Category Archives: Biodiversity

Bees, moths, cockroaches and some unfortunate flies: the surprising pollination of a European milkweed

Back in October 2021, I wrote a post about tracking down a rather elusive plant on the Danish coast. Karin and I had been cycling around the Odsherred region when I spotted the distinctive yellowing autumn foliage of Vincetoxicum hirundinaria growing close to the little fishing harbour at Klint.

I was particularly pleased to find it because Vincetoxicum is a genus of plants that I’ve been interested in for many years. It’s part of the milkweed and dogbane family, Apocynaceae, and belongs to the subfamily Asclepiadoideae, a group with some of the most extraordinarily complex flowers in the plant kingdom.

At the end of that post I asked a question: what was pollinating those flowers on the exposed Danish shoreline? At the time there were surprisingly few published records of pollinators for this widespread European species.

Five years later, I’m delighted to say that a new paper just published in AoB PLANTS, led by my Chinese colleague Wujian Xiong, and to which I’ve contributed, goes a considerable way towards answering that question. Not specifically for the Danish population, but for several populations studied in and around Salzburg, Austria. And the answer is rather more complicated, and interesting, than we might have expected.

A very specialised flower with remarkably unspecialised tastes

The flowers of Vincetoxicum hirundinaria are small, pale cream or ivory-coloured, and about 12 mm across. Like other members of the Asclepiadoideae, they have a highly specialised mechanism for transferring pollen.

Rather than producing individual loose pollen grains, the pollen is packaged into structures called pollinia, which are joined together by a central attachment mechanism to form a pollinarium. When an insect visits a flower, part of its body can become caught in a narrow guiding structure. As it pulls away, a pollinarium may become attached to its mouthparts, legs or other appendages. If all goes well, the insect carries that pollen to another flower and deposits it into a specialised receptive chamber.

It’s an extraordinarily precise mechanism, and one might reasonably expect that only a rather narrow range of insects could operate it successfully.But that’s not what we found.

Over 50 hours of observations, Wujian recorded almost 3,000 visits to the flowers, involving 81 species or morphospecies of insects from seven different orders. These included bees, wasps, ants, flies, beetles, moths, bugs, scorpionflies and cockroaches.

Of those 81 types of flower visitor, 37, representing five insect orders, were found carrying pollinaria. These included 15 types of flies, seven bees and wasps, seven moths, six ants, one beetle and one cockroach. That’s a remarkably diverse assemblage for a plant with such specialised floral structures. Indeed, it’s the most diverse pollinator assemblage so far documented for any member of Vincetoxicum.

What makes this particularly interesting is that the pollinaria were attached to different parts of the insects’ bodies. In flies they might be on the proboscis or legs; in bees and wasps on the mouthparts; in moths on the proboscis; and in cockroaches on the labial palps.

In other words, the same elaborate floral mechanism can operate with insects that differ enormously in their anatomy, size and behaviour.

Another chapter in the story of cockroach pollination

There’s a particularly satisfying connection here with another paper that Wujian and I published back in 2020. That study demonstrated specialised cockroach pollination in the rare and endangered Chinese species Vincetoxicum hainanense. Its small flowers open at night, produce an unpleasant scent reminiscent of rotting fruit, and are principally pollinated by a particular species of cockroach, Blattella bisignata.

It was the first documented example of specialised cockroach pollination in the Apocynaceae, and one of only a handful of well-studied examples from anywhere in the world. Now we have another species of Vincetoxicum that is visited and potentially pollinated by cockroaches, but in this case they are just one component of a much more diverse assemblage.

The contrast between the two species is striking. They share the same basic type of highly specialised floral architecture, yet one has a relatively narrow pollination system while the other is much more generalised. Clearly, floral complexity doesn’t necessarily translate into ecological specialisation.

Day and night, and the importance of place

Another interesting finding concerns when the flowers are available to pollinators. The flowers of V. hirundinaria open in the late afternoon, between about 4pm and 6pm, and remain open for an average of six days. They produce nectar and a strong, sweet scent that we describe in the paper as reminiscent of red wine.

That long floral lifespan means that they are potentially accessible to both daytime and nocturnal insects. But the relative importance of those insects varied considerably between the study populations. At most sites, pollen removal and deposition took place during the day, with bees, wasps and other diurnal insects playing the main roles. However, at the higher-elevation Gaisberg site, nocturnal moths and cockroaches were much more prominent, and pollen transfer during the night was particularly effective.

It’s a useful reminder that a plant doesn’t necessarily have one fixed set of pollinators across its geographical range. The animals available to visit flowers can change with location, habitat, season and time of day. This connects with some of the arguments that I’ve made recently about geographical variation in plant–pollinator specialisation. Ecological specialisation isn’t always an intrinsic, unchanging characteristic of a species, it can depend very much on the circumstances in which that species is studied.

But not every visitor is a useful pollinator

There’s a rather darker side to this story, however: the flowers of Vincetoxicum hirundinaria can be surprisingly dangerous places for small insects. We recorded 16 types of flies that became trapped by the flowers, often fatally, when their mouthparts became lodged in the structures associated with the pollinarium. Some were simply too small or weak to pull themselves free.

This phenomenon is known from other milkweeds, including the notorious moth-catching vine Araujia sericifera. However, the extent of insect trapping that we observed in V. hirundinaria was quite remarkable. There’s an important distinction here between attracting an insect to a flower and actually using it to transfer pollen successfully, even for those species that don’t get trapped.

Ants, for example, accounted for about a quarter of all recorded flower visits, but only 3.5% of the sampled ants carried pollinaria. In contrast, between 40% and 60% of individuals in several other insect groups carried pollinaria.

Even carrying a pollinarium doesn’t necessarily mean that an insect will successfully deposit pollen into another flower. That requires the insect to visit a suitable flower and position its body in just the right way. I’ve discussed this issue before in relation to what we actually mean by a pollination system. Counting floral visitors is relatively straightforward. Demonstrating that those visitors contribute to successful plant reproduction is much more difficult.

In the present study, the evidence for functional generalisation is strong in terms of the diverse insect groups capable of removing and transporting pollinaria. However, we cannot yet say precisely how much each group contributes to successful pollen deposition and seed production.

Indeed, overall pollen transfer efficiency was quite low. At most sites only about 3–9% of the pollen removed from flowers was successfully transferred to another flower’s receptive structures during the day. At Gaisberg, nocturnal pollen transfer efficiency averaged around 11%. Nor did the plants produce many fruits. Only 2% of naturally pollinated flowers produced mature fruits, compared with 14% following experimental cross-pollination. The plants were also self-incompatible in the populations we studied, meaning that pollen transferred between flowers of the same individual wouldn’t result in successful reproduction.

So despite attracting a remarkably diverse range of insects, V. hirundinaria appears to have substantial limitations on its reproductive success. Pollination isn’t simply a numbers game.

Back to the bigger picture

Readers of this blog will know that the diversity and evolution of pollination systems in Apocynaceae has been a major theme of my research for many years. In our large international review published in 2019, we brought together information on the pollinators of 567 species and showed that the family encompasses everything from extremely specialised relationships to highly generalised systems involving many different kinds of animals.

More recently, I’ve written about the extraordinary pollination biology of Heterostemma ficoides, which deceives fig wasps while providing genuine breeding opportunities for tiny flies. And only last week I discussed our new study of eleven co-occurring Apocynaceae species in Argentina, showing how very different flowers can share overlapping assemblages of visitors.

The new Vincetoxicum study adds another piece to this evolutionary and ecological puzzle. It demonstrates that even one of the most mechanically sophisticated types of flower can function with a very wide range of potential pollinators. Floral specialisation and pollinator specialisation are not the same thing, and we shouldn’t assume that one necessarily predicts the other.

But it also raises some intriguing questions. Why do the flowers attract so many insects that seem poorly suited to pollinating them? Is this simply an unavoidable consequence of producing nectar and scent signals that appeal to many different animals? Or might the presence of trapped insects actually reduce reproductive success by blocking access to the flowers’ reproductive structures?

And, returning to where I started, what exactly is pollinating those plants on the Danish shoreline at Klint? Are they visited by the same diversity of insects as the Austrian populations, or has their rather exposed coastal environment resulted in a very different assemblage? I still don’t know the answer to that question. But it would be fascinating to find out.

One of the things I enjoy most about working on the Apocynaceae is that each new study seems to reveal another layer of complexity in this extraordinary family. Almost 40 years after I began studying their flowers and pollinators, they continue to surprise me. My thanks to Wujian and all of our colleagues for helping to unravel another part of that story.

Here’s the full reference, with a link to the open-access paper:

Xiong, W., Ollerton, J., Liede-Schumann, S., Meve, U., Wagner, H.C., You, W., Mu, J. & Dötterl, S. (2026) Generalized pollination of morphologically specialized Vincetoxicum hirundinaria (Apocynaceae, Asclepiadoideae). AoB PLANTS 18(5): plag048. https://doi.org/10.1093/aobpla/plag048

Images above (c) Wujian Xiong

Moths, wasps, hummingbirds and nectar: exploring the pollination ecology of Apocynaceae in Argentina

1 Reply

One of the things that has fascinated me throughout my career is the extraordinary diversity of ways in which flowering plants interact with their animal visitors. Nowhere is this better illustrated than in the milkweed and dogbane family, the Apocynaceae, a group that I’ve been studying for almost 40 years.

Only last month I wrote about the remarkable pollination biology of Heterostemma ficoides, a species from Thailand that deceives fig wasps while offering a genuine breeding site to tiny flies. That was an example of an unusual pollination system involving a single species. But what happens when we look at a whole community of Apocynaceae growing together in the same habitat? Do they share pollinators, or have they evolved different ways of attracting and using animals?

Those are some of the questions that my Argentinian colleagues Carolina Torres and Leonardo Galetto and I have addressed in a new paper just published in the journal Taxon. The study focuses on eleven species of Apocynaceae growing together in the Chaco forests of central Argentina, a region characterised by strongly seasonal rainfall and a distinctive assemblage of plants and animals.

What’s particularly interesting is that the original field observations were made by Carolina and Leonardo over three flowering seasons between 1997 and 2000. We have now brought those data together, including detailed measurements of floral nectar, and added analyses of the networks of interactions between plants and their visitors. It’s a nice example of how carefully collected field data can continue to yield new insights many years later.

Different flowers, different rewards

The eleven species represent three major evolutionary lineages within the Apocynaceae: one rauvolfioid, three apocynoids (Mandevilla species), and seven members of the Asclepiadoideae, the group formerly recognised as the separate family Asclepiadaceae.

Their flowers vary enormously in size and structure. At one extreme are the relatively small flowers of species such as Araujia, Oxypetalum and Philibertia. At the other are the large, tubular flowers of Mandevilla, including M. petraea, in which the functional floral tube is more than 10 cm long.

These differences are reflected in the nectar that the flowers produce. The three Mandevilla species generally offer relatively large quantities of nectar containing substantial amounts of sugar, predominantly sucrose. They are visited by animals capable of accessing these rewards, including long-tongued hawkmoths, bumblebees and, in the case of M. pentlandiana, hummingbirds. By contrast, most of the smaller-flowered species produce much less nectar and show greater variation in the proportions of the three main nectar sugars: sucrose, glucose and fructose.

But these relationships are not quite as straightforward as we might expect. For example, two closely related species, Araujia brachystephana and A. odorata, have rather similar-looking flowers, yet the former produces nectar dominated by sucrose while the latter produces nectar consisting almost entirely of glucose and fructose. Both are visited primarily by wasps, although A. brachystephana attracts a broader range of insects.

In other words, even closely related species living in the same habitat may differ substantially in the rewards they offer and the animals that visit them.

There’s another intriguing result concerning the tree Aspidosperma quebracho-blanco. Previous research described its flowers as deceptive, attracting visitors without offering nectar. Yet in our study we found small quantities of nectar containing mainly glucose and fructose. Does nectar production vary between populations, or were there differences in the methods used to study the flowers? It’s an interesting question that deserves further investigation.

Sharing the flowering season

The timing of flowering is another important part of this ecological story. Most of the species flowered during the warmer, wetter months, particularly December and January, with flowering often continuing for three or four months. Seed dispersal, by contrast, was concentrated mainly between February and June, when conditions become progressively drier.

This makes ecological sense. Water availability is likely to be an important trigger for flowering in this seasonal environment, while the wind-dispersed seeds of these species are released mainly during the drier part of the year. But it also means that many of the species are flowering simultaneously and potentially competing for the attention of the same animals.

One possibility is that plants avoid this competition through differences in flowering time. Another is that they coexist by attracting different groups of pollinators, or by using the same pollinators in subtly different ways. For example, two Mandevilla species may share hawkmoth visitors but place their pollen on different parts of the moth’s proboscis, potentially reducing interference between their reproductive systems.

Among the smaller-flowered asclepiads, differences in the intricate structures that attach pollen masses to visiting insects may perform a similar function. These are fascinating examples of how floral morphology can influence ecological interactions, even when the same animals visit several plant species.

What do the interaction networks tell us?

As readers of this blog will know, I’ve been interested in plant–pollinator networks for many years. One reason for analysing these networks is that they allow us to move beyond individual plant species and examine how an entire assemblage of organisms is connected.

In the Chaco study, we found a mixture of apparently specialised and generalised relationships. Some species were visited by just one major functional group, such as wasps or butterflies, while others attracted a much wider range of insects. Wasps were particularly conspicuous. Indeed, two social wasps, Polybia ignobilis and Brachygastra lecheguana, were among the visitors that connected different parts of the interaction network. The introduced western honey bee, Apis mellifera, also played a connecting role.

It’s another reminder that when we talk about pollinators, we really need to look beyond bees. Wasps, flies, moths, butterflies and other animals can be important components of these ecological communities.

One intriguing outcome of our analyses concerns the overall structure of the network. At first glance, the interactions appear to be organised into distinct groups, or modules, with particular sets of plants sharing particular sets of visitors. However, when we compared the observed network with randomly generated networks that preserved the number of interaction partners of each species, neither its modularity nor its nestedness was significantly greater than expected.

In other words, although some groups of interactions look rather specialised, there is limited statistical evidence for strong compartmentalisation within this particular assemblage. The patterns may largely reflect differences in how many partners each species has, rather than the existence of clearly separated ecological groups.

There is an important qualification here. Our network includes only eleven plant species, all from the same family, rather than the full diversity of flowering plants in the Chaco forests. A more comprehensive network might reveal additional structure that we cannot detect at this scale.

This connects rather neatly with my recent discussion of global patterns in plant–pollinator specialisation. One of the lessons from that research is that specialisation depends partly on the spatial, taxonomic and ecological scale at which interactions are examined. What looks specialised at one scale may be much more generalised at another.

I should also emphasise that this is a network of floral visitors, not necessarily effective pollinators. We recorded animals capable of contacting the reproductive structures of the flowers, and of carrying pollinaria, but that does not establish that they successfully transferred pollen between plants. As I discussed in an earlier post about what we actually mean by a pollination system, that distinction really matters.

Back to the bigger picture

For me, this study makes an interesting contribution to the broader understanding of Apocynaceae pollination ecology that my colleagues and I developed in our 2019 global review of the family’s pollination systems. In that paper, involving 75 authors, we showed that Apocynaceae exhibit an extraordinary diversity of pollination systems, with strong evolutionary and biogeographical patterns. But we also emphasised how much remains unknown, with reliable pollination information available for only a little over 10% of the family’s species. The new Chaco study adds something that large-scale reviews cannot provide on their own: a detailed picture of how related species interact with their floral visitors within the same ecological community.

There’s also a connection to our recent review of edible Apocynaceae, which also involved Carolina and Leonardo. One of the striking results of that study was how little we know about the pollination biology of many species whose fruits, seeds or other parts are used by people. Understanding how these plants reproduce is important not only from an evolutionary perspective but also for their conservation and sustainable use.

Ultimately, what I find most interesting about this new paper is the contrast between the extraordinary diversity of floral forms and rewards, and the extent to which different species can share the same visitors. Specialised flowers do not necessarily mean exclusive relationships with particular pollinators. And generalised patterns of visitation do not mean that floral morphology and nectar chemistry are unimportant. The reality is more complicated, shaped by evolutionary history, seasonality, floral structure and the behaviour of the animals themselves.

It’s also a reminder of the continuing value of detailed natural history. Large-scale databases and sophisticated network analyses are enormously useful, but they depend ultimately on people spending time in the field, watching what visits flowers, measuring nectar, and recording when plants flower and produce seeds. Sometimes those observations take decades to reach their full potential. But that doesn’t make them any less valuable.

Here’s the full reference, with a link to the paper:

Torres, C., Ollerton, J. & Galetto, L. (2026) Reproductive phenology, flower-nectar traits, and floral visitors of eleven co-occurring species of Apocynaceae from Chaco forests (Argentina). Taxon 75(5): e70279. https://doi.org/10.1002/tax.70279

Listen to Karin and I on The Gardener’s Lodge podcast: pollinators, gardening and relationships

Leave a reply

Back in the summer, Karin and I were interviewed, one after the other, by garden designer and podcaster Mykal Hoare for his The Gardener’s Lodge podcast. Both interviews have now been released, covering rather different aspects of gardening and our relationship with the natural world.

My interview, published on 17 September, was a wide-ranging conversation about pollinators, their relationships with flowering plants, and the role that gardens can play in supporting biodiversity. We discussed everything from bees and butterflies to birds, the effects of climate change and this summer’s unusually hot, dry weather, and the perennial question of whether gardeners should favour native or non-native plants. We also talked about why providing flowers is only part of the story when it comes to conserving pollinators.

Karin’s interview, published on 1 October, approached gardening from a very different perspective. Drawing on her experience as a relationship therapist and author, she and Mykal explored how gardens and gardening can contribute to mental wellbeing, mindfulness and stronger relationships. In particular, they discussed how gardening together can reveal differences between couples, provide opportunities for communication and compromise, and strengthen our connections with nature. Apparently, our own sometimes contrasting ideas about how tidy or wild a garden should be provided some useful examples!

There’s quite a bit of common ground between the two conversations, particularly around the importance of observing and appreciating the wildlife that shares our gardens.

Some readers may remember that Karin has spoken about gardening and relationships on a podcast before, so this is an interesting continuation of that discussion.

Many thanks to Mykal for inviting us both onto the podcast and for two very enjoyable conversations!

Are plant-pollinator interactions more specialised in the tropics? A new global analysis suggests that it’s complicated…

2 Replies

Some scientific questions have a habit of following you around for much of your career. Back in 2002, Louise Cranmer and I published a paper in the journal Oikos asking a deceptively simple question: are tropical plants more specialised in terms of their pollinators?

At the time there was a widespread assumption in ecology that interactions between species become increasingly specialised as you move from the poles towards the tropics. It’s an attractive idea which had some (quite vague) theoretical support. Tropical ecosystems contain more species, have experienced relatively stable climates over long periods of evolutionary time, and seem like just the sort of places where tightly specialised interactions ought to evolve.

The trouble was that the evidence wasn’t especially convincing, in part because it hadn’t actually been well studied. In preparing the 2002 paper, Louise and I found only a handful of empirical studies that had addressed the question, not just for plants and pollinators, but for ALL kinds of interactions.

The question of tropical specialisation has subsequently cropped up repeatedly in my research. In 2014, Angela Moles and I went considerably further and asked whether the broader claim that species interactions are stronger and more specialised in the tropics might actually be a “zombie idea” – one of those scientific notions that persists despite accumulating evidence that the reality is considerably more complicated. That guest post on the Dynamic Ecology blog was then written up as an invited paper for Biotropica see: Tropical Zombies: Moles & Ollerton (2016) is now published.

Twenty-four years after that first Oikos paper, we can now address the question with a dataset that Louise and I could scarcely have imagined in 2002. A huge international collaboration led by Sailee Sakhalkar and Robert Tropek has just published “Global patterns in plant–pollinator specialization” in Nature Ecology & Evolution.

I’m delighted to be one of the co-authors and I find the answer to the question that we posed 24 years ago to be fascinating.

More than 110,000 interactions

The scale of the new study is worth emphasising. It assembled 3,415 quantitative plant-pollinator networks from 162 studies, covering 110,571 pairwise interactions between 5,343 pollinator species and 6,126 plant species. The data span almost all of the world’s terrestrial biogeographic regions, from 43.6°S to 81°N and from sea level to more than 4,000 metres elevation.

That makes this the most comprehensive attempt yet to examine the geography of specialisation in plant-pollinator interactions. But simply throwing thousands of networks into a statistical analysis would potentially create as many problems as it solves. Different researchers study pollination networks in very different ways. Some attempt to sample whole communities, while others concentrate on particular groups such as bees, hummingbirds or particular plants.

So an important feature of this study was that we explicitly separated these different types of datasets and accounted for sampling completeness and methodological differences. We also looked separately at different functional groups of pollinators rather than assuming that bees, beetles, butterflies, moths, flies and birds should all behave in the same way. That turned out to be important in ways that we could not predict.

So, are interactions more specialised in the tropics?

The short answer is no, not in any simple sense. There is geographical structure to plant-pollinator specialisation. But there is no straightforward increase in specialisation as you approach the Equator – nor, for that matter, a straightforward decrease. Instead, the patterns are distinctly nonlinear.

Network-level specialisation and pollinator specialisation tended to peak at relatively low northern latitudes, at around 20°–30°N, close to the tropical-subtropical boundary. I find this particularly interesting as I think the subtropics have been a bit neglected as a focus of ecological research: no one ever refers to themselves as a “subtropical ecologist” and there is no Journal of Subtropical Ecology. Yet the subtropics are not simply a transitional zone between the tropical and temperate worlds. They encompass an extraordinary range of ecosystems, including deserts, Mediterranean-type vegetation, subtropical forests, and some major centres of biodiversity and endemism. This includes the distinctive habitats in Yunnan that I’m exploring during my visits to China. Perhaps we need to start thinking of the subtropics as an important biogeographical region in their own right, rather than as the ecological space between two better-studied zones?

Plant specialisation showed a different and weak hemispherically asymmetric pattern. And once individual pollinator groups were examined separately, the picture became more complicated still. Hoverflies, for example, came closest to the traditional expectation of increasing specialisation towards the Equator. Birds showed something closer to the opposite pattern, while moths did something much more complicated.

In other words, there isn’t a latitudinal specialisation gradient. There are several overlapping geographical patterns whose shapes depend upon what organisms you study and what measure of specialisation you are considering.

That conclusion feels satisfying given the arguments Angela Moles and I made ten years ago. We suggested that attempts to find a universal latitudinal gradient could obscure the fact that different groups of organisms have different evolutionary histories, physiologies and ecologies. That is very close to what emerges from this much larger analysis.

Perhaps latitude was never really the interesting variable

One of the most intriguing results is what happens when we stop asking about latitude itself. We compared latitude with a suite of possible explanatory variables, including temperature, precipitation, environmental productivity, and plant and pollinator diversity. Climate generally did a better job of explaining specialisation than latitude, biodiversity or productivity. Again, however, there wasn’t a single climatic rule.

At the whole-network level, specialisation generally declined as mean annual temperature increased, which was not expected at all and runs counter to the argument that the tropics are always more specialised. Precipitation produced strongly nonlinear relationships, and plants and different pollinator groups responded differently. Birds, insects and even different groups of insects could show contrasting responses to the same climatic gradient.

There are some plausible biological explanations for these patterns. Warmer conditions may extend flowering periods and pollinator activity, creating more opportunities for species to encounter alternative partners and therefore encouraging generalisation. Colder or more seasonal environments might compress flowering and pollinator activity into shorter periods, producing tighter temporal matching between species. Rainfall adds another layer of complexity because it can influence flowering but also directly constrain pollinator activity.

These are hypotheses rather than demonstrated mechanisms, however. As we stress in the paper, these are correlational global patterns and climate covaries strongly with latitude. The results shouldn’t therefore be interpreted as demonstrating simple causal effects of temperature or rainfall.

Specialisation depends on how you look at it

There’s another connection here to research I’ve discussed previously on the blog. In work with Danish and Brazilian colleagues, for example, we showed that the apparent specialisation of plants and pollinators depends partly on the spatial scale at which interactions are examined – see Local and regional specialization in plant–pollinator networks. More recently I’ve written about why scale matters when analysing plant-pollinator networks.

And earlier this year, our global analysis of plant-pollinator interactions in gardens found that the factors explaining species richness were not necessarily those explaining interaction specialisation: Global drivers of plant-pollinator interaction specialization in gardens.

There’s also a nice connection to another study that I wrote about here in 2022. With Pablo Gorostiague and Pablo Ortega-Baes, I looked specifically at whether cactus pollination systems become more specialised towards the tropics. Our answer was yes and no: tropical cacti were visited by somewhat fewer pollinator species, but there was no latitudinal pattern in the number of functional pollinator groups that they used. As I wrote at the time, the answer depended on what we meant by “specialised”: Are cactus pollination systems more specialised in the tropics? A new study suggests yes…and no!

All of these studies point towards a similar conclusion: “specialisation” is not a single, fixed property of an ecological community. What we detect depends upon biological scale, geographical scale, the organisms being considered, and how the interactions have been sampled. That doesn’t make specialisation meaningless. Quite the opposite, it means we need to be precise about what sort of specialisation we’re talking about and at what scale.

The new global study puts that result into a much broader context. Different groups really can display different geographical patterns of specialisation, so studies focused on a particular lineage aren’t necessarily contradicting one another when they produce different answers. They may simply be measuring different pieces of a much larger ecological puzzle.

Why does any of this matter?

Specialisation is not just an esoteric property of ecological networks. It affects how plants and pollinators partition resources, how species coexist, how communities are assembled, and potentially how ecological systems respond when species disappear.

That becomes especially important in a changing climate. Our results suggest that changing temperature and rainfall regimes could restructure plant-pollinator networks, but they are unlikely to do so uniformly. Different pollinator groups and different parts of the world may respond in very different ways. Highly specialised interactions may also be particularly vulnerable to changes in phenology or the loss of interaction partners. Predicting the consequences therefore requires us to move beyond simple statements about “tropical” and “temperate” ecosystems.

For me, though, there is also a broader lesson here about how science progresses. In 2002, Louise Cranmer and I questioned whether the available evidence really supported the assumption that tropical plant-pollinator interactions were more specialised. Fourteen years later, Angela Moles and I argued that the wider claim about stronger and more specialised tropical interactions might be a zombie idea. Now, with more than 110,000 plant-pollinator interactions available for analysis, the answer is much clearer. Latitude matters; climate matters; the identity of the organisms matters; and ecology, as usual, refuses to obey a simple rule.

My particular thanks to Sailee and Rob for the monumental efforts they made to pull this work together, and cudos to all of the co-authors who helped to make the study possible.

Want to conserve birds and mammals? Then conserve their insects – our new paper in Nature Reviews Biodiversity sets out the details

1 Reply

It’s sometimes easy to forget quite how much of the living world depends upon insects.

That might seem an odd thing for someone who has spent much of his career studying insects and their relationships with plants to say. But conservation, ecological research, and public interest are still disproportionately focused on vertebrates: the birds, mammals, reptiles, amphibians and fishes that most people think of when they hear the word “wildlife”.

Yet those animals exist within ecological networks in which insects are frequently central players, not walk-on extras. A new review that I’ve co-authored with Rob Cooke, Eliza Grames, Chris Elphick and 18 other colleagues, published in Nature Reviews Biodiversity, tries to bring together what we know about those relationships – if you follow this link it’s free to read:

Cooke, R., Grames, E.M., Ollerton, J. et al. (2026) The importance of insect biodiversity for vertebrates. Nature Reviews Biodiversity DOI: 10.1038/s44358-026-00201-w.

The starting point is food, and the numbers are striking. We assembled diet information for 19,750 species of terrestrial and freshwater vertebrates. Of those for which data are available, 73% eat invertebrates and 44% eat them exclusively. That alone ought to give us pause when thinking about the consequences of insect declines.

But one of the things that I particularly enjoyed about writing this review is exploring with colleagues how food is only the beginning of the story. All of us involved in it, I think, learned just as much as we contributed.

Much more than something to eat

The relationships between insects and vertebrates are extraordinarily varied, and some are wonderfully unexpected. Termite mounds and ant nests provide nesting sites, refuges and places to lay eggs for mammals, birds, reptiles and amphibians. Some bats use cavities created by beetle larvae. Birds incorporate insect and spider silk into their nests. Other birds deliberately nest close to wasps or ants, apparently gaining protection from predators.

Then there are the chemical relationships. Poison frogs obtain many of their defensive alkaloids from the insects and other arthropods that they eat. Some snakes acquire defensive toxins from fireflies. Birds, mammals, reptiles and amphibians engage in various forms of “anting” or anointing, applying ants and other invertebrates to their bodies, probably for a variety of functions including parasite control.

Insects can even provide sanitation services. Beetles remove vertebrate dung and can thereby reduce the survival of parasites, while moths living in the nests of some birds consume faeces and other organic debris. There is fossil evidence that similar relationships involving feather-feeding beetles occurred in dinosaur nests during the Cretaceous.

One of the figures in the paper brings together examples from around the world, and is worth spending some time exploring. There are chimpanzees applying insects to wounds, parrots nesting in termite mounds, geckos “milking” honeydew from planthoppers, frogs sheltering in ant nests, herons using insects as fishing bait, and much more. Ecological interactions are rarely dull!

Insects connect ecosystems

There is also a much bigger ecological story within this review, because insect relationships with vertebrates move enormous quantities of energy and nutrients through ecosystems. Aquatic insects emerge from streams and lakes and become food for terrestrial animals; terrestrial insects fall into rivers and feed fishes. In some forest birds, aquatic insects can provide more than a quarter of their energy budget, rising seasonally to 50–90%. Conversely, terrestrial invertebrates can provide approximately half of the annual energy budget of freshwater fishes.

And then there are the indirect effects. As readers of this blog will know, a lot of my research and writing is related to plant pollination, but insect pollination doesn’t just matter to plants. When an insect pollinates a plant whose fruits or seeds are subsequently eaten by a bird or mammal, the insect is indirectly supporting that vertebrate too.

One striking example concerns figs. Fig trees depend upon their tiny agaonid wasp pollinators, while figs themselves are eaten by at least 1,274 species of birds and mammals. Remove the wasps and the consequences potentially propagate far beyond the insects and plants themselves.

This is something I’ve discussed from a different direction in my book Birds & Flowers, where the focus is largely on birds as pollinators rather than insects supporting birds, though all nectar-feeding birds also consume insects. But the broader lesson is the same: species do not exist independently of the ecological relationships that sustain them. I’ve written previously about some of the extraordinary diversity of these pollination relationships on this blog, including in my post on the limits to pollinator diversity.

What happens if the insects disappear?

This brings us to the more difficult question addressed by the review. There is now substantial evidence that insect abundance, biomass and diversity are changing, although, as I’ve argued on this blog before, the magnitude, geography and taxonomic extent of those changes need to be discussed carefully rather than reduced to simplistic claims of an impending “insect apocalypse”.

Nevertheless, widespread changes in insect populations inevitably raise the question of what happens to the vertebrates that depend upon them. Here the evidence becomes surprisingly patchy.

There are good reasons to expect effects: a meta-analysis of 125 studies, for example, found strong positive relationships between insect prey availability and bird reproductive success and chick condition. European insectivorous birds declined by 13% between 1990 and 2015 while omnivorous bird populations remained stable. There are also examples involving bats, small mammals and other groups.

But demonstrating that declining insects actually caused a particular vertebrate population to decline is much harder. Insects and vertebrates can respond simultaneously to habitat destruction, pesticides, climate change and other pressures, making cause and effect difficult to disentangle.

That’s an important distinction. Ecology is full of plausible stories; demonstrating the causal links is another matter. There are nevertheless cases where dependence on particular insects is difficult to ignore. The Critically Endangered Mountain Pygmy Possum in Australia depends heavily on Bogong Moths, whose numbers have crashed. The probable extinction of the Eskimo Curlew has likewise been linked partly to the disappearance of the Rocky Mountain Locust, formerly an important component of its diet.

A surprisingly large gap in our knowledge

Perhaps the thing that surprised me most while working on this review was how poorly documented many insect–vertebrate interactions remain. Even something as apparently straightforward as “what insects does this bird eat?” often turns out not to have a satisfactory answer. Studies may report prey simply as “invertebrates”, “arthropods”, or “insects”, with little or no taxonomic resolution. Birds are much better studied than most other vertebrates, yet detailed quantitative diet information is still available for only a small fraction of species.

Non-feeding relationships are even more scattered through the literature. This is part of what ecologists call the “Eltonian Shortfall“: our incomplete knowledge of who interacts with whom in nature.

One conclusion of the review is therefore that we need much better ways of recording these interactions: common data standards, better taxonomic resolution, long-term monitoring, experiments, larger databases, and greater collaboration between entomologists and vertebrate ecologists. Citizen science, computer vision, metabarcoding and environmental DNA all have potentially important roles to play.

There is also a plea here for something rather unfashionable: natural history. Observations of animals eating, nesting, sheltering, grooming and otherwise interacting with insects are not trivial anecdotes. Properly documented and brought together, they are the raw material from which larger ecological patterns emerge.

Conserving relationships, not just species

For me, one of the broader messages from this review is that conservation needs to pay much more attention to what I’ve referred to in the past as the “biodiversity of species interactions“.

We naturally talk about conserving species and habitats. But an ecosystem is not merely a list of species occupying the same place. It is also the vast network of feeding, pollination, seed dispersal, decomposition, competition, parasitism, facilitation and other interactions connecting those species.

Lose enough of those connections and an ecosystem can change profoundly even before many species have disappeared. That has a very practical implication: insect conservation is also vertebrate conservation. If we want landscapes that support healthy populations of birds, bats, amphibians, fishes and other vertebrates, maintaining abundant and diverse insect communities is not an optional extra. It is part of the ecological infrastructure upon which those animals depend. And, as our review makes abundantly clear, we have probably only begun to understand the full extent of that dependence. That’s why nature conservation policies and regulations, such as Biodiversity Net Gain (BNG) in England for instance, need to look carefully at their implications for insects and other invertebrates.

Finally, I’d like to thank all of the contributors to this paper for working so diligently on the review. It’s an output from the Vertebrates and Insects Working Group, which was formed by the Status of Insects Research Coordination Network – check out the website and get involved!

A century of genetic erosion in a British bumblebee – a new study just published

Leave a reply

Back in 2022 I wrote about a project in which I’d been involved with colleagues at the Natural History Museum, Imperial College London and several regional museums. We were using preserved bumblebees to ask a deceptively simple question: what can insects collected decades ago tell us about the environmental changes that their populations have experienced?

The first papers from that project showed two important things. One found increasing signs of developmental stress in the wings of museum bumblebees over the course of the 20th century, associated particularly with hotter and wetter conditions. The other demonstrated that it was possible to recover and sequence surprisingly good-quality DNA from insects collected more than a century ago, in some cases using just a single leg.

At the time I wrote that this opened up the possibility of looking at how the genomes of bee populations had changed through time. Well, we’ve now done it.

Published this week in Molecular Biology and Evolution, our new study focuses on the Moss Carder Bumblebee (Bombus muscorum), a beautiful species of flower-rich grasslands, heaths, marshes and coastal habitats. Once much more widespread across Britain, it is now largely concentrated in Scotland and the Scottish islands, with scattered populations surviving around the coasts of England and Wales.

Led by Victoria Mullin, the team generated whole-genome data from museum specimens collected between 1894 and 2019. The results provide a rather sobering picture of what has happened inside these bee populations as their geographical distribution has contracted.

Comparing bees collected before 1926 with those collected after 1975, we found about a 25% reduction in genome-wide genetic diversity. Looking at the full time series gives the same basic message: heterozygosity declined progressively through the 20th century, with no obvious sign of recovery.

That matters because losing a species is not the only way that biodiversity can decline. Long before extinction, populations can become smaller and more isolated and begin to lose genetic variation. That variation is part of the raw material that allows populations to respond to environmental change, disease and other challenges.

There was also a striking geographical pattern. In England and Wales, the amount of the genome occurring in long stretches of homozygosity increased by about 186% – in other words, almost threefold. These “runs of homozygosity” are consistent with populations becoming smaller, more fragmented and increasingly isolated from one another. Scotland showed a much weaker change, reflecting the fact that B. muscorum remains considerably more widespread there.

Importantly, however, the population structure is not simply “Scottish bees versus English bees”. The genomic data reveal a broad north-to-south gradient across Britain – what population geneticists call isolation by distance – rather than a set of sharply separated populations. That has implications for conservation because fragmentation may be breaking what was once a more continuously connected population into increasingly isolated regional groups.

There’s another interesting connection here with work that I published with colleagues more than a decade ago. In our 2014 Science paper we used historical records to show that Britain had lost 23 species of bees and flower-visiting wasps, with the fastest rate of extinction occurring from roughly the late 1920s to the late 1950s. We argued that large-scale changes in British agriculture, beginning just after the First World War, were an important part of that story.

The new study is looking at genetic change within a surviving species rather than the disappearance of whole species and it cannot pin that change on a single cause. But the parallels are difficult to ignore. The loss and fragmentation of flower-rich habitats associated with agricultural intensification is a likely contributor, alongside pesticides, climate change and other environmental pressures. The genomic erosion appears to have accumulated progressively across much of the same century in which Britain’s pollinator landscapes were being radically transformed.

One unexpected finding was also a useful warning about museum data. Of the 130 specimens originally labelled as Bombus muscorum, genetic analysis showed that 29 (almost a quarter) were actually two other, morphologically similar bumblebee species. Museum collections are extraordinarily valuable, but identifications cannot always simply be taken at face value, particularly for difficult groups. Genomics provides both a historical record and a way of checking the identity of the specimens supplying that record.

And that brings me back to what I think is one of the wider messages of this whole project: natural history collections are not dusty archives of dead organisms, they are biological time series.

A pinned bumblebee collected more than a century ago preserves information about where a species occurred, what environmental stresses it experienced, and, as we can now show, the genetic diversity of the population from which it came. Combined with modern genomic techniques, the millions of insects sitting in museum drawers around the world represent an extraordinary resource for understanding biological change.

The conservation message for the Moss Carder Bumblebee itself is fairly straightforward. Protecting the remaining populations is not enough if they become progressively smaller and more isolated. Restoring large areas of suitable, flower-rich habitat and reconnecting populations should help maintain population sizes and gene flow. More interventionist approaches such as translocation or assisted gene flow may eventually have a role, but would need careful assessment because of issues such as local genetic structure and disease transmission.

Four years ago, when I blogged about the first papers from this project, I finished with the idea that museum genomics might allow us to understand how insect populations had adapted, or failed to adapt, to a century of environmental change. We now have part of that answer. And it shows that beneath the visible contraction of a species’ range, another quieter form of biodiversity loss can be taking place: the erosion of diversity within its genome.

The paper is open access and you can download a copy here:

Mullin, V.E., Merchant, H.N., Arce, A.N., Nash, W. et al. (2026) Museum specimens reveal the genomic consequences of long-term population decline in an insect pollinator. Molecular Biology and Evolution 43: 1-11.

Fake figs, fooled wasps, and another extraordinary Apocynaceae pollination system

7 Replies

One of the things I love about working on the milkweed and dogbane family (Apocynaceae) is that it has an apparently inexhaustible capacity to produce strange and wonderful interactions with animals. Just when you think that the family’s pollination biology has yielded most of its surprises, along comes something like this.

A paper just out in New Phytologist, led by Aroonrat Kidyoo, describes the pollination of Heterostemma ficoides, an unusual climbing plant from northern Thailand. The title rather neatly gives away the story: “Fake figs, fooled fig wasps and rewarded phorid flies: brood-site mimicry and mutualism in Heterostemma ficoides (Apocynaceae)”.

And “fake figs” is not journalistic exaggeration, as you can see from the photos above, which are taken from the paper. The flowers of H. ficoides are globose, fleshy structures, about 1.5–2 cm across, with a small opening at the top. Superficially they look remarkably like the enclosed inflorescences (the syconia) that we normally call figs. Indeed, when the species was first described in 2019, its resemblance to a fig was obvious enough to inspire the name ficoides: fig-like. But looking like a fig is only the beginning of the story.

Fooling one of nature’s most specialised pollinators

The relationship between figs (Ficus) and their tiny agaonid wasps is one of the classic examples of an intimate plant–pollinator mutualism. Female fig wasps locate receptive figs using their scent, squeeze through the tiny opening or ostiole, and pollinate the enclosed flowers while attempting to reproduce inside. These relationships can be remarkably specific, in some cases representing rare examples of one-to-one plant-pollinator interactions.

So it was genuinely surprising to discover female Kradibia fig wasps entering flowers of Heterostemma ficoides. These wasps normally pollinate Ficus heterophylla, not an Apocynaceae, and the authors believe that nothing comparable, i.e., a fig wasp being fooled into visiting a completely unrelated flowering plant, has previously been documented. I have to concur, this relationship is (so far) unique.

The authors have neatly demonstrated that the deception of these wasps is not simply visual. The chemical bouquet produced by H. ficoides is closer to that of receptive Ficus heterophylla figs than it is to the flowers of its relative H. siamicum. More importantly, behavioural experiments showed that Kradibia wasps were significantly attracted to the scent of open H. ficoides flowers, though not to unopened flower buds. In other words, the plant seems to switch on its misleading olfactory advertisement at just the right time.

Once inside, however, the wasps discover that they have made a serious mistake. They wander around probing the inner surface, searching for egg-laying opportunities that simply aren’t there. Some die inside the flower, although more than 80% of observed individuals managed to get out again, and often went on to visit another Heterostemma flower.

In the process, pollen can be transferred between flowers, but as in other members of the Asclepiadoideae, Heterostemma does not dust its visitors with loose pollen grains. Instead, pollen is packaged into pollinia which become attached to the legs of the fig wasps.

So from the fig wasp’s perspective this is straightforward deception: the plant promises a breeding site that doesn’t exist, and the flowers get pollinated in return. But that is only half of the system.

The flies are not being fooled by the fake fig

The most important pollinator of H. ficoides appears to be a tiny phorid fly in the genus Megaselia, and its relationship with the plant is quite different. Female Megaselia enter the flowers and lay their eggs there. Mating may even take place on the flower beforehand. Their larvae then develop inside the fallen flowers, eventually completing their development in the decomposing floral tissue on the forest floor. For these flies the promise made by the flower is therefore an honest one: this really is a brood site.

Pollinaria become attached to the flies’ mouthparts, then the flies carry them to another flower, where pollen can be deposited. Interestingly, Megaselia are also important pollinators of some species of the related Ceropegia in Africa (both genera are members of the tribe Ceropegieae).

This creates an ecological combination that is one of the most distinct examples of a “bimodal pollination system” that I’m aware of, in which the same flower operates simultaneously as a brood-site mutualism and as a brood-site deception, involving two completely distinct groups of insects.

Back to our 2019 Apocynaceae review

For me, there is a satisfying connection here with the large review of Apocynaceae pollination systems that we published in Annals of Botany in 2019: The diversity and evolution of pollination systems in large plant clades: Apocynaceae as a case study. Aroonrat Kidyoo, lead author of this new study, was also one of the large international team involved in that paper.

We assembled usable pollination information for 567 species (a bit more than 10% of the family) and found an extraordinary diversity of interactions. About 73% of the species for which we had information were associated predominantly with a single broad pollinator group, whether bees and wasps, butterflies and moths, flies, beetles or birds. However, about 7% of the species had distinct bimodal pollination systems, such as bee + butterfly. Whether this is an unusually high proportion for such a family is currently unclear as there are few comparative statistics.

One of our broader conclusions from the study was that this interaction diversity was not randomly scattered across the Apocynaceae evolutionary tree. Pollination systems show strong phylogenetic structure and some transitions between pollinator types have occurred much more readily than others. The new Heterostemma study puts some wonderfully detailed natural history onto that much larger evolutionary picture.

Aroonrat and her colleagues suggest that Heterostemma ficoides may represent a shift away from the more open, nectar-producing flowers found in other parts of the genus. In section Oianthus, to which H. ficoides belongs, the flowers have become progressively more closed. In the extreme case represented here, the result is a globose chamber entered through a small opening, nectar production has apparently disappeared, and the plant has moved towards brood-site pollination by flies and deception of fig wasps.

Whether that evolutionary scenario survives a proper phylogeny and detailed pollination studies of the remaining species is an obvious question for the future. As the authors themselves acknowledge, both are currently lacking. But it is exactly the sort of transition that our 2019 study suggested we should be looking for.

And what does any of this have to do with eating Apocynaceae?

Quite a lot, in a roundabout way. Earlier this year my colleagues and I published another large synthesis of the family, this time asking a very different question: which Apocynaceae do people eat?

We documented 440 edible species, spread widely across the evolutionary tree of the family. That sounds like a lot, although statistically Apocynaceae are actually under-represented among the world’s food plants, probably in part because the family is so well endowed with toxic and pharmacologically active chemistry.

One point that emerged strongly from that work was that our knowledge of how those useful plants reproduce is remarkably incomplete. Many edible Apocynaceae are harvested for their fruits and seeds, which of course ultimately means that successful reproduction matters. Yet for most of those species we still know little or nothing about their pollinators.

That matters if we are serious about conserving wild food plants or developing some of them as future crops. You cannot fully understand the biology of a fruit-producing plant without understanding how the flowers get pollinated in the first place.

There are no Heterostemma species documented in our edible-plant study (though it appears that we missed one – see the comment below), so H. ficoides is not some promising new food species. But that isn’t my point. Rather, this new research is a reminder of just how much biological complexity can lie behind the production of an Apocynaceae seed. And we still know the pollination biology of only a fraction of this enormous family.

That, perhaps, is the broader lesson from all three papers. Large databases and family-wide syntheses are valuable because they reveal general patterns and, equally importantly, the gaps in our knowledge. But filling those gaps still depends on researchers spending time watching individual plants and individual insects, dissecting flowers, rearing larvae and asking exactly what the animals are doing. Sometimes what they find is stranger than anyone would have predicted.

Join me for “Pollinator Conservation: Threats and Opportunities” in Bedfordshire on Tuesday 15th September

Leave a reply

Bedfordshire Natural History Society (BNHS) has invited me to give a talk on the theme of “Pollinator Conservation: Threats and Opportunities” at Maulden Village Hall (MK45 2DP) on Tuesday 15th September. It starts at 8pm and non-members are encouraged to attend.

More details are available at the BNHS web site.

I’ll have signed copies of my books available to purchase, so if you’re in the area, please do come along.

Nature and national security: some recent developments

1 Reply

Back in January I wrote a post entitled Pollination as a matter of national security, prompted by the UK Government’s rather extraordinary publication of this report: Global biodiversity loss, ecosystem collapse and national security: a national security assessment. My argument was that although pollination is obviously not a national security issue in the traditional sense of defending a country against hostile states or terrorist attacks, the loss of pollinators threatens food production, supply chains and the functioning of ecosystems upon which societies ultimately depend. In other words, it is a threat to national resilience, which we can define a country’s capacity to anticipate, withstand, adapt to, and recover from major shocks or long-term pressures while maintaining the essential functions of society, the economy, government, and critical infrastructure. I’ll come back to this national-level framing at the end.

I concluded that post by suggesting that the real question was not whether pollination could be regarded as a national security issue, but whether national security thinking had fully adapted to the biological foundations on which societies depend.

Since January, several reports, papers and policy developments have convinced me that this is becoming a much wider and more serious discussion. Pollination is only one component of it. What is emerging is a view of nature itself as part of the infrastructure that makes societies secure.

Nature as infrastructure

Perhaps the clearest expression of this idea comes from a paper by Bradley Cardinale, Emmett Duffy and Rod Schoonover entitled Nature’s role in national security, published this summer in the journal Nature-Based Solutions.

The authors consider five areas where ecological disruption can generate national security risks: food security, water scarcity, health security, protection from natural disasters, and environmental crime. In each case, environmental degradation can create or amplify economic and social pressures which, under some circumstances, translate into political instability and security problems.

What I particularly like about the paper is its use of the term “natural infrastructure”. Functioning ecosystems, the authors argue, operate rather like electricity grids, communications systems and transport networks. They provide services and, crucially, buffer societies against shocks.

This is an important shift in emphasis. We usually talk about the benefits or ecosystem services that nature provides. Both are perfectly useful concepts, but describing ecosystems as infrastructure changes how we think about their loss. A degraded wetland is then not simply a conservation problem. It may represent the deterioration of flood-control infrastructure. Similarly, a damaged watershed is a weakening of water-security infrastructure, whilst a decline in pollinators can be understood as damage to parts of our food-production infrastructure.

Seen in that way, restoring ecosystems begins to look less like discretionary environmental spending and more like preventive investment in resilience.

Rewilding as defence

An even more striking example appeared in April in The RUSI Journal, published by the Royal United Services Institute, one of the world’s oldest defence and security think tanks.

Sam Jelliman, Brian Schmidt and Alan Chandler propose the concept of “defensive rewilding”: strategically restoring forests, wetlands, peatlands and other ecosystems in locations where they could also impede, divert or slow an invading military force. This infographic from the paper outlines their argument quite persuasively:

At first sight this sounds rather eccentric, but the logic is straightforward. Military planners since at least the time of Alexander the Great have always understood that terrain matters. At the Battle of Issus in 333 BCE, for example, Alexander fought the much larger Persian army on a narrow strip of land between the sea and the mountains, restricting its ability to manoeuvre and helping to neutralise its numerical advantage. Wetlands, forests, rivers, hedgerows and boggy ground can also provide advantage by restricting visibility and movement and make the passage of vehicles and troops much more difficult. We have seen examples of this during the war in Ukraine, just as we did in numerous earlier conflicts.

What Jelliman and colleagues suggest is that this defensive function could sometimes be deliberately combined with ecological restoration. Rewetting peatlands or restoring floodplains near vulnerable borders might simultaneously enhance biodiversity, store carbon, reduce flooding and create terrain that is difficult for armoured vehicles to cross. A win-win for both national defence and this wider concept of national security.

Clearly this is a concept that requires much more testing, and ecological restoration should not be distorted into a military land-management programme. There would also be difficult questions about land ownership, farming and the ecological appropriateness of particular interventions. But that is almost beside the point. The fact that rewilding is now being seriously discussed in a defence journal as potential security infrastructure is itself significant.

It challenges the assumption that spending on defence and spending on nature necessarily compete with one another, though we’ve long known that some areas of the UK than are used by the military are also important wildlife sites, Salisbury Plain being an obvious example, the largest known expanse of unimproved chalk grassland in north-west Europe precisely because it is used for military purposes rather than being converted to agriculture.

Losing nature can make countries poorer and less financially secure

Another recent study extends the argument into the financial system. Matthew Agarwala and colleagues published a paper in Nature Ecology & Evolution examining how losses of ecosystem services might affect the creditworthiness of nations, which is the degree to which lenders and investors judge that a country’s government can reliably meet its debt obligations in full and on time. Their model incorporated scenarios involving the deterioration of three services – wild pollination, marine fisheries and tropical timber – into sovereign credit assessments for 23 countries containing about 5.5 billion people.

The numbers produced by their scenarios are striking. Under partial ecosystem collapse, additional annual sovereign interest payments across the countries studied could reach around US$162 billion. India and China alone could experience tens of billions of dollars in increased annual debt-servicing costs. Some countries could suffer very large reductions in GDP.

These are modelled scenarios rather than predictions, of course, and the ecological assumptions that sit beneath such economic modelling need careful scrutiny. But the basic mechanism is compelling.

Governments depend upon functioning economies to raise taxes, borrow money and respond to crises. If ecological degradation reduces agricultural output, fisheries production and other forms of economic activity, the fiscal position of a country deteriorates. Borrowing becomes more expensive precisely when governments may need to spend more money adapting to environmental change.

This summer’s record breaking drought and wildfires in the UK, and the promise of further disruption due to a mega El Nino event, make such adaptation all the more urgent. Nature loss can create a nasty feedback loop: ecological degradation weakens economies, weaker economies have less capacity to invest in resilience and restoration, and that in turn increases their vulnerability to further ecological degradation.

All of this moves biodiversity loss well beyond the conventional environmental policy box and there are positive signs that these ideas are becoming embedded in government.

The UK security machinery is beginning to respond

In June the UK Government established its first Climate Security Taskforce, bringing together experts from the security, military, academic and environmental communities. Significantly, its remit explicitly encompasses not only climate change but nature loss. This is important as many of us have been arguing for years that you cannot disentangle the two and they should not be treated as separate domains of environmental concern. The Government’s announcement stated that climate change and nature loss are drivers of instability, economic disruption and security risk, and linked the Taskforce to the broader approach set out in the UK’s National Security Strategy.

That matters because one of the perennial problems with biodiversity has been institutional fragmentation. Nature tends to be dealt with by environment ministries, while finance ministries think about economic stability, agriculture departments think about food production, health ministries think about disease, and defence and intelligence agencies think about security. But ecological risks do not respect departmental boundaries.

A collapsing ecosystem can simultaneously affect food prices, water availability, infectious disease risks, migration, trade, government finances and geopolitical stability. Treating each of these consequences separately misses the common ecological cause.

China, where I’ve been spending quite a bit of time in recent years, provides an interesting comparison with the UK. Whereas the explicit framing of biodiversity loss as a national-security risk is relatively new in Britain, China has for some years incorporated “ecological security” into its much broader concept of national security. Its first national-security white paper, published in 2025, explicitly listed ecological security alongside military, economic, food, resource, biological and other forms of security. This has been reinforced in 2026 by China’s new Five-Year Plans for ecological protection and the construction of a “Beautiful China”, both of which link biodiversity conservation, ecosystem restoration and resilience to national ecological security. Chinese researchers have meanwhile developed an extensive literature on “ecological security patterns”, identifying the habitats, corridors and ecosystem services required to maintain resilient landscapes. The Chinese use of the term is not identical to the emerging British idea of “nature security”, but the overlap is striking: both ultimately recognise that the security of a state rests partly upon the ecological systems that support it.

From an intelligence assessment to Parliament

The UK Government’s January assessment has continued to reverberate through Westminster. On 2 September, the House of Commons Library published a new briefing ahead of a parliamentary debate on the assessment. The briefing does not constitute a new security assessment, but usefully brings together the original findings with subsequent parliamentary scrutiny, stakeholder responses and government action, including the creation of the Climate Security Taskforce and a new three-year Nature Security R&D Programme.

The briefing rightly emphasises an important point about the original security assessment. It is based on a reasonable worst-case scenario. It is not predicting that these ecosystems will suddenly collapse tomorrow. National security planning routinely explores high-impact possibilities precisely because governments need to understand what would happen if severe events occurred.

On 3 September the subject was then debated in Westminster Hall, following a cross-party initiative led by Adrian Ramsay MP. Whatever ultimately comes from that debate, this represents an interesting political transition. In January, biodiversity loss appeared in a national security assessment, and eight months later it had become the subject of a dedicated parliamentary research briefing and debate. That does not mean that nature has suddenly moved to the centre of British security policy. Far from it. But it suggests that the language of nature security is beginning to acquire institutional traction.

And then there is food

As if to underline the point, today (4 September) the National Audit Office has published a report on the Resilience of the food supply chain to disruptions. This is not primarily a biodiversity report. It examines the resilience of the UK food system to a much wider range of threats, including extreme weather, energy disruption, cyber-attacks, pandemics and animal and plant diseases. But there is a direct connection with the nature-security debate.

Food is one of the UK’s 13 designated Critical National Infrastructure sectors. The NAO concludes that although the food system has proved resilient to previous disruptions, the ability of the sector to withstand catastrophic events has not been tested and that Defra’s actions have not yet been sufficient to ensure resilience to such events.

This brings us back to where I started in January with pollination. Food security is not simply about having enough farms. It depends on soils, water, climatic stability, pollinators, pest regulation, genetic diversity, functioning supply chains, energy, transport and international trade. Pull at enough of those threads simultaneously and resilience begins to unravel.

So where does this leave us?

What interests me most about these developments is that they come from very different intellectual and institutional directions. Ecologists are describing ecosystems as natural infrastructure. Economists are beginning to incorporate biodiversity into sovereign financial risk. Defence researchers are thinking about restored landscapes as strategic assets. Government has created a taskforce dealing with climate and nature security. Parliament is discussing ecosystem collapse as a national security problem. And the National Audit Office is warning that one of the country’s most fundamental pieces of critical infrastructure, its food system, needs greater resilience.

These are not separate conversations, they are pieces of the same emerging picture.

There is, of course, a danger in pushing the national-security framing too far. Nature deserves protection for many reasons that have nothing to do with its usefulness to states, economies or human societies. I would be very uncomfortable with biodiversity conservation being justified solely because it makes Britain, or any other country, more secure.

But that is not the argument I am making. Rather, I think we have systematically underestimated the extent to which national security is built upon ecological stability. Forests, soils, rivers, wetlands, oceans, pollinators and the millions of other species that make ecosystems function are not scenery surrounding the human economy, they are part of its physical foundations.

Earlier this year I argued that conserving pollinators could be regarded as a form of preventive security investment. The developments since then suggest that the same argument can be made much more broadly: Protecting and restoring nature is not an alternative to investing in national resilience. It is one of the ways in which we invest in national resilience.

That feels like an idea whose time has arrived.

There is, however, an obvious limitation to framing all of this in purely national terms. Modern states are embedded in a globalised economy in which food, energy, raw materials, finance, manufactured goods and ecological impacts cross borders continuously. A country may strengthen the resilience of its own landscapes and infrastructure yet remain highly vulnerable to ecosystem degradation, crop failures or political instability elsewhere. In that sense, “national resilience” is also dependent on international ecological resilience: security at home increasingly rests on the stability of environmental and economic systems far beyond national borders. Just as individual trees in a forest are dependent upon one another to create the conditions in which all can thrive, regardless of species, so the too are countries mutually dependent. The leaders of the world’s nations would do well to remember that.

Keeping up with pollination: using ChatGPT as a research alert

3 Replies

There is too much science.

That is not a complaint about the quantity of research being done. Quite the opposite: it is remarkable how much good work is now being published across the world. But the sheer volume creates a serious practical problem. How can any researcher keep track of what is relevant to their field, let alone read and properly absorb it?

For someone interested in pollination, the difficulty is compounded by the extraordinary breadth of the subject. Pollination is not confined to a single academic discipline, or even to a small cluster of them. It cuts across botany, zoology, ecology, evolution, conservation biology, agriculture, economics, food security, national security, palaeontology, biogeography, genetics, behaviour, climate science and many other areas.

A paper that changes how I think about pollination might appear in a specialist plant journal, an entomological publication, or journals covering conservation, agricultural economics, or palaeontology. It might concern the structure of ecological networks, the nutritional quality of crops, the evolution of flowers, pesticide regulation, the movements of migratory birds or the fossil record of insects. It may not even use the word “pollination” prominently in its title or abstract.

That breadth is one of the great attractions of a subject that has kept me fascinated for 35 years, and which I tried to capture in my book Pollinators & Pollination: Nature and Society. It is also what makes keeping up with it so challenging.

The limitations of conventional alerts

There are, of course, many ways to receive information about new research. Journals send tables of contents. Google Scholar provides alerts based on keywords or authors. ResearchGate regularly tells me that someone has cited one of my publications.

These services have their place, but they tend to provide a rather narrow window onto the literature.

For example, Google Scholar can alert researchers to newly indexed material matching a search query. This is useful, but it is neither a comprehensive record of everything published nor a carefully curated selection: broad searches produce noise, while narrow ones inevitably miss relevant work.

A citation alert from ResearchGate tells me about research connected to work that I have already published. That can be valuable, and occasionally flattering, but it inevitably looks backwards. It shows me the expanding wake of my own research rather than offering a broad view of where the subject is moving.

Keyword alerts have a different problem. They can generate large quantities of material with very little discrimination. A search for “pollination”, “pollinator” or “plant–pollinator interactions” will retrieve many relevant papers, but also conference notices, marginally related studies, duplicate records and work of highly variable importance.

More restrictive searches reduce the noise but risk excluding the unexpected paper that turns out to be especially interesting.

What I need is not simply a larger stream of titles. I want something closer to an informed research assistant: a system that could search widely, exercise some judgement, explain why particular items might matter, and alter its approach in response to my comments.

I have therefore been experimenting with ChatGPT’s scheduling function as a weekly research-alert system.

A scheduled conversation

My current instruction is for ChatGPT to provide a shortlist every Friday morning of the most worthwhile new papers, preprints, reports and substantive analyses relating to plant–pollinator interactions, pollination networks, bird pollination and related broader biodiversity topics.

I have asked it to include no more than ten items, and fewer when the available material is weak. That final qualification is important: I do not need ten references merely to fill ten spaces, I’d rather receive four genuinely interesting papers than a padded list containing six that I will never read.

For each item, the system can provide the citation, a short account of the main finding and an explanation of why it may be relevant to my interests. It can also distinguish between peer-reviewed research, preprints, reports and other forms of analysis.

In that sense, the alert is already more useful than a conventional automated search. But the most important difference is that it is a conversation.

I can tell it that a particular paper was especially useful and ask it to look for more work of that kind. I can point out that another item was only marginally relevant. I can ask it to widen its search into palaeontology, ecological economics or agricultural policy, or to pay closer attention to a particular taxonomic group.

I can also tell it what not to do. When a major European pollinator-research white paper appeared, for example, the alert quite reasonably identified it as relevant. But I was one of its co-authors and did not need an artificial intelligence system to introduce it to me as a new discovery. I could therefore instruct the system to recognise my own publications and either omit them or flag them only when they were strategically relevant.

That adaptability is difficult to reproduce with conventional keyword alerts. The scheduled search is therefore not a fixed filter, it can be refined as my interests, projects and frustrations change.

From retrieval to assessment

The distinction between finding research and assessing it is also important.

A long list of newly published papers transfers the problem of selection from the search engine to the researcher. A useful alert should do more than retrieve documents. It should offer some preliminary judgement about novelty, relevance and significance.

Does a paper introduce a genuinely new idea, or does it repackage a familiar concept in new terminology? Is a striking conclusion supported by a strong study design? Does a paper matter because of its empirical results, its methods, its conceptual framework or its policy implications? Is it directly relevant to my work, or merely adjacent to it?

ChatGPT cannot answer such questions infallibly. Nor should its assessment be accepted without scrutiny. But it can help to triage the literature and identify which papers deserve closer attention.

This is particularly valuable outside one’s immediate specialism. I can usually make a rapid initial judgement about a field study of flower visitors or a paper on pollination networks. I may need more assistance in deciding whether a new economic analysis, remote-sensing method or palaeontological reconstruction is likely to be important.

The purpose is not to delegate scientific judgement, it is to direct that judgement more efficiently.

A few necessary cautions

There are obvious limitations.

An AI-generated research brief is only as good as the literature it can locate and the instructions it has been given. It may miss relevant work, misunderstand a paper, exaggerate novelty or place too much weight on a fashionable topic.

There is also a danger of creating an intellectual echo chamber. If I repeatedly tell the system what interests me, it may become increasingly good at confirming those interests while overlooking research that sits outside them. A useful alert needs some permission to surprise.

For that reason, I think the instructions should combine a clear core remit with an explicit request to include occasional unexpected items that could change how the field is viewed.

The summaries themselves must also remain starting points. Important papers need to be read, methods inspected and conclusions considered in relation to the evidence. An articulate paragraph generated by ChatGPT is not a substitute for engaging with the original study.

A better kind of alert

Used carefully, however, scheduled ChatGPT alerts offer something that conventional notification services do not: an iterative relationship between the researcher and the search process.

The system can learn that I am interested not only in pollination as an ecological interaction, but also in its agricultural, economic, evolutionary, historical and political dimensions. It can respond when I say that a particular area is becoming more important to a project. It can explain its selections, accept criticism and search differently the following week.

That makes it less like an electronic table of contents and more like a research assistant who arrives on Friday morning with a small pile of papers and a cup of coffee and says: “These are the ones I think you should look at, and this is why.”

Like any research assistant, it will sometimes be wrong. It may overlook something obvious, become overenthusiastic about a weak paper or misunderstand why a study matters. But it can be corrected. The conversation continues, and the next briefing can be better.

For a field as broad and rapidly expanding as pollination research, that may be the most useful feature of all.

I will finish with an example. Today’s alert arrived at the scheduled 08:00 (though I had to make my own coffee…) and included four items for me to look at; this was the fourth, which I’m reproducing verbatim:

4. Agarwala et al. — Biodiversity loss will decrease the future creditworthiness of nations

The authors incorporate projected losses of wild pollination, tropical timber and marine fisheries into sovereign credit-risk models for 23 countries. Their scenarios suggest that partial ecosystem collapse could substantially increase national borrowing costs and expose a major omission in conventional credit ratings.

Why it matters: This is one of the clearer attempts to translate pollination loss from ecosystem-service valuation into a decision framework already used by governments and financial markets. For ecological-economics work, the methodological bridge—from spatial ecosystem-service scenarios through macroeconomic models to sovereign credit ratings—is more important than the headline monetary totals.

The ecological assumptions deserve scrutiny. “Wild pollination” is compressed into a tractable economic input, so uncertainty in crop dependence, substitution, adaptation and spatial mismatch may be obscured once propagated through the financial model.

Assessment: Strategically consequential interdisciplinary analysis, not new pollination biology.

The paper actually appeared on 4th June but I missed it, even though it cites this paper by Millard et al., in which I was involved, and presumably appeared on a ResearchGate citation alert last month. ChatGPT rightly spotted that it was something which would interest me and that’s what’s so fascinating about this approach to keeping up with the literature – this system is intelligent enough to “understand” my needs, even if it’s not conscious of precisely why it understands them.

As always, I’d be pleased to hear your views on this topic – feel free to comment below.