Urban Ecosystems

If you it, they will come: quantifying attractiveness of crop for winter-active flower visitors in community gardens --Manuscript Draft--

Manuscript Number: UECO-D-19-00111 Full Title: If you plant it, they will come: quantifying attractiveness of crop plants for winter-active flower visitors in community gardens Article Type: Manuscript Keywords: winter pollination; urban conservation; visitor network; urban garden; urban ecology; pollinators; Syrphidae; Hymenoptera Corresponding Author: Tanya Latty University of Sydney Eveleigh, NSW AUSTRALIA Corresponding Author Secondary Information: Corresponding Author's Institution: University of Sydney Corresponding Author's Secondary Institution: First Author: Perrin Tasker First Author Secondary Information: Order of Authors: Perrin Tasker Chris Reid Andrew D. Young Caragh G Threlfall Tanya Latty Order of Authors Secondary Information: Funding Information: Abstract: Urban community gardens are potentially important sites for urban pollinator conservation because of their high density, diversity of flowering plants, and low pesticide use (relative to agricultural spaces). Selective planting of attractive crop plants is a simple and cost-effective strategy for attracting flower visitors to urban green spaces, however, there is limited empirical data about which plants are most attractive. Here, we identified key plant that were important for supporting flower visitors using a network-based approach that combined metrics of flower visitor abundance and diversity on different crop species. We included a metric of ‘popularity’ which assessed how frequently a particular crop plant appeared within community garden. We also determined the impact of garden characteristics such as size, flower species richness, and flower species density on the abundance and diversity of flower visitors. We found two plant species, Brassica rapa and Ocimum basilicum were identified as being key species for supporting flower visitor populations on all three of our metrics. Flower species richness had a strong positive effect on both the abundance and diversity of flower visitors. We suggest that gardeners can maximise the conservation value of their gardens by planting a wide variety of flowering plants including highly attractive plants such as B. rapa and O. basilicum. Suggested Reviewers: Katherine Baldock University of Bristol [email protected] Francis Ratnieks University of Sussex [email protected]

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation Mikhail Garbuzov University of Sussex [email protected] Ania Majewska University of Georgia [email protected] Ken Thomson University of Sussex [email protected]

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation Manuscript Click here to download Manuscript Tasker et al for UE 25 April.docx Click here to view linked References

1 If you plant it, they will come: quantifying attractiveness of crop plants for winter-active 2 3 flower visitors in community gardens 4 5 6 1 2 4, 5 1, 3 1 7 Perrin Tasker , Chris Reid Andrew D. Young , Caragh G Threlfall Tanya Latty 8 9 1. School of Life and Environmental Sciences, University of Sydney, Sydney, NSW 2006, 10 11 Australia 12 13 1. Department of Biological Sciences, Macquarie University, Sydney, NSW 2109, 14 Australia 15 2. School of Ecosystem & Forest Sciences, The University of Melbourne, Parkville, VIC 16 17 3010 18 3. Plant Pest Diagnostics Center, California Department of Food and Agriculture, 3294 19 Meadowview Road, Sacramento, CA 95832, USA 20 21 4. Department of Entomology and Nematology, UC Davis, Briggs Hall, Davis, CA 95616- 22 5270, USA 23 24 25 26 Corresponding author: Tanya Latty: [email protected] 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 1 63 64 65 Abstract 1 Urban community gardens are potentially important sites for urban pollinator conservation 2 3 4 because of their high density, diversity of flowering plants, and low pesticide use (relative to 5 6 7 agricultural spaces). Selective planting of attractive crop plants is a simple and cost-effective 8 9 strategy for attracting flower visitors to urban green spaces, however, there is limited 10 11 12 empirical data about which plants are most attractive. Here, we identified key plant species 13 14 that were important for supporting flower visitors using a network-based approach that 15 16 17 combined metrics of flower visitor abundance and diversity on different crop species. We 18 19 included a metric of ‘popularity’ which assessed how frequently a particular crop plant 20 21 22 appeared within community garden. We also determined the impact of garden 23 24 25 characteristics such as size, flower species richness, and flower species density on the 26 27 abundance and diversity of flower visitors. We found two plant species, Brassica rapa and 28 29 30 Ocimum basilicum were identified as being key species for supporting flower visitor 31 32 populations on all three of our metrics. Flower species richness had a strong positive effect 33 34 35 on both the abundance and diversity of flower visitors. We suggest that gardeners can 36 37 38 maximise the conservation value of their gardens by planting a wide variety of flowering 39 40 plants including highly attractive plants such as B. rapa and O. basilicum. 41 42 43 44 45 46 47 Keywords: winter pollination, urban conservation, visitor network, urban garden, urban 48 49 ecology, pollinators, Syrphidae, Hymenoptera 50 51 52 53 54 55 56 57 58 59 60 61 62 2 63 64 65 Introduction 1 Urbanisation is one of the major drivers of environmental change and has both direct 2 3 4 (habitat loss, land conversion) and indirect (heat island effect, changes in resource availability) 5 6 7 impacts on biodiversity. Given the pace of land conversion, developing strategies to conserve 8 9 biodiversity within cities is imperative. While conserving tracts of high-quality habitat (‘land 10 11 12 sparing’) is likely beneficial, it is largely not possible in the inner city where land changes are 13 14 largely irreversible. Enacting strategies that support target organisms within existing urban 15 16 17 land uses (‘land sharing’) by providing key resources may lead to better outcomes for 18 19 biodiversity in highly urbanised environments (Soga et al. 2014). 20 21 22 23 The increasing recognition of the conservation value of gardens within cities 24 25 (Baldock et al. 2019) coincides with interest in strategies for designing ‘pollinator friendly 26 27 28 gardens’ more broadly (Majewska and Altizer 2019). A large metanalysis of 178 studies found 29 30 31 that within-garden features such as flower diversity had a stronger impact on pollinator 32 33 populations than did landscape level features such as degree of urbanisation or distance to 34 35 36 agricultural fields(Majewska and Altizer 2019). (Kevin C Matteson & Langellotto, 2011) found 37 38 that the best predictors of bee and butterfly richness in US community gardens were floral 39 40 41 area and sunlight availability, while a study of domestic gardens in the UK identified native 42 43 plant species richness, the number of surrounding houses, and the extent of low canopy 44 45 46 vegetation as key drivers of solitary bee diversity (Smith, Warren, Thompson, & Gaston, 47 48 49 2006). In Australia, increases in the proportion of flowering native plants in a green space had 50 51 a positive impact on the density of non-native honey bees (Apis mellifera), while native, 52 53 54 ground dwelling Homalictus bees were most abundant in areas with low flower diversity and 55 56 small amounts of surrounding impervious surfaces (Threlfall et al. 2015). Taken together, 57 58 59 60 61 62 3 63 64 65 these studies suggest that flower communities are key to creating gardens that support strong 1 2 3 populations of flower-visiting . 4 5 6 Gardens are essentially constructed ecosystems, where the plant community is 7 8 explicitly selected and arranged by gardeners. This allows tremendous potential for selecting 9 10 11 plants that optimally support pollinators. Indeed, several studies of bees in urban gardens 12 13 14 have suggested (but not tested) that bee diversity may be largely attributable to the presence 15 16 or absence of particularly attractive flower species (Gunnarsson and Federsel 2014; Makinson 17 18 19 et al. 2017). 20 21 22 While there are many recommendations for choosing pollinator-friendly crop species, 23 24 25 the evidence base for flower selection is primarily anecdotal (Ratnieks and Garbuzov 2014). 26 27 Lists of pollinator friendly species available to the public often include poor 28 29 30 recommendations, a very large number of species, and are generally not based on empirical 31 32 data (Ratnieks and Garbuzov 2014; but see Garbuzov et al. 2017; Garbuzov and Ratnieks 2013; 33 34 35 Garbuzov et al. 2015 for exceptions). A notable exception is the experiment of (Garbuzov 36 37 38 and Ratnieks 2013) which investigated the relative attractiveness of 32 different plant species 39 40 in a specially planted pollinator garden. They found clear differences in plant attractiveness 41 42 43 and suggest that plant selection could be a powerful and cost effective strategy for making 44 45 gardens more pollinator-friendly. 46 47 48 49 Our primary aim in the present study was to identify pollinator-friendly plant species 50 51 that have the potential to attract pollinators to urban community gardens. We were also 52 53 54 interested in local-scale drivers of flower visitor abundance and diversity such as garden size, 55 56 flower density and flower abundance. Community gardens are becoming increasingly popular 57 58 59 in major cities around the world as people become more interested in local food production 60 61 62 4 63 64 65 and the social and health benefits of gardening (Guitart et al. 2012). Community gardens 1 2 3 typically use little to no pesticides (Pawelek et al. 2009) and have a wide variety of flowering 4 5 plants, both of which may benefit pollinating insects. Many of the plants commonly grown in 6 7 8 community gardens are pollinated (Baldock et al. 2015; Hennig and Ghazoul 2011; 9 10 Lawson 2005; Makinson et al. 2017) so gardeners should be highly motivated to conserve the 11 12 13 insects upon which crop productivity relies. Selective planting of highly attractive crop species 14 15 could be an easy way for gardeners to increase the conservation and food production value 16 17 18 of their gardens. 19 20 21 Previous attempts to quantify the attractiveness of plants in gardens have typically 22 23 24 used abundance of flower visitors as their primary metric (eg. Garbuzov and Ratnieks 2013). 25 26 27 Here, we developed a protocol for selecting ‘key’ crop species by combining measures of 28 29 flower visitor diversity and abundance with a popularity score, which characterised how often 30 31 32 a crop was found in community gardens. Plant popularity is an important consideration, as 33 34 urban planting strategies have the best chance of succeeding when there is substantial buy in 35 36 37 from stakeholders (Turo and Gardiner 2019). 38 39 40 41 42 43 44 Methods 45 46 47 Study sites 48 49 We sampled 20 community gardens across inner city Sydney, NSW, Australia including the 50 51 local government areas of Strathfield, Ku-ring-gai, Sydney City, Leichhardt, Marrickville and 52 53 Balmain. Sydney is a highly urbanised metropolitan area (population ~ 4 million) located in 54 55 coastal south eastern Australia. Prior to European colonisation in 1788, our study area’s 56 57 vegetation would have been primarily a mix of sandstone woodland and heath and blue gum 58 59 60 61 62 5 63 64 65 forest (Benson and Howell 1994), although it is now highly fragmented with a few small 1 2 pockets of native bushland remaining in the inner city. 3 4 5 Sydney enjoys a humid subtropical climate with warm to hot summers and mild to cool 6 winters (Köppen climate classification). Sampling was conducted from June through to 7 8 August, 2017 during the coldest part of Australian winter with temperature ranging between 9 10 13 C and 24 C, with an average mean temperature of 17 C during our sampling period. 11 12 13 We sampled during winter because winter pollination has been largely overlooked in tropical 14 15 and subtropical areas, where insects and flowers are active all year. Winter-active flower 16 17 visitors such as , wasps and the European honey bee (Apis mellifera) are important for 18 19 the many agricultural crops which are grown over the cooler winter months, including 20 21 brassicas (when grown for seed production), broad beans, and peas (Dabney et al. 2001). 22 23 Gardens used in this study were variable in size ranging from our smallest garden (Rowley 24 25 street garden) which measured 45 m2 to our largest garden (White’s Creek garden) at 1373 26 27 M2. Mean garden size was 482m 2± 90.8. 28 29 30 We obtained the permission of local councils and/or community garden prior to the use of 31 32 33 their gardens. 34 35 36 Insect and flower sampling 37 38 We sampled gardens between 1000 and 1600 hrs on days when it was not raining or 39 40 41 excessively windy. The weather conditions and temperature were recorded during each 42 43 sampling period. Each garden was sampled twice: once early in the season (June/mid-July) 44 45 46 and again in late winter (late-July to late-August). 47 48 49 We first surveyed flowering plants by identifying, recording and numbering all flowering 50 51 52 plants under 2 m tall within each garden. We considered same-species groups growing within 53 54 15 cm of one another to be a single ‘patch’. Patches that did not receive full-sun were avoided 55 56 57 if possible, based on evidence suggesting that sun-availability influences the abundance of 58 59 60 61 62 6 63 64 65 floral visitors (K.C. Matteson & Langellotto, 2010). For species represented by two or more 1 2 3 patches, we randomly selected one patch for insect sampling. 4 5 6 Each flowering patch was observed for 20 minutes during which observers walked around the 7 8 focal patch photographing all flower visitors from multiple angles using a Canon 1000D DSLR 9 10 11 camera. We used photography instead of hand netting because our pilot trials found that the 12 13 14 act of catching insects disturbed flower visitors (particularly flies) which did not always return 15 16 within the sampling period. A non-lethal sampling approach also made it easier to obtain 17 18 19 permission from community garden committees and local councils, who were generally 20 21 opposed to lethal sampling methods such as pan trapping. 22 23 24 25 Insects were occasionally missed if they flew away before a photograph could be taken. If 26 27 possible, escaping insects were identified by eye in the field. Based on estimations and field 28 29 30 records, we successfully photographed approximately 90% of all floral visitors. 31 32 33 All insects were identified to the lowest possible taxonomic level. Hover flies of the 34 35 36 Melangyna were initially identified as belonging to five species (Melangyna damastor, 37 38 Melangyna collatus, Melangyna viridiceps, Melangyna. sp.1, Melangyna. sp.2) however, 39 40 41 recent molecular evidence suggests that these morphological characteristics cannot be used 42 43 to differentiate between Melangyna species, and that these ‘species’ likely represent either 44 45 46 a single polymorphic species, or a complex of closely related species that are not currently 47 48 49 distinguishable morphologically (A. D. Young, unpublished COI data). All five Melangyna 50 51 ‘species’ were therefore grouped into one morphospecies, ‘Melangyna sp.’ 52 53 54 55 All flowering plants were identified to the lowest possible taxonomic level. 56 57 58 59 60 61 62 7 63 64 65 Data analysis 1 2 3 Identifying key plants 4 5 We used floral visitor-plant networks to identify plants that played a key role in supporting 6 7 8 flower visitors. We visualised visitor-plant systems as bipartite networks, where plants and 9 10 11 insects are nodes and links represent interactions (visits by insects to plants). We used the 12 13 ‘bipartite’ package in R to visualise and analyse networks (Dormann et al. 2009). We 14 15 2 16 calculated three metrics: normalised degree, visitation density (visitors per m ) and the total 17 18 number of gardens the focal plant was observed in. Normalised degree is a network- derived 19 20 21 metric that calculates the total proportion of available partners that a focal plant interacted 22 23 with. We calculated normalised degree using the ‘specieslevel’ function in the ‘bipartite’ 24 25 26 package in R. A normalised degree of ‘1’ would indicate a plant or visitor that interacted with 27 28 29 all potential partners in the garden. The visitation density was the number of pollinators 30 31 observed per m2 of plant and was selected to control for differences in plant patch size. Last, 32 33 34 we noted how many gardens a plant was observed in, as an indication of the plant’s popularity 35 36 with gardeners. We selected the top ten plant species in each category and present them in 37 38 39 Table 1. We consider key plants to be those that scored within the top 10 on all three key 40 41 42 metrics. 43 44 45 Identifying drivers of insect abundance and diversity 46 47 48 We examined the effect of flower species richness (number of species), the number of flower 49 50 2 51 species per m (flower species density) and garden size on floral visitor abundance (total 52 53 number of individuals) and diversity (species richness) using a generalised linear mixed model 54 55 56 with ‘garden’ included as a random effect. Variance inflation factors (VIFs) were checked to 57 58 59 ensure that the assumption of independence was met. VIFs greater than 10 suggest strong 60 61 62 8 63 64 65 collinearity (Quinn and Keough 2002). Consequently, we conservatively rejected variables 1 2 3 from our model if they had a VIF greater than 4 (Latty and Beekman 2009; Quinn and Keough 4 5 2002). We visually examined residual plots to ensure that the assumption of homogeneity of 6 7 8 variance was met (Latty and Beekman 2009; Quinn and Keough 2002), and we tested for 9 10 normally distributed residuals using a Shapiro-Wilk goodness of fit test. 11 12 13 14 Results 15 16 17 Survey results 18 19 20 We photographed 2242 incidents of floral visitation by 23 insect species in community 21 22 gardens over winter (Appendices 1 and 2). The floral visitors included a wide range of insect 23 24 25 families, with the most diverse and abundant family being the hover flies (Syrphidae; Fig. 1, 26 27 28 Appendix 1). At the species level, the honey bee Apis mellifera and the hover Melangyna 29 30 sp. were by far the most abundant floral visitors across the study (Fig. 2). Other floral visitors 31 32 33 included native bees, hover flies, other flies, butterflies, moths, true bugs, wasps and sawflies 34 35 (Appendix 1). 36 37 38 39 Community gardens supported a diverse array of exotic flowering plants (Appendix 2). 40 41 Gardens contained a mean 11.7 (±0.89) species of flowering plants, with a minimum of 4 42 43 44 and a maximum of 29. All plant species sampled were non-native to the Sydney Region 45 46 (Robinson 1991). We observed 71 plant species across 20 plant families. The most common 47 48 49 plant family was (daisy-like) followed by (mint) and Brassicaceae 50 51 52 (brassicas) (Appendix 2). The three most popular plant species were elegans, Ocimum 53 54 basilicum and bipinnatus 55 56 57 Key Plant species 58 59 60 61 62 9 63 64 65 Brassica rapa and Ocimum basilicum appeared in all 3 top-10 lists suggesting they are 1 2 3 attractive to a variety of floral visitors and popular amongst gardeners (Table 2). Five other 4 5 plant species featured in two top-10 lists and 10 species featured in a single top-10 list. 6 7 8 Factors influencing the species richness and abundance of flower visitors 9 10 11 Flower species richness had a significant positive effect on flower visitor abundance and 12 13 species richness (Fig. 3, Table 1). None of the other factors had a significant effect. 14 15 16 17 18 Discussion 19 20 21 We identified two key plant species which ranked in the top ten on all three of our 22 23 24 metrics: Ocimum basillicum and Brassica rapa. Ocimum basillicum (sweet basil) is a popular 25 26 herb grown for its aromatic . It is known to be highly attractive for bees and has been 27 28 29 used to attract pollinators to bell pepper crops resulting in increased pepper yield (Pereira et 30 31 al. 2015). Brassica rapa (field mustard) is commonly grown as a root vegetable (turnip), for 32 33 34 its leaves (eg. Mizuna), and for its seed oil. A study from Saudi Arabia found that Brassica rapa 35 36 flowers were attractive to a wide variety of bee and fly species (Shakeel et al. 2018). Both O. 37 38 39 basillicum and B. rapa produce nectar and pollen, although data about the nutritional content 40 41 42 of nectar and pollen in these species is scarce. 43 44 45 Brassicas in general were well represented in the top ten lists with three species 46 47 48 (Brassica rapa, Brassica oleracea and Brassica juncea) appearing at least once. While the 49 50 presence of brassicas was not particularly surprising given their popularity as a winter crop, 51 52 53 the fact that so many were in flower was unusual because Brassicas are usually harvested 54 55 before flowering. The abundance of flowering Brassicas may be indicative of a lower 56 57 58 investment in maintenance on the part of the gardeners. Rather than being a detriment, 59 60 61 62 10 63 64 65 however, our results support the idea that ‘messy’ gardens can often provide a greater 1 2 3 benefit to biodiversity than tidy, highly-maintained gardens. We suggest that allowing at least 4 5 some brassicas to go to flower is an effective strategy for supporting flower visitors. 6 7 8 Thirty percent of the top-10 most popular flowering crops also scored highly on at 9 10 11 least one of our two metrics of insect visitation/abundance. This is promising, as it is suggests 12 13 14 that many attractive species are also prized by gardeners. It is likely easier to convince 15 16 gardeners to plant more of a species which they already enjoy. 17 18 19 Interestingly, plants commonly described as ‘bee-friendly’ like Salvia elegans, Borago, 20 21 22 and Lavandula did not score highly on our insect-pollinator metrics but were planted in many 23 24 25 gardens. It is possible that these plants are more attractive to summer-active insects such as 26 27 solitary bees than to the fly-dominated winter assemblage. Our research highlights the need 28 29 30 for season-specific planting recommendations. 31 32 33 Gardens that had high flower species diversity also supported a high abundance and 34 35 36 diversity of flower visitors. Our results are in agreement with previous studies which have 37 38 largely identified local-scale flower diversity/abundance as key factors determining flower 39 40 41 visitor abundance and diversity in urban environments (Majewska and Altizer 2019; Makinson 42 43 et al. 2017; Matteson and Langellotto 2010; Pardee and Philpott 2014; Threlfall et al. 2015). 44 45 46 Previous work in Sydney community gardens found that landscape scale variables such as 47 48 49 amount of surrounding green space and distance to a forest fragment did not impact the 50 51 abundance and diversity of flower visitors during the summer (Makinson et al. 2017). Taken 52 53 54 together with previous studies, our work strongly supports the simple idea that urban 55 56 gardeners wishing to support flower visitors in their gardens can do so simply by planting a 57 58 59 greater variety of flower species. 60 61 62 11 63 64 65 Our results highlight the importance of flies as floral visitors in winter; of the 23 insect 1 2 3 species we observed, 15 were flies with the majority (10 species) belonging to the hover flies 4 5 (Syrphidae). This is unsurprising, as flies are commonly the dominant pollinators in colder 6 7 8 climates, such as Arctic (Elberling and Olesen 1999; Kevan 1972) and alpine regions (Lefebvre 9 10 et al. 2014). The role of flies as pollinators has historically received far less attention than that 11 12 13 of bees, despite the fact that flies are known to visit at least 555 species and 14 15 are pollinators of at least 100 cultivated plant species including cacao, onions and mango 16 17 18 (Larson, Inouye and Kevan, unpubl., as cited in (Ssymank et al. 2008). 19 20 21 Hover flies in particular should be encouraged in gardens as the larvae of species 22 23 24 within the subfamily , including Melangyna, Betasyrphus, Eupeodes, and other 25 26 27 common urban genera, are predators of . They can therefore deliver a double benefit 28 29 to gardens by providing both pollination and pest control services. Three of the syrphid floral 30 31 32 visitors observed in our study - Melangyna sp., Betasyrphus serarius and Eupeodes confrater 33 34 - have been identified as biological control agents of populations in Australian and 35 36 37 South-East Asian crops (Bowie 1999; Irshad 2014; Joshi and Ballal 2013). Further research 38 39 quantifying the ecosystem services provided by Syrphidae in particular and flies in general 40 41 42 could have widespread impacts on urban agriculture. 43 44 45 Our results suggest that urban-adapted native species of flower visitor are supported 46 47 48 by the exotic crop plant species found in Sydney Community Gardens. These plants might be 49 50 51 important resources for winter-active insects, providing nectar and pollen resources at a 52 53 time of year when the majority of native Australian plant species are not actively flowering 54 55 56 (Cunningham et al. 2002; Heard and Hendrikz 1993). Research in urban areas in Europe has 57 58 found that the presence of exotic winter-flowering plant species enables bumblebees and 59 60 61 62 12 63 64 65 to effectively forage over winter (Salisbury et al. 2015; Senapathi et al. 2017; Stelzer 1 2 3 et al. 2010). 4 5 6 Although we did not compare species richness in our community gardens to other 7 8 urban green spaces, we suggest that community gardens have the potential to be hotspots 9 10 11 for urban insect conservation because of high, year-round flower abundance and diversity, 12 13 14 low insecticide use and highly motivated stakeholders whose interests naturally align with 15 16 those of pollinating insects. Plant assemblages within community gardens can be easily 17 18 19 manipulated - compared to other urban green spaces like parks - allowing gardeners the 20 21 freedom to plant with insect conservation in mind. Here we have developed a technique for 22 23 24 identifying crop plants that support a variety of insect flower visitors. Future research is 25 26 27 needed to determine the extent to which pollinator-friendly plantings can impact the health, 28 29 reproduction and population growth of flower-visiting insects. 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 13 63 64 65 References 1 2 3 Baldock KC et al. (2019) A systems approach reveals urban pollinator hotspots and 4 5 conservation opportunities Nature ecology & evolution:1 6 7 Baldock KCR et al. (2015) Where is the UK's pollinator biodiversity? The importance of urban 8 9 areas for flower-visiting insects Proceedings of the Royal Society B: Biological 10 Sciences 282 doi:10.1098/rspb.2014.2849 11 12 Benson D, Howell J (1994) The natural vegetation of the Sydney 1: 100 000 map sheet. Royal 13 14 Botanic Gardens Sydney 15 16 Bowie MH (1999) Effects of distance from field edge on aphidophagous insects in a wheat 17 18 crop and observations on trap design and placement International Journal of Pest 19 Management 45:69-73 doi:10.1080/096708799228076 20 21 Cunningham SA, FitzGibbon F, Heard TA (2002) The future of pollinators for Australian 22 23 agriculture Australian Journal of Agricultural Research 53:893-900 24 doi:https://doi.org/10.1071/AR01186 25 26 27 Dabney SM, Delgado JA, Reeves DW (2001) Using winter cover crops to improve soil and 28 water quality Communications in Soil Science and Plant Analysis 32:1221-1250 29 doi:10.1081/css-100104110 30 31 32 Dormann CF, Fründ J, Blüthgen N, Gruber B (2009) Indices, graphs and null models: 33 analyzing bipartite ecological networks The Open Ecology Journal 2 34 35 36 Elberling H, Olesen JM (1999) The structure of a high latitude plant‐flower visitor system: 37 the dominance of flies Ecography 22:314-323 38 39 40 Garbuzov M, Alton K, Ratnieks FLW (2017) Most ornamental plants on sale in garden centres 41 are unattractive to flower-visiting insects PeerJ 5:e3066 doi:10.7717/peerj.3066 42 43 44 Garbuzov M, Ratnieks FL (2013) Quantifying variation among garden plants in attractiveness 45 to bees and other flower‐visiting insects Functional Ecology 46 47 Garbuzov M, Samuelson EEW, Ratnieks FLW (2015) Survey of insect visitation of ornamental 48 49 flowers in Southover Grange garden, Lewes, UK Insect Science 22:700-705 50 doi:10.1111/1744-7917.12162 51 52 53 Guitart D, Pickering C, Byrne J (2012) Past results and future directions in urban community 54 gardens research Urban Forestry & Urban Greening 11:364-373 55 doi:http://dx.doi.org/10.1016/j.ufug.2012.06.007 56 57 58 Gunnarsson B, Federsel LM (2014) Bumblebees in the city: abundance, species richness and 59 diversity in two urban habitats Journal of Insect Conservation 18:1185-1191 60 61 62 14 63 64 65 Heard T, Hendrikz JK (1993) Factors Influencing Flight Activity of Colonies of the Stingless 1 Bee Trigona-Carbonaria (Hymenoptera, Apidae) vol 41. doi:10.1071/ZO9930343 2 3 4 Hennig E, Ghazoul J (2011) Pollinating in the urban environment vol 15. 5 doi:10.1007/s11252-011-0202-7 6 7 8 Irshad M (2014) Role of Syrphids (Diptera: Syrphidae) as Biotic Agents and Pollinators in 9 Pakistan Journal of Bioresource Management 1:2 10 11 12 Joshi S, Ballal CR (2013) Syrphid predators for biological control of aphids Journal of 13 Biological Control 27:151-170 14 15 Kevan PG (1972) Insect pollination of high arctic flowers The Journal of Ecology:831-847 16 17 Latty T, Beekman M (2009) Food quality affects search strategy in the acellular slime mould, 18 Physarum polycephalum Behavioral Ecology 20:1160-1167 19 doi:10.1093/beheco/arp111 20 21 22 Lawson LJ (2005) City bountiful A Century of Community Gardening in America Berkeley and 23 Los Angeles, California and London, England: University of California Press, Ltd 24 25 26 Lefebvre V, Fontaine C, Villemant C, Daugeron C (2014) Are empidine dance flies major 27 flower visitors in alpine environments? A case study in the Alps, France Biology 28 Letters 10:20140742 29 30 31 Majewska A, Altizer S (2019) Planting gardens to support insect pollinators. Conservation 32 Biology. doi:10.1111/cobi.13271 33 34 35 Makinson JC, Threlfall CG, Latty T (2017) Bee-friendly community gardens: Impact of 36 environmental variables on the richness and abundance of exotic and native bees 37 Urban Ecosystems 20:463-476 38 39 40 Matteson KC, Langellotto GA (2010) Determinates of inner city butterfly and bee species 41 richness Urban Ecosystems 13:333-347 42 43 44 Pardee GL, Philpott SM (2014) Native plants are the bee’s knees: local and landscape 45 predictors of bee richness and abundance in backyard gardens Urban Ecosystems 46 47 17:641-659 48 49 Pawelek JC, Frankie GW, Thorp RW (2009) Modification of a community garden to attract 50 native bee pollinators in urban San Luis Obispo, California 51 52 53 Pereira ALC, Taques TC, Valim JO, Madureira AP, Campos WG (2015) The management of 54 bee communities by intercropping with flowering basil (Ocimum basilicum) enhances 55 56 pollination and yield of bell pepper (Capsicum annuum) Journal of insect 57 conservation 19:479-486 58 59 60 61 62 15 63 64 65 Quinn GP, Keough MJ (2002) Experimental design and data analysis for biologists. Cambrige 1 University Press., Cambrige, UK 2 3 4 Ratnieks FLW, Garbuzov M (2014) Listmania: The Strengths and Weaknesses of Lists of 5 Garden Plants to Help Pollinators BioScience 64:1019-1026 6 doi:10.1093/biosci/biu150 7 8 9 Robinson L (1991) Field guide to the native plants of Sydney. Kangaroo Press, 10 11 12 Salisbury A, Armitage J, Bostock H, Perry J, Tatchell M, Thompson K (2015) EDITOR'S CHOICE: 13 Enhancing gardens as habitats for flower-visiting aerial insects (pollinators): should 14 we plant native or exotic species? Journal of Applied Ecology 52:1156-1164 15 doi:10.1111/1365-2664.12499 16 17 18 Senapathi D, Goddard MA, Kunin WE, Baldock KCR (2017) Landscape impacts on pollinator 19 communities in temperate systems: evidence and knowledge gaps Functional 20 21 Ecology 31:26-37 doi:10.1111/1365-2435.12809 22 23 Shakeel M et al. (2018) Insect pollinators diversity and abundance in Eruca sativa Mill. 24 25 (Arugula) and Brassica rapa L. (Field mustard) crops Saudi Journal of Biological 26 Sciences doi:https://doi.org/10.1016/j.sjbs.2018.08.012 27 28 Soga M, Yamaura Y, Koike S, Gaston KJ (2014) Land sharing vs. land sparing: does the 29 30 compact city reconcile urban development and biodiversity conservation? Journal of 31 Applied Ecology 51:1378-1386 32 33 34 Ssymank A, Kearns CA, Pape T, Thompson FC (2008) Pollinating flies (Diptera): a major 35 contribution to plant diversity and agricultural production Biodiversity 9:86-89 36 37 Stelzer RJ, Chittka L, Carlton M, Ings TC (2010) Winter Active Bumblebees (Bombus 38 39 terrestris) Achieve High Foraging Rates in Urban Britain Plos One 5:7 40 doi:10.1371/journal.pone.0009559 41 42 43 Threlfall CG, Walker K, Williams NSG, Hahs AK, Mata L, Stork N, Livesley SJ (2015) The 44 conservation value of urban green space habitats for Australian native bee 45 communities Biological Conservation 187:240-248 46 47 doi:https://doi.org/10.1016/j.biocon.2015.05.003 48 49 50 Turo KJ, Gardiner MM (2019) From potential to practical: conserving bees in urban public 51 green spaces Frontiers in Ecology and the Environment 52 53 54 55 56 57 58 59 60 61 62 16 63 64 65 Table 1. The results of a general linear mixed model with flower visitor abundance and 1 2 3 diversity (species richness) as the dependent variables. Factors significant at the 0.05 level 4 5 are indicated by an asterisk. 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 17 63 64 65 Table 2. Top 10 flowers ranked highest to lowest in three metrics: normalised degree, 1 2 3 visitation rate and number of gardens. Species names in bold appear on all three top-10 4 5 lists. Those with an underline appear in at least two lists. 6 7 8 Rank Species Normalised Species Visitation Species Number of 9 10 degree rate gardens 11 1 Salvia elegans 0.484 Brassica 20.67 Ocimum 18 12 rapa basilicum 13 14 2 Ocimum 0.475 Brassica 19.83 Lavandula sp 13 15 basilicum juncea 16 3 Cosmos 0.444 Salvia 19.5 Tropaeolum 12 17 18 bipinnatus hispanica majus 19 4 Brassica rapa 0.443 17.33 Borago 11 20 polisa officinalis 21 22 5 Eruca vesicaria 0.440 Veronica 17 Rosmarinus 11 23 persica officinalis 24 6 Lobularia 0.433 Stellaria 16 Tanacetum 11 25 26 maritima media parthenium 27 7 Plectranthus 0.403 Ocimum 9.27 Brassica rapa 10 28 caninus basilicum 29 30 8 Salvia leucantha 0.395 Brassica 9 Salvia elegans 10 31 oleracea 32 9 Tanacetum 0.393 Lobularia 8.29 Taraxacum 8 33 parthenium maritima officinale 34 35 10 Brassica juncea 0.375 Clematis 8 Brassica 7 36 terniflora oleracea 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 18 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Figure 1. Map of the Sydney metropolitan area. Circles show the location of community 29 30 31 gardens used in this study. The numbers correspond to the garden name in the legend. Note 32 33 that the majority of gardens cluster in the inner city, where back or front yards are generally 34 35 36 small or on-existent. 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 19 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Figure 2. Visualisation of insect-flower network showing interactions between plants 27 28 29 30 (upper boxes) and insects (lower boxes) in all gardens combined. The width of the lines 31 32 connecting species is scaled to the number of interactions. Insect orders are identified with 33 34 35 the small silhouette symbol. The six most connected flower species and two most connected 36 37 insect species are labelled. The arrows link boxes to species names for clarity. 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 20 63 64 65

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Figure 3. The effect of flower species richness on (a) the species richness and (b) abundance 34 35 36 of flower visitors. The shaded areas indicate the 95% confidence intervals. 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 21 63 64 65 Appendix 1: All species of floral visitor observed throughout the study period and the total 1 2 3 number of floral visits recorded for each species. 4 5 6 Species Family Floral visits 7 Apis mellifera Apidae 932 8 9 Melangyna sp. Syrphidae 896 10 Betasyrphus serarius Syrphidae 95 11 Tetragonula carbonaria Apidae 65 12 13 Eupeodes confrater Syrphidae 63 14 Lucilia sp. Calliphoridae 41 15 Muscidae sp. Muscidae 35 16 17 Pieris rapae Pieridae 32 18 Exoneura sp. Apidae 13 19 Vespidae sp. Vespidae 13 20 21 Pentatomidae sp. Pentatomidae 10 22 Meliscaeva sp. Syrphidae 8 23 Drosophila melanogaster Drosophilidae 7 24 25 Diptera sp. ? 6 26 Eristalinus punctulatus Syrphidae 6 27 Sphaerophoria macrogaster Syrphidae 5 28 29 Syrphidae sp. Syrphidae 3 30 Tipulidae sp. Tipulidae 3 31 Eumerus Syphidae 3 32 33 Choreutidae sp. Choreutidae 2 34 Eristalinus aeneus Syrphidae 2 35 grandicornis Syrphidae 1 36 37 Pergidae sp. Pergidae 1 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 22 63 64 65 Appendix 2: All flowering plant species observed throughout the study period and the total 1 2 3 number of floral visits (if any) recorded. 4 5 6 Species Family Floral visits 7 Brassica rapa Brassicaceae 310 8 9 Ocimum basilicum Lamiaceae 269 10 Lavandula sp. Lamiaceae 150 11 Borago officinalis Boraginaceae 137 12 13 Veronica persica Plantaginaceae 119 14 Brassica juncea Brassicaceae 119 15 Lobularia maritima Brassicaceae 116 16 17 Salvia elegans Lamiaceae 91 18 Stellaria media Caryophyllaceae 80 19 Brassica oleracea Brassicaceae 72 20 21 Taraxacum officinale Asteraceae 59 22 Bidens pilosa Asteraceae 52 23 Chysanthemum parthenium Asteraceae 48 24 25 Eruca vesicaria Brassicaceae 45 26 Vicia faba Fabaceae 43 27 Rosmarinus officinalis Lamiaceae 41 28 29 Plectranthus caninus Lamiaceae 40 30 Salvia hispanica Lamiaceae 39 31 Tropaeolum majus Tropaeolaceae 32 32 33 Clematis terniflora Ranunculaceae 32 34 Brassica oleracea (Kohlrabi) Brassicaceae 26 35 Brassica oleracea var. sabellica Brassicaceae 25 36 37 Cosmos bipinnatus Asteraceae 23 38 Argyranthemum frutescens Asteraceae 19 39 Fragaria x annassa Rosaceae 17 40 41 Nepeta x faassenii Lamiaceae 16 42 Calendula officinalis Asteraceae 14 43 Tagetes erecta Asteraceae 14 44 45 Buddleja madagascariensis Scrophulariaceae 14 46 Malcolmia maritima Brassicaceae 14 47 Filipendula ulmaria Rosaceae 13 48 49 Trifolium repens Fabaceae 12 50 Brassica oleracea var. alboglabra Brassicaceae 12 51 Trifolium pratense Fabaceae 11 52 53 Osteospermum ecklonis Asteraceae 11 54 spiralis 10 55 Pisum sativum Fabaceae 7 56 57 Helianthus annuus Asteraceae 7 58 Salvia leucantha Lamiaceae 7 59 Viola tricolor Violaceae 7 60 61 62 23 63 64 65 Achillea millefolium Asteraceae 7 1 Solanum nigrum Solanaceae 7 2 3 Eruca sativa Brassicaceae 6 4 Chrysanthemum morifolium Asteraceae 6 5 Chrysanthemum coronarium Asteraceae 6 6 7 Crassula ovata Crassulaceae 6 8 Argyrathemum frutescens (Pink) Asteraceae 6 9 Tancetum vulgare Asteraceae 6 10 11 Oxalis crassipes rosea Oxalidaceae 4 12 Jonquil erlicheer Amaryllidaceae 4 13 Coriandrum sativum Apiaceae 4 14 15 Aster frikartii Asteraceae 2 16 Salvia officinalis Lamiaceae 1 17 Salvia farinacea Lamiaceae 1 18 19 Cosmos sulphureus Asteraceae 1 20 Pelargonium x hortorum Geraniaceae 1 21 Myosotis scorpiodes Boraginaceae 1 22 23 Solanum lycopersicum Solanaceae 0 24 Echinacea angustifolia Asteraceae 0 25 Salvia microphylla Lamiaceae 0 26 27 Allium tuberosum Amaryllidaceae 0 28 Origanum vulgare Lamiaceae 0 29 Helianthus tuberosus Asteraceae 0 30 31 Salvia dorisiana Lamiaceae 0 32 Salvia leucantha x Salvia chiapensis Lamiaceae 0 33 Salvia splendens Lamiaceae 0 34 35 Pelargonium inquinans Geraniaceae 0 36 Papaveraceae 0 37 Gazania rigens Asteraceae 0 38 39 Ribes uva-crispa Grossulariaceae 0 40 Capsicum annuum Solanaceae 0 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 24 63 64 65