A mean field model for competition: from neutral ecology to the Red Queen

O’Dwyer & Chisholm 2014 A mean field model for competition: from neutral ecology to the Red Queen. Ecology Letters 17: 961-969

I'm reliably informed that this is actually quite simple to understand! Equation 2 from O'Dwyer & Chisholm.

I’m reliably informed that this is actually quite simple to understand! Equation 2 from O’Dwyer & Chisholm.

Will Pearse

To spoil the punch-line, I’m not sure I completely agree that the model in this paper is biological defensible, but I’m quite certain this is a very important contribution. The authors have found a way to incorporate species differences into neutral theory: the most recently speciated species out-competes all others, and as a consequence phylogenetic branching times become more reasonable. Much ink has been spent suggesting Neutral Theory will form the building blocks of models that incorporate species differences, and this (finally!) is an extremely important piece of such work. My main concern is that I think, to be biologically defensible, there has to be some kind of inheritance of fitness from the new species’ ancestor. The only way I can see the youngest species being the best is if the driving force of biology is pathogens – the authors point towards this, and somewhere Ricklefs is jumping for joy – but I just can’t see it. To me, this would require that we change (again) the scope of Neutral models from covering species within the same guild to covering the same ‘pathological guild’. Moreover, I find it hard to believe that each speciation event is coupled with a magic trait that pathogens must evolve, from scratch. Surely a species in a large clade, presumably with an equally large body of pathogens, is even less likely to have such a trait evolve, and we have yet another way for diversity-dependence to rear its head. That said, all is certainly not lost, and this is an (extremely impressive) start. If a similar model could have multiple guilds nested within itself, and allow some degree of exchange between the guilds, I would have little trouble getting behind it. Using diffuse competition to approximate the competitive hierarchy was a wonderful moment in the paper, and it’s fantastic to see an argument used to defend Neutral Theory extending, not defending, it. If we can use these kind of approximations to bring even more niche-based concepts into Neutral Theory, things are looking up!

Lynsey McInnes

Lynsey Bunnefeld

I am still on the fence about this paper. On the one hand, I admire how they have taken neutral theory and changed it a bit in order to produce predictions that more closely match what we see in reality (specifically, they assume new species are fitter than all older species and that leads to more realistic distributions of species ages than the hardcore neutral theory where all individuals are equivalent). This is impressive. On the other hand, this is a bit of a weird tweak to make and doesn’t really seem to be biologically defensible, at least not generally.

I like the idea of neutral theory. I like its simplicity and the fact that it is remarkably good at predicting a lot of recurrent patterns we see in nature. I even like the way it sometimes fails and I really like that it can really irritate people. It is fun to watch people get stressed out about it. I agree that if it is to continue to have relevance, it needs to be continually scrutinised and tweaks applied and tested. This paper provides a remarkably simple and tractable tweak to deal with one of the outstanding issues with neutral theory – that is tends to predict unrealistically long species ages. By making new species fitter than older ones, the authors are able to purge communities of older species more quickly and so reproduce patterns that more closely match those observed in nature. Neat.

But wait? Are new species typically fitter than old ones? The authors’ argument for yes appears to hinge on the new species being free, or at least relatively more free, of nasties that could hold them back. I’m no expert, but my intuition is that not all, or indeed many, new species are ‘free’ in this way. Indeed, don’t most species come about through divergence in geographically or ecologically distinct arenas and might be really quite similar to their close relatives apart from in a key few traits (and not even that sometimes). Indeed, there seems to be mixed evidence at best that you shed the majority of your parasites upon speciation.

OK, but if the assumptions of this model sit uneasily, what other tweaks might be made to neutral theory to reign in unrealistically old species ages? At this is when the authors’ ideas become harder to put down. They have recognised that you need to find something that ‘gets rid’ of older species and their idea seems at least a bit better than species just having an ‘intrinsic’ life span (cycle of life style). An idea that has been bandied about, but with lots of quite robust refutation too. What else might do it? Some kind of slowing down of adaptation to changing environment? Some kind of competitive density effect? Some kind of lag in competitive interactions (your enemies catch up with you and get rid of you?). None of these sound particularly promising.

And so, while I might not agree with the authors’ model I’m pretty happy that people are producing such models and refining and refuting things further. One day we might be able to figure out where these recurrent patterns in biodiversity are coming from and the relative importance of niche and neutral processes. We won’t get there without trying.


About will.pearse
Ecology / evolutionary biologist

5 Responses to A mean field model for competition: from neutral ecology to the Red Queen

  1. I had a very superficial look at the paper and was quite surprised that they didn’t cite the idea of protracted speciation http://onlinelibrary.wiley.com/doi/10.1111/j.1461-0248.2010.01463.x/full , which (to me) seemed to solve the same problem in the same journal.

    The core of the problem (and the obvious route to potential solutions) seems the unrealistic assumption of point speciation in the UNTB, which requires very hight speciation rates in order to give “newcomers” a sufficient chance to establish.

    I would argue that there are a number of options to change this, the one proposed here (which is not neutral, however), but also protracted speciation or allopatric speciation. The reason why specially allopatric processes are not used more is that the models are more difficult to compute I guess.

    • will.pearse says:

      That’s a really good point. I always think of protracted speciation as something that makes branching times more reasonable (…and relaxes a somewhat unreasonable assumption), but I don’t think I ever would have made the link between these two approaches as you have there. Thanks!

  2. JP O'Dwyer says:

    Thanks for the interesting blog post! I agree with you both that our assumption for the precise form of the `landscape’ is less general than we would like. I also agree with your take that it serves as a proof of concept for adding fitness differences to neutral models and still obtaining approximate analytical results.

    In terms of protracted speciation—my understanding is that it changes expected (future) species lifetimes, primarily by significantly reducing the probability of immediate or early extinction. On the other hand, for the kinds of abundances folks have considered reasonable for the incipient species to `real’ species transition, I don’t think protracted speciation reduces (enough) the expected species age for a very abundant extant species.

    What is definitely in common is the idea of using dynamics rather than just snapshots of patterns to probe the differences between model predictions. But also what both approaches flag up is that using species ages/lifetimes alone may not be sufficiently well-defined or precisely known to distinguish between these various models. So solving the `species age’ or `species lifetime’ problem in neutral theory can definitely be a good motivation, but we need better ways to test these hypotheses.

    • will.pearse says:

      No worries; I loved the paper 😀 Are you working on some more things along these lines?

      I think you’re right that we need more ways of testing models. I think it’s something of a positive thing when you think about it: we’re saying we can explain the things we previously cared about, and now we need even more power to differentiate.

      • JP O'Dwyer says:

        Great! And yes, more to come on both model development and parametrization/testing with empirical data.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: