bush dog

We often talk about the South American wild dogs. The South American wild dogs are a sister group to Canis and its allies, and in South America, evolution allowed dogs to go many unusual directions.

When the dog family phylogenetic tree was drawn from sequencing the genome of a boxer, I was amazed that it put the bush dog, which is sort of a wild version of the dachshund, as being a sister species to the maned wolf, which has very long legs.

There was actually a big debate as to where the bush dog actually fit. When I was learning about dog evolution as a child, I had books that told me that the dhole, African wild dog, and the bush dog were all closely related because of their trenchant heel dentition. One of their carnassials has a single, blade-like cusp that increases their ability to bolt shear meat.

The phylogenetic tree that was created from the dog genome sequence pretty much ended this discussion. The dhole and African wild dog were both found to be closely related to Canis, more so than the side-striped and black-backed jackals. The bush dog was with the maned wolf and the other South American canids, and the trenchant heel dentition was the result of convergent evolution.

End of story.

Or so I thought.

In 2012, a study was released that that was meant to update the divergence times with all extant carnivora. The researchers used large samples of DNA and other characters to determine when these animals diverged from each other. Some of these “other characters were things like vocalization and scent gland similarities.

Its phylogenetic tree for Canidae is similar to that in the aforementioned paper on the domestic dog genome, except that it greatly increases the divergence time between species. For example, it has the Urocyon foxes diverging from the rest of Canidae 15-16 million years ago, instead of the 9-10 million years that the dog genome paper found.  It also has the golden jackal and the coyote being sister species, and the wolf is not the closest relative of the coyote. We now know this is very much in error, and it probably comes from the non-genetic “source trees” that were used in the analysis. It has has the Tibetan fox as being related to the extinct Falkland Islands wolf, which happened because there are almost no genetic studies on the Tibetan fox. Both the Tibetan fox and the Falkland Islands wolf had kind of weird squared off bodies, though, and this type of analysis does use morphology.

It has the dhole as being closely related to the wolf, golden jackal, and coyote.

But it has the African wild dog splitting off much sooner from this clade, and what’s more, it has the bush dog as its sister species!

One should be skeptical of this finding, because of its use of so many non-genetic “source trees,” it is going to miss the problem that occurs so much with dog species. Convergent evolution and phenotypic plasticity run riot in the family, and it is really hard to figure out relationships between species using just morphology and behavior alone.

This would make a lot of sense if it were confirmed with better genetic studies. Bush dogs are very weird animals. They are the only South American canids that hunt in packs. They really don’t have a rich fossil record, and it is pretty hard to connect them to other South American wild dogs.

It is tempting that they might be something that weird, but we need more evidence.

If they really did turn out not to be part of the South American clade of wild dogs and to be closer to the African wild dog, it would be a real shocker.

But not entirely.

The questions that would arise from it would be how it evolved.  We have evidence of Xenocyon coming into North America. Xenocyon is traditionally thought of as the ancestor of the African wild dog and the dhole, but it may not be. But there is also evidence of dholes or dhole-like dogs that are actually closer to the AWD coming into North America and making it as far south as Mexico.

So maybe there is something to it.

But this sort of study does have its limits. It’s trying to morph both classical and molecular techniques for taxonomy, and those tend not to hold up very well.

But I still think it’s worth examining.









tibetan fox corsac

L: Tibetan fox (Vulpes ferrilata) R: Corsac fox (Vulpes corsac)

I’m sure that just about everyone has seen the footage of the Tibetan fox. Its blocky head is really unique among canids, and one theory about why it has such an unusual head is that the head actually makes the fox more camouflaged around rocks as it stalks its prey.

If you’ve not seen the footage, here it is:

The Tibetan fox is found in the Tibetan and Ladakh Plateaus, and it pretty much found only in high grasslands. It is a specialist of this environment.

The mitochondrial genome of this species was sequenced recently. Preliminary studies had indicated that its closest relative is the corsac fox, which ranges in the steppe country that runs from the eastern edge of the Caucasus to northeastern China.  It lives in arid, semi-arid, and grassland habitats to the north of the core range of the Tibetan fox.

The corsac is the basic fox. It is smaller than most red foxes, and it comes in gray or reddish brown.

And to be completely honest with you, I have a very hard time telling corsac foxes from North American swift foxes (Vulpes velox). I should note that even though the two look alike, the corsac is more closely related to the red fox, while the swift fox* is very close to the arctic fox (Vulpes lagogpus). Both swift and corsac foxes are adapted to grassland ecosystems, so their similarities can be chalked up to simple convergent evolution.

The researchers who sequenced the Tibetan fox mitochondrial genome also looked at other dog species and found that modern dog species radiated very rapidly. The authors estimate less than 5 million years separates Canis species from the maned wolf, which is actually quite about half the time that has been estimated for the divergence between Canis and its allies and the South American wild dogs through the sequencing of the dog genome.

Mitochondrial DNA studies can lead us astray, but this is still an interesting find.

The researchers found that the corsac and Tibetan foxes split from each other about 1 million years ago.  Although they don’t say so in the paper, it’s pretty likely that the corsac is the ancestor of the Tibetan fox. Maybe it happened like this:

A population of corsac foxes, whose range had been pushed south during the Pleistocene, were able to roam into the Tibetan Plateau during a warming period. They were then cut off by glaciers and began their journey toward speciation, adapting the high grassland habitat in ways that made them quite different from their ancestors. The squared off heads gave the foxes marginal advantages in hunting in the high, rocky  grasslands.

The authors didn’t find any genetic evidence of any sort of adaptations in the mitochondrial genome for higher altitude living among Tibetan foxes. Those genes probably do exist, but they don’t exist in the mitochondrial genome.

So what the researchers found is that Tibetan fox is very closely related to the corsac fox. This is not a surprising find, but I have wondered where this fox really does fit in the dog family. It’s from such an isolated area that virtually no studies have been done on it.

The squareness of its head makes the Tibetan fox a bit of a celebrity in the digital age. If BBC cameras hadn’t filmed it (twice!), it wouldn’t really be known. I’d only ever heard of it from a children’s book, which included a bad painting of one. It was just a nondescript fox as far as I was concerned.

So the Tibetan fox is a modification on the basic fox.

It evolved from the banal to the bizarre.

Bizarre enough to be on the BBC.


*The study looked at kit foxes. Kit foxes are closely related to swift foxes. Both are very closely related to the arctic fox.




brush wolf

It cannot be overstated how much the discovery that coyotes are not as distantly related to wolves as we believed ultimately questions our entire understanding of the evolution of the Canis species.

The traditional understanding Canis species evolved from some form of Eucyon dog some six million years ago. Wang and Tedford, who wrote the most important book on the paleontology of the dog family, believe this was Eucyon davisi, which was the first of its genus to enter Eurasia. The genus Eucyon is where the common ancestor of the Canis dogs (including Lycaon and Cuon) and the South American wild dogs would be located. Eucyon dogs were small. Imagine them as being something like a black-backed jackal or a Hoary fox rather than a coyote.

Then, 5 million years later in the Southwestern US and northern Mexico, a coyote-like Canis evolved, which was called Canis lepophagus. This animal is sometimes considered the common ancestor of wolves and coyotes. It may be, but considering how close we now know wolves and coyotes are now, it’s not the most recent common ancestor. Canis lepophagus did migrate into Eurasia, where it either founded or is identical to Canis arnensis.

In Eurasia, several smaller jackal-to-coyote forms evolved. One of these was Canis estruscus,  which then evolved into Canis mosbachensis (which is called Canis variabilis in China).

Ron Nowak believed the red wolf was an offshoot of this wolf that wound up colonizing North America and then becoming isolated from the rest of Canis mobachensis when the ice sheets expanded. There was also a competing view that the red wolf was actually a remnant version of Canis edwardii or Canis priscolatrans (which were probably the same animal). This animal was roughly the size of a red wolf, but Nowak rejected it as a red wolf ancestor because it lived too early for what he thought were red wolf fossils.

The Eurasian wolf species evolved mosbachensis-variabilis, but the two forms of wolf shared habitat and likely exchanged genes, making it very difficult

The coyote’s evolution was never clear. It was thought to have evolved out of Canis lepophagus. It was thought that lepophagus evolved into edwardii, and then it began to become more gracile and smaller, eventually becoming the now coyote.  It’s now pretty clear that it evolved out of the Eurasian Canis lupus and not these endemic North American “wolves.”

It either evolved from the modern wolf, which evolved into roughly its current form 800,000 years ago, or it came from a late surviving mosbachensis-type wolves that were regularly crossing with modern wolves before they came into this continent. Maybe the remains that Nowak had been considering “red wolves,” were actually these ancestral wolves that were evolving into the modern coyote.

Maybe when this wave of wolves came back across from Eurasia, perhaps 50,000-100,000 years ago, it came into a world already dominated by a dire wolves, which already occupied the niche for large, pack hunting canids and this wave of Canis lupus evolved as the American jackal.  After all, the bobcat is just a diminutive Eurasian lynx that found itself in a very similar position when it came into this continent, and it evolved to be a smaller animal that generally hunts smaller quarry than its larger ancestor. Of course, the modern bobcat didn’t reach its current form until about 20,000 years ago, but it still was forced to adapt to a slightly different niche than its Eurasian ancestor.

In literature on the paleontology of Canis, there is a heated debate as to how these animals all fit. The conventional view is that the wolf evolved from Canis mosbachensis/variabilis through Canis etruscus, which may be the same thing as Canis edwardii/ Canis priscolatrans. Wang and Tedford contend that the coyote and wolf split from Eucyon.  The modern wolf evolved from Canis chihliensis, which was a large wolf-like canid. It spread into North America to found Canis armbrusteri, which then evolved into the dire wolf (Canis dirus) in North America and Canis gezi and Canis nehringi in South America.  In the Old World, another offshoot of chihliensis gave rise to Canis falconeri, which the supposedly gave rise to the Xencyon, which is supposed ancestor of the dhole and African wild dog. Another view holds that the Armbruster’s wolf (C armbrusteri) is descended from edwardii/priscolatrans (which may be the same as etruscus). This lineage then gave rise to the dire wolf and the two sister species in South America, thus descending solely from North America wolves.

All of these ideas come from paleontology, and they pretty much are done without looking very deeply into the studies that are examining the DNA of these species. It is pretty obvious from that literature that the notion that coyotes and wolves split at the time of the Eucyon ancestor is quite wrong. For that hypothesis to work, African of  wild dogs and dholes would have to be genetically closer to wolves than coyotes and golden jackals are. They aren’t.

But if the genome-wide analysis shows that coyotes are so much more closely related to wolves is true, then all these fossil and subfossil canids that are said to be the most recent common ancestor of wolves and coyotes simply aren’t.  Instead, all of these species that are classified in Canis are likely a mix of evolutionary dead ends, like the dire and Armbruster’s wolf, or could be hidden ancestors of extant canids that aren’t wolves or coyotes.

For example, black-backed and side-striped jackals diverged from the rest of Canis and its allies at about the same time that Eucyon was diverging from Canis. It is possible that there are many relatives of these particular dogs that are hidden in this vast sea of Canis fossils.

The new discovery about the coyote’s split from the wolf also means that any remains of North American canid that are listed as coyote that date to 1 million years before present are not coyotes. What they actually were is a very good question.

We’ve spent a lot of time assuming that coyotes and wolves were quite divergent. We know now that they really aren’t, but when we look into the past at all the “wolves” and “coyotes” that came before, we see how this genus became so successful. It can easily evolve into big game-hunting forms, but the real success is in its ability to assume the size and shape of the generalist predator. Phenotypic plasticity is a wonderful thing for a lineage to possess.

But the real message of the new discovery about wolves and coyotes should be is a cautionary tale about paleontology. Paleontology is a wonderful science, and it makes amazing discoveries every day, but when its faced with a lineage of animals where phenotypic plasticity and tendencies toward parallel and convergent evolution are commonplace, it is bound to make errors. Paleontologists aren’t examining flesh and blood that can have its molecules tested for relationships. They are measuring anatomical characters and determining phylogenetic relationships based upon the similarities of these characters.

Which works well.

Until you get something like wolves and coyotes, where there are many ancient fossil and subfossil remains that look like they could be ancestors of either.

But the DNA says they aren’t.

And paleontology would have problem catching the inverse. There are many species that we’ve discovered only through DNA testing. African butterfly fish in the Congo and Niger basins look identical to each other, but they have been isolated from each other for 57 million years. I have yet to see this species split into two, but if they were mammals, you could bet they would be placed in distinct species in heartbeat.

Paleontology is missing some really important things we’ve since found out through molecular analyses.

And paleontologists know this.

They are working with the data they have, and by definition, it’s going to be more incomplete than genetic studies.

Science is provisional. Different disciplines and methodologies are going to come up with different answers. It’s pretty amazing that one genome-wide assay study can wipe out so much literature in paleontology.

These debates have been raging for years.

And it turns out that everyone was actually wrong.

Update 21 August 2016:  It turns out that I missed a paper that actually did some limited DNA analysis and found that Canis nehringi was pretty much a South American dire wolf, as in it was likely the same species as the North American dire wolf. Canis gezi, however, was  more closely related to the modern maned wolf and had been incorrectly identified as a wolf. So let this stand as a correction to the error above.


gray fox west

A few weeks ago, we lost at least two canid species. Analysis of whole genome sequences indicated that the red and Eastern wolves are recent hybrids between wolves and coyotes. Indeed, this study also showed that the genetic variance between coyotes and wolves is equivalent to the variance between wolf populations, which actually calls into question whether coyotes are a valid species as well.

But this finding does not mean that there aren’t new cryptic species to be found in North America’s endemic canids.

I was just perusing some of the literature on gray foxes, when I came across this study in PLOS ONE. The authors sequenced mitochondrial DNA from 169 gray foxes from California and Georgia, as well as 11 “island foxes”  and added in a sample from an aberrant gray fox that wound up in Washington State.

The authors were trying to figure out if California and the American Southeast represented a kind of “glacial refugia” for the species during the Last Glacial Maximum.

What they found was a deep divide between Eastern and Western populations. The California and Washington samples and those from the “island foxes”  were estimated to have separated from the Georgia samples some 500,000 years ago. That’s actually greater than current genetic distance between Old World and North American red foxes, which separated 400,000 years ago, and are currently being proposed as distinct species.

We know from previous studies on Eastern gray fox mitochondrial DNA studies that the gray foxes of the Northeast are relatively recent colonizers from the Southeast.  So my guess is that we’d find a similar divergence between gray foxes from New York and Ohio and those from California as was seen in this study with Georgia and Western gray fox samples.

Now, this study looked at only mitochondrial DNA, and this is only a tiny part of the genome. More detailed genetic studies are needed to determine the exact time of divergence between the two gray fox populations. Further, because this study included foxes from only California, Washington, and Georgia, it doesn’t really show us where the divide between these two lineages exists on the North American continent.  More samples from across the range of gray foxes could give us that answer. My guess is there is a hybrid zone between the two lineages either in the Southwest or in the South-Central US.

But this assumes that there really this genetic divergence is confirmed with nuclear DNA sampling. It could be that the Western population just has an old mitochondrial DNA sequence that wound up surviving, even though the majority of the gray fox genome comes from same source as the Eastern gray fox.

It could be, but there is still a very strong possibility that Western gray foxes do represent a distinct species from the Eastern gray fox, and this question can be answered. We just need analysis from a bigger part of the genome from a broader cross section of gray foxes.

If there actually is a distinct species of Western gray fox, then it would be obvious that the island foxes, which have only been on the Channel Islands for 7,100-9,200 years, should be classified as part  of that species. The authors found that no extant population of gray fox in California actually gave rise to the island fox, but there are similarities between island foxes and those in Northern California. But they were still part of this Western gray fox division.

I’ve thought it very odd that gray foxes live in Minnesota quite well, but in the West, they don’t come as far north as western Oregon. The Washington sample in this study was the first gray fox found north of the Columbia River, and western Washington has a much, much milder climate than Minnesota.

Maybe the differences in range reflect a difference between species. Maybe the gray fox of Minnesota is the same as the gray fox of Georgia, and this species has evolved more cold tolerance than the Western species.

There are just so many questions that arise from one study that has largely been overlooked.

And if there are two species of gray fox on the North American mainland, there could be several cryptic species of gray fox in Mexico and Central America. Maybe the isolated populations of gray fox in Colombia and Venezuela are also different species.

The Urocyon foxes are really interesting animals. They are the most basal of all canids, and among North American canids, they are the only one without any connection to Eurasia.

Most taxonomists divide the genus into two species: the gray fox and the island fox. The island fox was recently removed from the Endangered Species List, but I’ve always been very doubtful that it actually is a species. Most of the evidence now shows that it was actually introduced by people. Something very similar could happen with red foxes in Australia, which are now reproductively isolated from the rest of the Old World red foxes. Maybe in 9,000, they will be morphologically distinct enough for someone to declare them the “Australian fox” and work to preserve them as a distinct species.

But in focusing so much on this odd insular population, could we have missed the really big story about the urocyon?  Maybe there were two species after all, but we never bothered to look into it.

Maybe one day, we’ll have Urocyon cinereoargenteus and Urocyon occidentalis as the two species of gray fox native to the United States.

So there is only one wolf species in North America.

But there could be two gray foxes on the mainland.

And that is pretty cool.






From the BBC’s The Hunt:


These two are aggressive nursers.

Coyotes are supposedly killing all the fawns, but they didn’t get these two.



Damaged antler

A buck with a damaged antler is the first to visit this year’s feeding station:

The antler could very well continue to grow out at this weird angle. The antler is still in velvet and will be for the next month or so.


Get every new post delivered to your Inbox.

Join 1,293 other followers

%d bloggers like this: