Monday, November 28, 2011

Supertramps play checkers on remote tropical islands


Biogeographers have a not-so-secret love affair with archipelagos.  Inarguably, islands are sexy.  But, when a bunch of islands get together they become the pieces in an irresistible puzzle of “who lives where and why?”

Consider the facts: Islands are hard to get to; for land-dwelling plants and animals they are stepping stones of potential habitat amidst an endless sea of death.  Therefore, islands close together (such as in an archipelago) should have the same plants and animals living on them, because the hardest part of setting up shop in a group of islands is traversing the distance from the mainland.  However, contrary to expectation, islands in archipelagos often have different biotas and the two-hundred-year-old question is “why?”

You may have heard of Charles Darwin- his answer to the “why” for finches in the Galapagos was evolution by natural selection after a long trip from mainland South America.  Someone you may not have heard of is Jared Diamond, but he also sparked a ferocious debate when he claimed that different bird species in the Bismarck Archipelago live on different islands because they compete with each other.

In 1975, Diamond noticed that certain pairs of bird species make a checkerboard pattern across the archipelago- the two species never live on the same island but do live on neighboring islands.  He suggested that one explanation is that the species are so similar in the types of food or habitat that they use that they compete and one kicks the other off the island.  Of course, there are lots of other reasons why two species don’t occur together on the same island: maybe the two are different and need different types of habitat, or maybe they live in different parts of the archipelago that are two difficult to travel between.  A new study this month in the Journal of Biogeography revisits Diamond’s original bird data from the Bismarck Archipelago and looks more closely at whether his checkerboard patterns really do result from competition.

It’s time for some math.  There were 154 bird species found across all the islands, which means that there are (154 x 153)/2 = 11, 781 possible pairs of bird species.  The researchers found that 1,516 pairs of species followed a checkerboard pattern, but that only 27 of these were between closely related species that are likely to compete with each other.  This number seems small relative to the total number of possible pairs; how do we know that this isn’t the number of checkerboard pairs that would happen just by chance? By using a null model, of course (see previous post), where a computer puts bird species on islands at random and then calculates the number of pairs of species that don’t occur together.  This showed that for several types of birds, there were actually more checkerboard pairs than expected by chance- which means that competition could be the explanation.  Or, it might not be.

When glaciers expanded and the sea-level dropped back in the Pleistocene, many of the Bismarck islands became connected by land, but there were still groups of islands that remained separate.  In many of the checkerboard pairs of species, the species live in different parts of the archipelago that were separated by water during the Pleistocene- this shows that species may not live together because they have not been able to get to the same islands.

Another peculiar commonality of the 27 checkerboard pairs of closely related species is that in almost all cases one of the species lives only on small islands where there aren’t many other bird species.  These ‘supertramps’, as Diamond called them, may be weak competitors that can only live in places not already usurped by other species.  The prevalence of supertramps in pairs of species that form checkerboards means that, despite 35 year of controversy, competition may actually be a good explanation, at least in part, for why islands have different bird species.  As for what’s responsible for the other (much larger) part, geography, evolution, and chance appear to be the major culprits.


You can find this article at:

ResearchBlogging.orgCollins, M., Simberloff, D., & Connor, E. (2011). Binary matrices and checkerboard distributions of birds in the Bismarck Archipelago Journal of Biogeography, 38 (12), 2373-2383 DOI: 10.1111/j.1365-2699.2011.02506.x

Sunday, November 20, 2011

Salmon like it cold, catfish like it hot


How many fish can you name?  Five? Fifty? How about all 829 species native to the rivers and lakes of the continental U.S. and Canada?  Naming all those species is impressive, but the fifteen-or-so minutes that it takes to do so would be a bit like watching water boil… only longer.  Names only become interesting if we know something interesting about the species they pertain to.

In the case of North American freshwater fish, one thing we do know is where each of them live (or lived, for 19 species that recently kicked the bucket); Lawrence Page and Brooks Burr recently published a brand-new field guide that has maps of the ranges of all of these fish species.  But still, this isn’t all that earth-shattering to anyone but an ichthyologist or sport fisher with a checklist.  What’s actually interesting is what Page and another biologist (Jason Knouft) did with these range maps.

By laying all of the maps on top of each other (using a computer, of course) they were able to show that native fish follow both a latitudinal and longitudinal diversity gradient; there tend to be more species of freshwater fish in the lakes and rivers in the southern and eastern parts of the continent than in the northern and western parts.  People have been studying latitudinal diversity gradients for a long time and have come to realize that, even though it’s a cool pattern, the fact that there are more species closer to the equator really doesn’t tell us much about why there are more species closer to the equator.

If the fish are split up into the different families to which they belong (e.g. basses, sculpins, catfishes, perch, trout, etc…) we can learn a lot more about why fish diversity varies so greatly between different places.  Different fish prefer different types of environments; salmon like it cold, but catfish like it hot.  When the researchers correlated temperature and other environmental variables with the diversity of each of these groups, it came as no surprise that trout and salmon (Salmonidae family) diversity is higher in the frigid north and that catfish diversity is higher in the balmy south.  In fact, each fish family had different aspects of the environment that they keyed in to.  Not many suckers live in the mountains, but minnows can be found high or low.

In the case of North American fish vs. Latitude, the latitudinal diversity gradient appears to be losing- the pattern appears to be more of an accident of which families of fish are most prevalent and less of a general pattern found across all types of fish.  Most of the differences in diversity between the north and south stem from differences in the environment and the ability of different groups of fish to survive and diversify in these environments.  In the end, latitude itself doesn’t actually have much to do with it.


You can find this article at:

ResearchBlogging.orgKnouft, J., & Page, L. (2011). Assessment of the relationships of geographic variation in species richness to climate and landscape variables within and among lineages of North American freshwater fishes Journal of Biogeography, 38 (12), 2259-2269 DOI: 10.1111/j.1365-2699.2011.02567.x

Sunday, November 13, 2011

Climate, not space, gives trees more room to range


Everything does not live everywhere; there are no baobab trees in Canada, nor caribou in Florida.  This is not a particularly profound statement.  But, for some of my former Malawian high school students, who had never traveled farther than 20km from their home or watched a nature documentary, it came as a revelation.  “But why, madam?” they would ask me.

Simplistically, the answer has two parts.  One, places are far away from each other, and two, places have different climates; therefore, living things have limited areas where they live because they either can’t get to new places or because when they get there they can’t survive.  Taking this one step further, it follows that species that can survive in a wider range of environments should also have larger ranges.  But, how can we actually determine whether this is true?

One pattern that seems consistent with this idea, is the observation that species that live farther north tend to have larger ranges.  This pattern has been called ‘Rapoport’s Rule’, though it is neither a general rule nor does is belong to Rapoport.  One explanation for why it (sometimes) occurs is that species living farther north have to be able to deal with a wider range of temperatures throughout the year.  This broader tolerance would allow them to live across a wider geographic area.  But again, how would we actually test this?

In the most recent issue of Ecography, Xavier Morin and Martin Lechowicz describe their clever way of figuring out whether the Rapoport effect they found across all 598 of North America’s native tree species actually resulted from tolerance of annual temperature variability, as opposed to just chance.  One alternative hypothesis they had to rule out was the possibility that the trees’ ranges were larger farther north simply because there is more land area in the northern parts of North America than there are in the southern parts.

On a computer, the scientists made an outline of North America and then randomly dropped the tree species onto the continent and let them spread.  When the trees were allowed to spread anywhere, the range sizes that came out of the simulation were much different than the actual ranges sizes for real North American trees and there was no relationship between the size of the range and how far north it was.  However, when the simulation was constrained so that each fake species was given a temperature tolerance and only allowed to spread to places within that temperature range, then the simulation produced range sizes that were larger further north.

This type of null modeling, where a computer makes a simulation that is used to test whether patterns observed in nature could come about just because of chance, is becoming much more popular with all kinds of scientists now that fast computers are cheaper and easier to use. 

You can find this article at:

ResearchBlogging.orgMorin, X., & Lechowicz, M. (2011). Geographical and ecological patterns of range size in North American trees Ecography, 34 (5), 738-750 DOI: 10.1111/j.1600-0587.2010.06854.x