Choosing to be picky: an evidence-based system for selecting pollinator-friendly garden plants to support biodiversity

Abstract

Pollinator declines are a global concern, and garden plants can help mitigate declines in biodiversity. The availability, functionality, variety, and desirability of plants in cultivation have resulted in many lists of pollinator-friendly garden plants, but these often lack clear evidence, being based largely on author expertise. By reviewing the Royal Horticultural Society's Plants for Pollinators garden plants list, we evaluated 354 plant taxa using a decision tree and a working group. Over 90% of taxa have good evidence of being beneficial to pollinators, whereas 4% were excluded, and 5% required further investigation. This methodology provides a practical approach to creating, reviewing, and refining plant lists for pollinators, ensuring they are based on robust empirical data. This process has the potential to be adapted and applied globally to enhance pollinator-friendly planting schemes. We present a structured evidence-based methodology of assessing evidence for cultivated plants for pollinators.

The need to preserve biodiversity and increasing concern over pollinator declines, with potential economic losses and negative impacts on human health, has led to greater focus on supporting pollinators with international, national and local strategies, plans, and guidance (e.g., Llywodraeth Cymru Welsh Government 2013; England, DEFRA 2014; the United States, Bureau of Land Management 2022; international, IBES 2018; the European Union 2024; all of Ireland, National Biodiversity Data Centre 2021). These strategies often emphasize that residential domestic gardens (yards) and other similar cultivated spaces, particularly in urban areas, provide a resource that can support abundant, functioning diverse pollinator communities, with good evidence particularly in North America and Europe (e.g., Baldock et al. 2015, Martins et al. 2017, Marín et al. 2020, Staab et al. 2020, Tew et al. 2021, Poole et al. 2024). Although, individually, these gardens are often small, they make up 16%–36% of urban areas and are very variable, for example, in size, shape, topography, aspect, and soil type, in addition to management, which is dependent on individual gardeners (Loram et al. 2007, Goddard et al. 2013, Baldock et al. 2019, Philpott et al. 2020, ONS 2023). This creates a diversity of garden features that provide a variety of nesting sites and other habitat requirements for pollinating insects—for example, hedges and fences for bumblebees (Osborne et al. 2008). The major factor for pollinator success in gardens, however, is the amount of pollen and nectar (forage) available. This flower forage is demonstrated by nectar availability; domestic gardens provide an estimated 85% of nectar in urban areas (Tew et al. 2021), and there is evidence that they reduce seasonal hunger gaps for farmland pollinators in rural areas (Timberlake et al. 2024). Both the abundance and species composition of flowering plants vary considerably among gardens and with their variety of native and nonnative plants, gardens can have far greater species richness than seminatural habitats (Loram et al. 2008, Goddard et al. 2013, Tew et al. 2022). This richness is directly related to the stability of nectar production in urban areas, reflects the desire of gardeners to have season-round flowering plants, and illustrates the value of nonnative cultivated ornamental plants (Salisbury et al. 2015, Staab et al. 2020, Tew et al. 2022). These factors signify that, although nesting resources and the needs of nonpollinating life stages should not be ignored, it is a high priority to focus on the forage for pollinators; indeed, the importance of plant selection in making gardens and similar cultivated spaces more pollinator-friendly by judicious planting has been recognized for some time (Garbuzov and Ratnieks 2014a).

Among managers of gardens (gardeners), who determine which plants are grown, there is a strong desire to support wildlife. Up to three quarters (67%–74%) of gardeners in the United Kingdom have indicated they engage in activities in their gardens that support wildlife (Thornton 2009, Potts et al. 2010). Further surveys showed that 84% of UK gardeners who considered they do “a fair amount” or “a lot” for wildlife planted flowers for wildlife (Thomas et al. 2021), and 54% of UK gardeners agreed with the statement, “I grow plants specifically to feed and provide homes for bees and other pollinators” (Sutcliffe et al. 2024). The desire to grow plants that provide resources for pollinators is catered for by lists of pollinator plants for gardens.

Why do lists of cultivated pollinator plants need reviewing?

Although lists of pollinator plants for gardens are popular and raise awareness, an analysis of 15 lists from Europe and North America found that they did not appear to be well grounded in empirical data, with limited evidence that they were anything other than based largely on personal observation, opinion, and uncritical recycling of other lists (Garbuzov and Ratnieks 2014a). The objective of this work is to address this lack of evidence by providing a practical method that uses robust criteria for ongoing revision and creation of lists of cultivated plants for pollinators.

Reviewing a list of cultivated garden plants for pollinators

One of the largest and most widely cited cultivated plant list schemes is the Royal Horticultural Society (RHS) Plants for Pollinators (Garbuzov and Ratnieks 2014a), which is referenced and promoted in the National Pollinator Strategy for England (Defra 2014). It was launched in 2011 (as RHS Perfect for Pollinators) and had undergone minor changes and refinements by the end of 2019 (RHS 2019), including the addition of some plants and the name change in 2015 to the RHS Plants for Pollinators. The purpose of the lists is to provide UK gardeners (and anyone planting or specifying plants within an ornamental cultivated space) with a list of garden plants whose flowers are attractive to pollinating insects. Plants on the list also qualify for labeling with a trademarked logo (figure 1) by UK plant suppliers at no charge from the scheme operators (the RHS). This does not reflect how the plants were produced but enables home gardeners and other users of cultivated plants to pick plants that provide resources for pollinators.

Figure 1.

RHS Plants for Pollinators trademarked logo.

At the end of 2019, the RHS Plants for Pollinators consisted of three lists: garden plants, wildflowers, and plants of the world. The wildflowers list consists of plants that are native to the British Isles or archaeophytes (nonnative plants intentionally or unintentionally introduced into Britain before approximately 1500 CE; Pyšek et al. 2004) and that are either suited for cultivated spaces or, where appropriate, for natural or seminatural habitat creations, such as wildflower meadows. The plants of the world list is a subset of the other two, created to aid the selection of pollinator-friendly plants from different regions of the world. We focus on the largest of these lists, garden plants, which is a selection of garden-worthy cultivated (defined below) plants, both native and nonnative to Britain.

The selected list contains taxa at several taxonomic levels: genus (including hybrid genus), species, and infraspecific (including subspecies, hybrid, variety, and cultivar). All taxa at the level below that listed are considered plants for pollinators, except where they provide significantly reduced floral resources. Plants with reduced access to or provision of pollen, nectar, or both are excluded from the list. A typical case would be double flowers, where petals are produced instead of the plant organs that support pollinators (Comba et al. 1999). For example, where a genus is listed, such as Rosa species, all Rosa species, subspecies, and other infraspecific taxonomic levels are considered plants for pollinators, but varieties such as Rosa ‘Paul's Himalayan Musk’ are excluded because double bloomed flowers have a reduced access and provision for pollinators. Therefore, despite exclusions, the number of plants that the Plants for Pollinators can apply to is much greater than that listed. At the end of 2019, this list consisted of 354 plant taxa (RHS 2019), the majority (281) of which were listed at the species level, 48 listed at genus, and the remaining taxa at hybrid genus (1), subspecies, and other taxa below species (25).

Structuring an evidence-based review of suitability of cultivated plants for pollinators

To show that plants are suitable for gardens and pollinators, the evidence should be assessed in a structured manner. To achieve this the evaluation, carried out in 2022, consisted of two parts: a decision tree (flowchart) based on searches of peer reviewed academic papers and gray literature and a working group to verify outcomes from part 1.

Part 1: The decision tree was divided into four (figure 2)

Figure 2.

The decision tree (flowchart), part 1 of a two-part evidence-based review of suitability of cultivated plants for pollinators. The evidence to meet the criteria is based on searches of peer reviewed academic papers and gray literature.

Decision 1: Is the plant suitable for garden cultivation in the current UK climate? The first decision was divided into three sections: availability, the ability to grow, and invasiveness.

First, is the plant available? Availability was based on whether the taxon was listed in RHS Plant Finder (Cubey et al. 2022) and RHS Find a Plant (www.rhs.org.uk/plants/search-form): Some plant groups such as bedding are not well represented in these publications, and so Internet searches of major UK suppliers were undertaken when no suppliers were listed.

Can it be cultivated outdoors in the United Kingdom, and is it therefore accessible to pollinators? The answer was based on the UK hardiness rating (David 2013, RHS 2022). If the taxa had not been allocated a hardiness rating, additional information was gained from online published cultivation notes from respected sources such as the RHS website. Plants with a rating lower than H3 were considered unsuitable for outdoor cultivation in the United Kingdom. H3 is defined as “half-hardy—likely to survive in an unheated glasshouse in a mild winter—Hardy in coastal [or] mild areas except in hard winters and at risk from sudden (early) frosts. May be hardy elsewhere with shelter or good microclimate. Can survive with artificial winter protection”; this is roughly equivalent to US Department of Agriculture hardiness rating of 9b/10 (Widrlechner et al. 2012). Perennial plants with a hardiness rating lower than H3 but frequently grown as annuals outdoors in the United Kingdom were considered to meet this criterion, provided that they produce flowers and resource for pollinators—for example, Nicotiana langsdorffii Weinm (RHS 2024).

Is it an invasive species? A plant was rejected if it was included in schedule 9 or 14 of the Wildlife and Countryside Act (Defra 2011, UK government 2022) or retained EU legislation (EU 2014), which list nonnative animals and plants representing a threat to the natural fauna and flora in England and Wales.

Plants that met the suitability for growing in the UK criteria progressed to decision 2; those that did not were excluded from the list.

Decision 2: Is there good evidence that the flowers of the plants listed are frequently visited by UK pollinators, making the plants a suitable resource? Publications were found by searching Google Scholar in January 2022 using the plant name (genus, binomial, and vernacular) and the terms pollinator, bee, and insect visit or visitation. These searches returned 424 relevant peer-reviewed publications and 33 independent gray sources (see the “Sources” section of supplement S1). High evidence was defined as having high-quality data, where a plant was listed in at least one peer-reviewed paper and there was agreement across more than one independent source. Where there was no good evidence in peer-reviewed publications, reference in multiple (two or more) independent edited (gray literature, as it was defined by Farace and Schöpfel 2010) sources was considered as medium evidence. Low evidence was where plants were also included on lists other than the RHS Plants for Pollinators. No evidence was where plants were only listed on RHS Plants for Pollinators and no other evidence was found. Frequently visited was agreed where either quantitative data in the reference showed visitation frequency as particularly high or a description suggested the same thing. Plants would fail to satisfy this criterion if there was evidence of only occasional visitation or an absence of quantitative evidence for visit frequency despite evidence of some visitation occurring. Although the present work is focused on plants being grown in the United Kingdom, data from studies outside the United Kingdom was used provided that the work covered the same broad insect groups that occur in the United Kingdom (e.g., bees, Hymenoptera; flies, Diptera; butterflies and moths, Lepidoptera; beetles, Coleoptera; but not bats or hummingbirds). This is acceptable because of the opportunism and generalization that is found in pollinator networks (Waser et al. 1996). If this criterion was met, the plant was considered suitable for inclusion. Plants failing to meet this specification were not excluded but moved to decision 3.

Decision 3: Is there good evidence that the flowers are occasionally visited by UK pollinators? Decision 3 used the high, medium, low, and no evidence criteria as for decision 2, but where visitation was not noted as being particularly high. If there was no evidence of a plant being occasionally visited by pollinators, the plant was recommended for rejection. If there was evidence that the plant was occasionally visited by pollinators, this was not by itself considered evidence that the plant was suitable for inclusion, so that plant was moved to decision 4.

Decision 4: Is the plant important to UK pollinators for an additional reason? The answer was defined as meeting one or more of four criteria and applied only to plants that had evidence of being occasionally visited by UK pollinators (decision 3): They must be high in floral resources; They must exhibit published data showing a high level of nectar, pollen or floral oils produced by flowers or inflorescences—for example, Tew and colleagues (2023). They must have an important flowering period; If the published flowering period fell between September and April this was considered important as there would usually be limited floral resources during this time—for example, for winter-active bumblebees (Stelzer et al. 2010). They must bloom prolifically; Evidence that a plant produces a large number of flowers, or blooms, with substantial overall resources for pollinators. It signifies abundant and often rapid flowering, indicating a high level of fertility and productivity in terms of flowering. The term prolific bloom is used to describe plants that bloom extensively, creating a visually striking display, and includes plants such as Hydrangea and Echinacea. They must attract (benefit) more than one insect order; There must be published reports indicating that many different types of pollinating insects visit a plant, including those from at least two of the insect orders Hymenoptera (excluding ants), Lepidoptera, Diptera and Coleoptera (excluding pollen beetles). Plants that met at least one of these criteria were considered suitable for inclusion on the list; those that did not were moved to the list of plants recommended for exclusion.

Part 1 of the review (the decision tree) gave one of three recommendations: to accept the plant as a suitable resource for pollinators (no change); raise the taxon level (e.g., from infraspecific to species or species to genus), where there was good evidence that most plants in that species or genus had high visitation (decision 2) or met at least one of the criteria in question 3 and therefore merited inclusion; and to remove the plant because there is not enough evidence for the plant taxon to remain on the list

Part 2: Working group

The working group was set up to add an extra layer of rigor. The internal (RHS) working group consisted of experienced botanists, horticulturists, and entomologists (the present authors AS, HB, SB, SH, RM, and MT). The recommended changes from part 1, the decision tree, were reviewed. This was carried out at in-person meetings and included further web and literature searches and personal experience. The working group had clearly defined terms of reference; it could not reject the proposed changes in part 1 without finding additional good evidence, as is defined in decision 2. The working group had one additional option—to flag the plant for more investigation (observe), a recommendation to obtain further evidence to support inclusion or reject the taxon. In these cases, the taxon was retained on the list while under review.

There is evidence that most plants on the RHS garden plants list provide resource for pollinators

All 354 taxa on the RHS Plants for Pollinators garden plants list were reviewed, and more than 90% met the criteria. Outcomes of the process are given in tables 1, 2, and 3 (the full decision matrix outcomes are available as supplement S1, and the working group outcomes are in supplement S2) and the final published list (RHS 2025a; www.rhs.org.uk/plantsforpollinators). In part 1 (decision tree) of the review, all plants on the list passed decision 1; that is, they were all available and suitable for cultivation in the United Kingdom, and none were legally defined as invasive or a species of special concern in England and Wales (DEFRA 2011, EU 2014, UK Gov 2022). The majority (292, 82%) met our definition of medium or good quality of evidence for decisions 2 and 3 (evidence of frequently or occasionally visited by pollinators), with 59 plants (17%) having low evidence and just two (less than 1%) with no evidence. Within this, there was evidence that just over one-third (127, 36%) of the plant taxa on the list were frequently visited by pollinators and were therefore accepted to the list at decision 2. Nearly two-thirds of the plants on the list (224, 63%) had evidence of being occasionally visited by pollinators and so were required to meet at least one of the criteria in decision 4. Of these occasionally visited plants 199 (89%) met at least one of the decision 4 criteria, with 75 (33%) being high in floral resources, 118 (53%) having a key flowering season, 87 being prolific blooming, and 101 (45%) benefiting multiple insect orders. Therefore 323 of the 354 (91%) taxa on the list in 2019 met the criteria at part 1 of the decision tree or the review.

Table 1.

Summary of the outcome of the plants for pollinators review (excluding raise to genera recommendations) of the 354 plant taxa on the 2019 list.

	Decision 1 a					Decision 2	Decision 3	Decision 4
		Weight of evidence b				Frequently visited	Occasionally visited	High resource		Key season		Prolific blooming		Benefits multiple orders		Met at least 1 at decision 4		Decision tree outcome
Taxon level	Suitable for consideration	High	Medium	Low	None			Number c	Frequently visited	Number	Frequently visited	Number	Frequently visited	Number	Frequently visited	Number	Frequently visited	Total met criteria	Not meeting criteria	Working group outcome	New list before raising to genus
Genus	48	36	11	-	1	25	22	34	16	31	15	10	5	37	17	48	23	47	1	1 observe	48
Species	281	109	120	50	-	95	184	107	54	137	92	98	72	143	82	246	161	252	29	13 remove 16 observe	268
Infraspecific taxad	25	4	12	9	-	7	16	9	4	15	11	12	10	3	2	21	16	24	1	1 remove	24
Total	354	149	143	59	1	127	224	150	74	183	118	120	87	183	101	317	200	323	31	14 remove 17 observe	340

^aConsisted of three criteria that all plants on the 2019 list met; they were found to be suitable for consideration i.e., they were all available, suitable for cultivation in the UK climate and none were legally scheduled as invasive in England and Wales.

^b High was defined as either high-quality data where a plant was listed in at least one peer-reviewed paper or there was agreement across more than one independent source. Where there was no good evidence in peer-reviewed publications, reference in multiple (two or more) independent edited (gray literature) sources was considered as medium evidence. Low evidence was where a plant was listed in one or two online lists without any published robust methods. None was where plants were listed on RHS Plants for Pollinators in 2019, and no other evidence was found.

^cTotal meeting criteria of those occasionally visited.

^dIncluding subspecies, hybrid, variety, and cultivar.

Table 2.

Plant taxa listed on the 2019 RHS Plants for Pollinators list at species (or lower taxa) which during the process of review, were found that there was good evidence to replace entries at species level or below with new entries at genus level.

	Numbers on 2019 list
Genus	Species	Infraspecific taxa a
Anchusa	2	0
Campanula	7	0
Centaurea	7	0
Doronicum	0	1
Elaeagnus	2	1
Erica	4	0
Eryngium	3	1
Euphorbia	8	1
Gypsophila	1	0
Hedera	1	0
Helianthus	1	1
Lavandula	2	1
Lavatera	2	0
Malus	2	0
Origanum	1	1
Pieris	2	0
Prunus	10	2
Salix	3	1
Sarcococca	2	0
Total (19 genera)	60	11

^aIncluding subspecies, hybrid, variety, and cultivar.

Table 3.

Summary of changes made to the 2019 RHS Plants for Pollinators list following evidence-based review.

Taxon level	2019 list	Removed (lack of evidence)	Observe (further evidence required)	Effect of raising to genus	Final after review, 2024 list
Genus a	48	0	1	+19	68
Species	281	13	16	–60	208
Infraspecific taxab	25	1	0	–11	13
Total	354	14	17	–65	289

^aIncluding hybrid genera.

^bIncluding subspecies, hybrid, variety, and cultivar.

Of the 31 (9%) plant taxa that did not meet the criteria of the decision tree, the working group (part 2) agreed that 14 (4%) of the taxa should be removed because of the lack of evidence. The remaining 17 plant taxa were added to a list of plants where investigation would be required (observe) but remain on the list as a plant for pollinators while under investigation (See supplement S2), because, in all cases, the working group considered that, although published evidence was lacking, the plants were likely to be a suitable resource for pollinators. Therefore, 340 (96%) of the 354 plant taxa on the list of 2019 were retained.

During the review, it was determined that there was good evidence that 71 species-level (or below) entries could be consolidated into 19 genus-level entries (table 2). For example, four eryngiums listed in 2019 (Eryngium alpinum, Eryngium giganteum, Eryngium planum, Eryngium × tripartitum) were replaced with “Eryngium species.” The number of eryngium species of hardiness rating H3 or above and listed in the RHS Plant Finder so available to UK gardeners is 33 (Cubey et al. 2022). In one case, the working group did not agree with a recommendation resulting from part 1 that all trees in the genus Tilia (known as lime trees in the United Kingdom and linden trees elsewhere) should be included; this was because of bee deaths associated with Tilia tomentosa Moench, although it is doubtful that these are the result of toxic nectar (Fossen et al. 2019). In the final list, the number of taxa listed at the genus level rose by 19 to 68, whereas the number of taxa listed at species level was reduced by 60 and lower taxa by 11. These changes led to an overall reduction in the plant names on the list from 354 to 289 (table 3). This has the effect that, although the list may appear shorter, there has, in fact, been an increase in the number of plants to which the descriptor Plants for Pollinators can be applied.

Evidence-based refinement of pollinator plant lists is worthwhile

The present review has enabled a refinement of a popular list and demonstrated a simple decision process for plants to be evaluated for inclusion as pollinator plants, which could be applied worldwide. This could be used on lists that already exist, such as Xerces Society (2011) in North America or the International Bee Research Association (IBRA 2008) for Britain and Western Europe, or for creating new lists. By doing this, we have developed a process that counters one of the greatest criticisms of these lists: that they are not backed by empirical data (Garbuzov and Ratnieks 2014a).

Over 90% of the plant taxa listed on the RHS Plants for Pollinators garden plant list met the evidence-based criteria for inclusion as plants good for pollinators, indicating that the list was on a sound footing and provided good recommendations for cultivated spaces for pollinators in the United Kingdom before the review. Because of significant crossover between pollinator lists (Garbuzov and Ratnieks 2014a), this by implication suggests that other lists are also likely to be good lists of plants for pollinators. Even so, the 9% recommended for removal, of which 4% were accepted by a working group and the remaining 5% recommended for further observation, show that a methodical review of plant lists for pollinators is worthwhile and will further increase confidence in the efficacy of the plants by anyone using them. This review goes further and demonstrates that lists of cultivated plants promoted as providing a resource for biodiversity, an environmental benefit or ecosystem service, can be evaluated in a systematic manner. Ideally, this should be done before the list is launched, but it can be applied retrospectively, provided changes and methodology are transparent and are clearly communicated to end users and the industry.

Considering new evidence and application outside of the United Kingdom

Although we have demonstrated a process that will lead to improved lists, there is an ongoing need to reevaluate the criteria as new evidence emerges for individual plants but also to account for factors such as climate change, which will affect the geographical range where plants are suitable for cultivation (Webster 2017). The number of plants in cultivation and their availability is also constantly changing; for example the 2022 edition of RHS Plant Finder contained 69,000 plants, of which 4200 were new additions (Cubey et al. 2022). Some further developments and refinement can also be considered, and processes should not remain static. In part 1, the first decision (“Is the plant suitable for garden cultivation in a UK climate?”) does not directly relate to the properties of the plant as a resource for pollinators and is region (UK) specific. It is, however, an essential prerequisite for plants that are being considered for a cultivated garden, excluding plants that can't readily be obtained or grown outdoors or that pose a threat to biodiversity. Although all plants considered met the three criteria that were specified, some further refinements could be considered, and appropriate modifications to the criteria and evidence would enable application outside of the United Kingdom.

Availability was based on whether the taxon was listed in RHS Plant Finder (Cubey et al. 2022). Although this publication and its online equivalent (RHS Find a Plant) are considered the industry standard for suppliers in the United Kingdom, it is reliant on nurseries supplying information and underrepresents some plants—for example, bedding plants. This necessitated the use of wider Internet searches of commercial suppliers to check the availability of some plants. In regions without an equivalent of RHS Plant Finder, Internet searches may be the only viable option to assess availability. This gives a good indication of the availability of plants at the time of the search but gives no indication of future availability and so should be reviewed regularly, especially for plants listed at infraspecific taxonomic levels (e.g., cultivars). Some lists have relied on plants being native as a criterion for wildlife friendliness, which can result in plants being difficult to obtain from the industry supply chains or not aesthetically suited to gardeners’ tastes (Thompson 2006). Although availability does not necessarily equate to the attractiveness of a plant to gardeners, it suggests a degree of popularity. Plants that are already popular with gardeners are more likely to be planted (and therefore contribute to pollinator forage), although it should be noted that lists can help influence plant popularity trends because a plant with a pollinator friendly label is more likely to be purchased than one without (Wignall et al. 2019).

Can it be cultivated outdoors in the United Kingdom, and is it therefore accessible to pollinators? To provide a guide, the answer to this question was based on a UK hardiness rating of H3 or above (plants likely to survive mild winters) and included plants that can be treated as annuals in the United Kingdom (David 2013). This relied on published information on the RHS website (RHS 2022). There remains the need for gardeners to consider their own location and situation. For example, in some parts of the United Kingdom, only plants that are fully hardy (H7) are currently likely to survive—for instance, Crataegus monogyna in northern Scotland. This also highlights that there is an opportunity to review and refine lists to consider other requirements—for example, for specific regions of the United Kingdom and conditions such as soil type, waterlogging, and drought tolerance, thereby improving the recommendations for gardener and garden situation. This was however beyond the scope of this review. For other regions of the world, the US hardiness ratings can be used as a basis for assessing if plants are possible to cultivate out of doors—for example, PlantsMaps (2025).

Is it an invasive species? A plant was rejected if it met a legal definition—that is, as is listed on schedules 9 and 14 of the Wildlife and Countryside Act (UK government 2022 and DEFRA 2011) or a species of special concern (EU 2014). Invasive plants have been assessed worldwide and there are lists that would enable this criterion to be considered for regions outside of the United Kingdom, such as EPPO list of invasive alien plants (EPPO 2025). This criterion should also be reviewed as plants are added to regulatory lists or the taxonomy of plants is refined. For example, the discovery that Gunnera manicata was a hybrid of Gunnera tinctoria meant that the hybrid is now defined as an invasive species (Edwards et al. 2023). Although this criterion ensured invasive plants covered by legislation in England and Wales were not included, it was beyond the scope of this review to consider nonlegislated plants that are of concern as identified by the GB Non-native Species Secretariat's nonnative risk assessment scheme (NNSS 2024), or that could become invasive in the future.

The second and third questions of the decision tree used published, both peer review and gray literature to assess the value of the plant taxa on the lists for pollinators. The Google Scholar searches for plant name (binomial and vernacular) along with the terms pollinator, bee, and insect visit or visitation returned 424 relevant peer-reviewed publications and 33 independent sources. Although it is possible that additional search terms such as fly may have found further reference sources, the search terms used enabled an assessment as either high or medium evidence of a high proportion (82%) of plant taxa on the list. This indicates that the search was reliable and gave a good representation of the available data on the suitability of the cultivated plants reviewed for pollinators. Any method for assessing large numbers of cultivated plants as a resource for pollinators needs to be practical to use, especially given the number of plants in cultivation in the United Kingdom, estimated to be between 100,000 and 400,000 different types of plants (Griffiths 2023). The validity of this approach was confirmed by part 2, the working group, where further searches were unable to find additional evidence for the 9% of plants that were recommended for removal, although 17 plant taxa were identified as needing additional research (observation) before a decision on these plants can be finalized. In most cases the working group considered the plants were likely to be visited by pollinators but found data was lacking (supplement S2).

Assessment of the evidence in the literature was somewhat subjective, because publications rarely shared comparable methodologies. It was therefore necessary for the reviewers to make a judgement on the basis of the description in the publication. There was quantitative data in some publications—for example, ranking plants (Garbuzov et al. 2017) or large studies where visitor frequency was explored, such as an insect visit every 10 minutes or less (Ruppel et al. 2019); however, many publications indicated that a plant was especially frequently visited compared with others but did not provide data. This factor was also evident for decisions 2 and 3 of the decision tree (is there evidence the plant is frequently or occasionally visited) and further highlights the need for quantified processes to identify plants that provide for pollinators.

Although the assumption that a plant with noted high visitation rates is a plant that can be considered a plant for pollinators, it is important to include plants that may be visited less often, because they can also provide a valuable resource. Failure to include these plants risks an oversimplification of what qualifies as a good pollinator resource and constrains gardeners and designers to a narrower plant palette. For plants noted as being visited less often, the four criteria at decision 3 identified those that provided resources at an important time of year, provided prolific blooming, were visited by multiple orders of insects or were high in floral resources.

The added value of a working group: Reliance on a decision tree can lead to errors

The working group (part 2) was vital in making assessments of the plants and where appropriate confirming recommended changes from the decision tree (part 1). As we previously highlighted, for some cultivated plants, information is not readily available in the literature or via other information sources. For example, to find suppliers or hardiness information, additional searches and expert experience were required to inform decisions. This stage did not rely solely on empirical data and highlighted considerations that were not covered in the decision tree and that may not have come to light without expert knowledge. To address the potential subjectivity of decisions made by a working group clear terms of reference need to be provided; in the present study, the result of the decision tree could not be rejected without evidence meeting our definition of good, although plants could be flagged for further investigation. An example would be the recommendation that all lime or linden trees (Tilia taxa) should be included in the lists, raising the current five species listed to genus level. Without the working group this would have led to the inclusion of T. tomentosa on the list. However, because there is an association with bee deaths (Fossen et al. 2019), it is an inappropriate plant to include on a list of pollinator-friendly plants. Although, in this case, it would be feasible to add a decision to the flowchart, this would either need to be very specific (e.g., is there any known toxicity to pollinators?) or vague (e.g., are there any other known issues with the plant?). The latter is undesirable, because the question is open; the former is too specific and would add to the complexity of the flowchart. It is important to consider the membership of a working group: A range of views, expertise, and experience is required. As the lists are of garden plants for pollinators, a working group should include experienced practitioners (i.e., horticulturists) and those with botanical and taxonomic knowledge of cultivated plants, as well as pollination or entomological specialists. For example, before accepting a raise to genus level, a knowledge of the whole genus is essential; therefore, it would not have been possible to confirm that Euphorbia should be raised to genus level without the knowledge of the working group. This pooling of empirical expertise will be important when criteria regarding hardiness and effects of climate change and potential invasiveness are refined. These examples highlight that a group of experts, be this a steering, advisory, or working group, adds extra rigor and enables consideration of additional factors that a structured methodology such as a decision tree does not.

Adding pollinator plants and subjectivity

The 17 plants highlighted as requiring additional information (observation) and the raising of plants to higher taxa levels indicate that many other plants may be worthy of inclusion but lack evidence in the literature or simply had not been considered by 2019. Indeed, a criticism of pollinator plant lists is that they omit many good plants, and no list (including the RHS Plants for Pollinators) can ever be complete (Garbuzov and Ratnieks 2014a). The process presented in the present article not only enables a retrospective review of plants on lists but provides a basis by which additional plants can be assessed. The subjectivity in the literature mentioned above highlights that there would, however, be a benefit in establishing protocols that gather standardized data—for example, using or adapting methodologies such as the Pollinator Monitoring Scheme's FIT (flower insect timed) Count (https://ukpoms.org.uk/fit-counts), other citizen science projects (Garbuzov and Ratnieks 2014a), or comparative recording across a range of plants (Smitley et al. 2024). It was beyond the scope of this review to cover variation in what plants provide for the different pollinators that use them; nor have we covered general traits that may make a plant good for pollinators (Potter et al. 2019). Our focus in the present article is to assess cultivated plant taxa individually for value to pollinators, gardens are communities of plants and therefore planting combinations should be considered. The benefit that these combinations of cultivated plants can have has been highlighted, for example, by the evidence that gardens can reduce seasonal hunger gaps for farmland pollinators (Timberlake et al. 2024) and the support botanical collections can give to the recovery of the pollinator communities of natural and seminatural habitats (Vilella-Arnizaut et al. 2022). Some combinations that meet growing conditions and the goals of gardeners are available (e.g., RHS 2025b), and these would benefit from an improved evidence base and research to identify and enhance the pollinator communities they support. It is known that there is great variation in pollinator visits to different wildflower seed mixes and within one plant species cultivars can vary a hundredfold (Garbuzov and Ratnieks 2014b, Smitley et al. 2024). A method of ranking the relative attractiveness of plants, based on visitation, and so producing a guide to those plants that are of more value to pollinators than others has been identified as a knowledge gap (Garbuzov and Ratnieks 2014a). There should also be recognition that some plants may support specialists but not necessarily attract a large number or wide range of pollinators or that flower structure or traits may affect the suitability for different pollinator groups (Tew et al. 2022, Harris and Ratnieks 2024). The importance for pollinator communities and identity of relevant traits would benefit from further study; for example, work with red clover (Trifolium pratense) indicated that of the traits that were measured (inflorescence density, florets per inflorescence, floret length, and nectar sugar production), only inflorescence density was significantly related to insect abundance and species richness (Harris and Ratnieks 2024). Although there are generalizations around flowering time, such exotic plants can extend flowering season when native plants have finished flowering (Salisbury et al. 2015) other traits such as the value to crepuscular and nocturnal pollinators, and the season of flowering would also benefit from further investigation. The potential that plant collections in public gardens have to identify and quantify traits that could be important has already been highlighted (Garbuzov and Ratnieks 2015). In addition, the identification of such traits could help plant breeders select for plants that are worthy of a pollinator friendly label, which may lead to greater sales (Wignall et al. 2019). All these factors underscore that a strength of an extensive list, such as the RHS Plants for Pollinators garden plant list, is to promote diverse planting and therefore preserve and enhance the diversity of plant–pollinator interactions.

Conclusions

We have demonstrated that lists of cultivated garden plants, recommended as being suitable for pollinators, are supported by empirical evidence and that this can be determined by reviewing an established list. Over 90% of the plants on the RHS Plants for Pollinators garden plant list met the criteria for being a good resource for pollinators. The two-part process used information based on published evidence and consisted of a flowchart (decision tree) and a working group. The process was structured, evidence based, and efficient, which is essential, given the substantial number of plant taxa in cultivation. The flowchart used criteria on plant suitability for gardens and value as a resource to pollinators. Although largely objective and repeatable, the decision tree relied on published evidence, which contained a range of sometimes contrasting data types and was of varying quality, so there was a degree of subjectivity in interpreting the evidence. Applying an objective set of criteria to this information does not always give a clear result, and there are also occasional additional factors that the process did not identify. Input from an expert working group enabled a final decision. The process not only facilitates a retrospective review of plants already on such lists but can provide a basis on which additional plants can be assessed before inclusion and, ideally, can be carried out before lists are launched. This review goes further and demonstrates that lists of cultivated plants promoted as providing a resource for biodiversity and potentially other environmental benefits can be evaluated in a systematic manner. It is likely that many of the garden plants identified in the present article would also be excellent resources in similar biomes, although the criteria (availability, cultivation requirements, invasiveness, and value to pollinators) would need to be evaluated. This methodical approach to reviewing garden plant taxa provides a framework that can be adapted and applied to pollinator-friendly lists within the floras of other countries.

Author Biography

The authors are all affiliated or have been affiliated with the Royal Horticultural Society (RHS), in London, England, in the United Kingdom. Andrew Salisbury (andrewsalisbury@rhs.org.uk) is head of Plant Health and an entomologist. Stephanie Bird is a senior entomologist at the RHS. Helen Bostock is a senior wildlife specialist. Carol Barrie is a project manager for the RHS Plants for Purpose project. Saskia Harris is a horticultural taxonomist and a botanist in the Cultivated Plant Diversity and Advice team at RHS Garden Wisley. Rosalyn Marshall is trials project manager, coordinating assessments of plants for garden worthiness. Richard Sanford is a senior horticultural data officer, working on RHS Plant Finder and the RHS horticultural database. Mark Tuson is a horticulturist at RHS Garden Wisley. Nick Tew is a postdoctoral researcher based at the University of California Davis, in Davis, California, in the United States; he held a fixed-term position at the RHS working to review the Plants for Pollinators list.

Author contributions

Andrew Salisbury (Conceptualization, Methodology, Project administration, Writing - original draft, Writing - review & editing), Stephanie Bird (Conceptualization, Methodology, Writing - original draft, Writing - review & editing), Helen Bostock (Conceptualization, Methodology, Writing - original draft, Writing - review & editing), Carol Barrie (Writing - original draft, Writing - review & editing), Saskia Harris (Methodology, Writing - original draft, Writing - review & editing), Rosalyn Marshall (Methodology, Writing - original draft, Writing - review & editing), Richard Sanford (Writing - original draft, Writing - review & editing), Mark Tuson (Methodology, Writing - review & editing), Nick Tew (Conceptualization, Methodology, Writing - original draft, Writing - review & editing).