Philosophy, DNA, Morphology and Taxonomy

[apistomaster]

Sorry, but I have more experience "collecting trout" than tropical fish and the same issues we are discussing here are the subject of often heated discussions among their experts and even endangered species legislation. These same problems with DNA, morphology and taxonomy have many parallels and extend to many taxa. Debates and discussions such as this one extend beyond fish to birds, insects, frogs, to name just a few; for some real nightmares, plants.

[raglanroad]

Personally I think studying both DNA and moprphological differences is not enough.

What would be enough ? As we can see, regulations based on Barcoding data results will be permitting what fish you get in the hobby - or for dinner - pretty soon.
and here's negative scientific rant.

http://www.citeulike.org/group/4555/article/4537543

DNA barcoding ("barcoding") has been proposed as a rapid and practical molecular method to identify species via diagnostic variation in short orthologous DNA sequences from one or a few universal genomic regions. It seeks to address in a rapid and simple way the "taxonomic impediment" of a greater need for taxonomic identifications than can be supplied by taxonomists. Using a complicated plant group, Solanum sect. Petota (wild potatoes), I tested barcoding with the most variable and frequently suggested plant barcoding regions: the internal nontranscribed spacer of nuclear ribosomal DNA (ITS) and the plastid markers trnH-psbA intergenic spacer and matK. These DNA regions fail to provide species-specific markers in sect. Petota because the ITS has too much intraspecific variation and the plastid markers lack sufficient polymorphism. The complications seen in wild potatoes are common in many plant groups, but they have not been assessed with barcoding. Barcoding is a retroactive procedure that relies on well-defined species to function, is based solely on a limited number of DNA sequences that are often inappropriate at the species level, has been poorly tested with geographically well-dispersed replicate samples from difficult taxonomic groups, and discounts substantial practical and theoretical problems in defining species. 10.3732/ajb.0800246

OK, so I won't send in any wild potatoes.

[raglanroad]

The reason there's no published DNA sequences for P. altum is because the species is a royal pain in the ass to sequence. I know.

Also even if it wasn't so diverged that hardly any primers fit it, it still wouldn't matter that the type specimen can't be sequenced - most type material can't, it's too old or it's been treated with formalin, or both - as you simply have to show that the fresh fish you've captured (ideally at the type locality) conforms to the type, then sequence the fresh fish. That's pretty much what type material is for.

Mike, have you barcode tested any Discus ?

[Mike_Noren]

Mike, have you barcode tested any Discus ?

They're easy to sequence if that's what you're asking; I've never identified a discus using barcodes so I don't know if COI is informative.
EDIT: a quick look at FishBOL suggests that the different species of Symphysodon differ by less than half a percent, so my guess is that COI will not be able to conclusively identify species of discus, but may be able to exclude some possibilities.

[raglanroad]

here they list altum and discus
http://www.fishbol.org/progress_list_2. ... &list=some

[grokefish]

There's a really good discussion I nearly missed here about modern systematics.

Mods, would it be possible to take the time to separate this from the "nonsense" about this Suzuki character, which is not particularly relevant?

Would be nice, so the evaluation of the science doesn't get buried.

Lots of remarks to discuss though. I will sleep, and then try to address some of the critical points.

Oh dear, that made me chuckle, well said Racoll!

Matt

[racoll]

First of all, sincere apologies for such a long post, but this thread is moving so fast....

Oh dear, that made me chuckle, well said Racoll!

I will edit the title to something more suitable - "Philosophy, DNA, Morphology and Taxonomy"

Thanks Mats

EDIT: a quick look at FishBOL suggests that the different species of Symphysodon differ by less than half a percent, so my guess is that COI will not be able to conclusively identify species of discus, but may be able to exclude some possibilities.

Yes, this a good case of where barcoding may fail. As the studies by Ready et al (2006) and Farias & Hrbek (2008) reported, these fishes are incredibly young, and exactly as Larry pointed out earlier, discus are in the process of diversification:

The usual mtDNA markers, as used in many phylogeographical studies, were not sensitive enough to pick up monophyletic groupings in discus. The "species" were a polyphyletic assemblage. I don't expect COI to be much different, although Farias & Hrbek got slightly better resolution with the control region, elucidating the blue discus clade, and they sampled 334 individuals from 24 locations. raglanroad, I don't think increasing the number of specimens will make any difference in this case, the group cannot be resolved into monophyletic groupings. That much is clear.

With cases like this the taxonomy will always be contentious, as the certain aspects of morphology (such as colour pattern) can be so plastic in such a short amount of time (due to selection, as discussed already), but Farias & Hrbek's data could not reject the hypothesis of a single variable species under a variety of species concepts.

I would like to read the Bleher et al (2007) Aqua paper, but I can't get it.

Anyway enough of the discus digression and back to barcoding...

However, the cases (such as above) where barcoding fails to identify species is rare, and the biological causes are fairly well understood. COI barcoding works well in the overwhelming majority of cases.

Raglanroad, you mention the potato paper; plants present their own unique problems, and researchers are just beginning to get to grips with how to solve that problem. A couple of important papers were published on this recently.

As we can see, regulations based on Barcoding data results will be permitting what fish you get in the hobby

Yes indeed. This is beginning to start happening as regulations are being applied to the hobby on the export and import side of the trade.

so about our Type Material for Hypancistrus zebra; How was it collected by the biologists ?

Point is that institutions take specimens for vouchering on faith -when they want to. Chain of possession could start with "local fish market" .

It is quite common for researchers to purchase fishes from local fishermen, on which to base their descriptions. These locals after all know the habitats and fishes best, so are able to catch them much more effectively.

I'm not saying barcoding and molecular data is useless, I consider it a very powerful tool, but that one should be aware that molecular data is not inherently truer or less likely to be wrong than morphological data. Molecular data do have pitfalls, they just happen to be different ones from those of morphology. No magic bullet.

My view exactly.

My main problem with the barcoding crowd is the tendency to confuse identification of species with definition of species.

It seems that we are in agreement in that DNA barcoding should be used in conjunction with other techniques.

Yes there was at least a few years ago, a considerable confusion. Barcoding is a tool for routine species identification and rapid heuristic biodiversity assessment. It is not, and never really has been seriously argued as, a replacement to "traditional" taxonomy or DNA taxonomy.

Barcoding is not DNA taxonomy, although DNA taxonomy often uses the COI barcode locus together with other multi-locus data.

A modern taxonomist will use whatever data are appropriate, morphological or molecular, to answer his or her questions.

I can tell you one thing, though, some scientists, especially the morphological systematists, are opposed to the idea of cryptic species.

I guess that depends on the species concept one follows

[raglanroad]

It is quite common for researchers to purchase fishes from local fishermen, on which to base their descriptions. These locals after all know the habitats and fishes best, so are able to catch them much more effectively.

Is it not also quite common that fishermen do not accurately reveal the true location of their fishing ? Or that SA markets may contain fish shipped from SA to Miami and then back to major SA markets ? Sellers might mislabel their stock.
http://www.journals.uchicago.edu/doi/abs/10.1086/339216
Scientists unfamiliar with the fish they study, do buy the fish from the market and do scientific studies, which are then cited and cited, or perhaps somewhat like Corydaras julii from the Rio Negro - so well suited to the ion-poor waters - it's just mistaken ID by the scientists ? If the fish were purchased and were indeed Corydoras julii (shipped along who-knows-which trade routes)...the researchers then watched the julii re-adjust to the de-ionized or Rio Negro water ! Or maybe they were actually trilineatus collected just earlier.
Rio Madeira discus anyone ?

raglanroad, I don't think increasing the number of specimens will make any difference in this case, the group cannot be resolved into monophyletic groupings. That much is clear.

I was pointing to the 23 specimens that rewrote Discus, so that extrememly high numbers of specimens was not proposed as a necessity/handy roadblock.

Barcoding is not DNA taxonomy, although DNA taxonomy often uses the COI barcode locus together with other multi-locus data.

Getting an automated online divorce is not the study of Law.
How would Hypancistrus zebra go down with barcoding ? You know, you could just use all the squashed fish or contact Jeff's customers.

[Suckermouth]

However, the cases (such as above) where barcoding fails to identify species is rare, and the biological causes are fairly well understood. COI barcoding works well in the overwhelming majority of cases.

For now the cases where barcoding may fail might be rare, but as research continues in DNA barcoding I'm sure these "rare" instances will keep popping up. I'm sure that quite a number of Neotropical species are quite young, which may make identification through DNA barcoding troublesome for certain species pairs. One study found that Potamotrygon motoro, P. scobina, and P. orbignyi "share haplotypes extensively," although they did find separation in the other Potamotrygon species tested in the study (Toffoli, 2008). Has DNA barcoding been applied extensively to Neotropical fishes and to (the other big freshwater diversification event) African cichlids? Have they found that it has been accurate at the same rate (up to 98%) as studies in other taxa? I'm not asking to argue against barcoding, I'm posing the question since I don't know.

I can tell you one thing, though, some scientists, especially the morphological systematists, are opposed to the idea of cryptic species.

I guess that depends on the species concept one follows

True. It's likely annoying to a traditional taxonomist to have to identify a species a species by testing it in a lab rather than just by looking at it while they're pulling it out of a net. Descriptions of species based on external morphology makes species identification inexpensive and without a lot of complicated and expensive equipment.

[apistomaster]

To some extent,local food fish markets and to a greater extent, the aggregators of tropical fish collected from over a very broad area do serve a purpose. The large export centers often are the first point where someone will discover a new species. That often leads to heightened interest in learning where they can find more of the same new discovery.

For the sake of this discussion, I do have a low opinion of classifying or naming a new discovery without using actual type localities although I realize classification and type locality are not mutually exclusive requirements to proceed with classification. There is often a human bias to name new discoveries expediently for a variety of reasons, without first bothering to find out where the fish came from. Just seems a little lazy if not sloppy depending on how many specimens are available to that first author.
One or two specimens makes a pretty thin data set.

The wide range and often extensive difference between populations of a species or how ancient or how recent the divergence can make having statistically significant numbers of specimens from as many different locations has to be helpful in the long run even if it causes some confusion to fish keepers early on. We fish keepers/breeders generally share the philosophy of doing our best to breed fish together that are from the same population to preserve their distinctiveness at the very least and avoid genetic incompatibilities at the worst. The latter is not rare but often does not become an evident problem until proceeding into successive generations. It can take some time before breeders discover they have a problem with breeding different populations of the putative same species. This is widely seen among the Cyprinodontiformes, more generally called killiefish by aquarists.
Accurate collection data and maintaining the pedigree of the aquarium bred generations is a very practical matter for some aquarium fish. I have been rather surprised at how frequently some of the Corydoras hybridize in aquariums. They are one example but I see plenty of opportunity for similar problems arising with Ancstrini Catfish and many of the Cichlidae, Apistogramma. The latter is even more problematic due to their considerable sexual dimorphism and strong resemblance of females of different species.
Ornamental tropical fish breeders have parochial concerns that are of no great importance to science in general.

[raglanroad]

fishbol lists one furunculus barcoding

[jimoo]

For the sake of this discussion, I do have a low opinion of classifying or naming a new discovery without using actual type localities although I realize classification and type locality are not mutually exclusive requirements to proceed with classification. There is often a human bias to name new discoveries expediently for a variety of reasons, without first bothering to find out where the fish came from. Just seems a little lazy if not sloppy depending on how many specimens are available to that first author.
One or two specimens makes a pretty thin data set.

It has been (gasp) about 15 years since I had mammology and ichthyology, but are not paratypes/holotypes with a collection locality not required for classification? Or is it indeed ok to use a "_______ market, Cambodia" for a location?

[Mike_Noren]

However, the cases (such as above) where barcoding fails to identify species is rare, and the biological causes are fairly well understood. COI barcoding works well in the overwhelming majority of cases.

The exact same argument, and usually in the exact same cases, can be made for morphology alone.

This is beginning to start happening as regulations are being applied to the hobby on the export and import side of the trade.

I don't think barcoding is anywhere near cheap & fast enough for primetime use with live fish yet. Sure one could develop a microarray to identify (many) Hypancistrus, but who would sink a million dollars into a project like that? That kind of thing works great when people are importing tonnes of canned tuna, but when you're importing 200 live catfish with shaky ID's? The reality is that the trade will be stuck with fin clippings and a days work for a lab tech at a cost of about $100 per sample, meaning that any more than random sampling is probably impractical.

I guess that depends on the species concept one follows

Somewhat ironically the "standard" morphological species concept is homologous to the "molecular" species concept: two groups of animals are sufficiently different in some character that a competent taxonomist consider them separate species.
The dislike of molecular species is really an instinctive dislike of species which are impossible to identify without access to a molecular lab.

[raglanroad]

is it indeed ok to use a "_______ market, Cambodia" for a location?

Yeti skull puchased in market would do if it were written up in a scientific manner, I suppose.
then National Geographic could get on the case.

[Suckermouth]

is it indeed ok to use a "_______ market, Cambodia" for a location?

Yeti skull puchased in market would do if it were written up in a scientific manner, I suppose.

Unfortunately, a comparable situation to this has occurred in the past:
http://en.wikipedia.org/wiki/Piltdown_Man

Science marches on, though.

[Mike_Noren]

is it indeed ok to use a "_______ market, Cambodia" for a location?

If that's all you've got. The locality should be as detailed as you can and according to the best of your knowledge, but if you're not likely to pin it down better than "_____ market, Cambodia", then that's the locality in the description, and other researchers will have to take it from there.

See e.g. this: http://www.nytimes.com/2005/05/12/science/12rodent.html

This kind of thing was much more common in the past, nowadays you really want a reasonable capture locality for a description, not least because museums will no longer accept specimens without a paper trail (due to CITES and other regulations).

Yeti skull puchased in market would do if it were written up in a scientific manner, I suppose.

For the locality.

[raglanroad]

Axelrod could handle the job.

[raglanroad]

is it indeed ok to use a "_______ market, Cambodia" for a location?

Yeti skull puchased in market would do if it were written up in a scientific manner, I suppose.

Unfortunately, a comparable situation to this has occurred in the past:
http://en.wikipedia.org/wiki/Piltdown_Man

Science marches on, though.

http://www.indiaenvironmentportal.org.in/node/24127

[Suckermouth]

http://www.indiaenvironmentportal.org.in/node/24127

Well, at least the description wasn't published in a peer-reviewed journal. Or at least, both Science and Nature didn't publish it.

[raglanroad]

And what happens in case of a very very rare creature ? do you need one in a bottle in an institution in order to get a binomial ?

[MatsP]

Searching the Cat-eLog for "market" in the "type locality" shows several fish that are indeed from "____ Market ____".

Schilbe uranoscopus:
Fish market in Cairo, Egypt.

Wallago micropogon:
Stung Treng morning market, 13°30.0'N, 105°58.0'E, Cambodia.

Pterycryptis anomala:
Market in Hong Kong.

Hemibagrus olyroides:
Sintang Market (reportedly caught in Sungai Kebian), Borneo, Indonesia.

Hemibagrus chrysops:
Serian wet market, from Sadong R., Sarawak, Borneo.

Hemibagrus velox:
Sungai Dareh, Pulau Punjung market, Sumatera Barat, Sumatra.

Mystus castaneus:
Serian market, from Sungai Sadong, Sarawak, Borneo.

Mystus mysticetus:
Nakorn Phanom market, Thailand.

Mystus albolineatus:
Prachinburi market, Bangpakong basin, Thailand.

Mystus multiradiatus:
Prachinburi market, Thailand.

Mystus falcarius:
Myitkyina market, Kachin State, Myanmar.

Batasio fasciolatus:
Market at Malbazar, 26°32'30"N, 88°44'17"E, West Bengal, India.

Sperata aciculuris:
South Oak-ka-lar-pa market (e. Yangon), Yangon Division, Myanmar.

Bagarius rutilus:
Market in Hanoi, Viet Nam.

--
Mats

[Suckermouth]

And what happens in case of a very very rare creature ? do you need one in a bottle in an institution in order to get a binomial ?

It's a rule that there must be a holotype for species descriptions after 1999, yes.

[raglanroad]

In a bottle ? No specimens from the undead allowed ? Just the binomial, that's what we're after with BiGFoOt.
And what of those described before 1999, then ? No problemo ?

[jimoo]

Well excellent, I'm heading to Cambodia next month, will be keeping an eye out in the markets!

[raglanroad]

name it something good, like "pentasuzukiorum".

[MatsP]

In a bottle ? No specimens from the undead allowed ? Just the binomial, that's what we're after with BiGFoOt.
And what of those described before 1999, then ? No problemo ?

Well, it needs to be possible to go back and find it in the same condition [hopefully - there are quite a few examples of 200 year old samples that are a bit naff nowadays, as they haven't been looked after properly] as it was when it was used for the description - that's pretty hard if you happen to release it back into nature, or if you keep it in a tank and it grows bigger, or in a zoo or whatever.

--
Mats

[racoll]

Scientists unfamiliar with the fish they study, do buy the fish from the market and do scientific studies
Is it not also quite common that fishermen do not accurately reveal the true location of their fishing ?

raglanroad, unfortunately in the real world, researchers do the best they can given the time and resources available. This might mean less than perfect locality data or sample size, but at least the species is named and recognised. If scientists waited until their data were absolutely perfect, nothing would ever get published or achieved. With so much undescribed diversity and so few researchers doing this work, they need to just be able to get on with it.

Going back to the discus, Ready et al. did not describe a new species, they merely applied an available name (Symphysodon tarzoo) to the green western Amazon discus, which despite the others not being the resolved, were recovered as monophyletic in respect to one another, in both genetic and morphological data. This is perfectly biologically valid.

Where the contention lies, is with which name should be applied to which population, which is a nomenclatural rather than biological issue.

Has DNA barcoding been applied extensively to Neotropical fishes and to (the other big freshwater diversification event) African c*****ds?

No not at all Milton, and I would imagine that barcoding will struggle with many of the Lake Victoria cichlids for example. These rapid radiations will trouble any identification and classification system.

Descriptions of species based on external morphology makes species identification inexpensive and without a lot of complicated and expensive equipment.

Not quite correct in my opinion. How much do you think an experienced taxonomist charges per hour for biodiversity assessment work? For a large sample, it will also likely require a team of experts all trained in many different groups a long time to accurately identify everything. Even if a parataxonomist wanted to carry out the work, often keys do not exist, or literature too obscure to access.

The dislike of molecular species is really an instinctive dislike of species which are impossible to identify without access to a molecular lab.

As Mike_Noren said earlier though, in many cases, these cryptic species can actually be differentiated upon closer inspection, and differences previously ascribed to intraspecific variation become more obvious when group hypotheses can be formulated a priori with barcode data.

[raglanroad]

In a bottle ? No specimens from the undead allowed ? Just the binomial, that's what we're after with BiGFoOt.
And what of those described before 1999, then ? No problemo ?

Well, it needs to be possible to go back and find it in the same condition [hopefully - there are quite a few examples of 200 year old samples that are a bit naff nowadays, as they haven't been looked after properly] as it was when it was used for the description - that's pretty hard if you happen to release it back into nature, or if you keep it in a tank and it grows bigger, or in a zoo or whatever.

--
Mats

that's the rationale behind keeping a jar..but is it the code ? And if prior to modern times ( circa 1999) there was no requirement to have an actual critter, how do you go back and look at it, in that case ? So even though there are 200 year old specimens, that doesn't mean that it's necessary to have one back then, or before modern 1999 times. Thus, if a name was given without a critter in a jar, then retroactive satisfying of the need for availability of said non-existent specimen, in order to go back and re-check, was not a requirement for getting or possessing a a scientific name then, and cannot be said to be one for possessing a name now - if I can find any scientifically valid name with no corresponding creature in a jar.

[raglanroad]

Scientists unfamiliar with the fish they study, do buy the fish from the market and do scientific studies
Is it not also quite common that fishermen do not accurately reveal the true location of their fishing ?

raglanroad, unfortunately in the real world, researchers do the best they can given the time and resources available.

They're wonderful people.

This might mean less than perfect locality data or sample size

"Doing the best they can", sounds like there is some room to wiggle ! I like it when the they say "Just do the best you can".

but at least the species is named and recognised

The Sea Serpent hunters will be pleased to hear this.

If scientists waited until their data were absolutely perfect, nothing would ever get published or achieved.

exactly

With so much undescribed diversity and so few researchers doing this work, they need to just be able to get on with it.

bottle or no bottle, locale or no locale.

Going back to the discus, Ready et al. did not describe a new species, they merely applied an available name (Symphysodon tarzoo)

The green western red spotted Discus was a species that just had never been named ? It was a recognized species before this ?
i think it's easier to deal with, if we say a green red-spotted fish had been described, but it had not been scientifically described as a species. Sub species, perhaps, but I don't see whether or not Lyons specifically used the necessary words "form" or "variant", to make it a legit subspecies name, and if it were accepted or rejected and so on..it's apparently not so straightforward as all that ! It would take a bit of untangling to make clear sense.
fishbase in need of reworking says

http://www.fishbase.org/nomenclature/Sp ... ymphysodon

http://www.fishbase.org/summary/Species ... ifasciatus

http://www.fishbase.org/museum/Occurren ... ifasciatus
not much about the western green spotted discus species
perhaps they needed to un-synonymize it after un-sub speciating it after un-synonymizing it

http://www.aqua-aquapress.com/volume_12.html

here it's under "blue discus", but called Symphysodon discus

Symphysodon tarzoo Lyons, 1959
Lyons, E. 1959. "Symphysodon discus tarzoo. New Blue discus electrify aquarium world". Tropicals Magazine; 4; pp. Cover; 6–8;10.
Sinônimos: Symphysodon discus tarzoo Lyons, 1959
Origem: Colômbia
Brasil
Peru

but the naming shuffle is not so important as geting a name without a bottled critter. Can it be done under any circumstance at all ? Binomial but no bottle ? If they locate Nessie, the bastards must kill her ? Can't they get away with just a plug of flesh, 300 eyewitnesses including 45 zoologists, who measured her, and 10 hours of clear frolicking video ?

[racoll]

They're wonderful people.

"Doing the best they can", sounds like there is some room to wiggle ! I like it when the they say "Just do the best you can".

The Sea Serpent hunters will be pleased to hear this.

bottle or no bottle, locale or no locale.

You make it sound like ichthyologists are a bunch of incompetents . It is simply not realistic to describe a species based on 500 specimens collected from surveys of the whole geographical range. The overwhelming majority (if not all) of modern studies in peer-reviewed journals are of excellent quality and more than adequately test the hypotheses proposed. It is unfortunate that Heok Hee (a taxonomist) no longer contributes in this forum, as I am sure he would have something to say on the matter. Sure, historically there were some less than ideal practices, but we learn from that and correct mistakes in the future.

that's the rationale behind keeping a jar..but is it the code ?

You can find the online copy of the ICZN code here.

uhh, if the name was never used as binomial - rejected- would it be available in the circumstances you describe ?

The name was deemed to be available by Ready et al., as they found Lyons' description to fulfill the requirements of the code. The disagreements with Bleher seem to revolve around this, as well as the various publication dates and obscure papers/books.

I am not in a position to state who is correct here, as I don't have the time to do the research.

If they locate Nessie, the bastards must kill her ?

I'm sure there are exceptions for very rare and endangered creatures. It will be in the Code. Have a read.