Munchity crunchity

Synopsis : 5%, 40%, 80%? What proportion of the Asian hornet’s diet is honey bees? It depends where and how you look, but we need to know to help mitigate the impact of the hornet invasion.

Introduction

Munchity crunchity are the words that usually come to mind {{1}} when I watch a dragonfly chewing on whatever hapless insect it has just caught. There’s the crunching as the chitinous exoskeleton is pared away and discarded and the munching through the soft abdominal tissues and thoracic musculature {{2}}

Golden-ringed dragonfly

I’m not sure whether munchity crunchity is onomatopoeic {{3}} or possibly just a phonestheme {{4}}.

Whatever it is, it’s a term that nicely describes the sight – and sound if you’re close enough – of eating something crispy on the outside and squidgy in the middle, like a dragonfly eating a wasp, an otter eating a sea urchin, or a dinner guest scoffing the last of the dark chocolate pralines.

Munchity crunchity is not ((Yet.)) the accepted technical term describing how an Asian hornet (Vespa velutina) worker dismembers a honey bee to prepare a protein-rich pellet to take back to the nest to feed the developing larvae, but perhaps it should be?

Although most of the press coverage (and inevitably most of the beekeeping coverage) has been about the decimation of honey bee colonies by Asian hornets, their success as an invasive insect is because they have catholic tastes and are opportunistic predators.

So what does the Asian hornet eat, how much do they eat, and does our focus on honey bees reflect our inherent and understandable bias, or truly represent their predilection for our favourite insect?

Sugar and protein

Adult Asian hornets eat sugar-rich liquids and a protein-rich ’soup’ regurgitated by developing larvae.

Nice.

The attraction of sugar-rich liquids to queens recently emerged from hibernation is reflected in the choice of baits for early-season trapping (a topic I’ll cover in the future) which contain things like fruit juice, syrups and beer.

Once the developing colony starts rearing brood they need significant amounts of protein to feed the larvae. The workers catch their prey, dismembering it and removing almost everything except the flight muscles {{5}}. This nutritious flesh pellet is carried back to the nest in the mandibles. This need for a high protein diet explains the use of shrimp or trout baits when trapping hornets in the summer.

Whilst developing the larvae does not defecate, the gut contents being eliminated at the prepupal stage. The released faeces remains in the bottom of the cell once the adult hornet emerges.

Dietary studies

So, if you wanted to analyse the diet of the Asian hornet – and therefore its potential impact on honey bees or other insects in the environment – there are a range of things you could study, including the:

• observation of hunting hornets
• prey pellet carried by returning workers to the hornet nest
• larval gut contents
• faecal samples from the vacated cell

The first two of these may require little or no specialised techniques other than good observation (and lots of time). The last two would require some sort of molecular analysis as the contents will have been digested.

Direct observation of hunting hornets requires being in the right place at the right time. That’s not a problem when they’re hawking in front of a bee hive, but next to impossible if they’re picking off spiders in the tree canopy.

Faecal samples are inevitably much older than other samples. It’s therefore likely that they have undergone some (or possibly extensive) degradation, making identification even using advanced methods problematic.

In this post I’ll discuss some recent publications that have mainly analysed flesh pellets and larval gut contents to determine the diet of Asian hornets.

Don’t expect studies like this to provide a definitive answer such as ”25% of the Asian hornet’s diet is honey bees” that’s applicable everywhere and all the time. After all, if there are no honey bees locally they won’t be able to catch and eat them.

I consider the studies are indicative, but that some of the results are interesting.

France

I’ll briefly cover three studies on Asian hornet diet, conducted using a range of methodologies in different countries {{6}}. The first, both here and chronologically, was published early in 2021 by Rome and colleagues (Rome et al, 2021) who studied 16 nests in the Dordogne region of South-west France from 2008-2010 {{7}}.

Throughout the season (or as long as the landowner allowed the nest to exist) returning workers were captured using a sweep net during periodic sampling. Any prey pellets they were carrying were preserved for analysis, both morphologically and using molecular methods.

Morphological analysis simply means trying to identify the remains using a microscope … not always easy if the prey pellet is a chunk of thoracic muscle. The authors therefore identified and partially sequenced the cytochrome oxidase I gene (COI).

That last sentence risked losing half the readership … don’t go!

Cytochrome oxidase is a protein present in the cells of all animals and is encoded by the COI gene. This gene is very widely conserved but not identical.

Therefore, if you determine the sequence of the COI gene from an unknown insect and compare it with a database of all known cytochrome oxidase genes you should be able to find a similar, or perhaps identical, match. COI genes of wasps are more similar to each other than they are to COI genes of bumble bees which are more similar to those of wasps than they are to butterflies etc.

Think of it as a sort of fingerprinting method. The more points of similarity, the more likely the gene (and hence prey pellet) is to have originated from a bee, or fly, spider or shrimp.

What do they munchity crunchity?

Over 12,000 hornets were captured in about 200 hours of sampling. Of these ~2,000 hornets were carrying prey pellets {{8}} from which 141 species were identified by COI gene sequencing and up to a further 18 species identified morphologically.

These included 43 families of insects, 3 families of spiders and 4 families of vertebrates {{9}}.

That’s a very diverse diet.

The majority (~60%) of the prey pellets were from Hymenoptera (bees and wasps), of which ~38% were honey bees and ~20% social wasps. The next largest group of insects were a very diverse range of Dipterans (flies), which comprised ~30% of the prey pellets.

Inevitably intermittent sampling missed a lot of hornets and their prey pellets. Based upon sampling effort, pellet numbers collected and the range of prey pellets identified it was possible to predict the total number of different prey species predated by Asian hornets … an eye-watering 411 species.

This emphasises just what an opportunistic generalised predator Vespa velutina is. If it’s there and they can find it and catch it, they will eat it.

Other interesting aspects of the study

Two other things caught my eye from the Rome et al., study.

Sampling hornets continued throughout the season. Therefore, the temporal distribution of prey pellets gives an indication of how ‘strong’ predatory activity is through the season. Using a similar approach it was possible to show that the proportion of prey pellets from honey bees or flies or wasps varied during the season (I only show honey bees below).

Temporal predation pressure (left) and proportion of honey bees in diet (right)

Predatory activity of Asian hornets peaked in the first week of October but honey bees (Apis mellifera) made up a smaller proportion of the diet as the season progressed. Remember, this is in the Dordogne region of France, so likely to be appreciably warmer than the UK by October. How this influences nest development and predatory activity elsewhere remains to be determined.

The second interesting thing was the sort of ‘back of an envelope’ calculation {{10}} of the total weight of insects consumed by an Asian hornet colony. By weighing prey pellets and developing Asian hornet pupae they could calculate approximately how many pellets each pupa had consumed. This, coupled with the average hornet nest size and the average honey bee weight, means a colony would consume 11.32 kg of insects in a season.

There are a lot of assumptions made in this calculation so treat it with considerable caution … suffice to say, an Asian hornet nest will consume a lot of a wide variety of insects in its environment.

Just one more thing

Bumble bees and solitary bees made up only 0.02% of the prey recorded. It’s clear that the Asian hornet is opportunistic and primarily preys on locally concentrated insects, such as honey bees and flies around carrion or cattle.

That being the case, the Asian hornet is unlikely to threaten locally rare species … if there are few about the hornet is unlikely to find and catch them.

The paper closes with a brief discussion of the impact (or lack of it) of the hornet on some of these rare insect species compared with the unselective trapping of insects using sugar/beer baits designed for hornets. In these as little as 1% of the total catch are hornets (Lioy et al., 2020).

This suggests that there’s a real risk that the methods to control the Asian hornet might be a greater threat to biodiversity than the hornet itself.

Portugal

Verdasca and colleagues (Verdasca et al., 2021) used a molecular approach to specifically investigate predation of Asian hornets on honey bees. This study was conducted in northern and central Portugal in mid/late 2018. It’s a relatively small study and is notable because they analysed nests after ‘inactivation’. This enabled the scientists to sample the jaws and stomachs of workers and the faecal pellet left by larvae days or weeks after the nest was poisoned.

I’m not going to discuss the methodology (metabarcoding) other than to say that they used an approach that looked at all the DNA present and determined the proportion that was from honey bees (or flies, spiders, moths etc.). There are several limitations to this strategy because it makes lots of assumptions that the amplification/detection methods employed work equally well irrespective of the target species. They probably do not … nevertheless, as a first approximation it is a useful approach.

Lots of honey bees in the diet

All nests analysed contained DNA from honey bees. The proportion of honey bee DNA vs. other species DNA in hornet worker jaw, stomachs or in faecal pellets varied between nests and samples. Jaws and faecal pellets contained much greater amounts of honey bee DNA, presumably because workers do not eat the prey, but instead carry it back to the nest for the larvae to consume.

Overall ~80% of the diet was from honey bees, with flies being the next most common prey. As in the Rome et al., (2021) study bumble bees and solitary bees were rare {{11}}.

I suspect this study overestimates the proportion of honey bees in the diet of the Asian hornet. The reduced proportion of honey bees in the diet late in the season (see above) is not discussed, but is also not apparent from the raw data presented.

Both the Verdasca and Rome studies were conducted in mainland Europe where there are significant differences in climate, local insect abundance and biodiversity from conditions in the UK where the Asian hornet is or may be becoming established.

What’s happening closer to home?

Great Britain

Eleanor Jones and colleagues from the National Bee Unit and University of Newcastle (Stainton et al., 2023) ‘cut to the chase’ and analysed the gut contents of Asian hornet larvae recovered soon after nest destruction. Nests were from two locations in England and three in the Channel Islands and were collected in August, September or October (2016-2020).

One advantage of analysing larvae is that they represent pooled samples of prey, following feeding by successive workers which may have brought back different prey species. However, that is also a disadvantage in that some digestion and hence degradation of the samples would have occurred, potentially biasing the results.

For example, you would expect that the DNA from prey pellets consumed earlier would degrade more than the more recently consumed material. Sampling a larva that was fed anything but honey bee bits for a week but that then received lots of honey bee protein would likely overestimate the proportion of honey bee material in the diet.

Again this study used a metabarcoding approach, but this time I can’t avoid describing it (in a very simplistic and superficial manner).

Essentially all the DNA from the larval gut was extracted and chopped randomly into little pieces. Each of these pieces was sequenced which allows the species (or at least the genus) to be determined e.g. you might be able to determine it’s from one of the bee-mimicking Volucella hover flies, without being able to distinguish between V. bombylans and V. zonaria.

Jigsaws

Each of these individual sequences is called a ‘read’ and the total number of honey bee reads vs the total number that match Volucella give an indication of the proportion of the different species in the diet.

Metabarcoding … sort of

Think of it as a jigsaw puzzle. If there were lots of blue pieces then the picture has a lot of sky in it. Or sea. But the shade of blue would give an indication whether it was a seascape or landscape.

Of course, it’s not quite that straightforward and there are those assumptions I’ve already discussed (and a number I’m quietly ignoring to make everyones life easier) but, again, to a first approximation and with suitable quality control of the material analysed, it provides an indication of the range of species and their relative abundance in the sample.

Metabarcoding results

Unfortunately, several of the samples failed these quality control tests. Despite this, data was obtained from all 5 nests analysed, summarised below:

Wasps were consistently the most commonly found species in samples; they were present in all nests sampled and in 26/38 individual larvae analysed.

Honey bee sequences were also present in the majority of sampled larvae (25/38) but the overall abundance i.e. the proportion of sequences {{12}} derived from honey bees was low (0.3 – 17%, where present). This is shown graphically below:

Proportion of different insect types in diet from metabarcoding studies

As in the previous two studies discussed it appears as though bumble bees and solitary bees do not feature in the diet of Asian hornets; no reads from these species were present in the datasets.

This study shows the benefit of sampling the pooled material from Asian hornet larvae. However, it is limited in terms of the overall number of larvae sampled, the total number of nests analysed, and both the geographic and temporal separation of the samples. For example, no conclusions can be drawn about changes in the diet during the season, or differences in the diet dependent upon latitude or local environment.

Notwithstanding these caveats it again shows that the Asian hornet preys on a wide range of other insect species {{13}}, in particular social wasps, a wide range of flies, spiders as well as honey bees.

Honey bees on the menu

Since this is a beekeeping blog I’ll close with some more general comments on what I think we can conclude from these studies about the impact of the Asian hornet on honey bees and other insects.

The Asian hornet is an opportunistic generalised predator. If it can find and catch the prey, it will. The spiders in the Stainton et al., (2023) study were predominantly orb-web spiders, often very large and obvious in early autumn.

Along came a hornet

Potential prey items that are small, inconspicuous, rare or difficult to catch are unlikely to feature in the hornet’s diet. It is unlikely that the hornet will reduce biodiversity; as a species declines in number it will be found less often.

Conversely, prey that is large, obvious, numerous and concentrated are likely to be targeted. Flies around cattle or on carrion are two examples, but so are the 20-30,000 honey bees in a hive … or half a million in an apiary.

What explains the large differences between the proportion of honey bees in the diet of French (~38%) and Portuguese (~80%) Asian hornets and those from the UK and Channel Islands (0-17%, average 5%)?

It’s too early to be sure, but one significant difference is that the density of Asian hornet nests in France and Portugal was much higher as the hornets were already established there. In contrast, those in the UK and Channel Islands were either singletons or becoming established (and so presumably at much lower densities).

Perhaps with more Asian hornets colonies/nests competing with each other honey bees, being a common and concentrated source, will feature more heavily in the diet?

Finally, I would assume that in environments depleted already of insects e.g. some urban and intensively farmed areas, honey bees will be the plat du jour more often.

Future studies

One notable thing missing from all these studies is the impact on native insect biodiversity and abundance. It might be inferred, but it needs to be measured {{14}}. This, coupled with an analysis of hornet predation of honey bees in insect-poor vs. insect-rich environments and in areas with different hornet densities, will provide a much better guide to what beekeepers and ecologists can expect.

If hornet predation on honey bees is inversely related to the insect-richness of the environment then beekeepers should encourage landscape improvement that favour insect numbers and diversity {{15}}. Large numbers of hives in an otherwise barren environment will inevitably invite predation by Asian hornets.

Similarly, notwithstanding the negligible impact on solitary and bumble bees, other pollinators may be depleted by Asian hornets; simply trying to fill the gap with imported honey bees may well not work. Agricultural strategies that improve the environment for all pollinators could be a better approach … and one that both directly and indirectly benefits honey bees.

The next few years are likely to be interesting for those living in areas where Vespa velutina incursions occur annually, or in areas where the species becomes established.

Remember that there are several meanings of the word interesting.

Notes

Munchity crunchity is also a track from the album Rattle on the Stovepipe by Vancouver’s Celtic funk band Mad Pudding. Enjoy!

A version of this post appeared in the BeeFarmer magazine in October 2023.

References

Lioy, S., Laurino, D., Capello, M., Romano, A., Manino, A., and Porporato, M. (2020) Effectiveness and Selectiveness of Traps and Baits for Catching the Invasive Hornet Vespa velutina. Insects 11: 706 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7602873/.

Rome, Q., Perrard, A., Muller, F., Fontaine, C., Quilès, A., Zuccon, D., and Villemant, C. (2021) Not just honeybees: predatory habits of Vespa velutina (Hymenoptera: Vespidae) in France. Annales de la Société entomologique de France (NS) 57: 1–11 https://doi.org/10.1080/00379271.2020.1867005.

Schmack, J.M., Lear, G., Astudillo-Garcia, C., Boyer, S., Ward, D.F., and Beggs, J.R. (2021) DNA metabarcoding of prey reveals spatial, temporal and diet partitioning of an island ecosystem by four invasive wasps. Journal of Applied Ecology 58: 1199–1211 https://onlinelibrary.wiley.com/doi/abs/10.1111/1365-2664.13856.

Stainton, K., McGreig, S., Conyers, C., Ponting, S., Butler, L., Brown, P., and Jones, E.P. (2023) Molecular Identification of Asian Hornet Vespa velutina nigrithorax Prey from Larval Gut Contents: A Promising Method to Study the Diet of an Invasive Pest. Animals 13: 511 https://www.mdpi.com/2076-2615/13/3/511.

{{1}}: Well, my mind at least.

{{2}}: Perhaps it should be crunchity munchity?

{{3}}: A word that directly imitates a sound.

{{4}}: A particular sound or sound sequence that (at least in a general way) suggests a certain meaning.

{{5}}: I’m assuming here that the hornet has caught a flying insect … use your imagination if it’s a spider, or shrimp scavenged from a market stall.

{{6}}: So, not necessarily directly comparable.

{{7}}: Note how long it can take from fieldwork to publication.

{{8}}: And a similar number were carrying wood pellets for nest building.

{{9}}: !

{{10}}: I love these …

{{11}}: Actually, so rare they failed to cross the detection threshold applied.

{{12}}: Jigsaw pieces.

{{13}}: And, in the case of the Tetbury sample, one larvae contained reads from the hedgehog, presumably scavenged!

{{14}}: It has been on some studies of invasive wasps in New Zealand (Schmack et al., 2021).

{{15}}: They should do this anyway!