In our comparison of mainland and island species turnover, we sampled mainland communities by randomly superimposing each of the Caribbean islands on the mainland five times. In each of these mainland island-shaped regions, we then tallied how many species were present and then measured turnover between regions.
Oceanic islands are famous for their high endemism. We all love Darwin’s finches, Hawaiian honeycreepers, Komodo Dragons, and all those weird things on Australia. These endemic wonders suggest that islands may be home to unique processes of diversification. However, there are many mainland regions that rival islands in their level of endemicity, especially in mountainous regions that have strong dispersal barriers and environmental gradients. Thus, you could ask, are islands truly special?
Anoles are perfect for asking this question because they have many species on both islands and the mainland. At the beginning of this year, I wrote a series of Anole Annals posts covering a few studies that have investigated island and mainland species diversity patterns in anoles. These papers, focusing on species richness, found that small isolated islands often had a species-poor communities relative to the mainland, concordant with traditional island biogeography theory. However, those islands that were big enough to support speciation broke the island biogeography predictions and in some cases resembled or surpassed the mainland in the degree of diversification and endemism.
Inspired by these studies, Adam Algar, Jonathan Losos, and I investigated a related but different measure of species diversity: species turnover. Species turnover is a measure of how many species two communities share (i.e., two communities that share most of their species have low turnover while two communities that share very few species have high turnover). Generally, it’s been shown that environmental dissimilarity and geographic distance are correlated with the degree of turnover between communities. We wondered how islands compared to the mainland in their species turnover patterns and hypothesized that islands would show higher turnover on average.
To ask our question, we first developed a model for species turnover among communities on the mainland that described how physical distance and environmental dissimilarity between mainland communities explain species turnover. In other words, given a distance and environmental dissimilarity between two mainland communities, we could estimate how much species turnover to expect on average. We did this for anoles and Eleutherodactylid frogs (Terrarana sensu Hedges et al. 2008) and for both, turnover increased with geographic distance and environmental dissimilarity between mainland communities. Then, we used this mainland model to predict turnover on islands, given their distance and environmental similarity to one another.
It turns out that the mainland model doesn’t predict island species turnover very well. Turnover, on average, is significantly higher between islands than between two communities on the mainland, all else equal. And turnover between islands and mainland communities is complete in nearly every case – much higher than would be predicted from geographic and environmental distance alone.
Most likely, this increased species turnover is driven by the fact that islands are surrounded by a truly inhospitable barrier – saltwater! – that effectively increases the distance between islands. It’s just much harder for two communities to exchange individuals or have gene flow over water. Intuitive, perhaps, but to our knowledge this paper is the first quantitative test of this island effect on species turnover. We find that islands do indeed add something unique to their assembly.
Yoel E. Stuart, Jonathan B. Losos, & Adam C. Algar (2012). The island-mainland species turnover relationship Proceedings of the Royal Society B DOI: 10.1098/rspb.2012.0816
- DNA Sequencing from Formalin-Preserved Specimens - December 16, 2014
- Sans the Claws, Anoles Present Under The Trees? - December 5, 2014
- Out Damned Spot! Or, How They’d Be Loving If They Weren’t Fighting - November 18, 2014
Martha Munoz
This is really cool, Yoel!
I was wondering – is there any way to control for time in your model? If island communities are younger, then presumably they have had fewer opportunities to exchange individuals than do older lineages on the mainland. Your point about saltwater dispersal, however, is certainly sound. I’m simply wondering if it the pattern is additionally confounded by time.
Yoel Stuart
Thanks Martha,
It’s a good point. We didn’t explicitly model time so it may be that all those islands will eventually homogenize more and show turnover patterns similar to the mainland. However, a lot of the islands and island blocks are 10s of millions of years old – as old as the Anolis radiation itself. So it seems that time shouldn’t play a big role here.
And, time cuts both ways – with less time, you have less opportunity for anagenetic speciation so you’d get less turnover.
Adam Algar
To, add because of the back colonisation of the mainland the island faunas are not necessarily younger. Actually the mix of older and younger clades on the mainland could provide an interesting framework to start trying to look at this . . . if we could get solid age estimates.
Skip Lazell
Great stuff! Hope to read more…. Thanks!
Yoel Stuart
Thanks Skip!!
Julián Velasco
Nice paper Yoel!
I´m curious about if the patterns that you reported still are maintained using “real” endemic areas for mainland species (this question would also apply for the paper of Algar & Losos 2011). I see that many mainland areas selected randomly (see fig 1) fallen in regions of low diversity for anole species (e.g., Amazonia, North Mexico and border of Colombia and Venezuela). It is probable that this random selection would underestimate the real richness of mainland areas? For example, Choco region maintains a high diversity of anole species (approx. 30 spp.) in contrast with the entire Amazonia (where probably occur less of 15 species). The same question would be valid for frogs which exhibit a high diversity in the Andean and Amazonian region. I think that an identification of areas of endemism for anoles species is necessary and opportune.
Adam Algar
Julián,
Great Question/comment. I think that would be a very interesting analysis. As you note in this (and related work), we’ve aimed to compare island patterns to randomly chosen mainland ones. Whether it underestimates the real richness of mainland areas depends on what we mean by real. If by real, we mean representative estimate, then I think our approach captures that. But if by real we mean maximal richness then I think you’re right that our approach underestimates. I agree that a mainland ‘hotspot’ to island comparison would be very timely – We should do it!
Ambika Kamath
Just for fun, is it possible to say what the saltwater barrier makes the effective distance between islands or between islands and the mainland? As in, “islands that are X kilometres apart are as different in species composition as two equally sized bits of mainland that are Y kilometres apart,” to get an intuitive sense of just how special islands are…
Yoel Stuart
That’s a great suggestion. We didn’t do that but it would be cool to, for both geographic and environmental distance. Let me look into it. Once we figure that out, I’ll write a little post about it. Thanks!
Skip Lazell
Great question, Ambika: Thanks! Skip
Ambika Kamath
Thank you!