As I was reading Freeman Dyson‘s recent collection Dreams of Earth and Sky, specifically his essay on democratizing genetic manipulation, I came across the following passage:

“For a plant growing in a hot climate, it is advantageous to reflect as much as possible of the sunlight that is not used for growth.  There is plenty of sunlight, and it is not important to use it with maximum efficiency.  The plants have evolved with chlorophyll in their leaves to absorb the useful red and blue components of sunlight and to reflect the green.  That is why it is reasonable for plants in tropical climates to be green.  But this logic does not explain why plants in cold climates where sunlight is scarce are also green.  We could imagine that in a place like Iceland, overheating would not be a problem, and plants with black leaves using sunlight more efficiently would have an evolutionary advantage.  For some reason that we do not understand, natural plants with black leaves never appeared.  Why not?  Perhaps we shall not understand why nature did not travel this route until we have traveled it ourselves.”

I was puzzled why he posed this question.  Dyson has previously addressed theories as to the origin of life (in fairness, perhaps I should quote a sentence from the beginning of that work: “This apology for a physicist venturing into biology will serve for me as well as for Schrodinger, although in my case the risk of the physicist making a fool of himself may be somewhat greater.”  Dyson makes very clear that he’s addressing these topics speculatively), so I would’ve expected a more nuanced treatment of evolution.

Evolution is not an engineer.  Evolution is a tinkerer.  If something is broken, or misfit, or maladjusted, it won’t be scrapped.  The energetic barrier is too high to expect nature to routinely start again from nothing and construct a suitable replacement.  If evolution stumbles across something that seems broken, it will be fixed, fixed poorly, fixed with duct tape and twine until it’s just barely adequate for its intended function.

For natural selection to give rise to a plant that uses silicon to harvest sunlight (which is why Dyson was discussing plants with black leaves above.  He was speculating upon the possibility that genetic manipulation could yield plants with silicon-based rather than chlorophyll-based sunlight collection, potentially resulting in a higher energetic yield), numerous processes must be optimized.  Extraction of silicon from the earth.  Incorporating it into an electron transport chain.  Etc.  But until all of that is somewhat functional, there won’t be any selective pressure.  Why would generation after generation of offspring be enriched for improved silicon uptake if there is no beneficial use for silicon?  Why would genetic sequences coding for protein folds able to cradle a silicon cofactor be enriched in the absence of silicon, or any way to use the energy thus captured?

(Hint: they won’t be.)

It’s possible that this sounds vaguely reminiscent of the claims made by creationists.  That evolutionary processes cannot enact macroscale changes because too much needs to go right all at once.  But there’s a difference.  In the absence of chlorophyll-producing plants, given sufficient time, a species might evolve a means to use silicon to capture light from the sun.  But, for a plant to do it?  This is dramatically more difficult than for an organism to start from scratch.  Plants already have a system that works well enough. That pre-existing system suppresses innovation.

For instance, I’m typing this essay on a QWERTY keyboard.  The layout was designed to slow my fingers down, ensure that the spindly arms of my typewriter wouldn’t collide with one another.  The arms are long gone by now.  And, sure, northern plants are free of heat constraints.  But with a good-enough system in place, evolution will not bother breaking out the twine to jury-rig a better one.  Actually, “good enough” isn’t even necessary; “barely adequate” is often sufficient to stifle innovation.  Probably the lowest-hanging analogy for that is the recent attempt to “revamp” the way healthcare is delivered in this country while still preserving the role of all our existing insurance providers.

The other thing Dyson’s passage made me curious about is, why would he expect a better sunlight harvester to arise only at high latitudes?  Yes, temperatures are cooler there, but overheating due to efficient sunlight capture is only a problem when exposed to high loads of radiation.  But I was under the impression that plants in the lower stories of rainforests, for instance, were exposed to very low amounts of sunlight — personally, if I thought a silicon-based sunlight capture system was going to evolve anywhere, I’d put my money on the rainforest.

This graph is from Campbell and Aarup’s study, “Photosynthetically available radiation at high latitudes.”  And, sure, over the course of a year the light levels at high latitude are lower than near the equator.  Not something you needed to scope a fancy graph to realize: it’s sunnier in Texas than in Alaska??  But I thought it might be worth seeing to get a sense of the numbers.  The axis gives values for number of photons per area per day, and if we’re interested in Dyson’s reasoning, it’s worth looking up some numbers for light levels in tropical rainforests.

Rainforest numbers shouldn’t match the equatorial line shown in the graph above.  Or, well, they should, but only for the tallest plants in any forest.  The majority of plants in tropical forests are exposed to very low levels of light in the underbrush.  As opposed to short scruffy plants in the glacially-flattened upper latitudes, which can expect to soak up the full quantity of incipient solar radiation, tropical plants often grow in shade.

Some numbers are tucked away in the methods section of Coste et al.’s “A cost-benefit analysis of acclimation to low irradiance in tropical rainforest tree seedlings: leaf life span and payback time for leaf deployment”  (unfortunately, they’re using the phrase “payback time” in its economics / game theory sense, not the action movie sense.  Their article features zero leaves embarking on a vengeful rampage): seems a plant striving through the undergrowth might receive on the order of 2 moles of photons per meter squared per day — well below the light levels at high latitude through almost the entire year.

All of which is to say that, well, I was a little surprised by Dyson’s evolutionary argument.  Because he’s clearly quite intelligent.  But evolution is a tricky topic, and there are a lot of subtle misconceptions out there.