The Origin of Modern Biodiversity: Coevolution of Flowers and Insects

29 06 2011

This talk is split into two major parts: the first will look at the general fossil record of insects, and the second will introduce the flowering plants and their interactions with insects.

Due to the constructive feedback received in the first two days of talks, I decided to do last-minute changes to the talk (applies to the next one as well). While these proved to be more accomodating to the audience, they also resulted in a loss of general structure. I hope it isn’t too noticeable though :)

So we begin by examining how insects can be fossilised. We find fossil insects in all sorts of deposits. The vast majority of insect fossils are isolated wings; due to the technical nature of discussing those, I will stick to the more spectacular body fossil deposits, all of which can rightly count as Konservat Lagerstätten.

Lacustrine (lake) deposits commonly contain insects. Here are some famous localities in Germany (which has many sites of exceptional preservation anyway). Rott is a locality with diatomitic rocks, preserving insects of all orders (not only aquatic ones), as well as spiders. Messel – most well-known from its fossil horse and Ida (the basal lemur) – also has its own insects, preserved with structural body colours. Enspel is more known for its mammals, but it also has its own share of insects.

Marine deposits rarely preserve insects. The Moler of Denmark is an exception, with thousands of insects of all orders, from crickets to bees to giant ants, being preserved there. Solnhofen, the locality where Archaeopteryx comes from, also has several insects.

Terrestrial deposits include many Ice Age and Quaternary-aged localities, but also the most famous insect fossils: the various ambers.

Amber is simply fossilised tree resin: some trees, when damaged or under stress, release a sticky resin that flows down the trunk. Anything on the tree trunk, or flying on to the resin, or lying on the ground near the tree and its roots will get stuck in it. This includes insects, but also feathers, hairs, snails, etc.

This also highlights that even amber has its preservational biases: only forest fauna is typically preserved in it. The one limitation to amber is that the inclusions cannot be prepared out. In fact, besides the chitin exoskeleton, the inclusion is hollow, meaning that even if the insect were to be extracted, it would fall apart immediately.

Besides incredible morphological detail, amber can also preserve behavioural interactions. Many copulations have been found in amber, both during the act (left) and in preparation for it (right).

Even cooler, parasitic interactions can also get preserved in amber. Top left is a fly with an endoparasitic nematode worm bursting out of its abdomen.

The centerpiece of this slide is something quite remarkable though. These two flies are clearly in copula, but they are not even of the same family: the male part is a chironomid, the female a simuliid.

That’s not all though. The chironomid is infected by nematodes, like the fly in the top left. Judging by the abdomen’s size, they are near the completion of their endoparasitic life cycle and about to burst out. A dipterologist will immediately point out another strange character in the chironomid: while it is a male, its antennae are of the female type (male chironomid antennae are bushy).

Such an alteration of sex is known nowadays from chironomids as the result of nematode parasitism (specifically, the Mermithidae family). That this is preserved in >40 Ma amber is not only spectacular, it also indicates that this parasitic interaction has been around for a very, very long time.

Source: Eichmann, F. 2003. Aus dem Leben im Bernsteinwald. Arbeitskreis Paläontologie Hannover 31, 89-94.

To finish off with amber, this diagram summarises the geographical and temporal distribution of most amber (and select other important insect fossil) localities. As you can see, we have a pretty good record through time and space just with these sites of exceptional preservation, allowing us to trace insect diversification from the time when flowering plants started rising in popularity.

Diagram source: Schmidt, A. R., Perrichot, V., Svojtka, M., Anderson, K. B., Belete, K. H., Bussert, R., Dörfelt, H., Jancke, S., Mohr, B., Mohrmann, E., Nascimbene, P. C., Nel, A., Nel, P., Ragazzi, E., Roghi, G., Saupe, E. E., Schmidt, K., Schneider, H., Selden, P. A. & Vávra, N. 2010. Cretaceous African life captured in amber. PNAS 107, 7329-7334.

The pictures above show the earliest body fossils of flowering plants (angiosperms). They are both from the same Cretaceous locality of Yixian, China.

The specimen on the left, Leefructus mirus, is newly-described and is noteworthy due to it being relatively derived, being placed confidently on the stem-group of the buttercups.

The one on the right is Archaefructus, long known as the earliest angiosperm.

Leefructus: Sun, G., Dilcher, D. L., Wang, H. & Chen, Z. 2011. A eudicot from Early Cretaceous of China. Nature 471, 625-628.

Archaefructus picture source: Friis, E. M., Doyle, J. A., Endress, P. K. & Leng, Q. 2003. Archaefructus – angiosperm precursor or specialized early angiosperm? Trends in Plant Science 8, 369-373.

As is usually the case with plants, spores and pollen are by far more numerous than body fossils (mostly because they are much more easily preserved) and they show that by the mid-Cretaceous, angiosperms had become the dominant plants on the landscape, despite having seemingly originated at the beginning of the Cretaceous/Late Jurassic – they basically rose to dominance in a little over 20 million years.

Diagram source: Lidgard, S. & Crane, P. R. 1988. Quantitative analyses of the early angiosperm radiation. Nature 331, 344-346.

Such a quick radiation requires an explanation, and the most sensible one is simply to say that angiosperms are much more suited for quick colonisation. The other dominant faunas of the time were cycads and conifers, both slow-growing plants. On the other hand, angiosperms can grow very quickly, allowing them to reach sexual maturity quickly so they can spread their offspring into a wider area where they can settle and colonise in record time. They basically managed to kick all the other plants out, as can be seen by the sharp decline in cycads and pteridophytes (both gymnosperms) as angiosperm numbers rose.

Angiosperms nowadays are the most species-rich plant group. However, only select groups are especially species-rich, and these are the ones that have the following traits:

  • Biotic pollination: Using insects as pollinators (some use bats or birds)
  • Herbaceous: grow quickly
  • Defensive: produce chemicals to ward off predators
  • Flexible: Can tolerate fluctuating climate conditions

These traits support the scenario already outlined for the angiosperm radiation, and is termed the palaeoherb hypothesis.

The palaeoherb hypothesis stands in contrast to the magnoliid hypothesis of angiosperm origin, where the magnoliids are supposed to be the most basal angiosperms. However, the basalmost angiosperms, as can be seen above, are not the magnoliids, but instead are what are referred to as the ANITA grade angiosperms.

As a sidenote, which gymnosperm group the angiosperms originated from is, at the moment, still contentious, though most authors gravitate towards the bennettitaleans, an extinct group of cycad-like plants.

This is just further confirmation of the rapid-life-cycle scenario, as the ANITA grade flowers are weed-like herbs.

These are the ANITA grade flowers.

The basalmost is Amborella, an endemic of the New Caledonian cloud forests. A familiar flower is Nuphar, a water lilly.

As can be seen, the arrangements of the flowers here are very simple – compare these to the complex elegance of a rose, for example.

Flowers are a model system for developmental biology and evo-devo due to the ABC model. The letters refer to master developmental genes (akin to animalian Hox genes) that control aspects of floral development:

  • A genes control sepal development
  • B + A genes control petal development
  • B+C genes control stamen development
  • C genes control carpel development

How these evolved and their expression patterns changed is the key to unlocking flower evolution.

One thing that is very important to understand is that flowers ostensibly have no purpose but to attract organisms. And that this function has been preserved throughout their evolution by these master genes is further proof that coevolution is the largest factor that allowed flowering plants to diversify so much.

Diagram source: Friis, E. M., Pedersen, K. R. & Crane, P. R. 2006. Cretaceous angiosperm flowers: Innovation and evolution in plant reproduction. Palaeo3 232, 251-293.

And naturally, the flowers coevolved with the most diverse taxonomic group on the planet, the insects. On the picture above, you will notice that the most diverse insect groups are also the ones that contain pollinators.

And the relation is easy to see. Flowers are colourful not only to us, but to insects as well. In fact, insects see flowers dfferently than we do: flowers that appear completely white to us may appear to insects as having dark lines. These are painted on with pigments that reflect UV light, which insect eyes can pick up, but not human ones.

These dark lines guide the insect towards the inside of the flower, where they (most of the time) get a reward of nectar, a sweet sugary drink that many insects love and need (very energetic).

Besides looks, flowers also smell nice, another way to appeal to insects.

From an evolutionary biology point of view, coevolutions come in many patterns. One that is seen in the insect-flower coevolution is an antagonistic coevolution pattern termed the escape and radiate pattern.

This is exemplified by the leaf beetle genus Blepharida and its host plant Bursera. The initial Blepharida species came across Blepharida and decided to take it on as a host plant, and radiated as no other insects were using it.

Eventually, Bursera develops a new chemical defence to deter the Blepharida species, allowing it to escape and radiate, free from the danger and stress posed by Blepharida.

However, eventuallysome Blepharida will develop a resistance to the novel chemical defence, allowing it and its progeny to colonise the new Bursera species and radiate.

Repeat ad nauseam.

However, the pattern most commonly seen in the insect-flower coevolution is the mutualistic one, exemplified by pollination.

As an outstanding example of this, take the case of Darwin’s orchid, pictured right. When Charles Darwin first encountered this species (Angraecum sesquipetale), he predicted that a moth species would be discovered as its sole pollinator – that green tube that drops down to the ground in the picture is where the moth would stick its proboscis to get to the nectar.

Darwin’s prediction was completely right: in 1902, Xanthopan morganii praedicta was discovered. Its proboscis was the longest in the insect world, it was also endemic to Madagascar and was observed pollinating the orchid.

How did Darwin know this moth would exist? The fact is that, as I said already, flowers are structured around their insect pollinators and both flower and pollinator need each other to survive, hence the mutualism. Some flowers are more specific than others, orchids being the more extreme cases of this.

And it is so that 85% of all angiosperms are pollinated by insects. I’ve already discussed the flower side of things. The insects also have several adaptations to the lifestyle, again some more specific than others.

For example, corbiculate bees, including the honeybees, are characterised by distinct basket-like structures on their hindlegs, called the corbicula. These are basically pollen baskets and serve no other purpose than to transport pollen. This is besides the general fluffiness of the bee body – insect-pollinated flower pollen sticks to those too.

Pollinator beetles take that hair to an extreme (of sorts), as can be seen in the middle picture.

It may seem that lepidopterans (moths and butterflies) are selfish and only suck out the nectar with their long proboscis, but they too pick up pollen on the way.

As for the fossil record of pollination, that is available as well. For a more spectacular example, consider the one above.

The pictures in the left column are a fossil from the Isle of Wight; on the right are modern analogues. These are fig wasps carrying fig pollen.

The fig that you eat is not a fruit, it’s a a collection of tiny flowers inside a fleshy compartment. The fig tree is pollinated only by the fig wasp. The dig “fruit” has a hole in it which only the fig wasp can widen just enough for her to lay her eggs inside. The larvae develop inside the “fruit” – i.e. inside a flower field – getting covered with pollen the whole time, while they feed on the fleshy outside bit. When they reach the adult stage, they fly off to mate. The females find other fig “fruit” to lay their eggs in. In the process of doing that, they will pollinate the fig tree, since the pollen is still stuck on their bodies.

This is another case of fossils showing just how ancient some associations are. In this case, 34 Ma.

Source: Compton, S. G., Ball, A. D., Collinson, M. E., Hayes, P., Rasnitsyn, A. P. & Ross, A. J. 2010. Ancient fig wasps indicate at least 34 Myr of stasis in their mutualism with fig trees. Biology Letters 6, 838-842.

For a more general case of insect-flower specificity, consider the many orchids that are sexually deceptive. If you notice the bee above, it’s going into the flower the wrong way – if it wanted nectar, it would have gone in head-first. What this bee is doing is trying to copulate with the orchid (pseudocopulation).

These orchids have evolved to mimic the females of their pollinators either by look or by smell (by releasing chemicals identical to sex pheromones). These associations are species- (or species-group) specific. The diversity of the orchids is directly linked to this sexual deception.

And we can discuss such specific interactions for hours, until we’re done with every plant species. The one thing they have in common is that they are all the result of the coevolution of flowers and insects, an event that is still ongoing and will continue to go on until the angiosperms die out (because, let’s face it, insects are unlikely to ever die out as long as the planet is still here).

Jump to: The Origin of Life, The Rise of Animals, Terrestrialisation, Mesozoic Vertebrates, The Tertiary

Research Blogging necessities :)
Compton, S., Ball, A., Collinson, M., Hayes, P., Rasnitsyn, A., & Ross, A. (2010). Ancient fig wasps indicate at least 34 Myr of stasis in their mutualism with fig trees Biology Letters, 6 (6), 838-842 DOI: 10.1098/rsbl.2010.0389

Lidgard, S., & Crane, P. (1988). Quantitative analyses of the early angiosperm radiation Nature, 331 (6154), 344-346 DOI: 10.1038/331344a0

Friedhelm Eichmann (2003). Aus dem Leben im Bernsteinwald Arbeitskreis Paläontologie Hannover, 31 (4), 89-94



16 responses

29 06 2011

What is Coevolution in Biology?…

From the Wikipedia article ( ), we get an excellent definition and numerous examples. > is “the change of a biological object triggered by the change of a related object.”[1] Coevolution can occur at many biolo…

19 11 2011

What is it like to attend the Universität Bonn?…

At least for the sciences, it’s a great place to be. The sciences have their own campus (with the exception of a couple of buildings, such as the Institute for Evolutionary Biology and the Max Planck Institute for Radioastronomy), and the humanities a…

31 12 2011
2011: Summary « Teaching Biology

[…] The Origin of Modern Biodiversity: Coevolution of Flowers and Insects. Also takes the award for longest post […]

15 02 2012
Top Papers of the Week: 06.02-12.02 « Teaching Biology

[…] What we recognise as butterflies are actually members of three superfamilies, collectively referred to as the Rhopalocera: the Papilionoidea (most butterflies), Hesperioidea (skippers), and Hedyloidea (“moth-butterflies”). This study confirms that this three superfamily splitting is erroneous, with Papilionidae (member of the Papilionoidea) lying at the base of a clade that includes all three superfamilies. Hence, the correct way to refer to split up the butterflies is to refer to all of them as Papillionoidea, including the hesperioids and hedyloids, rendering the term Rhopalocera synonymous with Papilionoidea. Personally, I think the analysis is fairly robust, and I trust this result. The other big part of the paper concerns their diversification through time, employing the molecular clock. The results largely agree with what the fossil record has to say, with both this paper and the record suggesting extensive diversification after the KT event, most probably due to the associated angiosperm radiation. […]

20 02 2012
Top Papers of the Week: 13.02-19.02 « Teaching Biology

[…] chambers are known in the fossil record, including from the critical Cretaceous period when insect biodiversity exploded. I can’t give a list of how many insects pupate in wood, but all holometabolous orders have […]

8 10 2012
Most Interesting Papers of the Week « Teaching Biology

[…] The coevolution between insects and flowering plants is a well-known phenomenon, given that it’s what led to the bulk of today’s biodiversity. This paper goes to the population level of a coevolution, finding that plants have an evolutionary response to insect predation even on short-term ecological timescales (i.e. “immediately”). Very cool. See a commenatry on it here. […]

28 10 2012
Most Interesting Papers of the Week « Teaching Biology

[…] mentioned fig pollination as an example of co-evolution in my insect-flower co-evolution post, and this spectacularly thorough paper reinforces the example’s status as the best and most […]

2 12 2012
Interesting Papers of the Week « Teaching Biology

[…] can check this post for information on the rise of flowering plants. This paper adds a new dimension as to why […]

23 12 2012
Most Interesting Papers of the Week « Teaching Biology

[…] gave an overview of the origin of angiosperms (flowering plants) in this post. Tl;dr version: the earliest angiosperms had very quick life cycles and so grew faster than the […]

29 12 2012
Top Research of 2012: Botany « Teaching Biology

[…] fossils from Europe to propose a temporal framework for their megadiversification in the Cretaceous in which they outcompeted the conifers and other land plants. The scenario Coiffard et al. synchs up with the view from the North American record, so it can be […]

13 02 2013
Daily Factoid: Some insects parasitise plants « Teaching Biology

[…] What’s as interesting as this modification of plant structure by parasitic insects (which is, in principle, similar to behavioural modifications by animal parasites), many studies suggest that the involved cytokinins are not of plant origin, but injected directly by the insects. Going even deeper, the cytokinins injected by the insects aren’t produced by the insects, but by bacterial symbionts. Draw that on a flowchart to see how cool this is: it’s a triple-layered interaction that’s been most likely moulded by millions of years of coevolution. […]

5 05 2013
business continuity

There is definately a great deal to learn about this issue.
I like all of the points you have made.

2 06 2013
Evolutionary Biology: Introduction To Evolution As A Science | Teaching Biology

[…] conception of insect behaviour and evolution would vary – we wouldn’t know just how deep the coevolution between flowers and insects […]

4 01 2014
Pieridae (Lepidoptera) as Model Organisms | Teaching Biology

[…] reason is for one of my pet topics, coevolution of plants and insects. The cladogram above is enhanced with two additional data sources: one is a molecular clock-based […]

20 08 2016

Hi Marc Srour! Thank you for your approach to this topic concerning both insects and flowers. I am delighted by the breadth and detail of your post. If you would like to guest-post on my website(s), please let me know.

I have recently published an article of a similar topic to my website at, and would appreciate your leaving me some feedback there.

Thank you in anticipation.

20 08 2016

By the way, an almost identical article is also available in my other website at

Leave a Reply

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

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

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s

%d bloggers like this: