In the spirit of advancing contrarian ecology, here’s an idea for a provocative review paper which I think a lot of people would find really interesting and useful, but that I don’t have time to write myself. Someone should review the theoretical literature and synthesize all the reasons why competing species might be expected to be similar to one another, rather than different.
Here’s why someone needs to write this paper (the following text, suitably reworded and fleshed out, could even comprise the first few paragraphs of the Introduction. Don’t say I never did anything for you). Ecologists and evolutionary biologists are always out to explain diversity. There are 30 million species in the world*, and in most contexts in everyday life, 30 million is a huge number. And those species often differ from one another in obvious ways. So all that diversity cries out for an explanation, right? Well, yes, it does. But on the other hand, how come there aren’t 60 million species, or 30 billion? Thirty million can’t possibly be a hard upper limit. And while lots of those species do indeed differ from one another in obvious ways, many of them also share obvious similarities. For instance, as Hutchinson famously pointed out, all the algae in a lake are pretty similar in many respects—they’re all only ever limited by the same few resources, they’re all consumed by many of the same zooplankton, etc. And while classic theory from Robert MacArthur emphasizes that coexisting competitors need to differ by some minimum amount, modern theory shows that such “limiting similarity” is actually expected only in very specific circumstances. Similarly (no pun intended), classic evolutionary theory emphasizing that interspecific competition selects for divergence among competitors (character displacement) makes various restrictive assumptions. Character displacement theory exists alongside a substantial but less well-known body of theory identifying conditions under which competitors are not expected to diverge. The possibilities that coexisting competitors will be similar rather than different, and that selection will generate convergence (or at least fail to promote divergence), are often regarded as unusual “exceptions to the rule”. But there’s really no reason to think of divergence and diversity as the rule and convergence and similarity as the exception—they’re simply two ends of a continuum, one end of which has received a great deal of attention from researchers while the other end has not. The review I’m suggesting would help to remedy that imbalance.
Just off the top of my head, here are some of the ideas that should be included in this review (again, don’t say I never did anything for you):
- Habitat filtering. Webb et al. (2002) suggest that coexisting species might be more phenotypically similar than expected by chance if species with a certain phenotype are the only ones that can grow in the local habitat. For instance, if you don’t use water efficiently, you can’t grow in a desert, so all desert plants should have high (and thus similar) water use efficiency. The analogous idea in evolution is that abiotic conditions can be a source of stabilizing selection—the only fit phenotypes are those close to some environmentally-defined optimum. Unfortunately, much work in phylogenetic community ecology and “functional trait” ecology implicitly or explicitly assumes that habitat filtering is the only reason why coexisting species might be phenotypically similar, and so treats a finding of coexisting, similar species as evidence for habitat filtering. One purpose of this review would be to expose this reasoning as fundamentally flawed.
- Fitness equalization. Modern coexistence theory (developed in its most general form by Peter Chesson) points out that coexisting species must always be similar in relevant respects as well as different in relevant respects. For instance, if species A is much better at acquiring the limiting resource than species B, species A is different than species B, and collectively those two species exhibit a diversity of resource acquisition abilities. But that difference inhibits rather than promotes coexistence, and competition will therefore tend to eliminate that diversity in the long run. Interspecific differences promote coexistence only when they strengthen intraspecific density-dependence relative to interspecific density-dependence. Interspecific similarities promote coexistence only when they reduce or eliminate “fitness inequalities”, so that no species is too superior to the others. Analogously, in evolution, phenotypic differences that lead to directional selection eliminate rather than promote diversity. Phenotypic differences that lead to negative frequency-dependent selection are the only differences that generate and maintain diversity. One implication here is that coexisting species should be expected to exhibit divergence in certain traits, and convergence in other traits.
- Obstacles to character displacement. The theoretical literature on character displacement, much of it reviewed in Schluter (2000), identifies lots of obstacles which might prevent divergence of competitors (e.g., range of available resources too narrow to provide an “ecological opportunity” for divergence).
- Character convergence and parallel shifts. In many circumstances, such as when different resources are not nutritionally equivalent, competition actually selects for convergence or parallel shifts in the resource use traits of competing species. This is in contrast to competition selecting for divergence, but some other factor prevents it. Basically, in many quite common situations, avoiding competition doesn’t (or doesn’t necessarily) increase your fitness, which means competition doesn’t (necessarily) select for divergence. For instance, competition for light doesn’t generate selection for plants to diversify and use other resources instead, because there are no substitutes for light. Further, recent theory shows that character convergence and parallel shifts actually promote, or at least don’t inhibit, stable ecological coexistence; they don’t lead to neutral stability, competitive exclusion, or priority effects (Fox and Vasseur 2008, Vasseur and Fox in press).
- Phenotypic clustering. Even the classic MacArthur-type resource overlap model doesn’t predict that, if you have a whole bunch of competitors, they will space themselves as widely and evenly as possible over the gradient of available resources. In fact, what often evolves is “clusters” of very similar species, with the clusters being widely and evenly spaced (Scheffer and van Nes 2006).
- “Apparent” habitat filtering. That is, something that looks like habitat filtering but is not, unless you’re prepared to use the word “habitat” in a pointlessly-broad and empty way. For instance, Leibold (1998) considers a simple model of a bunch of species which interact via resource and apparent competition. Depending on the productivity of the basal resource, you can have highly-competitive but vulnerable species coexist (low productivity), weakly-competitive but predation-resistant species coexist (high productivity), or species with intermediate competitive ability and predation resistance coexist (intermediate productivity). In other words, you always have similar species coexisting, although which species it is depends on productivity. It’s resource and apparent competition that produce this pattern: each species would be able to grow and persist at all productivity levels in the absence of the others, which is why it’s really stretching to call this “habitat filtering”.
The above is by no means a complete list. I’m sure I’m forgetting or not aware of a bunch of stuff Peter Abrams has done, and a bunch of other stuff too.
If you really wanted to make this review thorough, you could also review the empirical evidence for these ideas. For some, there won’t be any evidence, because nobody’s looked for any (e.g., character convergence).
Grad students: if you agree that this would be a really cool review to write, but you find the possibility daunting (if only because you, like me, are short of time), make it a group project. Rope in some of your friends, and start a reading group on this topic. Each week somebody else has to come up with a paper from the literature for the group to read and discuss. Take notes on your discussions. By the end of the semester, you’ll already be well on the way to pulling together the literature and mapping out how it all fits together, and you’ll have pulled in a bunch of co-conspirators who will be keen to start drafting chunks of what will eventually become a really nice, group-authored review. This is how my friend Dan Bolnick and some of his fellow students developed their review of individual-level specialization in “generalist” species (Bolnick et al. 2003). By the way, that paper won the Mercer Award from the Ecological Society of America as the best ecological paper of the year. Your journal club can be much more than a fun way to pass the lunch hour.
And if you do write this review, I hope you’ll submit it to Oikos (seriously). 😉
*footnote: I’m well aware that estimates of the number of species in the world vary widely, even before we start worrying about how to define bacterial and archael species. My point doesn’t depend on the precise number. If you don’t like my number, mentally substitute your own.