Why expect trade-offs in ecology and evolution? (UPDATED)

Evolutionary biologists and ecologists believe that evolving organisms are subject to trade-offs. You can’t have a ‘supergenotype’ or ‘superspecies’ that’s optimized to do everything, whose fitness (both in absolute terms, and relative to competing genotypes or species) is high in all possible circumstances. A corollary is that we believe ‘a jack of all trades is a master of none’–organisms that are specialized to be good at one thing (e.g., to survive in a particular environment) should be better at it than generalist organisms. The assumption of trade-offs underpins standard thinking on a whole range of issues, from speciation and adaptive radiation to competitive coexistence.

But why do we expect trade-offs? I mean why do we expect them theoretically; empirically we expect them because they’ve often been observed. If you’re reading this blog, I’m sure a number of standard answers immediately occur to you. Allocation is a big one: a unit of energy, materials, or time that an organism uses to do one thing (say, build a defensive structure) cannot be used to do something else (say, reproduce). Antagonistic pleiotropy is another big one: a mutation that improves one biological function often degrades another (although this is presumably not an ultimate explanation for trade offs, since one can ask why we expect antagonistic pleiotropy). Mutational decay is a third: a mutation that degrades an unused function (think cave fish eyes) will be selectively neutral. Over evolutionary time, we only expect organisms to remain good at what they’re selected to be good at. And the laws of physics are a fourth. For instance, for fundamental physical reasons there’s a trade-off between the rate and efficiency with which a physical system can perform work. There’s a trade-off between ‘fast-but-inefficient’ and ‘slow-but-efficient’  life history strategies for the same ultimate reason there’s a trade-off between power and fuel efficiency of car engines.

The common theme of all of the above is constraints on phenotypes. Allocation, antagonistic pleiotropy, mutational decay, and the laws of physics are all reasons why organisms can’t build (or can’t maintain, over evolutionary time) certain phenotypes. Trade-offs reflect constraints on the range of phenotypic options that are open to organisms.

But, as this article by Nick Rowe points out, there’s a different, and perhaps more fundamental reason, to expect trade-offs in ecology and evolution, one that has nothing to do with constraints on organisms’ phenotypic options. The mere fact that organisms are subject to natural selection, so that we only expect to observe organisms with maximal (or nearly maximal) relative fitness, implies that we’ll see trade-offs. The rest of what follows is a shameless steal from Nick’s article, justified only by the fact that Nick’s article is about trade-offs in economics. I’m merely translating his argument into an eco-evolutionary context (and very little translation is required). Do read the original version; Nick’s a terrific explainer.

Imagine an organism with two traits, X and Y, that determine fitness (the same argument holds if more than two traits matter). Assume that fitness increases with increasing values of both X and Y (the same argument holds if fitness decreases with increasing values of X and/or Y). And there are no constraints on the possible combinations of X and Y. To use Nick’s analogy, imagine shooting a shotgun at a target: every hole in the target represents a possible combination of X and Y. There is no trade-off between X and Y, no reason why having a high value of trait X constrains your ability to build or maintain a high value of trait Y. (I ignore  the hugely unlikely possibility that the pellets would just happen to hit the target in, say, a straight line with negative slope (\), thereby creating the appearance of a constraint on possible values of X and Y)

But if we go out and sample real organisms, we still expect to observe a trade-off between X and Y, because X and Y have fitness consequences. Here, we assumed that high values of X and Y promote high (relative) fitness. So in a plot of Y vs. X, only the points in the upper-right of the plot denote high-fitness individuals (or highly competitive species, if you want to think in ecological rather than evolutionary terms). The other points, those below and to the left of the upper-right boundary of our ‘shotgun blast’, denote individuals which are, by assumption, less fit than individuals on the boundary. And we don’t generally expect to observe less-fit individuals.

And that upper-right boundary is almost guaranteed to have a negative slope. The only plausible way it wouldn’t (under my assumptions) would be if there happened to be a single fittest type, a single point that was above and to the right of all the others. Remember, we’re assuming that X-Y phenotypes are determined by a shotgun blast–there are no constraints on where the individual pellets can hit the target. If you think the upper-right boundary is likely to have something other than a negative slope, you’re implicitly assuming some sort of constraint on possible X-Y phenotypes.

A corollary of this argument is that, when some other factor (such as migration) interferes with the ability of selection to purge less-fit types, we shouldn’t expect to see trade-offs, except due to constraints.

Shorter version: we expect to see trade-offs because they’re a ‘ghost of competition past‘ (Connell 1980 Oikos 35:131-8). Competition among genotypes (or species) purges the less-fit, so that only the most-fit remain. And the phenotypes of the most-fit are likely to exhibit a trade-off (or cluster near a single optimum), even in the absence of constraints.

UPDATE: Rees Kassen emailed me to point out that this argument is even more general. The same argument applies if you think of X and Y as two different environments in which an individual organism might live. If fitness in environments X and Y is determined by a shotgun blast, then selection will weed out all individuals which don’t have high fitness in at least one of the two environments. The only remaining individuals will be those along the upper-right boundary of the shotgun blast, which will be a negatively-sloped curve defining a trade-off in fitness across different environments.

Rees also notes that, as an empirical matter, such trade-offs in fitness across environments aren’t nearly as common and strong as we tend to think they ought to be. Perhaps because, as I suggested in the comments on this post, there are various factors, like migration, that can prevent the fittest type (or in ecology, the competitively-dominant type) from going to fixation…