A new modern mammal for Madagascar

ResearchBlogging.org

The fact that new, modern-day mammal species are discovered on a fairly regular basis should most definitely not be a surprise to the average Tet Zoo reader. These are not all 'cryptic species' distinguishable only on the basis of DNA: many are morphologically distinctive, honest-to-goodness new animals discovered either in the field or in museum collections. And they're not all bats and rodents: new monkeys, lemurs, sloths and hoofed mammals (peccaries, deer and bovids) have been named in recent years.

i-68eac88075968767d6d502d2231cba08-Gerald-Durrell-portrait-edit-Sept-2010.jpg

Mention 'new mammals' and 'Madagascar', and most people will assume that you're talking about lemurs (loads of new lemurs have been named over the past decade). However, all Madagascan mammals aren't lemurs, and the island continent has a fascinating assemblage of tenrecs, rodents, bats and carnivorans too. Madagascar's endemic carnivorans include a motley assortment of weird, unique species (like the Fossa Cryptoprocta ferox and Falanouc Eupleres goudotii) as well as a number of bushy-tailed, often striped 'mongooses'.

Traditionally, the Fossa, Falanouc and so on were regarded as peculiar civets (that is, they were included in Viverridae) while the Madagascan 'mongooses' were classified as, well, mongooses (as part of Herpestidae). Rather than representing the product of two, three, four or more separate invasion events from Africa, it now seems that all of these animals are part of a single, endemic clade (termed Eupleridae) that originated in the Oligocene and is close to mongooses proper (Yoder et al. 2003, Yoder & Goodman 2003) [diagram from Yoder et al. (2003) shown below].

i-4ef5656c82bb557d6ef17b03996a86e6-Yoder-et-al-euplerid-phylogeny-nature_Aug-2010.jpg

Anyway, I digress somewhat. The point of this article is to announce the publication of a new, living euplerid species: Salanoia durrelli Durbin et al., 2010, discovered in the marshes around Lake Alaotra in central-eastern Madagascar. Its announcement means that two Salanoia species are now known - the other one is the poorly known Brown-tailed mongoose or Salano S. concolor, named in 1839 [a stuffed specimen is shown below: I borrowed this image from here. It has previously appeared in Nowak (1999) and Garbutt (1999)].

i-97bb9760177e83c815704dc6dcde428d-Salanoia-concolor-Garbutt-1999-Aug-2010.gif

As is so often the case, the discovery and recognition of S. durrelli is pretty interesting. Hints of its existence have actually been in the literature for a while: in Mammals of Madagascar [shown below: an excellent book], Nick Garbutt (1999) referred to the rumoured existence of "a small carnivore in the reedbeds around [Lake Alaotra]", the identity of which "remains a mystery" (p. 140). These reports described an animal similar to the Brown-tailed mongoose. During a 2004 wildlife census, members of the Durrell Wildlife Conservation Trust (DWCT) observed a small carnivoran swimming in Lake Alaotra. They captured it, photographed it, and released it (I'd love to see these photos: would have been nice to use them here). While it looked like a Brown-tailed mongoose, it differed in colour (being grizzled yellowish brown: the Brown-tailed mongoose is usually reddish brown).

The DWCT researchers were already familiar with Garbutt's 'mystery carnivore' of the Lake Alaotra marshes and were confident that their swimming mongoose was one and the same thing. In 2005, they made a special effort to discovery the identity of this animal. And they succeeded, capturing two individuals - one male and one female - in traps set on floating vegetation at Andreba, at the eastern edge of the Alaotra wetlands. The female was euthanised and now serves as the holotype for the species (reminder: a species doesn't get conservation status until it officially exists, and 'existence' is best demonstrated by the procuring of a specimen).

i-c09b4f6fef5166206cc96bc2a5911416-Garbutt-1999-Mammals-of-Madagascar-Aug-2010-2.jpg

Molecular analysis revealed that the genetic distance between the Alaotra mongoose and the Brown-tailed mongoose is small (less than that observed within some other euplerid taxa uncontroversially regarded as monotypic species). In fact, the authors state that - if they were relying on genetic data alone - separate species status might not be awarded to the new animal. However, it isn't clear that low levels of genetic variation necessarily disprove species status for any given population: indeed, some mammals universally regarded as 'good' species (like Mule deer Odocoileus hemionus and White-tailed deer O. virginianus) are extremely close genetically. And it's increasingly recognised that we can't rely on one line of evidence alone anymore: molecular data needs to be taken into account when examining possible new species, but so does morphological and behavioural data, and it is - hopefully - well known by now that there is no clear, objective way of determining 'how distinct' a population has to be before it can be recognised as one of those entities that we term 'species' anyway.

Statistical analyses of morphology showed the Alaotra mongoose to be well separated from Brown-tailed mongooses in skull and tooth proportions, and in the size and shape of its feet and foot pads. Compared to the Brown-tailed mongoose, the Alaotra animal has a paler, more olivaceous, grizzled pelage, it may be slightly smaller, it has broader feet with larger palmar and plantar pads, a broader muzzle, and a more robust lower jaw and dentition [image used at very top, from Durbin et al. (2010), compares the skulls and lower jaws of the two Salanoia species]. Many details of the dentition are distinctive: in comparison with the Brown-tailed mongoose, the surface area of the first upper molar is large in the Alaotra animal, its second upper molar is large compared to its first upper molar, its fourth upper premolar is broader, its canine is more robust, and so on.

In view of these obvious morphological differences, it seems appropriate to recognise the Alaotra mongoose as a new species. The name Salanoia durrelli of course honours Gerald Durrell (1925-1995), "inspirational writer and conservationist" (Durbin et al. 2010, p. 6) [photo of Durrell used at very top from here]. This is far from the first animal to be named after Durrell: Durrell's night gecko Nactus durrelli (originally named as a subspecies), endemic to Round Island, was described in 1994, Durrell's tadpole goby Benthophilus durrelli was described from the Don River Basin (in Ukraine) in 2004, the new Russian cossoid moth Kotchevnik durrelli was described in 2004, the Sri Lankan Durrell's freshwater crab Ceylonthelphusa durrelli was named in 2005, and the new Ecuadorian glassfrog Centrolene durrellorum was named after both Gerald and Lee Durrell in 2005.

i-ecc0a382c0aa63d3b2f77b3144a481ee-Marsh-mongoose-Trevor-Hardaker-Aug-2010.jpg

While virtually nothing is known of the natural history of S. durrelli, its association with a marsh habitat and its relatively robust jaws and dentition suggest that it preys on crustaceans and molluscs, and perhaps also on amphibians, fish and so on. Durbin et al. (2010) note that S. durrelli may be a Madagascan equivalent of the African Marsh mongoose Atilax paludinosus: a wetland species [shown above; photo © Trevor Hardaker, from here] that also uses floating mats of vegetation for feeding and resting, and also has relatively robust jaws and teeth that help it to eat molluscs, crustaceans and small vertebrates. The Brown-tailed mongoose, in contrast, is more insectivorous, though it does also feed on frogs, small reptiles and rodents.

It goes without saying that the Alaotra mongoose is deserving of immediate conservation priority. Lake Alaotra - a unique habitat, inhabited by an endemic lemur (the only primate that spends most or all of its life living above water) as well as highly endangered birds - has been ruined by introduced fish and plants, overfishing, and by acidification and silting-up resulting from local deforestation. The Alaotra grebe Tachybaptus rufolavatus hasn't been observed since the late 1980s and recent assessments have led to the conclusion that it's now extinct. Durbin et al. (2010) speculate that rats, dogs, cats and civets introduced to the Alaotra area may well have had an impact on the ecology of the new mongoose: like the other endemic species, it's likely to be threatened and declining, but more data is urgently needed.

For more on recently discovered living mammal species, see...

Lake Alaotra and its endangered birds were previously discussed in...

And, for more on euplerids and other cat-branch carnivorans, see...

Refs - -

Durbin, J., Funk, S. M., Hawkins, F., Hills, D. M., Jenkins. P. D., Moncrieff, C. B., Ralainasolo, F. B. 2010. Investigations into the status of a new taxon of Salanoia (Mammalia: Carnivora: Eupleridae) from the marshes of Lac Alaotra, Madagascar. Systematics and Biodiversity 10.1080/14772001003756751

Garbutt, N. 1999. Mammals of Madagascar. Pica Press, Mountfield.

Nowak, R. M. 1999. Walker's Mammals of the World, Sixth Edition. The Johns Hopkins University Press (Baltimore and London)

Yoder AD, Burns MM, Zehr S, Delefosse T, Veron G, Goodman SM, & Flynn JJ (2003). Single origin of Malagasy Carnivora from an African ancestor. Nature, 421 (6924), 734-7 PMID: 12610623

- . & Flynn, J. J. 2003. Origin of Malagasy Carnivora. In Goodman, S. M. & Benstead, J. P. (eds) The Natural History of Madagascar. University of Chicago Press, Chicago, pp. 1253-1256.

Tags
Categories

More like this

This is probably a pretty heretic comment, but at what point does a small local inbred population become a distinct species worthy of preservation? I mean, it's nice that we have the polka dotted mongoose, distinct from the yellow and blue striped one, but if we reestablish a breeding habitat connecting the two, would the differences be preserved or naturally blend back to the original drab brown mongoose?

It's an interesting thought, although re-establishing connected ranges wouldn't necessarily result in cross-breeding if the niche specialisation has gone far enough. Especially if hybrid offspring were less suited to either niche than the relevant specialist.

By stripey_cat (not verified) on 02 Aug 2010 #permalink

I see you posit that the divergence of the euplerids took place in the Oligocene. That might well be, but that would suggest the clade was distinct before reaching Madagascar. Intuitively, I'd assume that divergence would not have occurred until they became isolated. In that case, their divergence could not possibly date back beyond the Miocene as no terrestrial carnivorans are known to have populated mainland Africa before that date.
Of course, they could have been distinct before reaching Madagascar, but this would require subsequent extinction on the mainland. Also, as far as I know, fossil euplerids are unknown outside Madagascar.

Last findings at Late Oligocene in Kenya (Losodok) shows the presence of a pre-Miocenic Feliformia, Mioprionodon hodopeus, so it's not impossible that archaic feliforms (or proto-feliforms miacoids) reached Africa in Late Eocene or Early Oligocene, following the same biogeographical pattern of anthracotheres. This would explain why Feliforms is so diversified and apparently nested in Africa.

By J.S. Lopes (not verified) on 02 Aug 2010 #permalink

Absolutely thrilling stuff. The world is an awe-inspiring place. Although I take news like this with a big dose of reality--Madagascar doesn't have the best record of protecting its wildlife, a lesson we Americans have had to learn the hard way, too.

"the Fossa Cryptoprocta ferox"

Does that binominal translate as I think it translates?

By Andreas Johansson (not verified) on 02 Aug 2010 #permalink

Andreas: Yes.

By Jelle Zijlstra (not verified) on 02 Aug 2010 #permalink

Darren:

Mention 'new mammals' and 'Madagascar', and most people will assume that you're talking about lemurs

Actually, I suspect that most people, alas, will assume that you're talking about yet another sequel to that kids' movie with the talking CGI animals... /cynic mode off

Salanoia durrelli Durbin et al., 2010, discovered in the marshes around Lake Alaotra

Lake Alaotra and its surrounding wetlands aren't exactly tiny, but this still seems like an astonishingly restricted natural range for a carnivorous mammal. Is there currently any reason to believe that S. durrelli was formerly more widespread?

Durrell's tadpole goby Benthophilus durrelli was described from the Don River Basin (in Ukraine) in 2004, the new Russian cossoid moth Kotchevnik durrelli was described in 2004

Those two honorifics are, I strongly suspect, a legacy of Gerald Durrell's famous trip to the Soviet Union in 1984. That trip, which resulted in a book and a television documentary series that were both titled (slightly inaccurately) Durrell in Russia, was a pretty big deal back then. Before Durrell's visit, very few Western television crews had been allowed to travel and film so freely across the USSR. From a PR point of view, Durrell's visit was a success on both sides of the Iron Curtain, and he is to this day still relatively well known in Russia and the former Soviet republics.

Brian:

I see you posit that the divergence of the euplerids took place in the Oligocene. That might well be, but that would suggest the clade was distinct before reaching Madagascar.

Eizirik et al. (2010) also posit an - admittedly late - Oligocene divergence of euplerids from herpestids (25.5 MYA, to be precise).

Reference:

Eizirik, E., Murphy, W.J., Koepfli, K.-P., Johnson, W.E., Dragoo, J.W., Wayne, R.K. & OâBrien, S.J. 2010. Pattern and timing of diversification of the mammalian order Carnivora inferred from multiple nuclear gene sequences. Molecular Phylogenetics and Evolution 56, 49-63.

@1 & 2 - Something of the sort happened with the Atitlan grebe Podilymbus gigas -- extinct largely because of human changes in the lake ecosystem, but hybridization with the Pied-billed grebe Podilymbus podiceps seems to have contributed to its extinction. (Which raises the question, even if there are no Atitlan grebes alive, if its genes survive, is it really completely extinct?) The Alaotra grebe Darren mentioned seems to have had a similar relationship with the Little grebe.

By William Miller (not verified) on 02 Aug 2010 #permalink

from Wikipedia: "The Fossa was formally described in 1833 by Edward Turner Bennett. The genus name Cryptoprocta refers to how the animal's anus is hidden by its anal pouch (crypto- is Greek for "hidden", and procta is Greek for "anus")"

Interesting critter. However, all the data show that it is better treated as a new subspecies: Salanoia concolor durrelli.

We all like Gerald Durrell in any case!

Minor little semantics quibble: instead of saying
"However, all Madagascan mammals aren't lemurs"
you mean to say "However, not all Madagascan mammals are lemurs" otherwise you are saying that there AREN'T any lemurs on Madagascar.

And I still enjoy your blogs, I just don't get to them much: this one was pointed out to me by Chad Arment.

However, all the data show that it is better treated as a new subspecies: Salanoia concolor durrelli.

Why? The authors explained at length why they chose to treat it as a species; I'd be interested to hear a contrary argument, but just asserting it is of little use. By the way, it does seem Durbin et al. also changed their mind: there are a few places where they actually call it a subspecies.

By Jelle Zijlstra (not verified) on 03 Aug 2010 #permalink

Lets all hope He has throwback weekend at some point and hides a couple dinosaurs in the Everglade swamps.

Surely Mr Bayngor is kidding. No way anyone is that retarded.

By Bradley F (not verified) on 03 Aug 2010 #permalink

@Wolfgang Bayngor: Heresy! Surely it's not a post-Creation creation but just a new member of an existing baramin rapidly micro-evolved into being after the Flood. /snark

I am perfectly serious. You have two clearly visible "points" ( Salanoia concolor & Salanoia durrelli ) and you insist there is a "line" between them, because there "must" be a line between them according to your believe system (which you think of as "science"). That there was some sort of "transition" (I think you guys like to call it "evolution") between these two individual examples of The Creator's Work which occurred outside of your observation and over a long period of time is a BIG assumption which requires a massive dose of BELIEF (not science) to embrace. And thanks for the compliment, Bradley, but I am still a "pretard" and will only achieve "retard" status after I more thoroughly study and embrace your Evolutionary Belief System.
Mu, you may have to wait millions and millions of years for the dinosaurs to show up in the Everglades because The Creator works in that grand time scale. However, if you keep your eyes on the news, you will see from time to time that He does re-create, or more correctly, re-introduce, previously-removed species from His Earth back into His Earth. Witness the "discovery" of extinct prehistoric coelacanths off the Comoro Islands in 1938. This is only one example of, as you like to phrase it, The Creator's "Throwback Weekend". He does it all the time, occaisonally a knowledgable person will realize what he is looking at and the event makes Big News, but that knoweldable person still cannot make the correct connection as to the cause of the occurrence - A Benevolent Act of The Creator which has been repeated milions of times over a span of miliions of years.

Regards, Wolf

@Wolfgang
I certainly have impression that earlier species of hominids are walking around from time to time.

Sudden disappearance of dragons, fairies and ropens may be just an opposite act of Him.

@Jelle Zijlstra
Why just a subspecies? Small genetic distance, morphological differences are small and quantitative. We also don't know if
marshland habitat is really distinctive (it might live or recently lived in nearby forests).

On top of it, sample size of Durrell's mongoose is 2 (1 female specimen and 1 live male) so there is no way to estimate individual variability.

BTW, I think I commented before: biologists must learn to work in a world where natural habitats are fragmented, so most clinal variability is artificaly gone. Otherwise we will see more and more "species" being artifically created as there are fewer and more distant populations.

Ahhh Wolf, how correct you are - evolutionary science is all just one big belief system, just like religion.

We don't present EVIDENCE, we merely rely on BELIEF. Yup, no evidence. Definitely no organisms that are intermediate between other species, or groups of species. Definitely no new species, observed emerging from ancestors in the modern day. Definitely no transitional anatomical features, literally half-way between two opposing conditions. Definitely no transitional fossils, logically interpreted as intermediate between groups. Definitely no relicts of evolution scattered all over the place (like vestigial organs, illogical bits of useless DNA, redundant behaviors). And definitely no means, mechanisms or processes behind the most important unifying theory in the history of science.

Yeah, all we scientists do is cross our fingers, consult our favourite books, and blindly believe in something we all know to be a lie.

Of course, as a creationist you are religious. Allow me to put it to you that creationism is bad religion. The creationist God is a pretty poor one - not thinking to allow 'his' creations to change with conditions.

Now that it looks like you're serious...

ard[e]nt Creationist

Hence the expression "The stupid! It burns!".

(You know what "ardent" means literally?)

You have two clearly visible "points" ( Salanoia concolor & Salanoia durrelli ) and you insist there is a "line" between them

The other way around. We have two clouds, and we're debating (see comment 14) whether they still overlap or whether they've stopped doing so.

That there was some sort of "transition" (I think you guys like to call it "evolution") between these two individual examples of The Creator's Work which occurred outside of your observation and over a long period of time is a BIG assumption which requires a massive dose of BELIEF (not science) to embrace.

Nope, it's the simplest explanation for what we see.

1) There is inheritance; it is good but not perfect. Children are similar to their parents, but they're not exact copies, and they're not even exact combinations of traits of their parents â every human being (for instance) has on average 150 mutations with respect to their parents.

2) Some of these mutations have effects on the organism. (Most don't, but some do.)

3) Depending on the environment, these effects can (however indirectly!) be beneficial or detrimental to the reproductive success of the organism in question.

4) This means that the next generation will contain more descendants of organisms with beneficial mutations and fewer of those with detrimental ones. (Doesn't need to be a large effect, though it sometimes is.) This is called natural selection.

Any questions?

However, if you keep your eyes on the news, you will see from time to time that He does re-create, or more correctly, re-introduce, previously-removed species from His Earth back into His Earth. Witness the "discovery" of extinct prehistoric coelacanths off the Comoro Islands in 1938. This is only one example of, as you like to phrase it, The Creator's "Throwback Weekend".

Why do you make arguments based on a lack of knowledge?

You assume that Latimeria chalumnae literally didn't exist before 1938. It is nothing short of laughable to assume we had investigated every square meter of the planet back then, or even nowadays.

You further assume that Latimeria chalumnae is identical to any species we know from fossils. This, too, is not the case. Instead, it shows you simply don't know what you're talking about. No expert on coelacanths has any trouble distinguishing Latimeria from Mawsonia, Macropoma, Megalocoelacanthus, Axelrodichthys or what have you. They're similar (except in size), yes, but they're by no means indistinguishable even for laypeople.

================================

Why just a subspecies?

This depends on what you feel like today.

Seriously. The term "species" had 147 different definitions as of February 2008; depending on the definition, there are from 101 to 249 endemic bird species in Mexico. Some of these definitions do not allow for the existence of subspecies, which is a term that has a long list of contradictory definitions, too...

The International Code of Zoological Nomenclature does not say anything about how to define species & subspecies. Anything. It lets each and every one of us do whatever the fuck we want in this respect.

By David MarjanoviÄ (not verified) on 03 Aug 2010 #permalink

As an ardant Creationist I believe this animal is merely another example of Our Creator adding new species to His Earth periodically.

Come to think of it, I believe I did hear a *poof*

By darvolution pr… (not verified) on 04 Aug 2010 #permalink

Witness the "discovery" of extinct prehistoric coelacanths off the Comoro Islands in 1938.

At least get your facts right. The 1938 discovery of Latimeria chalumnae was made off the coast of South Africa. The Comoro Islands population wasn't discovered until years later. (Discovered by Westerners, that is; the Comorian people, of course, already knew about the existence of the gombessa.)

I see that Creator recreated not only coleacanth, but a common troll.

The mounted specimen of Salanoia that appears on this page is not from a website but is an unattributed photo of a specimen on exhibit at the Field Museum of Natural History in Chicago. It was probably "borrowed" from Nowak's Walker's Mammals of the World.

By Bruce Patterson (not verified) on 04 Aug 2010 #permalink

I took it from this website, and, yes, it must have been scanned at some stage from Nowak (the little label at bottom left is the give-away). It's p. 771 of the sixth edition of Nowak.

By the way, two very relevant articles that should have been cited above (sorry)...

Patterson, B. D. 2000. Patterns and trends in the discovery of new Neotropical mammals. Diversity and Distributions 6, 145-151.

- . 2002. On the continuing need for scientific collecting of mammals. Journal of Neotropical Mammalogy 9, 253-262.

In my defence, I have cited them before :)

That The Creator's Earth is littered with the numerous remains of a vast array of removed ("extinct"), re-introduced and current species is apparent to any observer. There is no question of their existence. The dispute is over the relationship between the different species.

You guys offer the "simple" explanation that one group "evolved" from a prior group, we Creationists offer an even simpler (and to us more obvious) explanation - The Creator simply put them on His Earth with the ability to reproduce THEMSELVES, until He decides to remove them either permanently or temporarily from His Earth. You Evolutionists would have us Creationists believe that two cats can make a dog, or two dogs can make a wolf, etc. And yet such has NEVER been demonstrated in one single scientific experiment. Certainly, as David suggests, there is intra-species, generation-to-generation variations. That is obvious to even us Ardent Creationists. But it is equally obvious that two dogs cannot make a wolf, no matter how hard "science" tries to do so, or how persistently (and stubbornly) you guys insist that the dogs could do so if they had enough time.

(That it may soon be possible to CONVERT a dog to a wolf using DNA engineering will no doubt provide more "evidence" or "proof" to yourselves of your cherished Evolution Belief System.....but that's a topic for another discussion so I'll leave it alone for now.)

This site is devoted to the science of zoology, not to theological views of the world that should have died out in the Middle Ages. As individuals, we are all entitled to 'believe' what we want. But claims that demonstrate nothing more than ignorance cannot be tolerated, and attempts to propagate a personal, fairy-tale view of the world should be exposed as the hilarious hokum they are. The fact of evolution has been observed and demonstrated many times in modern organisms, and if you dispute this you need to go away and do your homework. Creationism is not a logical 'alternative' point of view: it is mythological nonsense based on sheer belief in magic, on a twisted and very dumb interpretation of ancient religious texts, and on wilful ignorance.

I hate it when Tet Zoo becomes a sounding-board for nonsense, so please let me know if you want to see Mr Bayngor banned from the site. There are plenty of other places on the internet where creationism can be discussed.

Darren,
I think Mr. Bangor is making some good points. Calling you to task on the core principles of evolution science. Have two dogs ever been bred to make a wolf (or other non-dog species) in a scientific laboratory? Or is there another repeatable, controlled scientific experiment you can direct him to that irrefutably demonstrates the truth of evolution?

Ban him. Does this blog have the capability to add a 'banhammer' graphic to the offending posts? If so, I guess it would be appropriate to have a thagomizer or something...

Woodrow:

I sincerely hope you're a Poe. Anyways, evolution is not just speciation as you suggest, but the change in inherited traits in a population through generations. This can be easily observed in repeatable, controlled scientific experiments such as the evolution of citrate use in E. coli and change in coloration of guppies introduced to novel environments. Of course, Creationists love to move the goalposts and claim this somehow doesn't count.

Arguing with creationists would use up much / all the time you might otherwise use writing for the blog. I do not come here to read about creationism but about SCIENCE. So, ban if needed.

Thank You, Mr. Bonewright !
I don't much care whether these Evolution Believers come over to the Creationist view of the world. I'd just like them to admit that what they are presenting as "science" is founded on BELIEF (and a certain measure of self-delusion) not repeatable, solid, demonstrable science.

Of all the millions of generations of diptera (house flies) that have been bred in scientific labs for the past 150 years, has anything but a housefly ever hatched out of housefly eggs? How long should we wait for that to happen? Have you ever dropped a ball from your hand and have it "fall" up? No. How many times would you repeat that experiement to "prove" that it is impossible for a ball to "fall" up on His Earth?

PS: Darren, no need to take a vote. I'll leave you guys alone to intellectually stroke each other's egos for another post-Darwin century or two.

Regards,
Wolf

Typical creationist idiot.

'Woodrow T. Bonewright's' IP address: 66.216.160.118
Wolfgang Bayngor's IP address: 66.216.160.118

Jerzy: That's better; I certainly agree there's room for argument that it is not a species by itself. I think it's more distinctive than you imply, because there are some qualitative differences (presence of a natatory fringe, m2 occlusal morphology). As you noted, sample sizes are small, but it's outside the range of variation of _S. concolor_ in many traits. On the other hand, the mtDNA is as we know virtually identical among the two species (but that conclusion is based on even smaller sample sizes).

David Marjanovic helpfully gave us the number of possible species concepts, and whether or not this euplerid is really a species will depend on the one you use. I think it is best understood as a highly specialized local isolate of _S. concolor_; whether it should be regarded as a species or not is an interesting debate, but it may be one of semantics rather than science.

By Jelle Zijlstra (not verified) on 04 Aug 2010 #permalink

Sockpuppetry is one hell of a drug!

Anyways, thanks for the information about the Oligocene African feliform. It did make me wonder though: Feliforms display a seriously odd pattern of biogeography.
- The most basal living taxon, *Nandinia*, is an African endemic and as far as I know no definitive fossil relatives are known.
- A clade consisting of hyaenids, viverrids, herpestids and euplerids is probably African in origin with *Mioprionodon* suggesting this clade might date back to the Oligocene.
- Felidae is most likely Asian in origin, and the same is probably true of the prionodontids.So far, nothing really odd since Oligocene felids are known.
- Barbourofelidae is probably sister to either Felidae or Felidae + *Prionodon*. However, this clade does not appear until the lower Miocene in Africa.
- Percrocutidae, probably closest to the viverroid clade, is probably palaearctic in origin and does not appear until the early Miocene.
- Nimravids are probably very basal feliforms dating back to the Oligocene and are Asian in origin.

The oddness of all this lies with the fact that *Nandinia* and *Miopronodon* suggest that Africa was colonised by feliforms in the Oligocene *twice*, as the sistergroup to the herpestoids (+ *Mioprionodon* and percrocutids?) are decidedly Eurasian, with *Nandinia* being the sistergroup to the entire clade.
Also, barbourofelids evolving in the African Miocene seems odd with felids already existing in the Eurasian Oligocene, especially if barbourofelids are sister to Felidae + *Prionodon*.

David has certainly impressive knowledge of evolutionary concepts. I was interested in criteria practically used in field mammalology, which are more narrow and somewhat subjective. Forms with similar differences in other extant carnivores are generally considered races not species.

Which doesn't mean it is not interesting or worth of study and protection.

Darren, ban him and direct him towards the Endless Thread on Pharyngula, or keep him around for our entertainment and your increase in page views -- whatever you feel like. :-)

Of course, sockpuppeting is a bannable offense on Pharyngula.

You guys offer the "simple" explanation that one group "evolved" from a prior group, we Creationists offer an even simpler (and to us more obvious) explanation - The Creator simply put them on His Earth

For this, you first have to assume the existence of a creator.

That is a massive, huge assumption at which Ockham's Razor just laughs.

with the ability to reproduce THEMSELVES

That's another massive, huge assumption. What in the fuck is supposed to stop mutation and selection (as I have described them -- I note you did not dispute that description at all!) from accumulating over time? What unknown force is supposed to tell a water molecule "no, you must not react with this guanine residue, because then the DNA repair apparatus would replace it by an adenine residue, and that would be one mutation too many"? Did you ever think about this?

You Evolutionists would have us Creationists believe that two cats can make a dog,

That's a plain lie. The last common ancestor of cats and dogs must have looked more like some sort of intermediate between a marten and a palm civet.

or two dogs can make a wolf, etc.

The opposite is true -- dogs have been bred from wolves. By humans. Artificial instead of natural selection.

Of all the millions of generations of diptera (house flies) that have been bred in scientific labs for the past 150 years, has anything but a housefly ever hatched out of housefly eggs?

No selection pressure ( = breeding) was applied to make something new of those fruit flies.

Fruit flies, not house flies.

And Diptera (yes, it starts with capital letter) is the name of a very large group that includes all flies and all midges, including mosquitoes. Come on. How many errors can one pack into a single sentence?

whether it should be regarded as a species or not is an interesting debate, but it may be one of semantics rather than science.

As long as we don't agree on a single species concept, it is one of semantics. I don't see any scientific reason to agree on a single species concept. :-)

Barbourofelidae is probably sister to either Felidae or Felidae + *Prionodon*.

Has any phylogenetic analysis been done on this problem? One with more than, like, eleven characters???

Same for nimravids. IIRC, someone has argued they're caniforms, and someone else has argued they're not carnivorans at all; I don't know if the barbourofelids were supposed to be included in Nimravidae either time.

By David MarjanoviÄ (not verified) on 04 Aug 2010 #permalink

'Woodrow T. Bonewright's' IP address: 66.216.160.118

Wolfgang Bayngor's IP address: 66.216.160.118

That alone amply justifies banning the creotroll. Please do so, Darren, if you haven't already.

Of course if you do ban Wolfboner then we won't get see David open another rhetorical Tin of Whoopass on his head.

Brian:

The most basal living taxon, *Nandinia*, is an African endemic and as far as I know no definitive fossil relatives are known.

Nandinia sp. is known from the late Miocene of Kenya (Morales et al., 2005). But that, of course, doesn't tell us much about the deeper evolutionary origin of this taxon.

However, there are a couple of other fossil contenders for the title of being the closest relative of Nandinia. Véron (1995) investigated the phylogenetic affinities of the extant fossa Cryptoprocta ferox by performing a cladistic analysis (with 77 mostly hard-tissue characters) of more than 30 species of extant feliforms (including, incidentally, Salanoia concolor), a few canoid outgroup taxa, as well as a couple of fossil taxa. One of these fossil taxa, Palaeoprionodon lamandini from the Oligocene of France, formed a clade with Nandinia binotata, and these two taxa, in turn, formed a sister clade to all other feliforms. (It should perhaps be noted, however, that some of Véron's other results were quite odd. For example, she didn't recover a monophyletic euplerid clade; notably, Cryptoprocta turned out to be the sister taxon of the extant Felidae in her analysis!) On the other hand, Peigné & de Bonis (1999) noted certain morphological resemblances between Nandinia and Stenoplesictis, and suggested that these two together with Palaeoprionodon might be the most basal feliform taxa.

It would seem, though, that both Palaeoprionodon and especially Stenoplesictis are to some extent wastebasket taxa. A number of Late Paleogene species have been assigned to both of these taxa, and both genera are supposedly present in both Western Europe (France) and Central Asia (Mongolia), respectively (Peigné & de Bonis, 1999; Karl et al., 2007). Stenoplesictis is supposedly also present in the early Miocene of East Africa (Kenya), although the African species, "Stenoplesictis" muhoronii, should perhaps in fact rather be assigned to Palaeoprionodon (Peigné & de Bonis, 1999).

In other words, based on current evidence, both Palaeoprionodon sensu lato and Stenoplesictis sensu lato do seem to be very basal feliforms (and thus likely close to Nandinia), but both taxa are clearly in need of a thorough revision.

The oddness of all this lies with the fact that *Nandinia* and *Miopronodon* suggest that Africa was colonised by feliforms in the Oligocene *twice*

Two separate feliform colonisation events from Eurasia to Africa in the Oligocene? That doesn't seem so terribly odd or unlikely to me, considering how many other non-flying tetrapods have managed to disperse from Eurasia to Africa (or vice versa), both then and later.

References:

Karl, H.-V., Gröning, E. & Brauckmann, C. 2007. Comment on a fossil civet skull from the Lower Oligocene of the Weisselster Basin (Saxonia, Germany). Studia Geologica Salmanticensia 43, 215-225.

Morales, J., Pickford, M. & Soria, D. 2005. Carnivores from the late Miocene and basal Pliocene of the Tugen Hills, Kenya. Revista de la Sociedad Geológica de España 18, 39-61.

Peigné, S. & de Bonis, L. 1999. The genus Stenoplesictis Filhol (Mammalia, Carnivora) from the Oligocene deposits of the phosphorites of Quercy, France. Journal of Vertebrate Paleontology 19, 566-575.

Véron, G. 1995. La position systématique de Cryptoprocta ferox (Carnivora). Analyse cladistique des caractères morphologiques de carnivores Aeluroidea actuels et fossiles. Mammalia 59, 551-582.

When Mehmet Koseman and I made the Tet Zoo movie, I spoke about the community that's built up here - and David was specifically referred to as the Tet Zoo Bulldog. We should release a version with director's commentary some time :)

the new Ecuadorian glassfrog Centrolene durrellorum was named after both Gerald and Lee Russell in 2005.

Typo alert.

Witness the "discovery" of extinct prehistoric coelacanths off the Comoro Islands in 1938.

As Dartian has already alluded, the coelacanth was not unknown to the local people before this; it was a very rare bycatch of trolling for oilfish (probably a couple pulled up in a year). It was not a popular catch: a landed coelacanth apparently thrashes all over the place madly snapping at everything (and they are big fish) and, once subdued, has almost no practical usage whatsoever (like many sharks, the flesh is practically inedible; about the best that could be done with them is using the skin for sandpaper).

Reading these comments, I suddenly remembered something a friend at WWF once told me about the media response to other recent new discoveries. Journalists were phoning up and asking "So with all these new species appearing, do we have to worry about the ones that are going extinct?" They clearly thought that these 'new' species were not just newly discovered but actually new. I think they thought that they were 'newly evolved' but it's not that disssimilar to our creationist friend above.
Maybe a lot of the confusion that surrounds these announcements has to do with the fact that we're very fixated on the idea of nature being divisible into species an idea which was, at the end of the day, invented by creationists.
Wolfie T Bone or whatever his name said that it was obvious that two dogs could never make a wolf. It isn't obvious at all if you look at them - I doubt many laymen would be able to pick the wolf out of a line-up of domestic dog breeds. But it's somehow obvious because we know wolves and dogs are different species. Because we learned that when we were kids.
And if we want to know if two euplerids are different species we work it out by deciding if they're as different as other things we already 'know' are different species. And if they are we make a big press release because it will help raise conservation money. I'm not saying it's not big news but I feel that, to most people, it sounds even bigger than it actually is.

By farandfew (not verified) on 05 Aug 2010 #permalink

I agree that there's a lot of confusion attending the announcement of new species but I disagree that the main problem is in the species concept itself*. Rather, I would focus on the point that most people think that we know a lot more about the world's biodiversity than we do. The people you refer to are quite possibly thinking that these must be newly-appeared species because, if they were there before, surely someone would have named them already?

*Not that that isn't a problem; whatever your preferred definition, species are a lot fuzzier in real life than most people realise.

Koseman

Actually... Koseman or Kösemen? I think I've seen both versions.

trolling for oilfish

Trawling :-]

By David MarjanoviÄ (not verified) on 06 Aug 2010 #permalink

David:

Trawling

Trolling and trawling are two different fishing methods. In this case, I'm pretty sure that 'trolling' was right.

Ironically enough, I have a feeling (I may be wrong) that internet 'trolls' received their title by analogy to the practice of trolling - you drop a line with a hook down deep and see if you get a bite.

Trolling and trawling are two different fishing methods.

...WTF.

:-o

Must be like the lede of a newspaper article. ~:-|

internet 'trolls' received their title by analogy to the practice of trolling - you drop a line with a hook down deep and see if you get a bite.

I've read this several times; it's probably correct â what little I know of the trolls of Norse mythology, or even Discworld, doesn't fit them at all.

By David MarjanoviÄ (not verified) on 06 Aug 2010 #permalink

Must be like the lede of a newspaper article.

I don't see how, unless I'm overlooking how the words sound in other accents. When I say them, the two words have very different vowel sounds* (one rhymes with "roll", the other with "shawl").

*Yeah, that's right, two vowel sounds that sound more distinct when spoken in a New Zealand accent. Take that, John Scanlon!**

**Retaliation for a comment made at Tet Zoo perhaps more than two years ago. Sometimes I have a very long memory...

:D :D BTW, :D Norse trolls are based on very real animals. What tetrapod looks like a gray boulder, is small, hairy, bipedal, round-headed, big eyes, doesn't run away when scared, and is very strange and mean? I got this explanation from the book of late Swedish naturalist Carl Lindblad and find it very likely. ;)

"Ironically enough, I have a feeling (I may be wrong) that internet 'trolls' received their title by analogy to the practice of trolling - you drop a line with a hook down deep and see if you get a bite."

That's exactly correct - the internet usage of the term is a verb, although it's become extended to the bored chumps who make a habit of it as well. Always fun to see an attempt deflected by simple facts, the lack of anger and emotion in the responses is like Kryptonite to them.

Of course, there's an outside chance that our sock-puppeting buddy is actually serious, which is way too funny for words. Hopefully it's just a bored /b/ kid.

Great blogging as always, Mr. Naish. Thanks!

(one rhymes with "roll", the other with "shawl").

<facepalm>

Of course.

Yeah, that's right, two vowel sounds that sound more distinct when spoken in a New Zealand accent. Take that, John Scanlon!

LOL!

By David MarjanoviÄ (not verified) on 07 Aug 2010 #permalink

Jerzy:

What tetrapod looks like a gray boulder, is small, hairy, bipedal, round-headed, big eyes, doesn't run away when scared, and is very strange and mean?

Any young owl, but especially a young eagle owl Bubo bubo.

Swedish naturalist Carl Lindblad

Don't you mean Jan Lindblad?

Yes, Jan. Sorry.

Sounds like yet another case of using the so-called "Phylogenetic species concept" as an excuse to describe all well-defined subspecies as new species. Pseudoscience, but very fashionable, especially among birdwatching guides and primatologists.

BTW, naming Russian species after Durrell has nothing to do with his visits to the Soviet Union. His books have been immensely popular in the USSR since the 1960-s; Three Tickets to Adventure was the first "zoological" book I've read (in 1973, at the age of 4). There is still a bit of a cult following in Russia.

Sounds like yet another case of using the so-called "Phylogenetic species concept" as an excuse to describe all well-defined subspecies as new species. Pseudoscience

Why?

How you define "species" isn't a matter of science. It's one of convention â or, rather, of 147 different and competing conventions (as of February 2008).

By David MarjanoviÄ (not verified) on 14 Aug 2010 #permalink

For my part, I could just as easily (if not more so) accuse the so-called Biological Species Concept of representing pseudoscience, because of its claim to be representing something that is hardly ever actually tested, and usually somehow equivocated away in the face of conflicting evidence on the rare occasion that it does emerge (*cough* flickers *cough* Canis *cough* shall-I-go-on...) Species concepts are a hell of a lot fuzzier than most people think.

Vladimir's post made me think why field biologists have so clear ideas what 'should' and what 'shouldn't' be a species.

So, back to basics. Why we classify animals? To distinguish them for practical purposes. It makes no sense to name every individual separately, because many names would describe entities identical for all purposes. Nor to name everything as one species, because entities would have too many distinguishing characters to be useful.

So, what is ideal definition of a species? One which groups individuals so that one species optimises maximum shared characteristics within species and maximum different characteristics between species.

Field biologists understand this intuitively. Species should be alike, but different from others. So that is why Vladimir and most other field biologists reject species with minimal differences.

This also explains why paleontologists talk more in terms of genus than species. They talk "Tyrannosaurus eats..." like field biologist would say "tiger eats..." but not "Panthera sp. eats..." Most characteristics which distinguish species within a genus are unperceptible in the paleotological specimen (eg. color, fur). So the name shifts to higher taxonomic entity. And paleontologists talk about genera of dinosaurs as they were a species.

David: because there is absolutely no valid reason to replace the concept of subspecies with the nearly identical PSC, other than making birders' life lists longer and description papers better-looking in one's resume.

Christopher: the advantage of BSC is that it establishes some kind of a bottom line, while PSC doesn't. If you use PSC, you can describe any human ethnic group as a species, for example.

Jerzy: your last paragraph is true for birds and reptiles, but not for mammals, as mammalian systematics have historically been mostly skull systematics.

For those interested, I have a little essay on the subject: http://dinets.travel.ru/birdsys.htm :-)

David: because there is absolutely no valid reason to replace the concept of subspecies with the nearly identical PSC

As I've said elsewhere, it's not a question of 'replacing' the Biological Species Concept. The BSC has simply never been a real contender except in the field of rhetoric; in practice, the definition of a species has always been closer to 'what the reigning expert chooses to recognise as a species'.

Christopher: the advantage of BSC is that it establishes some kind of a bottom line, while PSC doesn't.

But that is exactly what the Biological Species Concept doesn't do. In practice, establishing real reproductive compatibilities is almost never a viable possibility. And compatibility/incompatibility is often not a sharp line (which, after all, is what one would expect of the evolutionary process); in at least some groups (flowering plants* and fishes are examples I've commonly heard referred to), such blurry boundaries are often in the majority.

*Botanists seem to have regarded the BSC as a red herring right from the start.

Christopher: PSC replaces the subspecies concept, not BSC.
Establishing the degree of reproductive isolation is often possible and has been done for numerous taxa. Anyway, I am not trying to push BSC here. It is pretty much obvious by now that no species concept can be universally good. What I am saying is that PSC is mostly bad, and that using it to lower the plank of what is a valid species is bad science.

Vladimir:

well-defined subspecies

Who has determined what a 'well-defined subspecies' is? Every serious attempt to define subspecies that I've seen seems to admit that subspecies are to a significant extent non-discrete taxonomic units (and this regardless of whether morphological or molecular criteria have been used). If we are supposed to retain the subspecies concept, then we also must come up with an objective way to establish the subspecific status of organisms. If we can't come up with such a way - and I'm not particularly confident that we can, given the difficulty virtual impossibility to agree on which species concept to use - then what scientific utility is served by recognising subspecies?

Jerzy:

Field biologists understand this intuitively.

Any species concept that requires intuition does not qualify as science. At the risk of pointing out the bleedin' obvious: your intuition is different from my intuition. And my intuition is different from Darren's, whose intuition is different from David MarjanoviÄ's, whose intuition sure as hell is different from that of the creationist type who appeared earlier in this thread, and so on, and so on, ad infinitum. Anything that's supposed to be scientific must have some universal and above all objective standards.

field biologists reject species with minimal differences

I hope that's not a position that you personally hold. You certainly know more than enough about zoology to realise that "I don't believe that X and Y are different species because I can't tell them apart" is just another version of the logical fallacy known as the Argument from personal incredulity.

Because what's a 'minimal difference', anyway? How can you be so sure that two animals that seem alike to you, a human observer, don't appear as very different to the animals themselves? How do you know (without asking for help from those pesky 'lab biologists' or whatever you like to call them) that they don't differ significantly regarding their vocalisations? Their odour? Their pheromone compositions? Their chromosome count and structure? What if they look different in some crucial detail in UV light? And so on. Who's to say that such characters, even if they can't be detected in the field by human observers, can't be every bit as taxonomically diagnostic as the way the organisms look? There's an awful lot of things of that kind that we simply do not yet know about even regarding very common organisms, never mind newly discovered ones.*

* For the record: in this current instance I tend to share your scepticism regarding the specific distinctiveness of Salanoia durrelli. I am inclined to believe that there is not yet sufficient data to make any good judgement on that; I suspect that we don't yet know enough even about S. concolor to be able to say with confidence that the new euplerid falls outside its interspecific variation. But I admit that I haven't been able to read the Durbin et al. paper yet (Jelle's link didn't work for me), so for the time being I'm withholding judgement.

Vladimir again:

Establishing the degree of reproductive isolation is often possible and has been done for numerous taxa.

But you can not apply the BSC to the vast majority of taxa for the simple reason that the vast majority of taxa that have ever lived are extinct. The PSC, on the other hand, can at least in principle be applied to almost any organisms, living or extinct. That's a huge and, as far as I can tell, a unique advantage of this particular species concept.

Dartian,
In ornithology, a well-defined subspecies is one that an average birdwatcher can identify. In other taxa, any subspecies with consistent characteristics can theoretically be recognized as a full species under PSC. If you apply PSC sincerely and thoroughly, you have to split the deer mouse (Peromyscus maniculatus) into at least 160 species - one on each isolated mountain range in each basin.
How are you going to apply PSC to an extinct group? Would you consider Neanderthals a full species or not? The answer depends on their ultimate fate: were they killed off by H. s. sapiens, or were they swamped genetically? You never no such things with certainty for closely related extinct species.

In ornithology, a well-defined subspecies is one that an average birdwatcher can identify.

Phylloscopus warblers. Seagulls. Empidonax flycatchers. Snipes... I'm inclined to doubt the claim that birdwatchers are a controlling interest in bird taxonomy.

Tangential to this discussion, and not directly relevant to the preference of one species concept over another: I should point out that, contrary to common assumption, reproductive isolation does not directly correlate with morphological divergence. Studies in reproductive incompatibility have found that it can be due to a relatively small mutation in a single gene, or even whether or not the mating couple are carrying compatible intracellular parasites (in which case, it may be possible to experimentally 'cure' speciation). It is entirely possible for two populations to be almost indistinguishable and yet completely distinct species under the BSC.

Vladimir:

In ornithology, a well-defined subspecies is one that an average birdwatcher can identify.

Leaving aside the matter of who exactly is 'an average birdwatcher', that still leaves the question of what, then, to do with all the not-so-well-defined subspecies. Are they to be discarded, and if they are, who shall make that decision and on what grounds?

If you apply PSC sincerely and thoroughly, you have to split the deer mouse (Peromyscus maniculatus) into at least 160 species - one on each isolated mountain range in each basin.

I am not particularly familiar with systematic studies of that taxon; do you mean that someone has performed, say, a multivariate analysis and established that different P. maculatus populations really form 160 distinct clusters? If that's the case, reference, please.

And, besides, even if consistency demanded that we should recognise >160 new Peromyscus species, would that really be such a big deal? A botanist wouldn't flinch at something like that; just to take one example, the dandelions Taraxacum of Scandinavia alone can be split into about 900(!) apomictic species. In my opinion, the argument that applying the PSC would result in 'too many' species does not really hold water. Unless one thinks that vertebrate/tetrapod/mammal classification should follow different rules than the classification of other organism groups (which, I suppose, is a view that one might defend on purely pragmatic grounds).

How are you going to apply PSC to an extinct group?

The same basic way that it can, and should, be applied to extant taxa: by quantifying their similarity by using cladistic methods. When analysing fossil taxa, one typically has much fewer characters to work with (for example, one usually doesn't have access to their DNA). But the principle is the same.

Would you consider Neanderthals a full species or not?

I'm not a paleoanthropologist, but based on the available evidence, and as far as I'm able to judge it: yes. Most modern workers consider them full species, and they've done so for many years now. (Luckily, the Neanderthals are among those few fossil taxa for which we do have molecular data too.)

The answer depends on their ultimate fate: were they killed off by H. s. sapiens, or were they swamped genetically?

Hybridisation is not some magic test of conspecificity. Some undoubted sibling taxa hybridise freely with each other, others apparently not at all. Yet others hybridise with more distantly related taxa, but not with what appears to be their closest relatives.

Christopher:

contrary to common assumption, reproductive isolation does not directly correlate with morphological divergence. [...] It is entirely possible for two populations to be almost indistinguishable and yet completely distinct species under the BSC.

QFT.

So, what is ideal definition of a species? One which groups individuals so that one species optimises maximum shared characteristics within species and maximum different characteristics between species.

"Nothing in biology makes sense except in the light of evolution" (Dobzhansky 1973); "Nothing makes sense in evolution without a phylogeny" (Gould & MacFadden 2002, 2004).

It follows that the nomenclature we use should be suited as well as possible to talking about phylogeny -- no matter which, if any, rank we assign to any taxon.

Field biologists understand this intuitively. Species should be alike, but different from others. So that is why Vladimir and most other field biologists reject species with minimal differences.

So you are arguing for a phenetic species concept, neither a "biological" nor a phylogenetic one.

David: because there is absolutely no valid reason to replace the concept of subspecies with the nearly identical PSC, other than making birders' life lists longer and description papers better-looking in one's resume.

What do you mean by "the concept of subspecies"? There are several, for the same reason why there are so many species concepts.

(There aren't as many, because some species concepts don't allow species to be divisible. But that still leaves a lot.)

If you use PSC, you can describe any human ethnic group as a species, for example.

This is wrong.

We are so panmictic that only two ethnic groups have ever been isolated for long enough that they might have constituted discoverable clades: the Easter Islanders for 400 years (I think you agree that's a very short time), and the Tasmanians for apparently 14,000 years. That's it.

Who's to say that such characters, even if they can't be detected in the field by human observers, can't be every bit as taxonomically diagnostic as the way the organisms look?

Indeed, there's a phenetic species of Paramecium that turned out to be 27 "biological" species that never interbreed. Yes, twenty-seven.

But you can not apply the BSC to the vast majority of taxa for the simple reason that the vast majority of taxa that have ever lived are extinct.

And even among the living, you can't apply it to asexual organisms. Ernst Mayr knew this full well and wrote that such organisms "do not form species". Except that the codes of nomenclature force us to pretend otherwise anyway.

In ornithology, a well-defined subspecies is one that an average birdwatcher can identify.

That's a phenetic concept, not a phylogenetic one.

In other taxa, any subspecies with consistent characteristics can theoretically be recognized as a full species under PSC.

No, because phenetically distinguishable assemblages do not need to be clades.

How are you going to apply PSC to an extinct group? Would you consider Neanderthals a full species or not?

Under the PSC, yes, because:

The answer depends on their ultimate fate: were they killed off by H. s. sapiens, or were they swamped genetically?

This doesn't matter at all. A phylogenetic species can end by extinction as well as by species fusion. Only under the BSC is species fusion by definition impossible.

do you mean that someone has performed, say, a multivariate analysis and established that different P. maculatus populations really form 160 distinct clusters?

This does not matter. You should have asked if someone has performed a phylogenetic analysis and established that the populations form 160 distinct clades, distinct in the sense that the phylogenetic analysis can tell them apart.

Phenetics is not phylogenetics.

by quantifying their similarity by using cladistic methods

No. Phenetic methods quantify total similarity. Phylogenetic methods count shared derived character states; this often leads to totally different results.

Hybridisation is not some magic test of conspecificity.

Under the BSC, it is, by definition. Under the others, however, hybridization is completely irrelevant.

(Except that I bet there are many species concepts which combine the BSC with others. But I don't know that.)

Refs:

Theodosius Dobzhansky (1973): Nothing in biology makes sense except in the light of evolution. The American Biology Teacher 135: 125 -- 129.

Gina C. Gould & Bruce J. MacFadden (2002): Gigantism, dwarfism [argh], and Cope's rule: "Nothing in evolution makes sense without a phylogeny". Journal of Vertebrate Paleontology 22 (supplement to issue 3): 60A.

Gina C. Gould & Bruce J. MacFadden (2004): same title, Bulletin of the American Museum of Natural History 285: 219 -- 237.

By David MarjanoviÄ (not verified) on 17 Aug 2010 #permalink

This is really interesting. I get the impression everyone's arguing from a different standpoint. It seems like everyone including Vladimir agrees that the BSC is not workable in practice and, as Christopher Taylor points out, when it is workable it doesn't actually give us the results we expect and want.
But Vladimir and Jerzy are talking about the species we know and recognise and I think what Vladimir is suggesting is that there are two taxonomic levels (at least for tetrapods), one which is represented by 'traditional' species and one which was, until recently, occupied by 'subspecies'. As David M and Dartian have pointed out, there hasn't traditionally been any respectable way of classifying subspecies. Well maybe now there is - the PSC. But Vladimir is asking whether we can't keep the 'traditional' species at a higher taxonomic level.
Unfortunately there are very strong political reasons for doing this when any of these taxa are potentially threatened. Field conservationists are generally very keen to produce new species and press releases, so there's strong political pressure on them to be splitters. S. durrelli will attract a bit more conservation attention to Lake Aloatra. This sounds like a good thing and its often said that conservation focus on the 'species' level is arbitrary and that subspecies should be worthy of conservation attention. But the fact is that we're way short of being able to mobilise the support needed to conserve the currently recognised species of tetrapod. If that number ends up being hugely increased and, if the additions are mostly indistinguishable to 'the average birder' from known forms then the whole idea of a species as something worth conserving gets eroded. And there's no guarantee we'll be able to establish anything else in its place.
So are we quite sure that the list of species we'd have recognised in 1990, on vaguely articulated phenetic terms is really a completely subjective product of the human mind? Even if it is, I still wish that the PSC had got called the Phylogenetic Sub-species Concept, rendering sub-species the objectively defined rank and leaving species, like genus, as something more vague.

By farandfew (not verified) on 17 Aug 2010 #permalink

David:

This does not matter. You should have asked if someone has performed a phylogenetic analysis and established that the populations form 160 distinct clades, distinct in the sense that the phylogenetic analysis can tell them apart.

Actually, what I primarily wanted to know was where that '160 Peromyscus maniculatus subspecies' came from in the first place. I had a casual look at a few sources (e.g., Wilson & Reeder's Mammal Species of the World, Third edition), but I didn't find any that list anywhere near that many deer mouse subspecies. A proper phylogenetic analysis of said taxon would, as you say, of course be a different thing altogether.

No. Phenetic methods quantify total similarity. Phylogenetic methods count shared derived character states; this often leads to totally different results.

My bad. I used 'similarity' in that context as a sloppy shorthand expression; it didn't occur to me that it can be interpreted as support for phenetics. Thanks for the correction.

dwarfism [argh]

Why '[argh]'?

But Vladimir is asking whether we can't keep the 'traditional' species at a higher taxonomic level.

Phenetic species can be larger or smaller than phylogenetic ones (or, of course, coincide with them). The ICZN doesn't allow subspecies to contain species...

Why '[argh]'?

Because I find that word deeply ridiculous. (If you translated it into German, the only reasonable conclusion from that would be that it's named after someone whose surname was Zwerg.) What happened to "nanism"?

By David MarjanoviÄ (not verified) on 18 Aug 2010 #permalink

Sorry, I am at work and can't keep up with this rapidly spreading discussion. I will just point out that PSC defines species as "A group of organisms that shares an ancestor; a lineage that maintains its integrity with respect to other lineages through both time and space. At some point in the progress of such a group, members may diverge from one another: when such a divergence becomes sufficiently clear, the two populations are regarded as separate species." (Wikipedia). So if the Neanderthals were killed off by modern humans, they do qualify; if they were swamped genetically through interbreeding, they don't. There's nothing about cluster analysis in the way the concept was formulated: all it requires is a "separate evolutionary trajectory" - which is extremely vague and virtually never known with certainty.

Hi folks,

There is a new publication in birds about species delimitation:

http://onlinelibrary.wiley.com/doi/10.1111/j.1474-919X.2010.01051.x/abs…

BTW. Some of you mentioned that species is better for conservation than subspecies. Development circles in eg. Indonesia are now aware of this publicity-motivated splitting and it generally makes them see sciencists with mistrust. What I expected some years ago.

[from Darren: sorry, delayed by filter]

I will just point out that PSC defines species as "A group of organisms that shares an ancestor; a lineage that maintains its integrity with respect to other lineages through both time and space. At some point in the progress of such a group, members may diverge from one another: when such a divergence becomes sufficiently clear, the two populations are regarded as separate species." (Wikipedia).

Or, alternatively, you could try reading the actual primary literature rather than Wikipedia's garbling of it. Nixon & Wheeler (1990) provided one of the most influential Phylogenetic Species definitions: "the smallest aggregation of populations (sexual) or lineages (asexual) diagnosable by a unique combination of character states in comparable individuals". So, for instance, the question of whether modern humans could be divided into separate species depends on whether diagnostic characters could be identified that are not shared between populations (as David already noted, this does not seem likely). Phylogenetic species are not directly related to "separate evolutionary trajectories"; this is of course what they are attempting to identify (as is the case with all species concepts, including the Biological Species Concept) but there will always be a time lag of varying length between two populations becoming evolutionarily separated and acquiring differences allowing for their separation to be recognised (Taylor et al., 2000 [no relation]).

Nixon, K. C., & Q. D. 2008. An amplification of the phylogenetic species concept. Cladistics 6 (3): 211-223.

Taylor, J. W. 2000. Phylogenetic species recognition and species concepts in Fungi. Fungal Genetics and Biology 31: 21-32.

re Phenetic species can be larger or smaller than phylogenetic ones (or, of course, coincide with them).

People always complain about PSC-motivated splitting. Have there been any examples of PSC-motivated lumping?

(sorry I am too stupid to quote other commenters in the proper way with the little line of dots on the left)

By farandfew (not verified) on 18 Aug 2010 #permalink

I'm not aware of any cases where the PSC has been specifically cited as justification for lumping. But of course there are countless cases where morphologically distinct forms described as separate 'species' have been synonymised due to turning out to represent extremes in a very variable population, different life stages or sexes, or fixed intra-population variants (such as with intrasexual dimorphism).

It is true that the PSC will naturally tend towards finer species divisions than the BSC. I tried to think a while ago of cases where the Biolgical Species Concept would separate taxa that the Phylogenetic Species Concept would synonymise; the only possibilities I could think of were biological species whose reproductive isolation was due to factors such as polyploidy or small genetic changes (such as chirality in snails in the paper cited below) where the two forms were not directly interfertile, but gene flow might still be maintained by regular spontaneous mutations from one state to another. I don't know of any real-life examples of this affecting taxonomy, though.

Ueshima, R., & T. Asami. 2003. Single-gene speciation by left-right reversal. Nature 425: 679.

So if the Neanderthals were killed off by modern humans, they do qualify; if they were swamped genetically through interbreeding, they don't.

No, in the latter case they count as a species that was a species for 600,000 years and then stopped being a species due to species fusion.

Nixon & Wheeler (1990) provided one of the most influential Phylogenetic Species definitions: "the smallest aggregation of populations (sexual) or lineages (asexual) diagnosable by a unique combination of character states in comparable individuals".

Do you see any reference to phylogeny in this wording? All I see is a purely phenetic species concept. Such a species is even allowed to be polyphyletic.

I am too stupid to quote other commenters in the proper way with the little line of dots on the left

<blockquote>quote goes here</blockquote>

By David MarjanoviÄ (not verified) on 19 Aug 2010 #permalink

Jerzy:

Development circles in eg. Indonesia are now aware of this publicity-motivated splitting and it generally makes them see sciencists with mistrust.

Are you suggesting that we should let non-biologists' opinions decide what is and what isn't a species?

Do you see any reference to phylogeny in this wording? All I see is a purely phenetic species concept. Such a species is even allowed to be polyphyletic.

A Phylogenetic Species Concept is not a species determined by phylogeny (not in the most common usage, anyway), it is the species that provides the most appropriate unit for phylogenetic analysis.

The species level is supposedly where phylogenetic processes of ancestor-descendant become more significant than tokogenetic processes of reticulate descent. Terms such as 'monophyly' are pretty much meaningless at species level because it is only above species level that monophyly becomes possible.

OK, so, here we have a canonical definition of PS (although there's a bunch of others, very different ones, which is why I used Wiki as source). The problems with it are that:
a) In practice in can lead to indefinite splitting. For example, you can split any dog breed, any anthropologically distinct ethnic group, and even some bloodlines with unique mutations (i.e. the Habsburgs) into separate species.
b) PSC splitting is usually unfalsifiable. You can virtually always find some difference - at least a single-locus polymorphism - between any two lineages. If you can't, there's still something to hang on, like difference in habitat (which is always present if you compare geographical populations).

Dartian: you are absolutely correct. We should never let non-biologists' opinions decide what is and what isn't a species! And that's precisely why splitting for conservation purposes is unacceptable. And equally unacceptable is catering to special interests (birdwatchers, insect collectors, etc.), like in maintaining Northern Parula and Western Gull as separate (from Tropical Parula and Glaucous-winged Gull, respectively) species despite scientific evidence that they should be lumped.

it is the species that provides the most appropriate unit for phylogenetic analysis.

If you have a nonmonophyletic OTU in your analysis, you're asking for trouble.

For example, you can split any dog breed, any anthropologically distinct ethnic group, and even some bloodlines with unique mutations (i.e. the Habsburgs) into separate species.

Dog breeds, yes, as long as they are carefully kept separate.

Any human ethnic group other than the aboriginal Tasmanians (from 14,000 years ago till colonization) and perhaps the Easter Islanders (till colonization from 400 years earlier)? No. We're way too panmictic. Way too much tokogenesis for phylogenesis to become visible.

This also takes care of your point b).

By David MarjanoviÄ (not verified) on 19 Aug 2010 #permalink

If you have a nonmonophyletic OTU in your analysis, you're asking for trouble.

But the point is that Least Inclusive Taxonomic Units (to use a perhaps less ambiguous term) cannot practically be demonstrated to be monophyletic or otherwise in any real sense because of the obscuring effect of reticulate heredity.

David: ethnic groups differ in facial features and vocal signals. A load of marmozet species have been recently split based only on facial features, without any molecular data or any evidence of at least partial reproductive isolation. The sole argument is that they are isolated by rivers, although everybody knows that rivers in the Amazon frequently change there course. I strongly suspect that these new "species" are completely bogus, just as most of recent lemur splits are.

But the point is that Least Inclusive Taxonomic Units (to use a perhaps less ambiguous term) cannot practically be demonstrated to be monophyletic or otherwise in any real sense because of the obscuring effect of reticulate heredity.

If so, they aren't LITUs. LITU = smallest recognizable clade.

(And yes, I like LITUs a lot.)

A load of marmozet species have been recently split based only on facial features, without any molecular data or any evidence of at least partial reproductive isolation. The sole argument is that they are isolated by rivers

Are the facial features at least phylogenetically informative? I mean, is there a tree-shaped signal in the data?

If not, I agree it's all bogus.

By David MarjanoviÄ (not verified) on 21 Aug 2010 #permalink

David: they certainly are in humans (major races can be split into sub-races, and so on), but probably not in marmosets, as all those splits are based on a very small number of decorative features that seem to show up rather randomly in different populations.

Striking facial differences between otherwise similar populations or closely related species seem to be a general pattern in primates. There is obviously some interesting evolutionary mechanism at work, something about kin recognition. I am actually considering writing it up as a research grant application, although for now it's just a very vague idea.

major races can be split into sub-races, and so on

Ha! It's not even possible to define the major races, because the variation is all clinal!

Seriously. I've seen books that claim 3 races, and ones that claim 66. Lots of people have tried to divide humanity into races, but consensus was never reached even in those times when everyone agreed it had to be possible somehow. It just doesn't work. There's a lot of literature that you apparently need to have a look at!

I don't know about the marmosets, but humans are almost panmictic.

By David MarjanoviÄ (not verified) on 22 Aug 2010 #permalink

No, it is not all clinal, not at all. Some borders between races have been maintained for centuries, even millenia, by assortative mating despite immediate proximity. The Negritos still look strikingly different from other Filipinos, the Bushmen look absolutely different from the Bantu peoples, the Burmese look nothing like the Bengalis, and the upland Ethiopians look completely different from Nilotic people despite having lived in adjacent regions for at least 2,500 years.

the Bushmen look absolutely different from the Bantu peoples

Zulu and Xhosa look a lot more like Bushmen than the average Tanzanian does.

the Burmese look nothing like the Bengalis

There the cline is entirely within Assam and maybe Arunachal Pradesh.

the upland Ethiopians look completely different from Nilotic people

I'll need to look into this.

By David MarjanoviÄ (not verified) on 24 Aug 2010 #permalink

Vladimir: which new marmoset species have been named for facial features only? Thanks.

By Bradley F (not verified) on 25 Aug 2010 #permalink

Thanks Vladimir. I assumed you had new information. I cannot see that you are correct. The species included in that list on wikipedia have not only been split on facial features as you state: on the contrary, morphological information from elsewhere and molecular data has also been used to support validity of these species. For more, see:

Nagamachi, C. Y., Pieczarka, J. C., Barros, R. M. S., Schwarz, M., Muniz, J. A. P. C., & Mattevi, M. S. (1996). Chromosomal relationships and phylogenetic clustering analyses on genus Callithrix, group argentata (Callitrichidae, Primates). Cytogenetics and Cell Genetics, 72, 331â338.

Sena, L., Vallinoto, M., Sampaio, I., Schneider, H., Ferrari, S. F., & Schneider, M. P. C. (2002).
Mitochondrial COII gene sequences provide new insights into the phylogeny of marmoset species groups (Callitrichidae, Primates). Folia Primatologica, 73, 240â251.

Ferrari, S. F., Sena, L., Schneider, M. P. C., & Silva Júnior, J. S. (2009). Rondonâs Marmoset, Mico rondoni sp. n., from Southwestern Brazilian Amazonia. International Journal of Primatology DOI 10.1007/s10764-010-9422-6

Tagliaro, C. H., Schneider, M. P. C., Schneider, H., Sampaio, I., & Stanhope, M. J. (2000). Molecular studies of Callithrix pygmaea (Primates, Platyrrhini) based on transferrin intronic and ND1 regions: implications for taxonomy and conservation. Genetics and Molecular Biology, 23, 729â737.

By Bradley F (not verified) on 25 Aug 2010 #permalink

Bradley,
Sorry, I am in the field now and can't go through the sources, so I can't continue with this discussion. I wish I could... Keywords "Phylogenetic clustering analysis" are a reliable warning sign for bogus splits. But this is probably not an appropriate place for such a lengthy exchange on such a complicated subject, anyway.

OK Vladimir. I look forwards to hearing more from you on this subject. But when you have time I would like to see you move from a general comment to the specific details. You have been saying that virtually all recently named marmosets are not "real" species and that they have only been separated on the basis of facial colouration. I would like to see the specific details, as I just cannot find a single species for which this is true.

By Bradley F (not verified) on 25 Aug 2010 #permalink

And the racial divide is still very obvious.

Yes, because migration is involved. The agriculture- and iron-driven Bantu expansion started somewhere around the northwestern corner of the Congo basin and only reached South Africa a few hundred years ago; that wasn't enough to establish a new cline.

Besides, Tanzanians are not pure Bantu

Few if any humans are pure anything, especially when migration is involved (as it is here, as I just mentioned).

By David MarjanoviÄ (not verified) on 26 Aug 2010 #permalink

I never thought i'd be posting the 100th comment on a Zoology blog. This is a damn fine blog by the way congrats!

By Carlos Pelaez (not verified) on 13 Oct 2010 #permalink

Yeah, I know this is an old post. But I just discovered this blog and had to point something out. I see there's been some discussion of species concepts above. Possible some of my ideas are expressed above, but a quick scan did not indicate this to be the case. So, we've got this paragraph here:

"Molecular analysis revealed that the genetic distance between the Alaotra mongoose and the Brown-tailed mongoose is small (less than that observed within some other euplerid taxa uncontroversially regarded as monotypic species). In fact, the authors state that - if they were relying on genetic data alone - separate species status might not be awarded to the new animal. However, it isn't clear that low levels of genetic variation necessarily disprove species status for any given population: indeed, some mammals universally regarded as 'good' species (like Mule deer Odocoileus hemionus and White-tailed deer O. virginianus) are extremely close genetically. And it's increasingly recognised that we can't rely on one line of evidence alone anymore: molecular data needs to be taken into account when examining possible new species, but so does morphological and behavioural data, and it is - hopefully - well known by now that there is no clear, objective way of determining 'how distinct' a population has to be before it can be recognised as one of those entities that we term 'species' anyway."

So far as I can tell--as I am not willing to pay $41 for the article!--what we have is genetic distance for a mitochondrial gene (cytochrome B). What, exactly, would we do with that data anyways? Trying to apply a single genetic distance measure to the differentiating species doesn't really make any sense under any of the prevailing species concepts. It implies that we've got some kind of handy mtDNA yardstick. Below X% difference, they're the same species; above X% difference, they're different species. This is simply bunk. Whether there's agreement on "how distinct is enough" is beside the point (although you're right, there isn't); there's at least general agreement that we're looking for distinctness. That's not what a cyt-b distance address.

Suppose we were looking at morphology. The question, basically, is: do characters show a single uniform distribution among the populations in question; or do some of them show bimodal (or trimodal, etc.) distributions with gaps between groups of populations? So you measure a bunch of characters and get some kind of answer. Then someone comes in and tries to use a distance approach: we just measure the length of the right hind leg. If the difference between right hind leg lengths is greater than 1", they're different species. That's the morphological analogue of using cyt-b distance, and it's stupid.

It's not a question of morphological vs. molecular data, it's question of good data vs. crap.

Patrick:

So far as I can tell--as I am not willing to pay $41 for the article!

You haven't read the paper, and thus don't know what it actually says, but are still going to criticise it harshly?

(As for not wanting to pay an admittedly outrageous sum for a single paper, I can sympathise - but there are alternatives, you know. You can go to your nearest science library. Or you can contact the authors directly (Paulina Jenkins is the corresponding author; you can find her e-mail by using Google Scholar) and ask for a PDF.)

It implies that we've got some kind of handy mtDNA yardstick. Below X% difference, they're the same species; above X% difference, they're different species. This is simply bunk.

Who are you arguing against? Nobody in this thread has defended any such yardstick claims, and the authors of the paper explicitly state that mtDNA difference is inconclusive evidence as far as justifying species-level separation is concerned.

Suppose we were looking at morphology.

The authors of the paper are also looking at morphology.

That's the morphological analogue of using cyt-b distance, and it's stupid.

Strawman arguments are pretty stupid, yes.

I confess that I was waiting for someone else to respond to Patrick's message, as I don't quite understand the issue he sees with the paper (which, yeah, he admits to not having read. Not a good starting point for a convincing argument, Patrick). As evidenced by the many comments above - and, I hope, by the text of my original article, too - we are far off from the time when everyone in the community agrees on 'how distinct' a given population has to be before it can be regarded as one of those sacred entities we term 'species'. Durbin et al. (2010) use both genetic and morphological data to show that what they call Salanoia durrelli is distinct from other salano populations, and - following the PSC - warrants recognition as a species. They used 55 different morphological measurements. So, "good data vs crap"? Really?

"You haven't read the paper, and thus don't know what it actually says, but are still going to criticise it harshly?"

So, have you read it? If not, why defend it? :-) And, if so, perhaps you could point out if the impression I got of their genetic methods from the abstract is accurate.

As for my local science library--New Mexico State University does not have a subscription to the journal. I didn't feel strongly enough to contact authors.

"Who are you arguing against? Nobody in this thread has defended any such yardstick claims, and the authors of the paper explicitly state that mtDNA difference is inconclusive evidence as far as justifying species-level separation is concerned."

Yes, the authors apparently know it doesn't really mean anything; but why exactly would you publish an article that uses a certain technique if you're aware that the technique is bunk?

"The authors of the paper are also looking at morphology."

So I guess you missed that I was making an analogy there. :-)

"As evidenced by the many comments above - and, I hope, by the text of my original article, too - we are far off from the time when everyone in the community agrees on 'how distinct' a given population has to be before it can be regarded as one of those sacred entities we term 'species'."

As I mentioned, I don't think that uncertainty is relevant here. Using an mtDNA distance value is well outside the realm of that uncertainty.

"Durbin et al. (2010) use both genetic and morphological data to show that what they call Salanoia durrelli is distinct from other salano populations, and - following the PSC - warrants recognition as a species. They used 55 different morphological measurements. So, "good data vs crap"? Really?"

Yes, they've apparently got good morphological data, and crappy genetic data. This precludes meaningful comparison between the two kinds of data, and undermines statements like "it's increasingly recognised that we can't rely on one line of evidence alone anymore". If the second line of evidence is useless and discarded, we're back to relying on one line of evidence.

I should mention that this struck a nerve with me because there's been a trend recently in US herpetology of similarly advocating multiple lines of evidence, but having one (or occasionally both) of them be either insufficient or too poorly analyzed to provide useful inference.

So, have you read it?

Yes, I have, thankyouverymuch.

why defend it?

Because you were misrepresenting the research and the researchers.

perhaps you could point out if the impression I got of their genetic methods from the abstract is accurate

You cherry-picked the molecular results in order to casually dismiss the validity of the new taxon, and completely ignored (and indeed weren't even aware of) the morphological evidence.

As for my local science library--New Mexico State University does not have a subscription to the journal.

And did you ask the librarians if they could order it for you?

the authors apparently know it doesn't really mean anything

A strong similarity in mitochondrial cytochrome b sequences between two specimens does not disprove the hypothesis that they are different species; it just means that this line of evidence doesn't offer support to that hypothesis (that's why I used the word 'inconclusive' in comment #102). The authors got the result that they did, and they reported it.

the technique is bunk

As in, mtDNA comparisons are never informative? Peer-reviewed citation needed.

So I guess you missed that I was making an analogy there.

No, I pointed out that you used the analogy to make an ill-informed (and therefore unjustified) strawman attack on the paper.

"You cherry-picked the molecular results in order to casually dismiss the validity of the new taxon, and completely ignored (and indeed weren't even aware of) the morphological evidence."

I attacked the molecular approach used. Nowhere did I dismiss the validity of the new taxon. And, yes, I was aware of the morphological evidence. I just didn't mention it because it wasn't relevant to the point I was making (again, my mention of "suppose we were looking at morphology" was an analogy, not a suggestion that the authors -didn't- also look at morphology).

"A strong similarity in mitochondrial cytochrome b sequences between two specimens does not disprove the hypothesis that they are different species; it just means that this line of evidence doesn't offer support to that hypothesis (that's why I used the word 'inconclusive' in comment #102). The authors got the result that they did, and they reported it."

Conversely, relatively high dissimilarity in cyt-b probably wouldn't have been informative either (unless it were just -way- the hell out there, I suppose). Some species are more variable than others; rates of sequence evolution vary between lineages and between portions of the genome. An mtDNA distance value just isn't a useful approach.

" the technique is bunk

As in, mtDNA comparisons are never informative? Peer-reviewed citation needed."

Some applications of mtDNA can be useful, although I suspect that uncritical acceptance of mtDNA results does frequently lead to incorrect inferences. One substantial problem is that incongruence between mitochondrial (or, in plants, plastid) data and nuclear data is fairly frequent; a recent paper on this topic, addressing possible causes of this, is here:

http://sysbio.oxfordjournals.org/content/60/1/87.short

There are several other problems. For instance, mitochondrial relationships are expected to be substantially over-resolved relative to the genome as a whole (i.e., mtDNA coalescence times are far shorter than nuclear coalescence times):

http://www.genetics.org/cgi/content/abstract/103/3/513

And the recovery of mtDNA clades does not indicate whether there is or is not interbreeding between those clades. Some set or another of mtDNA relationships will exist either way.

Those particular concerns, however, are probably not relevant in this case, as we're just looking at a distance measure (unless this is misreported in the abstract, perhaps?). As I suggested earlier, this is analogous to simply presenting the difference between two numbers and then inferring that those two numbers are, or are not, part of the same normal distribution. Having a general idea of what the expected variance in your data might be in other species helps a bit, but it's still just a bad approach to the question.

"No, I pointed out that you used the analogy to make an ill-informed (and therefore unjustified) strawman attack on the paper."

You've yet to point out in what way my criticism is ill-informed. :-)

Patrick:

I was aware of the morphological evidence

How could you be, if you don't have access to the paper? (How do you know what, exactly, the authors have measured/analysed and how? What if it's their morphological data that are, to use your expression, 'bunk'?)

"suppose we were looking at morphology" was an analogy

It was not a particularly good analogy. You wrote

If the difference between right hind leg lengths is greater than 1"[*], they're different species. That's the morphological analogue of using cyt-b distance, and it's stupid.

How do you know that that analogue is 'stupid'? Answer: you don't - you are just taking it for granted, because it 'makes sense' (to you) that such variations in right hind length must be irrelevant. But it should be obvious that as a matter of principle, there is no a priori reason to assume that a consistent difference in right leg length could not be just as useful a character for distinguishing different species as any other.

* This is a science blog. Please use metric units only.

uncritical acceptance of mtDNA

Yes, uncritical acceptance of conclusions drawn from mtDNA (or indeed any) data is silly, but... nobody who has been participating in this thread has uncritically accepted the mtDNA evidence! It was you who started ranting about that 'yardstick' stuff from out of the blue.

You've yet to point out in what way my criticism is ill-informed

Are you for real? Does it really need to be pointed out to you (repeatedly, no less!) that criticising a paper that you haven't read is about as ill-informed as it gets?

This is a science blog. Please use metric units only.

Hm, I got a bit carried away there; this is Darren's blog, after all, and he makes the rules here. My bad.

Oy. I'm getting the distinct impression you aren't reading my posts any more than absolutely necessary to come up with apparently outraged replies so I'll leave it alone except:

"Are you for real? Does it really need to be pointed out to you (repeatedly, no less!) that criticising a paper that you haven't read is about as ill-informed as it gets?"

Allow me to clarify slightly: you haven't shown that my not having read the paper has any impact on my comments regarding the applicability of a distance-based mtDNA method. So, my ignorance may just be irrelevant.

P.S.

Inches.
Feet.
Yards.
Miles.
Gallons!
F a h r e n h e i t...

And, yes, that last was childish, but it made me giggle so what the hell. :-)

So I guess you missed that I was making an analogy there. :-)

The first thing you learn when you write your first scientific paper is that when you're misunderstood, it is your fault.

Long version:

1) no matter what you do, somebody will misunderstand you;
2) therefore you need to work hard to minimize the opportunities for misunderstandings to occur.

As I mentioned, I don't think that uncertainty is relevant here. Using an mtDNA distance value is well outside the realm of that uncertainty.

I don't understand this. Please explain.

As for my local science library--New Mexico State University does not have a subscription to the journal.

And did you ask the librarians if they could order it for you?

Come on, Dartian, that's ridiculous. I really don't think US state universities are any richer than European universities -- and the U of Vienna has been cutting subscriptions for years. I don't have access to the Journal of Paleontology, for crying out loud! The subscription to the Canadian Journal of Earth Sciences was ended with 2000, the Journal of Systematic Palaeontology (which is insultingly expensive) was never ordered at all, and so on.

Patrick should ask the authors, who will gladly provide the pdf for free.

By David MarjanoviÄ (not verified) on 08 Apr 2011 #permalink

[from Darren: sorry, reported as spam].

"The first thing you learn when you write your first scientific paper is that when you're misunderstood, it is your fault.

Long version:

1) no matter what you do, somebody will misunderstand you;
2) therefore you need to work hard to minimize the opportunities for misunderstandings to occur."

Alas, you're too late on that advice. However, I certainly don't apply the same standards to blog commenting as I do to writing for peer-reviewed journals.

" "As I mentioned, I don't think that uncertainty is relevant here. Using an mtDNA distance value is well outside the realm of that uncertainty."

I don't understand this. Please explain."

Well, I'm not sure how much more I can do to explain why I think this is an inappropriate method. The gist is that distinction between species is fundamentally a matter of form groups that are disjunct from each other in character space. An mtDNA distance value is inappropriate to addressing this problem. For instance, if we envision a single continuously variable character and a few individuals of a putative new species, an appropriate way of addressing distinctness would be to estimate the mean and variance of the character in the closely related known species, and then assess whether or not the putative new species falls within that distribution. A distance value alone cannot do this.

A more sociological indication that this is a problematic approach is that the authors themselves apparently do not have faith in the results. This is a bit of a red flag. Whenever authors pursue a method and then disregard its results, I tend to assume that either: 1) they're hiding something (the method is valid but in opposition to a preferred hypothesis); or 2) the method is bunk (and the authors know this but did it anyways). Since I already think the method is bunk, I favor the latter interpretation.

Also, difficulties with assessing species boundaries based on mtDNA distances (or, for that matter, other assessments of mtDNA variation, as I indicated in a previous comment) have been pointed out repeatedly in the literature, largely in the context of DNA barcoding. For instance, here are a couple that pop up in a quick search:

http://sysbio.oxfordjournals.org/content/55/5/715.short

http://sysbio.oxfordjournals.org/content/55/5/729.short

And here's another on the general questionability of mtDNA variation as an assessment of population structure or species boundaries:

http://onlinelibrary.wiley.com/doi/10.1111/j.0014-3820.2002.tb01364.x/a…

As for asking the authors--my interest in this particular paper is basically exhausted at this point. I don't think Dartian really wants me to read it, anyways; it's too convenient a stick to beat me over the head with. :-) Well, fine with me.

Just to point out a simple example of why mtDNA is not a metric for species: there's a whole group of congeneric mushroom that have (so far as known) identical ITS2 sequence. There is no question that they are distinct species. It's simply that, for whatever reason, their morphological and chemical diversity is not mirrored by their ITS2 sequences.

Actually, no one gene (or combination of genes) is the gold standard for differentiating species. Another plant example is Eriogonum (a large genus of North American buckwheats). I read one heart-breaking thesis about a decade ago where the grad student sequenced three separate genes (nuclear and chloroplast) from a bunch of species throughout Eriogonum and related genera. She got massive combs with only minimal resolution in a few branches. The only thing she could conclude was that a) Eriogonum needed to be broken up (she had sufficient resolution to demonstrate it was paraphyletic, if not polyphyletic), and b) most of its sister genera needed to be redefined as well (they showed up all through Eriogonum sensu latu). Other than that, she could not get sufficient resolution with any of the genes (or combination of the genes) to determine the group's phylogeny. So far as I know, she left botany after publishing that, and she did everything right.

By heteromeles (not verified) on 08 Apr 2011 #permalink

Patrick:

my not having read the paper has any impact on my comments regarding the applicability of a distance-based mtDNA method. So, my ignorance may just be irrelevant.

Which is actually fair enough (and was never in dispute, at least not by me) - but it then follows that your whole original comment was largely irrelevant in the context of this discussion. If I may try my hand at the analogy-thingy: it would be comparable to a situation where someone shows up on a Star Wars movie discussion forum only to post a comment saying that Star Trek sucks. (Such a sentiment might, for all we know, be shared by many reglars in that hypothetical Star Wars forum - but sharing it there most likely wouldn't add much of substance to the discussion.)

David:

Come on, Dartian, that's ridiculous. I really don't think US state universities are any richer than European universities

I didn't express myself clearly there; I meant that Patrick could try and ask the librarians to order a copy of that specific article for him. I did not mean that they should start subscribing to the journal!

Patrick again:

difficulties with assessing species boundaries based on mtDNA distances (or, for that matter, other assessments of mtDNA variation, as I indicated in a previous comment) have been pointed out repeatedly in the literature

Not that I want to press the point too much, but for the record, there are also recent large-scale phylogenetic studies (e.g., Agnarsson & May-Collado, 2008) that have found the results of mitochondrial Cyt-b comparisons to agree well with results obtained by other methods. So mtDNA comparison does still have its proponents too.

I don't think Dartian really wants me to read it

Why wouldn't I want that? The more people read science (and about science), the better for everybody in every way.

it's too convenient a stick to beat me over the head with

The discussions can get a bit heated here sometimes (although if you want to see what a real piranha pool looks like, you should visit Pharyngula). But by and large, this is a pretty amicable commenting environment. One just needs to realise that if one is going to make blanket statements (of any kind) here, one must also be prepared for the eventuality that they will be challenged. And one should then keep in mind the maxim 'only science can refute science'.

As for 'beating you with a stick' - I may be a poopyhead sometimes, but I try (and surprisingly often successfully too) not to invest too much of my ego into blog discussions. Meaning, among other things, that I (usually) don't bear grudges afterwards, even if I would have 'lost' a debate. As long as Science is the winner, I'm cool.

---

Reference:

Agnarsson, I. & May-Collado, L.J. 2008. The phylogeny of Cetartiodactyla: the importance of dense taxon sampling, missing data, and the remarkable promise of cytochrome b to provide reliable species-level phylogenies. Molecular Phylogenetics and Evolution 48, 964-985.

I certainly don't apply the same standards to blog commenting as I do to writing for peer-reviewed journals.

Why? They have the exact same purpose: to explain what you think and why you think so.

Just to point out a simple example of why mtDNA is not a metric for species: there's a whole group of congeneric mushroom that have (so far as known) identical ITS2 sequence. There is no question that they are distinct species.

This depends on your species concept.

This depends entirely on the species concept you happen to have chosen.

You apparently favor a phenetic species concept that takes several morphological traits into account. DNA barcoders favor another phenetic species concept, one that only takes COX1 into account. Yet other people sneer at phenetic species concepts, and so on and so forth...

The difficulty in understanding what you meant lay in the fact that you seem to take your preferred species concept for granted (...as, BTW, the barcoders do with their own). In the words of Jack Slater: "Mighty big mistake."

Actually, no one gene (or combination of genes) is the gold standard for differentiating species.

Whatever you want to be the gold standard is the gold standard -- for you.

I am not being sarcastic.

I meant that Patrick could try and ask the librarians to order a copy of that specific article for him.

...They'd do that???

if you want to see what a real piranha pool looks like, you should visit Pharyngula

And that's still nothing. Just yesterday, a newbie "mens'-rights activist" complained about the "vitriol" there, and a regular just laughed at him and told him what she got in discussions elsewhere on the Internet was far worse, ranging from being called "a worthless piece of trash" to being told she should be raped.

By David MarjanoviÄ (not verified) on 08 Apr 2011 #permalink

They'd do that???

I used to do it occasionally when I was a student and didn't have access to certain journals. I asked the staff at my university if they could order papers for me from other university libraries (within the same country) and there was usually no problem with that, provided that the papers I wanted weren't too many pages long; they arrived as photocopies, via snail mail (this was a number of years ago, before the Age of the PDF). I should perhaps add that my experience may not be typical; I was on very good terms with our librarians, so it's possible that I got some preferential treatment... ;)

Anyway, generally speaking, it would surely at least be worth a try to ask the librarians if they could assist you; they are there to help you, after all.

I asked the staff at my university

...'library'.

David:

discussions elsewhere on the Internet was far worse

Ahem.

I've just emailed a pdf of the article to Patrick. I love libraries, but these days you don't need to visit one should you be in quest of a new paper. Yeah, like you didn't know that :)

Darren:

these days you don't need to visit one should you be in quest of a new paper

I should add that in my case, the stuff that I needed help to track down was usually old (from the 19th century, in some cases); publications that to this day do not exist in electronic form, by authors who died many decades ago.

Yeah, like you didn't know that

*nods*

Hm. More's happened here since I wandered off. So... in order...

heteromeles--I've heard that the case in Astragalus (the genus of plants, mind you, not the bone; not being too familiar with skeletons, I was a bit baffled the first time I stumbled across that one) is much the same as in Eriogonum; multiple markers have been applied without yielding any substantial resolution. I've been working on another genus (Boechera, in Brassicaceae) with similar problems; 7 nuclear loci give good resolution of a good chunk of the relationships, but there are still a fair number of big polytomies & poorly-supported nodes.

Dartian--"Which is actually fair enough (and was never in dispute, at least not by me) - but it then follows that your whole original comment was largely irrelevant in the context of this discussion."

Well, I was responding to stuff mentioned in Darren's article. Seemed relevant.

Regarding mtDNA--it certainly does provide good phylogenetic inferences some, probably most, of the time*. However, there are a number of limitations that I think are not recognized as often as they should be.

*In animals, I should say. In plants mtDNA is basically useless. Very, very, very little variation.

"The discussions can get a bit heated here sometimes (although if you want to see what a real piranha pool looks like, you should visit Pharyngula). But by and large, this is a pretty amicable commenting environment. One just needs to realise that if one is going to make blanket statements (of any kind) here, one must also be prepared for the eventuality that they will be challenged. And one should then keep in mind the maxim 'only science can refute science'."

I don't know what you mean about Pharyngula. I've always found discussions there to be polite and uncontroversial. Regarding my earlier comment, though, I was mostly being facetious.

David--Regarding different standards for papers vs. blog comments: the gist is, my time is finite. If yours is not, perhaps you could explain your secret. :-)

"This depends on your species concept.

This depends entirely on the species concept you happen to have chosen.

You apparently favor a phenetic species concept that takes several morphological traits into account. DNA barcoders favor another phenetic species concept, one that only takes COX1 into account. Yet other people sneer at phenetic species concepts, and so on and so forth..."

It is not my impression that those in the bar-coding field conceive themselves to be adopting a distinct species concept, but that they think they are using a method that will give results that are in accordance with other species concepts (e.g., identifying a degree of mtDNA divergence that corresponds with reproductive isolation, or morphological / genetic clustering of individuals). However, I haven't read the bar-coding literature in too much detail. Creating a species concept in which a pairwise mtDNA distance is theoretically coherent would require, I think, some pretty dramatic modification of our understanding of the word; if you're familiar with any papers that attempt to do this, perhaps you could point me in the right direction?

It is also worth mentioning that there is no particular reason you couldn't apply a reasonable phenetic species concept to mtDNA data (although I still wouldn't advise this as a particularly useful approach!). You could score sequence variation in a form that would allow you to just stick it in a PCA. And of course you can (more reasonably) use other kinds of genetic data (microsatellites, AFLPs, etc.) in multivariate statistical clustering. The gist being, I'm not advocating anything that is necessarily morphological.

Darren--Thanks for emailing that! It does indeed confirm that they're using what I consider a worst-case scenario for the use of mtDNA in assessing species boundaries.

Actually, ITS2 second fold works pretty well as a barcode region. I was investigating it as a potential tool for sorting out soil fungi and oomycetes. There was nothing "phenetic" about this. We were simply looking for a simple way to identify what was growing in a soil sample, and we thought that soil DNA extraction might work, if we had a sequence to focus on.

While ITS2 works wonderfully 99% of the time (you could use it to unambiguously identify species of most groups of fungi and oomycetes to the species level), it totally fails to differentiate one small branch of basidiomycetes. I understand that it also generally works for insects and other animals.

ITS2 isn't a gene. The RNA folds together (four folds) to bring together bits of the ribosome together so they can work. The length and sequence of the folding regions (particularly on the second fold) doesn't particularly matter, and so it accumulates mutations. Contrast that with discussions of species barriers, which typically revolve around things that keep individuals of different species from breeding. Mating barriers have nothing to do with ITS2 sequences whatsoever. It's just that, most of the time, mutations on ITS2 occur at approximately the same rate that pre- and post-zygotic mating isolation evolves. As with all bardcoding techniques, this technique depends on correlation, not causation. Because of that, it's going to fail sometimes.

By heteromeles (not verified) on 09 Apr 2011 #permalink

@122: I'm not surprised on Astragalus. I currently deal with some of the Astragalus from a conservation angle, and that's genus is my poster child for bizarre substrate preferences. I mean, a species that prefers fossil Cenozoic reefs high in the mountains, but only those where the soil is shallow enough to keep gophers out? And its nearest relative apparently prefers salt marshes enriched by oil washing in from nearby seeps? Weird plants.

I've got a list of horror stories from the plant phylogenetics, so I sympathize. The problem isn't necessarily the genes or species concepts. Sometimes it's the bosses and gatekeepers who insist that they require perfect resolution before they'll consider letting you publish.

By heteromeles (not verified) on 09 Apr 2011 #permalink

Ahem.

Thanks, bookmarked.

David--Regarding different standards for papers vs. blog comments: the gist is, my time is finite. If yours is not, perhaps you could explain your secret. :-)

1) Time, young padawan, isn't something you have. It's something you steal.

2) If you don't have time to write a comment, don't write one. :-|

It is not my impression that those in the bar-coding field conceive themselves to be adopting a distinct species concept,

That's their problem, not mine.

but that they think they are using a method that will give results that are in accordance with other species concepts

This works fairly well in many cases, but not in others, as everyone should have expected.

Creating a species concept in which a pairwise mtDNA distance is theoretically coherent

What do you mean by "coherent"?

It is also worth mentioning that there is no particular reason you couldn't apply a reasonable phenetic species concept to mtDNA data (although I still wouldn't advise this as a particularly useful approach!).

What do you mean by "useful"?

bosses and gatekeepers who insist that they require perfect resolution before they'll consider letting you publish

What? Perfect resolution? No polytomies at all allowed? That's moronic.

By David MarjanoviÄ (not verified) on 10 Apr 2011 #permalink

@125: There's a difference between combs and polytomies. There are whole groups that are career destroyers.

What happens?
--naive young grad student sees nothing published on a group. Ahh, a subject is found!
--If naive young grad student is lucky, they find out what happened to previous grad students who assailed that group, and find something else to work on. Otherwise...
--Naive young grad student tries every trick in the book to develop a phylogeny for the problem group. Morphology, mitochondrial genes, plastid genes, nuclear genes (for many years, it was difficult/impossible to get nuclear genes out of most monocots), etc, etc, etc, ad nauseum. The grad students dig into their own savings when the grants run out, they use their inheritances from dead relatives to cover lab costs. They take out personal loans to get data (yes, I've seen all this happen).
--Finally, they either drop out, or they have a long fight with their advisor, attempt to publish the combs or minimal resolution cladograms they have found, and attempt to get a job. Or perhaps they go to dental school.

Why should we publish combs and other negative results? Because it's humane, for one thing. For another, it lets people know where the real problem groups are. Right now, too many phylogeneticists use grad students as human mine detectors, and let the combs stay out of sight so that they will trip up competitors.

While the surplus of grad students makes this approach feasible, I don't think it's good science. The people who succeed in phylogenetics are those who are lucky as much as skillful. Luck in finding a tractable group doesn't train people to solve problems, it just helps them think they know what they know what they're talking about. That's not good.

By heteromeles (not verified) on 10 Apr 2011 #permalink

David MarjanoviÄ:

2) If you don't have time to write a comment, don't write one. :-|

Nope, sorry, you'll just have to put up with my comments not being as well-written as the stuff I send to peer-reviewed journals. :-)

It is not my impression that those in the bar-coding field conceive themselves to be adopting a distinct species concept,

That's their problem, not mine.

However, it seems that the goals researchers describe for their work are a reasonable place to start from in evaluating or criticizing their work.

but that they think they are using a method that will give results that are in accordance with other species concepts

This works fairly well in many cases, but not in others, as everyone should have expected.

I am skeptical on this point. On the other hand, a technique that doesn't make much sense would not be appropriately justified by arguing that it works well a lot of the time.

Creating a species concept in which a pairwise mtDNA distance is theoretically coherent

What do you mean by "coherent"?

How about "logically consistent"? For instance, that reproductive relationships play a fundamental role in determining the distribution of morphological traits forms a generally good (though not perfect) justification for the use of a morphological discontinuity as a criterion by advocates of the biological species concept. Some kind of similar link between what we think species are and/or how they form and the use of pairwise mtDNA distances in delineating them seems necessary.

What do you mean by "useful"?

Aw, come on. :-)

There's a difference between combs and polytomies.

What do you mean by "combs", then? Hennig combs -- large paraphyletic series? If so, what's bad about them?

it seems that the goals researchers describe for their work are a reasonable place to start from in evaluating or criticizing their work.

I'm more interested in what they do than in what they want to do.

but that they think they are using a method that will give results that are in accordance with other species concepts

This works fairly well in many cases, but not in others, as everyone should have expected.

I am skeptical on this point.

Why? To begin with, other species concepts commonly don't agree with each other. Depending on the species concept, there are from 101 to 249 endemic bird species in Mexico.

What do you mean by "coherent"?

How about "logically consistent"?

That would only mean it doesn't lead to self-contradictory results.

For instance, that reproductive relationships play a fundamental role in determining the distribution of morphological traits forms a generally good (though not perfect) justification for the use of a morphological discontinuity as a criterion by advocates of the biological species concept.

Not just "not perfect", but in many cases quite bad. The correlation between morphological characters and pre- and postzygotic barriers is commonly low enough that morphological and biological species concepts give quite different results (cryptic species, Darwin finches...). Besides, phenetic morphological species concepts can allow polyphyletic species as well as para- and monophyletic ones; biological ones don't allow polyphyletic ones; phylogenetic morphological species concepts (think apomorphy-based ones) allow only monophyletic ones.

It's just too risky to try to use one species concept as a proxy for another.

Some kind of similar link between what we think species are and/or how they form and the use of pairwise mtDNA distances in delineating them seems necessary.

This is only possible if you -- explicitly or implicitly, consciously or not -- use pairwise mtDNA distances as a species concept.

What do you mean by "useful"?

Aw, come on. :-)

Come on, tell me. It's not at all obvious what you mean. What is your use for a species concept? What do you intend to do once you have a species-level classification? What are species concepts good for in your book? Different people have very different answers for these questions, and I have no idea what yours might be.

By David MarjanoviÄ (not verified) on 11 Apr 2011 #permalink