Adaptations evolve in populations

Organisms evolve adaptations to increase their fitness

There are few ideas in science that explain as much of the natural world as does natural selection, but there are few ideas in science that are more frequently misunderstood.

Often the misunderstandings are deliberate or disingenuous, but I’ve seen quotes like the one above even in undergraduate essays and popular science books.

There are two subtle problems with what is written above.

Firstly, adaptations do not evolve in individual organisms, because evolution isn’t something that happens to individual organisms. Evolution is something that happens to populations over time. A mutation that predisposes pea plants to making flowers that are purple will initially arise in a single, particular plant cell. But only if this plant’s offspring prosper (at the expense of peas with differently coloured flowers) will the population as a whole change over time.

Adaptations like flowers that attract pollinators – and the mutations that underlie them – arise in particular individuals, but they can only become more (or less) common in populations considered as a whole. Individuals mutate, but only populations evolve.

Biologists usually think of evolution as the change in frequency of an allele in a population over time. An allele is one of the various forms a particular gene can take. In the peas studied by Gregor Mendel, flower colour was determined by a single gene, which came in two variants: one variant – one allele – resulted in purple flowers, the other allele resulted in white flowers.

If – perhaps – the purple flowers were more attractive to pollinators like bees than the white, then the purple-flowered peas would tend to be fertilised more often, and would leave more offspring. The next generation of pea plants would then inherit a larger proportion of alleles causing purple flowers, and a reduced proportion of alleles causing white flowers. The purple allele would therefore increase in frequency at the expense of the white, and the population as a whole would become more purple over time.

Pisum sativum (white) [CC-BY-SA-3.0 Rasbak]

White-flowered peas (Pisum sativum) [CC-BY-SA-3.0 Rasbak]

Pisum sativum (purple) [CC-BY-SA-3.0 Rasbak]

Purple-flowered peas (Pisum sativum)  [CC-BY-SA-3.0 Rasbak]

For similar reasons, fitness is not a property of individual pea plants either; it is an average property of purple-flowered peas versus white-flowered peas. Any particular purple-flowered pea plant might get trampled by a cow and produce hardly any pea-pods; any particular white-flowered pea plant might get lucky and find itself in a cosy, well-fertilised spot near a beehive. But on average, the purple-flowered plants will do better and they will bequeath their purple-flower alleles disproportionately to the next generation: that is what we mean by the purple-flowered plants being ‘fitter’.

The second problem with the original statement is actually the smallest word: it’s that seemingly insignificant ‘to’. That ‘to’ is short for ‘in order to’, and there’s the rub. ‘In order to’ implies teleological agency and long-term purpose, which are not properties that you can reasonably attribute to natural selection.

Natural selection is a mindless algorithm, not a purposeful process. In essence natural selection is just this:

  • Organisms vary, and some of that variation is inheritable. Some peas have purple flowers, some have white, and this is (partly) due to purple- and white-flowered peas having different alleles of a particular flower-colour gene.
  • More offspring are produced than can survive and reproduce, so they will compete, and some variants will – on average – have greater success at making offspring than others. Pea plants produce more offspring peas than are ever likely to germinate, grow and get fertilised themselves. Some variants in the offspring may be better at competing for light, nutrients or pollinators: e.g. some flower colours may be more attractive to pollinators than others.
  • Those variants will contribute disproportionately to the next generation, and the frequency of their alleles will increase over time. If the purple-flowered peas are more attractive to pollinators, they will get fertilised more often, and contribute a larger number of seeds into the next generation. These seeds – and the plants that germinate from them – will inherit their parent’s purple-flower alleles, so we will see an increasing proportion of purple flowers as the generations tick over.
  • Rinse, repeat.

The match between the purple flowers and the purple preference of pollinators strikes us as clever and sensible and non-random, and it is all those things. But the process of natural selection that created this apparent design is just the mindless consequence of variation, heredity and differential reproductive success: it has no goals, no aims, no purpose.

Equally, this is a passive process for the pea population: they are subject to evolution by natural selection, but they are not active participants in the process.The peas do not strive supernaturally to make more purple flowers. The peas do not get magically instructed by the bees to make purple flowers. The peas certainly do not know what DNA sequences they would need to invent in order to make their flowers purple: new flower-colour alleles arise by mutation, which is just as mindless as natural selection, and even more heedless of what might be useful.

Populations do not evolve in order to do anything. Populations evolve, period. Purple flowers arise and become more common in pea populations because purple flowers happen to be attractive to bees; peas do not choose to evolve purple flowers in order to attract more pollinators. Mutation just throws up new alleles, and natural selection (and other important processes) make them more or less common over time.

It can become tedious describing this process passively and long-windedly every time: “purple flowers have become more common in the pea population because flower colour is highly heritable, and purple flowers are fertilised more frequently by bees”. If you want to write this more succinctly, you could say “purple flowers have evolved in peas because purple flowers are more attractive to bees” without loosing too much meaning.

Unfortunately, you might well see someone reword this as “peas have evolved purple flowers because bees prefer them”. This isn’t exactly wrong, but the ‘have evolved’ makes it sound very much like the peas are doing something actively, rather than having something done to them passively, and it’s all too easy to interpret this phrasing incorrectly. This puts us on the slippery slope to misconception.

If this is then reworded further to “peas evolved purple flowers to attract bees”, it sounds very much like like peas are active players in the process, and that they have a long term plan of improving their attractiveness to bees. This is almost guaranteed to mislead, unless you have your evolutionary biology filter turned up to maximum.

Do not write:

Some species evolved an adaptation to do some task

Write this instead:

An adaptation that does some task evolved in [a population of] some species

Given human nature (particularly on this subject), some people will still misinterpret your meaning, but at least you’re not implicitly misleading them with your language.


Recursion [CC-BY-SA-3.0 Steve Cook, Andreas Thomson, Frank Vinzent]

Recursion [CC-BY-SA-3.0 Steve Cook, Andreas Thomson, Frank Vinzent]


Bark [CC-BY-SA-3.0 Steve Cook]

Top left to bottom right: Betula utilis var. jacquemontii, Pinus sylvestris, Prunus serrula, Pinus pinaster, Quercus suber, Araucaria araucana, Metasequoia glyptostroboides, Pinus bungeana, Betula ermanii × pubescens, Pinus nigra ssp. laricio, Betula albosinensis cv. Red Panda, Platanus × acerifolia.

Wisley in Winter

It’s not really a botanic garden, but the Royal Horticultural Society’s gardens at Wisley is near enough as makes no difference. We visited in what should have been the dead of winter, but which in reality was this weird sprautumn mash-up that is now December in the UK. The heather garden was particularly pretty, despite the wind:

RHS Wisley heather garden [CC-BY-SA-3.0 Steve Cook]

RHS Wisley heather garden

At £12, the entry price is a bit steep: even Kew does reduced rates when most of the plants are underground, sleeping off the summer’s excesses.

Botanerd highlights included the biggest pine-cone I have yet added to my collection (yes, it is a collection, not a sickness)…

Pinus × holfordiana cone [CC-BY-SA-3.0 Steve Cook]

Holford pine (Pinus × holfordiana) cone – nothing satisfies like a 12 inch Pinus

…and the – apparently legal – purchase of class A drugs from the attached garden centre:

Lophophora williamsii [CC-BY-SA-3.0 Steve Cook]

Peyote cactus (Lophophora williamsii)

From what I gather, owning a peyote cactus is legal in the UK unless you ‘prepare it for consumption’, which magically transforms it into a 7 year jail term. This presumably includes accidentally letting it die and dry out into a button whilst on holiday. I may be having a teesy bit of buyers’ regret.


Most years the pond-water microscopy practical throws up an exciting ciliate (or two, or three), but this year, the only ones we saw were duplicates, or too bloody fast to photograph. Ho hum.

So this year you’ll have to make do with an imposter. It’s still got cilia, but it is not a ciliate, and although it’s no bigger than a single-celled organism, it is in fact a full-blown multicellular animal:

Euchlanis rotifer [CC-BY-SA-3.0]

Euchlanis rotifer

This is a rotifer (‘wheel animal’), which are famous for two things: being ridiculously small, and – in one infamous subgroup – being ridiculously averse to having sex. They’re also really cute when they turn on their feeding wheels. Happy New Year.

Bagging botanical Brussels

The last time we went to Brussels, I got terribly excited that the hotel we were staying in was right next door to the Botanical Garden of Brussels. Unfortunately – as we discovered in short order – at some point in the 1930s the plants had mostly been shipped off elsewhere, leaving the garden not very botanical, and Dr Cook not very impressed.

That elsewhere was the Plantentuin Meise / Jardin Botanique Meise, which is to the north of Brussels, conveniently situated – for the purposes of this second attempt to visit them – between Brussels airport and Antwerp.

Brussels botanic gardens [CC-BY-SA-3.0 Steve Cook]

Botanic Gardens, Meise

We were visiting in autumn, and they were in the middle of renovating part of their main glasshouse – the Plant Palace – but there was still plenty to see, including a good arboretum, an orchid exhibition, and some lovely Japanese beautyberries:

Callicarpa japonica [CC-BY-SA-3.0 Steve Cook]

Japanese beautyberry (Callicarpa japonica)

The entry fee (€ 7) is very reasonable, and the restaurant helpfully ensures all sandwiches come with a colonic depth-charge of salad; just the ticket to cure the faecal impaction resulting from the ‘food’ served to us on the flight out with <insert awful British airline here>. When we were there, there was also an art installation by Roos Van de Velde in praise of the best plant in the world – the maidenhair tree Ginkgo biloba  but you’ve only got until mid-January to catch it:

Gingko biloba Meise [CC-BY-SA-3.0 Steve Cook]

Roos Van de Velde’s Gingko biloba installation

Botanerd highlights:

  1. Whereas Kew’s evolution house (currently being renovated as part of the Temperate House restoration) always felt like a holding pen for a few cycads and ferns that didn’t physically fit into the Palm House, the Meise version was much more complete, with a good selection of horsetails, ferns, and high-quality fakes of fossil plants like Cooksonia, Lepidodendron, etc. That said, I would very much like to see these ‘evolution house’ efforts re-branded as ‘phylodiversity houses’ or similar, as the usual Scala Naturae bullshit of using modern mosses, lycopods, ferns, and conifers as stand-ins for prehistoric forms gets right up my arse.
  2. The pinetum is small, but has a good selection of species, including quite a few weird cultivars of Japanese cedar (Cryptomeria japonica) that I’ve not seen before.

Cryptomeria japonica cv. Cristata [CC-BY-SA-3.0 Steve Cook]

Cryptomeria japonica cv. Cristata

Previous baggings…

Organism of the week #29 – Galling

Today is the first day of the new (academic) year at $WORK, but – aside from a couple of intro lectures – this is the calm before the real storm, which arrives in the form of a deluge of biological chemistry in November. If you’re lucky, you’ll get a few mugshots of some weird ciliate or other around then, otherwise, expect tumbleweed until the new (calendar) year…

Insects and flowering plants have been co-evolving for 200 million years or so, so it’s unsurprising that many of their interactions have become very intimate. The relationships between gall-forming insects and their host plants are particularly interesting. Gall-forming insects lay their eggs on plants, but when the larvae hatch, instead of munching through the plant in the usual fashion, the larvae make the plant build bizarre little houses for them. These are the galls. Galls provide protection from weather, some protection from things that might like to eat the gall-former itself, and food for the developing larva.

In the UK, there are many gall-forming insects that target oak trees. These are mostly very small wasps. One of the most obvious is the knopper gall, which are the sticky crater-shaped things you often see attached to acorns around August.

Rather like the fluke we met a few weeks ago, knopper gall wasps have a complicated life-cycle involving several different hosts, in this case two different species of oak: the English (Quercus robur) and the Turkey (Quercus cerris). Female wasps emerge in spring from the galls that fell from English oaks the previous autumn. These females then lay asexual eggs on the catkins of the Turkey oak, which hatch out, and form rather inconspicuous galls. The males and females that hatch from these galls in April then mate, and the females lay their eggs on the developing acorns of English oaks, where they develop into the knopper galls in late summer.

Andricus quercuscalicis on Quercus robur [CC-BY-SA-3.0 Steve Cook]

Knopper galls caused by the larvae of the wasp Andricus quercuscalicis, on oak (Quercus robur)

Like so many pests and diseases – the knopper gall is actually a fairly recent introduction: the first galls in the UK were recorded only in the 1950s.

Smaller but creepier, the silk-button spangle gall is a trypophobes’ delight. Like the knopper gall, they’re caused by small wasps, but are much less obvious until you flip a leaf over:

Neuroterus numismalis and Neuroterus quercusbaccarum on Quercus robur [CC-BY-SA-3.0 Steve Cook]

Mostly silk-button spangle galls caused by the larvae of the wasp Neuroterus numismalis, also on oak

The less creepy common spangle gall isn’t as likely to make you feel like you have maggots crawling under your skin, but like the silk-button spangles, that’s exactly what the plant has got.

Neuroterus quercusbaccarum on Quercus robur [CC-BY-SA-3.0 Steve Cook]

Common spangle-galls (Neuroterus quercusbaccarum) on – you’ll never guess what. There are a few older reddish spangles in the silk-button image above too.

The most famous and most useful gall found on oak is the oak apple. A mixture of mashed up oak apple galls and iron salts was – for a millennium and a half – what you dunked a quill pen into, if you were lucky enough to know how to write.

Oak apple [CC-BY-SA-2.0 Bob Embleton]

Oak apple gall: these are caused by the larvae several different species of by cynipid wasps [CC-BY-SA-2.0 Bob Embleton]

Like the knopper and spangle galls, oak apples are caused by the larvae of a tiny wasp, but it’s not just wasps that can cause galls, or indeed just insects. Roundworms (nematodes) can also form galls, usually on the roots of plants; as can some fungiMites, which are more closely related to related to spiders than to insects, are responsible for many galls of sycamores and other maples:

Vasates quadripedes on Acer saccharinum [CC-BY-SA-3.0 Steve Cook]

Bladder galls caused by the mite Vasates quadripedes on – gasp, not oak! – sugar maple (Acer saccharinum)

Aceria macrorhynchus on Acer pseudoplatanus [CC-BY-SA-2.0 Lairich Rig]

Mite galls caused by Aceria macrorhynchus on sycamore (Acer pseudoplatanus) [CC-BY-SA-2.0 Lairich Rig]

Why would a plant build a house for a parasitic insect or mite? The long answer is very long, but the short answer is that the parasite secretes plant hormones called cytokinins to manipulate the plant into forming the gall. In plants, cytokinins cause cell division and growth. They also act as “feed me” markers: sugars and other nutrients are preferentially directed towards tissues that have cytokinin in them. Cytokinins also cause plant cells to make red pigments. By secreting cytokinins (and other plant hormones) gall-forming insects, mites, fungi and nematodes manipulate the plant into building small – often red! – houses for them, and to supply those houses with sugar.

Zeatin [Public domain]

cytokinin hormone: the structure is a modified form of the DNA base adenine

The galls formed by insects and mites are amazing structures, but at a molecular level, the most extraordinary galls of all are formed not by insects or mites, but by bacteria:

Crown gall [CC-BY-SA-3.0 Bhai]

Crown gall on Kalanchoe, caused by the bacterium Rhizobium radiobacter (Agrobacterium tumefaciens) [CC-BY-SA-3.0 Bhai]

Crown galls may not look so impressive, but the way in which their inhabitants cause the gall to be formed most certainly is. The galls are formed by the bacterium Rhizobium radiobacter[For any biologists doing a double-take at this point: until recently Rhizobium radiobacter was known as Agrobacterium tumefaciens , but it’s been reclassified. This is Brontosaurus all over again…]

Rhizobium radiobacter is the world’s smallest genetic engineer. Rather than simply secreting plant hormones to make the plant direct food at it, like the gall-forming insects do, this bacterium actually injects a piece of DNA into the plant’s cells to bend them to its will. The DNA encodes enzymes for plant hormone synthesis, which causes the plant cells to grow into a cancerous mass: the crown gall.

But even that’s not enough for this bacterium. The nutrients directed into the gall by the plant are not quite to the bacterium’s taste, so the DNA it injects also encodes genes that turn plant cells into factories for bacteria-friendly food. This food even leaks out of the plant’s roots, feeding the bacterium’s sisters in the soil.

It took until 1982 for humans to develop the technology to genetically manipulate a crop plant to make better food. A mindless single-celled organism has been pulling off the same trick for millions of years. There’s nothing new under the sun.

Organism of the week #28 – Fractal art

Flowers are essentially tarts. Prostitutes for the bees. (Uncle Monty, Withnail and I)

Our tiny garden has only passing acquaintance with sunshine, so about the only plants that really thrive in its dingy clutches are shade-loving ferns. This Japanese painted fern is my current favourite: who needs flowers anyway, when leaves look like this?

Athyrium niponicum cv. Pictum [CC-BY-SA-3.0 Steve Cook]

Japanese painted fern (Athyrium niponicum cv. Pictum)

The colour is spectacular, but I also love the shape of the leaves. I have a rabbit’s-foot fern on my windowsill at $WORK, which branches not once, not twice, not thrice, but fwice, with each leaflet at each level looking very much like a miniaturised version of the whole leaf:

Humata tyermannii [CC-BY-SA-3.0 Steve Cook]

Humata tyermannii

Self-similar shapes like this are called fractals. This fern’s leaf is not a true mathematical fractal, because its self-similarity goes only four levels deep, rather than infinitely so. Also, from a developmental point of view, the leaves of ferns are produced by interactions between signalling molecules that have very little in common with the maths of fractals. However, it is remarkable how easy it is to produce a fractal that looks very much like a fern – especially when someone else has done all the hard work of discovering one for you. This one is called the Barnsley fern fractal, after Michael Barnsley:

Barnsley fern [Public domain]

Barnsley fern

This rather cute simulator allows you to play about with the Barnsley fern without going to the hassle of coding up any R yourself, but for those who enjoy such things, here’s my version of the code written in R, based on the coefficients from its Wikipedia article. An entirely imaginary prize is available for anyone who can produce a good simulation of either of the two ferns above.

fm <- vector( "list", 4 )
fc <- vector( "list", 4 )
fp <- vector( "list", 4 )

# Stem
fm[[1]] <- matrix( c( 0.00, 0.00, 0.00, 0.16 ), nrow=2, byrow=TRUE )
fc[[1]] <- matrix( c( 0.00, 0.00 ), nrow=2 )
fp[[1]] <- 0.01

# Leaflets
fm[[2]] <- matrix( c( 0.85, 0.04, -0.04, 0.85 ), nrow=2, byrow=TRUE )
fc[[2]] <- matrix( c( 0.00, 1.60 ), nrow=2 )
fp[[2]] <- 0.85

# Left largest leaflet
fm[[3]] <- matrix( c( 0.20, -0.26, 0.23, 0.22 ), nrow=2, byrow=TRUE )
fc[[3]] <- matrix( c( 0.00, 1.60 ), nrow=2 )
fp[[3]] <- 0.07

# Right largest leaflet
fm[[4]] <- matrix( c( -0.15, 0.28, 0.26, 0.24 ), nrow=2, byrow=TRUE )
fc[[4]] <- matrix( c( 0.00, 0.44 ), nrow=2 )
fp[[4]] <- 0.07

n<-250000 # you might want to make this smaller if it takes forever to render
x <- numeric( n )
y <- numeric( n )
x[1] <- 0
y[1] <- 0

for( i in 1:(n-1) ) {
  choice  <- sample( 1:4, prob=fp, size=1 ) <- fm[[ choice ]] %*% c( x[i], y[i] ) + fc[[ choice ]]
  x[i+1]  <-[1]
  y[i+1]  <-[2]

plot( x, y, cex=0.05, col="dark green" )

Living dangerously

Towards the end of the last millennium, I spent a lot of time befriending arsenic. The last two years of my PhD involved measuring how much iodine had fallen off a chemical used as a wood preservative day-in, day-out, and arsenic trioxide was the most exciting component of an otherwise excruciatingly dull test for iodide ions.

My little bottles of arsenic warned me to beware of pirates…


…for there were enough pirates in each 10 gram bottle to kill me ten times over. I apparently managed to handle the stuff safely, as – at least so far – I’m not dead even one time over.

Arsenic became so popular as a detergent for getting stubborn relatives out of the family that by the late nineteenth century people started referring to it as “inheritance powder”.

A (not very) cheap and (exceptionally) nasty way to work out how poisonous something like arsenic is, is to take ten cages containing ten rats each, and give each cageful of rats an increasingly large dose of the poison in question. The individual rats will differ in how easy they are to kill with the poison, but the rats in the cages that get very small doses will probably all survive, and the rats in the cages that get very large doses will probably all die. Somewhere in the middle you’ll probably find a cage containing 5 corpses and 5 survivors. The dose of poison that kills half the rats in a cage is called the “lethal dose, 50%” or LD50. The LD50 for arsenic trioxide is about 10 milligrams for rats.

Humans are a lot heavier than rats: I’m 60-odd kilograms on a good day, a rat only half a kilo or so. Bigger animals are can survive higher doses of poisons  because the poison ends up diluted into their larger bodies. So  LD50 values are usually scaled for the size of their victim. If 10 mg kills a 0.5 kg rat, the LD50 is 20 mg kg−1. By maths and educated guesswork about the relative susceptibilities of rats and humans to arsenic (apparently the definitive experiment is unethical), the LD50 for a human (this human, specifically) is just over 1 gram.

Arsenic is pretty poisonous, but some other poisons beat it hands-down. Botulinum toxin, the active ingredient in Botox injections, has an LD50 of 1 ng kg−1 which is about 10 million times more lethal. But, as even 16th century proto-doctors noted, it’s the dose that makes the poison. Everything is poisonous in sufficient quantity, even things as seemingly innocuous as water.

Even the most child-proofed kitchen could kill you three times over. You just need to put in a little effort:

Water LD50 [CC-BY-SA-3.0 Steve Cook]

Lethal dose of water: 6 L

The LD50 of water is about 100 mL kg−1, at least in rats. For a runt like me, that means around 6 L of water drunk over a short period of time would be lethal, mostly by diluting the salts in my blood and tissue fluid to the point where water would flood into my cells, making them swell and pop.

The opposite effect could be achieved by eating too much common table salt, which would suck water out of my cells, causing them to shrivel. With an LD50 is about 3 g kg−1, pretty much every kitchen in the land has a pot of salt in it containing at least one lethal dose:

Salt LD50 [CC-BY-SA-3.0 Steve Cook]

Lethal dose of salt: 200 g

It’s pretty difficult to eat even a teaspoon of neat salt without puking, so I shouldn’t worry about anyone trying to get their inheritance early by force-feeding you table salt rather than good old arsenic, but it might just be possible to get someone to eat the necessary dose of sugar. With an LD50 of 30 g kg−1, that would mean eating 2 kg of the stuff, but it’s certainly a sweeter way to go than many.

Sucrose LD50 [CC-BY-SA-3.0 Steve Cook]

Lethal dose of sugar: 2 kg

Organism of the week #27 – Alien haemorrhoids

A quickie this week: magnolia fruit look like something from another planet.

Magnolia ×soulangeana fruit [CC-BY-SA-3.0 Steve Cook]

Magnolia ×soulangeana fruit

Load more