Data from Himalayan songbirds suggest that the rate-limiting step in biodiversity production may not be the speed of speciation, but rather the speed at which new niches are created. See Letter p.222
There are currently about 10,000 named bird species on the planet, but only half as many mammals. What dictates the number of species in any given group? In this issue, Price et al.1 (page 222) marry the ecologies of bird species living in close proximity in northern India with their ages to argue that it is not the rate at which new species form but the rate at which new niches are created that limits how many species actually persist. By building a new evolutionary tree linking all 358 songbird species inhabiting a Himalayan slope, the authors find that it has taken an average of 7 million years for related species to end up together on the landscape.
This seems a long time: bird speciation generally involves geographic separation and, from beginning to end, it usually takes less than 3 million years for species-level mutual disdain to evolve; new species can populate adjacent landscapes and join their close relatives in the blink of a geological era. That said, species that occupy the same niche rarely occur together in nature. So Price and colleagues argue that the roughly 4 million extra years it seems to take for species to establish themselves alongside their relatives means they have had to wait that long for available niches to arise. This suggests that finding a new role in nature may be the limiting step in diversification.
Given the ambiguous signals that evolutionary trees have offered us so far, using a partial evolutionary tree linking coexisting species to make an argument about large-scale processes is audacious. Previous studies that have compared the lengths of recent and older branches on evolutionary trees to measure diversification2 generally have found that more-recent branches are too long relative to expectations from simple evolutionary models. Such data often suggest both that extinction is non-existent (which, of course, is not true) and that diversification has slowed towards the present3.
There are at least three reasons for this odd pattern of slowing diversification, and Price et al. consider them all. Mundanely, inferring trees is hard, and recent branches on evolutionary trees may be biased to be too long relative to older branches4. Luckily, the authors could compare their tree to a recent and independently derived one that includes nearly all the same species5. Crucially, although the two trees differ greatly in total age and shape, they agree almost exactly on the 7-million-year time lag for the coexistence of relatives.
More interestingly, an apparent slowdown in diversification is expected because speciation takes time: we miss more-recent diversification events because we do not recognize separated populations that will eventually become new species as new species6. This means that recent speciation events have not yet been recorded, so recent branches will be longer than they should be. Many evolutionary trees are well described by such a process7. However, Price and colleagues sidestep this issue with a particularly nimble piece of logic. Because most speciation takes place when populations are separated geographically, and given that lineages that occupy the same niche rarely coexist, there can be few or no unrecognized species in a sample of species living in proximity. Nascent species may exist elsewhere, but the recent branches of these authors' tree are certainly not too long because of a species-recognition problem.
Most intriguing is to consider the possibility that slowdown is a common fate in diversification. Following the invasion of relatively empty niche space — by dispersal to a new region or by the evolution of new ways to make a living — an initial flurry of speciation fills up an initial set of niches, but then new species become established only when others go extinct or new niches are created8.
Such an ecological-demand model (in contrast to evolutionary supply) also fits many evolutionary trees well3, and Price and colleagues' data on songbirds tally with this story (Fig. 1). Other data from the region suggest that recent ecological opportunity may be linked to changes in elevation, and Price et al. indeed find that close relatives differ most often in the elevation at which they live. The well-known pattern of highest species richness at mid-elevations9 is, in their data, associated with more food (and so more niches) rather than with more speciation. And songbirds span a greater range of sizes and shapes at species-poor high elevations, perhaps owing to release from competition with other groups of birds that could not establish themselves.
There is little question that an ecological-demand model of biodiversity offers up a grand narrative: nascent species arise in ample supply as isolated populations, but these expand their ranges and fill landscapes only if, and only when, those landscapes have ecological room for them. Many new species might therefore wither on the vine and never reach this expansion phase. This model implies a decoupling of the rate of speciation from the rate of subsequent establishment, a novel pattern that has recently been documented for flies and birds10.
An ecological-demand model also raises several questions. Evolutionary biologists have long appreciated that close relatives compete strongly11. However, we have only recently re-appreciated that speciation is often intimately linked to diverging ecologies12. Therefore, to evaluate the ecological-demand model, we need to know the relative contributions of ecological13 and non-ecological speciation14 to biodiversity production. And, given that the model suggests that the number of species in a group is related to niche availability10, we need to understand both why there are more niches for some groups (say, birds) than for others (say, mammals), and whether it is really true that, at any one time, many landscapes are full to the brim with biodiversity.
Price, T. et al. Nature 509, 222–225 (2014).
Nee, S., Mooers, A. O. & Harvey, P. H. Proc. Natl Acad. Sci. USA 89, 8322–8326 (1992).
Morlon, H., Potts, M. D. & Plotkin, J. B. PLoS Biol. 8, e1000493 (2010).
Revell, L. J., Harmon, L. J. & Glor, R. E. Syst. Biol. 54, 973–983 (2005).
Jetz, W., Thomas, G. H., Joy, J. B., Hartmann, K. & Mooers, A. O. Nature 491, 444–448 (2012).
Purvis, A., Orme, C. D. L., Toomey, M. H. & Pearson, P. N. in Speciation and Patterns of Diversity (eds Butlin, R., Schluter, D. & Bridle, J.) 278–300 (Cambridge Univ. Press, 2009).
Etienne, R. S. & Rosindell, J. Syst. Biol. 61, 204–213 (2012).
Rabosky, D. L. & Lovette, I. J. Proc. R. Soc. B 275, 2363–2371 (2008).
Rahbek, C. Ecography 18, 200–205 (1995).
Rabosky, D. L. & Matute, D. R. Proc. Natl Acad. Sci. USA 110, 15354–15359 (2013).
Darwin, C. On the Origin of Species (Murray, 1859).
Schluter, D. The Ecology of Adaptive Radiation (Oxford Univ. Press, 2000).
Nosil, P. Ecological Speciation (Oxford Univ. Press, 2012).
Svenson, E. I. Organisms Diversity Evol. 12, 229–240 (2012).
About this article
Cell Host & Microbe (2020)