The cost of leafing

This week's Nature has a news feature by me on leaves (why have they made 'plantecology' one word?). Specifically, it's about the patterns in leaf construction seen across all land plants, what causes them, and the consequences that they have for our understanding of the living world. (Lovers of waffle will not want to miss hearing me talk about this on the Nature podcast.)

[M]ost of the variation in the physical and biochemical properties of leaves can be represented on a single axis running, to put it crudely, from quick and juicy to slow and tough. [This is] the 'worldwide leaf economics spectrum', and it embodies many of the trade-offs that govern how plants deploy their resources within the limits that physics places on biological possibility.

The work … has attracted the attention of everyone, from plant physiologists studying how leaves work to biogeochemists looking at the cycling of nutrients on a global scale. In part, the paper is so popular because of the size and scope of the database that underlies the work; but the popularity also reflects the intellectual excitement that surrounds the discovery that so much can be explained by so little. This has given some ecologists hope that by looking at the large-scale patterns in how organisms work, they can gain a general understanding of why species live where they do, and why some are common and others are rare. Such findings are not of purely academic interest: climate researchers are using them to improve their models of the consequences of global warming.

For me, one of the most interesting aspects of writing this feature was investigating an idea among some ecologists that the best way to understand why species live where they do, and why some are common and others rare is to think not about species, but traits — such as leaf biology, seed size and number, and so on.

In some ways, this is counterintuitive. Most, perhaps all, cultures name the plants and animals around them, and recognize that they split into groups of similar kinds, i.e. species. Many ecologists come to the science through a love of natural history, and identifying and naming stuff. And consequently, many theories of biodiversity are rooted in what's been called 'nomenclatural ecology':

To try to understand things such as what determines the number of species that can coexist in a place, how numerous each species is, and how productive the system as a whole is, ecologists have traditionally looked at what species are present, how they interact, and how their abundance affects that of the others. Such an approach is an extension of ecology's roots in natural history, says Brian McGill of McGill University in Montreal, Canada. "People become ecologists because they love to go outdoors and look at the woods. They get attached to putting names on things, and get focused on knowing lots about particular organisms."

But this approach soon becomes intractably knotty, as the number of possible interactions between species rises geometrically with the number of species. "We don't have the capacity to learn as much as we need to know by studying one species at a time. Studying interactions between species, and then trying to build that up, hasn't panned out. It's too complicated," explains McGill.

Traits — such as, for leaves, mass-per-area, or photosynthetic rate — are measurable, and comparable in a way that species names aren't, and also allow one to quantify natural variation. I think of this as dropping below the species level, to look at the components of biology — how organisms work, and how they differ — and I've a hunch it might offer a way out of the current morass of different theories to explain the origin and maintenance of biodiversity. Of which there are tons — it did my head in trying to get to grips with this area.

Of course, to make satisfying science, one wants to be able to turn trait studies back into predictions about species, because as human beings that's how we perceive the world.

But this looks like it might be possible — Science recently published an extremely cool paper by Bill Shipley and colleagues using trait measurements and, to my satisfaction, maximum entropy theory to predict the abundance and distribution of plants in abandoned French vineyards with 94% accuracy. Which is pretty damn good.

I confess, I really should have covered this in ITBOAH, but it didn't cross my path while I was writing this. Very sorry, ITBOAH-readers.