Almonds are one of the most economically valuable crops globally and require pollination by insects to optimise the production of high quality, marketable nuts. The Western honey bee (Apis mellifera L.) is an efficient pollinator of almond, and hives are often placed in almond orchards to provide pollination services.

While several studies have investigated pollen collection by honey bees, little is known about the usage of almond and other pollen sources by individual hives during almond bloom.

Protective covers are commonly employed in agricultural systems to reduce the impacts of extreme weather events, pest species and to control the environmental conditions in which crop plants are grown. As protected cropping systems are expanding rapidly, there is an urgent need to better understand how variations in netting practices might impact pollination service delivery by wild and managed insects to pollinator dependent crops. We used southern highbush blueberry (Vaccinium corymbosum L. interspecific hybrid) crops to investigate (i) how variations in protected cropping structures (fully netted, partially netted and unnetted blocks) influence the amount and composition of pollen deposited on crop stigmas; (ii) to what extent blueberry floral abundance and plant richness in remnant vegetation influence pollen composition on crop stigmas; and (iii) the difference between stigmatic pollen load composition in the middle and at the edge of crop blocks. We collected data from 15 field blocks of 6 different cultivars distributed on 10 farms. We collected blueberry stigmas to analyse the pollen load and measured blueberry floral abundance and richness of flowering plant taxa in remnant vegetation every two weeks. Our results indicate that blueberry pollen abundance on stigmas was reduced by up to 81% under full netting and 36% by partial netting. On blueberry stigmas, we identified a total of 31 morphospecies of non-blueberry pollen from 20 plant families. There was no relationship between blueberry stigmatic pollen loads and blueberry floral abundance. Moreover, the composition of non-blueberry pollen on stigmas differed between blueberry blocks under different netting categories. However, there was no relationship between plant taxa present in the surrounding remnant vegetation of each block and the pollen load on the stigmas of each block. Combining all netting treatments, stigmas located at the edge of the blocks received a greater amount of both conspecific (5% more) and heterospecific (40% more) pollen grains than those within the middle of blocks. Pollen flow in fields is reduced under netting structures as well as in the middle of blocks. Reduced blueberry pollen flow under nets may be detrimental to fruit yield and quality for some varieties of pollinator dependent crops, particularly those that are self-incompatible.

1. Most pollinator policy initiatives are focused on habitat restoration and increas-ing the availability of floral resources, yet the choice of plant species is not always compatible with farming system cultivation and management needs.

2. In this paper, we developed a framework for selecting plants to specifically meet stakeholder needs. We trialled 19 plant species and collected observational data on plant-insect visitors, plant survival in the orchard environment and potential risks to crops and the environment. We used this framework to identify plants suitable to incorporate into blueberry cropping systems.

3. Practical implication: Our framework ensured plant choice based on informed decisions and allowed the selection of two plant species that aligned well with industry needs. Different plants may be optimal for different conservation aims, hence plants selected need ideally to be evaluated for their use by the flower-visiting taxa, as well as align with industry growing practices and needs.

1. Despite the benefits of a diverse approach to crop pollination, global food production remains reliant on a low diversity of managed pollinators, especially the European honey bee (Apis mellifera). To facilitate more robust pollinator management and improve the resilience of the production system, it is necessary to understand regional variation in the pollination ecology of global food crops. Watermelon (Citrullus lanatus [Thunb.] Matsum & Nakai) is a highly insect pollinator-dependent crop and even though it is grown globally across many different climate zones, little is known about its pollination ecology across the diverse growing regions of Australia, spanning from the tropics to the arid zone.
2. We compared the species composition, visitation rates and effectiveness of the dominant floral visitors on 15 farms across five major watermelon-growing regions of Australia.
3. We found that insect species composition differed significantly among regions, but honey bees were the dominant watermelon flower visitor, with relative abundance varying from 73% to 94%. However, native bees (including stingless bees Tetragonula sp., and bees from families Megachilidae and Halictidae such as Lasioglossum, Homalictus and Lipotriches) and flies (particularly Syrphidae sp.) also visited and transferred pollen onto watermelon flowers.
4. In particular, native stingless bees were common visitors in several growing regions and deposited similar amounts of pollen to honey bees.
5. Our findings indicate that the Australian watermelon industry utilizes honey bees, but the diverse assemblage of available native pollinating taxa provides an additional opportunity for growers in specific growing regions. Pollination service delivery could be increased by deploying managed populations (e.g., native stingless bee colonies), employing pollinator-safe land management practices as well as exploring methods for increasing the efficiency of managed honey bee colonies.

Almond is one of the world's most economically valuable crops and many varieties require cross pollination for optimal fruit set. For this reason, western honey bee (Apis mellifera L.) hives are often placed in almond orchards. However, little is known about the usage of almond and other pollen sources by individual hives during almond bloom. Here, we investigated the timing, identity and quantity of pollen collection associated with almond floral abundance and spatial location of individual hives by sampling 440 individual pollen tray samples and counting 45,072 pollen grains from 13,200 pollen pellets collected from 80 individual hives across the flowering season in Victoria, south-eastern Australia. A large proportion of hives collected non-almond pollen in addition to almond pollen (63/80 = 79%). The weight of almond pollen collected by the hives at each sampling time was positively related to the number of concurrently open almond flowers. However, non-almond pollen richness and abundance was not related to the number of almond flowers but had a positive relationship with the weight of almond pollen collected. There was no relationship between the distance among hives and identity of pollen collected. Yet, three plant families in the study area were found to account for a high percentage of the non-almond pollen collected, Euphorbiaceae, Fabaceae and Asteraceae. Understanding crop and non-crop pollen collection could inform honey bee diet needs and identify the plant species of importance to inform best practice bee management during almond flowering.

Hybrid crop production is more reliant on pollinators compared to open-pollinated crops because they require cross-pollination between a male-fertile and a male-sterile line. Little is known about how stigma receipt of pollen from male-sterile genotypes affects reproduction in hybrids. Non-viable and non-compatible pollen cannot fertilise plant ovules, but may still interfere with pollination success. Here we used seedless watermelon (Citrullus lanatus (Thunb.) Matsum. & Nakai) as a model hybrid plant, to evaluate the morphology, physiology, and movement of pollen from inter-planted genotypes (diploids and triploids). We found that pollen from triploids ('Exclamation' and 'Royal Armada') and diploids ('SP-6', 'Summer Flavor 800', and 'Tiger') was visually distinguishable. Pollen in triploids had more deformities (42.4–46%), tetrads (43–44%), and abnormal growth of callose plugs in pollen tubes. The amount of pollen in triploids to germinate on stigmas was low (8 ± 3%), and few pollen grains produced pollen tubes (6.5 ± 2%). Still, contrary to previous reports our results suggest that some viable pollen grains are produced by triploid watermelons. However, whilst honey bees can collect and deposit pollen from triploids onto stigmas, its effect on hybrid watermelon reproduction is likely to be minimal due to its low germination rate.

Managed and wild pollinator taxa are responsible for greater than one third of global food production and their services benefit more than 85% of wild plants. As the production of pollinator-dependent crops expands globally, an understanding of the mechanisms underlying pollinator behaviour and pollination service delivery is becoming increasingly important. This thesis investigates the efficiency and behaviour of insect pollinators that visit global food crops with a focus on watermelon (Citrullus lanatus (Thunb.) Matsum. & Nakai" family: Cucurbitaceae) as a model crop species.

First, I reviewed the literature concerning watermelon pollination biology to ascertain existing knowledge gaps. Second, I collated and synthesized the literature on watermelon floral visitors and their efficiency across different growing regions globally through a systematic literature and meta-analysis. Third, I used field surveys and pollen deposition trials, to explore floral visitor composition and effectiveness of watermelon insect visitors across 15 commercial farms in five geographically distinct growing regions" Riverina (NSW), Katherine (NT), Gumlu (QLD), Lakeland (QLD) and Chinchilla (QLD) in Australia. Fourth, I investigated the pollen morphological traits across diploid and triploid watermelon genotypes and compared the fate and outcome of pollen grains in inter-planted genotypes (hand pollinated crosses" triploid x triploid and triploid x diploid) in a hybrid cropping system. Finally, I examined how insect flower visit duration relates to foraging behaviour, pollination success and the level of crop pollinator dependency using field data collected from 11 insect-pollinated crop species, including watermelon, across 56 farms in Australia, New Zealand, and USA.

I found that knowledge gaps exist concerning the identity of pollinator assemblages across cultivars, genotypes, and growing regions. In particular, few published studies have focused on seedless cultivars due to their relatively recent cultivation. The results of this review indicated that different watermelon genotypes require specific management to ensure optimal production. In the global meta-analysis of watermelon floral visitors, I recorded over 265 watermelon floral visitors (including 5 orders, 18 families and 75 genera) from 54 records across 17 countries. Honey bees (Apis mellifera L.) and other bees were equally effective at depositing pollen on stigmas, but varied in their total contribution to fruit and seed set. In Australian field surveys of watermelon flower visitors, I found that the insect species community composition differed significantly among regions, but honey bees were the dominant watermelon flower visitor at all sites, with relative abundance varying from 73 - 94%. However, native bees (Tetragonula spp., and bees from the Families Megachilidae, and Halictidae including the genera Lasioglossum, Homalictus, Lipotriches), and flies (Syrphidae spp.) also visited and transferred pollen onto watermelon stigmas. In particular, native stingless bees deposited similar amounts of pollen as honey bees in some regions. In my investigation of pollen morphological traits across diploid and triploid watermelon genotypes, I found that pollen from triploids and diploids was visually distinguishable, as pollen in triploids had more deformities (44%), tetrads (43.5%), and abnormal growth of callose plugs in their pollen tubes. Contrary to previous reports, I found that triploid watermelons produced low numbers of viable pollen grains that geminated (8±3%) and produced pollen tubes (6.5±2%). Whilst honey bees can collect and deposit pollen from triploids onto stigmas, the impact of this behaviour on hybrid watermelon reproduction is likely minimal due to the low germination rate of pollen from triploids. Finally, I found that flower visit duration varies across insect taxa and crop species and was generally not associated with either the number of pollen grains deposited on the stigma or fruit set. Flower visit duration was only marginally positively related to pollen tube growth. However, there was an interaction between taxa and visit duration, indicating that the outcome upon plant reproductive success as a result of visit duration, varied according to different floral visitor groups. I also found that floral visitors spent more time on flowers when foraging for nectar compared to pollen and their visit duration was greater with increasing pollinator dependency of the crop species.

This thesis demonstrates the importance of understanding local, regional and global variation in pollinator assemblages across crop genotypes. Identifying wild pollinator taxa, deploying managed populations of native taxa (e.g. native stingless bee colonies in Australia), employing pollinator-safe land management practices, and exploring methods for increasing the efficiency of managed honey bee colonies, are all measures likely to contribute to improving watermelon production in Australia and globally. Greater understanding of pollinator behaviour, efficiency and plant-pollinator interactions is necessary to ensure ongoing effective and resilient pollination services to watermelon cropping systems, and to develop best practice pollination management for pollinator dependent crops to improve global food security.

For most food crops the identity and efficiency of pollinators across key growing regions remains a significant knowledge gap that needs to be addressed before we can develop crop-specific approaches for pollination service delivery. Here, we conducted a systematic literature review and meta-analysis on watermelon (Citrullus lanatus (Thunb. Matsum. & Nakai)), a globally important fruit crop, to identify the floral visitors and their efficiency across different growing regions. We found that 265 insect species visit watermelon flowers (including 5 orders, 18 families and 75 genera) across 17 countries and 6 continents. Bees and flies were the most abundant flower visitors overall, but show distinct regional differences. Honey bees were the majority visitor in 53% of growing regions (range: 0 - 94%), whilst wild bee species were more abundant in 42% of regions (range: 3.4 - 100%). Honey bees and other bees were equally effective at depositing pollen on stigmas, but varied in effectiveness for fruit set and seed set. Pollination data from global studies appear to be limited for the largest-scale watermelon producers, namely: China, Turkey, and India, with the majority (56%) of data available from North America. This synthesis identified four key themes for improving pollination in watermelon: increasing honey bee densities on crops where local polices and environmental conditions are suitable; introducing other managed pollinators; identifying key wild pollinator taxa to encourage within crops; and improving local and landscape management practices to support pollinators.

Pollinators face multiple pressures and there is evidence of populations in decline. As demand for insect-pollinated crops increases, crop production is threatened by shortfalls in pollination services. Understanding the extent of current yield deficits due to pollination and identifying opportunities to protect or improve crop yield and quality through pollination management is therefore of international importance. To explore the extent of “pollination deficits,” where maximum yield is not being achieved due to insufficient pollination, we used an extensive dataset on a globally important crop, apples. We quantified how these deficits vary between orchards and countries and we compared “pollinator dependence” across different apple varieties. We found evidence of pollination deficits and, in some cases, risks of overpollination were even apparent for which fruit quality could be reduced by too much pollination. In almost all regions studied we found some orchards performing significantly better than others in terms of avoiding a pollination deficit and crop yield shortfalls due to suboptimal pollination. This represents an opportunity to improve production through better pollinator and crop management. Our findings also demonstrated that pollinator dependence varies considerably between apple varieties in terms of fruit number and fruit quality. We propose that assessments of pollination service and deficits in crops can be used to quantify supply and demand for pollinators and help to target local management to address deficits although crop variety has a strong influence on the role of pollinators.

The dataset consists of six tabs (each referring to a data chapter of the thesis). Data related to this dataset was collected online from Scopus research database (Chapter 2), and in the field from Lake Powell, Victoria, Australia (Chapter 5), and from the East-North Coast of New South Wales, Australia (chapters 3, 4, 6, and 7). Each tab has a spreadsheet with data from each thesis research chapter, and the content of each tab is explained below:

Chapter 2: A review of honey bee (Apis mellifera L.) interactions with other pollinators. The dataset consists of information extracted from reviewed articles, such as the title of the article, the authors, the year of publication, the country in which the research was performed, the place of study (i.e., open fields, or enclosed areas like glasshouses, cages, etc.), whether the study aimed to test competition or not, whether the study found and/or discussed competition or not, the type of interaction (i.e., direct or indirect) between honey bees and other species, the main methods used to perform the research, the interacting species with honey bees, and the behaviour performed by honey bees during the interaction.

Chapter 3: Observation of birds foraging on raspberry orchards. The dataset consists of the number of instances insectivorous or nectarivorous birds were seen perched on cages containing brown blowflies, within a raspberry orchard block. The first column is the days since flies were released inside cages until birds were seen inside the same cages, column two is the number of observational hours on each day of data collection, the following four columns have the bird species name and the number of individuals of each species, and the last column is the total number of individuals.

Chapter 4: Effectiveness of brown honeyeater (Lichmera indistincta) in pollinating blueberry flowers compared to insect pollinators. The dataset consists of stigmas collected after a single visit by one pollinator species. The first column is the pollinator ID, the second column is the identification of the stigma sampled, and the last column is the number of conspecific (i.e., blueberry) pollen grains deposited.

Chapter 5: Pollen collection by honey bee hives in almond orchards. The data consists of pollen pellets removed from different hives placed in almond orchards for pollination. The first column is the almond flower abundance during the flowering season (measures were always taken from the same tree branch throughout the study period), column two is the day of data collection (11 days in total), column three is the apiary identification, column four is the hive identification, column five is the heterospecific pollen richness (number of species), column six is the abundance of heterospecific pollen, column seven is the weight of heterospecific pollen collected from each hive, column eight is the weight of almond pollen collected from each hive, column nine is the total weight.

Chapter 6: Impacts of protective nets on pollen flow in blueberry orchards. The dataset consists of analyses of pollen deposition on blueberry stigmas under different netting treatments. The first column is the blueberry plant variety stigma was sampled from, column two the orchard block identification, column three is the netting treatment (i.e., covered - block completely covered by nets, partially - block partially covered by nets with sides open, and open - no nets), column four is the week of data collection (seven weeks in total), column five is the abundance of blueberry flower within each block (measures were always taken from the same blueberry plant), column six is the conspecific pollen abundance, column seven is the heterospecific pollen abundance on the stigmas, column eight is the heterospecific pollen richness in the stigmas, column nine is the point of data collection within the orchards (i.e., edge or center), and the last column is the number of non-blueberry flowers present in the remnant vegetation surrounding the orchard blocks.

Chapter 7: Effects of multiple visits on pollen deposition in blueberry, blackberry, and raspberry flowers. The dataset consists of pollen deposition analyses after multiple visits by different insect taxa in blueberry, blackberry, and raspberry flowers. The first column is the sample identification, column two is the crop species (i.e., Rubus sp - raspberry, rabbiteye - blueberry, or blackberry), column three is the crop species variety, column four is the block identification, column five is the date of sample collection, column six is the row number within the block that the sample was collected from, column seven is the identification tag used for the sample during collection, column eight is the total number if visits received by the sample (i.e., stigma), column nine is the combination of species that visited the stigma (i.e., MX - mix of species, SB - stingless bees, HB - honey bees, CR - carpenter bees), column ten is the number of conspecific pollen counted on the stigma, column eleven is the number of heterospecific pollen counted on the stigma, column twelve is the richness of heterospecific pollen counted on the stigma, column thirteen is the total visit duration time in seconds, column fourteen to thirty four is the visit duration of each subsequent visit.