Why should species diversity be highest where organismal performance is highest? (UPDATEDx2)

The question in the post title has bugged me for years. I’ve read and thought about it enough that I think it’s a good question, meaning a question without an obvious answer. But it’s on a topic on which I’m far from an expert, so there might well be some non-obvious answer that I’m unaware of. So I’m going to pose the question and look forward to learning from the comments.

I was reminded of this question most recently by Grady et al. (2019 Science). I emphasize that this paper is merely one among many that could be used to illustrate my question, and that I’m not criticizing this paper or its authors. This isn’t a post-publication review of Grady et al. (2019). This is me thinking out loud about a broad issue, and picking a recently-published paper as an illustration purely because it would be work to look up a bunch of additional illustrations and I am lazy this is only a blog post.

Anyway, Grady et al. report a bunch of data establishing two main claims. First, species diversity of large-bodied ectothermic marine predators is highest in the tropics, but species diversity of large-bodied endothermic marine predators is highest in the subtropics and temperate zones. Second, endotherms outperform ectotherms in colder water–they can swim faster, their neurons fire faster, they can consume more food relative to their metabolic demands, etc.

Grady et al. then argue that the performance pattern explains the diversity pattern–that species diversity of ectotherms is maximized where it is because those are the places where ectotherm individual performance is maximized. And similarly for endotherms.

But why should that be the case? Why should correlation indicate causation here?

Grady et al. suggest that high individual performance causes high species diversity because “[H]igher abundances and foraging success reduce extinction rates and permit specialization, which promotes speciation” and “species with high abundances tend to have large ranges and subsequent fragmentation may promote allopatric speciation, particularly across ocean basins or hemispheres”. To which, maybe! But that strikes me more as plausible hand-waving than anything. Which is fine! It’s a starting point, and you have to start somewhere. But I dunno, I also feel like I could dream up other plausible hand-wavy explanations for the correlation between performance patterns and diversity patterns. And I bet I could dream up plausible scenarios in which performance and diversity are uncorrelated, or even negatively correlated, because high performance does not cause high species diversity, or even causes reduced species diversity.* The point is that I don’t trust my own intuitions. Especially not when I’ve already seen the data, because whatever data I’ve seen are going to make whatever post-hoc causal explanation I dream up seem totally plausible to me. Had the data come out differently, I’d have dreamed up a different plausible-seeming post-hoc explanation. Like Duncan Watts said, everything is “obvious” once you know the answer. Post-hoc intuition should either be the starting point for further research, or the summary of an already-established conclusion. Not a key plank in an argument aiming to establish a conclusion.

Returning to the question in the post title: I have the vague, admittedly-anecdotal sense that many biologists, especially those working in macroecology and macroevolution, are quick to make the leap from “organisms with trait(s) X perform well in environment(s) Y” to “there should be many species with trait(s) X in environment(s) Y”. But my brain stumbles at that leap. What apparently seems like an intuitive leap to many people strikes me as a gap in the argument, one that could use a lot more theoretical and empirical research to fill it in.**

And maybe that gap has been filled in in many cases, and I just don’t know about it! It certainly seems like a really interesting and important line of research to pursue, which probably means it has been pursued and I’m just ignorant of it. But if I’m ignorant of it, well, maybe others are too. So you tell me: what are the best worked-out case studies of high organismal performance causing high diversity? Looking forward to learning from your comments.

UPDATE: See the comments, where Mark Vellend reminds me that he made the same point in his very fine book. Apologies for having forgotten that, Mark! More broadly, I’m sure this point must’ve been made by others too.

UPDATE #2: John Grady was kind enough to take the time to comment. Thank you to John for contributing to the very thoughtful and interesting conversation we have going in the comments.

*And it’s sounds like Grady et al. could too. They also write “In addition, high metabolism may increase range size and reduce allopatric speciation, and respiratory constraints may limit utilization of benthic resources near coastlines.” That is, they suggest reasons why the high performance of endotherms might actually reduce rather than increase their propensity to speciate in subtropical and temperate waters.

**Part of why I stumble over this gap is that the theories of speciation I’m most familiar with don’t invoke high absolute individual performance. They invoke frequency-dependent individual performance (e.g., character displacement). They also involve mechanisms having nothing to do with individual performance, like different selection pressures in different locations plus barriers to gene flow between locations. Conversely, I tend to think of high performance as the culmination of adaptation, the achievement of a (frequency-independent) “peak” on the “fitness landscape”. Which leaves me wondering why achieving a frequency-independent fitness peak would promote diversification. But maybe all this just shows how ignorant I am of the enormous theoretical and empirical literature on speciation.

20 thoughts on “Why should species diversity be highest where organismal performance is highest? (UPDATEDx2)

  1. Excellent. This is precisely my criticism of my own field in fish morphology > performance > niche distribution. Paul Webb generated a plausible functional model to explain the patterns that were already seen. This was really nice work. Then others used simple, qualitative biomechanical models to give a bit more biomechanical support for Webb’s models but 1) the particular biomechanical models were searched in biomechanical model space with a particular goal and 2) the models are interpreted in hindsight (and at least some could be interpreted in opposite ways if that was the pattern to explain). The entire field has used Webb’s models as the causal explanation when these should be treated as the hypothesis that still need to be tested. It’s now been 30+ years. The fluid dynamic modeling tools are here now (and have been for maybe 15 years) but its still not been addressed this way. Maybe others have this criticism but apart from a few small comments in my own papers this criticism doesn’t seem to be well known.

  2. I am equally intrigued by this question. I recall some long-ago MTE paper(s) in which somehow high species diversity became a consequence of high temperature, but I also recall not being convinced (i.e., the model didn’t “naturally” make this prediction, you had to insert an assumption that more diversity would be generated at high temperatures, and so of course it did). I think there are some macroecological studies that make a compelling argument that productivity/energy lies somewhere on a causal pathway to diversity, but the underlying reasons are not clear. And it’s not a universal: for plants, some important global diversity hotspots are in Mediterranean climates, where conditions for plant growth in a general sense are not especially favorable (compared to warm, wet places).

    To quote from my book (p. 179):

    “Most post hoc explanations for ecological patterns have been expressed verbally rather than mathematically, often with rather vague logical links. Using the latitudinal gradient as an example, commonly invoked causal factors include historical disturbance (e.g., glaciation), productivity, environmental stress, climatic stability, environmental heterogeneity, and species interactions (Lomolino et al. 2010). On their own, only one of these is even close to being self-explanatory: higher environmental heterogeneity ought to support greater species diversity via spatially variable selection. Otherwise, it is not clear how these factors might relate to species diversity per se. To quote from a widely used ecology textbook (Krebs 2009), “Climate determines energy availability, and the key variables for terrestrial plants and animals are solar radiation, temperature, and water. Climates that are more stable are more favorable, cause higher productivity, and all these factors work together to support more species.” This “explanation” is difficult to make sense of. It says only that many individual plants or animals might thrive best when conditions are consistently warm and moist, but it says nothing about why more _species_ should be found under such conditions. I think that articulating how such low-level factors might be expected to influence the high-level processes that generate diversity (speciation, dispersal) or those that help maintain or reduce diversity (selection and drift) would be of great benefit to the conceptual coherence of community ecology.”

    • Good point re: Mediterranean climates being a hotspot for plant diversity. And yes to the MTE model of the latitudinal diversity gradient being another example here.

      Similar arguments get made often in the paleo literature too, I think (at least, that’s my casual impression).

      Sincere apologies for having forgotten that passage from your book! I recall it now and thought it was spot on when I read it (as I’m sure you can tell from the post). I’ll update the post to point to it.

      • No doubt I failed to cite others who have said the same thing previously. When it comes to expression of thoughts and ideas, it’s impossible to recall all of who said what (especially when a “new” idea is new in pretty subtle ways)…

  3. What kinds of models are typically used for speciation? I’m familiar with some partial differential equation drift-diffusion models that model density peaks over time and space, where space represents some fitness or niche landscape, but my impression was that these were more interesting to mathematicians and were generally too cumbersome and technical to convince many ecologists.

    I’m quite interested in questions like this one, where causal and mechanistic insights could play a role, but unfortunately I don’t know what sorts of canonical models are used, or how seriously different communities take different kinds of theoretical approaches.

      • Oh I understand. There seems to be a substantial literature in several important facets of research here. I was just asking about the theoretical frameworks because I don’t know what is typically used, whereas for dispersal and other phenomena I have some experience or at least a sense of the tools used.

        The connections between individuals and their behaviour, and population dynamics on larger scales, seems to be reasonably clear. But I don’t know a good way to understand speciation. Perhaps I should also read this book. 😉

  4. On a different note, this post is a good illustration of how my blogging style has changed over the years. 2011-2012 me would’ve probably called this post “The Underpants Gnome Theory of Speciation” and snarkily ripped all of macroecology for purported ignorance of speciation research in evolutionary biology.

    That version of this post probably would’ve gotten more eyeballs, and some readers would’ve found it more fun to read and preferred its bluntness. But it also would’ve put off other readers, and risked coming off as a personal attack on the authors of Grady et al. On balance, I think this version of the post is a better stylistic choice than version 2011-12 me would’ve written (especially given the size of the audience this blog has now). But it’s not a totally obvious decision (e.g., there might’ve been a way for 2011-12 me to write a post on “the Underpants Gnome Theory of Speciation” so that it came across as funny rather than as snarky and aggressive.)

  5. You’re not alone. I encountered the same apparent logical gap when writing the latitudinal gradients chapter in Natural Systems, and filled it in with some woolly phrasing (an undergraduate textbook isn’t really the place to start an argument). I’ve never liked the assumption that larger populations or ranges lead to speciation, because doesn’t that automatically imply that the resultant species have smaller populations and ranges? Isn’t the latter more typical of tropical species anyway? I haven’t followed this through to any conclusion yet but I share your sense of being unsatisfied.

    • Yeah, I had the same thought about the argument that larger geographic ranges = speciation. I’m reassured to discover I’m not alone in that!

      There certainly is a sensible argument that larger geographic ranges buffer against (some) causes of extinction. But the causal relationship between geographic range size and speciation rate seems trickier to work out. But again, I haven’t read Coyne & Orr so what do I know?

      Re: your textbook, no, an undergrad textbook isn’t the place to start an argument. But can’t it be a place to highlight a gap in knowledge? Or a place to “teach the controversy”? I’m thinking for instance of this: https://dynamicecology.wordpress.com/2016/08/17/zombie-ideas-in-ecology-textbooks-now-teach-the-controversy/

  6. I can go one further. Why do we always expect that species will be must abundant (or more or less equivalently have highest occupancy) in places where they perform best? They don’t (https://academic.oup.com/jpe/article/5/1/46/1293933). Yet this assumption is central to much applied ecology including species distribution modelling and habitat assessment.

    This is all implicit in the idea of shared preference niches in papers by Colwell & Fuentes, Irene Wisheu & Mike Rosenzweig but seems to have been completely forgotten by ecologists still obsessed with niche differences.

    • “Why do we always expect that species will be must abundant (or more or less equivalently have highest occupancy) in places where they perform best? ”

      Yup! I was thinking of bringing this up in the post, but couldn’t work it in without either interrupting the logical flow or resorting to an epic footnote. So I decided to just wait for a commenter to bring it up, and I wasn’t disappointed. 🙂

      I’ll take the excuse to shamelessly plug my own little theoretical paper, showing that there are only limited circumstances in which we should expect the species that perform best in environment X to be the most abundant species in environment X: https://www.frontiersin.org/articles/10.3389/fmicb.2012.00268/full.

      Credit where it’s due: I used to naively think that *of course* the most abundant species in environment X would be the ones best adapted to environment X. But then I tried to explain my intuition on this to Graham Bell, and he brought me up short with a (very fair and polite!) one-sentence reply that totally exposed the gaping hole in my intuition. The paper of mine I shamelessly linked to was basically me writing a whole paper to try to convey to others what Graham Bell conveyed to me in one sentence.

  7. Two people have steered me to your site within the past hour and I’ve enjoyed discovering the blog. Its a good general question you pose, and I agree that many papers are a bit cavalier about connecting performance to diversity. However, I feel compelled to point out that in the Grady el al paper you mention, we do go beyond higher performance = high diversity. Broadly, our logic + data was that higher relative performance = higher collective consumption (resource capture) & abundance = higher diversity. That second piece helps fill the gap you mention. Collective consumption was something we spelled out mathematically and provide global data to support predictions (eg, Fig 5). Marine mammal abundance increases nearly two orders of magnitude as their performance increases, so there is an important dot connected in the movement from performance to diversity. Our framework is undoubtedly simplistic but it does try to make some of the connections you want to see.

    Although we synthesize and analyze lots of diversity data, our model validation mostly focus on the performance -> division of resources side. I think more work remains connecting the division of resources to diversity. The idea that more resource use and abundance promotes diversity is not new (e.g. the ‘more individuals hypothesis’), seems intuitively reasonable, and does have some supporting data. But I think its also the most difficult to formally link, and there are many steps between collectively consuming more of the available resources to having more species, relating to things like niche partitioning and gene flow. In the paper we write, ‘Higher abundances and foraging success reduce extinction rates and permit specialization, which promotes speciation (34). With higher relative performance in cold waters, endotherms can consume a higher fraction of their preferred prey, expand their dietary breadth, or specialize on a subset of their potential prey base…In addition, species with high abundances tend to have large ranges (36) and subsequent fragmentation may promote allopatric speciation, particularly across ocean basins or hemispheres (37).’ And then we add some marine examples, like incipient speciation with diet specialization in orcas, and the number of sister species living on different sides of the globe. Of course, a lot of work remains in this area and other factors matter. Things like habitat complexity and environmental stability are likely to be just as important.

    Extrapolating from this work, a general approach for answering your question may be to show how performance affects competitive success and resource division (1), how 1 drives abundance and niche diversity (2), and how 2 shapes diversity. Still a lot of ground to cover, but less mysterious than going straight from performance to diversity.

    • Thanks for sharing your thoughts here John. I think the gap you highlight–between “high total abundance” and “high species diversity” is the one that needs filling in.

      Re: whether high resource availability or high resource consumption promotes diet specialization, with diet specialization then leading to speciation, I’m not so sure that’s generally true. I’m thinking for instance of Darwin’s finches. My recollection of Peter and Rosemary Grant’s work is that it’s during years of *low* seed availability–drought years, basically–that selection most strongly favors those individuals with the most extreme beak sizes. The rare individuals with the extremely large beaks can eat (and specialize on) the extremely large seed species for which there’s little competition, and the rare individuals with the extremely small beaks can eat the extremely small seed species for which there’s little competition. And so those are the years in which there’s strong selection favoring assortative mating. You really do not want to mate with a bird with a very different beak size than you because your offspring would have intermediate beak sizes and would likely starve. But in wet years, there are plenty of seeds in all size ranges and so birds with all beak sizes have no trouble finding food. And so it’s in wet years birds of all beak sizes adopt very generalized diets and just eat whatever seeds they happen to encounter in the size ranges they’re capable of consuming. And in wet years there’s rampant hybridization even between different species with different mean beak sizes. If wet years ever become the rule it seems like the Galapagos would become much more productive and rampant hybridization would collapse many currently-fairly-distinct species of Darwin’s finches into an indistinct hybrid swarm or single generalized species. Going by memory of the Grants’ work here, might be getting some details wrong, but pretty sure that’s the gist.

      • James Stegen and I tried to articulate when you would expect links from energy -> community abundance -> species richness in our paper “When should species richness be energy limited and how would we know?” here:


        The required assumptions are:
        1) Per species extinction probability is a function of population size,
        2) The focal assemblage operates under a zero-sum resource constraint (and therefore extinction rate is richness-dependent), and
        3) There is sufficient time for an equilibrium to emerge.

        Discussion about speciation or speciation rate can be a red herring since it’s entirely possible (although many disagree!) that speciation events are not what is limiting the total number of species in an area.

      • Thanks for this Allen. I agree that it seems easier to justify an extinction rate-based argument to close this gap than a speciation rate-based argument. Will follow up your link with interest.

        EDIT: Allen, please also accept my apologies for referring you to as Stuart when I first commented. That was very rude and sloppy of me. I’ve edited my original comment to correct my mistake.

  8. You raise an interesting question Jeremy which I think illustrates a more general point: that many of the broad patterns we see in biogeography, and which are widely believed to be well understood, in fact have hardly been rigorously tested at all. I note that the Grady et al. paper quote invokes greater specialization as a mechanism for speciation: “[H]igher abundances and foraging success reduce extinction rates and permit specialization, which promotes speciation”. But as Angela Moles and I pointed out in our guest DE post some years ago, this is hugely problematic, and by no means are tropical species, and their interactions interactions, always more specialized:


    Specialization per se is not a necessary prerequisite for speciation: if it was then we would not see radiations of clades full of generalist species, as we do in some plant families that have highly generalised interactions with pollinators such as Asteraceae, Brassicaceae or North American _Asclepias_. All that speciation requires is for a population to be _different_ not necessarily more specialized.

    Slightly tangentially, but as an illustration of the power of blogging: as you know, that guest post with Angela led to the piece being published in Biotropica in 2016, and it’s gained some traction (43 citations to date). However it also encouraged Angela and I to work on a grant proposal to the Australian Research Council for some joint research, and we discovered just before Christmas that this had been successful. So thanks for encouraging us to work together in the first place 🙂

Leave a Comment

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.