Plant and insect-pollinator interactions in agricultural and semi-natural landscapes of southern irrigated zone of Punjab, Pakistan

By

Muhammad Asif Sajjad M.Sc. (Hons) Agri. Entomology

Ph.D Thesis

IN AGRICULTURAL ENTOMOLOGY

Department of Entomology Faculty of Agricultural Sciences and Technology BAHAUDDIN ZAKARIYA UNIVERSITY MULTAN PAKISTAN 2012

Plant and insect-pollinator interactions in agricultural and semi-natural landscapes of southern irrigated zone of Punjab, Pakistan

A thesis submitted in fulfillment of the requirement for the degree of

Doctor of Philosophy

In

Agricultural Entomology

By

MUHAMMAD ASIF SAJJAD M.Sc. (Hons) Agri. Entomology Regd. No. 99-ag-37

Department of Entomology Faculty of Agricultural Sciences and Technology Bahauddin Zakariya University Multan,Pakistan 2012

To

The Controller Examinations, Bahauddin Zakariya University Multan.

I find the thesis entitled “Plant and insect-pollinator interactions in agricultural and semi-natural landscapes of southern irrigated zone of Punjab, Pakistan” submitted by Mr. Muhammad Asif Sajjad, satisfactory and recommend that it be processed for evaluation by the external examiner(s) for the award of PhD degree.

Supervisor ______Dr. Shafqat Saeed Assistant Professor

Statement and declaration

The work submitted in this thesis under the title “Plant and insect-pollinator interactions in agricultural and semi-natural landscapes of southern irrigated zone of Punjab, Pakistan” is fulfillment of the requirements for the degree of Doctor of Philosophy.

I declare that this work is the result of my own investigations and has not already been accepted in substance for any degree, nor it is currently being submitted for any other degree.

All authors work referred to in this thesis has been fully acknowledged.

______Muhammad Asif Sajjad (Candidate)

I certify that the above statement is correct. ______

Dr. Shafqat Saeed Assistant Professor (Supervisor) Acknowledgement

All praises to Almighty Allah alone, the Omnipotent, the Omnipresent, the most merciful and compassionate for the best owing upon me the sense of inquiry and requisite potential for the successful accomplishment of this piece of research work. My special praise to Holy Prophet, Muhammad (peace be upon him) who is the most perfect and forever a torch of guidance for humanity as a whole. I feel great pleasure and honour to express my heartiest and sincerest sense of gratitude and appreciations to my supervisor, Dr. Shafqat Saeed, Assistant Professor, Department of Entomology, Faculty of Agricultural Sciences and Technology Bahauddin Zakariya University Multan,Pakistan I am highly thankful to Dr. Nazim Hussain, Chairman Department of Entomology, for his cooperation, philanthropic behavior and encouragement throughout my studies for his technical and moral support for the execution of research described in the dissertation. I am highly indebted to (late) Dr. Mumtaz H. Bukhari, Professor of Botany, for his help in identification of plant species, Dr. Claus Claussen (Germany) for Syrphid fly species, Dr. Ather Rafi (NARC, Islamabad) for butterfly species, Dr. Andrew Whittington (Scotland) for fly species and Dr. Terry Grisworld (USA) for bee species. I found no words to express my gratitude and profound admirations to my cherished and affectionate parents, brother and sisters for their prays, spiritual and intellectual inspiration, financial support and encouraging attitude to carry me at the destination and for their stretching hands of sincerity towards me to achieve higher goals of my life. I don’t find words to thank my wife for her noble sacrifice, prays and endless cooperation. Without her love, encouragement and understanding, this endeavor would not have been possible. I am greatly indebted to my Ph.D. Colleagues Mr. Asad Masood, Mr. Naeem Iqbal, Mr. Mudssar Ali, Mr. Wali Muhammad for their amicable attitude and cooperation for my research work. The study was funded by Higher Education Commission of Pakistan as an Indigenous PhD fellowship.

MUHAMMAD ASIF SAJJAD

Dedications

I dedicate this dissertation

To

My parents

Contents Page #

Chapter 1 Summary 1

Chapter 2 Introduction 5

Chapter 3 Pollination systems based on frequencies of insect visiting guilds 13

Chapter 4 A yearlong plant-pollinator network 40

Chapter 5 Variation in abundance and composition of Hoverflies (Diptera: 60 Syrphidae) around the year

Chapter 6 Effect of habitat on pollinator biodiversity and plant reproductive 78 success

Chapter 7 Identification of potential native crop pollinators. 95

References 115

List of Publications 136

Copy of Published papers 137

i

List of Tables Chapter 3 Page #

Table 1 Plant species, their floral characters, sampling efforts and floral visitors 34 summary. Table 2 Abundance of 9 functional groups of pollinators on 150 plant species. 38 Chapter 4

Table 1 Fitted models to the distribution of interactions per species for the plant- 52 flower visitor network of the southern Punjab, Pakistan. Chapter 5

Table 1 Syrphid fly species and their abundance on flowering plants at Bahauddin 66 Zakariya University Campus Multan, Pakistan during January to December, 2008.

Table 2 Monthly syrphid fly abundance (number of census) on flowering plants 76 from January to December, 2008 at Bahauddin Zakariya University Campus Multan, Pakistan. Chapter 6

Table 1 Properties and types of three studied natural reserves at southern Punjab, 82 Pakistan.

Table 2 Among-location differences in pollinator abundance and diversity. Means 85 followed by different letters are statistically different at α 0.05.

Table 3 Comparison of reproductive success of three locations. The mean sharing 91 similar letters are statistically similar at α 0.05.

Table 4 Insect visiting P. juliflora during April to June, 2009 in three selected 92 locations along with percentage of visits (relative to the total number of visits in a particular location). Chapter 7

Table 1 List of floral visitors of onion (A. cepa L.) in Multan, Pakistan. 104

Table 2 Foraging behavior (stay time/visit/umbel; Mean + SD; n= 8 for each 106 treatment) and average abundance of different species of insects on flowers of A. cepa L. Mean values sharing similar letters in respective columns show non-significant differences (P<0.05) by using LSD test.

Table 1b Yield attributing components of S. sesban ( Mean + SD; n=12 for each 109 treatment) Mean values sharing similar letters in respective columns show non-significant differences (P<0.05) by using LSD test.

ii

Table 2b Average pollinator attendance, visitation index (IVR), and visitation 109 behavior of different insect pollinators.

Table 3b Linear regession analysis between Number of flowers 110 visited/inflorescence/visit and Time spent (seconds)/inflorescence/visit.

iii

List of Figures Chapter 2 Page #

Flow diagram depicting the various processes that are thought to be Fig 1 11 important in plant and pollinator interactions at local or landscape levels.

Chapter 3 Fig 1 The ordination of floral visitors groups according to their visiting profiles 21 (Kruskal's stress-1= 0.232). Note: (Bee S= Short tongue bees, Bees L= Long tongue bees). Fig 2 The ordination of plant species according to their flower visitor 22 frequencies (Kruskal’s stress-1 = 0.276). Binomial codes for floral morphology: uppercase (Italic, normal) + uppercase (Italic, normal) = Umbel, upper case (bold) + lowercase (normal) = Gullet, uppercase (bold) + uppercase (normal) = Flag, uppercase (normal) + uppercase (bold) = Tube, uppercase (normal) + lowercase (normal) = brush, uppercase (normal) + uppercase (normal) = Head, uppercase (bold) + uppercase (bold) = Bowl, lowercase (normal) + lowercase (normal) = Disc, lowercase (Italic) + lowercase (Italic) = Dish shaped. Zygomorphic flowers are underlined and flowers with nectar guides are in parenthesis. Binomial codes for plant species are given in table 1. Fig 3 Dendrogram of 9 functional groups of the cluster analysis with Bray and 24 Curtis distance, based on their visiting frequencies on 55 plant species. The dotted line is the cut off point for determining the separate clusters.

Fig 4 Dendrogram of 55 plant species of the cluster analysis with Bray and 26 Curtis distance, based on floral visitor frequencies, the dotted line is the cut off point for determining the separate clusters. The clusters are given umbers, indicated on the line. The two main branches given the letters A and B.

Fig 5 The number of plant species with certain flower types per cluster. 28

Fig 6 Visitor composition of 55 plant species in the study. The floral visitor 30 species were divided into 8 groups, shown as percentage of the total number of flower visitors.

Fig 7 The mean composition of functional groups of 8 identified clusters based 32 on number of species (spec) in that functional group and their individuals (ind).

iv

Fig 8 Representative flowers from various flower morphs. (A) bowl shaped 33 Oxystelma esculenta (B) Dish shaped Convolvulus arvensis (C) Gullet shaped Capparis decidua (D) Disc shaped Anagallis arvensis (E) Umbuliferous Torilis japonica (F) Tube shaped Lantana camara (G) Flag shaped Sesbania sesban (H) Brush shaped Eucalyptus camaldulensis (I) Austeraceous head Launaea procumbens.

Chapter 4 Fig 1 Frequency distribution of generalization level of plant hosts at Southern 48 Punjab, Pakistan. A normal fit corresponding to sample values is shown in each histogram plot.

Fig 2 Frequency distribution of generalization level of insect visitors at 48 Southern Punjab, Pakistan. A normal fit corresponding to sample values is shown in each histogram plot.

Fig 3 Seasonal variation in composition and abundance of diverse insect orders 49 at southern Punjab, Pakistan.

Fig 4 Relationship between annual interactions and system size in Southern 50 Punjab, Pakistan. The best fit regression equation is Y = 22.4994 + 8.336X, P< 0.0001, R²= 0.982.

Fig 5 Relationship between connectance and system size in the Southern 52 Punjab, Pakistan. Solid circles, winter; open triangles, autumn; solid rectangles, summer; open circles, spring, overall connectance value.

Fig 6 Monthly variation of six properties of a year-long plant-pollinator 56 network. For each property data points represent the values of individual monthly network (month 1 = January).

Fig 7 Variation of six properties of a year-long plant-pollinator network as a 57 function of the number of the merged monthly data considered for calculation. Chapter 5

Fig 1 Monthly mean temperature °C and Relative humidity (%) at Bahauddin 65 Zakariya University Campus Multan, Pakistan, during January to December, 2008.

Fig 2 Monthly abundance of the members of Sub-family Milesinae at 67 Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

v

Fig 3 Monthly abundance of the members of Sub-family Syrphinae at 67 Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

Fig 4 Monthly dynamics of number of flowering plant species, number of 68 hoverfly species and their abundance at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

Fig 5 Plant species visited by more than 6 syrphid fly species along with their 69 number and abundance at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

Fig 6 Multidimensional scaling analysis of 10 syphid species with Bray and 69 Curtis distance based on their abundance on 59 flowering plants at Bahauddin Zakariya University Multan Campus, Pakistan, from January to December, 2008.

Fig 7 Relationship between abiotic factors (average temperature °C and relative 71 humidity %) and number of syprhid fly species and their abundance over the year at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

Fig 8 Relationship between biotic factors (number of flowering plant species 72 and floral abundance per month) and number of syprhid fly species and their abundance over the year at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

Fig 9 Relation between number of syrphid fly species and their abundance at 73 Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

Fig 10 Relation between flowering plant species and their floral abundance at 75 Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

Chapter 6

Fig 1 Percentage of species and individuals of related insect orders at three 86 locations in southern Punjab during 2008.

Fig 2 Rarefaction curves of three locations i.e. (a) Perowal (b) Chechawatni (c) 86 Chak-Katora, based on individual rarefaction method showing the expected number of species as a function of sample size.

Fig 3 Rank-abundance curves of pollinator species visiting the three locations 89 (A: Perowal; B: Chechawatni; C: Chak-Katora) in Southern Punjab. The

vi

names of the species accounting for at least 10% of the visits at a given location are provided.

Fig 4 Relationship between abundance, species richness and diversity 91 (Shannon-Wiener) and plant performance in P. juliflora at southern Punjab, Pakistan. Solid circles, Chak-Katora; open circles, Chechawatni; open triangles, Perowal.

Chapter 7 Fig 1 Diurnal activity pattern of A. dorsata and A. florea on the flowers of A. 105 cepa L. in Multan, Pakistan. Fig 2 Abundance of different species of Syrphid flies on the flowers of A. cepa 105 L. in Multan, Pakistan. Fig 3 Abundance of non-Syrphid flies on the flowers of A. cepa L. in Multan, 105 Pakistan.

Fig 4 Number of seeds produced per umbel per 20 minute of visit by different 107 pollinators in A. cepa L. Mean values sharing similar letters show non- significant differences (P<0.05) by using LSD test.

Fig 5 Average number of seeds produced in open-pollinated and self-pollinated 108 umbels.

Fig 6 Population dynamics of different pollinators in afternoon (1530 hrs to 111 1830 hrs).

vii

Summary

Keeping in view the importance of pollination as the most important ecosystem service performed by the insects, current thesis deals with some aspects of plant and pollinator interactions under natural conditions i.e. pollination syndromes search among native flora, yearlong plant and pollinator network, seasonal variations in abundance and diversity of pollinators, effect of habitat types on pollinator diversity and plant reproductive performance and some aspects of crop pollination. In conjunction with the efforts to search pollination syndromes at plant community level, research was conducted in southern Punjab of Pakistan. A plant community with a variety of floral morphs was observed for their floral visitors round the year. Relationship between plant species and insect functional groups on the basis of their visiting frequencies was explored with two most often used statistical techniques i.e. multidimensional scaling analysis and hierarchical cluster analysis. Both the techniques showed almost similar results. Three more or less distinct clusters of floral visitor functional groups were obtained i.e. (short tongue bees + long tongue bees), (syrphid flies + flies excluding syrphid flies) and (butterflies + moths). There was great degree of generalization of floral visitors among actinomorphic flowers i.e. Austeraceous heads, umbels, bowl shaped, brush shaped, disc shaped and dish shaped flowers. Zygomorphic i.e. gullet and flag shaped flowers, exhibited strong affinity towards short and long tongue bees. A single tube shaped flower Lantana camara showed the classical butterfly pollination syndromes. Short tongue bee pollination syndrome could not be separated from long tongue bee pollination syndrome solely on the basis of qualitative floral characters. This study was unable to predict accurately, the classical pollination syndromes in native flora of Punjab, Pakistan. However, some evidences were there regarding bee pollination syndromes (zygomorphic gullet and flag shaped flowers) and butterfly pollination syndromes (tube and bowl shaped flowers). The network properties of flowering plants and their floral visitors were studied year around in a 20 hectare planted forest and adjacent agricultural landscape at Bahauddin Zakariya University Multan, Pakistan. Eighty plant species and 162 insect species resulted in 1573 interactions with a cumulative connectance of 12.31%. The frequency distribution of floral visitor interactions was less uniform than those of plants. Network size was smallest in winter and late autumn when interactions were low while it was the maximum in mid and late spring, the time when the maximum number of interactions was also observed. Connectance ranged from 9.63% to 52.38% (average=19.43%) and was negatively associated with system size. The variation in the number of interactions followed the changes in the number of plant mutualists more strongly and significantly than that of animal mutualists. System size increased with the increase in number of interactions. Cumulative distribution of number of interactions [p(k)] per plant and their floral visitor (k) showed a better fit to a truncated power law curve. The results also fit well to an exponential function when plotted connectance against system size. Interaction evenness increased with decrease in interaction diversity. Weighted nestedness followed the interaction diversity. The connectance and interaction evenness were least affected while nestedness, weighted nestedness and interaction diversity were most affected by merging temporal data. Species composition and population dynamics of hoverflies (Diptera: Syrphidae) in relation to some abiotic and biotic factors were studied over a year long period in the District Multan, Pakistan. The community of hoverflies was composed of 14 species which were recorded from 59 plant species. Among Syrphinae, Ischiodon scutellaris, Episyrphus balteatus and Sphaerophoria bengalensis were the most abundant whereas among Milesiinae, Eristalinus aeneus and Eristalinus laetus were the most frequent floral visitors. The peak abundance and richness of hoverflies was observed in spring (March- April), the time when the maximum numbers (35) of plant species were in flowering. Only four species i.e. E. aeneus, E. laetus, Mesembrius bengalensis and Paragus serratus remained active all through the year in low or high abundance. Among agricultural and wild plant species, Mangifera indica and Launaea procumbens were visited by the maximum number of syrphid species or in highest abundance, respectively. On the basis of similarity in floral host plant visitation frequencies, Syrphinae could easily be distinguished from Miliesiinae. Abundance of hoverflies was positively correlated with the floral abundance and flowering plant species, while temperature and relative humidity were negatively or only weakly correlated. This study was conducted in an effort to understand the effects of spatial variations in pollinator assemblage due to habitat isolation on the reproductive

2 performance of perennial plant species. Variations in pollinator assemblage structure (abundance, diversity and Shannon-Wiener index) were studied at three widely isolated (100 to 200 km apart) nature reserves of Southern Punjab, Pakistan, in order to explore its effects on reproductive performance of Prosopis juliflora. Species richness and abundance were highest in Pirowal Sanctuary followed by Chichawatni Sanctuary and Chak Katora forest reserve. The pollination system of P. juliflora was highly generalized with 74 insect visitor species in four orders among all the three sites. However, pollinator assemblage varied significantly in composition among the sites. Out of the 4 reproductive parameters considered, the number of pods per raceme and germination varied significantly among the three locations. The reproductive performance of P. juliflora in terms of number of pods per raceme and germination improved with abundance of

pollinators. In conjunction with the efforts to discover the pollinator community of crops and the best pollinators for crop production, some experiments were conducted at the research farm of University College of Agriculture, Bahauddin Zakariya University, Multan, Pakistan. Two cross-pollinated crops were selected i.e. onion (Allium cepa L.) and janter (Sesbania sesban). 1) The community of pollinators in A. cepa was composed of two bees (Order: Hymenoptera) and eight true flies (Order Diptera). Among bees, Apis dorsata proved to be an abundant pollinator (2.85±1.57 individuals/25 plants) while among true flies Episyrphus balteatus (Syrphidae) had the maximum abundance of 14.00±4.61 individuals/25 plants. All the insect pollinator species reached peak activity during 10:00 to 12:00 h. Eupeodes corollae (Syrphidae) exhibited the most efficient foraging behavior by visiting 17.14±1.38 flowers in 147.5±8.14 seconds on an umbel. A. dorsata was revealed as the most effective pollinator, however, based on seed setting results for visits by single species over 20 minutes and which produced 506 seeds/umbel/20 minute visit. 2) Different yield attributing components of S. sesban were measured in three types of pollination treatments i.e., wind pollination, self pollination and cross pollination by insects. Cross pollinated inflorescences produced maximum number of pods (4.41) and maximum seed weight (1.1 g) as compared to wind pollinated (2.42 pods and 0.72 g seeds/ inflorescence) and self-pollinated inflorescences (2.25 pods and 0.46 g seeds/

3 inflorescence). Germination was also better in cross pollinated inflorescences as compared to self and wind pollinated inflorescences. Megachlie bicolor and A. dorsata exhibited most efficient foraging behavior, whereas A. florea proved to be the most abundant pollinator with highest visitation index of 0.74. Most of the pollinator activity took place in afternoon (1530 h–1830 h).

4 Chapter 2

INTRODUCTION

Biodiversity and Ecosystem Functioning

During the last 150 years human has brought about changes in the ecosystem in a way unrepresented in any other period of the human history (Solecki et al. 1999). Due to this change in ecosystem the plant’s biodiversity has decline and many species have been extinct or at the verge of extension. The millennium Ecosystem Assessment (MA) Synthesis report (UNEP, 2005) unambiguously mentioned the biodiversity loss as one cause in a complex of causes for change of ecosystem functioning and declining ecosystem services. Decreased biodiversity in an ecosystem can lead to instability of an ecosystem by decreasing the production of vegetation (Lehman and Tilmar, 2000), affecting the food web structure (McCann, 2000; Dunne et al., 2002) and lowering the resistance to invasive species (Kennedy et al., 2002). An ecosystem is a complex of many processes which take place simultaneously; therefore studying the effect of biodiversity on one particular process may provide better understandings than trying to study a whole ecosystem. Insect pollination is one of the most important ecological processes which involve two parties (plants and insects) that can mutually benefit from each other (Procter, 1978; Faegri and Van der Pijl, 1979). Insects actively or passively transfer pollen grains, containing the male gamete, from the stamens to stigma which lead to the process of fertilization in ovaries. It is a kind of mutualistic interaction in which insects usually receives pollen and/or nectar from plants in return for their pollination services. In this way, plants and pollinators have been coevolved through millions of years (Faegri and Van der Pijl, 1979). The loss of species have not spared plant and their pollinators; many plant and pollinator species have been extinct or at the verge of extinction (Buchmann and Nabhan, 1996). Great concern exists for the negative consequences this may have for reproduction of wild plants and crops (Kearns and Inouye, 1997; Fisher, 1998; Kearns et al., 1998; Cox and Elmqvist, 2000). For example in the Europe, up to 83% of the 264 species grown as crops are animal-pollinated (Williams, 1996).

5

The nature of biotic pollination There is a tremendous diversity of pollination systems with great variations in the degree of specialization and generalization for pollinators among plants and for flowers among pollinators (Proctor, 1978; Waser et al., 1996). Flowers can be very simple with easily assessable pollen and nectar resources, or very complicated which may limit the number of possible visitor species. Pollinators can be as different as insects, birds, bats and even lizards (Traveset and Saez, 1997; NyHagen et al., 2001). Pollination systems can be classified into types of plant species with functionally similar flowers and similar pollinator compositions (guilds). These types are called pollination syndromes (Faegri and Van der Pijl, 1979). Insects are dominant pollinators on the earth (Faegri and Van der Pijl, 1979); at least 70% of all angiosperms are insect pollinated (Kearn and Inouye, 1997). Specialization versus generalization The majority of plant-pollinators are generalized in nature, and only the minority of species interactions is specialized (Jordano 1987; Ellis and Ellis-Adam 1993; Memmott 1999). One-to-one relationship between single plant and animal species are extremely rare, particularly in temperate climates (Kwak et al., 1998). The degree of specialization can be regarded in two ways: evolutionary and ecological specialization (Armbuster et al., 2000). Evolutionary specialization is a process with a direction, i.e. from many to fewer pollinating taxa. The ecological specialization of a species is a state, referring to having few pollinators (from the same or different functional groups) relative to other plant species, or visiting few functional types of flowers compared to other insects in case of the flower visitors. Furthermore specialization and generalization are not a dichotomy, but a continuum (Armbuster et al., 2000; Johnson and Steiner, 2000). The distribution of interactions between plants and flower visitors is highly asymmetrical (Jordano, 1987; Bronstein, 1995; Memmott, 1999; Olesen and Jordano, 2002). Many specialist plants are visited by generalist insects, whereas many specialist insects visit plant species that are mostly visited by generalist insects. Ollerton and Watts (2000) elucidated it by putting the example of true flies (Diptera) which prefer dull colored flowers; it only means that true flies prefer dull

6

colored flowers statistically, since many brightly colored flowers are exclusively pollinated by true flies. Flowers having traits representing particular pollination syndromes reflect the impact of long-term natural selection by particular guilds of pollinators (Armbruster et al., 2000). However, the refinement of current concepts of pollination syndromes needs to incorporate ecological context, specially the role of temporal and spatial variation in the role of pollinators (Fenster and Dudash, 2001). Flowers may show specialized traits, but are paradoxically visited by a large number of species. It may be that only a small proportion of the visitors are actual pollinators (affectively transfer pollen to stigma in single visit, seed or fruit set per visit etc.), functioning as a selective force (Johnson and Steiner, 2000; Ollerton and Watts, 2000). Another theory may be that flowers can also be adapted to relatively less effective pollinators when this adaptation causes little loss in the fitness contribution of a more effective pollinator. Biodiversity and pollination Diversity of both plants and pollinators is important for pollination in an area. Changes in species richness and population size of both plant and pollinators can affect pollination at community level (Rathcke and Jules, 1993). Pollination consists of quantity components i.e. number of visits per pollinator species per plant species and qualitative components i.e. co and heterospecific pollen deposition (Kwak et al., 1998). Therefore for many plant species the loss of pollinator species will result in lower seed set (Kearns and Inouye, 1997; Tomimtsu and Ohara, 2002) and/or inbreeding depression (Oostermeijer et al., 1994; Luijten, 2001) particularly in the plant species having high degree of crossing. In many plant species the negative effects of cross-pollination can be overcome by self-pollination (autogamy), long term seed banks or clonal propagations (Spira, 2001). Plant diversity in shape of species richness and composition can have various effects on plant-pollinator interactions. Generally the number of plant species is positively correlated to the number of pollinator species and individuals (Corbet, 1997; Potts et al., 2003). Plants can also facilitate each other in time; insects usually have longer phonological life span than the flowering period of particular plant species and therefore need several sequentially flowering plant species during their life span (Bronstein, 1995).

7

Insects and plants not only morphologically co-evolved but phonologically, they have also been adapted for each other. Plants can also have negative effects on each other in shape of competition towards the number of floral visitors and their visitation rates. This competition may result in heterospecific pollen deposition and pollen loss which can result in reduced seed set (Waser, 1978; Campbell and Motten, 1985; Brown et al., 2002) or reduced pollen flow distance (Campbell, 1985). In nature this constraint is overcome by the phenomenon of floral constancy in pollinators; in which a pollinator tend to visit a single plant species even in the presence of many other flowering plant species in that particular area. Floral constancy is a characteristic foraging behavior of the bees. Goulson and Wright (1997) also reported a marked floral constancy in syrphid flies (Diptera: Syrphidae). The balance between competition and facilitation depends on the type of plant species, temporal and spatial variations in species composition and flower abundance and density (Kwak et al., 1998). Competiotion and facilitation interactions mostly occur between plant species with shared pollinator fauna (Feinsinger, 1987). Majority of the plant species are more generalist and the nature of plant pollinator food web is rather complex (Mammott, 1999). Therefore the extinction of one of the interacting species does not directly lead to the extinction of other species (Kearn et al., 1998). It is the nature of interaction, either facultative or obligatory, which determines the significance of loss of partners. Generalist plant species may be resilient to a pollinator species loss, because other pollinator species may fulfill the deficiency of the lost one (Kwak, 1994). However the pollination potential (pollen harvesting, pollen flow distance and pollen deposition on stigma) of the alternate pollinator species determine their exchangeability. Plant species that depend on single or few pollinator species are said to be the most valueable in the loss of pollinator species may leave only a few or no alternatives (Johnson and Steiner, 2000; Spira, 2001). Unfortunately, data showing whether the plant species with various degrees of specialization differ in their vulnerability to pollinator loss are currently lacking.

8

Status of pollinators and pollination studies in Indian sub-continent Indian sub-continent is represented by three countries i.e. India, Pakistan and Bangladesh. World’s 7.28 % of the biodiversity including 70,000 insect species are present in this part of the world (Gupta, 2004). Most of the parts of the sub-continent come under tropical to sub-tropical climatic conditions. After the series of work done by British authors during the late 19th and early 20th century namely “The Fauna of British India” , no considerable additions or revision have been done except from India. Where the biodiversity of insects has been well studied and revised after mid 20th century. For example beetles (Coleoptera); represented by 17036 known species form India (Gupta, 2004). Similarly, more than 700 bee species are known from this part of the world. On the other hand, In Pakistan, only a few studies have been done resulting in only few known bee species. The diversity of beetles, wasps and dipterous flies still remain a mystery in Pakistan. Only the diversity of hoverflies have been studied well (Saleem et al., 2001; Arif et al., 2001; Irshad, 2001; Sajjad et al., 2008; Sajjad and Saeed, 2009; Sajjad et al., 2010). During the last thirty years, many new records of different insect species have been made by many authors. These newly recorded insect species need compilation. Crop pollination by honey bees (Apis mellifera, A. dorsata and A. florea) have been the focus of scientists on a limited scale in Pakistan. However, the trend of crop pollination by non-apis bees is still lacking among pollination biologists. Species decline in Pakistan The ecological threat of great concern in Pakistan is the continuous loss of natural and modified habitats due to the process of fragmentation and degradation i.e. forest and fresh water ecosystems (Anonymous, 2000). The loss of species is also positively correlated with the loss of natural habitats. Today, in Pakistan, hundreds of species of plants and animals are declining and many are on the verge of extinction. Some species have already been extinct, many are reckoned as internationally threatened and more still are of national concern. As an example of species loss, at least six mammal species are known to have disappeared from Pakistan within last 400 years: Tiger (Panthera tigris); Swamp Deer (Cervus duvauceo) Lion (Panthera leo); Indian One-horned Rhinoceros

9

(Rhinoceros unicornis); Asiatic Cheetah (Acinonyx jubatus venaticus); and Hangul (Cervus elaphus hanglu). According to IUCN red list of threatened animals, 37 species and 14 sub-species of internationally threatened or near-threatened mammals occur in Pakistan (Anonymous, 2000). There is little or no data available to demonstrate the trends of species decline in Pakistan. A few preliminary attempts have been made to draw a list of 500 native plants species which are believed to be nationally rare or threatened (Davis et al., 1986). A rapid loss of natural habitats has seen during last few decades which is still continuing. It is estimated that Pakistan’s woody biomass is depleting at the rate of 4 to 6 percent per year (GOP, 1992; Hosier, 1993). Consumption (primarily for fuel wood and household) is expected to increase in the line with human population growth at about 3% per year. Given the widespread conversion of natural habitats into agricultural habitats, almost all remaining natural or modified ecosystems in Pakistan are now critically threatened. No systematic and comprehensive assessments have yet been made aiming to rank the importance of biodiversity of Pakistan’s remaining natural ecosystems and habitats. The situation becomes worst in case of lower animals i.e. insects, mollusks, fungi etc. which play a very crucial role in the stability of any ecosystem. Research Area The research area is in the Southern Irrigated Zone of Punjab, Pakistan (PARC, 1980). Climate of the area is sub-tropical with hot summer and cold winter; the mean daily maximum and minimum temperatures are in the range of 38 to 46 °C and 8 to 12 °C, respectively while the mean monthly summer rainfall is ca. 18 mm (Khan et al., 2010). Geographically, it is an alluvial plain with fertile soil deposited by the flood regime of the rivers through thousands of years. Most of the land is cultivated and irrigated by canals or underground water. A variety of crops are grown but cotton and wheat rotation mostly is the tradition. Four districts of Southern Irrigated Zone of Punjab are studied for our objectives i.e. Multan, Muzafargarh, Bahawalpur and Khanewal. There are three wildlife parks i.e. Bahawalpur, Khanewal and Muzafargarh, and many planted forests in the region. The greater part of the area is under agricultural use. Topographically, there are no marked variations in this part of Pakistan.

10

Aims and outlines of the thesis In this thesis, the status of pollinators and pollination is assessed in southern irrigated zone of Punjab by studying different aspects of plant and pollinator interactions. Community diversity is the local species richness and abundance of flowering plants and flower visiting insects. Functional diversity is the number of different pollination systems or syndromes reflected as guilds of plants with similar floral morphologies and plant preferences. The relation between community and functional diversities is investigated. Plants and their floral visitors are also studied in their food web round the year. This also includes temporal and spatial variations in their abundance and diversity. Ultimately the effect of habitat types on the diversity of pollinator types is studied with particular reference to plant reproductive success. Finally, experiments were performed in conjunction with the efforts to identify the role of insects in crop production and selecting the best pollinators in terms of single visit efficacy.

Habitat types

Pollinator abundance Pollination Plants abundance and diversity system and diversity

Food web

Tempo-spatial variations

Pollination

Plant reproductive success

Fig 1. Flow diagram depicting the various processes that are thought to be important in plant and pollinator interactions at local or landscape levels.

11

The main subject of this thesis is therefore to address the importance of biodiversity of insect pollinators for the pollination of entomophilous plants. It includes different aspects from landscape and community level to pollination and seed set. Following were the specific objectives of the thesis.

1) To study the biodiversity of flowering plants and flower visiting insects in agriculture and semi-natural landscapes of southern irrigated zone of Punjab. 2) To study various pollination systems based on frequencies of insect visiting guilds. 3) To determine the food web interaction between flowering plant and insect visitors communities along the year. 4) To determine variations in abundance and diversity of flower visiting insects around the year. 5) To determine the effect of habitat on pollinators diversity and plant reproductive success. 6) To discover the potential native pollinators for crop production.

12

Chapter 3

Pollination Syndromes Based on Frequencies of Insect Visiting Guilds Summary

In conjunction with the efforts to search pollination syndromes at plant community level, research was conducted in southern Punjab of Pakistan. A plant community with a variety of floral morphs was observed for their floral visitors round the year. Relationship between plant species and insect functional groups on the basis of their visiting frequencies was explored with two most often used statistical techniques i.e. multidimensional scaling analysis and hierarchical cluster analysis. Both the techniques showed almost similar results. Three more or less distinct clusters of floral visitor functional groups were obtained i.e. (short tongue bees + long tongue bees), (syrphid flies + flies excluding syrphid flies) and (butterflies + moths). There was great degree of generalization of floral visitors among actinomorphic flowers i.e. Austeraceous heads, umbels, bowl shaped, brush shaped, disc shaped and dish shaped flowers. Zygomorphic i.e. gullet and flag shaped flowers, exhibited strong affinity towards short and long tongue bees. A single tube shaped flower Lantana camara showed the classical butterfly pollination syndromes. Short tongue bee pollination syndrome could not be separated from long tongue bee pollination syndrome solely on the basis of qualitative floral characters. This study was unable to predict accurately, the classical pollination syndromes in native flora of Punjab, Pakistan. However, some evidences were there regarding bee pollination syndromes (zygomorphic gullet and flag shaped flowers) and butterfly pollination syndromes (tube and bowl shaped flowers).

Key words: Pollination syndromes, floral visitor guilds, floral traits, flora of Pakistan

13

Introduction In perspective of floral evolution studies, it is believed that the diverse array of floral morphologies reflect selection pressures exerted by different pollinating animal groups. This selection pressure could be in the form of visitation frequencies, feeding habits, mouth part types and body morphologies of these pollinating animals. This has originated the idea of specialization in plant and their pollinator plants and their pollinator interactions. For the first time Kölreuter (1761) and Sprengel (1793) described this specialized interaction between pollinator types and floral traits. Afterwards, it was studied in depth by different evolutionary biologists (Müller, 1883; Knuth, 1906; Fǽgri and Ven der Pijl, 1966; Delpino, 1968; Stebbins, 1970). Delpino (1868) used the term “adaptational groups” for the flowers with morphologically distinct traits which is now known as “pollination syndromes”. Fenster et al., (2004) defined pollination syndrome as a suit of floral traits including reward, associated with attraction and utilization of a specific group of animals and pollinators. During the last two decades, pollination syndrome concept has widely been tested on variety of flowering plant species around the world. Since then, it has been subject of debate due to great variability among the results. An important first step for the study of evolutionary consequences of pollination syndromes, one must recognize that the concept implies that pollinators are clustered into functional groups (e.g. long-tongue flies or small, nectar-collecting bees) that behaves in similar ways on a flower and exerts similar presser, which in turn, generates correlation among floral traits (e.g. long and narrow corolla tubes, pollen presented in certain way, or particular nectar quantities and concentrations) (Waser et al., 1996; Ambruster et al., 2000). Such pollinator-driven floral evolution can proceed with or without the animals coevolving (Kiester et al., 1984). There is dire need to organize pollinator communities by functional groups and then understate more fully the degree to which these groups overlap in their selective pressure they exert on floral traits (Ollerton and Watts, 2000). The description and characterization of pollination syndromes has been a prominent thing for evolutionary biologists. There is no doubt that in some cases, one can predict the pollinators of a

14 particular flower from its floral biology e.g. color, scent, reward, anthesis timings and symmetry etc (Ollerton and Watts, 2000). The current trend in research work on pollination syndrome is mostly based on testing the list of characters given by Fǽgri and Ven der Pijl (1979). This aroused only the idea of specialization among different floral traits towards their specific type of pollinators. This situation becomes more critical for the new students learning the functional diversity of flowers (Wilson et at., 2004). The reason being pollination syndromes concept has yet been included in the introductory biology text books (Johnson and Steiner, 2000). As a matter of fact there is great deal of overlap among functional groups on a single flowering plant species (Ollerton and Watts, 2000). Fǽgri and Ven der Pijl (1979) have acknowledged that there are no hard and fast rules in determining pollination syndromes. Ollerton and Watts (2000) elucidated it by putting the example of true flies (Diptera) which preferred dull colourd flowers; it only means that true flies prefer dull colored flowers statistically, since many brightly colored flowers are exclusively pollinated by true flies. For example Euphrasia striata, which has shown high affinity towards true flies though it has brightly colored petals (Hingston and McQuillan, 2000). As a matter of fact, pollination syndromes represent floral evolution through selection pressure exerted by single or multiple functional groups (Fenster et al., 2004). The pollinators which are relatively most frequent and relatively most effective will usually be the most important selective force (Stebbins, 1970). Therefore the visitation frequencies of functional groups alone may not reflect the clear picture of floral evolution. Classifying plants according to syndromes is a time consuming job. Collecting data on pollinators is even more difficult because not all the floral visitors are pollinators and the actual pollinators vary in their effectiveness. Furthermore the temporal and spatial variations in abundance and diversity of pollinators may also undermine our judgment of pollination syndrome search. Floral phenologies therefore should be kept in mind, as plant species with prolonged flowering may be visited by different pollinators in different intervals along the season. Yet a more quantitative approach of assessment of nature of pollination syndromes can be achieved (Ollerton and Watts, 2000).

15

Pollination syndromes long have been regarded as excellent examples of convergent evolution (Baker, 1961; Grant and Grant, 1965; Stebbins, 1970; Fenster et al., 2004). The existence of similar suites of plant reproductive traits in unrelated groups suggests a certain degree of specialization of flowers to guilds of pollinators. Yet the pollination syndrome concept has recently attracted some criticism and some biologists have recently questioned the degree to which plant-pollinator interaction interactions are evolutionarily specialized. Some have suggested instead that extreme specialization in nature is, in fact, rare (Herrera, 1996; Waser et al., 1996). Many an efforts have been made to identify pollination syndromes based on pollination visiting guilds around the world. Most of the studies focused single family or genera with similar flowers (Armbuster and Herzing, 1984; Wilson et al., 2004; Figueroa-Castro and Cano-Santana, 2004; Wolfe and Sowell, 2006). Only a few workers focused a wide range of plant families at a time (Hingston and McQuillan, 2000; Ollerton and Watts, 2000; Devy and Davidar, 2003) with a variety of floral traits. Most of the work on testing the hypothesis of pollination syndromes has been done in North America, South Africa, Australia, Mexico and Europe. In South Asia, this concept is poorly tested on native plant species both at individual or community level. The purpose of current study was of twofold: (1) Do native insect-pollinated plants fit into traditional pollination syndromes? (2) Do these syndromes successfully predict the most frequent pollinators of the flowers?

16

Materials and Methods Study Site: The study was done in the Southern irrigated zone of Punjab (PARC, 1980), Pakistan. Climate of area is subtropical with a hot summer and cold winter; the mean daily maximum and minimum temperatures are in range of 38 to 50Ċ and 8 to 12 Ċ, respectively with the mean monthly summer rainfall is ca. 18mm (Khan et al., 2010). Geographically it is an alluvial plain which fertile soil deposited by the flood regime of the rivers trough thousands of years. Most of the land is cultivated and irrigated by canals or underground waters. A variety of crops are grown but cotton and wheat rotation mostly is the tradition. We focused on three cities of Southern irrigated zone of Punjab for our objectives i.e. Multan, Muzafargarh and Khanewal. Muzafargarh is situated 40 Km in east of Multan and Khanewal is 50 Km in west. In Multan city (30.255°N; 71.513°E; 114 ± 6 meter above sea level), I chose two localities, a twenty year old planted forest of 20 hectares in Bahauddin Zakariya University (BZU) campus and the other one in the research farm of Central Cotton Research Station (CRS) . In CRS, I also investigated unmanaged margins of watercourses having a variety of wild plants. In Muzafargarh (30.075°N; 71.192°E; 400 ± 25 meter above sea level), I selected the bank of Chanab River. I focused on the least disturbed areas along the river verges and some adjacent agricultural landscape. In Khanewal city (30.341 °N; 72.035 °E; 437±16.5 mete above sea level), I selected a wildlife sanctuary (7212 hectare) and adjacent agricultural lands for study. I attributed planted forest and unmanaged river or water course verges as a semi-natural landscape and wildlife sanctuary area as natural landscape. Floral visitor census: The liaison of plant species and their floral visitors was investigated in a large array of vegetation types. In semi-natural and natural habitats I used random walks in the study areas and considered only single plant species at one time during its anthesis. I focused only plant species with diurnal anthesis and recorded only diurnal floral visitors. For the plants with prolong anthesis (day and night), I took observations only in the day. Fifteen individuals per plant species were randomly selected and each plant was observed for 60 seconds. In this way there was a total of 15 minute observation per plant species in one census. For each individual plant, I counted the number of visiting individuals per floral visitor species by visual observations. Fortnightly census of each flowering plant species was done from the very begin to the

17

end of its flowering period. Rainy and cloudy days were avoided for data recording. I defined the floral units for each plant species separately and counted floral visitors from these floral units, since different plant species had different inflorescence types. All plants Species were identified by local well renowned experts (see acknowledgement). I recorded the identities of all the visitors excluding ants since these were notorious as nectar thieves and containing some anti-pollen fungal spores on their bodies. All the insect species were first morphotyped and then identified to lower taxonomic level whenever possible. For family level identification, I used keys in Borror et al. (1981). The genus of the syrphid flies (Syrphidae: Diptera) were identified with the help of key given in Vockeroth (1996) and species were identified by related expert (see acknowledgement). Bees were identified only up to genus level using the key of Michnere (2007). Butterflies were identified to species level by local well renowned taxonomist (see acknowledgement). The insects which could not be identified to species level were grouped according to morphospecies (a taxonomic species based wholly on morphological differences from the related species). Voucher specimens were deposited at Agricultural Museum of the University College of Agriculture, Bahauddin Zakariya University Multan. Data analysis: The floral visitors were grouped into nine functional groups for analysis i.e. beetles, wasps, short tongue bees, long tongue bees, syrphid flies, flies, butterflies, moths and birds. As I did not measure any morphological features of the insects, the subdivision was mainly based on taxonomy. I separated syrphid flies from other flies because of its higher abundance (composed 45% of total Diptera). As I supposed that on account of their greater abundance, they might have different selection pressure than rest of the Diptera. Some species of skippers were also observed but I merged them with moths due to their close resemblance. I did not measure any morphological features of the plants and therefore subdivided plant species into nine flower types. I based flower types on the six "structural blossom classes" (dish-bowl, bell-funnel, head-brush, gullet, flag and tube) as described by Fagri and van der Pijl (1979). In total, I had nine flower types: disc, gullet, flag, umbel, bowl, dish, heads, brush and tube.

18

Relationship among plant species and among insect functional groups on the basis of visiting functional group frequencies were explored with two most often used statistical techniques i.e. Multidimensional Scaling analysis (MDS) and hierarchical cluster analysis (HCA). MDS for binary data was used to look for the data matrix in multidimensional space which is expressed by two dimensional arrangements of variables when the data are graphed. Cluster analysis also look for similarities in the suits of traits which characterize syndromes but suing several different mathematical techniques at computer software XLSTAT (XLSTAT, 2008). The resulting ordination plots and hierarchical dendrogram were based on similarity of visitor profiles. Bray-Curtis index of similarity was used instead of Euclidean distance as input to the both of the analysis because many of the cells in the data matrix were zero (Beals, 1984). For cluster analysis Ward’s method of clustering was used.

19

Results Results using Multidimensional Scaling Analysis A total of 73.75 hours of net observation was done on 3360 individuals of 55 plant species in 17 different families. The floral visitor profile was composed of 5407 individuals of 162 insect species in four orders. All the nine functional groups were significant predictors of the variation in their visitor profile in the phenotypic space (Fig 1). Three more or less distinct clusters were apparent: (short tongue bees + long tongue bees), (Syrphidae + Diptera) and (butterflies + moths). Wasps, birds and beetles were widely separated from each other. It is fair to assume that functional groups in each cluster exert similar selection pressure on respective floral morphs. The ordination of 55 plant species shows that bowl shaped flowers “Heliotropium europaeum, Phyla nodiflora, Verbena officinlis, Oxystelma esculenta and Physalis minima” were clustered on the top left of the ordination plot (Fig 2). These species were either associated with butterflies and moths or short and long tongued bees. P. minima was pale yellow in color and exclusively visited by short and long tongued bees. Another bowl shaped plant “Calotropis procera” was placed in the top right of the ordination plot due to its high affinity towards long tongue bees than butterflies and moths. A single tube shaped flower in this study i.e. Lantana camara also exhibited strong affinity towards butterflies and moths. Out of three yellow Austeraceous head species, Launaea procumbens and Pulicaria crispa were clustered on the top left of the plot showing the combined effect of bees and syrphids. Whereas Sonchus asper was placed in lower right half of the plot due to high Diptera visitation. Launaea procumbens was visited by the highest number of syrphid flies among all the tested flowering plants, since Syrphidae were closer to bees (Fig 1), L. procumbens was clustered near P. crispa which was also mainly visited by bees. Out of eight brush shaped flowering plants, five (Prosopis juliflora, Grewia subinaequalis, Ageratum conyzoides, Eucalyptus camaldulensis and Cirsium arvense) were clustered in the upper right side of the ordination plot. These five species have shown attractiveness towards syrphidae and butterflies in addition to bees. The other brush shaped Leucaena leucocephala and Acacia nilotica were scattered in the lower

20

right and left halves of the ordination plot, respectively. This is due to more visitations by beetles and birds in former and wasps in later. Disc shaped flowers were widely scattered across the ordination plot. There was no clear affinity of any functional group towards disc shaped flowers. Yet two loosely spaced clusters were obvious. First one in the top left and second one in the lower right half of the ordination plot. The first cluster represents more affinity towards bees and butterflies whereas the second cluster shows more affinity towards flies.

1

Moths Butterflies

0.5

Wasps Bees S Bees L 0

Dim2 -1.5 -1 -0.5 0 0.5 1 1.5 Beetles

Syrphidae -0.5 Diptera

Birds -1 Dim 1

Fig 1. The ordination of floral visitors groups according to their visiting profiles (Kruskal's stress-1= 0.232). (Bee S= Short tongue bees, Bees L= Long tongue bees, Diptera= Diptera excluding Syrphidae)

21

1.5

1 (CP) LP SS Co Ms PC LC 0.5 MI Pj PM cp tp Gs (tt) csp (pa) (PN)HE ct Ca ca Ta Ac Cd (OE) aa DC ct (AG) Ec VO 0 Ancv

Dim2 zj -1.5 -1Bv -0.5 0hr 0.5 1 1.5 mc mm Ds SA rm ai (ma) -0.5 Ll at eh La sf (aa) (Am) FI Obs54 (cp) -1 TJ Vs

-1.5 Dim1

Fig 2. The ordination of plant species according to their flower visitor frequencies (Kruskal’s stress-1 = 0.276). Binomial codes for floral morphology: uppercase (Italic, normal) + uppercase (Italic, normal) = Umbel, upper case (bold) + lowercase (normal) = Gullet, uppercase (bold) + uppercase (normal) = Flag, uppercase (normal) + uppercase (bold) = Tube, uppercase (normal) + lowercase (normal) = brush, uppercase (normal) + uppercase (normal) = Head, uppercase (bold) + uppercase (bold) = Bowl, lowercase (normal) + lowercase (normal) = Disc, lowercase (Italic) + lowercase (Italic) = Dish shaped. Zygomorphic flowers are underlined and flowers with nectar guides are in parenthesis. Binomial codes for plant species are given in table 1.

Gullet and flag shaped zygomorphic flowers were loosely placed in the central upper and lower left of the ordination plot. These plant species were mostly visited by short and long tongue bees and occasionally by butterflies. Only the Dalbergia sissoo (gullet shaped) was visited by more Syrphidae than bees, therefore placed in the lower right half of the ordination plot. The two umbel bearing plant species (Daucus carota and Torilis japonica) were widely separated from each other. In contrast of D. carota, T. japonica was not visited by even a single bee species.

22

The two dish shaped flowering plant species (Convolvulus arvensis and Convolvulus sp.) were placed in top right half of ordination plot and associated with short and long tongue bees, flies and Syrphidae but these two species were fairly separated from each other. This is due to the difference in beetles and wasps visitation i.e. C arvensis was not visited by wasps exclusively but frequently visited by beetles and vice versa in case of Convolvulus sp. The nectar guides significantly attracted butterflies than any other group. Only seven plant species had visible nectar guides in this study. These include C. procera, P. nodiflora and O. esculenta from actinomorphic bowl shaped flowers while Malcolmia africana and Anagalis arvensis from disc shaped actinomorphic flowers. The single tubular flower (L. camara) in this study had also dark yellow nectar guides on pink or pale back ground. P. nodiflora and O. esculenta were mostly visited by butterflies and wasps whereas C. procera was visited excessively by long tongue bees. Similarly M. africana and L. camara were mostly visited by butterflies and A. arvensis was visited by dipterous flies and Syrhidae in excess. The only flower with nectar guides among zygomorphic flowers was sp.16. Though it was a zygomorphic gullet shaped flower but mostly visited by Syrphidae and butterflies than bees. Results Using Cluster Analysis The column cluster statistics of nine self defined functional groups based on their visiting frequencies revealed five distinct guilds (clusters) i.e. (beetles), (birds), (wasps + butterflies + moths), (short tongue bees + long tongue bees) and (Diptera + Syrphidae) (Fig 3). It is fair to assume that functional group groups in each cluster exert similar selection pressure on their respective floral morphs. The row cluster statistics of 55 plant species resulted in five distinct clusters, that split into two main branches (Fig 4) i.e. branch A (Cluster 1) and branch B (Cluster 2-8). The flower types are not equally distributed over the clusters. Disc shaped flowers are widely distributed among the cluster except cluster 7. Likewise, brush shaped flowers are also widely distributed except cluster 1 and 7 (Fig 5).

23

1.4

1.2

1

0.8 0.6

Dissimilarity 0.4

0.2

0 L S

Birds Bees

Bees Moths Wasps Beetles Diptera Syrphidae Butterflies

Fig 3. Dendrogram of 9 functional groups of the cluster analysis with Bray and Curtis distance, based on their visiting frequencies on 55 plant species. The dotted line is the cut off point for determining the separate clusters. ). (Bee S= Short tongue bees, Bees L= Long tongue bees).

Flower types; bowl, brush, flag and tube are exclusively found in branch B. Cluster 7 is entirely composed of flag shaped flowers. The only tube shaped (L. camara) fall in cluster 6. Most of the zygomorphic flowers are restricted to cluster 3 and 7. The later is entirely composed of zygomorphic flowering plant species (L. aphaca, F. indica and Vicia sativa). Zygomorphic flowers (gullet and flag shaped) were predominantly visited by short and long tongue bees. On the other hand cluster 1, 2 and 6 are dominantly by Actinomorphic flowers i.e. disc, brush and bowl shaped flowers, respectively. Cluster 1 was dominated by Diptera and Syrphidae, cluster 2 was dominated by Diptera and bees while cluster 6 was occupied mostly by butterflies and moths. In contrast, the actinomorphic flowers were visited by a variety of pollinator groups in great variability. For example bowl shaped flowers (H. europaeum, O. esculenta, P. nodiflora and V. officinlis) are gathered in cluster 6 due to high visitation by butterflies and moths (Fig 6). Whereas two other bowl shaped flowering plants (C. procera and P. minima) were visited significantly by short or long tongue bees and fall in cluster 3 and 4, respectively.

24

The two dish shaped flowers in this study (C. arvensis and Convolvulus sp.) fall in cluster 2 by supporting almost all kind of pollinator groups. Brush shaped flowers are distributed in cluster 2, 4, 5, 6 and 8 but tend to be concentrated in cluster 2. The most frequent observed insect groups were long tongue and short tongue bees followed by Diptera and Syrphidae. Beetles, wasps and birds did not represent a majority of visitors for any plant species. The tube shaped L. camara was clustered in cluster 6 along with gullet, disc, bowl and brush shaped flowers. On the other hand five different functional groups of floral visitors were distributed in this cluster giving the sense of extreme generalization even in tube shaped L. camara which otherwise, exhibiting a typical butterfly pollination syndrome through MDS. The MDS is a latest technique to visualize the relative position of each plant species (with particular floral morph) in a phenotypic space. Therefore it is more explanatory than HCA.

25

3.5

3

A B 2.5

2

1.5

Dissimilarity 1 2 3 4 5 6 7 8 1

0.5 d 0 d f b u u Sp.ni Acanil Behvar Phynod Achasp Heleur Vicsat Lataph Abuin Fumin Malcor Centpul Dalsis Leule Torjap Malafr Anaarv Aspten Euphel Zizjuj Daucar Capdec Sonasp Marmin Ranmur Laupro Cirarv Agecon Gresu Paracu Projul Conarv Sp. Con. Casocc Sesses Pulicri Melind Triter Tripor Cucpro Calpro Medsat Phymin Atymin Tamaph Suafr Halrec Euccam Chrtin Clevis Alhgra Chotri Lancam Oxyesc Verof

Fig 4. Dendrogram of 55 plant species of the cluster analysis with Bray and Curtis distance, based on floral visitor frequencies, the dotted line is the cut off point for determining the separate clusters. The clusters are given numbers, indicated on the line. The two main branches given the letters A and B. The shorts for plant names are used which are given in table 1. The dotted line is the arbitrary cut off point for determining separate clusters.

26

Discussion The bees (long and short tongue) were clustered together. This suggests that it is not possible to distinguish short tongue bee pollination syndrome from long tongue bee pollination syndromes solely on the basis of qualitative floral traits, particularly when dealing plant species at community level. This might be because at community level, we usually do not consider floral traits quantitatively e.g. corolla length, nectar depth etc., instead qualitative traits like color, shape, orientation etc. are considered. For example Kite et al. (1998) reported Amorphophallus species (Araceae) exclusively pollinated by beetles but classically showing the fly pollination syndrome. But Hawk moth pollinated flowers surely can not be pollinated by carrion flies (Ollerton and Watts, 2000). When seen the relative position of syndromes in phenotypic space, syrphids are much closer to long tongue bees than any other group. In the current study, the genus Eristalinus was the most dominant among Syrphidae. The members in the genus Eristalinus (E. aenus, E. arvorum, E. laetus, E. taeniops and E. tenax) were greater than 10 mm in length with mouthparts ranging from 5mm to 7mm. This character makes them closer to long tongue bees in mode of feeding but flies can not manipulate the complex flowers which the bees can. Bowl shaped flowers have open nacteries and reproductive organs are in the centre of the flowers. These are the species with only a single or few flowers in loose inflorescence. In our studies there was a great degree of generalization in bowl shaped flowers. O. esculenta and V. officinlis were strongly related with butterflies and moths. But C. procera and P. minima showed great affinity towards bees. Dipterous flies, syrphidae, beetles and birds had limited affinity towards bowl shaped flowers. The bowl shaped C. procera and O. esculenta (both from family Asclepiadaceae) have same colored (Purple + white) and same sized flowers. O. esculenta showed strong affinity towards butterflies than bees but vice versa in case of C. procera although they were growing in the same focal plots. This is might be due to orientation from the main axis i.e. in O. esculenta flowers are hanging downward but in C. procera vertically upward. On the other hand, in P. minima, yellow flowers are also in hanging position but exclusively visited by long and short tongue bees, even in the presence of considerable population of butterflies and dipterous flies in the focal area. Fǽgri and Ven der Pijl

27

(1979) associated bowl shaped flowers with beetles but we observed few beetles only in the flowers of C. procera. Beetles were very rare floral visitors in our study. Fǽgri and Ven der Pijl (1979) themselves warned that names of the pollination syndrome are based on typical pollinators and should not be taken literally e.g. some solitary bees may exert similar selection pressure as that of beetles depending on their behavior in flowers.

10 Disc Umbel Bowl Brush Gu llet Flag Dish Head Tube 9 8 7 6 5 4 3

Number of species 2 1

0

12345678 Clusters Fig 5. The number of plant species with certain flower types per cluster.

The only tube shaped flower i.e. L. camara exhibited strong butterfly and skipper pollination syndrome as described by Fǽgri and Ven der Pijl (1979). The flowers have pinkish tube with dark yellow colored nectar guides and punching smell. It was also visited by some long tongue bees (Amagilla sp., Apis dorsata and Ceratina sp.) and one kind of long tongued Syrphidae (Mesembrius bengalensis). The tube length was 1.04+0.03 cm. No clear syndrome was observed in Astraceous head shaped flowers. Astraceous heads congregate numerous flowers together, enhancing the attractiveness for floral visitors and offering flat surface for them to land (Leppik, 1977). Because of their open structure, heads are visited by large array of insect species from all kinds of functional groups i.e. Coleoptera, Diptera, Hymenoptera and Lepidoptera (Bertin, 1989). These may include ineffective floral visitors and floral herbivores (Waser et al., 1996), which in turn may decrease plant reproductive success. Butterflies can take nectar from wide range of floral forms due to their long proboscis, including tubular corollas and complex

28 zygomorphic flowers to the open flowers with more exposed nectar e.g. actinomorphic Astraceous heads. Brush shaped actinomorphic flowers have exposed pollen and nectar rewards. In most of the studies, no any particular syndrome has been observed in brush shaped flowers. Rather these can even attract birds (Hingston and McQuillan, 2000) as seen in our study i.e. one species of song bird visited P. juliflora and L. leucocephala. Devey and Davidar (2003) have also documented bat pollination in brush shape flowers in western India. The disc shaped actinomorphic flowers in our study, have also shown the strong generalization towards their visitor guilds. There was a great variation in color and size of disc shaped flowers. Other variations may include floral reward, number of petals and presence of ultra violet or visible nectar guides. These all kind of variations may undermine our efforts towards searching a pollination syndrome in the same shaped flowers (Waser et al., 1996; Wilson et al., 2004). Like Astraceous heads, disc shaped flowers also provide comfortable foraging and landing platform to all kinds of insect guilds. Insects with both short and long mouth parts can easily forage these flowers. Gullet and flag shaped zygomorphic flower exhibited strong bee pollination syndromes though it was difficult to separate short tongue bees from long tongue bees in their pollination syndromes. This finding represents the true picture of classical bee- pollination syndrome of Fǽgri and Ven der Pijl (1979). The similar findings about zygomorphic flowers are also evident from various workers (Herrera, 1988; Hingston and McQuillan, 2000). Yellow and pale flag shaped flowers (Vicia sativa, Atylosia minima, Melilotus indica and Sesbania sesban) were more strongly correlated with bees than pink coloured Alhagi graecorum. This finding is in accordance with Hingston and McQuillan (2000). Dish (plate) shaped flowers are morphologically closer to bowl shaped flowers but with less depth than bowl shaped flowers. Like bowl shaped flowers no any particular syndrome was associated with dish shaped flowers. Disc shaped flowers were much closer to bowl shaped flowers therefore must be considered as bowl shaped flowers for the future studies.

29

Beetles Wasps Bees S Bees L Diptera Syrphidae Butterflies Moths Birds 100%

s 80%

60%

40%

Percentage of visitor of Percentage 20%

0% Zizjuj Projul Cirarv Gresub Paracu Daucar Sonasp Euphel Laupro Aspten Capdec Conarv Agecon sp. Con Ranmur Marmin Cluster 1 Cluster 2

Beetles Wasps Bees S Bees L Diptera Syrphidae Butterflies Moths Birds 100% s 80%

60%

40%

20% Percentage of visitor of Percentage

0% Pulcri Triter Suafru Clevis Chotri Halrec Chrtin Tripor Calpro Alhgra Melind Casocc Cucpro Medsat Phymn Atymin Euccam Tamaph Cluster 3 Cluster 4 Cluster 5

Beetles Wasps Bees S Bees L Diptera Syrphidae Butterflies Moths Birds 100%

s 80%

60%

40%

Percentage of visitor of Percentage 20%

0% Sp.16 Dalsis Vicsat Acanil Heleur Veroff Leuleu Malafr Cenpul Abuind Torjap Lataph Malcor Achasp Oxyesc Fumind Anaarv Phynod Lancam Sohanjna Cluster 6 Cluster 7 Cluster 8 Fig 6. Visitor composition of 55 plant species in the study. The floral visitor species were divided into 8 groups, shown as percentage of the total number of flower visitors.

30

Capparis deciduas was extensively visited by a song bird which did not hover like humming birds. C. deciduas has gullet shaped pinkish flowers which is provided with an ample quantity of nectar in specially evolved deep spoon shaped corolla. In most of the cases birds are reported to visit brush and tubular shaped dark colored flowers. But their visitation on gullet shaped zygomorphic flowers has poorly been reported previously. Both of the analysis techniques i.e. MDS and HCA gave almost similar results. This is due to similarity index i.e. Bray and Curtis distance in both the analysis. That is why, short and long tongue bees and syrphids and Diptera were clustered in similar fashion in both the statistical techniques. In contrast to HCA, beetles and birds are fairly separated from each other in MDS ordination plot. Similarly the row HCA and MDS ordination plot of plant species also gave almost the similar picture with few exceptions. In cluster analysis, A. minima, S. fruticosa, P. minima and T. aphylla are grouped in cluster 4 but in MDS ordination plot, these are widely separated in pairs so that the bowl shaped P. minima is well placed among the cluster of other bowl shaped flowers. MDS is a latest technique, most frequently used for pollination syndrome studies (Ollerton and Watts, 2000; Hingston and McQuillan, 2000; Wilson et al., 2004; Ollerton et al., 2009). Unlike other ordination techniques such as principal component analysis (PCA) or detrended correspondence analysis (DCA), MDS does not make assumptions as to the distribution of the variables (Minchin, 1987). Instead, MDS ordinate objects on the basis of rank distances, thus preserving ordered relationship, so that similar objects are closer to each other. Pollination Syndromes in this study: Both the techniques (MDS and HCA) showed almost similar results in this study i.e. generalization in actinomorphic (disc, brush, dish, bowl and umbel) flowers while some sort of specialization in zygomorphic flowers. Actinomorphic flowers with shallow nectarines did not fit in any of the traditional pollination syndrome while zygomorphic flag and gullet shaped flowers exhibited traditional bee pollination syndrome. The actinomorphic tube shaped flowers with deep nectarines exhibited traditional butterfly pollination syndrome.

31

Beetles Wasps Bees S Bees L Diptera Syrphidae Butterflies Moths Birds 100%

s 80%

60%

40%

Percentage of visitor of Percentage 20%

0% ind ind ind ind ind ind ind ind spec spec spec spec spec spec spec spec Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 5 Cluster 6 Cluster 7 Cluster 8

Fig 7. The mean composition of functional groups of 8 identified clusters based on number of species (spec) in that functional group and their individuals (ind).

There were few exceptions which deviated from my overview e.g. the bowl shaped actinomorphic P. minima was exclusively visited by short and long tongue bees while zygomorphic gullet shaped D. sissoo had strong association with syrphidae and Diptera. The short tongue bee pollination syndrome could not be separated from long tongue bee pollination syndromes whereas Syrphidae pollination syndrome could not be separated from Diptera pollination syndrome, therefore the strategy of separating short tongue bees from long tongue bees and Syrphidae from rest of the Diptera proved of no use. The strong similarity in floral preferences between short and long tongue bees; and Syrphidae and Diptera further supports our findings. In conclusion, this study was unable to predict accurately, the classical pollination syndromes in native flora of Punjab, Pakistan. However, some evidences were there regarding bee pollination syndromes (zygomorphic gullet and flag shaped flowers) and butterfly pollination syndromes (tube and bowl shaped flowers).

32

A B C

D E F

G H I

Fig 8. Representative flowers from various flower morphs. (A) bowl shaped Oxystelma esculenta (B) Dish shaped Convolvulus arvensis (C) Gullet shaped Capparis decidua (D) Disc shaped Anagallis arvensis (E) Tube shaped Lantana camara (F) Umbuliferous Torilis japonica (G) Brush shaped Eucalyptus camaldulensis (H) Flag shaped Sesbania sesban (I) Austeraceous head Launaea procumbens.

33

Table 1. Plant species, their floral characters, sampling efforts and floral visitors summary.

Duration of Binomial Floral Morph N N N N Plant Family/Species Shorts Color Orientation Observation code Shapes (A/Z)* locations census individuals Speceis (Minutes)** Aizoaceae Trianthema portulacastrum tp Tripor Disc A White Up/side 3 8 120 70 15 Amaranthaceae Achyranthes aspara aa Achasp Disc A Purple Down 2 6 90 36 12 Apiaceae Daucus carota DC Daucar Umbel A White Up 2 5 75 122 23 Torilis japonica TJ Torjap Umbel A White Up 1 2 30 14 7 Asclepiadaceae Calotropis procera (CP) Calpro Bowl A Purple Up 4 9 135 319 45 Oxystelma esculenta (OE) Oxyesc Bowl A Purple Down 2 8 120 60 11 Asphodelus tenuifolius at Aspten Disc A White Side 3 5 75 52 15 Asteraceae Sonchus asper SA Sonasp Head A Yellow Up 5 8 120 65 18 Cirsium arvense Ca Cirsarv Brush A Purple Up 4 4 60 108 37 Launaea procumbens LP Laupro Head A Yellow Up 6 11 165 196 28 Ageratum conyzoides Ac Agecon Brush A Purple Up 1 6 90 110 23 Pulicaria crispa PC Pulcri Head A Yellow Up 3 3 45 69 21 Boraginaceae Heliotropium europaeum HE Heleur Bowl A White Up 3 7 105 62 26 Brassicaceae Malcolmia africana (ma) Malafr Disc A Purple Up 2 3 45 27 8 Capparidaceae Capparis decidua Cd Capdec Gullet Z Pink Side 4 8 120 146 26 Cleome viscosa cv Clevis Disc Z Yellow Side 2 8 105 26 7

34 Table 1. continued Duration of Binomial Floral Morph N N N N Plant Family/Species Shorts Color Orientation Observation code Shapes (A/Z)* locations census individuals Speceis (Minutes)**

Chenopodiaceae Suaeda fruticosa sf Suafru Disc A Green Up/side 4 4 60 32 13 Haloxylon recurvun hr Halrec Disc A Yellow Up/side 2 2 30 55 13 Convolvulaceae 5 9 135 128 28 Convolvulus arvensis ca Conarv Dish A White Up 2 12 180 111 30 Convolvulus sp. csp Con sp. Dish A White Up

Cucurbitaceae

Cucumis prophetarum cp Cucpro Disc A Yellow Side 6 10 150 91 23

Euphorbiaceae Euphorbia helioscopia Eh Euphel Disc A Green Up 3 3 45 50 11 Chrozophora tinctoria ct Chrtin Disc A Yellow Up 2 3 45 43 24 Fabaceae 6 8 120 165 28 Prosopis juliflora Pj Projul Brush A Pale Side 2 2 30 36 8 Dalbergia sissoo Ds Dalsis Gullet Z Pale Side Ll Brush A Pale Up 1 4 60 43 12 Leucaena leucocephala Leuleu pa Disc A Yellow Up/side 1 3 45 65 18 Parkinsonia aculeata Paracu Flag 3 4 60 54 8 Melilotus indica MI Melind Z Yellow Side 2 2 30 33 10 Alhagi graecorum AG Alhgra Flag Z Pink Side 1 3 45 13 3 Atylosia minima AM Atymin Flag Z Yellow Side 2 2 30 2 2 Vicia sativa VS Vicsat Flag Z Purple Side 1 10 150 82 9 Cassia occidentalis Co Casocc Gullet Z Yellow Side 2 2 30 5 2 Lathyrus aphaca LA Lataph Flag Z Pale Side SS Flag Z Yellow Side 3 5 75 49 8 Sesbania sesban Sesses Ms Flag Z Pale Side 2 4 60 118 21 Medicago sativa Medsat

35 Table 1. continued Duration of Binomial Floral Morph N N N N Plant Family/Species Shorts Color Orientation Observation code Shapes (A/Z)* locations census individuals Speceis (Minutes)**

Fumariaceae 2 3 45 5 3 Fumaria indica Fi Fumind Gullet Z Purple Side

Gentianaceae Disc Up 2 3 45 16 9 Centaurium pulchellum (cp) Cenpul A Pink

Malvaceae Disc Side 3 7 105 47 16 Abutilon indicum ai Abuind A Yellow Malvastrum mc Disc A Yellow Side 2 9 135 55 22 coromandelianum Malcor Brush Up 1 3 45 95 28 Grewia subinaequalis Gs Gresub A Yellow Marsiliaceae Marsilia minuta mm Marmin Disc A Yellow Up 4 6 60 88 18

Mimosaceae Brush Side/up 3 8 120 33 21 Acacia nilotica An Acanil A Yellow

Myrtaceae Brush Side/up 2 3 45 56 11 Eucalyptus camaldulensis Ec Euccam A Pale

Primulaceae Disc Up 2 4 60 19 7 Anagallis arvensis (aa) Anaarv A Blue

Ranunculaceae Disc Side/up 4 4 60 88 14 Ranunculus muricatus rm Ranmur A Yellow

Rhamnaceae Disc Side 2 7 105 140 21 Ziziphus jujuba zj Zizjuj A Green

Solanaceae Bowl down 1 6 90 36 7 Physalis minima PM Phymin A Yellow

Tamaricaceae Brush Side 2 3 45 55 18 Tamarix aphylla Ta Tamaph A Pink

36

Table 1. continued Duration of Binomial Floral Morph N N N N Plant Family/Species Shorts Color Orientation Observation code Shapes (A/Z)* locations census individuals Speceis (Minutes)**

Tiliaeae 2 5 75 28 6 Chorchorus tridens cti Chotri Disc A Yellow Side

Verbenaceae Tube Up 3 9 135 133 22 Lantana camara (LC) Lancam A Yellow Bowl Up 5 7 105 73 20 Phyla nodiflora (PN) Phynod A White Bowl Up 2 8 120 48 14 Verbena officinlis VO Veroff A White

Zygophyllaceae Disc Up 6 9 135 136 24 Tribulus terrestris tt Triter A Yellow Sp-16 Sp-16 Gullet Z White Side/up 2 4 60 18 11 Bauhinia variegata Bv Bauvar Gullet Z White side 1 2 30 33 6

*A=Actenomorphic, Z=Zygomorphic **Since each individual plant was observed for one minute, therefore same figures are applied for the number of individuals per plant species observed.

37

Table 2. Bees Bees Plant Species Beetles Wasps Diptera Syrphidae Butterflies Moths Birds S L Aizoaceae Trianthema portulacastrum 1 19 29 5 15 1 Amaranthaceae Achyranthes aspara 2 11 15 1 2 5 Apiaceae Daucus carota 1 3 21 47 50 Torilis japonica 2 1 7 4 Asclepiadaceae Calotropis procera 2 24 84 163 13 13 6 14 Oxystelma esculenta 7 5 4 2 32 10 Asphodelus tenuifolius 2 7 2 35 2 4 Asteraceae Sonchus asper 7 10 29 16 2 1 Cirsium arvense 1 3 4 44 17 24 14 1 Launaea procumbens 2 4 42 16 7 103 13 9 Ageratum conyzoides 6 23 10 45 22 4 Pulicaria crispa 6 11 41 2 4 3 2 Boraginaceae Heliotropium europaeum 1 3 14 15 4 1 8 16 Malcolmia africana 1 4 8 1 13 Capparidaceae Capparis decidua 3 7 38 45 23 18 12 Cleome viscosa 1 3 19 1 1 1 Chenopodiaceae Suaeda fruticosa 1 19 2 5 4 1 Haloxylon recurvun 6 5 18 16 9 1 Convolvulaceae Convolvulus arvensis 10 19 26 26 33 2 12 Convolvulus sp. 13 19 33 5 27 10 4 Cucurbitaceae Cucumis prophetarum 20 24 7 32 8 Euphorbiaceae Euphorbia helioscopia 3 1 41 5 Chrozophora tinctoria 2 6 26 1 4 4 Fabaceae Prosopis juliflora 2 9 17 66 21 47 2 1 Dalbergia sissoo 1 15 3 17 Leucaena leucocephala 5 3 6 5 20 2 2 Parkinsonia aculeata 2 26 2 35 Melilotus indica 6 43 2 3 Alhagi graecorum 4 23 3 3 Atylosia minima 9 2 2

38 Table 2. Continued

Bees Bees Plant Species Beetles Wasps Diptera Syrphidae Butterflies Moths Birds S L Vicia sativa 1 1 Cassia occidentalis 32 47 1 2 Lathyrus aphaca 5 Sesbania sesban 48 1 Medicago sativa 2 82 5 4 23 2 Acacia nilotica 8 9 1 11 4 Bauhinia variegata 8 9 1 11 4 Fumariaceae Fumaria indica 1 4 Gentianaceae Centaurium pulchellum 2 4 10 Malvaceae Abutilon indicum 1 2 10 4 3 17 10 Malvastrum 2 9 10 2 12 19 1 coromandelianum Grewia subinaequalis 5 4 47 12 22 5 Marsiliaceae Marsilia minuta 8 14 59 6 1 Myrtaceae Eucalyptus camaldulensis 2 1 32 10 11 Primulaceae Anagallis arvensis 4 3 8 4 Ranunculaceae Ranunculus muricatus 2 2 8 54 22 Rhamnaceae Ziziphus jujuba 2 30 15 88 4 1 Solanaceae Physalis minima 22 14 Tamaricaceae Tamarix aphylla 1 9 19 19 6 1 Tiliaeae Chorchorus tridens 3 23 2 Verbenaceae Lantana camara 5 1 16 1 7 49 54 Phyla nodiflora 13 7 13 17 23 Verbena officinlis 3 2 6 1 33 3 Zygophyllaceae Tribulus terrestris 6 47 35 5 9 32 2 Sp-16 1 1 2 2 6 6

39 Chapter 4

Yearlong plant-pollinator network

Summary

The network properties of flowering plants and their floral visitors were studied year around in a 20 hectare planted forest and adjacent agricultural landscape at Bahauddin Zakariya University Multan, Pakistan. Eighty plant species and 162 insect species resulted in 1573 unique interactions with a cumulative connectance of 12.31%. The frequency distribution of floral visitor interactions was less uniform than those of plants. Network size was smallest in winter and late autumn when interactions were low while it was the maximum in mid and late spring, the time when the maximum number of interactions was also observed. Connectance ranged from 9.63% to 52.38% (average=19.43%) and was negatively associated with system size. The variation in the number of interactions followed changes in the number of plant mutualists more strongly than that of animal mutualists. System size increased with the increase in number of interactions. Cumulative distribution of number of interactions [p(k)] per plant and their floral visitor (k) showed a better fit to a truncated power law curve. The results also fit well to an exponential function when plotted connectance against system size. Nine properties were also studied for monthly networks: number of plant species, number of visitor species, number of unique interactions, connectance, weighted nestedness, interaction evenness, interaction diversity, network level of specialization and robustness. Most network properties (particularly the number of plants, visitors and unique interactions) varied markedly during the year and all, except for one (interaction evenness), were significantly affected by the aggregation of temporal data. Results support previous studies that suggest the need to explicitly consider temporal dynamics of networks when describing their structure and highlight the possible bias that merging temporal data may cause, particularly in temperate communities where both plants and insect species are active year-round.

40

Introduction

The study of mutualistic networks is important for understanding the organization of biodiversity in ecological systems. Some species interact strongly with each other, but others interact weakly (Hwang et al., 2008). In view of this great variation, a community- level plant-pollinator network will produce a more realistic description of the pollination relationships than individual autoecological studies (Waser et al., 1996; Medan et al., 2002). These networks describe the interactions that occur between flowers and flower visitors at a given place and time (Olesen and Jordano, 2002). Network analysis of plant-pollinator communities has recently been used to study the plant pollinator relationships, particularly to detect the changes in ecological gradient (Elberling and Olesen, 1999), to study the degree of specialization and generalization among mutualists (Medan et al., 2002; Basilio et al., 2006), to restore the damaged ecosystems (Forup and Memmott, 2005; Forup et al., 2008) and to identify functional compartments (Dicks et al., 2002; Fontaine et al., 2006). Mutualistic interactions between animals and plants form several intricate interaction webs (Memmott, 1999). Recent analysis of plant-pollinator interaction webs demonstrate that these contain a continuum from fully specialist to fully generalist species (Olesen and Jordano, 2002; Jordano et al., 2003; Lundgren and Olesen, 2005; Basilio et al., 2006). However, topology of these networks is nested i.e. specialists tend to interact with sets of partners that are subsets of the partners interacting with more generalists (Bascompte et al., 2003; Olesen et al., 2008; Petanidov et al., 2008). Several methods of studying specialization and generalization have been investigated e.g. Compartmentalization, power law distribution, truncated power law distribution, grouping on the basis of functional diversity etc. Compartmentalization describes the way a food web can be divided into subsections, such that organisms within a compartment interact more strongly with one another than with species in other compartment of the web (Dicks et al., 2002). A number of studies documented that the degree distribution of interactions in a community follows a power law distribution with exponential truncation (Jordano et al., 2003; Vázquez and Aizen, 2004; Lundgren and Olesen, 2005; Petanidou and Potts, 2006; Guimarães et al., 2007) which reflects that most species are specialists. However, few other studies have shown a power-law regime

41 for animal taxa and more clearly, a truncated power-law regime for plant taxa (Jordano et al., 2003; Basilio et al., 2006). The bulk of studies are static descriptions of the structure of networks, and a few studies explore the temporal dynamics of pollination networks (e.g., Petanidou, 1991; Medan et al., 2002; Lundgren and Olesen, 2005; Basilio et al., 2006; Kaiser, 2006), and all conclude that several variables have strong temporal dynamics, e.g., species number, species linkage level (number of links of a species to other species), total number of links in the network, network connectance and nestedness. Most of them explored the qualitative metrics and overlooked quantitative metrics. One noteworthy consequence of the temporal dynamics shown by networks is that most network descriptions will be biased in some way as they all aggregate temporal data to some extent. Basilio et al. (2006) illustrated the effect of aggregating temporal data by identifying three potential sources of bias when merging a whole year of observations into a single network. First, species that apparently have the potential to interact may actually have non-overlapping phenologies (Olesen and Jordano, 2002; Jordano et al., 2003; Basilio et al., 2006). Second; species with prolonged phenologies may appear to have more interactions than they typically have at any particular moment (Waser et al., 1996). Third, possible seasonal changes that may occur in system size, symmetry, connectance, species turnover and degree of generalization of the species involved may be overlooked (Basilio et al., 2006). A limitation shared by temporal descriptions published so far -except for (Baldock et al., 2011)- is that, because of the way the data was gathered in the field, only qualitative networks can be constructed. This limits our ability to understand changes in the quantitative aspects of these networks and, equally important, the effect of data aggregation on quantitative metrics. In this context, the aims of the present study are two-fold: 1) to describe the temporal dynamics of a quantitative plant-pollinator network along a whole year, and 2) to assess the effect of temporal data aggregation on qualitative and quantitative network properties. These aims were addressed in the context of a sub-tropical forest park in Punjab, Pakistan.

42

Materials and Methods Study Sites: The study was done at 30 hectare forest park of University campus of Bahauddin Zakariya University Multan (30.255°N; 71.513°E; 114 ± 6 meter above sea level). The forest has never been cultivated except some supplementary plantation of Eucalyptus camaldulensis, Dalbergia sissoo, Acacia nilotica, Albizzia procera and Leucaena leucocephala. Besides these trees, a large array of naturally growing annual or perennial weeds, shrubs and trees grow in the forest. The area is located in the Southern Irrigated Zone of Punjab (PARC, 1980). The mean monthly temperature ranges between maximum of 35 °C to 40 °C and minimum of 10°C to 20°C. The extreme maximum temperature of the region varies between 45°C and 51°C during the months of May and June, while the lowest minimum temperature varies between 0°C to -5°C during the month of January. The annual average monthly summer rainfall is ca. 18 mm (Khan et al., 2010). The region has four distinct seasons i.e. winter (December, January and February), spring (March, April and May), summer (June, July and August) and autumn (September, October and November). Geographically, it is an alluvial plain with fertile soils deposited by the flood regime of the rivers through thousands of years. Most of the land is cultivated and irrigated by canals or underground waters. A variety of crops are grown but cotton and wheat rotation mostly is the tradition. Floral visitor censuses: The interactions between plants and their floral visitors was investigated in a large array of vegetation types i.e. planted trees, annual wild plants, perennial shrubs and 5 crop plant species grown as a part of agro-forestry on 2 acres. Out of which Brassica napus, Raphanus stivus, Allium cepa flowered in spring while Helianthus annuus and Abelmoschus esculentus flowered in summer. Fortnightly census of each flowering plant species were done from the very beginning to the end of its flowering period. I defined the floral units for each plant species separately since different plant species had different inflorescence types. I recorded only diurnal floral visitors on the plants having diurnal anthesis. During a census, fifteen individuals of a given plant species were randomly selected and observed for 60 seconds. Thus, there was a total of 15 minutes of observations per plant species per census. For each individual plant, I counted the number

43

of visiting individuals per floral visitor species by visual observations. All the plant species were identified by local well renowned experts (see acknowledgement). All the insect specimens were first morphotyped and then identified to the lowest possible taxonomic level. The family level identification of Diptera and Coleoptera was done by using keys of Borror et al., (1981). The syrphid fly and butterfly species were identified by experts (shown in acknowledgement). Bee genera were identified by using the keys of Michener (2007). Voucher specimens were deposited at the Agricultural Museum of the University College of Agriculture, Bahauddin Zakariya University Multan. Data analysis: I computed connectance (C) as C = I/(v × h) × 100 for every month, where "I" stands for interactions, "v" for visitors and "h" stands for hosts in that particular month. A cumulative connectance for the complete study period (i.e. assuming that all mutualists were present all year round) was also calculated by the same formula. Degree distributions were calculated separately for plants and pollinators (following Jordano et al., 2003) for each month using the function ‘networklevel’ from the package bipartite (Dormann et al., 2008) run in R (R Development Core Team, 2008).

Objective 1: describe the monthly dynamics of a quantitative plant-pollinator network along a whole year To describe the temporal changes in network structure along the year, interaction data for each month was analyzed as a separate network. Nine properties were calculated for each monthly network: number of plant species, number of visitor species, number of unique interactions, connectance, weighted nestedness, interaction evenness, interaction diversity, network level of specialization and robustness to species loss. A brief description of each property is given below. The number of unique interactions was calculated as the number of non-zero cells in the matrix representation of the interaction network. Weighted nestedness was calculated following the method developed by Galeano et al. (2009), which, unlike the original nestedness (Bascompte et al. 2003), considers interaction frequencies in the calculation of nestedness. Nestedness can be visualized when plants are ranked from the most

44

specialized to the least specialized. This reordering will show that the set of animals a plant interacts with are contained in a larger set, which in turn is contained in a larger set, and so on. In other words, specialists (species holding few interactions) interact with species that form well-defined subsets of the species with which generalists (holding many interactions) interact. A deviation from this pattern (for instance, a specialist-to- specialist interaction) is considered to reduce nestedness, which ranges between 1 (perfect nestedness) and 0 (perfect chaos). Interaction evenness treats the interactions between each pair of species in a matrix, which occur at different frequencies, as “species” which occur at different “abundances”. In this way, applying the Shannon's evenness to an entire matrix calculates the evenness of interactions in the same way the evenness of a community would be calculated (Bersier et al. 2002; Blüthgen et al. 2006). It has been described as “a measure of the uniformity of energy flows along different pathways” in a network (Tylianakis et al. 2007). Interaction diversity, which is also based on Shannon’s diversity index, calculates the diversity of pair-wise interactions (Blüthgen et al. 2006; Blüthgen et al. 2008). The network-level of specialisation is a measure of specialisation based on the deviation of a species’ realized number of interactions and that expected from each species’ total number of interactions. This metric ranges between 0 (no specialization) and 1 (perfect specialisation for given interaction totals). For a full explanation on this metric please refer to Blüthgen et al. (2006). The robustness of a network measures its response to the loss of species. If a given fraction of species of one guild (for instance, the pollinators) are eliminated, a number of species of the other guild (e.g. plants) which depend on their interactions become extinct. In this way a curve can be built to describe the occurrence of secondary extinctions. Robustness ranges between 0 for minimum and 1 for maximum robustness. All network metrics were calculated with the R package ‘bipartite’ (Dormann et al. 2008). Monthly values of each metric were plotted for a whole year.

Objective 2: to estimate the effect of merging temporal data on network structure To assess the effect on network properties of the progressive aggregation of monthly interaction data we constructed networks encompassing time periods of increasing length,

45 from one to six months (following Medan et al. 2006). We started by merging interaction data from all possible pairs of consecutive months (January + February, February + March, etc.). Arithmetically, this means adding the corresponding pair of matrices. We repeated the process for periods of three to six months. For each network of merged data we computed the same properties that were used to describe the monthly networks. As a means to estimate the degree to which each network property was affected by data merger the coefficient of variation of the grouped means across all aggregation levels was calculated.

46

Results The overall size of our observed system was 80 plant species, 162 insect species and 1573 unique interactions. Cumulative connectance was 12.31%. The maximum number of insect species was recorded in the month of April while the plants were maximized in the months of March, April and May. Afterwards (Jun to December), both the trophic levels behaved in the same fashion of gradual decrease until December. Nineteen plant species had 13 or more interactions at least during one month. A maximum of 31 interactions in single month were recorded on Calotropis procera during the month of April and overall, it comprised 4.57% of the total interactions of the system followed by Malvastrum coromandelianum, Launaea procumbens, Tribulus terrestris and Cucumis prophetarum ca. 3% of interactions each. C. procera also had highest number of mutualists (46 visitor species). The number of mutualists of a plant species increased with the increase in length of its flowering period (Plant linkage level = 6.45+2.70 x flowering period in months, r²=0.38, p<0.001). Throughout the study, the most species-rich families among insects were Syrphidae (14 species), Apidae (10), Halictidae (5) and Megachilidae (8). Between month changes in the proportion of diverse insect orders were significant in all the four insect orders (χ² test, P<0.001). During winter, flies (Diptera) increased and bees and wasps (Hymenoptera) declined, but in spring, summer and autumn, bees and wasps remained dominant over flies (Fig 3). On an average, each insect species visited 7.2±9 plant species (mean±SD) and each plant species was visited by 15±8.6 (Mean±SD) insect species, thus the generalization level of insects was lower than that of plants. The frequency distribution of visitor interaction was less uniform than those of plants (Fig 1, 2). As a consequence, the distribution of animal interactions significantly departed from normality (Kolmogorov- Smirnov D–test, P<0.05).

47

0.012

0.01

0.008

0.006

0.004

species no.mutualistic of 0.002

0 0 100 200 300 400 no. of interactions per mutualistic species

Fig 1. Frequency distribution of generalization level of plant hosts at Southern Punjab, Pakistan. A normal fit corresponding to sample values is shown in each histogram plot.

0.018

0.015

0.012

0.009

0.006

no. of mutualistic species 0.003

0 0 100 200 300 400 500 no. of interactions per mutualistic species

Fig 2. Frequency distribution of generalization level of insect visitors at Southern Punjab, Pakistan. A normal fit corresponding to sample values is shown in each histogram plot.

48

Ten insect species had 13 or more interactions at least during one month. A maximum of 26 interactions in single month were recorded by Sphaerophoria bengalensis during the month of April but overall, it comprised only 1.8% of the total interactions of the system. The most interaction-rich flower visitors included Eristalinus aeneus, Halictus ps. Apis dorsata, Apis florea, Ischidon scutellaris and Ceratina sexmaculata (3 to 4.4% of the total interactions). These most interaction-rich flower visitors (except Syrphidae) had long activity periods (>9 months) and interacted with >40 mutualist plant species. The number of mutualists of a floral visitor species increased with the increase in length of its active period (months) (insect linkage level = - 1.477+2.66 x flowering period in months, r²=0.624, p<0.001).

Diptera Hymenoptera Lepidoptera Coleoptera 1800 1600 1400

s 1200

1000

800

Population 600 400 200 0 Winter Spring Summer Autumn Seasons

Fig 3. Seasonal variation in composition and abundance of diverse insect orders at southern Punjab, Pakistan.

Connectance ranged from 9.63% to 52.38% (average=19.43%) and was weakly negatively associated with system size (Connectance = 27.18-7.27[System size]; R²=0.307; p=0.062). The activity peaks (highest system sizes) of plant-visitor matrix were observed during the months of spring (March, April and May). The connectance value was lowest (11.67%) for spring (March, April and May) and the value for individual summer (June, July and August) (12.59%), winter (December, January and February) (38.79%) and autumn (September, October and November) (14.68%) months were greater than cumulative connectance (12.31%).

49

Cumulative distribution of number of interactions [p(k)] per plant and their floral visitor (k) showed a better fit to a truncated power law curve (Table 1) [p(k)=k-γ exp(- k/kx)] than to a power law curve [p(k)=k- γ] (in both expressions γ is the fitted constant and kx is the truncated value. The variation in the number of interactions followed the changes in the number of plant mutualists more strongly and significantly (Pearson's correlation coefficient= 0.750 at p=0.05) than that of animal mutualists (Pearson's correlation coefficient= 0.500 at p=0.05). System size increased with the increase in number of interactions along the year (Interactions = 55.95+7.10[System size]; R²=0.44) (Fig 4). The results also fit well to an exponential function when plotted connectance against system size Connectance = 43.26 x Exp (-1.57-03 x System size); R²=0.71] (Fig 5).

400 350

300 250

200

Interactions 150 100

50

0

0 1000 2000 3000 4000 System size

Fig 4. Relationship between annual interactions and system size in Southern Punjab, Pakistan. The best fit regression equation is Y = 22.4994 + 8.336X, P< 0.001, R²= 0.982

The maximum number of active insect species was recorded in April, while March, April and May were the months of highest plant richness (Fig. 2). Coincidentally, the highest numbers of unique interactions were recorded in the period March-June (Fig. 6). After the sharp increase in the period January-April, a gradual decline in the number of flowering plant species was observed between May and December. The number of

50

visitor species, however, remained at a plateau from June to October followed by a sharp decline in the following two months (Fig. 2). Seasonal variations in weighted nestedness and interaction diversity seemed to follow the variations in number of plant and insect visitor species. Interaction evenness was lower in December and January; thereafter it gradually increased until June. The maximum interaction diversity was recorded in December where there was a maximum number of plant and visitor species. The most specialized interactions occurred between October and February (Fig. 6). Regarding the effect of data aggregation, the number of animal and plant species and the number of unique interactions among them were the properties most affected by data aggregation (as shown by their high coefficients of variation across aggregation levels, Fig. 7). Interaction evenness was by far the property least affected by data aggregation. The rest of the properties measured (connectance, weighted nestedness, interaction diversity, specialization and robustness) showed intermediate coefficients of variation. Regarding the direction of the change in network properties caused by data aggregation, the number of plants, animals and unique interactions, weighted nestedness, interaction diversity and robustness increased, while connectance and specialization decreased with increasing data aggregation. Finally, interaction evenness showed no clear trend in response to data aggregation.

51

Table 1. Fitted models to the distribution of interactions per species for the plant-flower visitor network of the southern Punjab, Pakistan.

Power Truncated Truncated Power law law power law power law Months (Plants) (Insects) (Insects) (Plants) γ γ γ γ W1 - - - - W2 - - - - W3 0.48228 0.52447 0.22325* 0.02114* Sp1 0.56692 0.50918 -0.68279* -0.57917* Sp2 0.35558 0.39667 -0.39294* -0.45334* Sp3 0.37836 0.40426 -0.48277* -0.49796* Su1 0.48356 0.48094 -0.49352* -0.49077* Su2 0.47068 0.46503 -0.59464* -0.51106* Su3 0.50311* 0.67849 0.45297 0.19751* Au1 0.41318* 0.57929* -0.12839 - Au2 0.15866* 0.46163* - - Au3 - - - -

*Values with highest R² and lowest AIC (Akaike's Information Criterion) values. -Values are provided for each month except where data were insufficient for calculation.

60

50

40

30

Connectance 20

10

0 0 1000 2000 3000 System size

Fig 5. Relationship between connectance and system size in the Southern Punjab, Pakistan. Solid circles, winter; open triangles, autumn; solid rectangles, summer; open circles, spring, overall connectance value. An exponential fit is added to the plot.

52

Discussion

The concept of sequential webs has solved many topological problems of cumulative webs when studied for a longer periods of time (Basilio et al., 2006). Although cumulative webs have the advantage of including all the species of any community in only one diagram or table, but many important parameters are overlooked i.e. possible changes in system size, assemblage composition, species turnover, connectance and degree of specialization of the community pollination cycle (Basilio et al., 2006). On the other hand, studying the bimonthly (consecutive) webs yet has the possibility of including non-overlapping species, which could decrease the assemblage similarity between the months (Basilio et al., 2006).

The degree of specialization is undermined by the fact that species with extended phenologies may hold more attractions than they have at any particular time, which may lead to overstated generalization scores (Basilio et al., 2006). In sub-tropical areas of Pakistan, there are four distinct seasons (winter, spring, summer and autumn) with remarkable variation in abiotic factors and species turnover and composition and hence their activity duration (several days to several months) as seen in this study, consecutive webs reflect the pattern of interactions during a discrete time span, would describe interactions only among partners with coincident phenologies, and how the connectance and system size changes with discrete time span (Basilio et al., 2006).

To date, studies involving total community of interacting plants and flower visiting animals are much rare. Previously, not even a single community level plant- pollinator network has been reported from oriental region. This might be because of two important factors (Jordano, 1987) i.e. firstly it requires a labor-intensive sampling procedure and secondly several animal taxa pose taxonomic difficulties.

We observed more generalization among plant species than pollinator species. Olesen and Jordano (2002) analyzed 29 pollinator networks around the world and concluded that there is no hard and fast rule for determining generalization between plants versus pollinators. However, they observed some oceanic island networks with rather lower Animal to Plant ratios (A: P) than mainland networks but having marked

53 trend for decreasing plant generalization level, independent of any parallel trend in network size. In our study, A: P ratio was 2:1 but the generalization in plants was two fold than that of visitors. A high density and diversity of flower visitors may cause more interspecific competition, reduced niche overlap, and cause specie to become more specialized (Olesen and Jordano, 2002). However, a high density may not necessarily result in more specialization if the spectrum of exploited resources is increased concordantly (Olesen and Jordano, 2002).

Armbuster et al. (2000) distinguished between evolutionary and ecological specialization in plant-pollinator interactions. The first category is defined as the process of evolution towards more specialization, whereas the later is a state and refers to having one to few interactions. In food web and network studies, we are only dealing with ecological specialization (Olesen and Jordano, 2002) which actually may not represent the evolutionary specialization.

In our study, the distribution of interactions in insect species departed from normality. The non-normal distribution of the number of mutualistic interactions is a common phenomenon, at least for flower visitors (Basilio et al., 2006) but this is not the rule (Medan et al., 2002). This non-normal distribution of interactions is the result of a few mutulists having many interactions and many mutualists having few interactions (Moldenk and Lincoln, 1979; Elberling and Olesen, 1999; Jordano et al., 2003; Basilio et al., 2002). The non-normal distribution of flower visitor interactions is another proof of more specialization in flower visitors than plant species in this study. Among floral visitors, E. aeneus, A. dorsata and Halictus sp. were the most interaction-rich mutualists but had lower generalization level (≤ 40 plant mutualistic species) than that of interaction rich plant, C. procera (46 insects mutualistic species).

These interaction-rich species must be called as topological key-stone species (quantification of elements of any ecological network, Libralato et al., 2006), since they hold considerable proportion of pollinators in terms of their interactions and diversity. The information on topological key-stone species can provide basis for setting conservation priorities (Benedek et al., 2007). Key-stone species are more important to

54 be conserved since their deletion could cause a severe damage to the reliable functioning of any system by damaging co-evolutionary mutualistic counterparts (Jordan et al., 2002). Our results also exhibited exponential truncation both for higher (pollinators) and lower (plants) trophic levels. A number of studies documented that the degree of specialization in a plant-pollinator community follows a power law distribution with exponential truncation (Jordano et al., 2003; Vazquez Vázquez and Aizen, 2004; Lunaudgren and Olesen, 2005; Petanidouv and Potts, 2006; Guimarãaes et al., 2007). This reflects that most species are specialists (i.e. interacting with only one or just a few partners) whereas a few species have a number of interactions much higher than the average (Kallimanis et al., 2009). This might be because of the fact that interactions would not be possible between species (forbidden links) due to non-overlapping phenologies or non-matching morphologies. Jordano et al., (2003) suggest that this structural constraint would impose an upper limit on the number of links per node, causing the degree distribution to drop off more rapidly than under a power law. However, few other studies have shown a power-law regeime for animal taxa and more clearly, a truncated power-law regieme for plant taxa (Jordano et al., 2003; Basilio et al., 2006). Degree distribution is related to connectance. Networks having lower connectance, show power-law distributions while higher connectance webs often display less skewed exponential or uniform distributions (Dunne et al., 2002). The connectance in the discrete monthly webs in this study was lowest for spring season (average: 11.67%) and highest for winter season (average: 38.79%) but the networks followed the truncated power-law for both plant and animal taxa. This might be because of the much higher connectance values of our discrete seasonal webs, than webs reported by Dunne et al. (2002) i.e. 0.03% (minimum) and 0.3% (maximum).

55

Fig 6. Monthly variation of nine properties of a year-long plant-pollinator network. For each property data points represent the values of individual monthly network (month 1 = January).

56

Fig 7. Variation of nine properties of a year-long plant-pollinator network as a function of the number of the merged monthly data considered for calculation.

57

This study described monthly variations in the structure of a quantitative plant- visitor network and explored the effect of merging monthly data on network descriptors. This is the first study to provide a quantitative description of the seasonal changes shown by a plant-visitor interaction network. It is also the first study to assess the effect of temporal data aggregation on quantitative network descriptors. In this section we will first address the main limitations of this study then discuss our results in the context of our objectives. A potential limitation of our study is that sampling effects may have influenced observed network structure in at least two possible ways. First, some network properties such as connectance clearly scale with the number of species in the network analyzed (Jordano, 1987). This does not imply that the variations in connectance should be overlooked, but rather that it is reasonable to bare in mind that this change is an artifact and does not reflect a seasonal shift in behavior of the species in the network (Blüthgen et al., 2008). Second, previous studies that described the temporal dynamics of plant- pollinator networks (Basilio et al., 2006; Alarcón et al., 2008; Olesen et al., 2008; Petanidou et al., 2008; Baldock et al., 2011) have pointed out the role that the abundance and phenophase length (i.e. the length of a species’ period of activity) seem to have on network structure as they can influence the detection probability of pairwise interactions (Blüthgen et al., 2008; Vázquez et al., 2009). Thus, interactions involving abundant species with long phenophases may be easier to observe than those involving rare species with short phenophases (Vázquez et al., 2009), resulting in that the latter species may appear as more specialized simply because of this sampling bias. In our study, however, the most generalist species belonged to the Syrphidae, although species in this family tended to have rather short phenophases (March to April). In this case, the short phenophase is compensated by the large abundances recorded in the field by the species in this family (Stang et al., 2006; Dormann et al., 2009; Vázquez et al., 2009). In this study I have shown that all network properties, both qualitative and quantitative, showed significant seasonal variations. This is consistent with previous studies that show that network properties (connectance, nestedness, modularity, etc.) vary significantly within a single the year (Basilio et al., 2006). Our results, thus strengthen

58

the idea that a proper description of a network with year-round activity should include all seasons and not just the few months of spring and summer. Different studies have shown variable levels of temporal variations on network properties (Olesen et al., 2008; Petanidou et al., 2008) however, seasonal variations are more prominent than interannual variations (Olesen et al., 2008). In this study, the seasonal variations in nestedness and interaction diversity followed the variations in number of plant and animal species. The seasonal changes in weighted nestedness and network-level specialization suggest that species exhibited more specialized interactions during the time of the year when the number of species and interactions were at a minimum. Since specialization is not affected by system size and sampling intensity (Blüthgen et al., 2006) our finding regarding seasonal changes in specialization seem robust. Although counterintuitive, this could mean species behave as generalists and forage less selectively when resources are abundant, but become more selective when resource are scarce. Our simulations of data aggregation showed that most properties, both qualitative and quantitative, are affected by temporal aggregation. Our findings are in accordance with previously reported qualitative studies (Basilio et al., 2006; Medan et al., 2006). Their results suggest that the analysis of a system which is active along the whole year should be done at shorter, biologically relevant periods due to the introduction of “forbidden links” which may bias the results when longer periods are considered (Jordano et al., 2003). The aggregation of network data for long periods may give rise to many non-coincidental plants and pollinator phenologies, resulting in, for instance, decreased network connectance. Cumulative webs may bias the results whereas comparing seasonal or sequential webs more explicitly shows the network dynamics.

59

Chapter 5

Variation in abundance and composition of Hoverflies (Diptera: Syrphidae) around the year

Summary

Species composition and population dynamics of hoverflies (Diptera:Syrphidae) in relation to some abiotic and biotic factors were studied over a year long period in the District Multan, Pakistan. The community of hoverflies was composed of 14 species which were recorded from 59 plant species. Among Syrphinae, Ischiodon scutellaris, Episyrphus balteatus and Sphaerophoria bengalensis were the most abundant whereas among Milesiinae, Eristalinus aeneus and Eristalinus laetus were the most frequent floral visitors. The peak abundance and richness of hoverflies was observed in spring (March-April), the time when the maximum numbers (35) of plant species were in flowering. Only four species i.e. E. aeneus, E. laetus, Mesembrius bengalensis and Paragus serratus remained active all through the year in low or high abundance. Among agricultural and wild plant species, Mangifera indica and Launaea procumbens were visited by the maximum number of syrphid species or in highest abundance, respectively. On the basis of similarity in floral host plant visitation frequencies, Syrphinae could easily be distinguished from Miliesiinae. Abundance of hoverflies was positively correlated with the floral abundance and flowering plant species, while temperature and relative humidity were negatively or only weakly correlated.

60

Introduction

Hoverflies are a very important group of insects because their ecosystem services are manifold. Their larvae exhibit a variety of feeding modes i.e. aphidophagous, saprophagous, zoophagous and phytophagous (Sommaggio, 1999), whereas adults are floral visitors of hundreds of different plant species (e.g. Tooker et al., 2006). Adults use nectar for energy and/or pollen for proteins, lipids and vitamins (Fægri and Van Der Pijl, 1979). These floral resources enhance the longevity and fecundity of adult dipterous flies (Shahjahan, 1986).

Many syrphid species also have been documented as efficient crop pollinators (Sajjad et al., 2008). Hoverflies are a characteristic feature of spring season (March-April) in the sub- tropical areas of Pakistan where the average temperature and relative humidity ranges from 23°C to 59%, respectively.

The basic biology of most of the fly pollinators is poorly understood (Kearns, 2001), more so for oriental species, and which may be responsible for the minimal conservation focus for this order (Kearns, 2001). Knowledge of seasonal abundance and diversity of pollinators in relation with floral abundance and abiotic factors has generally been documented as helpful in setting up their conservation strategies (Souza-Silva et al., 2001; Hegland and Boeke, 2006;

Shebl et al., 2008). At the landscape level, positive relationships between the richness and abundance of floral resources and pollinator diversity and activity have been found (Dewenter et al., 2002; Potts et al., 2003). But on a micro scale very little is known about the overall activity patterns of pollinator in relation to the distribution of floral resources and ultimately its effect on the carrying capacity of pollinator populations (Hegeland and Boeke, 2006). Bees are the only taxon which is thoroughly considered in this context (Juker and Wolters, 2008) and very little attention has been given to Diptera pollinators (Kearns, 2001; Ssymank et al., 2008). The abundance of flower visitors varies with the seasons which could be linked to variation in abundance of floral resources (number of flowering plant species and availability of flowers) (Barret and Helenurm, 1987). Population dynamics of some pollinator species are positively correlated with wet and hot periods (Carvalho et al., 1991) though, many factors other than climate can influence the diversity of existing seasonal patterns, such as food abundance and predation (Wolda, 1988). Various methods of measuring species composition of syrphids associated with arable crops have been investigated (Sobota and Twardowski, 2004). These include yellow traps, direct collection and sweeping. Since our study was on a very small scale, I directly observed and identified syrphid species on the flowers. Previously there was no information available

61 representing the yearly variation in abundance and composition of syrphid fly species from any part of Pakistan. The current study was intended to determine: the year-long variations in abundance and composition of hoverfly species; the plant species which share the maximum number of hoverfly species in their highest abundance; the similarity of resources use among different hoverfly species; the relationship between abundance and species composition and abiotic (temperature and relative humidity) and biotic factors (number of flowering plant species and floral abundance).

62 Materials and Methods Study Site: The study was conducted from January to December, 2008 in a planted forest of 20 hectares and adjacent agricultural farm at Bahauddin Zakariya University campus Multan, Pakistan (30.255°N; 71.513°E; 114 ± 6 meter above sea level). The climate of the area is subtropical with mean daily maximum and minimum temperatures range from 38 to 46 °C and 8°C to 12°C, respectively (Fig 1). Plant species and floral units: Besides planted trees, a variety of natural vegetation grows in the forest including annual wild plants and perennial shrubs (Table 1). I focused on the available plant species in flower, including crop plants in the adjacent agricultural landscape during the full course of our study. As different plant species had different kinds of inflorescence types, I defined the floral units for each plant species separately and each time recorded observations from those floral units. Floral abundance was estimated by randomly selecting and tagging 15 plants of each plant species and counting total floral units fortnightly. Hoverfly visitor census: In the forest, I conducted random walks and focused only single plant species at a time during its anthesis. Fifteen plants of each plant species were randomly selected and each plant was observed for 60 seconds, recording syrphid visitation at its floral unit. In this way there was a total of 15 minute of observation per plant species in one census. For agricultural crops, 15 plants were selected randomly from the margins of the field. For each plant I counted the number of visiting individuals per syrphid species by visual observation. A fortnightly census of each flowering plant species was carried out throughout the flowering period. The observations were done on clear sunny days, while rainy or cloudy days were avoided. To avoid the phenomenon of floral constancy (insects tend to visit single plant species even in the presence of many other flowering plant species in that particular area) among syrphid flies (Goulson and Wright, 1998), I selected the plants of a particular species at a considerable distance from each other (>5m). Statistical analysis. I used linear regression analysis to find the relationship between the abundance and diversity of hoverflies during the year and to test the relationship between abundance and diversity of hoverflies, and the availability of floral resources (Number of plant species in flowering and floral density per month.

63 To estimate the similarity in the hoverfly guilds with similar floral food requirements, we used multidimensional scaling analysis. Bray-Curtis index of similarity was used instead of

Euclidean distance as input since many of the cells in the data matrix were zero (Beals, 1984). To achieve accuracy we excluded from analysis those syrphid species that visited two or fewer plant species. XLSTAT computer software (XLSTAT, 2008) was used for all analysis.

64 Results The community of hoverflies consisted of 14 species, representing two sub-families and 11 genera (Table 1). The members of the sub-family Syrphinae were in greater abundance (545 individuals) than members of Milesiinae (344 individuals) in the overall study period. All the seven species in sub-family Syrphinae belonged to seven different genera whereas the seven species of Milesiinae were attributed to only four genera. Among Syrphinae, Ischiodon scutellaris (Fabricius) proved to be the most frequent floral visitor followed by Sphaerophoria bengalensis (Macquart) and Episyrphus balteatus (Degeer) i.e. 37, 23 and 22% of total abundance, respectively. Likewise these three most abundant species also visited the maximum number of plant species i.e. 37, 26 and 30, respectively. Only 2 individuals of Scaeva latimaculata (Brunetti) were recorded during the full study period. The genus Eristalinus (E. aeneus, (Scopoli) E. laetus (Wiedemann), E. arvorum (Fabricius) and E. aetniops (Wiedemann) was comprised of 96 % of the total abundance of Milesinae. E. aeneus and E. laetus were the most abundant floral visitors and visited the maximum number of plant species i.e. 40 and 28, respectively. The remaining species of Milesiinae had much lower abundance and also visited only a few plant species.

40 100 Temperature °C Relative humidity(%) 35 90 80 ) 30 70 25 60 20 50 15 40 30 10 Temperature °C Temperature 20 (% humidity Relative 5 10 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Fig. 1. Monthly mean temperature °C and Relative humidity (%) at Bahauddin Zakariya University Campus Multan, Pakistan, during January to December, 2008.

The seasonal dynamics revealed that syrphids remain active around the year with remarkable fluctuations in abundance and species diversity. The peak in abundance and diversity was observed in the month of April (Figs 2 & 3). This coincided with the month when the maximum number (35) of plant species was in flower (Fig. 4).

65 Table 1. Syrphid fly species and their abundance on flowering plants at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008 Plant Sub-family Syrphid species Abundance species visited Ischiodon scutellaris (Fabricius, 1805) 206 37 Episyrphus balteatus ( Degeer, 1776) 122 30 Eupeodes corollae (Fabricius, 1794) 50 23 Syrphinae Sphaerophoria bengalensis (Macquart, 1842) 126 26 (Total=545) Melanostoma sp. ( Schiner, 1860) 25 9 Scaeva latimaculata (Brunetti, 1923) 2 2 Paragus serratus (Fabricius, 1805) 14 10 Eristalinus aeneus (Scopoli, 1763) 212 40 Eristalinus laetus (Wiedemann, 1824) 99 28 Eristalinus taeniops (Wiedemann, 1818) 2 2 Milesiinae Eristalinus arvorum (Fabricius, 1787) 17 7 (Total=344) Eristalis tenax (Linnaeus, 1758) 2 1 Syritta pipiens (Linnaeus, 1758) 3 2 Mesembrius bengalensis (Wiedemann, 1819) 9 3

Among Milesiinae, E. laetus and E. aeneus remained active throughout the year in variable abundance. Both the species attained their peak population in March to May. Though in the month of July E. laetus was not recorded but its population increased again in the later half of the year. Other less abundant Milesiinae i.e. Eristalis tenax (Linnaeus), E. taeniops and Syritta pipiens (Linnaeus) were active only in the month of April. Mesembrius bengalensis (Wiedemann) proved to be a more abundant floral visitor in the month of November than March, May and October. Like Milesiinae, most of the Sryphinae were restricted to the spring season (March- April) i.e. E. balteatus, Eupeodes corollae (Fabricius), S. bengalensis, Melanostoma sp. (Walker) and S. latimaculata. Although E. balteatus and S. bengalensis were highest in abundance but no longer observed after spring season. Only two species proved to be persistent throughout the year i.e. Paragus serratus (Fabricius) and I. scutellaris. Four observed species i.e. E. aeneus, E. laetus, P. serratus and I. scutellaris were observed to be weather tolerant to variable extents, since these were recorded foraging in the coldest (12 °C) months of December and January to the hottest (32°C) months of July and August.

66 Six plant species receiving greater than six syrphid visitor species were identified (Fig. 5). Out of these six species two were agricultural crops (Mangifera indica and Daucus carota) and remaining four (Launaea procumbens, Ageratum conyzoides, Convolvulus arvensis and Parkinsonia aculeate) were wild plant species. L. procumbens was visited by the highest number (9) of syrphid species with highest abundance. In all six selected plant species, the proportion of Syrphinae abundance was greater than that of Milesiinae (Fig. 5).

E aeneus E laetus E taeniops E arvorum E tenax S pipiens M bengalensis 1000

100 e

10 Abundanc

1 123456789101112 Months (January to December)

Fig. 2. Monthly abundance of the members of Sub-family Milesinae at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

I scutellaris E balteatus E corollae S bengalensis P serratus Melanostoma sp. S latimaculata 1000

100 e

Abundanc 10

1 123456789101112 Months (January to December)

Fig. 3. Monthly abundance of the members of Sub-family Syrphinae at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

On the basis of similarity in the floral host plant preferences, multidimensional scaling analysis separated the members of Milesiinae (in upper half of ordination plot) from Syrphinae (in the lower half of ordination plot) except P. serratus (Fig 6).

67 Flowering plant species No. of hoverfly species Abundance of hoverflies 40 500 35 450 400 30 350 25 300 20 250 15 200

hoverfly species hoverfly 150 10 100 Abundace of hoverflies 5 50 Flowering plant species+No. of of species+No. plant Flowering 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Fig. 4. Monthly dynamics of number of flowering plant species, number of hoverfly species and their abundance at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

The plant profile was composed of 59 plant species in 29 families (Table 1). Flowering periods of these species varied with a high degree of overlap. A large number (35) of plant species were observed flowering in the month of April (the peak spring season).This is the month in which there were maximum number of hoverfly species at peak abundance (Fig. 4). The winter season (Nov-Feb) proved to be unfavorable for syrphid flies since the average temperature did not exceed 20°C and relative humidity did not lower below 70%. Temperature had a weak negative effect on the number of syphird species, whereas it had a slightly positive relation to the abundance of syrphid flies throughout the year (Fig. 7). Likewise, relative humidity was negatively correlated with both the species abundance and species number of Syrphidae over the year (Fig. 7). The biotic factors (number of flowering plant species and flower abundance per month) were also positively correlated with the number of syrphid species and their per month abundance (Fig. 8). Floral abundance decreased with the decrease in number of flowering plant species whereas syrphid abundance also decreased with the decrease in their species number over the year (Figs 9-10).

68

Syrphinae Milesinae No. of syrphid fly species 90 10

80 9 8 s 70 7 60 6 50 5 40 4 30 3 20 2 Number species hoverfly of hoverflie of Abundance 10 1 0 0 Mangifera Daucus Launaea Ageratum Convolvulus Parkinsonia indica carota procumbens conyzoides arvensis aculeata Fig. 5. Plant species visited by more than 6 syrphid fly species along with their number and abundance at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

Configuration (Kruskal's stress (1) = 0.259) 0.8 E arvorum 0.6 M bengalensis 0.4 E laetus 0.2 E aeneus P serratus

0

Dim2 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 I scutellaris -0.2 E balteatus

-0.4 Melanostoma S bengalensis E corollae sp. -0.6

-0.8 Dim 1

Fig. 6. Multidimensional scaling analysis of 10 syphid species with Bray and Curtis distance based on their abundance on 59 flowering plants at Bahauddin Zakariya University Multan Campus, Pakistan, from January to December, 2008.

69 Discussion

All the species encountered from the sub-family Syrphinae were aphidophagous whereas all those from the sub-family Milesiinae were saprophagous. The abundance of aphidophagous species was greater than that of saprophagous species in this study. The abundance of any species in a particular habitat and particular interval of time depends on availability of breeding places and hosts. Most of the members of Milesiinae breed in marshes, damp places and rotting materials (Sommaggio, 1999) e.g. Hartley (1961) described E. aeneus as more frequent in estuarine marsh pools. Since the Melisiinae are entirely aphidophagous, they lay eggs near or in the aphid patches on the plants. Almost all the aphid species in the southern part of Punjab appear in spring (Feb-March), suggesting that larval diet is also very much important in determining population dynamics of aphidophagous species. Among Syrphinae, I. scutellaris, S. bengalensis and E. balteatus and among Milesinae, E. aeneus and E. laetus, had the greatest abundance. I. scutellaris, E. balteatus and E. aeneus are widely distributed over many parts of the world (Ghahari et al., 2008). These species have adapted to a wide range of geographic patterns, therefore they could be considered as the most successful syrphid species not only in terms of distribution but also in exhibiting the greatest host plant range in this study. In nature the organisms with more generalization in mutulistic interactions are more successful than the organisms with more specialized and limited types of mutulistic interactions (Fenster et al., 2004). The observations in the field might be undermined by well studied marked floral constancy among many hoverfly species (Goulson and Wright, 1998) and daily blooming pattern of the plant species (Freitas and Sazima, 2003). To overcome these problems I selected individual plants (at time of anthesis) at a considerable distance (>5m) from each other only during their anthesis. Paragus serratus is zoophagous and feeds on soft bodied insects other than aphids (Sommaggio, 1999) and I observed P. serratus throughout the year even in the months of winter (January, November and December); i.e. the time which is unfavorable for the development of almost all kinds of aphid species. The two other biotic factors in this study (the availability of flowering plant species and floral abundance) have shown a positive relation to syrphid abundance and diversity. Many pervious studies have also exhibited a positive correlation between availability of floral resources and abundance of pollinators (Barret and Helenurm, 1987; Wolda, 1988; Inouye and Kearns, 1993; Souza-Silva et al., 2001). As with landscape level studies, a positive relation between richness and abundance of floral resources and pollinator diversity and activity has also been observed at a micro scale i.e. within a plant community (Klein et al., 2003).

70 500 y=39.95+1.24x 500 r²=0.006 y=398.54-4.80x P=0.818 r²=0.203 e 400 e 400 p=0.141

300 300

200 200

Syrphid fly abundanc fly Syrphid Syrphid fly abundanc fly Syrphid 100 100

0 0 40 50 60 70 80 90 10 15 20 25 30 35 Relative humidity (%) Temperature °C 14 14 y=9.15-6.55x y=6.09-5.93x 12 r²=0.088 12 r²=0.029 p=0.348 p=0.594 10 10

8 8 6 6

4 Syrphid species fly 4

Syrphid fly species fly Syrphid 2 2

0 0 10 15 20 25 30 35 40 50 60 70 80 90 Relative humidity (%) Temperature °C

Fig. 7. Relationship between abiotic factors (average temperature °C and relative humidity %) and number of syprhid fly species and their abundance over the year at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

71

500 y=-76.45+9.05 500 r²=0.454 y= 170.25+3.45x p=0.016 r²=0.370 400 400 P=0.36 e e

300 300

200 200

Syrphid fly abundanc fly Syrphid Syrphid fly abundanc fly Syrphid 100 100

0 0 2 7 12 17 22 27 32 0 2000 4000 6000 8000 Flowering plant species Flower abundance 14 y=2.18+0.15x 14 y=2.56+5.19x r²=0.304 12 r²=0.194 p=0.063 12 p=0.152 10 10

8 8

6 6

Syrphid fly species fly Syrphid 4 4 Syrphid fly species fly Syrphid

2 2

0 0 2 7 12 17 22 27 32 0 2000 4000 6000 8000 Flowering plant species Flower abundance

Fig. 8. Relationship between biotic factors (number of flowering plant species and floral abundance per month) and number of syprhid fly species and their abundance over the year at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

72

The composition of floral community may also be an important factor in predicting diversity and abundance of pollinators (Potts et al., 2003). Likewise, Hegland and Boeke (2006) also reported that small scale spatial variations in the density and diversity of floral resources can positively affect pollinator activity. Syrphidae are typically generalists in their floral visitation (Memmot, 1999, Dupont et al., 2003) and therefore have a stronger positive response to floral abundance than to plant species richness. Another study (Steffan-Dewenter and Tscharntke, 1999) also failed to link syrphid fly activity to landscape scale variables. Altogether, this suggests that the availability of floral resources may be a poor predictor of syrphid activity. There may be some other spatial factors e.g. color and floral host preferences (Haslett, 1989; Sutherland et al., 1999) may also work at small scale that might explain the lack of response to the more general characteristics of plant richness or blossom density. All six plant species which were visited by a maximum number (greater than 6) of syrphid species had open type of flower morphs with shallow nectarines. According to pollination syndromes concept, dipterous flies prefer actenomporphic yellow or white coloured flowers with shallow nectar (Fægri and Van der Pijl, 1979). Flies unlike bees, can not manipulate the complex zygomorphic and tubular flowers with hidden nectar and deep nectarines. Milesiinae are more bulky than most Syrphinae with longer proboscis (5-10 mm) enabling them to feed from tubular flowers with rather deep nectars e.g. M. bengalensis in this study fed on tubular Lantana camara in which corolla depth was 10.4+0.03 mm.

600 y=-141.35+45.27x r²=0.878 500

e p=0.000

400

300

200

abundanc fly Syrphid 100

0 2468101214 Number of syrphid fly species Fig. 9. Relation between number of syrphid fly species and their abundance at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008.

73 The knowledge of similarity in the floral host plants among the pollinators can help us to develop conservation strategies for pollination or biological control. E. aenus and E. laetus were clustered together in the cluster analysis. This is because they have long mouthparts and can feed from almost all kinds of open flower morphs with corolla depth less than 10 mm. In this study, all the flowers had nectar depth less than 10 mm, except L. camara. There was no clear distinction between the floral plant preferences of Milesiinae and Syrphinae. This might be because the flies with longer mouthparts exploit nectar from both shallow and deeper corollae (Fenster et al., 2004). On the other hand, pollen feeding is more common in Syrphinae than Milesiinae, particularly from flowers with deeper nectar i.e. pollen placed well above the nectar which is easily assessable by Syrphinae (Holloway, 1976). Many Syrphinae usually feed nectar from very short and shallow flowers e.g. umbelliferous flowers (Apiaceae). Number of syrphid fly species and their abundance were weekly negatively associated with relative humidity. This might be because of proven negative association of relative humidity with aphid populations while positive association of syprhid with aphids populations (Devi et al., 2011). The relation between Diptera and environmental factors may vary with geographical distinction e.g. Carvolho et al., (1991) documented a positive relationship between temperature and number of Diptera in tropics but this relationship could be negative in hotter areas of the world e.g. sub-tropical areas of Pakistan. In hot and humid climates with no dry seasons, temperature and humidity do not significantly fluctuate as much round the year e.g. Abaete, Salvador (Brazil). Under such climatic conditions bees have been studied extensively (Viana and Kleinert, 2005) and found to be unaffected by temperature and humidity but positively related with the availability of floral resources. The seasonal variation in floral visitors is almost certainly related to resources availability. And higher species richness is positively correlated to higher resource diversity. Less frequently visitors are possibly related to specific plants or have a short activity period (Inouye and Kearns, 1993). In an ecosystem, the importance of any syrphid fly species cannot be estimated solely on the basis of its abundance. The pollination potential of the most abundant syrphid may be less than that of the least abundant syrphids. These facts suggest the importance of deeper studies of insect-plant interaction in this habitat with the perspective of maintaining natural and agricultural plant communities. Unfortunately, such detailed studies on pollination by Diptera are rare, especially in Pakistan, but hopefully the present work can serve as groundwork for further such research.

74

12000 y=757.0+203.4x r²=0.743 10000 p=0.00

e 8000

6000

4000

abundanc Flower 2000

0 2 7 12 17 22 27 32 Flowering plant species

Fig 10. Relation between flowering plant species and their floral abundance at Bahauddin Zakariya University Campus Multan, Pakistan during January to December, 2008

75 Table 2. Monthly syrphid fly abundance (number of census) on flowering plants from January to December, 2008 at Bahauddin Zakariya University Campus Multan, Pakistan.

Plant species Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Aizoaceae Trianthema 0(1) 1(3) 0(1) 0(1) 4(2) portulacastrum Sesuvium sesuvioides 0(1) 9(1) 0(1) Amaranthaceae Achyranthes aspara 1(1) 0(1) 0(1) 0(2) 0(1) 1(2) 1(1) Anacardiaceae Mangifera indica 9(1) 23(1) Apiaceae Coriandrum sativum 7(1) 33(2) Daucus carota 1 9(1) 17(1) 4(1) Torilis japonica 2(1) 5(1) Asclepiadaceae Calotropis procera 0(1) 19(2) 0(1) 0(1) 0(1) 0(1) 0(1) 0(1) Asphodelus tenuifolius 1(2) 1(2) Asteraceae Sonchus asper 3(3) 17(2) 0(1) 0(1) Cirsium arvense 9(2) 23(2) Launaea procumbens 3(2) 60(2) 1(2) 1(2) 0(2) 0(1) 0(1) Ageratum conyzoides 1(1) 34(3) 1(2) Helianthus annuus 7(1) 10(1) Conyza bonariensis 12(2) 0(2) 0(2) 0(2) 0(1) Carthamus persicus 5(2) 1(2) Pulicaria crispa 0(1) 4(2) Boraginaceae Heliotropium europaeum 1(1) 1(1) 7(3) 0(2) 0(1) Brassicaceae Brassica campestris 0(1) 15(2) Sisymbrium irio 0(1) 1(2) Malcolmia africana 1(2) 0(1) Raphanus stivus 1(1) 31(2) Capparidaceae Capparis decidua 3(1) 20(2) 0(1) 0(1) 0(1) 0(1) 0(1) Cleome viscosa 0(3) 0(2) 1(2) Caryophyllaceae Spergula arvensis 1(1) 2(1) Chenopodiaceae Chenopodium album 10(2) 0(1) 0(2) 0(1) Salsola baryosma 3(2) suaeda fruticosa 4(2) 0(1) 0(1) Haloxylon recurvun 9(2) Convolvulaceae Convolvulus arvensis 2(2) 16(2) 7(2) 2(2) 0(1) Convolvulus sp. 0(1) 13(2) 2(1) 1(1) 0(2) 1(1) 3(1) 2(2) Cucurbitaceae Momordica charantia 15(1) 14(2) 10(2) 2(1) Cucumis prophetrum 5(1) 0(2) 0(2) 1(2) 0(1) 1(2)

76 Table 2. Continued

Plant species Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Euphorbiaceae Euphorbia helioscopia 5(1) 0(1) Chrozophora tinctoria 5(2) 1(1) Fabaceae Prosopis juliflora 0(1) 15(1) 8(1) 0(1) 0(1) 0(1) 0(1) Albizzia procera 6(1) Dalbergia sissoo 4(1) 16(1) Medicago sativa 2(2) 1(2) Leucaena leucocephala 18(2) 3(1) Parkinsonia aculeata 35(1) 16(1) 9(1) 2(1) Melilotus indica 0(1) 3(3) Gentianaceae Centaurium pulchellum 0(1) 4(2) Liliaceae Allium cepa 8(1) 12(1) 5(1) Malvaceae Abutilon indicum 2(1) 5(2) 12(2) 0(1) Malvastrum 2(1) 1(1) 9(2) 2(1) 0(2) 0(1) 0(1) 0(1) 0(1) 0(1) coromandelianum Grewia subinaequalis 22(2) Marsiliaceae Marsilia minuta 0(1) 0(2) 0(1) 6(1) Meliaceae Melia azedarach 3(2) Mimosaceae Acacia nilotica 0(1) 0(2) 0(2) 1(2) 4(1) 3(1) Myrtaceae Eucalyptus 11(2) camaldulensis Portulacaeae Portulaca oleracea 0(1) 2(2) 0(1) Primulaceae Anagallis arvensis 0(1) 3(2) 1(1) Ranunculaceae Ranunculus muricatus 44(2) Rhamnaceae Ziziphus jujuba 2(2) 1(1) 0(1) 1(2) 1(1) Rutaceae Citrus medica 5(1) Solanaceae Solanum surattense 1(1) 0(2) 0(1) 0(1) 0(2) Verbenaceae Lantana camara 1(2) 0(2) 0(1) 0(1) 1(2) 1(1) 4(1) 2(1) Zygophyllaceae Tribulus terrestris 2(2) 2(2) 2(2) 0(1) 3(2) Total Monthly 5(4) 9(16) 232(50) 469(54) 91(45) 30(30) 7(28) 2(20) 31(28) 5(12) 10(6) 5(3) Note: the observation months for each plant species represent the flowering period of that particular plant species.

77 Chapter 6

Effect of habitat on pollinator biodiversity and plant reproductive success

Summary This study was conducted in an effort to understand the effects of spatial variations in pollinator assemblage due to habitat isolation on the reproductive performance of perennial plant species. Variations in pollinator assemblage structure (abundance, diversity and Shannon-Wiener index) were studied at three widely isolated (100 to 200 km apart) nature reserves of Southern Punjab, Pakistan, in order to explore its effects on reproductive performance of Prosopis juliflora. Species richness and abundance were highest in Pirowal Sanctuary followed by Chichawatni Sanctuary and Chak Katora forest reserve. The pollination system of P. juliflora was highly generalized with 74 insect visitor species in four orders among all the three sites. However, pollinator assemblage varied significantly in composition among the sites. Out of the four reproductive parameters considered, the number of pods per raceme and germination varied significantly among the three locations. The reproductive performance of P. juliflora in terms of number of pods per raceme and germination improved with abundance of pollinators.

78

Introduction Pollinator assemblage structure can have important influences on floral evolution and reproductive interactions among plant species (Moeller, 2005). Recent studies of structure of plant-pollinator networks have shown that they exhibit ‘nestedness’, i.e. specialist pollinators and plants are subsets of more generalised networks (Kallimanis et al., 2009). Following these ideas, an increasing number of community-level studies have concluded that generalization is the rule and specialization is rare (e.g., Olesen and Jordano, 2002). However, most studies have not distinguished between pollinators and floral visitors, despite the fact that a visitor may not actually pollinate the flowers. Pollination systems can be examined in the same general framework used to understand community structure and assembly. For example, the variation of visitor abundance and their spatial predictability within plant populations allow regular floral visitors to be distinguished from occasional or incidental floral visitors (Root, 1973). From the ecological point of view, generalization is considered a positive trait that may favour competitive ability, colonization capacity and invasion ability in plants (Richardson et al., 2000). In a geographical context, the spatial structure of variation in pollinator abundance and community composition can also have important implications for plant reproductive performance and ultimately floral evolution (Gomez et al., 2007). This spatial scale variation in pollinator communities remains poorly documented. Flower visitation activity of pollinators is often strongly affected by climatic conditions (McCall and Primack, 1992) and climate varies on both large and small spatial scales. The size and quality (availability of nesting places and floral resources) of any natural habitat also determines the richness and abundance of pollinator species (Cunningham, 2000). When large tracts of forests are subjected to fragmentation due to deforestation, the organisms remaining in discontinuous remnants are exposed to many biotic and abiotic changes (Saunders et al. 1991). The survival of any species under such conditions depends on its life history, ecological characteristics and mutualistic interactions (Rathcke and Jules, 1993). Habitat isolation and quality certainly affects pollinator diversity and hence reproductive success of plants, but seed set may also be influenced by other factors such as patch size, density of flowering plants, occurrence of competing alternative flowers (Jennersten and Nilsson, 1993; Kleijn and Langevelde, 2006), genetic variability (Van Treuren et al., 1994) and abundance of pollinators (Donaldson et al., 2002).

79

Previously, many studies assessed the affect of habitat quality and fragmentation or isolation on the diversity and abundance of pollinators and ultimately plant reproductive success (Steffan-Dewenter and Tscharntke, 1999; Donaldson et al., 2002; Murren, 2002). These studies were done on small scales, i.e. either on islands or in isolated patches of forest in single habitats suffering from deforestation. Our study was conducted on a different scale: The study locations in this research were relatively large (>1000 acres) and widely separated (100-200 km) from each other. Since there is no geographical barrier in the plains of Punjab, we can presume that thousands of years back, the entire study area may have been covered by continuous, similar vegetation and inhabited by a homogenous population of pollinator species. However, the distribution of pollinators may vary in space and time (Ollerton and Cranmer, 2002). Very few studies have explicitly tested for spatial variation in pollinator richness and diversity (Herrera, 2005). This study focuses on the invasive Prosopis juliflora (Sw.) DC, a multipurpose perennial tree native to northern South America. Its fast growth, drought resistance and salt tolerance has led to its successful naturalization in dry tropical and sub-tropical areas of Africa, south-east Asia, the Indian sub-continent and Australia. It was deliberately introduced to the sub-continent in 1857 (Luna, 1996) with view point of its qualities and uses, i.e. construction material, charcoal, soil conservation and rehabilitation of degraded and saline soils (Pasiecznik et al., 2001). Due to improper management, however, its potential benefits were undermined, because it becomes a serious invasive tree (Pasiecznik, 2001). Its invasive qualities have resulted in multiple negative effects on food security, livelihoods and environment in many parts of the world, e.g. Ethopia (Dubale, 2006), India and Kenya (Mwangi and Swallow, 2005) by declining livestock production and productivity due to its competition with palatable native trees. P. juliflora is self-incompatible (Simpson, 1977) and receives visitors from a variety of insect orders (Ward et al., 1977). Many pollination biologists have explored the reproductive biology of P. juliflora (DeOliveira and Pires, 1990; Iqbal and Shafiq, 1997; Zaitoun et al., 2009). About 300-400 flowers are arranged on a raceme (Zaitoun et al., 2009). An individual flower has petals only 3 mm wide while its anthers are 4 mm wide. The small size of the flower ensures contact of even small insect with reproductive organs and minimizes the risks of nectar robbing. Thus, it is fair to assume that all species recorded act as effective pollinators to some degree. We investigated P. juliflora populations in three widely separated wildlife reserves (ca. 100 to 200 km apart from each other) of the southern Punjab (Pakistan). The aim was to

80

study (i) variation in the pollinator community structure (species richness, diversity and abundance) among these reserves due to their isolation a hundred years ago; and (ii), whether the variation in pollinator community structure affects the reproductive performance of P. juliflora.

81

Materials and Methods Study area The study was conducted in the southern irrigated zone (PARC, 1980) of the Punjab in Pakistan. Geographically, it is an alluvial plain with fertile soil deposited by the floods over thousands of years. The Climate of the area is sub-tropical with cold winter and hot summer; the mean monthly temperature ranges between 25°C and 30°C, with 35°C to 40°C maxima, and 10°C to 20°C minima. The extreme maximum temperature of the region varies between 45°C and 51°C, recorded in May and June, while the lowest minimum temperature is 0°C to - 5°C, recorded in January. May and June are the hottest months, whereas January is the coldest month of the region (Khan et al., 2010). The annual average monthly summer rainfall is ca. 18 mm. Most of the land is cultivated, irrigated with canal or underground water. A variety of crops are grown, but a cotton-wheat rotation with mango is the tradition and very important. I selected three natural habitats: Chechawatni wildlife safari (District Sahiwal), Perowal safari (District Khanewal) and Chak Katora wildlife park (District Bahawalpur) (Table 1). Chechawatni is 100 km north-east from Perowal, while Chak Katora is 150 km south-east: between them are either cities or agricultural land. These natural habitats have never been cultivated. I assessed Perowal as undisturbed since it is protected (fenced), and according to the definition of a safari park, entry is prohibited without permission of the officials. Chechawatni is the least disturbed site since the area is larger and with few human interventions (grazing and wood cutting) relative to Chak Katora (disturbed).

Table 1. Properties and types of three studied nature reserves at southern Punjab, Pakistan. Type of Area Altitude Human Location GPS Soil type reserve (acres) (feet) intervention Perowal plantation, Wildlife safari 30º343” N 17823 437±16.5 No Silt loam District Khanewal (fenced) 72º035” E

Chichawatni Wildlife park 30º544” N 11530 205±22 small Silt loam plantation, (unfenced) 72º703” E District Sahiwal

Wildlife park 29º780” N Sandy Chak Katora Hasilpur, 1323 447±19.4 large/ big (unfenced) 72º548” E loam District Bahawalpur

82

Flower Visitor Census In each location I chose four experimental plots (ca. 2 acres each) with similar populations of P. juliflora (ca. 100 trees) and selected 25 same sized trees in each plot, widely separated from each other. On each tree, I tagged one branch ~10ft long originating from the main stem. Fortnightly the number of individuals of each species were counted visiting the flowers on these tagged branches (60 seconds/branch), from first week of April to end of June, 2009. I reselected another branch on the same tree if the chosen one was not flowering. The data was recorded on sunny days from 08:00 to 11:00 h (local time). Identification of Diptera and Coleoptera was done by using standard keys (Borror et al., 1981). Syrphids (Diptera: Syrphidae) and butterflies (Lepidoptera: Rhopalocera) were identified to species by experts (see Acknowledgement). Bee genera were identified by using the keys of Michener (2007). Insects which could not be identified to species level were grouped by morphospecies and voucher specimens deposited at the Agricultural Museum of the University College of Agriculture, Bahauddin Zakariya University Multan. Reproductive success Plant reproductive success was measured in terms of number of pods per raceme, seed weight per pod and number of seeds per pod. Since the pod-setting percentage is very low in P. juliflora (deOliveira and Pires, 1990), I added data from two more adjacent branches on each of the sampled trees. During each visitor census, I measured the number of flowers per branch and harvested the mature pods fortnightly. Number of pods per raceme, number of seeds per pod, pod length, 100-seed weight and germination (%) was recorded. Data analysis Two traits of the pollinator assemblage visiting p. juliflora flowers are analyzed in this study i.e. abundance and diversity (Magurran, 2004). I assessed pollinator diversity by calculating species richness, diversity, evenness and dominance. Species Richness was measured as number of insect visitor species in each location. Dominance was calculated as the relative abundance of the most abundant visitor species. I used pollinator rank-abundance plots as a way to visualize the structure of the pollinator communities (Magurran, 2004). Diversity was calculated as Shannon-Wiener index and Hurlbert’s Probability of an interspeciefic encounter (PIE) index (Colwell, 2005). Hurlbert is an evenness index that combines the two mechanistic factors affecting diversity i.e. dominance and species abundance which itself is the complement of Simpson's index. Among-location differences in pollinator abundance, richness, diversity (Shannon- Winere and Hurlbert indexs) and plant reproductive success were analyzed with one-way

83

ANOVA, followed by the multiple-comparison test called Least Significant Difference (LSD) test. Richness, dominance and diversity (Shannon-Winere and Hurlbert indexes) were compared among the locations using individual-based rarefaction curves generated by EcoSim (Gotelli and Ensminger, 2005); the graph was drawn in PAST software. Rarefaction allows for estimation of number of species (S) expected in a random samples of n individuals taken from a larger collection made up of N individuals and S species (Gotelli and Entsminger, 2005). The effect of pollinator diversity on plant reproductive success was explored with simple linear regression analysis between species richness, abundance and Shannon-Wiener index as predictors and components of reproductive success as responses.

84

Results A total of 800 insects belonging to 77 species in four orders were observed visiting the flowers of P. juliflora in the three isolated locations. The majority of species were Hymenoptera (41) and Diptera (17; Fig. 1). Only four species made up more than 6% of the total visits across all sites. These species were two long-tongued large bees (Apis dorsata and A. florea) and two syrphid flies (Eristalinus aeneus and Ischiodon scutellaris). Considering the seven most abundant floral visitors across all sites (Fig. 2), four species (A. florea, A. dorsata, Megachile sp. and E. aeneus) had long mouth parts (3.44, 6.14, 5.22 and 4 mm, respectively; Somanathan et al. 2009; Chaiyawong et al. 2004) and three species (I. scutellaris, Episyrphus balteatus and Agapostemon sp.) had shorter mouth parts (2-3 mm; Gilbert, 1981). Among these four most abundant floral visitors, E. balteatus and A. dorsata were the most abundant at Chichawatni (Table 4) while E. aeneus, I. scutellaris and A. florea at Khanewal. On the other hand, E. aeneus and I. scutellaris were not observed in Chak Katora, which was dominated by two bees Megachile sp. 2 and Agapostemon sp.

Table 2. Among-location differences in pollinator abundance and diversity. Means followed by different letters are statistically different at α 0.05.

Locations Abundance Dominance Shannon- Species Richness Hulbert Wiener Perowal 5.22 a 0.24 b 2.68 a 36.42 a 0.88 a

Chichawatni 3.50 ab 0.32 a 2.58 a 30.35 b 0.87 a Chak-Katora 1.59 b 0.27 ab 2.31 b 18.95 c 0.89 a

Pollinator assemblage structure was similar in three locations with few abundant species and high number of scarce species (Fig. 2), but the assemblage varied significantly in composition among the sites. The most abundant species in three locations also varied (Fig. 2). Eleven species (two fly, two bee, three wasp, two butterfly and two beetle species) were observed solely in Chichawatni (Table 4). Five species (one dipteran, one wasp, two butterfly and one beetle species) were restricted to Chak Katora while seventeen species (two flies, ten bees, three wasps, one butterflies and one moth) occurred only in Khanewal. Twelve species were commonly found in the three locations, i.e. four fly, four bee, one wasp and three butterfly species (Table 4).

85

Hymenoptera Diptera Lapidoptera Coleoptera 100%

80%

60%

40%

20%

0%

ind spe ind spe ind spe

Perowal Chechawatni Chak-Katora

Fig 1. Percentage of species (spe) and individuals (ind) of related insect orders at three locations in southern Punjab during 2008.

Fig 3. Rarefaction curves of three locations i.e. (a) Perowal (b) Chichawatni (c) Chak-Katora, based on individual rarefaction method showing the expected number of species as a function of sample size.

86

At all three sites Hymenoptera was the largest order in terms of abundance and number of species followed by Diptera (Fig. 1). There was significant variation in pollinator abundance among sites in terms of visitation frequency (F = 0.037, df = 126, P = 0.05). The maximum visitation frequency (5.22 individuals per 30 minutes) was observed in Pirowal followed by Chichawatni and Chak-Katora (3.50 and 1.59 individuals per 30 minutes). Randomization analyses showed that locations differed in Shannon-Wiener diversity indices (F = 5.58, df = 70.88, P = 0.005; Table 2). Pirowal and Chichawatni had more diverse and statistically similar visitor assemblages than Chak-Katora in terms of Shannon-Wiener index. At all three sites the Hurlbert index was always higher than 0.87 (more than 87% probability of two randomly selected insects being different species), and differences among the three locations were statistically non-significant (F = 0.30, df = 69.19, P = 0.736). The rarefaction curves of three locations based on individual rarefaction method (using the expected number of species as a function of sample size) are shown in Fig. 3. There was a significant difference (F = 35.79, df = 82.18, P = 0.00) in species richness among the three locations (Table 2) i.e. highest in Pirowal followed by Chichawatni and Chak-Katora. One-way ANOVAs showed that there were no significant differences in seeds per pod and seed weight of P. juliflora among the three sites, but there were significant differences in the number of pods per raceme and germination. Significantly more pods per raceme were recorded in Chichawatni and Pirowal than Chak-Katora (F = 70.33, df = 75, P < 0.001), whereas germination percentage showed a non-significant (df = 27, F = 2.35, P = 0.114) tendency being higher in Pirowal (51%) than Chichawatni (41.2) and Chak-Katora (40.8%). The number of pods per raceme and germination appeared pollinator limited (Table 3). Since there were no significant differences in seeds per pod and seed weight among the three locations, we did not relate them to species abundance, richness and Shannon-Wiener index. Only significantly varying components were selected for that purpose, i.e. number of pods per raceme and germination. Abundance was positively related to the number of pods per flower and germination. Species richness and the Shannon-Wiener index (Fig. 4) were not related.

87

Discussion The pollination system of P. juliflora is extremely generalized. Its brush-shaped flowers were visited by 74 insect species with different morphology, size and behaviour from a variety of functional groups, i.e. short-and long-tongued bees, butterflies, flies and beetles. Previous studies have also failed to relate any particular functional group of insects to the evolution of brush- shaped flowers (Hingston and McQuillan, 2000; Devy and Davidar, 2003). Given the morphology of P. juliflora flowers, contact of insect visitors with flower's reproductive organs is unavoidable. Muzaffar and Ahmad (1991) noted P. juliflora as a major nectar and pollen resource for bees in Pakistan. Prosopis juliflora is not native to the sub-continent, yet it received visits by a large array of potentially pollinating insects. Pollination success in P. juliflora is always low and very few legumes are produced despite large numbers of flowers per tree: DeOliveira and Pires (1990) reported a pollination efficiency of 29% in P. juliflora based on the number of inflorescences per tree. However, when related to the number of flowers, it dropped to 1.48% The low fruiting success under open pollination as well as under exclusion of pollinators indicates that P. juliflora is more or less incapable of autonomous selfing (DeOliveira and Pires, 1990). The low fruiting success is might also be because many plant species mass-flower to increase floral display and abort most of their flowers and may not have enough resources to develop fruits from all these flowers (Trueman and Wallace, 1999). The total number of seeds produced by one individual may, however, still be high enough for the species to survive (Karron and Mitchell, 2011). Several reasons have been documented for the low reproductive success by DeOliveira and Pires (1990): poor pollen viability, short period of pollen release or stigma receptivity, lack of synchronization between pollen release and pollen reception, flower sterility or high rates of ovary abortion, few pollinating insects at the time of maximum receptivity. Goel and Behl (1995) found maximum pollen production in P. juliflora at midday, but insects are less mobile during high temperatures at this time (ca 40ºC in this study). The pollinator assemblage in rank-abundance curves revealed A. dorsata and A. florea as the most abundant species in all three locations. A single colony of Apis may have thousands of workers and trees provide ideal nesting locations for them, making them the most dominant visitors, as in case of this study. Both the species have already been reported as the most

88

dominant floral visitors of various crops (Sajjad et al., 2008; Sajjad et al., 2009; Ali et al., 2011) in the study location.

Fig 2. Rank-abundance curves of pollinator species visiting the three locations (A: Perowal; B: Chichawatni; C: Chak-Katora) in Southern Punjab. The names of the species accounting for at least 10% of the visits at a given location are provided.

89

I also found a significant spatial variation among locations in pollinator assemblages in terms of species richness, diversity and dominance. The abundance-richness relationship is frequent in pollinator assemblages (Steffan-Dewenter et al., 2002), i.e. few abundant species and high number of scarce species in this study. The observed species richness ranged from 19 to 34 species. The highest species richness was recorded in the largest and least disturbed site, at Pirowal. Habitat quality includes factors that directly influence the life history of a pollinator species, i.e. availability of breeding and shelter places and host plants (Winfree et al., 2011). The surrounding quantity and quality of habitats is usually positively related to insect species richness and abundance (Oeckinger and Smith, 2007) but the discovery of these relationships may depend upon the scale of investigation.

Perennial plant species respond in different ways towards habitat fragmentation and it is difficult to find a clear rule in general (Donaldson et al., 2002), i.e. populations in small fragments do not always experience pollination deficits. However, many other studies suggest a better plant reproductive performance in larger reserves or mainland than smaller ones or on islands (Jennersten, 1988; Cunningham, 2000; Donaldson et al., 2002; Murren, 2002). Our data clearly suggests that for P. juliflora, the number of pods and germination rate of seeds rise with abundance of its flower visitors. Further, flower visitor abundance was a better predictor of plant reproductive performance than species richness and the Shannon-Wiener index. This might be because of less sensitivity of Shannon-Wiener index to actual site differences, i.e. the Shannon-Wiener index can be similar even if two sites are different in species richness. Only few studies give some empirical evidences regarding consequences of pollinator assemblage on plant reproductive success on account of pollen limitation. Diversity and identity (Gomez et al., 2010), and relative abundance of floral visitors (Sahli and Conner, 2006) have been reported as the important predictors of plant reproductive success. The sensitivity of pollinator assemblages to predict plant reproductive success may vary with plant species and the pollination effectiveness of available pollinator species (Talavera et al., 2001). In conclusion, flower visitor spectra of P. juliflora at Pirowal and Chichawatni were more species rich and abundant than at Chak-Katora possibly due to the former two sites being larger and least disturbed if not undisturbed compared to the latter. This ultimately affected insect visitation frequency and hence fruit set of P. juliflora. In contrast to flower visitor abundance,

90

species richness and the Shannon-Wiener index were no good predictors of reproductive performance of P. juliflora.

Table 3. Comparison of reproductive success of three locations. The mean sharing similar letters are statistically similar at α 0.05.

100-Seed No of No of Pod length Germination weight pods/raceme seed/pod (cm) (%) (gm)

Perowal 5.829 a 10.77 a 15.050 a 2.767 a 51.00 a

Chichawatni 6.413 a 10.95 a 15.312 a 2.778 a 41.20 b Chak-Katora 3.241 b 10.90 a 14.799 a 2.787 a 40.80 b

Fig 4. Relationship between abundance, species richness and diversity (Shannon-Wiener) and plant performance measured as germination rate (lower graphs) and fructification (number of pods, upper graphs) in P. juliflora at southern Punjab, Pakistan. Solid circles, Chak-Katora; open circles, Chichawatni; open triangles, Perowal.

91

Table 4. Insect visiting P. juliflora during April to June 2009 in three selected locations along with percentage of visits (relative to the total number of visits in a particular location).

Chichawatni Khanewal Hasilpur Insect visitors

No. % No. % No. % Diptera Calliphoridae sp.(calliphoridae) 2 0.74 1 0.79 Euphumosia sp1. (Calliphoridae) 2 0.74 Euphumosia sp2. (Calliphoridae) 3 1.10 4 1.00 3 2.36 Diptera sp. (Unidentified) 1 0.37 1 0.25 Heterostylodes(Anthomyiidae) 2 0.74 1 0.25 2 1.57 Sarcophaga (Sarcophagidae) 4 1.00 Odontomyia sp. (Stratiomyidae) 1 0.37 3 0.75 Diptera sp. (Unidentified) 1 0.79 Episyrphus Balteatus (Syrphidae) 19 6.99 2 0.50 6 4.72 Eupeodes corollae (Syrphidae) 2 0.74 2 0.50 Eristalinus aeneus (Syrphidae) 14 5.15 46 11.47 Eristalinus laetus (Syrphidae) 5 1.84 8 2.00 1 0.79 Eristalodes taeniops (Syrphidae) 1 0.25 Musca domestica (Syrphidae) 2 0.74 2 0.50 Melanostoma sp. (Syrphidae) 15 5.51 0 1 0.79 Sphaerophoria bengalensis (Syrphidae) 8 2.94 16 3.99 3 2.36 Scaeva latimaculata (Syrphidae) 1 0.37 Ischiodon scutellaris (Syrphidae) 13 4.78 42 10.47 Hymenoptera Apis dorsata (Apidae) 84 30.88 59 14.71 8 6.30 Apis florea (Apidae) 26 9.56 92 22.94 29 22.83 Megachile sp1. (Megachilidae) 2 0.50 Nomia Oxybeloiues (Halictidae) 1 0.25 Halictidae sp1.(Halictidae) 3 1.10 7 1.75 4 3.15 Halictidae sp2. (Halictidae) 5 1.25 Megachile sp. (Megachilidae) 1 0.37 1 0.25 Megachile sp2. (Megachilidae) 8 2.00 23 18.11 Amegilla sp. (Apidae) 11 4.04 5 1.25 Megachile sp3. (Megachilidae) 1 0.37 5 1.25 Halictidae sp3. (Halictidae) 2 0.74 1 0.25 Halictidae sp4. (Halictidae) 4 1.00 Halictidae sp5. (Halictidae) 4 1.00 Ceratina sp.(Apidae) 1 0.79 Apidae sp. (Apidae) 1 0.25 Osmia sp. (Megachilidae) 1 0.25

92

Table 4. Continued

Chichawatni Khanewal Hasilpur Insect visitors

No. % No. No. % Megachile sp4. (Megachilidae) 1 0.37 Halictidae sp6. (Halictidae) 1 0.79 Melissodes sp. (Apidae) 1 0.37 9 2.24 Halictidae sp7. (Halictidae) 2 0.50 Melissodes sp. (Apidae) 1 0.37 6 1.50 Agapostemon sp. (Halictidae) 8 2.94 7 1.75 17 13.39

Colletes sp. (Colletidae) 1 0.25 Thyreus sp. (Apidae) 2 0.50 Xylocopa basalis (Apidae) 1 0.37 Coelioxys sp. (Megachilidae) 2 0.50 3 2.36 Agapostemon sp. (Halictidae) 4 1.00 3 2.36

Halictidae sp8. (Halictidae) 2 0.74 Vespa orientalis (Vespidae) 2 0.74 Polistes olivaceus (Vespidae) 4 1.47 2 0.50 1 0.79 Crabronidae sp.(Crabronidae) 2 0.74 1 0.25 Tiphiidae sp. (Tiphiidae) 3 1.10 Wasp 1 (Unidentified) 1 0.25 Wasp 2 (Unidentified) 5 1.84

Wasp 3 (Unidentified) 2 0.74 1 0.79 Wasp 4 (Unidentified) 1 0.79 Wasp 5 (Unidentified) 2 0.50 wasp 6 (Unidentified) 1 0.25 wasp 7 (Unidentified) 2 0.50 2 1.57 Wasp 8 (Unidentified) 1 0.37 Wasp 9 (Unidentified) 1 0.25 Lepidoptera Danaus chrysippus (Danadidae) 3 1.10 Colotis amata (Pieridae) 1 0.25 Junonia almana (Nymphalidae) 2 0.74 2 0.50 1 0.79 Eurema hecabe (Pieridae) 3 2.36

Tarucus venosus (Lycaenidae) 2 0.50 Anaphais aurota (Pieridae) 1 0.79 Colotis vestalis (Pieridae) 3 1.10 2 0.50 Colotis etrida (Pieridae) 3 1.10 1 0.25 1 0.79 Pieris brassicae (Pieridae) 1 0.37 1 0.25 1 0.79 Moth1 (unidentified) 3 0.75

93

Table 4. Continued

Chichawatni Khanewal Hasilpur Insect visitors

No. % No. No. % Parnara guttata (Hespiriidae) 1 0.37 Heliothis armigera (Noctuidae) 2 0.74 2 1.57 Polyommatus eros (Lycaenidae) 11 2.74 2 1.57 Coleoptera Beet1 (unidentified) 4 3.15 Beet2 (unidentified) 1 0.37 Beet3 (unidentified) 1 0.37 5 1.25 Weevil1 (unidentified) 1 0.37 Weevil 2 (unidentified) 3 1.10 2 0.50

94

Chapter 7

In search of the best native pollinators for crop

production Summary In conjunction with the efforts to discover the pollinator community of crops and the best pollinators for crop production, experiments were performed at research farm of University College of Agriculture, Bahauddin Zakariya University, Multan, Pakistan. Two cross-pollinated crops were selected i.e. onion (Allium cepa L.) and janter (Sesbania sesban). 1) The community of pollinators in A. cepa was composed of two bees (Order: Hymenoptera) and eight true flies (Order Diptera). Among bees, Apis dorsata proved to be an abundant pollinator (2.85±1.57 individuals/25 plants) while among true flies Episyrphus balteatus (Syrphidae) had the maximum abundance of 14.00±4.61 individuals/25 plants. All the insect pollinator species reached peak activity during 10:00 to 12:00 h. Eupeodes corollae (Syrphidae) exhibited the most efficient foraging pattern by visiting 17.14±1.38 flowers in 147.5±8.14 seconds on an umbel. A. dorsata was revealed as the most effective pollinator, however, based on seed setting results for visits by single species over 20 minutes and which produced 506 seeds/umbel/20 minute visit. 2) Different yield attributing components of S. sesban were measured in three types of pollination treatments i.e., wind pollination, self pollination and cross pollination by insects. Cross pollinated inflorescences produced maximum number of pods (4.41) and maximum seed weight (1.1 g) as compared to wind pollinated (2.42 pods and 0.72 g seeds inflorescence-1) and self pollinated inflorescences (2.25 pods and 0.46 g seeds inflorescence-1). Germination was also better in cross pollinated inflorescences as compared to self and wind pollinated inflorescences. Megachlie bicolor and Apis dorsata exhibited most efficient foraging pattern, whereas A. florea proved to be the most abundant pollinator with highest visitation index of 0.74. Most of the pollinator activity took place in afternoon (1530 h–1830 h).

95

Introduction Pollination is a most important ecosystem service provided by insects, resulting in sustainability and continuity of life. Pollinators, particularly bees, pollinate about 66% of the world’s crop species (Roubik, 1995) and are directly or indirectly essential for an estimated 15–30% of food production (McGregor, 1976). The availability of natural insect pollinators is decreasing rapidly due to the continuous use of pesticides and decline of necessary habitat (Richards, 2001). Pollinators provide an essential ecosystem service that contributes to the maintenance of biodiversity and ensures the survival of plant species including crop plants. Insect pollination is necessary for many cross pollinated crops especially in the case of hybrid seed production e.g. onion (Allium cepa L.) (Mayer and Lunden, 2001). The role of managed honey bee (Apis mellifera L.) in pollination has widely been documented by many authors (Kumar et al. 1989; Rao and Suryanarayans, 1989; Ahmed and Abdalla, 1984; Mayer and Lunden, 2001; Tolon and Duman, 2003), but managed bee pollination is not always possible in all environments. Conserving alternate native pollinators can be a good option in areas which are very hot e.g. Southern Punjab, Pakistan where the average temperature in summer is 46ºC (PARC, 1980) or very cold and dry Balochistan province, Pakistan, where stationery bee keeping can not be practiced because of prevailing dry and cold climatic conditions and lack of forage during the large part of the year. Consequently, native pollinators should be assessed for their pollination potential, so as to conserve and manage the most efficient native pollinators to produce maximum crop yield. Pollinator species and their composition may vary with geographical area, latitude and time (Ollerton and Louise, 2002). For example, in the mountainous Hindu Kush Himalayan region Apis mellifera, A. dorsata, A. cerana and A. florea are the most frequent visitors (Chandal et al., 2004), but in the plains of Punjab, Pakistan, A. mellifera and A. cerana are poorly represented. Plant species relying on animal (e.g., insects) vectors for pollination may evolve specialized flowers as a result of mutualistic interactions with certain species (Herrera, 1989). If specialized visitors are sometimes scarce or absent (e.g., Herrera, 1990; Ågren, 1996) these plant species may be at a greater risk of pollinator limitation to seed

96 production (Jordano, 1990; Sun, 1997). Native bees are very important and specialized visitors of papilionoid flowers e.g. Sesbania sesban (Evans and Roter, 1987) therefore, the knowledge of native bee diversity and their role in crop reproductive success is very important in order to maximize the crop yield. The relative importance of floral visitors to the reproductive success of a plant species is a function of how effectively they deposit pollen on stigma in their single visit, as well as their relative visitation frequency. Fenster et al., (2004) analyzed the effectiveness of pollinator species in terms of their quality and quantity. The quality component refers to the amount of compatible pollen transferred on each visit, and quantity can be reflected by measuring the frequency with which the insect species visits the flowers. The quality and quantity of each visitor then contribute to their ‘‘pollinator importance’’ (Olsen, 1997). A single pollinator species may vary in its efficacy due to execution of different foraging pattern on a flower. For instance Honey bees can effectively pollinate or robe the nectar without pollinating the flowers in its single visit (Free, 1960). The most common pollinator therefore, may or may not be the most effective pollinator (Willmer et al., 1994; Fishbein and Venable, 1996; Stone, 1996; Olsen, 1997; Freitas and Paxton, 1998; Ivey et al., 2003). Therefore plant's pollination performance is largely determined by three factors: (i) the number of pollinators that visit the plant (visitation frequencies); (ii) the number of flowers each pollinator probes during its visit to the plant (visitation rates); (iii) the effectiveness of the pollinator in transferring appropriate pollen at each flower (Faegri and van der Pijl, 1971). Crop plant species vary significantly in their pollination requirements and hence their dependence on pollinating insects (Morse and Calderone, 2000). I tested three crop plant species i.e. onion (A. cepa), janter (S. sesban) and canola (Brassica napus) for their most efficient pollinator species. Most of the experiments on the onion have been done in caged conditions using different flies and bees e.g. Calliphora, Lucilia (Diptera: Calliphoridae), Eristalis sp. (Diptera: Syrphidae), Osmia rufa (Hymenoptera: Megachillidae) (Moffett, 1965; Bohart

97 et al., 1970; Currah and Ockendon, 1983; Free, 1993; Schittenhelm et al., 1997) but very few studies have been done in open field conditions. Onion flowers are protandrous and pollen is shed within 2-3 days before the stigma is receptive (Lesley and Ockendon, 1978), therefore, self-pollination within a flower is not possible. In order for pollination to occur, pollen must come from another flower of the same or a different plant (Zdzislaw et al., 2004). Thus, cross-pollination is common in onion (Chandel et al., 2004), which results in early seed set and higher yields. Wind is not a factor of significance in onion pollination (Erickson and Gabelman, 1956) and onion does not produce quality seed if insects do not visit the flowers (Chandel et al., 2004). Non-availability of pollinators during the flowering period of onion causes only 17% fruit setting and free availability of pollinators increased fruiting up to 73% (Rao and Sunyanarayana, 1989). Cross-pollination is obligatory in the fertilization of male-sterile onions used in hybrid seed production (Van der Meer and Van Bennekom, 1968). Onion suffers severe inbreeding depression with drastic decrease in growth bulb size, and seed production after only two cycles of self-pollination within a plant (Jones and Davis, 1944). In onion when the flowering begins, only a few flowers open each day on an umbel, but the number increases until at full bloom where 50 or more florets may be open on a single day (Moll, 1954). Apart from honeybees, onion flowers are visited by bumblebees, dipterans and butterflies (Jablonski et al., 1982). In various regions of India (Chandel et al., 2004), syrphids are important contributors in the process of pollination along with the most effective A. dorsata and A. florea. The lack of intense attractiveness of onions may cause the bees to neglect the crop (Franklin, 1970), particularly if another highly attractive crop is in flowering nearby. Perennial S. sesban is used in various agroecosystems around the world in a variety of ways. It can be used as fodder crop, green manure and fuel wood (Evans and Rotar, 1987; Macklin and Evans, 1990; Weigand et al., 1995) different parts are used in medicines for human and livestock (Woodward, 1988). It is native to Africa but subsequently distributed to almost every country of the world (Evans, 2001).

98

Despite of its wide adaptation and multiple uses, little research has been done on the mating behavior, pollination biology and ecology of this species (Brewbaker, 1990; Gebermariam et al., 2002). Knowledge of pollination biology, ecology and breeding system is essential to determine genetic structure of population (Brown and Allard, 1970; Heering, 1994). S. sesban was one of the five species included in the genetic improvement programme for agroforestry development in bimodal rainfall highlands of Eastren and Central Africa agro-ecological zones (Owuor and Owino, 1993). It is important to know the degree of self and cross pollination in S. sesban as it helps in developing some sampling techniques helpful for getting maximum information about genetic recombinations. Cross pollination in S. sesban takes place by a variety of bee species (Evans and Rotas, 1987). The current study was aimed to study the diversity of native floral visitors of A. cepa and S. sesban, their role in crop reproductive success in perspective of finding the best pollinators for future conservation. Their seasonal and diurnal dynamics and foraging pattern of pollinators were also observed in this study.

99

Materials and Methods Experiment no. 1 Study site The investigations were carried out at the research farm of the University College of Agriculture, Bahauddin Zakariya University, Multan (30.255º North Latitude and 071.513º East Longitude) during 2007 vegetative season (October-April). Climate of the area is sub-tropical with hot summer and cold winter; the mean daily maximum and minimum temperatures are in the range of 38 to 50ºC and 8 to 12ºC, respectively. The mean monthly summer rainfall is ca. 18 mm. The experimental material was onion (Allium cepa L.) seedlings of Var. “Dark Red”. Pollinator abundance and population dynamics Pollinator abundance and population dynamics were calculated by randomly observing 25 plants for 60 seconds/plant and counting the number of individuals visiting for each of the different pollinator species. Observations were made in two-hourly intervals from 6:00 to 18:00 h throughout the day and repeated in four-daily intervals during the full flowering period of 36 days. ). The genus of the syrphid flies (Syrphidae: Diptera) were identified with the help of key given in Vockeroth (1996) and species were identified by related expert (acknowledgement). Bees were identified up to genus level using the key of Michnere (2007) and the species were identified by comparing with identified specimens of Pusa collection placed at National Insect Museum, National Agriculture Research Center, Islamabad. Visitation pattern Visitation pattern was observed by counting the average stay time/umbel/visit and number of flowers visited/umbel/visit in different intervals of the day for different species. Role of different pollinators in reproductive success To determine the role of different pollinators in pollination, 50 primary umbels of the same age were randomly selected and veiled with nylon mesh bags before the flowers opened. When 50 percent of the flowers had opened in the middle of peak flowering period, which lasts for 4-10 days (Lesley and Ockendon, 1978), these umbels were unveiled during the peak foraging period of the day, i.e. 10:00 to 12:00 h. Only one

100 individual of single pollinator species was allowed to sit on the umbel at a time, for a total of 20 minutes. During this time the number of flowers visited/umble/20 minutes were visually counted, a stop watch being used to time the visit until the insect flew away. The umbel was re-veiled after completion of a 20 minutes visit. These veiled umbels were later counted for seed set and compared with that of open-pollinated and self-pollinated (veiled throughout the flowering duration) umbels. The data of visiting time/umbel, flowers visited and seed produced/umbel/20minute visit were subjected to the statistical analysis using analysis of variance (ANOVA). Means were separated by LSD test at P= 0.05 using MSTAT-C software (Steel and Torrie, 1980). For comparison of self-pollinated and open-pollinated yield, 25 umbels of same age were veiled with nylon mesh bags before opening of the flowers and another 25 umbels were tagged for open pollination, respectively. Experiment no. 2 Study site The study was carried out from August to November, 2007 vegetative season (August-November). The study site is located at research farm of University College of Agriculture, Bahauddin Zakariya University Multan, Pakistan (30.255º North Latitude and 071.513º East Latitude). Pollinator community Pollinator community on S. Sesban was recorded in whole flowering season. To identify the pollinator species, collection was done throughout the flowering season using hand net. The collected pollinators were identified to the lowest taxonomic level by using standard identification keys developed by Department of Biology, Valdosta State University. These keys are available at http://chiron.valdosta.edu/jbpascar/Intro.htm. Population dynamics of pollinators Population dynamic and attendance (mean activity rate) of different pollinators was recorded through out the flowering period with seven days interval. Data was recorded from 1500 h to 1800 h at time of opening of flowers. Twenty plants were selected randomly for this purpose and observed for 40 seconds each and counted for the number of individuals of each insect species per plant and the observation were repeated

101 with one hour of interval. Visitation rate of an insect species was also estimated in a relative fashion. I used visitation Index (IVR) (Talavera et al., 1999), defined as the product of its frequency at the flower (F) and its mean activity rate or attendance (AR) i.e., IVR= F x AR. Meteorological data of average temperature and relative humidity was also measured. Insect frequency at the flower (F) is defined as number of individuals of an insect species relative to the total number of insects included in the census. Stay time per inflorescence Ttime stayed by individual insect specie per inflorescence per visit was counted by using a stop watch in such a way that when the insect sat on the inflorescence, the time was started and counted for number of flowers visited by that insect specie. Linear regression analysis was done between number of flowers visited per inflorescence per visit and time spent per inflorescence per visit Xylocopa sp. was very less frequent visitor therefore it was omitted from this study. Pollination biology For the comparison between different pollination methods i.e., self pollination, wind pollination but no insect pollination and open pollination (wind and insect pollination) an experiment was performed. For this purpose 60 un-opened individual inflorescences of the same age were selected and counted for the number of flowers per inflorescence. Out of these 60 inflorescences, 20 individual inflorescences were veiled with butter paper bags (for self pollination). Other twenty individual inflorescences were veiled with nylon mesh bags (wind pollination but no insect pollination) and remaining 20 inflorescences were only tagged and kept exposed for insects and wind pollination. After two days, the inflorescences were un-veiled and counted for number of flowers shed and number of fertilized flowers. Then these were set to their fate up till crop matures. On the maturity of crop, pods were harvested and counted for number of pods matured per inflorescence, number of seeds per pod, pod length, pod weight and seed weight. To get more reliable yield contributed by the three pollination treatments, seed weight per inflorescence was calculated by the product of number of mature pods per inflorescence and seed weight per pod. Seed produced by the three pollination methods were tested for germination.

102

The data of number of average number of flowers per inflorescence, number of fertilized flowers, number of seeds per pod, pod length, pod weight, Seed weight per pod and seed weight per inflorescence were subjected to the statistical analysis using Analysis of Variance. Means were separated by Least Significant Difference (LSD) test at P= 0.05 using MSTAT-C software (Steel and Torrie, 1890).

103

Results Experiment no. 1 The community of pollinators was mainly composed of two bee species i.e. A. dorsata and A. florea and eight fly species i.e., Episyrphus balteatus, Eupeodes corollae., Sphaerophoria scripta (L), Mesembrius bengalensis, Eristalinus aeneus (all Syrphidae), Musca domestica (Muscidae), Calliphoridae sp. and Sarcophaga sp. (Sarcophagidae) (Table I). The spectrum of pollinator abundance was composed 87% of dipteran species and remaining 13 % of Hymenoptera species. Among the dipteran species, 72% were composed of syrphid flies and 28% of non-syrphids i.e. M. domestica, Calliphoridae sp. and Sarcophaga sp. Likewise, among hymenopteran species, 55% were composed of A. dorsata and 45% A. florea.

Table 1. List of floral visitors of onion (A. cepa L.) in Multan, Pakistan

Insect order Family Species

Apidae Apis dorsata Hymenoptera Apidae Apis florea

Syrphidae Episyrphus balteatus Syrphidae Eupeodes corollae Syrphidae Sphaerophoria scripta Syrphidae Mesembrius bengalensis Diptera Syrphidae Eristalinus aeneus

Muscidae Musca domestica Calliphoridae Calliphoridae sp. Sarcophagidae Sarcophaga sp.

In bees, A. florea was a frequent a pollinator having an average abundance of 3.57±2.57 compared to that of A. dorsata i.e., 2.85±1.57 (Table II). Diurnal abundance dynamics of bees i.e. A. dorsata and A. florea (Fig. 1) revealed that their activity started early in the morning i.e. 6:00 h. With the increase in temperature and decrease in humidity, as the day progressed, their activity increased, peaking between 10:00 h and 12:00 h, with the temperature ranging from 29.9 to 31.0°C. As the temperature rose to 33.0°C at 14:00 h, a sharp decline in activity was observed, which again started increasing at 16:00 h (30°C) and then again decrease to minimum up to 18:00 h (sunset).

104

Apis dorsata Apis florea 9 8 7

6

5

4

seconds 3 2

plants/60 Abundance/25 1 0 6:00 8:00 10:00 12:00 14:00 16:00 18:00 Time of the day

Fig. 1. Diurnal activity pattern of A. dorsata and A. florea on the flowers of A. cepa L. in Multan, Pakistan.

Episyrphus balteatus Eupeodes corollae Sphaerophoria scripta (L) Mesembrius bengalensis Eristalinus aeneus 25

20

15

10

5

0 secondsAbundance/25 plants/60 6:00 8:00 10:00 12:00 14:00 16:00 18:00 Time of the day Fig. 2. Abundance of different species of Syrphid flies on the flowers of A. cepa L. in Multan, Pakistan

Musca domestica Calliphoridae sp. Sarcophaga sp. 18 16

14

12

10 8 6 4 2

Abundance/25 plants/60 seconds plants/60 Abundance/25 0 6:00 8:00 10:00 12:00 14:00 16:00 18:00 . Time of the day Fig. 3. Abundance of non-Syrphid flies on the flowers of A. cepa L. in Multan, Pakistan. Among the syphids, E. balteatus remained the most frequent visitor having an average abundance of 14.00±4.61 followed by E. corollae (8.42±3.59) and E. aeneus 105

(8.00±5.62). S. scripta and M. bengalensis remained very low in abundance (Fig 2). The maximum activity of syrphids was observed between 10:00 to 12:00 h (29.9°C to 32.6°C). The activity pattern of the other flies (Fig. 3) also exhibited the same pattern as that of bees and syrphids. Maximum visits were made by M. domestica having an average abundance of 9.3±6.15 followed by Calliphoridae sp. (2.42±1.90) and Sarcophaga sp. (0.57±0.78) (Table 2).

Table 2. Visitation pattern (stay time/visit/umbel; Mean + SD; n= 8 for each treatment) and average abundance of different species of insects on flowers of A. cepa L. Mean values sharing similar letters in respective columns show non-significant differences (P<0.05) by using LSD test.

Stay time/ No. of flower Average abundance Pollinator Species visit/umbel visited/umbel/visit (Individuals/25 plant) (seconds) Apis dorsata 19.5+ 1.60 h 11.4 + 1.83 c 2.85+ 1.57 Apis florea 56.6 + 3.30 e 3.5 + 0.70 g 3.57+ 2.57 Episyrphus balteatus 52.4 + 5.41 f 6.5 + 1.30 ef 14.00 + 4.61 Eupeodes corollae 147.5 + 8.14 a 17.1 + 1.38 a 8.42 + 3.59 Sphaerophoria scripta 49.2 + 2.68 f 4.5 + 0.75 fg 1.57 + 0.78 Mesembrius bengalensis 37.2 + 3.64 g 7.0+ 1.20e 0.88 + 0.78 Eristalinus aeneus 63.7+ 4.85d 12.7 + 1.98 b 8.00 + 5.62 Musca domestica 66.4 + 1.64 d 3.5 + 0.84 g 9.3 + 6.15 Calliphoridae sp. 93.5 + 4.98b 16.7 + 1.45 a 2.42 + 1.90 Sarcophaga sp. 72 + 4.14 c 9.4 + 1.83 d 0.57 + 0.78

There was a significant difference in time spent/umbel/visit, F= 562.91; df= 63; p<0.001 and number of flowers visited/umbel/visit, F= 106.13; df= 63; p<0.001) of different pollinators. Among the flies, E. corollae stayed on an umbel for maximum time period (147.5±8.14 sec.) and visited maximum numbers of flowers i.e., 17.14±1.38) followed by Calliphoridae sp. (16.7±1.45 flowers in 93.5±4.98 sec.) (Table 2). Contrarily, M. domestica spent 37.2 seconds and covered minimum number of flowers (3.5±0.84) on an umbel. A. dorsata stayed for minimum time period of 19.5±1.60 seconds/umbel/visit and visited 11.4±1.83 flowers/umbel/visit. A. florea, E. balteatus and S. scripta were relatively lazy foragers with a stay time of 56.6±3.30, 52.4±5.41 and 49.2±2.68 seconds/umbel/visit, respectively and visited only 3.5±070, 6.5±1.30 and 4.5±0.75 flowers/umbel/visit, respectively. Similarly, M. domestica and E. aeneus also showed similar visitation pattern having a stay time of 66.4±1.64 and 63.7±4.85

106 seconds/umbel/visit, respectively but E. aeneus visited greater number of flowers (12.7±1.98 flowers/umbel/visit) than that of M. domestica i.e. 3.5±0.84 flowers/umbel/visit. Pollinator abundance, dynamics and visitation pattern were not good indicators of an insect being effective pollinator. Therefore, the role of different pollinators in seed setting was measured.

A. dorsata A. florea E. balteatus E. corollae M. domestica E. aeneus Calliphoridae sp. 600 506a

500 380b 400 277b 292b 277b 286b 293ab 300

200

minutes of visit 100

0 Different pollinators Numberof seeds produced/umbel/20

Fig. 4. Number of seeds produced per umbel per 20 minute of visit by different pollinators in A. cepa L. Mean values sharing similar letters show non-significant differences (P<0.05) by using LSD test.

There were significant differences among the insect species in seeds produced as a result of 20 minute visits (F= 67.6; df= 24; p<0.001) (Fig 4). The number of seeds produced per 20 minute visit revealed that A. dorsata resulted in maximum number of seed set (506±27.41 seeds/umbel/20 minute visit) followed by Calliphoridae sp., which produced 292.4±19.81 seeds/umbel/20 minute visit. The other five pollinators viz. A. florea, E. baleatus, E. corollae, E. aeneus and M. domestica resulted in statistically similar seed settings of approximately 280 seeds/umbel/20 minute visit). The comparison of yield of self-pollinated (caged throughout the flowering season/no insect visitation) and open-pollinated plants shows (Fig 5) that an average of 130 seeds/umbel were produced in caged plants whereas open-pollinated plants produced 932 seeds/umbel. This represents 616 % more seed set in open-pollinated plants whose credit solely goes to insect pollinators.

107

1200 932 1000

800

600 400 130 200

Average numberseeds/umbel of 0 Self pollinated Open pollinated

Fig. 5. Average number of seeds produced in open-pollinated and self-pollinated umbels.

Experiment no. 2

Pollinator community of the S. sesban was composed of six bee species. These include Apis florea, A. dorsata, Megachile bicolor, Megachile sp., Xylocopa sp. and a Nomia sp. Nomia sp. was very rare therefore excluded from our studies. A. florea proved to be the most frequent pollinator (Table 2) having average population of 1.28 individuals per plant per 40 sec and maximum visitation index (0.74). A. dorsata remained the second most frequent visitor (0.59 individuals per plant per 40 sec). The two Megachile sp. also remained active in low attendance. Xylocopa sp. Was very less frequent pollinator of S. sesban i.e., 0.06 individuals per plant per 40 sec with visitation index less than 0.01. The flowers visited–time spent per inflorescence relationship (Table 3) was linearly significant for four pollinator species. This suggests that each pollinator species visited different number of flowers with different stay times. M. bicolor proved to be the most efficient visitor having higher explicative value (R²= 0.87) followed by A. dorsata (R²= 0.36) and Megachile sp. (R²= 0.26). A. florea was very lazy forager (R²= 0.04) and covered only 1.34 flowers on average in 16.71 seconds of visit per inflorescence (Table 2).

108

Table 1. Yield attributing components of S. sesban ( Mean + SD; n=12 for each treatment) Mean values sharing similar letters in respective columns show non- significant differences (P<0.05) by using LSD test.

Fertilized Seed Seed weight/ Germination Flowers/ Mature pods/ Pod length Pod weight Treatments Flowers/ Seeds/pod weight/pod inflorescence (%) inflorescence inflorescence (cm) (g) inflorescence (g) (g) Nylon mesh

Self/wind 11.00 + 2.0 a 8.00 + 3.46a 2.42 + 1.38 b 23.74 + 4.94 a 17.07 + 2.60 a 0.48 + 0.14 a 0.30 + 0.07 a 0.72 + 0.38 b pollinated 10 + 2.7

Butter paper Self pollinated 12.00 + 2.08 a 8.50 + 2.77 a 2.25 + 0.75 b 18.83 + 5.01 b 13.71 + 3.04 b 0.39 + 0.11 a 0.21 + 0.07 b 0.46 + 0.15 b 9 + 2.1

Cross- 1.1 + 0.54 a 16 + 4.2 Pollinated 11.42 + 1.17 a 8.75 + 1.71 a 4.41 + 1.83 a 2 3.73 + 3.70 16.85 + 2.11 a 0.38 + 0.11 a 0.25 + 0.07 ab Insect pollination

Table 2. Average pollinator attendance, visitation index (IVR), and visitation pattern of different insect pollinators

Pollinator Visitation Time No. of flowers attendance Pollinators Index stayed/inflorescence/ visited/inflorescence/ (Population/plant/ IVR visit visit 40 seconds)

Apis dorsata 0.59 ± 0.21 0.16 12.8 ± 10.18 2.14 ± 0.53 Apis florea 1.28 ± 0.63 0.74 16.71 ± 8.88 1.34 ± 0.80 Xylocopa sp. 0.06 ± 0.02 <0.01 - - Megachile bicolor 0.24 ± 0.18 0.01 4.13 ± 2.03 1.73 ± 0.88 Megachile sp. 0.17 ± 0.04 0.01 4.20 ± 1.93 1.87 ± 0.74

109

As the flowers opened in afternoon, most of the insect visitation took place in afternoon (1530 h to 1830 h). Therefore data of population dynamics was recorded in this interval of the day. Results showed (Fig. 1) that A. florea started its activity earlier (1530 h) than A. dorsata (1630 h) and attained its peak at 1630 h. However, A. dorsata got its peak activity at 1730 h, which was the end point of A. florae activity. A. dorsata observed feeding in later hours (1830 h), near and shortly after sunset. Other three bees i.e., M. bicolor, Megachile sp. and Xylocopa sp. remained active in low population from 1630 h to 1730 h. There was no fluctuation in temperature and relative humidity around the flowering period and average temperature and relative humidity remained 28.25ºC and 73%, respectively. Weather also remained clear in all the dates of data record.

Table 3. Linear regession analysis between Number of flowers visited/inflorescence/visit and Time spent (seconds)/inflorescence/visit

Pollinators Linear model R² F P n

Apis dorsata y = 1.42 + 0.05 x 0.38 20.80 <0.0001 35

Apis florea y = 1.16 + 0.01 x 0.04 1.54 0.224 35

Megachile bicolor y = 0.048 + 0.40 x 0.87 93.02 <0.0001 15

Megachile sp. y = 1.03 + 0.19 x 0.26 4.73 0.049 15

Y=Number of flowers visited/inflorescence/visit, X= Time spent/inflorescence/visit

There was no significant difference between numbers of flowers per inflorescence among the three pollination treatments i.e., nylon mesh, butter paper and open pollinated. This is because inflorescences of the same age and same size were selected for the experiment. Flower and pod shedding is a common phenomenon of leguminous crops including S. sesban. Flower shedding was also on the same pattern in all the three pollination treatments as numbers of fertilized flowers were also statistically similar for the three treatments. This means that all the flowers, which did not receive insect visitation (nylon mesh) or air current (butter paper) were able to self pollinate as well. Comparison of means of number of mature pods per inflorescence (F= 10.34; P<0.001) revealed that the

110 maximum numbers of pods (4.41) were successfully matured in open pollinated inflorescences (Table I). Pod maturation success in other two treatments i.e., Nylon mesh and butter paper were statistically at par i.e., 2.42 and 2.25 pods per inflorescence.

A. dorsata A. florea M. bicolor Megachile sp. Xylocopa sp. 1.6 1.4 1.2 1 0.8

0.6

0.4

Population/plant/20 seconds 0.2

0 15:30 16:30 17:30 18:30 Time of the day

Fig .6 Population dynamics of different pollinators in afternoon (1530 hrs to 1830 hrs).

Statistically similar seeds per pod (F= 3.49; P= 0.048) were observed in nylon mesh and open pollinated treatments i.e., 23.74 and 23.73, respectively. Whereas only 18.83 seeds per pod were produced in inflorescences covered with butter paper. Likewise, pods produced in nylon mesh bags and open pollination were statistically similar in lengths (F= 6.64; P=0.0055) having mean lengths of 17.07 and 16.85 cm, respectively followed by pods produced in butter paper i.e., 13.71 cm. On the other hand, there was no significant difference in pod weights per inflorescences (F= 2.23; P=0.1311) were recorded in all the three treatments. Whereas a significant difference was observed in seed weight per pod among all the three treatments (F= 4.83; P=0.0182). The maximum seeds weight (0.30 g) per pod was recorded in nylon mesh cage followed by open pollinated and butter paper i.e., 0.25 and 0.21 g, respectively. Open pollinated inflorescences produced maximum number of seeds per inflorescences (F= 8.60; P=0.0017) i.e., 1.1 followed by other two treatments, which were statistically similar as well. Germination percentage was maximum (16%) in the seeds produced by open

111 pollination as compared to the seeds produced in butter paper and nylon mesh bags i.e., 9 and 10%, respectively. Discussions Experiment no. 1 The pollinator community in our experiment was mainly composed of Hymenoptera and Diptera. Apart from honey bees, onion flowers are visited by bumble bees, dipterans and butterflies (Jablonski et al., 1982). In various parts of India (Chandel et al., 2004) Apis dorsata, A. cerana, A. florea and A. mellifera are the most effective pollinators. Also dipterans from the syrphid (Syrphidae) family take part in onion pollination (Bohart et al., 1970). In our studies, only two native species of bee (A. dorsata and A. florea) were observed, with A. florea the most abundant pollinator followed by A. dorsata, This is contrary to the results of Chandel et al. (2004) in which A. dorsata was a more frequent pollinator of onion than A. florea and A. mellifera in mountainous areas of India. This change in Apis species composition and abundance might be in the result of competition for resources between introduced specie i.e., A. mellifera and native species. Moreover, the spatial variations due to habitat types and its utilization also determine pollinator assemblage by affecting the availability of nesting places and foraging resources (Sajjad et al., 2012). Other factors that may determine species composition and abundance include cropping pattern and land altitude, which need further study. Priti (1998) supports our observations that A. florea was the most abundant pollinator of onion followed by A. mellifera in lowland conditions. The foraging activity by bees and flies started at around 6:00 h, which is supported by the findings of Chandel et al. (2004). The foraging activity of A. dorsata, A. florea and all other dipteran pollinators peaked between 10:00-12:00 h, however, peak activity for A. dorsata and A. cerana was observed between 12:00-14:00 h (Partap and Verma, 1994; Priti, 1998; Chandel et al., 2004). Their findings might be due to difference in climate and latitude. In the present studies, the activity of all pollinators was comparatively low early in the morning and late in the evening. This diurnal pattern of abundance can be useful for scheduling pest control operations.

112

Foraging pattern and abundance alone are not good indicators of any floral visitor as an effective pollinator. Many other factors also contribute such as body size, shape of an insect, its thurst for nectar or pollen, or the chances to be in contact with the stigma of the flowers and pollen deposition. A. dorsata proved to be the most effective pollinator in our study, which is supported by the work of several authors (Kutjatnikova, 1969; Jefimockina, 1971; Martin, 1978; Lazic et al., 1985; Kumar et al., 1989; Priti, 1998; Chandel et al., 2004). On the other hand, from a closely related plant species, Stephen et al. (2007) harvested a higher Allium ampeloprasum L. seed yield (340.7g/47 umbels) contributed by Calliphora vicina as compared from M. domestica (70.5g/47 umbels). We also observed Calliphoridae sp. as the second most effective pollinator among all the pollinators and most effective among the Diptera. In our experiment, the comparison of yield obtained from open and self-pollinated crop represents a difference of 616% of the yield in open pollinated crop (Fig. 5). Similar results were also observed by Zdzislaw et al. (2004) who found 699% more yield in open pollinated crop than self pollinated crop. Experiment no. 2 We recorded only the bee species i.e., Apis dorsata, A florea, Megachile bicolor, Megachile sp., Nomia sp. And Xylocopa sp. Our this finding is in accordance with Gebremariam (2002); Heering (1994); Arroyo (1981); Brewbaker (1990); Evans and Roter (1987) who found that S. sesban flowers are visited by a wide range of bee species including A. mallifera, M. bituberculata (mason bee), Chalicodoma sp. (leaf cutting bee), Bombus canariensis, Xylocopa flavorufa and X. somalica. This is because floral structure of S. sesban with pollen and nectar concealed in the papilionoid flowers (Burbidge, 1965) seems to be suited for cross pollination by bees. Abundance of pollinators may increase the outcrossing rate (Gebremariam et al., 2002), which may vary greatly among and within population (Brown et al., 1989). In our study most of the pollinator visitation took place in afternoon (1530–1830 h). This is because S. sesban flowers open in afternoon (Heering, 1994) and pollinators are only attracted when the flowers are opened. Flower and pod abortion is a common phenomenon in S. sesban. Out-crossing is probably the preferred method of reproduction under natural conditions but selfing is clearly possible (Brewbaker, 1990; Heering, 1994). According to FAO (1961) S. sesban

113 does not require isolation for pure seed production due to its self-compatibility. Our results are also in agreement with this as there was no significant difference between number of fertilized flowers in cross pollinated (open pollinated) and self-pollinated (butter paper and mesh bags) inflorescences. Self-pollination may lead to inbreeding depression (Seavey and Bawa, 1986) and it may affect the latter stages of lifecycle. As in our study, pod abortion was greater in self-pollinated inflorescences (Nylon mesh and butter paper) as compared to open pollinated treatment. Contrary to our findings, Heering (1994) suggested that pod abortion is a common phenomenon and self or cross pollination is not a limiting factor in it. There were 23.73 seeds produced per pod under open pollination, where as self pollinated pods produced 18.83 seeds. Similar results were obtained by Heering (1994) who found 26 seeds from open pollinated pod and 20 seeds from self pollinated pod. Selective abortion is also indicated by the pattern of seeds produced within the pod. Bawa and Webb (1984) observed greater seed settings in the distal end of the pod. Self pollination occurs latter in the flowering period (Gebremariam et al., 2002) when pollinators fail to visit the flowers. Such a “delayed selfing” favors the plant when pollinators are limiting. In conclusion, bees played very important role in cross pollination of S. sesban and ultimately crop reproductive success by increasing seed weight per inflorescence (1.1 g) and germination (16%) as compared to other two treatments of self and wind pollination. In addition to cross pollination, further studies should be done on pollination potential of different native bee species, which may influence cross pollination by having different abilities of depositing pollen on stigma based on body size and shape. Since native pollinators have shown great potential in pollinating agricultural corps, their on-farm conservation can be the best option in the plains of Punjab. This study provides a baseline for future researches on conservation of native pollinators which should include exploring the biology of native pollinators and devising some pollinator-friendly practices for farmers for sustainable agriculture production.

114

References

Ågren, J., 1996. Population size, pollinator limitation and seed set in the self- incompatible herb Lythrum salicaria. Ecology, 77: 1779–1790. Ahmed, I.H. and A.A. Abdalla, 1984. The role of honeybees as pollinators on onion (Allium cepa L.) seed production. Acta Horticult., 143: 127-132.

Alarcón, R., N.M. Waser and J. Ollerton, 2008. Year-to-year variation in the topology of a plant-pollinator interaction network. Oikos. 117: 1796-1807. Anonymous, 2000. Biodiversity action plan for Pakistan. Imprint (pvt) Ltd., Rawalpindi Cantt., Pakistan. pp.1.

Arif, M.J., A. Suhail and M. Yousaf, 2001. Genus Sphaerophoria St. Farg et Serv. from Pakistan. Int. J. Agri. Biol., 3(2):260-261.

Armbruster, WS, C.B. Fenster and M.R. Dudash, 2000. Pollination "principles" revisited; specialization, pollination syndromes, and the evolution of flowers. Det Nor. Vidensk. Acad. I. Mat. Natur. Kl. Skr. Ny Ser., 39:179-200.

Armbruster W.S. and A.L. Herzig, 1984. Partitioning and sharing of pollinators by four sympatric species of Dalechampia (Euphorbiaceae) in Panama. Ann. Mo. Bot. Gard., 71:1–16.

Arroyo, M.T.K., 1981. Breeding systems and pollination biology in Leguminosae. In: Polhill, R.M. and P.H. Raven (eds.), Advances inLegume Systematics, Part 2, pp: 724–769. Royal Botanic Gardens, Kew, England.

Baker H.G., 1961. The adaptation of flowering plants to nocturnal and crepuscular pollinators. Q. Rev. Biol., 36:64-73.

Baldock, K.C.R., J. Memmott, J.C.R. Guajardo, D. Roze and G.N. Stone, 2011. Daily

temporal structure in African savannah flower-visitation networks, and

consequences for network sampling. Ecology. 92: 687-698.

Barrett, S.C.H. and K. Helenurm, 1987. The reproductive biology of boreal forest herbs. I. Breeding systems and pollination. Can. J. Bot., 65: 2036-2046.

115

Bascompte, J., P. Jordano, C. Melián and J.M. Olesen, 2003. The nested assembly of plant-animal mutualistic networks. Proceedings of the National Academy of Sciences, 100: 9383-9387. Basilio, A.M., D., Medan, J.P. Torretta and N.J. Bartoloni, 2006. A year-long plant- pollinator network. Austral Ecology, 31: 975–983.

Bawa, K.S. and C.J. Webb, 1984. Flower, fruit and seed abortion in tropical forest trees: implications for the evolution of paterual and maternal reproductive patterns. American J. Bot., 71: 736–751. Beals, E.W., 1984. Bray-Curtis ordination: an effective strategy for analysis of multivariate ecological data. Adv. Ecol. Res., 14: 1-55.

Benedek, Z., F. Jordan and A. Baldi, 2007. Topological keystone species complexes in ecological interaction networks. Community Ecology, 8(1): 1-7.

Bertin, R.I. 1989. Pollination biology, pp. 23-83. In W.G. Abrahamson (ed.), Plant- Animal interactions. McGraw Hill, New York.

Bhandari, M.M., 1978. Flora of Indian desert. Scientific Pub. Jodhpur.

Bingham, C.T., 1987. The Fauna of British India, including Ceylon and Burma, Hymenoptera. Vol. I Wasps and Bees. Taylor and Francies, London.

Blüthgen, N., J. Fründ, D.P. Vazquez and F. Menzel, 2008. What do interaction network metrics tell us about specialization and biological traits? Ecology, 89: 3387-3399. Blüthgen, N., F. Menzel and N. Blüthgen, 2006. Measuring specialization in species interaction networks. BMC Ecology, 6: 9. Bohart, G.E., W.P. Nye and L.R. Hawthorn, 1970. Onion pollination as affected by different levels of pollinator activity. Utah Agric. exp. Stn. Bull., 482.

Borror, D.J., D.M Long and C.A. Triplehorn, 1981. An Introduction to the Study of Insects. Saunders College Publishing, Philadelphia.

Brewbaker, J.L., 1990. Breeding systems and genetic improvement of perennial Sesbanias. In: Macklin, B. and D.O. Evans (eds.), Perennial Sesbania Species in Agroforestry Systems, Proceedings of a Workshop Held in Nairobi, Kenya, March

116

27-31, 1989, pp: 39–44. Nitrogen Fixing Tree Association, Waimanalo, Hawai, USA

Bronstein,J.L., 1995. The plant-pollinator landscape. In: Hansson,L., Fahrig,L. & Merriam,G. (eds.), Mosaic landscapes and ecological processes, pp. 256-288. Chapman & Hall, London, UK.

Brown, A.H.D., J.J. Burdon and A.M. Jarosz, 1989. Allozyme analysis of plant mating systems. In: Soltis, D.E. and P.S. Soltis (eds.), Isozymes in Plant Biology, pp: 73– 86. Chapman and Hall, London

Brown, B.J., R.J. Mitchell and S.A. Graham, 2002. Competition for pollination between an invasive species (Purple loosestrife) and a native congener. Ecology, 83: 2328- 2336.

Brown, A.H.D. and R.W. Allard, 1970. Estimation of the mating system in openpollinating maize populations using allozyme polymorphisms. Genetics, 66: 133–145.

Buchmann, S.L. and G.P. Nabhan, 1996. The forgotten pollinators. Island Press, Washington D.C., USA.

Burbidge, N.T., 1965. The Australian species of Sesbania Scopoli. (Leguminosae). Australian J. Bot., 13: 103–141.

Campbell, D.R. and A.F. Motten, 1985. The mechanism of competition for pollination between two forest herbs. Ecology. 66: 554-563.

Campbell, D.R., 1985. Pollen and gene dispersal: the influences of competition for pollination. Evolution, 39: 418-431.

Carvalho, A.M.C., J. Mendes, C.H. Marghiori and C. Lomônaco, 1991. Variação espacial e sazonal de dipterous muscóides em duas áreas de cerrado no município de Uberlândia - MG. I. Calliphoridae e Muscidae. R. Cent. Ci. Bioméd. Univ. Fed. Uberlândia., 7: 27-34.

117

Chandel R.S., R.K. Thakur, N.R. Bhardwaj and N. Pathania, 2004. Onion seed crop pollination: a missing dimension in mountain horticulture. Acta Horticult., 631: 79-86.

Colwell, R.K., 2005. EstimateS-statistical estimation of species richness and shared species from samples, version 7.5, at http://www.purl.oclc.org/estimates. Date of access: 20 Aug. 2009.

Corbet, S.A., 1997. Role of pollinators in species preservation, conservation, ecosystem stability and genetic diversity. Acta Horticult, 437: 219-228.

Cox, P.A. and T. Elmqvist, 2000. Pollinator extinction in the Pacific islands. Conservation Biology, 14: 1237-1239.

Cunningham, S.A., 2000. Depressed pollination in habitat fragments causes low fruit set. Proc. R. Soc. Lond. 567:1149-1152.

Currah, L. and D. J. Ockendon, 1983. Onion pollination by blowflies and honey bees in large cages. Ann. Biol., 103:497-506.

Davis, S.D. et al., 1986. Plants in Danger: What Do We Knoow? IUCN, Gland, Switzerland.

De Oliveira, V.R. and I.E. Pires, 1990. Pollination Efficiency of P. juliflora (Sw) DC in Petrolina, Pernambuco. In: The Current State of Knowledge on P. juliflora. M.A. Habit and J.C. Saavedra (Eds). FAO, Rome, Italy. pp. 233_239.

Delpino, F., 1868. Ulteriori osservazione sulla dicogamia nel regno vegetale. Atti della Societa Italiana di Scienze Naturali Milano, Vols. 1. devi Y.R., J. Kalita and T.K. Singh, 2011. Biological control potential of an aphidophagous syrphid, Episyrphus balteatus, De-Geer (Diptera : Syrphidae) on mustard aphid, Lipaphis erysimi (Kalt.) (Homoptera : Aphididae) on cabbage ecosystem in Manipur. Journal of Experimental Sciences, 2(12): 13-16.

118

Devy, M.S. and P. Davidar, 2003. Pollination systems of trees in Kakachi, a mid- elevation wet evergreen forest in Western Ghats, India. American Journal of Botany, 90(4): 650-657.

Dewenter, S. I. and T. Tscharntke, 1999. Effects of habitat isolation on pollinator communities and seed set. Oecologia, 121:432-440.

Dicks L.V., S.A. Corbet and R.F. Pywell, 2002. Compartmentalization in plant-insect flower visitor webs. J. Anim. Ecol., 71, 32–43.

Donaldson J., J. Nanni, C. Zachariades and J. Kemper, 2002. Effects of Habitat Fragmentation pollinator diversity and plant reproductive success in renosterveld shrublands of South Africa. Conservation biology, 16(5): 1267-1276.

Dormann, C.F., B. Gruber, J. Fründ , 2008. Introducing the bipartite Package: Analysing Ecological Networks. R News. 8: 8-11. Dormann, C.F., J. Fründ, N. Blüthgen, B. Gruber, 2009. Indices, graphs and null models: analyzing bipartite ecological networks. The Open Ecology Journal. 2: 7-24. Dunne,J.A., R.J. Williams and N.D. Martinez, 2002. Network structure and biodiversity loss in food webs: rubustness increases with connectance. Ecology Letters, 5: 558-56.

Dupont, Y.L., D.M. Hansen and J.M. Olesen, 2003. Structure of a plant-flower– visitor network in the high-altitude sub-alpine desert of Tenerife, Canary Islands. Ecography, 26: 301 – 310.

Elberling, H. and J.M. Olesen, 1999. The structure of a high latitude plant-flower visitor system: the dominance of flies. Ecography, 22: 314–23.

Ellis,W.N. and A.C. Ellis-Adam, 1993. To make a meadow it takes a clover and a bee: the entomophilous flora of N.W. Europe and its insects. Bijdragen tot de Dierkunde, 63: 193-220.

Erickson, H.T., and W.H. Gabelman, 1956. The effect of distance and direction on cross- pollination in onions. Am. Soc. Hortic. Sci. Proc., 68: 351-357.

Evans, D.O. and E.P. Rotar, 1987. Sesbania in Agriculture, Westview Tropical Agriculture Series; No. 8. Boulder, Colorado (USA), Westview Press . 119

Evans, D.O., 2001. Sesbania sesban: Widely distributed multipurpose NFT. In: Roshetko, J.M. (ed.) Agroforestry Species and Technologies . pp. 157- 158. Winrock International, Morilton, USA.

Fægri, K. and L. van der Pijl, 1966. The Principles of Pollination Ecology. Oxford: Pergamon. 248pp.

Faegri, K. and L. van der Pijl, 1971. The Principles of Pollination Ecology. 2nd. ed. Pergamon Press, Oxford, UK.

Fægri, K. and L. van der Pijl, 1979. The principles of pollination ecology. Pergamon Press, London, UK.

FAO, 1961. Agricultural and Horticultural Seeds; Their Productiom, Control and Distribution. FAO Agricultural Studies No. 55. Food and Agricultural Organization of the United Nations, Rome.

Feinsinger, P., 1987. Effects of plant species on each other's pollination: Is community structure influenced? Trends in Ecology and Evolution, 2: 123-126.

Fenster, C.B., W.S. Armbruster, P. Wilson, M.R. Dudash, and J.D. Thomson, 2004. Pollination syndromes and floral specialization. Annu. Rev. Ecol. Evol. S., 35: 375–403.

Fenster, C.B. and M.R. Dudash, 2001. Spatiotemporal variation in the role of hummingbirds as pollinators of Silene virginica. Ecology, 82: 844-851.

Figueroa-Castro, D.M. and Z. Cano-Santana, 2004. Floral visitor guilds of five Allochronic flowering Asteraceous species in xeric community in central Mexico. Environmental Entomology, 33(2):297-309.

Fishbein, M. and D.L. Venable, 1996. Diversity and temporal change in the effective pollinators of Asclepias tuberosa. Ecology, 77: 1061–1073.

Fisher, B.L., 1998. Insect behavior and ecology in conservation: preserving functional species interactions. Conservation and Biodiversity, 91: 156-158.

120

Fontaine C., I. Dajoz, J. Meriguet and M. Loreau, 2006. Functional diversity of plant- pollinator interaction webs enhances the persistence of plant community. Plos- biology, 4(1): 129-135.

Forup, M. L. and J. Memmott, 2005. The restoration of pollination processes in British meadows. Restoration Ecology, 13:265-274.

Forup, M.L., K.E. Henson and J. Memmott, 2008. The restoration of ecological interactions: plant-pollinator networks on ancient and restored heathlands. Journal of Applied Ecology, 45: 742–752.

Franklin, D.F., 1970. Problems in the production of vegetable seed. In: The indispensable pollinators. Ark. Agr. Ext. Serv. Misc. Pub., 127: 112-140.

Free, J.B., 1960. The behaviour of honeybees visiting the flowers of fruit trees. J. Anim. Ecol., 29: 385-395.

Free, J. B., 1993. Insect pollination of crops (2nd edition). Academic press inc. ltd., London.

Freitas, B. M. and R. J. Paxton, 1998. A comparison of two pollinators: the introduced honey bee Apis mellifera and an indigenous bee Centris tarsata on cashew

Anacardium occidentale in its native range of NE Brazil. J. Appl. Eco., 35: 109– 121.

Freitas, L. and M. Sazima, 2003. Daily blooming pattern and pollination by Syrphids in Sisyrinchium vaginatum (Iridaceae) in Southeastren Brazil. Journal of the Torrey Botanical Society, 130(2): 55-61.

Galeano, J., J.M. Pastor, J.M. Iriondo, 2009. Weighted-Interaction Nestedness Estimator (WINE): A new estimator to calculate over frequency matrices. Environ. Modell. Softw., 24: 1342-1346. Gebremariam G., S. Nemomissa, A. Demissie and J. Hanson, 2002. The mating system of Sesbania sesban (L.) Merr. (Leguminosae). Ethiopian J. Sci., 25(2): 177-190.

Ghahari, H., R. Hayat, M. Tabari and H. Ostovan, 2008. Hoverflies (Diptera:Syrphidae) from rice fields and around grasslands of Northern Iran. Mun. Ent. Zool., 3(1): 275-283. 121

Goel, V. L. and H. M. Behl, 1995. Propagation of Prosopis juliflora from rooted stem cuttings. International Tree Crops Journal, 8:193-201.

Gomez, J.M., J. Bosch, F. Perfectti, J. Fernandez and M. Abdelaziz, 2007. Pollinator diversity affects plant reproduction and recruitments: the tradeoffs of generalization. Oecologia, 153:597-605.

GOP, 1992. Forestry Sector Master Plan. Government of Pakistan.

Gotelli, N.J. and D.F. Entsminger, 2005. EcoSim, Null models software for ecology, v 7.72 Acquired Intelligence Inc and Kesey-Bearm, at http://www.homepages.together.net/~gentsmin/ecosim.htm.

Goulson, D. and N.P. Wright, 1998. Flower constancy in the hoverflies Episyrphus balteatus (Degeer) and Syrphus ribesii (L.) (Syrphidae). Behavioral Ecology, 9(3): 213-219.

Grant ,V. (1949). Pollination systems as isolating mechanisms in angiosperms. Evolution, 3:82-97.

Grant, V. and K.A. Grant, 1965. Flower pollination in the Phlox family. Columbia Universit University Press, New York. USA.

Guimaraes, P.R., G. Machado, M.A.M. de Aguiar, P. Jordano, J. Bascompte, , A. Pinheiro and S.F. dos Reis, 2007. Build-up mechanisms determining the topology of mutualistic networks. J. Theor. Biol., 249, 181–189.

Gupta, R.K., 2004. Advances in insect biodiversity. Agrobios, Jodhpur.

Hamer, K. C., J. K. Hill, N. Mustaffa, S. Benedick, T. N. Sherratt, V. K. Chey and M. Maryati, 2005. Temporal variation in abundance and diversity of butterflies in Bornean rain forests: opposite impacts of logging recorded in different seasons. Journal of Tropical Ecology., 21:417–425.

Hartley, J.C., 1961. A taxonomic account of the larvae of some British Syrphidae, Proc Zool. Soc. London.,. 136: 505-573.

Haslett, J.R., 1989. Interpreting patterns of resource utilization: randomness and selectivity in pollen feeding by adult hoverflies. Oecologia, 78: 433-442.

122

Heering, J.H., 1994. The reproductive biology of three perennial Sesbania species (Leguminosae). Euphytica, 74: 143–148.

Hegland, S.J. and L. Boeke, 2006. Relationship between the density and diversity of floral resources and flower visitor activity in a temperate grassland community. Ecological Entomology, 31:532-538.

Herrera, C. M., 1996. Floral traits and plant adaptation to insect pollinators: a devil's advocate approach. In: D.G. Lloyd and S. C.H. Barrett (Eds.), Floral Biology, pp. 65-87. Chapman and Hall, New York.

Herrera, C. M., 2005. Plant generalization on pollinators, species property or local phenomenon? Am. J. Bot., 92:13-20.

Herrera, C.M. 1988. Variation in mutualisms: the spatio-temporal mosaic of a pollinator assemblage. Biol. J. Linn. Soc. 35: 95-125.

Herrera, C.M., 1990. The adaptedness of the floral phenotype in a relict endemic, hawkmoth-pollinated violet. 1. Reproductive correlates of floral variation. Biological J. Linnean Soc., 40: 263–274

Hingston, A.B. and P.B. McQuillan, 2000. Are pollination syndromes useful predictors of floral visitors in Tasmania? Austral Ecology, 25:600-609.

Holloway, B.A., 1976. Pollen-feeding in hover-flies (Diptera:Syrphidae). N. Z. Journal of Zoology, 3: 339-350.

Hosier, R., 1993. Forest energy in Pakistan: the evidence for sustainability. Pakistan Household Energy Strategy Study (HESS). Government of Pakistan Energy Wing, Islamabad.

Hwang, J.K., K. E. Lee, S. E. Maeng and J. W. Lee, 2008. Scaling behaviors of plant- pollinator mutualistic networks. Journal of the Korean Physical Society, 53(6):3151-3155.

Inouye, D.W. and A. Kearns, 1993. Variation in dipteran pollination popullation: monitoring by malaise traps in tropics, pp. 264-265. Proceedings of the International Symposium in Tropics. University of Agricultural Sciences.

123

Iqbal, M.Z. and M. Shafiq, 1997. Seedling performance of two desert plant species (Prosopis juliflora and Blepharis sindica) grown under uniform edaphic conditions. Journal of Tropical Forestry, 9(4):458-464.

Irshad, M. 2001. Aphids and their biological control in Pakistan. Pak. J. Biol. Sci., 4(5): 537-541.

Ivey, C.T., P. Martinez, and R. Wyatt, 2003. Variation in pollinator effectiveness in swamp milkweed, Asclepias incarnata (Apocynaceae). Am. J. Bot., 90: 214–225.

Jablonski, B., J. Skowronek, H.W. Woyke and R.W. Doruchowski, 1982. Biologia kwitnienia, nektarowanie, zapylanie i owocowanie mêskosterylnych linii cebuli (Allim cepa L.). Pszczel. Zesz. Nauk, 26: 57-104.

Jauker, F. and V. Wolters, 2008. Hover flies are efficient pollinators of oil seed rape. Oecologia, 156: 819-823.

Jefimochkina, O.H., 1971. Spaskij Luk I sortouluchshytelnaja rabota s nim. Avtoreferat.

Jennersten, O. and S.G.Nilsson, 1993. Insect flower visitation frequency and seed production in relation to patch size of Viscaria vulgaris (Caryophyllaceae). Oikos, 68: 283-292.

Johnson, S.D. and K.E. Steiner, 2000. Generalization vs. specialization in plant pollination systems. Trends Ecol. Evol., 15:140–43.

Jones, H.A. And Davis, G.N., 1944. Inbreeding and heterosis and their relation to the development of new varieties of onions. USDA Tech. Bull., 874.

Jordan F., I. Scheuring and G. Vida, 2002. Speceis positions and extinction dynamics in simple food webs. Journal of Theoretical Biology, 215:441-448.

Jordano, P., J. Bascompte and J. M. Olesen, 2003. Invariant properties in coevolutionary networks of plant–animal interactions. Ecol. Lett., 6: 69–81.

Jordano, P., 1990. Biología de la reproducción de tres especies del género Lonicera (Caprifoliaceae) en la Sierra de Cazorla. Anales Del Jardín Botánico De Madrid, 448: 31–53.

124

Jordano, P. 1987. Patterns of mutualistic interactions in pollination and seed dispersal: Connectance, dependence asymmetries, and coevolution. The American Naturalist, 129: 657-677.

Kaiser, C. N., 2006. Functional integrity of plant–pollinator communities in restored habitats in Mauritius. Dissertation. University of Zu ¨rich, Switzerland. Kallimanis A. S., T. Petanidou, J. Tzanpoulos, J.D. Pantis and S.P. Sgardelis, 2009. Do plant-pollinator interaction networks result from stochastic process? Ecological Modelling, 220: 684-693.

Kearns, C.A., 2001. North American dipertran pollinators: assessing their value and conseration status. Consercation Ecology, 5(1):5. http://www.consecol.org/vol5/iss1/art5.

Kearns, C.A. and D.W. Inouye, 1997. Pollinators, flowering plants, and conservation biology. Bioscience, 47: 297-307.

Kearns, C.A., D.W. Inouye and N.M. Waser, 1998. Endangered mutualisms: The conservation of plant-pollinator interactions. Annual Review of Ecology and Systematics, 29: 83-112.

Kennedy, T.A., S. Naeem, K.M. Howe, J.M.H. Knops, D. Tilman and P.B. Reich, 2002. Biodiversity as a barrier to ecological invasion. Nature, 417: 636-638.

Kiester A.R., R. Lande and D.W. Schemske, 1984. Models of co-evolution and speciation in plants and their pollinators. Am. Nat., 124:220-43.

Kite, G.C., W.L.A. Hetterrscheid, M.J. Lewis, P.C. Boyce, J. Ollerton, E. Cocklin, A. Diaz and M.S.J. Simmonds, 1998. Inflorescence odours and pollinators of Arum and Amorphophallus (Araceae). pp. 295–315. In : Reproductive Biology. Owens S.J. and P.J. Rudall (eds.). Royal Botanic Gardens, Kew.

Khan, S.U., M. Hassan, F.K. Khan and A. Bari, 2010. Climate classification of Pakistan. Balwois 2010 Conference, Ohrid, Republic of Macedonia, pp.1-47. http://www.balwois.com/balwois/administration/full_paper/ffp-1295.pdf: 1–47

125

Kleijn D. and F. V. Langevelde, 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, I. Steffan-Dewenter and T. Tscharntke, 2003. Fruit set of highland coffee increases with the diversity of pollinating bees. Proceedings of the Royal Society of London. Series B. Biological Sciences, 270: 955 – 961.

Knuth, P. 1906. Handbook of Flower Pollination.Vol. I. Transl. JR Ainsworth Davis. Oxford: Clarendon. 382 pp.

Kölreuter, J.G., 1761. Vorl¨aufige Nachrichten von einigen das Geschlecht der Pflanzen betreffenden Versuchen und Beobachtungen.Leipzig: Gleditschischen Handlung.

Kumar, J., R. C. Mishra and J. K. Gupta, 1989. Effect of honeybee pollination on onion (Allium cepa L.) seed production. Indian Bee J., 51:3-5.

Kutjatnikova, R.J., 1969. Biologia cvenija repchatova luka v polupustinnoj zone Severnoho priaralja. Trudy po prikladnoj botanika genetike I selekcii, vol. 32, Moskova.

Kwak, M.M., 1994. Planten en bestuivers: achteruitgang leidt tot verschuivende relaties. Landschap 11: 29-39.

Kwak, M.M., O. Velterop, and J. van Andel, 1998. Pollen and gene flow in fragmented habitats. Applied Vegetation Science, 1: 37-54.

Lazic, B., D. Haker and R. Markov, 1985. Prilog proucavanju uticanu peela na oplodnju crnog luka, X jugoslovenski simpozijum o semenarstvu, Piltvicka jezera.

Lehman, C.L. and D. Tilman, 2000. Biodiversity, stability and productivity in competitive communities. The American Naturalist, 156: 534-552.

Leppik E.E. 1977. The evolution of capitulum types of the Compositae in the light of insect-flower interaction, pp. 61-89. In V.H. Heywood, J.B. Harborne, and B.L. Turner (eds.), The biology and chemistry of the Compositae, vol. 1. Academic, London, England.

126

Lesley, C. and D.C. Ockendon, 1978. Protandry and sequence of flower opening in the onion (Allium cepa L.). New Phytol., 81: 419-428.

Loreau, M., S. Naeem, and P. Inchausti, 2002. Biodiversity and ecosystem functioning. Oxford University Press, New York, USA.

Luijten, S. H., 2001. Reproduction and genetics of fragmented plant populations. Ph.D- thesis, University of Amsterdam, the Netherlands.

Lundgren, R. and J.M. Olesen, 2005. The dense and highly connected world of Greenland's plants and their pollinators. Arctic, Antarctic and Alpine Research, 37(4):514-520.

Macklin, B. and D.O. Evans, 1990. Perennial Sesbania species in agroforestry systems. Proceedings of a Workshop Held in Nairobi, Kenya, March 27-31, 1989. Nitrogen Fixing Tree Association, Waimanalo, Hawai, USA.

Magurran, A.E. 2004. Ecological diversity and measurements. 2nd edn. Princeton University Press, Princeton, N.J.

Martin, E.C., 1978. The use of bees for crop pollination. In: The hive and honey bee. Hamilton, Illinois.

Mayer, D.F and J.D. Lunden, 2001. Honey bee management and wild bees for pollination of hybrid onion seed. Acta Horticult., 561: 275-278.

McCall, C. and R.B. Primack, 1992. Influence of flower characteristics, weather, time of day, and season on insect visitation rates in three plant communities. Am. J. Bot., 79:434-442.

McCann, K.S., 2000. The diversity-stability debate. Nature, 405: 228-233.

McGregor, S.E. 1976. Insect pollination of cultivated crop plants. Agriculture Handbook No. 496, USDA-ARS, U.S. Gov. Print.

Medan, D., N.H. Montaldo, M. Devoto et al., 2002. Plant pollinator relationship at two altitudes in the Andes of Mendoza, Argentina. Arctic, Antarctic Alpine Res., 34, 233-41.

Memmott, J., 1999. The structure of a plant-pollinator food web. Ecol. Lett., 2: 276–80.

127

Mesquida, J., M. Renard and J.S. Pierre, 1988. Rapeseed (Brassica napus L.) productivity: The effect of honey bee (Apis mellifera L.) and different pollination condition in cage and field tests. Apidologie, 19: 51-72.

Michener, C.D., 2000. The Bees of the World- John Hopkins University Press, Baltimore.

Minchin, P.R., 1987. An evaluation of relative robustness of techniques for ecological ordination. Vegetatio, 69: 89-107.

Moffett, J.O., 1965. Pollinating experimental onion varieties. Am. Bee J., 105: 378.

Moldenke, A.R. and P.G. Lincoln, 1979. Pollination ecology in montane Colorado: a community analysis. Phytologia, 42:349–79.

Moll, R.H., 1954. Receptivity of the individual onion flower and some factors affecting its duration. Am. Soc. Hort. Sci. Proc., 64: 399-404.

Morse, R. and N.W. Calderone, 2000. The value of honey bees as pollinators of U.S. Crops in 2000. Bee Culture, 128: 1–15.

Murren, O.J. 2002. Effect of habitat fragmentation on pollination: pollinators, pollinia viability and reproductive success. Journal of Ecology, 90:100-107.

NyHagen, D.F., C. Kragelund, J.M. Olesen, and C.G. Jones, 2001. Insular interactions between lizards and flowers: flower visitation by an endemic Mauritian gecko. Journal of Tropical Ecology, 17: 755-761.

Ockinger E. and Smith H D. 2007. Semi-natural grasslands as population sources for pollination insects in agricultural landscapes. Journal of Applied Ecology. 44: 50- 59.

Olesen, J.M. and P. Jordano, 2002. Geographic patterns in plant-pollinator mutualistic networks. Ecology, 83: 2416-2424.

Olesen, J.M., J. Bascompte, H. Elberling and P. Jordano, 2008. Temporal dynamics in a pollination network. Ecology, 89: 1573-1582. Ollerton, J. and S. Watts, 2000. Phenotype space and floral typology—towards an objective assessment of pollination syndromes. Det.Nor. Vidensk. Acad. I. Mat. Natur. Kl. Skr.Ny Ser., 39:149–59.

128

Ollerton, J. and L. Cranmer, 2002. Latitudinal trends in plant-pollinator interactions: are tropical plants more specialized? Oikos, 98:340.

Ollerton, J. and C. Louise, 2002. Latitudinal trends in plant-pollinator interactions: are tropical plants more specialized? Oikos, 98: 340-350.

Olsen, K.M., 1997. Pollination effectiveness and pollinator importance in a population of Heterotheca subaxillaris (Asteraceae). Oecologia, 109: 114–121.

Oostermeijer, J.G.B., M.W. van Eijck and J.C.M. den Nijs, 1994. Offspring fitness in relation to population size and genetic variation in the rare perennial plant species Gentiana pneumonanthe (Gentianaceae). Oecologia, 97: 289-296.

Owuor, O.O and E. Owino, 1993. Control pollination and pollen management in Sesbania sesban (L.) Merr. Euphytica, 70: 161-165.

PARC, 1980. Agro-ecological regions of Pakistan. PARC, Islamabad

Partap, U. and L.R. Verma, 1994. Pollination of radish by Apis cerana. J. Apic. Res., 33: 237-241.

Petanidou, T. and S.G. Potts, 2006. Mutual use of resources in Mediterranean plant– pollinator communities, how specialized are pollination webs? In: Waser, N.M., Ollerton, J. (Eds.), Plant–Pollinator Interactions from Specialization to Generalization. University of Chicago Press, Chicago, pp. 220–244.

Petanidou, T., A.S. Kallimanis, J. Tzanopoulos, S.P. Sgardelis and J.D. Pantis, 2008. Long-term observation of a pollination network: fluctuation in species and interactions, relative invariance of network structure, and implications for estimates of specialization. Ecol. Lett., 11:564-575. Potts, S.G., B. Vulliamy, A. Dafni, G. Ne'eman, and P. Willmer, 2003. Linking bees and flowers: how do floral communities structure pollinator communities? Ecology, 84: 2628-2642.

Priti, 1998. Abundance and pollination efficiency of insect visitors of onion bloom. Indian Bee J., 60: 75-78.

129

Proctor, M.C.F., 1978. Insect pollination syndromes in an evolutionary and ecosystem context. In: Richards,A.J. (ed.), The pollination of flowers by insects, The Linnean Society Symposium Series 6, pp. 105-116. The Linnean Society of London, Academic Press, London, UK.

R Development Core Team, 2008. R: A language and envi-ronment for statistical computing . R Foundation for StatisticalComputing, Vienna, Austria. ISBN 3- 900051-07-0, URLhttp://www.R-project.org

Rao, G.M. and M.C. Suryanarayana, 1990. Studies on the foraging behaviour of honeybees and its effect on seed yield in Niger. Indian Bee J., 52: 31-33.

Rao, G.M. and M.C. Suryanarayana, 1989. Effect of honeybee pollination on seed yield in onion (Allium cepa L.). Indian Bee J., 51: 9-11.

Rathcke, B.J. and E.S. Jules, 1993. Habitat fragmentation and plant-pollinator interactions. Current Science, 65: 273-277.

Richards, A.J., 2001. Does low biodiversity result from modern agricultural practices affect crop pollination and yield? Annl. Bot., 88:165-172.

Richardson, D.M., N. Allsopp, C.M. D'Antonio, S.J. Milton and M. Rejmanek, 2000. Plant invasions-the role of mutualisms. Biol. Rev., 75:63-99.

Root, R.B., 1973. Organization of a plant-arthropod association in simple and diverse habitats: the fauna of collards (Brassica oleraceae). Ecol. Monogr., 43:95-124.

Roubik, D. W., 1995. Pollination of cultivated plants in the tropics, FAO Agricultural Services Bulletin 118.

Saeed, S., A. Sajjad, O. Kwon and Y.J. Kwon, 2008. Fidelity of Hymenoptera and Diptera pollinators in onion (Allium cepa L.) pollination. Entomol. Res., 38: 276- 280.

Sajjad, A., S. Saeed and A. Bashir, 2012. Spatial variation in spatial variation in pollinator communities and reproductive performance of Prosopis juliflora (fabaceae). Journal of Pollination Ecology, 8(9): 59-66.

130

Sajjad, A., S. Saeed and A. Masood, 2008. Pollinator community of onion (Allium cepa L.) and its role in crop reproductive success. Pak. J. Zool., 40(6): 451-456.

Sajjad, A., S. Saeed and M. Ashfaq, 2010. Seasonal variation in abundance and composition of hoverfly (Diptera: Syrphidae) communities in Multan, Pakistan. Pakistan J. Zool., 42(2): 105-115.

Sajjad, A. and S. Saeed, 2009. New records of Eristalinus (Syrphidae: Diptera) from Multan, Pakistan. Pakistan J. Zool., 41(3): 238-239.

Saleem, M., M.J. Arif and A. Suhail, 2001. Taxonomic studies of Syrphidae of Peshawer- Pakistan. Int. J. Agri. Biol., 3(4): 533-534.

Saunders, D.A., R.J. Hobbs and C.R. Margules, 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology, 5: 18-32. Schittenhelm, S., T. Gladis and V.R. Rao, 1997. Efficiency of various insects in germ plasm regeneraton of carrot, onion and turnip rape accessions. Plant Breed., 116: 369-375.

Seavey, S.R. and K.S. Bawa, 1986. Late acting self incompatibility in Angiosperms. Bot. Rev., 52: 195–219.

Shahjahan, M., 1986. Effect of diet on the longevity and fecundity of the adults of the tachinid parasite Trichopoda pennipes pilipes. J. Econ. Entomol., 61: 1102-1103.

Shebl, M.A., S.M. Kamel, T.A. Abuhashesh and M.A. Osman, 2008. Seasonal abundance of leafcutting bees (Megachile minutissima, Megachilidae, Hymenoptera). World J. Agric. Sci., 4(2):280-287.

Sobota, G. and J. Twardowski, 2004. Variation in species spectrum of hoverflies (Diptera:Syrphidae) in arable crops depending on the collection method. Electronic journal of Polish Agricultural Univrsties, 7(2):08, http://www.ejpau.media.pl/volume7/issue2/biology/art-08.html.

Solecki, W. D., J. Long, C. C. Harwell, V. Myers, E. Zubrow, T. Ankerson, C. Deren, C. Feanny, R. Hamann, L. Hornung, and G. Snyder, 1999. Human-environment interaction in South Florida’s Everglades region: Systems of ecological degradation and restoration. Urban Ecosystems, 3(3-4): 305-343. 131

Sommaggio, D., 1999. Syrphidae: can they be used as environmental bioindicators? Agriculture, Ecosystems and Environment, 74: 343-356.

Souza-Silva, M., Fontenelle, J.C.R. And Martins, R.P., 2001. Seasonoal abundance and speceis compositon of flower-visiting flies. Neotropical Entomology, 30(3): 351-359.

Spira, T.P., 2001. Plant-pollinators interactions: a threatened mutualism with implications for ecology and management of rare plants. Natural Areas Journal, 21: 78-88.

Sprengel, C.K., 1793. Das entdeckte Geheimnissder Natur im Bau und in der Befruchtung derBlumen. Berlin: Vieweg.

Ssymank, A., C.A. Kearns, T. Pape, and F.C. Thompson, 2008. Pollinating flies (Diptera): A major contribution to plant diversity and agricultural production. Tropical conservancy, 9: 86-89.

Stang, M., G.L. Klinkhamer, and E. van der Meijden, 2006. Size constraints and flower abundance determine the number of interactions in a plant-flower visitor web. Oikos. 112: 111-121. Stebbins, G.L., 1970. Adaptive radiation of reproductive characteristics in angiosperms. I: Pollination mechanisms. Annu. Rev. Ecol. Syst., 1:307–26.

Steel, R.G.D. and J.H. Torrie, 1980. Principles and Procedures of statistics: A Biomaterical Approach, pp: 232–58. Mc-Graw Hill Book Co., Inc., New York.

Steffan-Dewenter I. and T. Tscharntke, 1999. Effects of habitat isolation on pollinator communities and seed set. Oecologia, 121:432-440.

Steffan-Dewenter I., U. Munzenberg, C. Burger, C. Thies, T. Tscharntke, 2002. Scaled- dependent effects of landscape context on three pollinator guilds. Ecology, 83: 1421-1432.

132

Stephen, L.C., B.C. Hellier, L.R. Elberson, R.T. Staska and M.A. Evans, 2007. Flies (Diptera: Muscidae: Calliphoridae) are efficient pollinators of Allium ampeloprasum L. (Alliaceae) in field cages. J. econ. Ent., 100: 131-135.

Stone, J. L., 1996. Components of pollination effectiveness in Psychotria suerrensis, a tropical distylous shrub. Oecologia, 107: 504–512.

Sun, M., 1997. Genetic diversity in three colonizing orchids with contrasting mating systems. American J. Bot., 84: 224–232.

Sutherland, J.P., S.S. Matthew and G.M. Poppy, 1999. The influence of floral character on the foraging behavior of the hoverfly, Episyrphus balteatus. Entomologia Experimentalis et Applicata, 93: 157-164.

Talavera, S., F. Bastida, P.L. Ortiz and M. Arista, 1999. Pollinator attendance and reproductive success in Cistus libanotis L. (Cistaceae). Int. J. Plant Sci.,162: 343- 352.

Talavera, S., F. Bastida, P.L. Ortiz and M. Arista, 2001. Pollinator attendance and reproductive success in Cistus libanotis L. (Cistaceae). International Journal of Plant Sciences, 162:343-352. Tolon, B. and Duman, Y., 2003. The effect of pollination by honeybee (Apis mellifera) on onion (Allim cepa) seed production and quality. XXXVIII Apimondia International Apicultural Congress, pp. 554.

Tomimatsu, H. and M. Ohara, 2002. Effects of forest fragmentation on seed production of the understory herb Trillium camschatcense. Conservation Biology, 16: 1277- 1285.

Tooker, J.F., H. Hauser and L.M. Hanks, 2006. Floral host plants of Syrphidae and Tachinidae (Diptera) of Central Illinois. Ann. Entomol. Soc. Am., 99(1):96-112.

Totland O., A. Nielsen, A.L. Bjerknes and M. Ohlson, 2006. Effects of an exotic plant and habitat disturbance on pollinator visitation and reproduction in a boreal forest herb. American Journal of Botany, 93(6): 868–873.

133

Traveset, A. and E. Sáez, 1997. Pollination of Euphorbia dendroides by lizards and insects: Spatio-temporal variation in patterns of flower visitation. Oecologia, 111: 241-248.

UNEP, 2005. Millennium Ecosystem Assessment Synthesis Report. Pre-publication Final Draft Approved by MA Board on March 23, 2005. Millennium Ecosystem Assessment, UNEP (United Nations Environmental Programme).

Van Der Meer, Q. P. and J.L. Van Bennekom, 1968. Research in pollen distribution in onion seed fields. Euphytica, 17: 216-219.

Van Treuren R., R. Bijlsma, N.J. Ouborg and M.M. Kwak, 1994. Relationship between plant density, outcrossing rates and seed set in natural and experimental populations of Scabiosa columbaria. J. Ecol .Biol., 7:287-302.

Vázquez, D. P. and M. A. Aizen, 2004. Asymmetric specialization: a pervasive feature of plant–pollinator interactions. Ecology, 85: 1251–1257.

Vázquez, D.P., N. Chacoff and L. Cagnolo, 2009. Evaluating multiple determinants of

the structure of plant–animal mutualistic networks. Ecology. 90: 2039-2046.

Viana, B.F. and A.M.P. Kleinert, 2005. A community of flower visiting bees (Hymenoptera:Apoidea) in the coastal sand dunes of northeastern Brazil. Biota Neotropica, 5(2): 79-91.

Vockeroth, J.R., 1996. A revision of genera of Syrphini (Diptera: Syrphidae). Mem. Ent. Soc., Canada., 62: 176.

Waser N. M., L. Chittka, M.V. Price, N.M. Williams and J. Ollerton, 1996. Generalization in pollination systems, and why it matters. Ecology, 77: 1043–60.

Waser, N.M., 1978. Interspecific pollen transfer and competition between co-occurring plant species. Oecologia, 36: 236.

Weigand, R.O., J.D. Reed, A.N. Said and V.N. Ummuna, 1995. Proanthocyanidins (condensed tannins) and the use of leaves from Sesbania sesban and S. goetzei as protein supplements. Animal Feed Sci. Technol., 54: 175–192.

134

Williams, I.H., 1996. Aspects of bee diversity and crop pollination in the European Union. In: A. Matheson, S.L. Buchmann, C. O'Toole, P. Westrich and I.H. Williams (eds.), The conservation of bees, pp. 63-80. Academic Press, London, UK.

Willmer, P. G., A. A. M. Bataw, and J. P. Hughes, 1994. The superiority of bumblebees to honey bees as pollinators: insect visits to raspberry flowers. Ecol. Entomol., 19: 271–284.

Wilson P., M.C. Castellanos, J.N. Hogue, J.D. Thomson and W.S. Armbrustre, 2004. A multivariate search for pollination syndromes among penstemons. OIKOS, 104: 345-361.

Wolda, H., 1988. Seasonality and the community pp. 69-95. In Gee, J.H.R. and P.S. Giller (eds). Organization of communities. Past and present: Blackwell. Oxford, 365pp.

Wolfe L.M. and D.R. Sowell, 2006. Do pollination syndromes partition the pollinator community? A test using four sympatric morning glory species. Int. J. Plant Sci., 167(6):1169-1175.

Woodward, A., 1988. Chemical composition of browse in relation to relative consumption of species and nitrogen metabolism of livestock in Southern Ethiopia. Ph. D. Thesis, p: 195. Cornell University, Ithaca, N.Y., USA

XLSTAT, 2008. XLSTAT. (http://xlstat.com/en/download) Date of access: 20 Aug. 2008.

Zaitoun, S., A.A. Al-Ghzawi, N. Samarah and R. Mullen, 2009. Pod production of Prosopis juliflora (Sw.) DC. As affected by supplementary and honeybee pollination under arid conditions. Acta Agriculturae Scandinavica, 59: 349-356.

Zdzislaw, W., J. Giejdasz, and G. Proszynski, 2004. Effect of pollination on onion seeds under isolation by the mason bee (Osmia rufa L.) (Apoidea, Megachilidae) on the setting and quality of obtained seeds. J. Apic. Sci., 48:35-41.

135