0 - -100 - E. plantagineum 7.29 6.38 27.31 7.15 275 12 T. patula 7.29 - 0 - -100 - L. maritima 7.29 - 0 - -100 - M. officinalis 7.29 - 0 - -100 - A. majus 7.29 - 1.3 - -82 - C. sativum 7.29 6.38 0 0 -100 -100 External weeds 0 0 171.66 67.89 17166 6789

In both years of study the plants of B. officinalis, D. tenuifolia and C. cyanus appeared the following years but did not show weediness and maintained their area within the initial plots. On the other hand, seedlings from A. schoenoprasum, T. patula, L. maritima and M. officinalis did not emerge the following year.

Chapter IV 4. Discussion

In accordance with some previous studies (Colley and Luna, 2000; Frund et al., 2010) the results of this study showed differences in attractiveness among the plant species. In both years of study, the most attractive plants for bees were P. tanacetifolia and E. plantagineum, followed by D. tenuifolia and mixed plots. However, no consistent results were found for the other plant species from one year to another. Therefore, in 2011, B. officinalis was more frequently visited by honey bees than in 2012, while the flowers of C. cyanus and C. sativum had more bee visits in 2012 than in 2011. Some plants selected for this study, such as P. tanacetifolia and B. officinalis, have been previously studied in other regions, i.e. USA and UK, and are frequently cited as being attractive to bees (Carreck and Williams, 2002; Campbell et al., 2012a). In this study, the results confirmed that P. tanacetifolia was highly attractive to bees in Central Spain, while B. officinalis was the most attractive plant species to honey bees although being less visited in 2012 than in 2011.

On the other hand, the most attractive plant to hoverflies in 2011 was L. maritima, followed by C. sativum, C. arvensis, D. tenuifolia and mixed plots. In 2012 there were very few hoverflies in the study area, probably because of the cool and dry spring (Amoros-Jimenez et al., 2012). However, the flowers of D. tenuifolia were the most attractive to the few hoverfly visitors recorded. The results from 2011 confirmed the high attractiveness of L. maritima and C. sativum to hoverflies reported in other regions and under different field conditions (Ambrosino et al., 2006; Campbell et al., 2012a).

In addition to the attractiveness of plants to pollinators, a long flowering period is considered important (Adhikari and Adhikari, 2010). For that reason, the analysis of attractiveness efficiency of the plant species acquired important significance as it combined the total number of insect visits/minute and flowering duration, facilitating the choice of plants which are beneficial for insect conservation. Among the other plant species assessed, D. tenuifolia and E. plantagineum were attractive to bees for a long flowering period, reflecting their high attractiveness efficiency in both years of study. Similarly, the analysis of the attractiveness efficiency to hoverflies showed that L. maritima and D. tenuifolia maintained their attractiveness constantly during their long flowering period, while the short biological cycle of C. sativum reduced its

Chapter IV attractiveness efficiency in the field. Therefore, although attractive to hoverflies, a short flowering period limited C. sativum as a food source for hoverflies in agro-ecosystems. In contrast, D. tenuifolia which had a long flowering period and showed to be attractive for both groups of beneficial insects (bees and hoverflies) in both years of study could have an important potential in the conservation of pollinators within agro-ecosystems. In addition, D. tenuifolia could have an added economic value as it is widely used as rocket salad (D'Antuono et al., 2009; Hall et al., 2012).

The study pointed out that some inter-specific interactions between pollinators and non-pollinator insects can affect the attractiveness of plants through long flower occupation (Bosch, 1992) or damage (J. Barbir, personal observation). For example, H. ruficolis occupied the flowers of P. tanacetifolia and E. plantagineum, while P. viridicoerulea, described as a poor pollinator (Bartomeus et al., 2008), occupied the flowers of C. arvensis, reducing the approachability of the flowers to bees and hoverflies. Additionally, Tropinota squalida, described as a pest in agro-systems (Khater et al., 2003), negatively influenced the attractiveness of the studied flowers to beneficial insects by destroying the bottom part of corollas when reaching for the nectar of E. plantagineum (J. Barbir, personal observation). Nevertheless, in 2012 the number of beetle visits was significantly lower than in 2011, coinciding with an increase in the number of bee visits in P. tanacetifolia and E. plantagineum. As these inter-specific insect interactions can influence the final estimation of the attractiveness of the plants to bees and hoverflies, it would be advisable to be considered in further studies on the conservation of pollinators.

Regarding the management of plant species in a mixture, the hypothesis was that mixtures could combine the advantages of all species in terms of a longer flowering period and offering complementary resources relative to single species. However, the results of this study have revealed that, in general, mixed plots showed lower attractiveness efficiency than mono-specific plots of D. tenuifolia for both groups of pollinators and in both years of study. The lower number of visits of pollinators in mixed plots might be due to the lower floral density in these plots (Essenberg, 2012; Hudewenz et al., 2012). In contrast, although mixed plots were less visited by pollinators than some mono-specific plots, i.e. P. tanacetifolia, E. plantagineum or L. maritima, they showed a longer flowering period and higher diversity of floral traits. Furthermore, some plant species did not contribute efficiently to floral mixtures because

Chapter IV of their low attractiveness to pollinators (A. majus), low height (L. maritima) or both (A. schoenoprasum and T. patula). Therefore, plants such as L. maritima, which is attractive to hoverflies (Colley and Luna, 2000) and improves biological control in vineyards (Begum et al., 2006), has a greater potential as a beneficial plant, if planted as a mono-specific stand.

As the study area was relatively small, the results regarding the weediness of plants and their behaviour under field conditions should be considered as preliminary. Although tillage reduced the seedling emergence of C. arvensis within the plots, it was not effective in stopping its wide spatial expansion. This finding is in accordance with Saramago et al. (2012) who also described C. arvensis as weedy. On the other hand, seedlings of P. tanacetifolia also expanded beyond their original plots in the following years, but as the seedling emergence was significantly reduced after tillage treatment, this management practice may be appropriate as a control against its potential weediness. Indeed, previous studies (Rasmussen, 1991) suggested the positive weed- killing effect of harrowing on P. tanacetifolia. On the other hand, E. plantagineum expanded beyond its original plots only in 2011 and its self-seeding was successfully reduced by tillage. Although E. plantagineum has not been reported as weedy in the Mediterranean region, in other regions such as Australia, it has been described as invasive and weedy (Grigulis et al., 2001).

In conclusion, of twelve herbaceous plants included in this study D. tenuifolia can be implemented in agro-ecosystems of Central Spain as mono-specific stands, due to its high attractiveness efficiency to both bees and hoverflies and its efficient self- reproduction within the plots where it was planted. However, it can also be included in mixtures with other plant species recommended in this study, i.e. B. officinalis and C. sativum, as attractive to pollinators, not weedy, but with shorter flowering period than D. tenuifolia. On the other hand, C. arvensis, despite its attractiveness to hoverflies, is not recommendable as a beneficial plant in agro-ecosystems because of its potential weediness. Conversely, when considering the interactions among the studied insects, P. tanacetifolia and E. plantagineum, besides being attractive to bees, proved to be attractive to beetles that occupy the same feeding niche as local pollinators. In addition, those two plants also require tillage treatment to control their high self-seeding which can lead to potential weedines. Therefore, it is necessary to expand this study to a larger

Chapter IV scale in order to fully recommend the implementation of those beneficial plant species in agro-ecosystems of Central Spain.

CHAPTER V

Published online: Acta Agriculturae Scandinavica, Section B – Soil & Plant Science

(Accepted: 13/05/2014)

Jelena Barbir, José Dorado, César Fernández-Quintanilla, Tijana Blanusa, Cedo Maksimovic & Francisco R. Badenes-Pérez (2014). Wild rocket – effect of water deficit on growth, flowering and attractiveness to pollinators. Acta Agriculturae Scandinavica, Section B – Soil & Plant Science (DOI:10.1080/09064710.2014.925575)

Chapter V

V. Wild rocket - effect of water deficit on growth, flowering and attractiveness to pollinators

1. Introduction

Although fresh water is one of the most important natural resources in the world, it is not equally distributed, and in many regions is very limited and scarce. Considering that the largest part of available water is used for irrigation in agriculture, it is important to apply irrigation treatments that reduce water consumption and maximize water use efficiency (Howell, 2001; Christian-Smith et al., 2012). Regulated deficit irrigation (RDI), i.e. the deliberate water deficit applied to a crop at certain times (English, 1990), has been shown to save water while maintaining plant growth of various crops in areas with limited water sources (Chartzoulakis et al., 2002; Fereres and Soriano, 2007; Perez-Pastor et al., 2014; Perez-Perez et al., 2014). On the other hand, exposure of plants to water deficit can negatively affect plant development (Hsiao, 1973; Gutbrodt et al., 2011; Garcia-Tejero et al., 2012) and for this reason, while applying strategies that optimise water use in crop production, it is essential to mitigate the negative effects of water stress on plant development.

As it is closely related to crop production, the effect of water deficit on plant attraction to insect herbivores has been well-studied (Huberty and Denno, 2004). Preference for water-stressed over well-watered plants has been shown in some studies with certain insect herbivores (Showler and Moran, 2003), but other studies have shown the opposite (Hoffman and Hogg, 1992) or no effect (Badenes-Perez et al., 2005) of water deficit on herbivore preference. Considering that pollinators’ richness and abundance have been decreasing with a decrease in plant diversity (Biesmeijer et al., 2006), plants attractive to pollinators have taken an important role in their conservation (Batary et al., 2010; Carrie et al., 2012). Nevertheless, the effect of water deficit on the attraction of flowering plants to insect pollinators has not been studied yet.

Wild rocket, Diplotaxis tenuifolia (L.) DC. (Family: Brassicaceae), is one of the three cultivated rocket species whose demand and production has been increasing in Mediterranean countries (Bennett et al., 2007; Lamy et al., 2008; Znidarcic et al., 2011;

Chapter V

Hall et al., 2012). Besides being used as a crop, it can be used to attract bees and hoverflies over its long flowering season (Barbir et al., 2014a). Although D. tenuifolia is a C3 – C4 intermediate plant, with more efficient photosynthetic performance at high temperatures than other cruciferous plants with C3 photosynthesis (Schuster and Monson, 1990; Ueno et al., 2006), little is known about its ecology and physiology under water deficit. As pollinators are mainly nectar feeders and the effect of water deficit in insect preference is mediated by changes in plant physiology and plant chemistry (Huberty and Denno, 2004; Gutbrodt et al., 2011), it is important to test if water deficit could affect flower morphology and nectar composition in D. tenuifolia.

Considering the dry Mediterranean climate and its scarcity of water for irrigation, it is important to study the effect of water deficit on the growth and the attractiveness to pollinators of D. tenuifolia in order to evaluate its capability to cope with water scarcity without negative consequences for its use as a crop and a beneficial plant to pollinators. For this reason, the main objective of this study was to test different irrigation treatments to find the most efficient irrigation regime for D. tenuifolia, both as a crop and as a plant attractive to pollinators. In addition, two other objectives were to test the effect of different drought duration on growth, flower development and attractiveness to pollinators in D. tenuifolia; and to assess the effect of moderate deficit irrigation (MDI) and severe deficit irrigation (SDI) on growth, flower development and attractiveness to pollinators of D. tenuifolia.

2. Material and methods

The study was divided in two experiments. Experiment 1 was set up to test the effect of different drought treatments on growth and floral development of D. tenuifolia, and was conducted from May to August 2012 in Madrid, Spain. Experiment 2 was set up to test the effect of different RDI treatments on growth and floral development of D. tenuifolia, and it was conducted from August to October 2012 in Reading, UK. Both experiments also tested the effect of the different water deficit treatments on attractiveness of D. tenuifolia to bees and hoverflies.

Chapter V

2.1. Experiment 1 – Response of D. tenuifolia to drought

Seedlings of D. tenuifolia were transplanted to pots (10 cm × 13 cm) filled with 520 g of a mixture of peat-based compost and vermiculite (3:1) by volume, irrigated to container capacity and moved to the greenhouse. The minimal (night-time) temperature in the greenhouses during the experiment was 18-20 oC and the maximal (day-time) was 25-29 oC. The relative humidity varied from 50% to 80%.

Before starting the experiment, a preliminary study was conducted in order to estimate the duration of the drought treatments (D) and the number of replicates necessary to show significant differences in plant transpiration. For that purpose, plant transpiration was measured every 2-3 days in five plants that were not being irrigated (data not shown). All plants died after 16 days, so it was decided to apply drought treatments over a period of 14 days, in order to test if plants could recover by re- watering. Based on the preliminary differences in plant transpiration, it was decided to use 5 different drought treatments and a replication of six plants per treatment.

In the greenhouse, plants were randomly organized, over four tables and the drought treatments were applied after the plants of D. tenuifolia developed their first flowers. In the control treatment, each plant was under 100% potential evapotranspiration (ETp), i.e. adding all the water lost by evapotranspiration in the previous 24 h, while the other treatments did not receive any irrigation for 4, 8, 11 or 14 days (D-4, D-8, D-11 and D-14), respectively. Therefore, the D-14 treatment started first, followed by the D-11 treatment (three days later), the D-8 treatment (six days later) and the D-4 treatment (ten days later).

At the end of the experiment, three plants of each treatment were chosen for destructive analysis and other three plants were re-watered for 19 days to study if they could recover after drought treatments applied (R19). The measured parameters were plant height, biomass, number of leaves and open flowers, flower area (area/flower), total floral area (total area of all open flowers/plant, on the day of the measurements), flower diameter, corolla tube diameter, corolla tube length, and sugar concentrations in floral nectar. These parameters estimated the effect of water deficit on plant survival and development, while the effect of water deficit on the attractiveness of D. tenuifolia to pollinators was measured directly under field conditions.

Chapter V

2.2. Experiment 2 – Response of D. tenuifolia to RDI

Seedlings of D. tenuifolia were transplanted to individual pots (10 cm × 10 cm) filled with 400 g of peat-based compost, irrigated to container capacity and randomly organized over tables in a greenhouse with a minimal (night-time) temperature of 12-16 oC and a maximal (day-time) temperature 20-33 oC. The relative humidity varied from 50% to 80%.

Plants were exposed to three irrigation treatments (six plants per each) based on ETp. Control plants were grown under 100% ETp (all plants were weighed daily and watered back to the container capacity). Plants under moderate deficit irrigation (MDI) were grown under 60% ETp (plants received 60% of the 100% ETp values) while severe deficit irrigation (SDI) plants were under 30% ETp (plants received 30% of the 100% ETp values). Plants within this experiment were divided into two groups. In the first group, RDI was applied for a total of 30 days, 16 days before flowering (RDI-BF) and 14 days during flowering (RDI-DF); while in the second group, RDI started when all plants had developed their first three flowers (RDI-DF).

2.3. Water consumption

In order to estimate the quantity of water used for each irrigation treatment, total water consumption/plant for control and water deficit treatments (D-4, D-8, D-11, D-14, MDI-BF, SDI-BF, MDI-DF and SDI-DF) was calculated using following equation:

Total water consumption/plant = (CWD × No. daysWD) + (C100% ETp × No. days100% ETp)

Where CWD is the water consumption/plant under water deficit; No. daysWD is the number of days under water deficit; C100% ETp is the water consumption/plant under

100% ETp; and No. days100% ETp is the number of days under 100% ETp. Sum of the number of days under water deficit and the number of days under 100% ETp irrigation was 30 days for all treatments.

Chapter V

2.4. Plant transpiration and growth

As plants close their stomata as a result of water stress (Chaves et al., 2003), an AP4 Leaf Porometer (Delta-T Devices Ltd., Cambridge, UK) was used to measure leaf stomatal conductance (gs). Stomatal conductance measurements were made between 11:00 and 13:00 h every 3–5 days in two leaves per plant, randomly chosen among medium size leaves (6-10 cm length). Stomatal conductance was also used to assess apparent plant survival (and 0 mmolm-2s-1 indicating the cessation of transpiration by a plant).

The assessments of vegetative parameters were mainly focused on leaf number and plant biomass, because of their importance for commercial production of wild rocket (Znidarcic et al., 2011). Plant height and fresh and dry biomass of shoots (leaves, stems and flowers) were measured at the end of each experiment. Total number of dead and viable (living) leaves per plant was counted and the total leaf area (of living leaves) was measured using a Leaf Area Meter (Delta – T Devices Ltd., Cambridge, UK). In addition, leaves shorter than 10 cm (small leaves) and leaves longer than 10 cm (big leaves) were counted separately in order to estimate if the impact of water deficit varied depending on the size of the leaves.

2.5. Floral characteristics

In order to estimate total production of flowers per plant, open and senescent flowers were counted separately. At the end of each experiment, three open flowers per plant (the top youngest, the bottom oldest and one intermediate flower) of the highest stem were used to measure the diameter of the corolla and the corolla tube, and the corolla tube length using a Vernier calliper.

Total floral area per plant was measured using an Area Meter (Delta – T Devices Ltd., Cambridge, UK) by placing all open flowers from a single plant with petals facing the measurement pane. The average area per flower was calculated by dividing the total floral area with the number of flowers per plant.

Chapter V

2.6. Nectar analysis

As the effect of water deficit on insect preference is mediated by changes in plant chemistry (Gutbrodt et al., 2011), it was tested if water deficit could affect nectar composition. For that purpose, the nectar was extracted by using a modification of the rinsing method described in Petit et al. (2011). This method was preferred as it is not destructive, which facilitated the use of flowers for further measurements. Three flowers were used per plant, and each flower was washed twice with 2 μl of miliQ H2O. The content was transferred into an erlenmeyer and diluted to 5 ml with miliQ H2O. High- Performance Liquid Chromatography (Dionex, 4500i) was used to measure the content of glucose, fructose and sucrose in each sample. Separation was done on an ionic column (RCX-10, 250 × 4.6 mm, 7μm) using 5% sodium hydroxide 1N and 95% miliQ

H2O as solvents with a flow rate of 2 ml/minute (injection volume 50 μl).

2.7. Pollinator visits

In order to estimate if water deficit can affect the attractiveness of D. tenuifolia to pollinators, their visits to the flowers of plants grown under different water deficit treatments were monitored under field conditions. In both experiments, each insect visit was counted only when the insect had a direct contact with the reproductive structures of the flower (i.e., stamens and pistils). The observations were conducted on the last day of each experiment, in the morning (9.00 h – 14.00 h local time), when the weather conditions were preferable for the insects to fly and visit flowers (sunny days without wind and temperatures between 22oC and 32 oC). In each observation set (group of three plants) all opened flowers and insect visits per minute were counted (nine repetitions over a three-minute period for each observation set). Some insect species were identified in situ, while others were captured using a butterfly net for further identification.

In experiment 1, floral attractiveness of well-watered and drought-stressed plants to pollinators (bees and hoverflies) was estimated under field conditions in two sites: the CSIC experimental farm “La Poveda” (Arganda del Rey, Madrid, Spain) and the vegetable garden of the Center for Ecology and Education “Puente del Perdón” (Rascafría, Madrid, Spain). Data collection in the first site was conducted on the 16th of May 2012, while on the second site data collection was conducted on the 17th of July

Chapter V

2012. Plants were transported from the greenhouse to the field sites by van (driving distance < 1h). The experimental unit consisted of four observation sets (control, D-4, D-8, D-11) organized in a circle (Ø3 m). The experiment consisted of three experimental units separated 15 m from each other. Plants in the D-14 treatment were excluded from the experiment because none of them had open flowers.

In experiment 2, floral attractiveness of well-watered and RDI water stressed plants to pollinators was estimated under field conditions in the experimental garden of the University of Reading, UK. The experimental unit consisted of five observation sets (control, MDI-BF, SDI-BF, MDI-DF and SDI-DF) distributed in a circle (Ø3 m). The experiment consisted of two experimental units separated 15m from each other.

2.8. Statistical analysis

Data were analysed with univariate analysis of variance (ANOVA) using SPSS® Statistics19 software (IBM, Chicago, IL, USA). Prior to the ANOVA, the assumptions of normal distribution and homogeneity of variances were tested by Kolmogorov Smirnov and Levene’s tests, respectively. Differences among means were analyzed by Tukey’s B post hoc test (P ≤ 0.05). A bivariate two-tailed Pearson analysis was used to assess the correlation between floral parameters and visits of pollinators per minute.

3. Results

3.1. Response of D. tenuifolia to drought

Effect of drought on plant growth

Although gs decreased in all drought treatments (Fig. V-1A), significant differences were found only for the plants that were drought-stressed for 11 and 14 days

(F4,85=6.96, P ≤ 0.01). After six days of recovery, the gs of D-14 plants continued to be significantly lower than in other plants (F4,40=3.29, P = 0.02), while after the recovery period of 16 and 19 days, gs reached values similar to the control treatment (Fig. V-1B).

After finishing the drought treatment, plants with gs zero (no transpiration) appeared dead. However, all plants of control, D-4 and D-8 treatments, as well as 83% 63

Chapter V

of D-11 maintained their gs values above zero, while in 17% of D-11 plants and 100% of D-14 leave transpiration was reduced to zero. After 19 days of recovery, gs values in some plants recovered and the plants started to flower again (plant survival becoming 100% for D-11 and 33% for D-14 treatment at the end of experiment 1).

For each measurement (day), means with different letter are significantly different at P ≤ 0.05, as determined by the Tukey’s B multiple comparison test.

Furthermore, 14 days without irrigation (D-14 treatment) caused significant decrease in all vegetative parameters: plant height (F4,25=7.84, P ≤ 0.01), fresh biomass

(F4,10=23.77, P ≤ 0.01), dry biomass (F4,10=5.01, P = 0.02), total number of leaves

(F4,25=10.09, P ≤ 0.01), number of leaves ≤10 cm length (F 4,25=7.79, P ≤ 0.01), number of leaves >10 cm length (F4,25=7.75, P ≤ 0.01) while number of dead leaves significantly increased (F4,25=17.03, P ≤ 0.01). Eleven days without irrigation (D-11 treatment) caused a significant decrease of fresh biomass (F4,10=23.77, P ≤ 0.01), dry biomass

Chapter V

(F4,10=5.01, P = 0.02), total number of leaves (F4,25=7.75, P ≤ 0.01), number of leaves

>10 cm length (F4,25=7.75, P ≤ 0.01), while number of dead leaves significantly increased (F4,25=17.03, P ≤ 0.01). However, plant growth was not significantly influenced by 8 and 4 days without irrigation (Table V-1).

After 19 days of recovery, only fresh biomass weight was significantly lower in

D-14 plants compared to the control plants (F4,10=3.42, P = 0.05), while all other parameters were comparable to control levels (Table V-1).

Effect of drought on floral characteristics

After 14 days without irrigation, D. tenuifolia plants did not have any open flowers, so only senescent flowers and number of stems were counted. Significantly lower values for total floral area (F4,24=15.15, P ≤ 0.01) and number of open flowers

(F4,25=15.93, P ≤ 0.01) were found in D-8 and D-11 (Table V-2). The values of flower area (F4,24=66.59, P ≤ 0.01), flower diameter (F4,25=37.69, P ≤ 0.01), corolla tube diameter (F4,25=37.67, P ≤ 0.01), corolla tube length (F4,25=36.67, P ≤ 0.01) and number of stems were significantly lower in D-11 than in control, D-4 and D-8 plants

(F4,25=3.28, P = 0.05). After 19 days of recovery, no significant differences between treatments were found for any of the measured floral parameters in any of the treatments (Table V-2).

Chapter V

- * 14]) and of re 14 - -

D 7.18b 4.57a 9.33a 0.00a 28.00a 23.00a ≤ 0.01) ≤ 11 - P P D 4.40a 2.00a 13.80ab 24.00a 22.00a 11.33a - 8 - 11] or [D 14 days D ≤ 0.05; ** ** 0.05; ≤ 5.71a 1.00a - 16.24ab 26.33a 25.33a 18.33a P P

* At the end ofAt end the 19 days recovery 4 - D 5.70a 2.00a 8], 11 [D 14.30ab 28.50a 17.00a 11.00a - 2.33a 7.53a Control 22.32a 28.33a 26.00a 18.00a 4], 8 [D 4], - * ** ** ** ** ** ** 14 - D 1.00c 0.00b 1.00c 6.91b 4.65b 47.7b 27.00a 11 - D 5.05b 7.50ab 3.00bc 57.83ab 12.38b 10.5bc 13.67b 8 - D 5.68ab 6.17ab 3.33c 68.00a 23.60a 18.50ab 12.33a 4 - At the end of drought treatments D 6.34ab 6.17ab 4.00c 67.58a 28.79a 20.17ab 14.00a 7.73a 7.17a 3.83c 68.27a Control 32.07a 21.50a 14.33a ≤10cm) Effects of different drought treatments (no irrigation for 4 [D for treatments drought irrigation different (no of 1 . Effects -

Plant height (cm) Biomass fresh (g) (g) dry Biomass No. of leaves No. of ( leaves No. of (>10cm) leaves (dead) leaves of No. watering overon 19 days characteristicswatering morphological tenuifolia D. of plants. Table V Means within a row with different letter are significantly different byTukey’s multiple B comparison test (

Chapter V - 14 - 2.67a 3.33a 7.00a 0.84a D 1.96a 12.00a 75.67a 16.34a 11 - 14]) and of re D - 9.56a 1.48a 3.83a 4.44a 6.67a 6.33a 21.67a 150.33a ≤ ≤ 0.01). 8 - P P D 4.82a 1.24a 3.44a 4.33a 8.33a 3.67a 19.78a 196.00a 4 At the end ofAt end the 19 days recovery - ≤ ≤ 0.05; ** D 9.08a 1.16a 3.22a 4.11a 7.67a 7.33a 11] or [D 14 days 19.00a P P 124.67a -

* 1.20a 3.44a 4.00a 8.00a Control 8], 11 8], [D 10.01a 19.39a 12.33a - 267.00a ** ** ** ** ** * ** 4], 8 [D 4], - 14 ------D

5.83ab 62.83a

11 - D 1.08b 1.84b 0.32b 0.32b 5.20b 15.75cd 66.00a 15.00b 8 - D 1.46a 2.12a 0.40a 0.40a 6.17ab 25.54bc 66.83a 17.33b treatments 4 - drought D significantly different by Tukey’s B multiple comparison test ( test comparison multiple B Tukey’s by different significantly of 1.47a 2.16a 0.40a 0.41a 6.00ab 36.93ab 51.83a 25.17ab end the At 14 drought treatment did not have any open any have not did treatment drought 14 - Control 1.55a 2.22a 0.41a 0.41a 9.17a 50.23a 69.17a 32.17a under the D the under

) 2 ) Effect of treatments drought 4 [D different for Effect irrigation (no 2

2. -

Total floral area (cm area floral Total area (cm Flower diameter (cm) Flower (cm) diameter tube Corolla Corolla tube length (cm) No. stems (senescent) No. flowers (open) flowers No. Table V overon characteristics 19 days tenuifolia. floral watering D. the of Means within a row with different letter are Plants of D. tenuifolia

Chapter V

The concentration of sugars (glucose, fructose and sucrose) in nectar was measured directly after the drought treatments as well as after the recovery period. Although there is an apparent tendency for concentration of glucose and fructose to decrease in the plants exposed to longer period without irrigation, no significant differences were found between the treatments (Fig. V-2A). However, this tendency did not remain after re-watering the plants for 19 days (Fig. V-2B).

Fig. V-2. Concentration of sugars in nectar after different drought treatments (no irrigation for 4 [D-4], 8 [D-8], 11 [D-11] or 14 days [D-14]) (A) and after re-watering of 19 days (B).

No significant differences were found between drought treatments. The vertical bars represent standard errors. Concentrations of sugars in D-14 after re-watering are measured in only one plant that recovered after drought stress.

Chapter V

Effect of drought on pollinator visits

Due to the different environmental conditions between two sites (“La Poveda” and “Puente del Perdón”), there was some variation in the species of pollinators visiting the flowers of D. tenuifolia. In “La Poveda”, flowers of D. tenuifolia were mainly visited by bees (Anthophora spp., Andrena spp., and Lassioglossum spp.) and hoverflies (Sphaerophoria rueppellii and/or S. scripta, Scaeva sp. and Syrphus sp.). In “Puente del Padrón” both groups, hoverflies (Sphaerophoria rueppellii and/or S. scripta, Eupeodes corollae, Eristalis tenax, Melanostoma sp. and Episyrphus balteatus) and bees (Apis melifera, Megachile spp., Anthophora spp., Andrena spp., and Lassioglossum spp.) were more diverse.

At the site “La Poveda”, the mean number of bees visiting the flowers of D. tenuifolia was significantly higher in control plants and D-4 plants than in plants that were under longer drought (F3,4=16.94, P ≤ 0.01). However, the overall number of hoverflies was low and no significant differences were found between treatments (Fig. V-3A). Also, there was a strong positive correlation between number of flowers and number of visits by bees/min in the first experimental site (r=0.92, n=8, P ≤ 0.01), while there was no significant correlation between number of flowers and number of hoverflies visits (r=0.67, n=8, P = 0.07).

Chapter V

For each treatment, means that are significantly different at P ≤ 0.05 are marked with different letter (comparison within insect groups), as determined by the Tukey’s B multiple comparison test. The vertical bars represent standard errors.

At the site “Puente del Perdón”, the mean number of bees visiting the flowers of D. tenuifolia did not significantly differ between drought treatments (Fig. V-3B). Nevertheless, hoverflies apparently distinguished between different plants, visiting less the plants under drought (F3,8=7.29, P ≤ 0.01). There was no significant correlation

Chapter V between number of flowers and number of visits by bees (r=0.39, n=12, P = 0.21) and hoverflies (r=0.55, n=12, P = 0.07).

3.2. Response of D. tenuifolia to RDI

Effect of RDI on plant growth

Regulated deficit irrigation (RDI) applied before and during flowering significantly reduced gs in the D. tenuifolia plants (Fig. V-4). The gs of control plants -2 -1 averaged 330 and 300 mmolm s before and during flowering, respectively, while gs of the plants under the MDI and SDI were constantly lower than 150 mmolm-2s-1. On the first day of the experiment, there were no significant differences between treatments, but 4 and 6 days after RDI was applied, values of gs in MDI and SDI were significantly lower than in the control plants. In general, no significant differences in gs were found between MDI and SDI.

Chapter V

For each measurement, means with different letter are significantly different at P ≤ 0.05, as determined by the Tukey’s B multiple comparison test.

Applied before flowering, SDI significantly reduced height (F2,21=6.18, P ≤ 0.01), total leaf area (F2,17=43.63, P ≤ 0.01), fresh biomass (F2,17=23.36, P ≤ 0.01) and dry biomass (F2,17=6.98, P ≤ 0.01) of the plants, while the number of dead leaves increased after 30 days of SDI (F2,21=3.54, P = 0.12). However, no significant differences were found between control and MDI plants (Table V-3).

Chapter V

Table V-3. Effect of different regulated deficit irrigation (RDI) treatments (moderate deficit irrigation before flowering (MDI-BF) and during flowering (MDI-DF); and severe deficit irrigation before flowering (SDI-BF) and during flowering (SDI-DF) on the vegetative characteristics of D. tenuifolia. Measured after the treatments were done.

RDI (16 days BF + 14 days DF) RDI (14 days DF)

Control MDI-BF SDI-BF Control MDI-DF SDI-DF Plant height (cm) 43.50a 46.00a 33.40b ** 43.50a 42.17a 36.83a Biomass fresh (g) 22.81a 27.35a 8.14b ** 18.33a 20.67a 13.80b ** Biomass dry (g) 2.62a 2.59a 1.67b ** 2.62a 3.14a 2.60a No. leaves (tot) 35.25a 32.63a 26.88a 31.83ab 37.67a 25.70b ** No. leaves (≤10cm) 28.75a 26.25a 21.88a 24.83a 33.50a 24.67a No. leaves (>10cm) 6.50a 6.38a 5.00a 7.00a 4.83a 1.00b ** No. leaves (dead) 2.00b 4.25ab 4.88a * 2.67c 10.70b 17.70a ** Total leaf area (cm2) 219.22a 206.44a 140.11b ** 219.22a 214.05a 103.60b ** Means within a row with different letter are significantly different by Tukey’s B multiple comparison test (* P ≤ 0.05; ** P ≤ 0.01)

Applied during flowering, SDI significantly reduced fresh biomass (F2,15=13.32, P

≤ 0.01), number of big leaves (F2,15=11.41, P ≤ 0.01), number of dry leaves

(F2,15=27.26, P ≤ 0.01) and total number of leaves (F2,15=6.08, P ≤ 0.01), as well as total leaf area (F2,15=9.56, P ≤ 0.01). In contrast, MDI significantly increased the number of dry leaves when applied during flowering (Table V-3).

Effect of RDI on floral characteristics

While SDI applied before flowering influenced significantly growth and development of the vegetative parts of D. tenuifolia, it affected only the total floral area

(F2,17=4.78, P = 0.02) (Table V-4).

Chapter V

Table V-4. Effect of different regulated deficit irrigation (RDI) regimes (moderate deficit irrigation before flowering (MDI-BF) and during flowering (MDI-DF); and severe deficit irrigation before flowering (SDI-BF) and during flowering (SDI-DF) on the floral characteristics of D. tenuifolia. Measured after the treatments were done.

RDI (16 days BF + 14 days DF) RDI (14 days DF) Control MDI-BF SDI-BF Control MDI-DF SDI-DF Total floral area (cm2) 18.13ab 20.17a 9.06b * 18.13a 12.18ab 8.00b * Flower area (cm2) 1.78a 1.66a 1.42a 1.78a 1.66a 1.14b ** Flower diameter (cm) 2.19a 2.07a 1.88a 2.21a 2.13ab 1.90b * Corolla tube diameter (cm) 0.38a 0.39a 0.34a 0.38a 0.36a 0.34a Corolla tube length (cm) 0.41a 0.34a 0.36a 0.41a 0.36a 0.30b ** No. stems 3.88a 4.25a 2.50a 3.33a 2.67a 2.50a No. flowers (senescent) 25.13a 28.13a 18.63a 20.17a 13.67a 17.17a No. flowers (open) 10.00a 12.29a 6.43a 10.00a 7.50a 6.67a Means within a row with different letter are significantly different by Tukey’s B multiple comparison test (* P ≤ 0.05; ** P ≤ 0.01)

However, when applied during flowering, SDI caused significant reduction of total floral area (F2,15=3.50, P ≤ 0.05), flower area (F2,15=8.40, P ≤ 0.01), flower diameter (F2,15=5.53, P = 0.02), and corolla tube length (F2,15=10.32, P ≤ 0.01). Instead, MDI (when applied before and during flowering) did not have an effect on the number and characteristics of D. tenuifolia flowers compared to control plants (Table V-4).

Effect of RDI on pollinator visits

At the University of Reading (UK) site, both groups of pollinators (bees and hoverflies) were detected. Of all bees, two species of bumblebees were the most frequent visitors to the flowers of D. tenuifolia (Bombus terrestris and Bombus pascuorum). Conversely, the group of hoverflies was more diverse including four genera (Sphaerophoria sp., Eupeodes sp., Episyrphus sp. and Eristalis sp.).

Chapter V

For each treatment, means with different letter are significantly different at P ≤ 0.05 (comparison within insect groups), as determined by Tukey’s B multiple comparison test. The vertical bars represent standard errors.

Figure V-5 illustrates the mean number of visits of pollinators/min to the flowers of D. tenuifolia under different deficit irrigation treatments. Irrespective of whether MDI or SDI, were applied before or during flowering, there was no significant effect on the attractiveness of the plants to bees and hoverflies (Fig. V-5). However, no significant correlations between number of flowers and visits of bees (r=0.06, n=30, P = 0.74) and hoverflies (r=-0.07, n=30, P = 0.71) were found.

Chapter V

3.3. Water consumption

In order to assess the most suitable irrigation treatment for D. tenuifolia, the total water consumption was calculated for each treatment for 30 days (Table V-5).

Table V-5. Total water consumption per plant for irrigation of potted D. tenuifolia during 30 days experiment after drought treatments (D) of 4 [D-4], 8 [D-8], 11 [D-11] or 14 days [D-14]) and regulated deficit irrigation (RDI) treatments (moderate deficit irrigation before flowering [MDI-BF] and during flowering [MDI-DF]; and severe deficit irrigation before flowering [SDI-BF] and during flowering [SDI-DF]

Water deficit No. of days Total water Water Treatment No. Water under 100% ETp consumption/plant saved/plant days (g/day) irrigationa (g) (%) D-control - - 30 1467.6 0 D-4 4 - 26 1271.9 13 D-8 8 - 22 1076.2 27 D-11 11 - 21 929.5 37 D-14 14 - 16 782.7 47 RDI-control - - 30 1129.6 0 MDI-BF 30 22.6 - 677.7 40 SDI-BF 30 11.2 - 336.9 70 MDI-DF 11 22.6 19 963.8 15 SDI-DF 11 11.2 19 838.9 26 a 100% ETp irrigation (replacing all the water lost by evapotranspiration in the previous 24 h) in drought treatments was 48.92 g/day; while 100% ETp in RDI treatments was 37.65 g/day

The maximal water savings were measured for MDI applied before flowering, while the minimal water savings were measured in plants that were without irrigation for 4 days (D-4 treatment) and MDI applied during flowering (Table V-5).

Chapter V

4. Discussion

This study shows that drought periods of 11 and 14 days can reduce the gs to 0 mmolm-2s-1 in 33% and 100% of D. tenuifolia plants, respectively. However, after 19 days of re-watering most plants had gs values increased to the control plants’ values, except in 67% of D-14 plants that died. In plants that survived, drought stress of 11 and 14 days negatively influenced biomass and number of leaves, while no significant differences were found between control, D-4 and D-8 plants. Regardless of whether MDI was applied before or during flowering, it did not affect plant growth. In contrast, SDI applied before flowering significantly reduced biomass, total leaf area and plant height, which are all important parameters for commercial production of D. tenuifolia (Znidarcic et al., 2011). When applied during flowering, SDI significantly reduced biomass, total leaf area, total number of leaves, and caused a significant decrease in number of big leaves. Indeed, the negative effect of SDI on plant growth was less evident when applied before flowering. If produced as a crop under greenhouse conditions, D. tenuifolia can be under drought stress for no more than 8 days, or MDI can be applied without further negative effects on its commercial production. However, if produced under field conditions, additional experiments would be necessary as selection of an irrigation regime depends on the local climate (McDonald and Girvetz, 2013).

On the other hand, if the main purpose of D. tenuifolia is to attract pollinators, water deficit should not affect quantity and quality of its flowers. The results of this study show that all floral parameters, except sugar concentration in nectar, were negatively affected by drought of 11 or more days. These findings agree with Carroll et al. (2001), who also reported that drought stress of 12 days did not affect the concentration of sugars in the nectar of Epilobium angustifolium L. and with Cerekovic et al. (2013), reporting that drought stress of 12 days negatively affect development of leaves and flowers of blackcurrant (Ribes nigrum L.). In addition, this research shows that flower development in D. tenuifolia is more sensitive to drought stress than vegetative development, as drought of 8 days did not affect any of the measured vegetative parameters, while it reduced the number of open flowers and the total floral area. On the other hand, MDI and SDI did not influence floral traits when applied

Chapter V before flowering. However, when applied during flowering, SDI reduced floral area. As in other studies conducted with other plants (Mingeau et al., 2001; Cakir, 2004; Alvarez et al., 2013), these results show that D. tenuifolia responds differently to regulated deficit irrigation and water stress depending on the phenological stage of the plant.

Since pollinator visits are correlated with the number of open flowers and their total floral area (Conner and Rush, 1996), our results suggest that, under the temperature and relative humidity conditions of this experiment, D. tenuifolia could tolerate drought for 4 days without affecting its attractiveness to pollinators. Additionally, direct observations of pollinators’ visits to water-stressed plants were included in this study, confirming that the number of flowers can affect flower visitation by pollinators (Eckhart, 1992; Conner and Rush, 1996; Caraballo-Ortiz et al., 2011) as bees and hoverflies preferred to visit plants that had more opened flowers. This could be the main reason why there was no significant difference in the frequency of visits by pollinators to flowers of D. tenuifolia after RDI treatments, which did not affect the number of open flowers per plant. Nevertheless, the positive correlation between number of pollinator visits/min and number of flowers was not significant when the abundance of bees and hoverflies was low.

The simultaneous assessment of vegetative and floral parameters under different regimes of RDI and drought stress showed that RDI treatments caused less of a disturbance to the plant than 11 or 14 days of drought stress. Although not many studies have compared those two water deficit treatments, similar findings were found by Luo et al. (2007) and Cuevas et al. (2007) when measuring the effect of both, drought stress and RDI, on the growth and flowering of Eriobotrya japonica Lindl. Additionally, Cuevas et al. (2007) showed that MDI applied before flowering positively influenced the flowering of E. japonica, while if applied during flowering, it restricted the normal flower development.

Application of MDI before flowering was the most efficient water-saving technique for commercial production of D. tenuifolia, as it did not reduce plant growth and saved up to 40% of irrigation water. The longest drought period tested (D-14) reduced up to 50% the consumption of irrigation water, but resulted in death of most of the plants even after irrigation was reinstated. Although reducing vegetative growth,

Chapter V

SDI before flowering could save water consumption by 70% without affecting flowering and plant attractiveness of D. tenuifolia to pollinators.

In conclusion, in the environmental conditions of this study, potted D. tenuifolia can cope with drought of 4 days without having any consequences on its growth, flowering and attractiveness to pollinators. However, a longer drought period can negatively affect development of its flowers and reduce its vegetative growth. Additionally, this study shows that RDI can improve water use efficiency. Depending on the purpose of growing D. tenuifolia, water use can be reduced by 40% (if grown as a crop) or by 70% (if grown to attract pollinators).

CHAPTER VI

Submitted to: Agricultural and Forest Entomology (Submitted: 23/07/2014)

Barbir, J., Badenes-Pérez, F.R., Fernández-Quintanilla, C., Dorado, J. (submitted). Can proximity of floral field margins improve pollination and seed production in Coriandrum sativum L. (Apiaceae)?

Chapter VI

VI. Can proximity of floral field margins improve pollination and seed production of coriander crop, Coriandrum sativum L. (Apiaceae)?

1. Introduction

Coriander, Coriandrum sativum L. (Apiaceae), is a native plant in Mediterranean region, where it is mainly grown in a small-scale private gardens or fields (Carrubba et al., 2006). However, its cultivation has been widespread in a lot of regions all over the world (Western Europe, South and North America, India) and its production has become large-scaled (Maroufi et al., 2010). Coriander has been cultivated for its fruits (hereinafter referred to as “ seeds”), leaves and oils (Maroufi et al., 2010; Nadeem et al., 2013), characterized by high antioxidant activity (Wangensteen et al., 2004) desirable in healthy human diet.

Insect pollination is necessary for production of seeds in coriander, even though some pollen can be transmitted from one flower to another by wind (Koul et al., 1989; Bendifallah et al., 2013). As flowers of coriander are characterized for their exposed nectaries, abundant pollen production and zygomorphic flowers (Koul et al., 1989; Sinacori et al., 2014), they are attractive to flies (mainly hoverflies), bees and beetles (Colley and Luna, 2000; Bendifallah et al., 2013; Barbir et al., 2014a). However, foraging of pollinators in C. sativum flowers is limited due to its short bloom period (20 to 40 days, depending on the region). Therefore, there is a need to complement the agricultural landscapes with additional floral sources where pollinators can forage before and after C. sativum blooming.

Intensification of agriculture has reduced the availability of forage sources for pollinators (Tscharntke et al., 2005; Winfree et al., 2009; Le Feon et al., 2010), leading to the decline in abundance and diversity of pollinators in agro-ecosystems (Biesmeijer et al., 2006). This lack of pollinators in agricultural fields can cause poor pollination of crops, as well as a significant reduction in crop yields (Richards, 2001; Garibaldi et al., 2009). Therefore, in order to buffer this impact of intensive agriculture and to conserve wild pollinators within their natural habitats, it has been recommended to introduce

Chapter VI floral margins, as additional forage and shelter sources to pollinators, within monoculture landscapes (Landis et al., 2000; Marshall, 2005; Pywell et al., 2011; Rands and Whitney, 2011).

Many studies evaluated the importance of field margins in conservation of the beneficial insects in general (Landis et al., 2005; Olson and Wackers, 2007; Haaland et al., 2011) or natural enemies (Tinkeu et al., 1996; Steingrover et al., 2010) and pollinators (Pywell et al., 2011; Nicholls and Altieri, 2013; Scriven et al., 2013) in particular. In addition, some studies directly estimated how implementation of field margins reflected on the crop production by comparing yield/ha in the fields with and without floral margins (Carvalheiro et al., 2012). However, when assessing the effect of field margins on crop production in large-scale studies, the possibility for additional external factors (i.e. rainfall and temperature, physical and chemical properties of soil or plant development) biasing the results, is high. Therefore, the assessment of the relationship between yield, presence of field margins, and visits of pollinators to flowers in margins and crop should be preformed, excluding major external factors that can influence plant development and yield production. These factors can be minimized by potting plants of coriander (Ghazoul, 2006) and reducing the scale of the study, and in that way studying only a direct relation between pollinators visits and seed production. The selection of coriander as a target crop for this experiment was based on its requirement for insect pollination to produce seeds (even though some pollen can be transmitted from one flower to another by wind) as well as the easy way to estimate its total seed production under semi-field conditions.

In the previous study it has been shown how Diplotaxis tenuifolia L. DC. is equally attractive for two groups of the most abundant pollinators in Central Spain, i.e. bees and hoverflies, (Barbir et al., 2014a). Accordingly, it has been hypothesized that a margin with this floral species can attract as much pollinators and improve seed production in coriander to the same extend as a mixed margin including various plant species. Hence, the main objective of this study was to estimate the impact of field margins on seed production of coriander, where two different types of field margins were tested: mono-specific margin (D. tenuifolia) and mixed margin (six annual herbaceous species). For that reason, it was tested: i) if open pollination, i.e. control (without field margins), vicinity of mono-specific margins and mixed margins, increases the seed production of coriander when compared with no-pollination, i.e. 82

Chapter VI covered inflorescences of coriander, under field conditions; ii) if frequency of pollinator visitation to the flowers of coriander was higher in the presence of field margins than in control; and iii) if the crop seed production was higher in presence of field margins than in control without floral margins.

2. Material and methods

2.1. Study area and experimental design

This study was conducted in the spring of 2013 at the experimental farm La Poveda (40º 19' N and 3º 29' W, elevation 536 m) sited in Arganda del Rey (Madrid, Spain). The study area is characterized by a continental Mediterranean climate with cold winters and hot summers and very scant precipitation (about 400 mm per year). In general, floral areas are scarce within the experimental farm, as in the last four years maize, barley and poplars were mainly produced. The experiments were randomly distributed within those fields.

In order to test how the proximity of different field margins affects visits of pollinators and seed production in coriander, three different treatments were applied: the vicinity of mixed margin (mixture of six herbaceous species), mono-specific margin (D. tenuifolia) and control (absence of floral margin). Mixed margin (10 m × 1.5 m), consisting of six herbaceous species attractive to bees and hoverflies in Central Spain (Phacelia tanacetifolia Benth., Echium plantagineum L., D. tenuifolia, Centaurea cyanus L., Calendula arvensis L. and Borago officinalis L.), was sown in fall 2012. This floral mixture has been found as efficient and attractive to pollinators in the region (Barbir et al., 2014a).The mono-specific margins of D. tenuifolia were planted in February 2013, where 18 plants were planted in a row, with the separation of 0.5 m between them. From the previous experience with D. tenuifolia this distribution of the plants forms 10 m × 1.5 m margin (J. Barbir, unpublished results). Each experiment with a different type of margin was separated from others at least 200 m, while minimum separation between three repetitions within each treatment was 20 m.

In February 2012, seedlings of coriander were transplanted to pots (10 cm × 13 cm) filled with 520 g of a mixture of peat-based compost and vermiculite (3:1 by

Chapter VI volume), irrigated to container capacity and distributed over tables in the greenhouse. The minimum temperature in the greenhouse was 18-20 oC and the maximum temperature was 25-29 oC. The relative humidity varied from 50% to 80%. In April, one month before blooming, plants of coriander were transported to the field. Fifteen plants of coriander were organized in a row (separation between plants was 0.5 m) at 1.5 m distance from the margins and were buried in the ground together with their pots. The plants were potted in order to avoid the differences in soil properties which could affect crop development and seed production (Moore and Lawrence, 2013). During the experiment, each potted plant was irrigated 4 times/week with 0.5 l of water and each margin was irrigated 4 times/week with 10 litres of water each time.

2.2. Visitation of pollinators

In order to estimate the attractiveness of the margins to pollinators and its effect on the pollination of coriander, the mean number of visits of bees and hoverflies per minute (Campbell et al., 2012a) was counted in the flowers of the margins and coriander plants. Insect visits to the flowers of the margins were counted twice per week (9 min/margin) in three marked areas of 0.5 m × 0.5 m (3min/marked area), located in the centre and both ends of each margin. Insect visits to the flowers of coriander were also counted twice per week in four randomly chosen plants/row (3 min/plant). Visits were counted only when the pollinator had a direct contact with the reproductive structures of the flower (i.e., stamens and pistils). The observations were conducted always in the morning (9.00 h – 13.00 h local time, Greenwich Mean Time + 2:00) when the weather conditions were preferable for the insects to fly and visit flowers (sunny days without wind).

2.3. Floral assessment

The flower density (number of flowers) was counted every observation day within the marked area (0.5 m × 0.5 m) of all margins in order to investigate its influence on the attractiveness of the plants (Stout et al., 1998). Each capitulum of species belonging to the Asteraceae family (i.e., C. cyanus and C. arvensis) was considered as an individual flower. In addition, the total number of flowers/plant in coriander was

Chapter VI estimated by multiplying the mean number of flowers in 10 inflorescences randomly chosen and the total number of inflorescences found in the plants selected for observation of insect visits.

2.4. Seed production

In order to study how the presence of different field margins can affect pollination and seed production in coriander, potted plants of coriander were managed, i.e. regularly irrigated, under field conditions until all seeds were formed. At harvest time, all 135 plants were unburied from the ground and transported to the laboratory where the seeds were counted and weighted. Differently than in other studies which measured the number of seeds/plant to estimate the efficiency of insect pollination in coriander (Al-Fattah, 2001; Chaudhary and Singh, 2007), this study used number of seeds/inflorescence because the number of stems and inflorescences can vary from one plant to another. Therefore, for each plant, number of inflorescences, number of formed seeds/inflorescence (actual SN/inf) and number of dry flowers/inflorescence (flowers that did not develop its fruit) were counted in order to estimate the percentage of actual seed production in coriander. Sum of dry flowers/inflorescence and actual SN/inf represented potential of one inflorescence to produce seeds (potential SN/inf).

Seed production in coriander with and without open pollination

In order to estimate the grade of auto pollination in coriander and to study how different treatments of open pollination, i.e. control (without margins), vicinity of mixed margins and mono-specific margins, and no pollination can affect its seed production, a minimum of three inflorescences per plant were covered to prevent insect pollination. Therefore, three inflorescences per plant were covered 3-5 days before flowering while other three inflorescences (at the same plant and its stage of development) were marked with a blue string as not-covered. In total 220 inflorescences were covered (no pollination) and 220 were not-covered (open pollination).

Chapter VI

Seed quality

Seed weight (SW). Seeds from each plant were collected, counted and weighted using a precision balance. The mean weight per seed was calculated for each treatment. In addition, the seed weight/plant (SW/plant) was used as a parameter to evaluate whether the presence of floral margins improves the pollination in coriander, by calculating the percentage increase compared to control without margins.

Germination rate of coriander seeds. This parameter has been used to estimate the seed quality (Putievsky, 1980). Since there were no more than 30 seeds in some replications of covered inflorescences, the sample size in germination tests including covered inflorescences was limited to 10 seeds per petri dish in order to perform three repetitions per treatment. In contrast, the germination tests that did not include covered inflorescences (i.e., the comparison between controls without margin and crops with margins) did not have this limitation and therefore a greater number of seeds (e.g., four replications with 25 seeds per petri dish) were used.

2.5. Statistical analysis

The data were analysed with SPSS® Statistics19 software (IBM, Chicago, IL, USA). A simple general linear model (GLM) was used in order to assess the ability of different insects to discriminate between flower species. The mean value of insect visits per minute within the observation area was used as response variable for these analyses. Prior to an ANOVA analysis, the assumptions of normal distribution (Kolmogorov Smirnov test) and homogeneity of variances (Levene’s test) were tested. Differences among means were explored by Tukey’s B post hoc test. Univariate ANOVA was also used to test whether there were significant differences in seed production and seed germination rate between the treatments. Finally, a bivariate two-tailed Pearson analysis was used to assess the correlation between floral density and visits of pollinators/minute.

Chapter VI

3. Results

3.1. Visits of pollinators to the flowers

Visits of pollinators to the flowers of margins

The results showed no significant differences in the mean number of opened flowers per marked area between mono-specific and mixed margins (Table VI-1). However, frequency of pollinator visits was significantly higher in mono-specific than in mixed margins for all groups of pollinators. Indeed, the results showed that visits of bees (F1,34 = 10.54, P ≤ 0.01), hoverflies (F1,34 = 5.48, P ≤ 0.01) and beetles (F1,34 = 12.09, P ≤ 0.001) were more frequent in mono-specific margins than in mixed margins. Consequently, the mean number of visits/minute of total pollinators was significantly higher (F1,34 = 12.45, P ≤ 0.001) in the mono-specific margins than in mixed margins (Table VI-1).

Table VI-1. Mean (± S.E.) flower density and mean (± S.E.) number of visits of pollinators to the flowers of mono-specific and mixed floral margins

Mono-specific Mixed No. flowers/marked area 38.44 ± 2.73a 41.60 ± 9.63a No. beetles/min*** 0.68 ± 0.17a 0.08 ± 0.02b No. bees/min*** 0.51 ± 0.17a 0.13 ± 0.05b No. hoverflies/min** 0.36 ± 0.17a 0.16 ± 0.05b Total No. pollinators/min*** 1.55 ± 0.17a 0.37 ± 0.08b ***significant differences between the treatments at P≤ 0.001; the same letter in a row represents no significant differences between treatments

Visits of pollinators to the flowers of coriander

The results showed that number of inflorescences/plant and number of flowers/plant of coriander did not significantly differ between treatments (Table VI-2).

Chapter VI

Table VI-2. Mean (± S.E.) flower density and mean (± S.E.) number of visits of pollinators to the flowers of coriander without proximate floral margins (control) and in vicinity of mono-specific and mixed margins

Control Mono-specific Mixed No. inflorescence/plant 33.10 ± 4.42a 35.07 ± 3.14a 36.50 ± 3.66a No. flowers/plant 546.15 ± 72.86a 578.60 ± 51.85a 602.25 ± 60.46a No. beetles/min*** 0.07 ± 0.02b 0.52 ± 0.14a 0.87 ± 0.25a No. bees/min 0.11 ± 0.04a 0.12 ± 0.05a 0.30 ± 0.12a No. hoverflies/min*** 0.06 ± 0.03b 0.58 ± 0.12a 0.46 ± 0.15a Total No. pollinators/min*** 0.24 ± 0.06b 1.22 ± 0.19a 1.63 ± 0.36a ***significant differences between the treatments at P≤ 0.001; the same letter in a row represents no significant differences between treatments

However, number of beetles/min, number of hoverflies/min and number of total pollinators/min was significantly higher in mono-specific and mixed margins treatments

(F2,42 = 5.82, P ≤ 0.01; F2,42 = 5.97, P ≤ 0.01; F2,42 = 9.06, P ≤ 0.001) than in control without margins.

Correlations between number of flowers and number of pollinators in coriander plants and margins

Significant correlations between number of flowers and number of visits of pollinators/min were found for mono-specific and mixed margins treatments (Table VI- 3). In contrast, no significant correlations were found for coriander plants in the control treatment.

The number of flowers in mixed margins was positively correlated with the number of pollinators visiting flowers of both, mixed margins and coriander in the vicinity of mixed margins. Therefore, these results suggest an increase in the number of pollinators visiting flowers of mixed margins with the increase in number of flowers of coriander and vice versa. On the other hand, a strong positive correlation was found between number of flowers in mono-specific margins and number of pollinators visiting them.

Chapter VI

Table VI-3. Correlations between number of flowers and number of pollinators' visits/min

No. pollinators Coriander Margins No. flowers Control Mono-specific Mixed Mono-specific Mixed Control 0.291 - - - - Coriander Mono-specific - 0.145 - - 0.571* - Mixed - - 0.376 - 0.522* Mono-specific - 0.219 - 0.695** - Margins Mixed - - 0.528* - 0.552* *significant correlation at P≤ 0.05; **significant correlation at P≤ 0.01; (-) no correlation

Although no significant correlation was found between number of flowers in mono-specific margins and visits of pollinators to the adjacent coriander flowers, a negative correlation between number of coriander flowers and number of pollinators visiting flowers of mono-specific margins was detected (Table IV-3).

3.2. Seed production in coriander

Seed production in coriander with and without pollination

The results in Table VI-4 showed that the number of seeds/inflorescence formed in plants in the control was higher than in plants in the mono-specific and mixed margin treatments (F2,217 = 15.43, P ≤ 0.001).

Table VI-4. Potential (± S.E.), actual (± S.E.) and percentage (± S.E.) of actual seed number (SN) per inflorescence in covered inflorescences of coriander under control, mono-specific and mixed margin treatments

Control Mono-specific Mixed

Potential SN/inf.*** 15.68 ± 0.46a 12.54 ± 0.31c 14.38 ± 0.34b Actual SN/inf.*** 1.28 ± 0.23a 0.20 ± 0.06b 0.42 ± 0.11b Actual SN/inf. (%)*** 7.99 ± 1.45a 1.59 ± 0.52b 2.85 ± 0.72b ***significant differences between the treatments at P≤ 0.001; the same letter in a row represents no significant differences between treatments

In addition, the total potential of inflorescences to produce seeds was significantly higher in the control plants of coriander than in coriander plants in vicinity of the mixed 89

Chapter VI

and the mono-specific margins (F2,217 = 18.73, P ≤ 0.001). As these inflorescences of coriander were covered and insects did not get in contact with the flowers, consequently actual percentage of coriander seed number per inflorescence was higher in control than in the other two treatments (F2,217 = 12.81, P ≤ 0.001) (Table VI-4).

Considering the effect of insect pollination, the results showed a significantly lower number of seeds produced in the covered inflorescences of coriander than in inflorescences exposed to open pollination, e.g. control, mono-specific and mixed margin treatments (Table VI-5). Significant differences were also found between open pollination treatments, where coriander plants under control and mono-specific margins treatments produced significantly less seeds/inflorescence than the plants located next to mixed margins (F2,205 = 4.85, P ≤ 0.01).

Table VI-5. Potential (± S.E.), actual (± S.E.) and percentage (± S.E.) of actual seed number (SN) per inflorescence of coriander in not-covered (control, mono-specific and mixed margin treatments) vs. covered inflorescences

Control Mono-specific Mixed Covered inf.

Potential SN/inf.*** 16.13 ± 0.31a 13.55 ± 0.31c 14.78 ± 0.34b 14.10 ± 0.23bc Actual SN/inf.** 2.57 ± 0.28b 2.83 ± 0.28b 4.00 ± 0.38a 0.60 ± 0.09c Actual SN/inf. (%)** 15.86 ± 1.85b 20.15 ± 1.85b 25.68 ± 2.24a 3.99 ± 0.57c **significant differences between the treatments at P≤ 0.01; ***significant differences between the treatments at P≤ 0.001; the same letter in a row represents no significant differences between treatments

In contrast, the highest potential seed production was found in the control plants, while the lowest was found in mono-specific margin treatment (F2,205 = 13.47, P ≤ 0.001). Considering the relative value of actual production of seeds, the highest percentage of actual seed number was found in coriander plants next to the mixed margins (F2,205 = 5.32, P ≤ 0.01), while less than 4% of potential SN reached maturity in treatments with covered inflorescences (Table VI-5).

Seed quality

After counting the potential and actual number of seeds per inflorescence in all plants included in the experiment (4,426 inflorescences in total), significant differences

Chapter VI in number of seeds produced per inflorescence were found between control, mono- specific and mixed margin treatments (F2,4426 = 26.17, P ≤ 0.001) (Table VI-6).

Table VI-6. Potential (± S.E.), actual (± S.E.) and percentage (± S.E.) of actual seed number (SN) per inflorescence of coriander, and seed weight (SW) per seed (± S.E.), SW per plant (± S.E.) and percentage of SW per plant (± S.E.) under control, mono- specific and mixed margin treatments

Control Mono-specific Mixed

Potential SN/inf.*** 16.68 ± 0.11a 14.89 ± 0.08b 14.96 ± 0.08b Actual SN/inf.*** 4.06 ± 0.10c 4.60 ± 0.09b 5.04 ± 0.09a Actual SN/inf. (%) *** 23.25 ± 0.55c 28.97 ± 0.48b 31.82 ± 0.49a SW/seed (mg)*** 7.8 ± 0.9b 12.8 ± 0.8a 13.3 ± 1.0a SW/plant (g)** 0.96b 2.12a 2.45a SW/plant (%) 100.00 220.83 255.21 **significant differences between the treatments at P≤ 0.01; ***significant differences between the treatments at P≤ 0.001; the same letter in a row represents no significant differences between treatments

The highest potential SN per inflorescence was found in control plants of coriander (F2,4426 = 115.04, P ≤ 0.001), while no differences were found in coriander plants under mono-specific and mixed margin treatments (Table VI-6). However, this potential production did not translate into actual production, with significantly lower values of percentage of SN per inflorescence in control plants than in coriander plants close to mixed margins (the highest percentages) and those under mono-specific treatment (F2,4426 = 65.83, P ≤ 0.001).

Seed weight. The results showed that the average weight of each seed was significantly higher in coriander plants next to margins treatment than in control plants

(F2,4426 = 11.66, P ≤ 0.001), indicating better seed quality. Seed weight per plant was also significantly higher in mono-specific and mixed margin treatments than in control plants of coriander (F2,15 = 6.36, P ≤ 0.01), increasing the seed production for more than 220 % and 255 %, respectively (Table VI-6).

Germination rate of coriander seeds. Seed quality was higher in plants next to floral margins since the results of germination tests showed higher values in the

Chapter VI

germination rate of these treatments relative to control plants (F2,21 = 10.76, P ≤ 0.001) (Figure VI-1A).

Fig. VI-1. Germination rate of coriander seeds. Covered inflorescences of coriander (covered) and not-covered inflorescences of coriander in vicinity of: mono-specific margins (NO - D. tenuifolia), mixed margin (NO - mixed) and without margins (NO - control). Means with same letter in each graph are not significantly different at P ≤ 0.05, as determined by Tukey’s B multiple comparison test. Seed germination in: (A) sample of 25 seeds x 4 repetitions and (B) sample of 10 seeds x 3 repetitions.

Chapter VI

In the second batch of seed germination tests including seeds from covered inflorescences, the results showed similar values for seeds from plants next to the field margins (mixed and mono-specific) (F3,20 = 9.59, P ≤ 0.001), while seed from covered inflorescences, i.e. not exposed to open pollination, showed lower quality similar to the control plants (Fig. VI-1B).

4. Discussion

Although many studies confirmed that number of pollinator visits increases with higher density of flowers (Conner and Rush, 1996; Dauber et al., 2010; Essenberg, 2012; Barbir et al., 2014b), in this study no significant differences in flower density were found between mono-specific and mixed margins. Still abundance of pollinator visitors (bees, hoverflies and beetles) was significantly higher in mono-specific margins than in mixed margins, which could be explained by the relatively higher attractiveness of D. tenuifolia flowers to local pollinators compared to the floral mixture (Barbir et al., 2014a). The attractive effect of floral margins to pollinators has also affected nearby plants of coriander. Indeed, when comparing the open pollination treatments, i.e. control, mono-specific and mixed margin, number of coriander flowers per plant did not significantly differ, but the frequency of hoverfly and beetle visits to their flowers was significantly higher in the vicinity of mono-specific and mixed margins than in control plants without margins. However, no significant differences in bee visits/min were found between the different treatments of coriander.

The possible explanation could be the different biology of these insect groups. The most frequent beetles (Heliotaurus ruficolis L. and Psilothrix viridicoerulea L.) are highly attracted to the species used in the margins (Barbir et al., 2014a) and prefer diverse floral areas over bare soil (Woodcock et al., 2012). Hoverflies, unlike bees, do not collect pollen and nectar for their offspring and disperse into landscapes in linear manner, whereas bees return to their brood cells after foraging (Kleijn and van Langevelde, 2006). On the other hand, most hoverflies, show an extremely specialized selection of microhabitat sites for larval development, usually not found in intensively managed agro-ecosystems (Jauker et al., 2009), which can explain higher abundance of hoverflies in vicinity of floral margins than in control plants of coriander without 93

Chapter VI proximate field margins and surrounded by bare soil. In addition, majority of bee species in the study area are solitary-mining bees (Barbir et al., 2014a), which prefer to nest in bare soil (Cane, 1997) and in vicinity of their host plants (Julier and Roulston, 2009). Since control plants of coriander were mainly surrounded by bare soil, it probably contributed to the nesting of bees in surrounding area, which resulted in higher abundance of bees in the control plants of coriander compared with other treatments.

Recently, many studies directly correlated the abundance and diversity of pollinators with crop production under controlled conditions, i.e. cages or greenhouses (Jauker and Wolters, 2008; Jauker et al., 2012) or in open fields by estimating yield/ha (Carvalheiro et al., 2012). However, this study mixed those two methodologies, as it was conducted under semi-controlled field conditions (potted plants and regulated irrigation) in order to directly correlate insect visits with seed production by eliminating the influence of some environmental factors, as for example the soil properties. Therefore, a strong positive correlation between number of flowers in mixed margins and pollinator visits to nearby plants of coriander has been found, confirming that visits of pollinators were more frequent in the flowers of crop located next to the mixed margins than in control without floral margins (Ghazoul, 2006).

On the other hand, the highly attractive mono-specific margins of D. tenuifolia did not contribute to the spillover of insects from the margin to the crop flowers. However, the negative correlation between number of coriander flowers and visits of pollinators to the margin showed that when blooming of coriander ends, i.e. number of flower decreases, pollinators switch flower-foraging from coriander flowers to D. tenuifolia flowers. Therefore, although not contributing directly to the attraction of the crop to pollinators, D. tenuifolia showed good potential in their conservation in the study area (Barbir et al., 2014a).

In accordance with Jacobs et al. (2009), number of seeds/inflorescence and percentage of seeds produced/inflorescence were significantly lower in covered inflorescences than in open pollinated ones, thus demonstrating the essential role of insect pollination in this crop. However, when comparing the number of seed produced in the different treatments of open pollination, the highest number of coriander seeds per inflorescence was found near mixed margins, while the lowest number of seeds/inflorescence was achieved in control plants without margins, despite their

Chapter VI significantly higher potential for seed production. Indeed, crop plants in the vicinity of mixed margins produced 31% of their potential seed number, crop plants close to the D. tenuifolia margins produced 29%, while control plants without margins produced only 23% of their potential seed number.

Average weight per seed was one of the measures for estimation of the seed quality, which resulted significantly higher in both floral margin treatments than in the control plants of coriander. In addition, when comparing seed production, i.e. seed weight per plant, the results showed that when floral margins were present in the field, the seed production of coriander increased by more than 200%. Although the number of seeds/inflorescence was slightly lower in mono-specific treatment than in the mixed margin treatment, it did not reflect on the differences in final seed production/plant as the quality of seed was equally good. Another parameter for seed quality estimation was germination rate of produced seeds. The seeds of coriander plants in the vicinity of floral margins showed significantly higher germination rate than seeds produced in the control treatment.

Finally, it can be concluded that the proximity of both floral margins (mixed and mono-specific) improved the seed quality of coriander plants, as seed weight and germination rate were significantly higher than in control plants. However, number of seeds produced was significantly higher in the crop grown near mixed margins than near mono-specific margin, probably due to stronger positive correlations in pollinator visits to coriander flowers and number of margins’ flowers. As the experiment was conducted under semi-controlled conditions, it can be concluded that pollinator visits was the main factor that biased the results, and that presence of both mixed or mono- specific (D. tenuifolia) margins improve the crop production for more than 200% in small-scaled gardens and, in addition, conserve the local pollinators.

Chapter VII

VII. General discussion

Agricultural landscapes, dominated by large monocultures, have caused alarming declines in farmland biodiversity, threatening the optimal functionality of agro- ecosystems (Krebs et al., 1999; Benton et al., 2003) and leaving beneficial insects to suffer from deficit of food sources, shelter and nesting sites (Cane, 2008; Pywell et al., 2011; Carvalheiro et al., 2012). In addition, the recent report of parallel decline of pollinators and insect-pollinated plants (Biesmeijer et al., 2006) encouraged further investigations in plant-pollinator interactions (Campbell et al., 2012a; Carrie et al., 2012; Wratten et al., 2012) and habitat management (Cranmer et al., 2012; Holzschuh et al., 2012; Wratten et al., 2012) to enhance the presence of beneficial insectary plants within agro-ecosystems and therefore attract beneficial insects.

In order to promote the conservation of beneficial insects in a specific region, there is a need to ensure sustainable restoration of pollinator habitats within agro- ecosystems by using native beneficial insectary plants that can offer better food sources and shelter to pollinators (Menz et al., 2011). However, to manage beneficial insectary plants within specific agro-ecosystems, it is essential to reflect on its climate conditions (Memmott et al., 2010), as beneficial insectary plants of one region can behave as weedy or invasive under different environmental conditions of another region. On the other hand, it is also important to be familiar with the abundance and diversity of different pollinator groups of the specific region, mainly because the habitat management for conservation of pollinators frequently vary among different groups. For example, in Central Spain where diversity and abundance of solitary bees (especially mining bees) is relatively high in comparison with other groups, the habitat management applied should differ from the UK landscape management where the most abundant pollinators are bumblebees and hoverflies. Therefore, it is recommendable to study the agricultural behaviour and the attractiveness to pollinators of introduced or native beneficial plants under the field conditions of a particular region (Comba et al., 1999; Corbet et al., 2001), before recommending their use.

In the Mediterranean region, use and availability of aromatic (Lamiaceae) plants is very common. As many species of the aromatic plants are native, highly attractive to pollinators and well adapted to the semi-arid conditions of the Mediterranean region

Chapter VII

(Petanidou and Vokou, 1993; Garcia, 2000), their use for conservation of pollinators in agro-ecosystems should be considered. However, before implementing these plants in the agro-ecosystems, it is required to be familiar with the most appropriate ways to manage them within agricultural fields. For this reason, six aromatic plant species, with different bloom periods, were pre-selected for this study, as their good adaptation to semi-arid conditions and ability to survive in the field during hot summer months (Petanidou and Vokou, 1993) showed good potential for creation of a functional long- flowering floral mixture (from early spring to mid-fall).

The results revealed a strong positive correlation between insect visits and floral density (Dauber et al., 2010; Scriven et al., 2013). Since, floral morphology did not affect the attractiveness of different plant species to pollinators, high floral density of Nepeta tuberosa and Hyssopus officinalis was probably responsible for the high attractiveness of these plants to pollinators, mainly visited by leafcutter bees and hoverflies. On the other hand, the lower floral density in mixed plots was probably the explanation for reduced number of pollinator visits in these plots (Essenberg, 2012; Hudewenz et al., 2012), mainly because within the mixed plots never occurred that all plant species were flowering at the same time. In addition, some plant species included in the mixture were less attractive and consequently not efficient in the floral mixture (Melissa officinalis and Thymbra capitata). Therefore, these plant species should be replaced with more efficient ones, in order to increase the floral density of the mixture, and optimize the supply of food for pollinators.

In addition, the results showed that under field conditions, besides floral density, other factors such are weather and plant competition, can negatively influence the efficiency of plants in the mixture to attract pollinators. Therefore, this study provides some interesting hints for the design of floral margins, such is to avoid the usage of low- growing plant species in floral mixtures, since pollinators cannot reach the flowers. For example, the reduced pollinator visitation of T. capitata in the floral mixture was detected because pollinators were unable to approach the flowers. On the other hand, in mono-specific plots this plant was highly attractive to small solitary bees and for that reason its use in mono-specific stands is preferred over mixed. Following this example, this study emphasizes the importance to find the most appropriate management for each plant species, in order to achieve the greatest benefit from its implementation within agro-ecosystems. 97

Chapter VII

With a similar aim, twelve spontaneous native or naturalized annual herbaceous plants were studied under the field conditions of Central Spain. As all pre-selected herbaceous plants are spring flowering species, their attractiveness to pollinators was simultaneously compared together with their floral phenology (Adhikari and Adhikari, 2010), using the attractiveness efficiency, i.e. combination of the total number of insect visits/minute and flowering duration, as a parameter for comparison. This parameter goes one step further in the evaluation of beneficial insectary plants for conservation of pollinators, showing that some plant species as Borago officinalis or Coriandrum sativum, frequently cited as highly attractive and efficient beneficial plants (Ambrosino et al., 2006; Campbell et al., 2012a), are less efficient in conservation of pollinators due to their short flowering period. In contrast, Diplotaxis tenuifolia, which had a long flowering period and showed to be attractive for bees and hoverflies in both years of study, showed to be highly efficient in both, mono-specific stands (due to its high attractiveness efficiency) and mixed stands (due to its high contribution to the functionality of the mixture).

The study also pointed out that some inter-specific interactions between studied pollinators and other insects can affect the attractiveness of plants through flower occupation (Bosch, 1992) or damage (Barbir et al., 2014a). For example, Heliotaurus ruficollis occupied the flowers of Phacelia tanacetifolia and Echium plantagineum, while Psilothrix viridicoerulea, described as a poor pollinator (Bartomeus et al., 2008), occupied the flowers of Calendula arvensis, reducing the approachability of the flowers to bees and hoverflies. Additionally, Tropinota squalida, described as a pest in agro- ecosystems (Khater et al., 2003), negatively influenced the attractiveness of the studied flowers to beneficial insects by destroying the bottom part of corollas when reaching for the nectar of E. plantagineum (J. Barbir, personal observation). As these inter-specific insect interactions can influence the final estimation of the attractiveness of the plants to bees and hoverflies, it would be advisable to be considered in further studies on the conservation of pollinators.

Although the plants assessed in these two studies are native and frequent in the study area, some of them showed a weedy behaviour as they intruded in the neighbouring crop. Most of aromatic plant species maintained their original stands during three years, with the exception of the reduction in the case of M. officinalis. Considering that perennial plants are recommendable as floral sources to pollinators in 98

Chapter VII agro-ecosystems due to their low maintenance (Tuell et al., 2008), the fact that the initial area of M. officinalis was reduced may represent a limitation for its usage in floral margins. On the other hand, some herbaceous plants, i.e. C. arvensis, P. tanacetifolia and E. plantagineum, showed wide spatial expansion, intruding the neighbouring crop. Although shallow tillage significantly reduced their self seeding, it did not stop weedy behaviour of C. arvensis (Saramago et al., 2012). Hence, the usage of this plant as a beneficial insectary plant in agro-ecosystems of Central Spain is not recommendable.

After analyzing the potential attractiveness of all pre-selected plants (perennial aromatics and annual herbaceous) to pollinators and their agricultural management in agro-ecosystems it was necessary to deepen the recommendations regarding the habitat management practices. Therefore, the scale of further experiments was expanded, the number of plant species included (using only the most efficient plants) was reduced, and the experiments were focused on the crops of interests. In this way, besides conserving pollinators in agro-ecosystems, it is also possible to improve the pollination of targeted crops.

Therefore, as D. tenuifolia showed high attractiveness efficiency, good self- seeding and no weediness under field conditions, it has been considered as a good beneficial plant for conservation of local pollinators. However, before recommending it to local farmers, it was important to estimate its capacity to cope with water deficit as well as to recommend the most appropriate regulated deficit irrigation (RDI) regime due to the semi-arid conditions of Central Spain. Since the study was conducted under controlled green house conditions, the exact RDI regime under field conditions could not be recommended, as it highly depends of precipitation and soil type (McDonald and Girvetz, 2013). However, it was evaluated how different RDI regimes can influence the growth and the attractiveness of D. tenuifolia to pollinators. The study showed how the development of its flowers is more sensitive to drought stress than vegetative development, as drought of 8 days did not affect any of the measured vegetative parameters, since it reduced the number of open flowers and the total floral area. Considering that pollinator visits are correlated with the number of open flowers and their total floral area (Conner and Rush, 1996), the results suggest that, under the temperature and relative humidity conditions of this experiment, D. tenuifolia could tolerate drought for 4 days without affecting its attractiveness to pollinators. However, those results should be considered with precaution, as the experiment was conducted 99

Chapter VII under controlled conditions and it is recommendable to estimate the effect of drought under field conditions.

In agreement with previous researches conducted with other plants (Mingeau et al., 2001; Cakir, 2004; Cuevas et al., 2007; Alvarez et al., 2013), this study showed that D. tenuifolia responds differently to RDI and water stress depending on the phenological stage of the plant. Although moderate deficit irrigation (MDI) did not affect plant growth or floral attractiveness to pollinators in any phenological stage, severe deficit irrigation (SDI) applied before flowering significantly reduced some vegetative parameters, but not floral; whereas when applied after flowering SDI negatively affected all vegetative parameters and surface area of individual flowers. The study revealed that the negative effect of drought was reflected mainly on floral characteristics, while the negative effect of RDI was primarily expressed on the vegetative parameters of the plant. Application of MDI before flowering was the most efficient water-saving technique for commercial production of D. tenuifolia, as it did not reduce plant growth and saved up to 40% of irrigation water. On the other hand, the longest drought period tested (14 days) reduced up to 50% the consumption of irrigation water, but as it resulted in death of most of the plants, even after irrigation was reinstated, its application was inadvisable. Although reducing vegetative growth, SDI before flowering could save water consumption by 70% without affecting flowering and plant attractiveness of D. tenuifolia to pollinators. Therefore, this irrigation regime could be used as sustainable and beneficial when managing the pollinator habitats in agro-ecosystems.

Once estimating the efficiency in attracting pollinators of all studied plants included in the study, choosing the most appropriate plant for efficient mono-specific stands, and optimizing the attractiveness of herbaceous and aromatic mixture, the effect of floral margins (mono-specific or mixed) on pollination and crop yield production was estimated under field conditions. The mixture consisted of the most attractive herbaceous plants chosen in the previous study. This spring blooming mixture was considered appropriate for the experiment as coriander is also a spring blooming crop. The experiment showed that the proximity of both types of floral margins improved quality (i.e. seed weight and seed germination rate) and quantity of coriander seeds. On the other hand, number of seeds produced was significantly higher in the crop growing near mixed margins, than near mono-specific margin, probably due to stronger 100

Chapter VII correlations in pollinator visits to coriander flowers and number of flowers in mixed margins (Ghazoul, 2006). As the experiment was conducted under semi-controlled conditions (potted plants of coriander), it can be concluded that pollinator visits was the main factor that biased the results and that presence of both mixed or mono-specific (D. tenuifolia) margins can improve the crop production for more than 200% and, in addition, conserve the local pollinators in agro-ecosystems. Nevertheless, it is recommendable to link the understanding of insect conservation practices with the existing information on agronomic production of insect-pollinated crops, in order to develop more sustainable agricultural practices. Therefore, further studies should be expanded to a larger scale and carried out with different insect-pollinated crops.

Finally, after evaluating potential of various aromatic and herbaceous plants for conservation of pollinators in agro-ecosystems of Central Spain, it has been found that some plants have higher value as beneficial insectary plants than others, and that some plants are more recommendable to be used for conservation of pollinators than others. However, it is important to apply the appropriate management when introducing the selected beneficial insectary plants in agro-ecosystems in order to provide benefits to both local pollinators and local farmers (by improving the agricultural production in the field through increased rate of pollination).

101

Chapter VIII

VIII. Conclusions

Conclusions on Objective 1:

Although floral morphology, e.g. flower length, flower height, corolla tube length, corolla height and calyx length, did not affect the attractiveness of six aromatic (Lamiaceae) plant species to pollinators, the floral density did, as plant species with higher floral density, i.e. Nepeta tuberosa and Hyssopus officinalis, were characterized as the most attractive plant species.

Of six aromatic (Lamiaceae) plant species, two summer species (Melissa officinalis and Thymbra capitata) did not efficiently contribute to the attractiveness of the floral mixture to pollinators, reducing its attractiveness during the summer period.

Conclusions on Objective 2:

Among twelve herbaceous annual plant species, the flowers of Borago officinalis, Echium plantagineum, Phacelia tanacetifolia and Diplotaxis tenuifolia were the most attractive to bees; while the flowers of Calendula arvensis, Coriandrum sativum, D. tenuifolia and Lobularia maritima were attractive to hoverflies; therefore, D. tenuifolia was the only species with high attractiveness efficiency to both bees and hoverflies.

The floral mixture resulted in lower attractiveness efficiency to pollinators than mono-specific stand of D. tenuifolia, but higher than most of the other mono- specific stands.

Conclusion on Objective 3:

Of six aromatic plant species studied, none showed weedy behaviour under field conditions. In contrast, some of the most attractive herbaceous plant species, such as P. tanacetifolia and C. arvensis, showed potential weediness under field conditions, but their self-seeding was significantly reduced by tillage which could be recommended as a suitable weed management.

102

Chapter VIII

Conclusions on Objective 4:

In controlled greenhouse conditions, plants of D. tenuifolia could be without irrigation for 4 days without affecting its growth, flowering and attractiveness to pollinators. However, lack of irrigation for 8 days or longer significantly reduced their vegetative growth and floral morphology, i.e. number of open flowers, total floral area, flower diameter, corolla tube diameter and corolla tube length.

Regulated deficit irrigation can improve water use efficiency, and depending on the purpose of growing D. tenuifolia, as a crop or as a beneficial plant to attract pollinators, it can reduce water consumption by 40% to 70% without affecting its vegetative and floral development and without reducing its attractiveness to pollinators.

Conclusions on Objective 5:

The proximity of floral margins improved the seed quality (e.g., seed weight and germination rate) of coriander crop plants with regard to the controls without floral margin. In addition, coriander crop production, i.e. number of seeds, was significantly higher in plants grown near floral mix margins than near mono- specific margins.

Since the experiments were conducted under semi-controlled field conditions, it can be concluded that pollinator visits was the main factor that biased the results, and that presence of both mixed or mono-specific (D. tenuifolia) margins can improve the coriander production for more than 200% in small-scale gardens and, in addition, conserve the local pollinators within agricultural fields.

103

Chapter IX

IX. Conclusiones

Conclusiones del Objetivo 1:

Aparentemente, las características morfológicas florales, e.g., diámetro y altura de la flor, diámetro y longitud del tubo de la corola, longitud del cáliz, no afectan significativamente a la atracción de las seis especies aromáticas (Lamiaceae) sobre los polinizadores. Por el contrario, si ha tenido un efecto significativo la densidad floral, resultando las especies con mayor densidad de flores (Nepeta tuberosa e Hyssopus officinalis) las especies más atractivas a polinizadores dentro de este estudio.

De las seis especies perennes aromáticas consideradas, las dos especies con fenología de verano (Melissa officinalis y Thymbra capitata) no han contribuido eficientemente al atractivo de la mezcla floral sobre los polinizadores, debido a la reducción de la atracción total durante el periodo estival.

Conclusiones del Objetivo 2:

De entre las doce especies herbáceas anuales consideradas en este estudio, las flores de Borago officinalis, Echium plantagineum, Phacelia tanacetifolia y Diplotaxis tenuifolia han resultado ser las más atractivas para las abejas, mientras que las flores de Calendula arvensis, Coriandrum sativum, D. tenuifolia y Lobularia maritima han demostrado ser las más atractivas para los sírfidos. Por lo tanto, D. tenuifolia es la única de las especies consideradas con elevada atracción a ambos grupos de polinizadores, abejas y sírfidos.

La mezcla floral ha tenido menor eficiencia de atracción a polinizadores que las parcelas mono-específicas de D. tenuifolia, aunque mayor que el resto de parcelas con cultivos mono-específicos.

Conclusiones del Objetivo 3:

Ninguna de las seis especies aromáticas perennes consideradas en este estudio ha mostrado tendencia invasora. Por el contrario, algunas de las especies herbáceas anuales con mayor atractivo a polinizadores (P. tanacetifolia and C.

104

Chapter IX

arvensis), se han comportado como invasoras en condiciones de campo. Sin embargo, este comportamiento como malas hierbas se redujo significativamente con el laboreo, una labor agronómica habitual que puede recomendarse para prevenir la expansión de estas especies.

Conclusiones del Objetivo 4:

En condiciones controladas de invernadero, se ha comprobado que D. tenuifolia puede mantenerse sin riego durante 4 días sin que ello afecte a su potencial atractivo a los polinizadores. Por otro lado, la falta de riego durante 8 días o más, reduce significativamente su crecimiento vegetativo y sus características morfológicas florales, e.g. número de flores, diámetro y área de la flor, diámetro y longitud del tubo de la corola.

Un Déficit de Riego Regulado puede mejorar la eficiencia del uso del agua, el cual, dependiendo del uso final de la especie D. tenuifolia, como cultivo para el consumo o como planta atractiva a polinizadores, puede reducir el consumo de agua entre un 40 y un 70 % sin que afecte su desarrollo vegetativo y floral, y sin reducir su atracción a polinizadores.

Conclusiones del Objetivo 5:

La proximidad de un margen floral junto al cultivo de cilantro mejora la calidad de sus semillas (peso y tasa de germinación) respecto a las plantas sin margen floral. Por otro lado, los márgenes compuestos por una mezcla floral incrementan el rendimiento del cultivo (numero de semillas) respecto a las parcelas junto a márgenes mono-específicos.

Puesto que los experimentos han sido desarrollados en condiciones de campo semi-controladas, se puede concluir que las visitas de polinizadores han sido el factor determinante en los resultados, y que la presencia de un margen floral (ya sea mono-especifico o de mezcla) puede aumentar la producción a pequeña escala del cultivo de cilantro en más de un 200%, ayudando además a la conservación de los polinizadores en los agro-ecosistemas.

105

Chapter X

X. References

Abrahamczyk, S., Kluge, J., Gareca, Y., Reichle, S. and Kessler, M. (2011) The influence of climatic seasonality on the diversity of different tropical pollinator groups. Plos One, 6, 9. Adhikari, S. and Adhikari, M. K. (2010) Floral phenology and pollination ecology of Punica granatum L. in Kathmandu, Nepal. Nepal Journal of Science and Technology, 11, 115-124. Al-Fattah, M. A. (2001) The ecological behaviour of honeybee foragers on flowers of certain aromatic crops and their role in seed production.The 37th International Apicultural Congress, Durban, South Africa Albrecht, M., Duelli, P., Mueller, C., Kleijn, D. and Schmid, B. (2007) The Swiss agri- environment scheme enhances pollinator diversity and plant reproductive success in nearby intensively managed farmland. Journal of Applied Ecology, 44, 813-822. Altieri, M. A. (1991) How best we can use biodiversity in agroecosystems. Outlook on Agriculture, 20, 15-23. Altieri, M. A. (1999) The ecological role of biodiversity in agroecosystems. Agriculture Ecosystems & Environment, 74, 19-31. Alvarez, S., Banon, S. and Sanchez-Blanco, M. J. (2013) Regulated deficit irrigation in different phenological stages of potted geranium plants: water consumption, water relations and ornamental quality. Acta Physiologiae Plantarum, 35, 1257- 1267. Ambrosino, M. D., Luna, J. M., Jepson, P. C. and Wratten, S. D. (2006) Relative frequencies of visits to selected insectary plants by predatory hoverflies (Diptera: Syrphidae), other beneficial insects, and herbivores. Environmental Entomology, 35, 394-400. Amoros-Jimenez, R., Pineda, A., Fereres, A. and Marcos-Garcia, M. A. (2012) Prey availability and abiotic requirements of immature stages of the aphid predator Sphaerophoria rueppellii. Biological Control, 63, 17-24. Andersson, G. K. S., Rundlof, M. and Smith, H. G. (2012) Organic farming improves pollination success in strawberries. PLoS One, 7. Badenes-Perez, F. R., Nault, B. A. and Shelton, A. M. (2005) Manipulating the attractiveness and suitability of hosts for diamondback moth (Lepidoptera : Plutellidae). Journal of Economic Entomology, 98, 836-844. Baker, H. G. and Baker, I. (1983) In Handbook of experimental pollen biology.(Ed, Jones, C. E. L. R. J.), pp. 117-141. Barbir, J., Badenes-Perez, F. R., Fernandez-Quintanilla, C. and Dorado, J. (2014a) The attractiveness of flowering herbaceous plants to bees (Hymenoptera: Apoidea) and hoverflies (Diptera: Syrphidae) in agro-ecosystems of Central Spain. Agricultural and Forest Entomology. Barbir, J., Dorado, J., Fernández-Quintanilla, C., Blanusa, T., Maksimovic, C. and 106

Chapter X

Badenes-Pérez, F. R. (2014b) Wild rocket – effect of water deficit on growth, flowering and attractiveness to pollinators. Acta Agriculturae Scandinavica, Section B – Soil & Plant Science. Bartomeus, I., Bosch, J. and Vila, M. (2008) High invasive pollen transfer, yet low deposition on native stigmas in a Carpobrotus-invaded community. Annals of Botany, 102, 417-424. Bartomeus, I., Potts, S. G., Steffan-Dewenter, I., Vaissiere, B. E., Woyciechowski, M., Krewenka, K. M., Tscheulin, T., Roberts, S. P. M., Szentgyorgyi, H., Westphal, C. and Bommarco, R. (2014) Contribution of insect pollinators to crop yield and quality varies with agricultural intensification. PeerJ, 2, e328. Bastian, O. (2013) The role of biodiversity in supporting ecosystem services in Natura 2000 sites. Ecological Indicators, 24, 12-22. Batary, P., Andras, B., Kleijn, D. and Tscharntke, T. (2011) Landscape-moderated biodiversity effects of agri-environmental management: a meta-analysis. Proceedings of the Royal Society B-Biological Sciences, 278, 1894-1902. Batary, P., Baldi, A., Saropataki, M., Kohler, F., Verhulst, J., Knop, E., Herzog, F. and Kleijn, D. (2010) Effect of conservation management on bees and insect- pollinated grassland plant communities in three European countries. Agriculture Ecosystems & Environment, 136, 35-39. Begum, M., Gurr, G. M., Wratten, S. D., Hedberg, P. R. and Nicol, H. I. (2006) Using selective food plants to maximize biological control of vineyard pests. Journal of Applied Ecology, 43, 547-554. Bendifallah, L., Louadi, K. and Doumandji, S. (2013) Bee fauna potential visitors of coriander flowers Coriandrum sativum L. (Apiaceae) in the Mitidja area (Algeria). Journal of Apicultural Science, 57, 59-70. Bennett, R. N., Carvalho, R., Mellon, F. A., Eagles, J. and Rosa, E. A. S. (2007) Identification and quantification of glucosinolates in sprouts derived from seeds of wild Eruca sativa L. (salad rocket) and Diplotaxis tenuifolia L. (wild rocket) from diverse geographical locations. Journal of Agricultural and Food Chemistry, 55, 67-74. Benton, T. G., Vickery, J. A. and Wilson, J. D. (2003) Farmland biodiversity: is habitat heterogeneity the key? Trends in Ecology & Evolution, 18, 182-188. Bernhardt, P. (2000) Convergent evolution and adaptive radiation of beetle-pollinated angiosperms. Plant Systematics and Evolution, 222, 293-320. Biesmeijer, J. C., Roberts, S. P. M., Reemer, M., Ohlemuller, R., Edwards, M., Peeters, T., Schaffers, A. P., Potts, S. G., Kleukers, R., Thomas, C. D., Settele, J. and Kunin, W. E. (2006) Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science, 313, 351-354. Billeter, R., Liira, J., Bailey, D., Bugter, R., Arens, P., Augenstein, I., Aviron, S., Baudry, J., Bukacek, R., Burel, F., Cerny, M., De Blust, G., De Cock, R., Diekotter, T., Dietz, H., Dirksen, J., Dormann, C., Durka, W., Frenzel, M., Hamersky, R., Hendrickx, F., Herzog, F., Klotz, S., Koolstra, B., Lausch, A., Le Coeur, D., Maelfait, J. P., Opdam, P., Roubalova, M., Schermann, A., Schermann, N., Schmidt, T., Schweiger, O., Smulders, M. J. M., Speelmans, M.,

107

Chapter X

Simova, P., Verboom, J., van Wingerden, W., Zobel, M. and Edwards, P. J. (2008) Indicators for biodiversity in agricultural landscapes: a pan-European study. Journal of Applied Ecology, 45, 141-150. Bosch, J. (1992) Floral biology and pollinators of three co-occurring Cistus species (Cistaceae). Botanical Journal of the Linnean Society, 109, 39-55. Bosch, J., Gonzalez, A. M. M., Rodrigo, A. and Navarro, D. (2009) Plant-pollinator networks: adding the pollinator's perspective. Ecology Letters, 12, 409-419. Branquart, E. and Hemptinne, J. L. (2000) Selectivity in the exploitation of floral resources by hoverflies (Diptera : Syrphinae). Ecography, 23, 732-742. Brittain, C. and Potts, S. G. (2011) The potential impacts of insecticides on the life- history traits of bees and the consequences for pollination. Basic and Applied Ecology, 12, 321-331. Buide, M. L. (2006) Pollination ecology of Silene acutifolia (Caryophyllaceae): Floral traits variation and pollinator attraction. Annals of Botany, 97, 289-297. Cakir, R. (2004) Effect of water stress at different development stages on vegetative and reproductive growth of corn. Field Crops Research, 89, 1-16. Campbell, A. J., Biesmeijer, J. C., Varma, V. and Wackers, F. L. (2012a) Realising multiple ecosystem services based on the response of three beneficial insect groups to floral traits and trait diversity. Basic and Applied Ecology, 13, 363- 370. Campbell, D. R., Bischoff, M., Lord, J. M. and Robertson, A. W. (2012b) Where have all the blue flowers gone: pollinator responses and selection on flower colour in New Zealand Wahlenbergia albomarginata. Journal of Evolutionary Biology, 25, 352-364. Cane, J. H. (1997) Ground-nesting bees: the neglected pollinator resource for agriculture. Acta Horticulturae (Wageningen), 437, 309-324. Cane, J. H. (2008) A native ground-nesting bee (Nomia melanderi) sustainably managed to pollinate alfalfa across an intensively agricultural landscape. Apidologie, 39, 315-323. Caraballo-Ortiz, M. A., Santiago-Valentin, E. and Carlo, T. A. (2011) Flower number and distance to neighbours affect the fecundity of Goetzea elegans (Solanaceae). Journal of Tropical Ecology, 27, 521-528. Carreck, N. L. and Williams, I. H. (2002) Food for insect pollinators on farmland: insect visits to flowers of annual seed mixtures. Journal of Insect Conservation, 6, 13- 23. Carrie, R. J. G., George, D. R. and Wackers, F. L. (2012) Selection of floral resources to optimise conservation of agriculturally-functional insect groups. Journal of Insect Conservation, 16, 635-640. Carroll, A. B., Pallardy, S. G. and Galen, C. (2001) Drought stress, plant water status, and floral trait expression in fireweed, Epilobium angustifolium (Onagraceae). American Journal of Botany, 88, 438-446. Carrubba, A., la Torre, R., Saiano, F. and Alonzo, G. (2006) Effect of sowing time on coriander performance in a semiarid Mediterranean environment. Crop Science,

108

Chapter X

46, 437-447. Carrubba, A. and Scalenghe, R. (2012) The scent of Mare Nostrum: medicinal and aromatic plants in Mediterranean soils. Journal of the Science of Food and Agriculture, 92, 1150-1170. Carvalheiro, L. G., Kunin, W. E., Keil, P., Aguirre-Gutierrez, J., Ellis, W. N., Fox, R., Groom, Q., Hennekens, S., Van Landuyt, W., Maes, D., Van de Meutter, F., Michez, D., Rasmont, P., Ode, B., Potts, S. G., Reemer, M., Roberts, S. P. M., Schaminee, J., WallisDeVries, M. F. and Biesmeijer, J. C. (2013) Species richness declines and biotic homogenisation have slowed down for NW- European pollinators and plants. Ecology Letters, 16, 870-878. Carvalheiro, L. G., Seymour, C. L., Nicolson, S. W. and Veldtman, R. (2012) Creating patches of native flowers facilitates crop pollination in large agricultural fields: mango as a case study. Journal of Applied Ecology, 49, 1373-1383. Carvell, C., Meek, W. R., Pywell, R. F., Goulson, D. and Nowakowski, M. (2007) Comparing the efficacy of agri-environment schemes to enhance bumble bee abundance and diversity on arable field margins. Journal of Applied Ecology, 44, 29-40. Castroviejo, S., Aedo, C., Cirujano, S., Laínz, M., Montserrat, P., Morales, R., Muñoz Garmendia, F., Navarro, C., Paiva, J., Soriano, C. (1993) Flora iberica 12, Real Jardín Botánico, CSIC, Madrid. Cerekovic, N., Pagter, M., Kristensen, H. L., Pedersen, H. L., Brennan, R. and Petersen, K. K. (2013) Effects of drought stress during flowering of two pot-grown blackcurrant (Ribes nigrum L.) cultivars. Scientia Horticulturae, 162, 365-373. Chambers, R. J. and Adams, T. H. L. (1986) Quantification of the impact of hoverflies (Diptera, Syrphidae) on cereal aphids in winter-wheat - an analysis of field populations. Journal of Applied Ecology, 23, 895-904. Chartzoulakis, K., Patakas, A., Kofidis, G., Bosabalidis, A. and Nastou, A. (2002) Water stress affects leaf anatomy, gas exchange, water relations and growth of two avocado cultivars. Scientia Horticulturae, 95, 39-50. Chaudhary, O. P. and Singh, J. (2007) Diversity, temporal abundance, foraging behaviour of floral visitors and effect of different modes of pollination on coriander (Coriandrum sativum L.). Journal of Spices and Aromatic Crops, 16, 8–14. Chaves, M. M., Maroco, J. P. and Pereira, J. S. (2003) Understanding plant responses to drought - from genes to the whole plant. Functional Plant Biology, 30, 239-264. Christian-Smith, J., Cooley, H. and Gleick, P. H. (2012) Potential water savings associated with agricultural water efficiency improvements: a case study of California, USA. Water Policy, 14, 194-213. Colley, M. R. and Luna, J. M. (2000) Relative attractiveness of potential beneficial insectary plants to aphidophagous hoverflies (Diptera : Syrphidae). Environmental Entomology, 29, 1054-1059. Comba, L., Corbet, S. A., Hunt, L. and Warren, B. (1999) Flowers, nectar and insect visits: Evaluating British plant species for pollinator-friendly gardens. Annals of Botany, 83, 369-383.

109

Chapter X

Conner, J. K. and Rush, S. (1996) Effects of flower size and number on pollinator visitation to wild radish, Raphanus raphanistrum. Oecologia, 105, 509-516. Corbet, S. A., Bee, J., Dasmahapatra, K., Gale, S., Gorringe, E., La Ferla, B., Moorhouse, T., Trevail, A., Van Bergen, Y. and Vorontsova, M. (2001) Native or exotic? Double or single? Evaluating plants for pollinator-friendly gardens. Annals of Botany, 87, 219-232. Cranmer, L., McCollin, D. and Ollerton, J. (2012) Landscape structure influences pollinator movements and directly affects plant reproductive success. Oikos, 121, 562-568. Crowder, D. W., Northfield, T. D., Gomulkiewicz, R. and Snyder, W. E. (2012) Conserving and promoting evenness: organic farming and fire-based wildland management as case studies. Ecology, 93, 2001-2007. Cuevas, J., Canete, M. L., Pinillos, V., Zapata, A. J., Fernandez, M. D., Gonzalez, M. and Hueso, J. J. (2007) Optimal dates for regulated deficit irrigation in 'Algerie' loquat (Eriobotrya japonica Lindl.) cultivated in Southeast Spain. Agricultural Water Management, 89, 131-136. D'Antuono, L. F., Elementi, S. and Neri, R. (2009) Exploring new potential health- promoting vegetables: glucosinolates and sensory attributes of rocket salads and related Diplotaxis and Eruca species. Journal of the Science of Food and Agriculture, 89, 713-722. Dauber, J., Biesmeijer, J. C., Gabriel, D., Kunin, W. E., Lamborn, E., Meyer, B., Nielsen, A., Potts, S. G., Roberts, S. P. M., Sober, V., Settele, J., Steffan- Dewenter, I., Stout, J. C., Teder, T., Tscheulin, T., Vivarelli, D. and Petanidou, T. (2010) Effects of patch size and density on flower visitation and seed set of wild plants: a pan-European approach. Journal of Ecology, 98, 188-196. Dobbs, T. L. and Pretty, J. N. (2004) Agri-environmental stewardship schemes and "multifunctionality". Review of Agricultural Economics, 26, 220-237. Dobson, A., Lodge, D., Alder, J., Cumming, G. S., Keymer, J., McGlade, J., Mooney, H., Rusak, J. A., Sala, O., Wolters, V., Wall, D., Winfree, R. and Xenopoulos, M. A. (2006) Habitat loss, trophic collapse, and the decline of ecosystem services. Ecology, 87, 1915-1924. Eckhart, V. M. (1992) Spatiotemporal variation in abundance and variation in foraging behavior of the pollinators of gynodioecious Phacelia linearis (Hydrophyllaceae). Oikos, 64, 573-586. Elts, J. and Lohmus, A. (2012) What do we lack in agri-environment schemes? The case of farmland birds in Estonia. Agriculture Ecosystems & Environment, 156, 89- 93. Engel, M. S. (1999) The taxonomy of recent and fossil honey bees (Hymenoptera: Apidae; Apis). Journal of Hymenoptera Research, 8, 165-196. Engel, M. S. (2000) Classification of the bee tribe Augochlorini (Hymenoptera: Halictidae). Bulletin of the American Museum of Natural History, 250, 1-89. Engel, M. S. and Dingemansbakels, F. (1980) Nectar and pollen resources for stingless bees (Meliponinae, Hymenoptera) in Surinam (South-America). Apidologie, 11, 341-350.

110

Chapter X

English, M. J., Musich, J. T., Murty, V. V. N. (1990) In Management of Farm Irrigation Systems(Ed, Hoffman, G. J., Howell, T. A., Solomon, K. H.) American Society of Agricultural Engineers, St. Joseph, MI, USA, pp. 631-663. Essenberg, C. J. (2012) Explaining variation in the effect of floral density on pollinator visitation. American Naturalist, 180, 153-166. Fereres, E. and Soriano, M. A. (2007) Deficit irrigation for reducing agricultural water use. Journal of Experimental Botany, 58, 147-159. Fontaine, C., Dajoz, I., Meriguet, J. and Loreau, M. (2006) Functional diversity of plant-pollinator interaction webs enhances the persistence of plant communities. Plos Biology, 4, 129-135. Frund, J., Linsenmair, K. E. and Bluthgen, N. (2010) Pollinator diversity and specialization in relation to flower diversity. Oikos, 119, 1581-1590. Garcia-Tejero, I., Duran-Zuazo, V. H., Arriaga-Sevilla, J. and Muriel-Fernandez, J. L. (2012) Impact of water stress on citrus yield. Agronomy for Sustainable Development, 32, 651-659. Garcia, J. D. D. (2000) Patterns of insect flower visitation in Lavandula bucchii Webb (Lamiaceae), an endemic shrub of Tenerife (Canary Islands). Journal of Natural History, 34, 2145-2155. Garibaldi, L. A., Aizen, M. A., Cunningham, S. and Klein, A. M. (2009) Pollinator shortage and global crop yield: looking at the whole spectrum of pollinator dependency. Communicative & Integrative Biology, 2, 37-39. Ghazoul, J. (2006) Floral diversity and the facilitation of pollination. Journal of Ecology, 94, 295-304. Gill, R. J., Ramos-Rodriguez, O. and Raine, N. E. (2012) Combined pesticide exposure severely affects individual- and colony-level traits in bees. Nature, 491, 105-8. Gomez, J. M., Bosch, J., Perfectti, F., Fernandez, J. D., Abdelaziz, M. and Camacho, J. P. M. (2008) Spatial variation in selection on corolla shape in a generalist plant is promoted by the preference patterns of its local pollinators. Proceedings of the Royal Society B-Biological Sciences, 275, 2241-2249. Goulson, D. (2003) Effects of introduced bees on native ecosystems. Annual Review of Ecology Evolution and Systematics, 34, 1-26. Goulson, D., Lepais, O., O'Connor, S., Osborne, J. L., Sanderson, R. A., Cussans, J., Goffe, L. and Darvill, B. (2010) Effects of land use at a landscape scale on bumblebee nest density and survival. Journal of Applied Ecology, 47, 1207- 1215. Grigulis, K., Sheppard, A. W., Ash, J. E. and Groves, R. H. (2001) The comparative demography of the pasture weed Echium plantagineum between its native and invaded ranges. Journal of Applied Ecology, 38, 281-290. Gurr, G. M., Wratten, S. D. and Luna, J. M. (2003) Multi-function agricultural biodiversity: pest management and other benefits. Basic and Applied Ecology, 4, 107-116. Gutbrodt, B., Mody, K. and Dorn, S. (2011) Drought changes plant chemistry and causes contrasting responses in lepidopteran herbivores. Oikos, 120, 1732-1740.

111

Chapter X

Haaland, C., Naisbit, R. E. and Bersier, L.-F. (2011) Sown wildflower strips for insect conservation: a review. Insect Conservation and Diversity, 4, 60-80. Hall, M. K. D., Jobling, J. J. and Rogers, G. S. (2012) Some perspectives on rocket as a vegetable crop: a review. Vegetable Crops Research Bulletin, 76, 21-41. Heard, M. S., Carvell, C., Carreck, N. L., Rothery, P., Osborne, J. L. and Bourke, A. F. G. (2007) Landscape context not patch size determines bumble-bee density on flower mixtures sown for agri-environment schemes. Biology Letters, 3, 638- 641. Herrera, C. M. (2001) Deconstructing a floral phenotype: do pollinators select for corolla integration in Lavandula latifolia? Journal of Evolutionary Biology, 14, 574-584. Hoffman, G. D. and Hogg, D. B. (1992) Effect of alfalfa water-stress on potato leafhopper (Homoptera, Cicadellidae) plant preference and oviposition rate. Annals of the Entomological Society of America, 85, 506-516. Hogg, B. N., Bugg, R. L. and Daane, K. M. (2011) Attractiveness of common insectary and harvestable floral resources to beneficial insects. Biological Control, 56, 76- 84. Holzschuh, A., Dudenhoffer, J. H. and Tscharntke, T. (2012) Landscapes with wild bee habitats enhance pollination, fruit set and yield of sweet cherry. Biological Conservation, 153, 101-107. Howell, T. A. (2001) Enhancing water use efficiency in irrigated agriculture. Agronomy Journal, 93, 281-289. Hsiao, T. C. (1973) Plant responses to water stress. Annual Review of Plant Physiology and Plant Molecular Biology, 24, 519-570. Huberty, A. F. and Denno, R. F. (2004) Plant water stress and its consequences for herbivorous insects: A new synthesis. Ecology, 85, 1383-1398. Hudewenz, A., Klein, A. M., Scherber, C., Stanke, L., Tscharntke, T., Vogel, A., Weigelt, A., Weisser, W. W. and Ebeling, A. (2012) Herbivore and pollinator responses to grassland management intensity along experimental changes in plant species richness. Biological Conservation, 150, 42-52. Hussain, A. I., Anwar, F., Nigam, P. S., Sarker, S. D., Moore, J. E., Rao, J. R. and Mazumdar, A. (2011) Antibacterial activity of some Lamiaceae essential oils using resazurin as an indicator of cell growth. Lwt-Food Science and Technology, 44, 1199-1206. ILC (2000) Safety and health in agriculture.International Labour Conference, Geneva Inouye, D. W. (1980) The terminology of floral larceny. Ecology, 61, 1251-1253. Jacobs, J. H., Clark, S. J., Denholm, I., Goulson, D., Stoate, C. and Osborne, J. L. (2009) Pollination biology of fruit-bearing hedgerow plants and the role of flower-visiting insects in fruit-set. Annals of Botany, 104, 1397-1404. Jauker, F., Bondarenko, B., Becker, H. C. and Steffan-Dewenter, I. (2012) Pollination efficiency of wild bees and hoverflies provided to oilseed rape. Agricultural and Forest Entomology, 14, 81-87. Jauker, F., Diekotter, T., Schwarzbach, F. and Wolters, V. (2009) Pollinator dispersal in 112

Chapter X

an agricultural matrix: opposing responses of wild bees and hoverflies to landscape structure and distance from main habitat. Landscape Ecology, 24, 547-555. Jauker, F. and Wolters, V. (2008) Hover flies are efficient pollinators of oilseed rape. Oecologia, 156, 819-823. Johnson, B. R. (2010) Division of labor in honey bees: form, function and proximate mechanisms. Behavioral Ecology and Sociobiology, 64, 305-316. Julier, H. E. and Roulston, T. H. (2009) Wild bee abundance and pollination service in cultivated pumpkins: Farm management, nesting behavior and landscape effects. Journal of Economic Entomology, 102, 563-573. Keasar, T., Pollak, G., Arnon, R., Cohen, D. and Shmida, A. (2006) Honesty of signaling and pollinator attraction: The case of flag-like bracts. Israel Journal of Plant Sciences, 54, 119-128. Kells, A. R., Holland, J. M. and Goulson, D. (2001) The value of uncropped field margins for foraging bumblebees. Journal of Insect Conservation, 5, 283-291. Khater, A. M., El-Zakardy, K. A. H. and Ebadah, I. M. A. (2003) The efficiency of honeybees and other insect pollinators in pollination of faba bean (Vicia faba L.). Bulletin of Faculty of Agriculture Cairo University, 54, 465-481. Kleijn, D. and van Langevelde, F. (2006) Interacting effects of landscape context and habitat quality on flower visiting insects in agricultural landscapes. Basic and Applied Ecology, 7, 201-214. Klein, A.-M., Vaissiere, B. E., Cane, J. H., Steffan-Dewenter, I., Cunningham, S. A., Kremen, C. and Tscharntke, T. (2007) Importance of pollinators in changing landscapes for world crops. Proceedings of the Royal Society Biological Sciences Series B, 274, 303-313. Klein, A. M., Steffan-Dewenter, I. and Tscharntke, T. (2003a) Fruit set of highland coffee increases with the diversity of pollinating bees. Proceedings of the Royal Society B-Biological Sciences, 270, 955-961. Klein, A. M., Steffan-Dewenter, I. and Tscharntke, T. (2003b) Pollination of Coffea canephora in relation to local and regional agroforestry management. Journal of Applied Ecology, 40, 837-845. Koul, A. K., Hamal, I. A. and Gupta, S. K. (1989) Pollination mechanism in Coriandrum sativum L. (Apiaceae). Proceedings of the Indian Academy of Sciences-Plant Sciences, 99, 509-515. Krebs, J. R., Wilson, J. D., Bradbury, R. B. and Siriwardena, G. M. (1999) The second silent spring? Nature, 400, 611-612. Lamy, E., Schroder, J., Paulus, S., Brenk, P., Stahl, T. and Mersch-Sundermann, V. (2008) Antigenotoxic properties of Eruca sativa (rocket plant), erucin and erysolin in human hepatoma (HepG2) cells towards benzo(a)pyrene and their mode of action. Food and Chemical Toxicology, 46, 2415-2421. Landis, D. A., Menalled, F. D., Costamagna, A. C. and Wilkinson, T. K. (2005) Manipulating plant resources to enhance beneficial arthropods in agricultural landscapes. Weed Science, 53, 902-908.

113

Chapter X

Landis, D. A., Wratten, S. D. and Gurr, G. M. (2000) Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology, 45, 175-201. LaSalle, J. and Gauld, I. D. (1993) Hymenoptera: their diversity, and their impact on the diversity of other organisms, CAB International. Le Feon, V., Schermann-Legionnet, A., Delettre, Y., Aviron, S., Billeter, R., Bugter, R., Hendrickx, F. and Burel, F. (2010) Intensification of agriculture, landscape composition and wild bee communities: A large scale study in four European countries. Agriculture Ecosystems & Environment, 137, 143-150. Luo, H. J., Zheng, Z. B., Luo, S., Pan, Y. S. and Liu, X. H. (2007) In Proceedings of the Second International Symposium on Loquat(Eds, Huang, X. M. and Janick, J.) International Society Horticultural Science, Leuven 1, pp. 417-422. Macukanovic-Jocic, M., Stevanovic, Z. D., Mladenovic, M. and Jocic, G. (2011) Flower morphophysiology of selected Lamiaceae species in relation to pollinator attraction. Journal of Apicultural Research, 50, 89-101. Maloof, J. E. (2001) The effects of a bumble bee nectar robber on plant reproductive success and pollinator behavior. American Journal of Botany, 88, 1960-1965. Markussen, M. V., Kulak, M., Smith, L. G., Nemecek, T. and Ostergard, H. (2014) Evaluating the sustainability of a small-scale low-input organic vegetable supply system in the United Kingdom. Sustainability, 6, 1913-1945. Maroufi, K., Farahani, H. A. and Darvishi, H. H. (2010) Importance of coriander (Coriandrum sativum L.) between the medicinal and aromatic plants. Advances in Environmental Biology, 4, 433-436. Marshall, E. J. P. (2005) Field margins in northern Europe: Integrating agricultural, environmental and biodiversity functions, Canadian Weed Science Soc, Soc Canadienne Malherbologie, Quebec. Marshall, E. J. P. and Moonen, A. C. (2002) Field margins in northern Europe: their functions and interactions with agriculture. Agriculture Ecosystems & Environment, 89, 5-21. Martin Gonzalez, A. M., Dalsgaard, B., Ollerton, J., Timmermann, A., Olesen, J. M., Andersen, L. and Tossas, A. G. (2009) Effects of climate on pollination networks in the West Indies. Journal of Tropical Ecology, 25, 493-506. McDonald, R. I. and Girvetz, E. H. (2013) Two challenges for u. S. Irrigation due to climate change: Increasing irrigated area in wet states and increasing irrigation rates in dry states. Plos One, 8, 10. Memmott, J., Carvell, C., Pywell, R. F. and Craze, P. G. (2010) The potential impact of global warming on the efficacy of field margins sown for the conservation of bumble-bees. Philosophical Transactions of the Royal Society B, 365, 2071- 2079. Menz, M. H. M., Phillips, R. D., Winfree, R., Kremen, C., Aizen, M. A., Johnson, S. D. and Dixon, K. W. (2011) Reconnecting plants and pollinators: challenges in the restoration of pollination mutualisms. Trends in Plant Science, 16, 4-12. Michener, C. D. (2007) The bees of the world, Johns Hopkins University Press.

114

Chapter X

Miller, R., Owens, S. J. and Rorslett, B. (2011) Plants and colour: Flowers and pollination. Optics and Laser Technology, 43, 282-294. Mingeau, M., Perrier, C. and Ameglio, T. (2001) Evidence of drought-sensitive periods from flowering to maturity on highbush blueberry. Scientia Horticulturae, 89, 23-40. Miñarro, M. and Prida, E. (2013) Hedgerows surrounding organic apple orchards in north-west Spain: potential to conserve beneficial insects. Agricultural and Forest Entomology, 15, 382-390. Moore, S. R. and Lawrence, K. S. (2013) The effect of soil texture and irrigation on Rotylenchulus reniformis and cotton. Journal of Nematology, 45, 99-105. Morandin, L. A. and Kremen, C. (2013) Bee preference for native versus exotic plants in restored agricultural hedgerows. Restoration Ecology, 21, 26-32. Morris, C. and Potter, C. (1995) Recruiting the new conservationists - farmers adoption of agri-environmental schemes in the uk. Journal of Rural Studies, 11, 51-63. Morse, R. A. and Calderone, N. W. (2000) In Bee Culture Magazine Vol. 128, pp. 1-14. Nadeem, M., Anjum, F. M., Khan, M. I., Tehseen, S., El-Ghorab, A. and Sultan, J. I. (2013) Nutritional and medicinal aspects of coriander (Coriandrum sativum L.) A review. British Food Journal, 115, 743-755. New, T. R. (2005) Invertebrate conservation and agricultural ecosystems, Cambridge University Press. Nicholls, C. I. and Altieri, M. A. (2013) Plant biodiversity enhances bees and other insect pollinators in agroecosystems. A review. Agronomy for Sustainable Development, 33, 257-274. Olson, D. M. and Wackers, F. L. (2007) Management of field margins to maximize multiple ecological services. Journal of Applied Ecology, 44, 13-21. Perez-Pastor, A., Ruiz-Sanchez, M. C. and Domingo, R. (2014) Effects of timing and intensity of deficit irrigation on vegetative and fruit growth of apricot trees. Agricultural Water Management, 134, 110-118. Perez-Perez, J. G., Robles, J. M. and Botia, P. (2014) Effects of deficit irrigation in different fruit growth stages on 'Star Ruby' grapefruit trees in semi-arid conditions. Agricultural Water Management, 133, 44-54. Petanidou, T. and Vokou, D. (1993) Pollination ecology of Labiatae in a phryganic (East Mediterranean) ecosystem. American Journal of Botany, 80, 892-899. Petit, S., Rubbo, N. and Schumann, R. (2011) Nectar collected with microcapillary tubes is less concentrated than total nectar in flowers with small nectar volumes. Australian Journal of Botany, 59, 593-599. Pfiffner, L. and Wyss, E. (2004) Use of sown wildflower strips to enhance natural enemies of agricultural pests, CSIRO Publishing. Philpott, S. M., Uno, S. and Maldonado, J. (2006) The importance of ants and high- shade management to coffee pollination and fruit weight in Chiapas, Mexico. Biodiversity and Conservation, 15, 487-501. Pineda, A. and Marcos-Garcia, M. A. (2008) Seasonal abundance of aphidophagous

115

Chapter X

hoverflies (Diptera : Syrphidae) and their population levels in and outside Mediterranean sweet pepper greenhouses. Annals of the Entomological Society of America, 101, 384-391. Pontin, D. R., Wade, M. R., Kehrli, P. and Wratten, S. D. (2006) Attractiveness of single and multiple species flower patches to beneficial insects in agroecosystems. Annals of Applied Biology, 148, 39-47. Potts, S. G., Biesmeijer, J. C., Kremen, C., Neumann, P., Schweiger, O. and Kunin, W. E. (2010) Global pollinator declines: trends, impacts and drivers. Trends in Ecology & Evolution, 25, 345-353. Power, A. G. (2010) Ecosystem services and agriculture: tradeoffs and synergies. Philosophical Transactions of the Royal Society B-Biological Sciences, 365, 2959-2971. Power, E. F. and Stout, J. C. (2011) Organic dairy farming: impacts on insect-flower interaction networks and pollination. Journal of Applied Ecology, 48, 561-569. Putievsky, E. (1980) Germination studies with seed of caraway, coriander and dill. Seed Science and Technology, 8, 245-254. Pywell, R. F., Meek, W. R., Hulmes, L., Hulmes, S., James, K. L., Nowakowski, M. and Carvell, C. (2011) Management to enhance pollen and nectar resources for bumblebees and butterflies within intensively farmed landscapes. Journal of Insect Conservation, 15, 853-864. Rader, R., Reilly, J., Bartomeus, I. and Winfree, R. (2013) Native bees buffer the negative impact of climate warming on honey bee pollination of watermelon crops. Global Change Biology, 19, 3103-3110. Rands, S. A. and Whitney, H. M. (2011) Field margins, foraging distances and their impacts on nesting pollinator success. Plos One, 6, e25971, 1-8. Rasmont, P. and Iserbyt, I. (2010) In Atlas HymenopteraMons, Gembloux. Rasmussen, J. (1991) A model for prediction of yield response in weed harrowing. Weed Research, 31, 401-408. Reith, M., Baumann, G., Classen-Bockhoff, R. and Speck, T. (2007) New insights into the functional morphology of the lever mechanism of Salvia pratensis (Lamiaceae). Annals of Botany, 100, 393-400. Reith, M., Classen-Bockhoff, R. and Speck, T. (2006) Biomechanics of salvia flowers: The role of lever and flower tube in specialization on pollinators, Crc Press- Taylor & Francis Group, Boca Raton. Richards, A. J. (2001) Does low biodiversity resulting from modern agricultural practice affect crop pollination and yield? Annals of Botany, 88, 165-172. Rundlof, M., Nilsson, H. and Smith, H. G. (2008) Interacting effects of farming practice and landscape context on bumblebees. Biological Conservation, 141, 417-426. Sandrock, C., Tanadini, L. G., Pettis, J. S., Biesmeijer, J. C., Potts, S. G. and Neumann, P. (2014) Sublethal neonicotinoid insecticide exposure reduces solitary bee reproductive success. Agricultural and Forest Entomology, 16, 119-128. Santos, S. A. B. d., Roselino, A. C. and Bego, L. R. (2008) Pollination of cucumber, Cucumis sativus L. (Cucurbitales: Cucurbitaceae), by the stingless bees 116

Chapter X

Scaptotrigona aff. depilis moure and Nannotrigona testaceicornis Lepeletier (Hymenoptera: Meliponini) in greenhouses. Neotropical Entomology, 37, 506- 12. Saramago, I. S., Regato, M. A., Guerreiro, I. M. and Regato, J. E. (2012) In Xxviii International Horticultural Congress on Science and Horticulture for People, Vol. 924 (Eds, Tous, J., Gucci, R. and Fevereiro, P.) Int Soc Horticultural Science, Leuven 1, pp. 255-259. Scriven, L. A., Sweet, M. J. and Port, G. R. (2013) Flower density is more important than habitat type for increasing flower visiting insect diversity. International Journal of Ecology, 2013, 1-12. Scheper, J., Holzschuh, A., Kuussaari, M., Potts, S. G., Rundlof, M., Smith, H. G. and Kleijn, D. (2013) Environmental factors driving the effectiveness of European agri-environmental measures in mitigating pollinator loss - a meta-analysis. Ecology Letters, 16, 912-20. Schiestl, F. P. and Johnson, S. D. (2013) Pollinator-mediated evolution of floral signals. Trends in Ecology & Evolution, 28, 307-315. Schuster, W. S. and Monson, R. K. (1990) An examination of the advantages of C3-C4 intermediate photosynthesis in warm environmnets. Plant Cell and Environment, 13, 903-912. Showler, A. and Moran, P. (2003) Effects of drought stressed cotton, Gossypium hirsutum L., on beet armyworm, Spodoptera exigua (Hubner), oviposition, and larval feeding prefernces and growth. Journal of Chemical Ecology, 29, 1997- 2011. SIAR (2014) Ministerio de agricultura, alimentación y madio ambiente, http://eportal.magrama.gob.es/websiar/SeleccionParametrosMap.aspx?dst=1. Sinacori, A., Carrubba, A., Oliveri, E. and Sinacori, M. (2014) Api e pronubi selvatici nella produzione di Coriandrum sativum L. (Famiglia Apiaceae) in Sicilia occidentale. Avenue media; IT. Smith, H. A. and Chaney, W. E. (2007) A survey of syrphid predators of Nasonovia ribisnigri in organic lettuce on the Central Coast of California. Journal of Economic Entomology, 100, 39-48. Song, B. Z., Wu, H. Y., Kong, Y., Zhang, J., Du, Y. L., Hu, J. H. and Yao, Y. C. (2010) Effects of intercropping with aromatic plants on the diversity and structure of an arthropod community in a pear orchard. Biocontrol, 55, 741-751. Steffan-Dewenter, I., Potts, S. G. and Packer, L. (2005) Pollinator diversity and crop pollination services are at risk. Trends in Ecology & Evolution, 20, 651-652. Steffan-Dewenter, I. and Tscharntke, T. (2000) Resource overlap and possible competition between honey bees and wild bees in central Europe. Oecologia, 122, 288-296. Steingrover, E. G., Geertsema, W. and van Wingerden, W. (2010) Designing agricultural landscapes for natural pest control: a transdisciplinary approach in the Hoeksche Waard (The Netherlands). Landscape Ecology, 25, 825-838. Stokstad, E. (2013) Pesticides under fire for risks to pollinators. Science, 340, 674-676.

117

Chapter X

Stout, J. C., Allen, J. A. and Goulson, D. (1998) The influence of relative plant density and floral morphological complexity on behaviour of bumblebees. Oecologia, 117, 543-550. Stout, J. C. and Goulson, D. (2002) The influence of nectar secretion rates on the responses of bumblebees (Bombus spp.) to previously visited flowers. Behavioral Ecology and Sociobiology, 52, 239-246. Stout, J. C. and Morales, C. L. (2009) Ecological impacts of invasive alien species on bees. Apidologie, 40, 388-409. Stubbs, A. E. and Falk, S. (2002) British hoverflies: an illustrated identification guide. Second edition, British Entomological and Natural History Society. Subba Reddi, C. and Meera Bai, G. (1984) Butterflies and pollination biology. Proceedings of the Indian Academy of Sciences Animal Sciences, 93, 391-396. Tinkeu, L. S. N., VanImpe, G., Hance, T. and Univ, G. (1996) In 48th International Symposium on Crop Protection, Pts I-Iv, Vol. 61, pp. 917-925. Tscharntke, T., Klein, A. M., Kruess, A., Steffan-Dewenter, I. and Thies, C. (2005) Landscape perspectives on agricultural intensification and biodiversity - ecosystem service management. Ecology Letters, 8, 857-874. Tuell, J. K., Fiedler, A. K., Landis, D. and Isaacs, R. (2008) Visitation by wild and managed bees (Hymenoptera : Apoidea) to eastern US native plants for use in conservation programs. Environmental Entomology, 37, 707-718. Ueno, O., Wada, Y., Wakai, M. and Bang, S. W. (2006) Evidence from photosynthetic characteristics for the hybrid origin of Diplotaxis muralis from a C-3-C-4 intermediate and a C-3 species. Plant Biology, 8, 253-259. Van der Sluijs, J. P., Simon-Delso, N., Goulson, D., Maxim, L., Bonmatin, J. M. and Belzunces, L. P. (2013) Neonicotinoids, bee disorders and the sustainability of pollinator services. Current Opinion in Environmental Sustainability, 5, 293- 305. Vanbergen, A. J. and Insect Pollinators Initiative (2013) Threats to an ecosystem service: pressures on pollinators. Frontiers in Ecology and the Environment, 11, 251-259. Walther-Hellwig, K., Fokul, G., Frankl, R., Buchler, R., Ekschmitt, K. and Wolters, V. (2006) Increased density of honeybee colonies affects foraging bumblebees. Apidologie, 37, 517-532. Wangensteen, H., Samuelsen, A. B. and Malterud, K. E. (2004) Antioxidant activity in extracts from coriander. Food Chemistry, 88, 293-297. Watling, J. I. and Donnelly, M. A. (2006) Fragments as islands: a synthesis of faunal responses to habitat patchiness. Conservation Biology, 20, 1016-1025. Winfree, R., Aguilar, R., Vazquez, D. P., LeBuhn, G. and Aizen, M. A. (2009) A meta- analysis of bees' responses to anthropogenic disturbance. Ecology, 90, 2068- 2076. Woodcock, B. A., Bullock, J. M., Nowakowski, M., Orr, R., Tallowin, J. R. B. and Pywell, R. F. (2012) Enhancing floral diversity to increase the robustness of grassland beetle assemblages to environmental change. Conservation Letters, 5,

118

Chapter X

459-469. Woodcock, B. A., Edwards, M., Redhead, J., Meek, W. R., Nuttall, P., Falk, S., Nowakowski, M. and Pywell, R. F. (2013) Crop flower visitation by honeybees, bumblebees and solitary bees: Behavioural differences and diversity responses to landscape. Agriculture Ecosystems & Environment, 171, 1-8. Wratten, S. D., Gillespie, M., Decourtye, A., Mader, E. and Desneux, N. (2012) Pollinator habitat enhancement: Benefits to other ecosystem services. Agriculture Ecosystems & Environment, 159, 112-122. Zhang, Y. W., Robert, G. W., Wang, Y. and Guo, Y. H. (2007) Nectar robbing of a carpenter bee and its effects on the reproductive fitness of Glechoma longituba (Lamiaceae). Plant Ecology, 193, 1-13. Zihlif, M., Afifi, F., Abu-Dahab, R., Majid, A., Somrain, H., Saleh, M. M., Nassar, Z. D. and Naffa, R. (2013) The antiangiogenic activities of ethanolic crude extracts of four Salvia species. Bmc Complementary and Alternative Medicine, 13. Znidarcic, D., Ban, D. A. and Sircelj, H. (2011) Carotenoid and chlorophyll composition of commonly consumed leafy vegetables in Mediterranean countries. Food Chemistry, 129, 1164-1168.

119

Chapter XI

XI. Appendices

1. List of abbreviations

AES – Agri-environmental Schemes

Ali – Alium schoenoprasum

Ant – Antirrhinum majus

ArcGIS – Arc Geographic Information System

Bor – Borago officinalis

Cal – Calendula arvensis

CATPCA – Categorical Partial Component Analysis

Cen – Centaurea cyanus

Cor – Coriandrum sativum

CSIC – Consejo Superior de Investigaciones Científicas

D – Drought

D-11 – Drought of 11 days

D-14 – Drought of 14 days

D-4 – Drought of 4 days

D-8 – Drought of 8 days

DCA – Discriminate Canonical Analysis

Dpl – Diplotaxis tenuifolia

Ech – Echium plantagineum

ETp – Potential Evapotranspiration

GLM – General Linear Model gs – Stomatal Conductance

Lob – Lobularia maritima 121

Chapter XI

M – Mixture of floral species

MASL – Meters above Sea Level

MDI – Moderate Deficit Irrigation

MDI-BF – Moderate Deficit Irrigation – before flowering

MDI-DF – Moderate Deficit Irrigation – during flowering

NE-SW – North-East – South-West

Pha – Phacelia tanacetifolia

R19 – recover of 19 days after drought treatments applied

RDI – Regulated Deficit Irrigation

RDI-BF – Regulated Deficit Irrigation – before flowering

RDI-DF – Regulated Deficit Irrigation – during flowering

RTK-GPS – Real Time Kinematic-the Global Positioning System

SDI – Severe Deficit Irrigation

SDI-BF – Severe Deficit Irrigation – before flowering

SDI-DF – Severe Deficit Irrigation – during flowering

SE – Standard Error

SN/inf – Seed Number per Inflorescence sp. – species spp. – species (plural)

SPSS – Statistical Package for the Social Sciences

SW – Seed Weight

Tag – Tagetes patula

UK – The United Kingdom

USA – The United States of America

122

Chapter XI

2. List of figures

I General introduction

Fig. I-1. Percentage of agricultural land relative to the total land area across the Earth Fig. I-2. Different groups of pollinators [1. Honey bee (Apidae: Apis sp.); 2. Bumblebee (Apidae: Bombus sp.); 3. Carpenter bee (Apidae: Xylocopa sp.); 4. Digger bee (Apidae: Amegilla sp.); 5. Leafcutter bee (Megachilidae: Megachile sp.); 6. Mining bee (Andrenidae: Andrena sp.; 7. Sweat bee (Halictidae: Halictus sp.; 8. Hoverfly (Syrphidae: Sphaerophoria scripta.); 9. Beetle (Coleoptera: Heliotaurus ruficollis)]

Fig. I-4. Sowed field margins: 1. Mono-specific field margin (Phacelia tanacetifolia Benth.), 2. Mixed field margin

Fig. I-5: Use of beneficial insectary plants in small-scale gardening

III Functionality of aromatic (Lamiaceae) floral mixture in conservation of pollinators in Central Spain

Fig. III-1. Experimental design (randomized block design - each number represents one Lamiaceae plant species (mono-specific plot), while M1 and M2 represent mixed plots)

Fig. III-2. Discrimination among plant species upon the first and second canonical axis

Fig. III-3. Relationship between morphological characteristics and number of insect visits in 2012 (A) and in 2013 (B)

Fig. III-4. Relationship between floral phenology and insect visitation

IV The attractiveness of flowering herbaceous plants to bees (Hymenoptera: Apoidea) and hoverflies (Diptera: Syrphidae) in agro-ecosystems of Central Spain

Fig. IV-1. Attractiveness efficiency of flower species. Schematic diagram depicting the relationship between the total number of insect visits/minute (y-axis) and flowering

123

Chapter XI duration (number of observation days) of plant species (x-axis): A (upper-right corner): high total number of insect visits/minute in a long observation period (high attractiveness efficiency); B (upper-left corner): high total number of insect visits/minute in a short observation period; C (bottom-right corner): low total number of insect visits/minute in a long observation period; and D (bottom-left corner): low total number of insect visits/minute in a short observation period

Fig. IV-2. Visits of bees, hoverflies and beetles to the flowers of the studied plant species. Means with same letter in each graph are not significantly different at P ≤ 0.05, as determined by Tukey’s B multiple comparison test. Abbreviations: 2011 (1): Bor= Borago officinalis; Cal= Calendula arvensis; Cen= Centaurea cyanus; Dpl= Diplotaxis tenuifolia; Pha= Phacelia tanacetifolia; Ali= Alium schoenoprasum; M1= mixture of those species. 2011 (2): Cor= Coriandrum sativum; Ech= Echium plantagineum; Lob= Lobularia maritima; Tag= Tagetes patula; Ant= Antirrhinum majus; M2= mixture of those species. 2012: M= mixture of the species used in 2012

Fig. IV - 3. Attractiveness efficiency of the studied plant species for bees and hoverflies in 2011 and 2012 (abbreviations of plant species as in Fig. IV-2.). Scatter-plots depict the relationship between the total number of insect visits/minute (y-axis) and flowering duration (number of observation days) of plant species (x-axis). Bars represent the standard deviation from 3 replicated plots

V Wild rocket - effect of water deficit on growth, flowering and attractiveness to pollinators

Fig. V-1. Mean stomatal conductance of D. tenuifolia under different drought treatments (no irrigation for 4 [D-4], 8 [D-8], 11 [D-11] or 14 days [D-14]) (A) and during a recovery period (well-irrigated) of 19 days (B). For each measurement (day), means with different letter are significantly different at P ≤ 0.05, as determined by the Tukey’s B multiple comparison test.

Fig. V-2. Concentration of sugars in nectar after different drought treatments (no irrigation for 4 [D-4], 8 [D-8], 11 [D-11] or 14 days [D-14]) (A) and after re-watering of 19 days (B). No significant differences were found between drought treatments. The

124

Chapter XI vertical bars represent standard errors. Concentrations of sugars in D-14 after re- watering are measured in only one plant that recovered after drought stress.

Fig. V-3. Number of pollinators visiting flowers of D. tenuifolia after different drought treatments (no irrigation for 4 [D-4], 8 [D-8] and 11 days [D-11]) in two sites of Central Spain: La Poveda (A) and La Sierra (B). For each treatment, means that are significantly different at P ≤ 0.05 are marked with different letter (comparison within insect groups), as determined by the Tukey’s B multiple comparison test. The vertical bars represent standard errors.

Fig. V-4. Mean stomatal conductance of D. tenuifolia under moderate deficit irrigation (MDI) and severe deficit irrigation (SDI). Deficit irrigation applied before and during flowering (A) and deficit irrigation applied only during flowering (B). For each measurement, means with different letter are significantly different at P ≤ 0.05, as determined by the Tukey’s B multiple comparison test.

Fig. V-5. Number of pollinators visiting flowers of D. tenuifolia after moderate deficit irrigation (MDI) and severe deficit irrigation (SDI). Deficit irrigation applied before and during flowering (A) and deficit irrigation applied only during flowering (B). For each treatment, means with different letter are significantly different at P ≤ 0.05 (comparison within insect groups), as determined by Tukey’s B multiple comparison test. The vertical bars represent standard errors.

VI Can proximity of floral field margins improve pollination and seed production in coriander, Coriandrum sativum L. (Apiaceae)?

125

Chapter XI

3. List of tables

III Functionality of aromatic (Lamiaceae) floral mixture in conservation of pollinators in Central Spain

Table III-1. Aromatic plant species used

Table III-2. Mean (± SE) insect visits per minute to flowers of selected aromatic plant species in 2012 and 2013

Table III-3. Morphological characteristics of flowers of the six Lamiaceae species studied

Table III-4. Correlation between discriminating variables and 1st and 2nd standardized canonical discriminant functions (80.04%)

Table III-5. Spatial increase (%) of the area occupied by plant species one year after planting

IV The attractiveness of flowering herbaceous plants to bees (Hymenoptera: Apoidea) and hoverflies (Diptera: Syrphidae) in agro-ecosystems of Central Spain

Table IV-1. Plant species and their flowering periods in 2011 and 2012

Table IV-2. List of insects visiting the studied plant species

Table IV-3. Differences in seedling emergence (Nº plants/m2) after applying two different managements at the end of plants cycle: shallow tillage and no-tillage

Table IV-4. Spatial increase (%) of the area occupied by plant species one year after planting

V Wild rocket - effect of water deficit on growth, flowering and attractiveness to pollinators

Table V-1. Effects of different drought treatments (no irrigation for 4 [D-4], 8 [D-8], 11 [D-11] or 14 days [D-14]) and of re-watering over 19 days on morphological characteristics of D. tenuifolia plants

126

Chapter XI

Table V-2. Effect of different drought treatments (no irrigation for 4 [D-4], 8 [D-8], 11 [D-11] or 14 days [D-14]) and of re-watering over 19 days on the floral characteristics of D. tenuifolia

VI Can proximity of floral field margins improve pollination and seed production in coriander, Coriandrum sativum L. (Apiaceae)?

Table VI-1: Mean (± S.E.) flower density and mean (± S.E.) number of visits of pollinators to the flowers of mono-specific and mixed floral margins

Table VI-3. Correlations between number of flowers and number of pollinators' visits/min

127

Chapter XI

128