Research-based wildflower seed mix is more attractive to pollinating insects than standard mixes

Key points

  • Wildflower margins or plots are popular agri-environment scheme (AES) options that provide habitat and forage for insects, but some flower species and mixes are not underpinned by robust research.
  • Scientists created a new mix based on their own primary field research, and another new mix based on scientific literature. These were trialled alongside two standard AES wildflower mixes and a fallow plot, to determine which mix was most attractive to insects.
  • The primary research mix attracted both the highest number of total insect visitors, and the highest wild bee abundance and richness. It attracted more bumblebees, and significantly greater solitary bee abundance, than the other habitats.
  • Dandelion, spear thistle, wild carrot and hedgerow cranesbill received the highest numbers of wild bee species visits.
  • Authors recommend that:
    • Cornfield annuals are added to mixtures.
    • Pollen and nectar legume mixes are updated to include a broader range of flower species.
    • The ultimate flower mix for wild bees consists of alsike clover, kidney vetch, birds’-foot trefoil, wild carrot, hedgerow cranesbill, dandelion and rough hawkbit, with the addition of poppy and corn marigold as annuals.

View Visual Summary →


Wildflower seed mixIntensive agriculture can negatively affect the biodiversity of farmland by reducing semi-natural habitats and weeds. Agri-environment schemes (AES) have been implemented widely in recent decades to restore biodiversity. Wildflower mixes, grown in strips or plots, are a popular AES option, providing food for wild pollinators and insect predators of crop pests. They provide pollen and nectar for adult insects (e.g. bees) and host plants for larvae (e.g. moths and butterflies).

Wildflower mixes are made of non-competitive grasses and wildflowers. They differ in the species that they contain and the insects they aim to attract. For example, a mix comprising a small number of clovers and vetch species may be targeted at specialist pollinators such as bumblebees, whereas a mix containing lots of umbellifers (e.g. wild carrot, yarrow) and daisy-types (e.g. smooth hawksbeard, ox-eye daisy) may attract a wider range of pollinating insects.

Different flower species vary in their nutritional content. Generally, the main nutrient in nectar is carbohydrate, while pollen is a source of protein and lipids. Insects need to gather plenty of these nutrients for their larvae, and for themselves.

Previous research has shown that different AES wildflower mixes attract different types of pollinating insects, but few have considered novel or bespoke seed mixes. This study addresses that knowledge gap by comparing two standard AES wildflower seed mixes with two novel wildflower seed mixes that were specially designed to attract wild bees.

They aimed to find out:

  • Which mix attracted the highest number of insect pollinators
  • Which mix attracted the highest number and diversity of wild bees
  • Identify which wildflower species were particularly important

What did they do?

The researchers created two new mixtures of wildflowers:

  • Wild Bee ‘WB’ – 16 plants chosen based on the authors’ previous research into wild bees
  • Literature ‘LT’ – 17 plants chosen after reviewing wider scientific literature

Each novel mix included four annual cornfield species (cornflower, corncockle, poppy and corn marigold) to act as a nursery in the first year. The mix contained plants that flower at different times of year to ensure resources from spring until late summer.

These two mixes were planted alongside plots of two standard flower mixes that are widely used in agri-environment schemes: 

  • Fabaceae legumes ‘FAB’ – 6 plants, predominantly clovers and vetches
  • Wild flowers ‘WF’ – 12 plants commonly sown as a mix in margins and plots

All the mixes had a ratio of 20:80 wildflowers to grasses, except the ‘FAB’ mix, which had 100% wildflower seeds.

The study took place in a single field on two farms in Oxfordshire and West Sussex, each with different soil characteristics and management. The four mixes and a fallow control were established in 20x5m plots. Each treatment was replicated five times, resulting in a total of 25 plots on each farm. All plots were managed in a similar way, but the cutting regime varied between farms depending on their individual circumstances.

Insects on each plot were identified and counted by walking the same route every 2-3 weeks between April and August in 2019, 2020 and 2021. The numbers and types of flowers within 2 metres on either side of the route were also recorded.

What they found

Over 4,000 insects were observed and identified during the project, from a variety of families: 28% of insects spotted were flies; 20% were beetles; 20% were bumblebees; 15% were solitary bees; 9% were hoverflies; and the remainder were honeybees, butterflies and solitary wasps.

A total of 79 different flower species were recorded in the plots, with 28 of those sown and 51 originating spontaneously from the existing seedbank in the soil.

Insects were observed visiting 55 flower species. Dandelion, spear thistle, wild carrot and hedgerow cranesbill received the highest numbers of wild bee species visits. 

Key plant species

Just 11 ‘key’ wildflower species were found to cater for all wild bee species recorded during the study. In order of popularity, these were dandelion, spear thistle, wild carrot, hedgerow cranesbill, smooth hawksbeard, rough hawkbit, hedge mustard, alsike clover, viper’s bugloss, common poppy and scentless mayweed.

Which mix is the winner?

The winning mix ‘WB’
Wild carrot
Hedgerow cranesbill
Rough hawkbit
Common poppy
Field scabious
Meadow cranesbill
Kidney vetch
Field bindweed
Musk mallow
Rough chervil
Corn marigold

The novel wild bee mix ‘WB’.

This mix, based on previous wild bee research, had more of the ‘key’ flowers per plot than any other mix and attracted the highest abundance and richness of wild bees compared to the others. It also had the highest total insect abundance, suggesting it was attractive to other taxa as well.

The legume mix ‘FAB’, standard wildflower mix ‘WF’ and mix based on literature ‘LT’ all had significantly fewer solitary bees and bumblebees than the ‘WB’ mix. Furthermore, the abundance of wild bees in these three lesser-performing mixes was similar to the fallow control plots.

What it means

The novel ‘WB’ mix of 16 species of wildflowers based on previous research was the most successful mix in terms of abundance and richness of wild bees, and it also attracted the highest total insect abundance.

‘WB’ plots held a significant number of the top 11 sown and spontaneous plant species for wild bees. Many of these are in the daisy family, which have high quantities of nectar and pollen due to their dense flowerhead containing lots of tiny florets. Another favoured group are umbellifers such as wild carrot, whose flowers are grouped en masse in umbrella-like flowerheads, and they provide large overall quantities of nectar. Plants in the daisy and carrot families therefore may be attractive because they save flight time and energy expenditure for adult foraging insects.

This study has demonstrated that a flower mix created through primary research performed well in terms of insect resources and visits, despite being sown on two farms with different soil types, management histories, and different seed banks. The species chosen were very attractive to their target insects, which encourages farmers to continue conservation efforts, and it also saves money and time by planting only the most useful species. Incorporating this research-based selection of flower species into agri-environment schemes should be considered if we are to better support the wider pollinator community.


Based on the findings of this research, authors recommend that:

  • The optimum wildflower mix for wild bees based on this and previous related studies should contain: alsike clover, kidney vetch, birds’-foot trefoil, wild carrot, hedgerow cranesbill, dandelion and rough hawkbit, with the addition of poppy and corn marigold as annuals.
  • Cornfield annuals are added to future wildflower seed mixes, because species such as common poppy and corn marigold proved attractive to wild bees and other insect pollinators. They can also act as a ‘nursery’, allowing the perennials to establish during the first year and suppressing the growth of weeds.
  • Legume-heavy pollinator mixes such as ‘FAB’ trialled in this study (and that are widely used by farmers) should be updated to include a broader range of species. The mix not only lost floral abundance rapidly as grasses invaded, but it performed poorly in terms of overall abundance of insects and wild bees.

Read the original paper 

Nichols, R.N., Goulson, D., & Holland, J.M. (2023). A novel farmland wildflower seed mix attracts a greater abundance and richness of pollinating insects than standard mixes. Insect Conservation and Diversity, 62: 190-204.

Thank you for reading this item. The GWCT conducts leading research, challenges misinformation and promotes effective strategies in the countryside. We are a small charity and every donation can make a big impact. It's quick and all cards, Direct Debit, Apple Pay, Google Pay and PayPal are accepted: