British phenological records indicate high diversity and extinction rates among late-summer-flying pollinators – a recently published study

Balfour et al Figure 1

Natural history records of plant flowering and pollinator foraging, much of them collected by well informed amateurs, have huge scientific importance. One of the values of such records to ecology is that it allows us to document where these species occur in space and when they are active in time. This can be done at a range of spatial and temporal scales, but large-scale patterns (for example at a country level) are, I think, especially useful because they provide scientific evidence that can inform national conservation strategies.

During 2017 I collaborated with a young early career researcher at the University of Sussex, Dr Nick Balfour, on an analysis of the phenologies of British pollinators and insect pollinated plants.  That study was recently published (see citation below) and I think that the results are fascinating.

Nick did most of the leg work on this, which involved assessing more than one million records that document the activity times of aculeate wasps, bees, butterflies and hoverflies held in the databases by three of the UK’s main insect recording organisations, the Bees, Wasps and Ants Recording Society (BWARS), the UK Butterfly Monitoring Scheme (UKBMS) and the Hoverfly Recording Scheme (HRS).  Information on flowering times was taken from a standard British flora (Clapham et al. 1990 – Flora of the British Isles. Cambridge University Press).

As well as looking at annual flight periods and flowering trends for these organisms we also focused on pollinator and plant species that were endangered or extinct. Here are some headline results and thoughts on what the work shows:

  • About two-thirds (62%) of pollinator species peak in their flight times in the late summer (July and August), though there was some variation between the different groups – see the figure from the paper above).  Particularly noticeable was the double peak of the bees, with the first peak denoting the activity of many early-emerging solitary bees, such as species of the genus Andrena, whilst the second peak is other solitary bees plus of course the bumblebees which by that time have built up their colonies.
  • A rather fixed phenological pattern with respect to different types of plants was also apparent, which I was not expecting at all: insect pollinated trees tend to flower first, followed by shrubs, then herbaceous species (again, refer to the figure above). This might be because larger plants such as trees and shrubs can store more resources from the previous year that will give them a head start in flowering the following year, but that idea needs testing.
  • Putting those first two points together, what it means is that trees tend to be pollinated by those earlier emerging bees and hoverflies, whereas the herbs are mainly pollinated by species that are active later.
  • When looking at the extinct and endangered pollinators, the large majority of them (83%) were species with a peak flight times in the late summer, a much larger proportion than would be expected given that 62% of all species are active at that time. However this was mainly influenced by extinct bee species and the same pattern was not observed in other groups.
  • The obvious explanation for that last point is that historical changes in land use have led to a dramatic reduction in late summer flowering herbaceous species and the subsequent loss of floral resources has been highly detrimental to those bees. But intriguingly no such pattern was apparent for the endangered pollinators and clearly there are complex reasons why pollinators should become rare or extinct, a point that I have discussed previously on the blog.
  • The lack of late summer flowering resources for pollinators is a contentious issue however as plant conservation groups have in the past recommend that meadows and road verges are cut in late summer to maximise plant species richness.  Mowing road verges once or twice a year certainly benefits plant diversity, as this recent review by Jakobsson et al. (2018) demonstrates.  But there’s very little data available that assesses how timing of cutting can affect pollinators.  The only study that I know of (and if I’ve missed any, please let me know) that has considered this is the PhD work of one of my former students, Dr Sam Tarrant who looked at pollinators and plants on restored landfill sites compared to nearby nature reserves.  In a paper that we published in the journal Restoration Ecology in 2012 we showed that on restored landfill sites the abundance of pollinators in autumn surveys (conducted September-October) was just as high as for summer surveys.  On nature reserves, which are routinely cut from mid-July onward, this was not the case.

Here’s the full citation of Nick’s study with a link to the publisher’s website, and a copy of the abstract is below.  If anyone wants a PDF, drop me a line:

Balfour, N., Ollerton, J., Castellanos, M.C., Ratnieks, F.L.W. (2018) British phenological records indicate high diversity and extinction rates among late-summer-flying pollinators. Biological Conservation 222: 278-283


The long-term decline of wild and managed insect pollinators is a threat to both agricultural output and biodiversity, and has been linked to decreasing floral resources. Further insight into the temporal relationships of pollinators and their flowering partners is required to inform conservation efforts. Here we examined the
phenology of British: (i) pollinator activity; (ii) insect-pollinated plant flowering; and (iii) extinct and endangered pollinator and plant species. Over 1 million records were collated from the historical databases of three British insect monitoring organisations, a global biodiversity database and an authoritative text covering the national flora. Almost two-thirds (62%) of pollinator species have peak flight observations during late-summer
(July and August). This was the case across three of the groups studied: aculeate wasps (71% of species), bees (60%), and butterflies (72%), the exception being hoverflies (49%). When species geographical range (a proxy for abundance) was accounted for, a clear late-summer peak was clear across all groups. By contrast, there is marked temporal partitioning in the flowering of the major plant groups: insect-pollinated tree species blossoming predominantly during May (74%), shrubs in June (69%), and herbs in July (83%). There was a positive correlation between the number of pollinator species on the wing and the richness of both flowering insect pollinated herbs and trees/shrubs species, per calendar month. In addition, significantly greater extinctions occurred in late-summer-flying pollinator species than expected (83% of extinct species vs. 62% of all species). This trend was driven primarily by bee extinctions (80% vs. 60%) and was not apparent in other groups. We contend that this is principally due to declines in late-summer resource supplies, which are almost entirely provisioned by herbs, a consequence of historical land-use change. We hypothesize that the seasonality of interspecific competition and the blooming of trees and mass-flowering crops may have partially buffered spring flying pollinators from the impacts of historical change.

11 thoughts on “British phenological records indicate high diversity and extinction rates among late-summer-flying pollinators – a recently published study

  1. Martin

    Tricky to comment because I think it simplifies things that ought not to be (but obv. has to be done for the sake of hypothesis, idea discussion, and importantly a short abstract).
    Is your measure total floral resources available in late-summer or total number of species that flower in late summer – the graphs are small but I think they are number of species? because the former may be important for some pollinators but the latter more important for particular pollinators (i.e. if the losses were of specialist feeders and so therefore were they associated with the loss of a particular plant species?). Similarly there could be more total floral tree resources now than there was at the start of your study (as less coppicing occurs now and there has been more recent tree-planting which could have increased total tree floral resources).
    Your study starting point is not the start of the only major agricultural change in England; the major losses of species-rich permanent grassland occurred during enclosure (as did much of the wetland and lowland heath) habitats which contained the majority of (in terms of quantity) the late-flowering herbaceous species and so some pollinator species may have been on a downward trend because of that and the final finishing point for them was where your study investigated.
    The lack of late summer-flowering resources is not contentious at all unless you have evidence from quantitative surveys that cutting road verges in late summer is occurring on a large enough scale to affect late-summer floral resources; evidence of campaign/advice is not the same as evidence of application of campaign/advice by councils. Also, the vast majority of floral resources in meadows will (depending on where you are in the UK) be gone by mid-July (there are early flowering ecotypes of some species, like common knapweed, which are found in meadows whereas the later flowering ones are found in grasslands originally derived from communal medieval grazing pasture).
    Similarly, Landfill sites can be very good for pollinators but much depends on the quality of the restoration work and its subsequent management rather than it being intrinsically a land-fill site.
    There are also outlier issues that would need a mention – loss of nesting/overwintering sites, loss of larval food sites, fragmentation of feeding/overwintering sites and genetic instability within smaller isolated populations – all of which are probably mentioned in the paper but not in the abstract (which is all I’ve read and usually all I get to read and I comment on that just to point out its modern-day importance regarding science communication to the interested non-academic digital reader).

    1. jeffollerton Post author

      Hi Martin – thanks for your comments. If you’d like a copy of the papers I’m happy to send them to you, just email me.

      Yes, clearly there’s not enough room in a blog post to deliver all of the detail, I use these posts simply to highlight that the work is out there.

      To answer your specific questions:

      (1) The measure of floral resources is number of species. Relative abundance is very variable between years and for most species we don’t know what they offer in terms of amount and quality of nectar and pollen. But note that there is a positive correlation between plant and pollinator diversity.

      (2) Enclosure certainly caused changes but whether it affected regional plant diversity & pollinator diversity is debatable. We simply don’t have good data from that period. Enclosure certainly didn’t have as great an effect as the changes from the early 20th century – see:

      (3) I meant contentious with respect to the advice that’s being given. But note from our landfill paper we know that nature reserves tend to be mown or grazed rather early in the year. I disagree with you about most floral resources being gone by mid-July though: it very much depends on the weather and in any case those late flowering plants that are still in flower become much more important for pollinators such as queen bumblebees and some butterflies and hoverflies that hibernate.

      (4) Agreed, landfill sites vary hugely in their quality. But our paper shows that even those that have been allowed to naturally colonise can support significant numbers of pollinators. As quite extensive and relatively undisturbed expanses of grasslands they certainly have conservation value.

      (5) Agreed, the other factors you mention are also important but it’s not possible to address all of these in a single study. However other studies have looked at these issues and collectively it forms a mosaic of research evidence.



  2. Emily Scott

    Interesting. The only flowers I have out in my garden at the moment are Japanese anemone, mint, oregano and echinops – think I need to get some more late flowering plants in.

  3. Miles King

    very interesting research Jeff.

    I’ve also often wondered what the late summer flowering resource would have been before the large-scale loss of flowers from the agricultural landscape, most of which took place during hte 20th century. Current practice recommends leaving flower-rich grasslands (including road verges) uncut until late in the season. But this would definitely not have happened under agricultural management. Quite the opposite.

    Historic management of hay meadows meant the hay was cut when it was at its most nutritious – while this would have varied from year to year, by altitude and regionally, hay cuts were taking place in June and July, but no meadow would ever have been left uncut until September. flower-rich pastures, including downlands, heathy grassland and neutral pastures would also have been grazed through the spring and into summer – reducing the amount of flowers available for insects, relative to modern conservation practice. Having said that, of course there would still have been a large resource of flowers, relative to the flower-free perennial rye grass pastures and silage fields of today. Later-flowering plants such as knapweeds and scabious species for example may have had the opportunity to flower in pastures where later summer grazing pressure tailed off (or where stock were moved from summer pastures) but this doesn’t quite feel right. Subsistence farmers use a resource up to the point where it is no longer usable, nothing is wasted. heathland affords more obvious late flower resources and it’s easy to forget how much more common heathland or heathy grassland was in agricultural landscapes, 100 years ago, than it is now.

    Two other thoughts come to mind. Several plants have second flowerings after they have been cut for hay or grazed off in spring and summer. Critically important plants for pollinators such as red clover and bird’s-foot trefoil will happily flower through into early autumn.

    The other thought is that there was a another potentially very important pollen and nectar resource – arable weeds. Before every field was sprayed to the point where no weeds now survive, arable weeds would have been plentiful in arable fields which had been harvested, or were lying fallow. Those species with shorter lifecycles could germinate flower and set seed after the harvest and before the autumn, while others eg the speedwells just carry on flowering as long as possible. Even arable weeds of the crop (the segetals) would have carried on flowering on headlands after the harvest.

    1. jeffollerton Post author

      Thanks for the comment Miles. I had similar thoughts with regards to meadows when I read George Peterken’s recent book, and also with regards to the absence of hedgerows across large areas per-enclosure.

      But I think you’ve answered the question already in the sense that the pre-industrial agricultural landscape that you are describing was a much more complex, heterogeneous, and florally rich than the kinds of landscapes we see in intensively farmed regions. So flower visiting insects could usually find resources of some sort somewhere in the landscape. I suspect that riparian vegetation was important in this regard for instance. If only we had a time machine!

  4. Miles King

    Thanks Jeff. We do have time machines of a kind. John Clare’s poetic descriptions of pre- (and post-) enclosure landscapes tell us they were full of flowers. Richard Jefferies describes the 60 different types of flower he found along one suburban lane (in Surbiton, Surrey.)

    As you say, for the average insect with the capacity to forage over an area of perhaps a hundred acres of farmland (including hedge, arable, pasture, meadow, riverbank, pond edge, hedgerow and woodland), there would have been a flower resource of one kind or another, right across the seasons.

  5. Bendifallah

    Hi Jeff,
    Interesting research!!!
    I would like a copy of the paper please. I have published a similar paper but on wild bees in Northern Algeria: FLOWERING PLANTS PREFERRED BY NATIVE WILD BEES (HYMENOPTERA, APOIDEA, APIFORMS) IN THE ALGERIAN LITTORAL REGION
    L. Bendifallah and Francisco Javier Ortiz-Sánchez.

    Kind regards
    Prof Leila Bendifallah
    Head of Department of Agronomy
    Faculty of Sciences
    University M’hamed Bougara, Boumerdes, Algeria


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s