I’ll start with the short version of this post, in the form of an old joke. Then I’ll elaborate.
Here’s the short version: Two hikers are walking in the woods. They round a bend and see a bear on the trail ahead of them. The bear charges towards them. They begin to run, but the bear is catching up. Suddenly, the first hiker sits down and starts changing his hiking boots for sneakers. The second hiker says, “What are you doing?! You’ll never outrun that bear, sneakers or no sneakers!” The first hiker replies, “I don’t have to outrun the bear, I just have to outrun you.” 🙂
The joke is about the difference between absolute and relative fitness, and the importance of not mixing them up. In evolution, it’s usually relative fitness that’s of interest, as that’s all that natural selection cares about. For instance, even if two alleles at the same locus are both unfit in some absolute sense, the fitter of the two will still increase in frequency at the expense of the other. The same is true in community ecology. If you’re trying to understand issues like competitive exclusion, coexistence, and the maintenance of diversity, then what matters is the relative abilities of different species to survive and reproduce. For instance, even if a species survives and reproduces well in some absolute sense, it will be excluded if a competing species survives and reproduces even better.
Do ecologists ever mix up absolute and relative fitness? I’m not sure. I’ve never run into a really clear-cut case of mixing the two up, but occasionally I’ve run into cases that seem ambiguous. Hopefully readers can chime in with other examples, or clear up any misunderstandings of mine on the examples I discuss below (the examples all come from areas of the literature with which I’m not super-familiar).
The closest thing to a clear mix-up of which I’m aware comes from the niche construction literature. In a book review, philosopher of biology Samir Okasha caught the leading advocates of niche construction, Laland et al., in a mismatch between their verbal definition of ‘positive’ niche construction, and their mathematical models of this phenomenon. Their verbal definition was based on relative fitness, but their mathematical models were based on absolute fitness. Laland et al. admitted the ambiguity. But not having actually read the book in question, I can’t comment further.
Okasha compares Laland et al.’s slip to failure to distinguish ‘strong’ from ‘weak’ altruism in evolution. Strong altruism refers to actions that reduce an organism’s absolute fitness while raising the absolute fitness of others. Weak altruism refers to actions that raise an organism’s absolute fitness while raising the absolute fitness of other organisms even more–thereby reducing the altruist’s relative fitness compared to non-altruists. Weak altruism can’t evolve in an unstructured panmictic population. But I’ve occasionally seen ecologists appearing to claim that the ecological equivalent can come about. I say “appearing” because the verbal arguments I’ve seen seem ambiguous to me, and so I’m unclear exactly what’s being claimed.
For instance, plants growing in certain ‘harsh’ environments, like alpine meadows, typically facilitate rather than compete with their neighbors, meaning that removing the neighbors of a focal plant reduces rather than increases adult survival, growth, and/or fecundity. That is, neighbors increase (some components of) one another’s absolute fitness.* Which has led some authors to speculate on whether plant-plant facilitation makes the ‘realized niche’ larger than the ‘fundamental niche’ and allows some species to persist in environments where they otherwise couldn’t. As I said, this speculation is ambiguous; not all possible versions of it can work, I don’t think.** In particular, just because your neighbors raise your absolute fitness does not mean they can’t competitively exclude you. Their relative fitness might still be higher than yours–perhaps in part because you also facilitate them!
Economist Nick Rowe has an analogy that, slightly modified, makes this point nicely. Imagine you’re sitting in a crowded theater, and you can’t see the stage. So you stand up, and then you can see better. But that doesn’t mean that everybody will see better if everybody stands up, or that you personally will see better if everybody stands up. We can slightly modify this analogy to explain why interspecific facilitation of the sort seen in alpine plant communities doesn’t necessarily allow a given plant species to grow where it otherwise couldn’t. Imagine people sitting in a crowded theater. Beneath everyone’s seat is a lever that can raise the seat. But nobody can reach the lever underneath their own seat; it can only be reached by their neighbors. If everybody pulls everybody else’s lever (i.e. everybody raises the absolute fitness of their neighbors), it’s not the case that everybody in the theater will be able to see better. Indeed, maybe nobody will see better, or some will see better and some will see worse. So if you’re interested in how facilitation affects plant coexistence or their distributions along environmental gradients, you need to put together a model that specifies how facilitation affects the relative, not (just) absolute, fitnesses of different species (More specifically, your model must explain how facilitation makes relative fitnesses negatively frequency-dependent). Which some people are starting to do, though it’s early days and I don’t think a very wide range of models has been explored yet.
A final context in which I wonder about this is in comparisons of how well introduced species perform in their native vs. introduced ranges. It’s my impression that many papers look at the absolute fitness, or components thereof, of introduced species in their native vs. introduced ranges. For instance, asking if an introduced plant species has higher survival, growth, or fecundity in its introduced range than it does in its native range. I’ve never quite understood why that’s the comparison of interest. For instance, what if the introduced range is just a better place for all plants to grow, so that the introduced species has a higher absolute fitness in its introduced range than in its native range, but a lower relative fitness?
An analogy to evolution helps to clarify my question here. Evolutionary biologists are careful to distinguish two distinct senses of the term ‘local adaptation’ (see Kawecki and Ebert 2004 for a good review). One sense, sometimes called ‘home vs. away’, compares the absolute fitness of a given genotype at its native (‘home’) site vs. at some other (‘away’) site where it isn’t ordinarily found. But that sense of local adaptation is consistent with the same genotype being favored by selection everywhere. If you want to understand spatial variation in selection, you need to study local adaptation in the sense of ‘home vs. foreign’ comparisons. That is, you need to measure the fitness of genotypes growing at their home site relative to the fitness of foreign genotypes that have been transplanted into that site. If you repeat this for a bunch of sites and find that, at each site, the home genotypes always have higher relative fitness than the foreign ones, then you’ve shown ‘local adaptation’ in an evolutionarily-relevant sense. That is, you’ve shown spatial variation in selection, with different genotypes being favored by selection at different sites. Analogously, if you’re interested in why an introduced species seems to be increasing rapidly in abundance in its introduced range and replacing native species there, shouldn’t you asking if its relative fitness is higher when it’s in a ‘foreign’ (i.e. introduced) site than in its ‘home’ (native) site? Not asking about its absolute fitness at ‘home’ vs. ‘away’? But I don’t know the introduced species literature well at all, so maybe I’m just totally missing the point of the studies to which I’m referring?
*This caveat about fitness components is actually quite important. Just because alpine plants generally increase the adult survival, growth, and/or fecundity of their neighbors does not mean that they actually raise rather than lower the absolute fitness of their neighbors, since seeds have to germinate and seedlings have to survive to adulthood for a seed plant to complete its life cycle. Indeed, it’s hard to imagine that you could have a system in which every individual actually did increase the absolute fitness of its neighbors. In such a system, absolute fitnesses would all be positively density-dependent, and total density would explode to infinity. But it’s perfectly possible to imagine a system in which, say, neighbors increase one another’s survival and fecundity, but actually decrease one another’s fitnesses by occupying sites that could otherwise be occupied by new offspring of other individuals. My point here isn’t to criticize empirical studies of facilitation in alpine plants and other systems for only measuring certain fitness components. I’m just making a conceptual point here, one which I suspect is familiar to many readers (including those working on facilitation), but perhaps not to all.
**One way (not necessarily the only way) this speculation could work is if you have one species (call it species X) that facilitates its neighbors but isn’t affected by them. For instance, maybe around each individual of species X is a “habitable zone” that would otherwise be lethal to other species if that individual of species X weren’t there. Individuals of species X are to neighboring individuals of other species what a beaver is to the aquatic species that live in a beaver pond, but couldn’t live in the stream or dry land that would be there if the beaver bond wasn’t there. Species X facilitates other species and allows them to persist where they otherwise couldn’t by creating patches of habitat that wouldn’t otherwise exist. I take it this is more or less what is thought to happen with some species of “nurse plants”, though I don’t really know the nurse plant literature at all.