No, I think it would be next to impossible to characterize a new species based on COI data. While I'm a molecular "junkie", I realize the importance of well trained taxonomists. Unfortunately, there are fewer and fewer of these.

I'm not certain I understand your question, but... Barcoding is an exceptionally powerful tool when done well. To do it well requires a few things. Firstly, "voucher" samples for the species must be obtained from taxonomists. With a few vouchers in hand, you then proceed to sequence the barcode region. Then you make certain that no know species has the same barcode. If you can meet all these conditions, then Barcoding works. Unfortunately, this can't always be done.

Thanks Mark. I asked  because I found some paper where the author described new species basing only upon COI sequence. Meanwhile no proper and exaustive morphological description was given. Is this Taxonomy?

No, I think it would be next to impossible to characterize a new species based on COI data. While I'm a molecular "junkie", I realize the importance of well trained taxonomists. Unfortunately, there are fewer and fewer of these.

Unfortunately there are some researchers who like to evaluate forms to species status based on just COI, or Cytb, or chromosome numbers without any morphology. And vice verse there are still researchers who describe "new species" based on some morphological traits of doubtful quality without any genetics. A well-balanced taxonomist should use as many factors as possible and value their significance.

This is an example of what I meant. I've been studying Zospeum (even anatomically) for years. Its shell variability is remarkable even in the same population. If the Author had any knowledge about that I think he would never described a n. sp. I don't bring into question his value as a scientist as a whole....only as a "taxonomist". Why not consulting a real taxonomist before publishing?

For Barcoding you need Classical taxonomist (expert in Morphology) to obtain or to confirm the specimens. Molecular part comes later other wise it may not be fruitful as you wanted.

DNA Barcoding is for SPECIMEN identification. Although the process of creating a DNA Barcoding sequence database might introduce doubt about the circumscription of particular species, DNA Barcoding is not a tool for Taxonomy. That would be DNA Taxonomy and its approaches of species delimitation.

I would like to add that another premise of DNA barcoding is the "acceleration"of species discovery. Sometimes complete species description are time-consuming and other times (more often in Fungi for instance) you find cryptic taxa with no observable morphological differences. I believe that DNA barcoding and taxonomy should be more about quickly documenting biodiversity, even if species descriptions are incomplete. Of course, barcodes of voucher specimens are necessary to clarify taxonomic conflict.

 It is not necessary that barcoding evaluate the validity of an organism on species level but this is certain that it can evaluate the validity of genus.

Santiago - I somewhat disagree with your view on accelerated species discovery. There are no formal criteria for species description based on barcoding. Consider the following example - DNA sequencing is performed on 5-6 voucher specimens from a few geographical locations, and a barcode is established. 2 years later, somebody finds a new specimen from a different area, and it doesn't match the barcode. Does this indicate a new species - NO! Here is the reason why; Imagine now that the new specimen had been found 2 years earlier. During the barcoding process, the variations would have been incorporated into the known barcode. It would have been known that species X has a barcode with variations at position Y and Z in the COI, for example.  

Further, I do not believe that there is a rational process for deciding what is a "species" on a barcode level. Where do we draw the line - 5 nucleotides difference, 10, 1? 

In a former job, I worked with unicellular green algae. There are two strains of Chlamydomonas that are completely interfertile, indistinguishable by light microscopy and each strain has a series "sub-strains" with various nutritional requirements. By all reasonable measurements, these two strains should be the same species. At the moment, however, they are considered different species because one strain has an intron in the CytB gene whereas the other does not - Does this really make sense? 

One final note, if we applied "loose" barcoding standards to humans, it is very likely that all races could be classified as distinct species. Nobody would ever consider doing that! So why do we do it with other animals, plants and bacteria? 

Unfortunately, I believe we are falling victim to "sexy sciences". It seems much sexier to say "I've discovered a new species" rather than "Here I describe a new variant of species X". - Just my two cents. 

I think that the term "barcoding" is often misapplied  / used as synonym for sequence analysis which has been around much longer than that term. Looking through the article, the author uses sequence information and resulting p-distance in addition to morphological data as evidence for a new distinct species. I think that this is a valid approach. If the presented morphological or genetic distance is enough for that depends on the used species concept which in the case of DNA Barcoding (animals and plants) is mainly the morphological species concept. My suggestion would be not to be too inclusive (methodological) when using the term DNA Barcoding or even better talk about details instead of "big words". Generalization can be useful, but it does not always help to understand each other :-)

Hi Mark, I understand your point. This is why barcoding is still an open debate. The situation you describe applies also to traditional morphology-based taxonomy, where someone finds a similar but unmatching specimen from a new location. Personally, I'm more in line with an integrative taxonomy view, where you tackle a taxonomic question from multiple biological angles (morphology, ecology, geography, population genetics, phylogenetics). Then again, gathering evidence can be time consuming. Just as traditional species descriptions, newly barcoded species can be viewed as hypothesis that can be tested with further evidence, for instance, with population genetic, phylogeographical, behavioral, ecological, or even novel morphological data. The idea is to put hands on biodiversity discovery, and later on, thoroughly look for more evidence. Barcoded species are definitively not written in stone.

Thanks everybody for the warm debate and the usefull informations provided.

Thomas, I'm glad you browsed the paper I attached. In this particular case, the Author used exclusively p-distance as an evidence for a new species, since his morpho-taxonomical approach is, at least, very weak if not reliable at all. The morphological differential diagnosis with the other species is scant, and totally insufficient...maybe because there are no differences! Who knows  about Zospeum can not deny it. This is a clear example where, given a p-distance, the Author forced himself to find (..at every cost) some morphological feature to justify the description of a new species derived only by this COI analisys. The only morphological-differential feature provided: a "dome-like second whorl" does not mean anything in Zospeum.

I'm asking again: is this Taxonomy?

My opinion is that a barcode only approach can be used for species identification when the reference library is well developed. 

However, when the goal is species delimitation or defining new species using only barcode data applied to a species concept is a very weak approach (in my opinion). I would say that barcode data is not inherently bad at this but that any single point of inference for determining species is weak. There are many examples in nature of morphology being weakly informative or even entirely misleading for species delimitation. A barcode gene is no different. I think the best approach would use multiple independent sources of data (e.g. COI + nuclear gene + genital morphology) to infer a biological pattern consistent with biological differentiation based on some criteria set beforehand in the form of a species concept.

See attached papers for my work in applying barcode data for species identification and determining local species richness.

Hope this contributes to the conversation!

Dear Willy, if we define taxonomy as description, identification and classification of organisms, I would say yes, it is taxonomy what the author applies.

"Species are the smallest groups that are consistently and persistently distinct, and distinguishable by ordinary means" Arthur Cronquist, 1978

If the two morphological characters (dome-­like structured 2nd whorl and columellar fold) are consistently and persistently distinct is another question. Are 8 samples enough to show that ? The genetic p-distance seems to be significant with 11-12 % (this is however the distance to the closest morphological match). 5.6 % as distance to the closest genetic match is above the (current) barcoding gap (3.2 %) of the group and with a "well developed" data backbone one could argue that it makes the group distinguishable by "ordinary" means.

In the end it is up to the editors and referees of the respective journal to decide if the manuscript does reflect current taxonomic standards. Furthermore do you think we can put general restriction on taxonomy without compromising the whole system ?

"...individual characters do not have a fixed, a priori importance; a character is only as important as it proves to be in each instance in marking a group that has been recognized on the basis of all the available information. The value of character is established a posteriori" Arthur Cronquist, 1988

Taxonomy is not a one step but an ongoing process. Like the example of Mark - with new data / evidence the value of character - morphological, genetic, or other - might change. 

Cheers Thomas

On barcoding as a tool to identify species by sequence matches, my answer is yes. The approach has been widely accepted by molecular taxonomists and animal/plant geneticists for species analysis. However, describing a new species just based on barcoding will require more caution as the approach is based on genetic distances at COI, and the distances at this gene region are known to vary widely among some taxa. For the description of a new species, a combination of morphological and barcode data can be a useful approach. Barcode of life datasystem (http://www.boldsystems.org/) is a very useful link to know more about barcoding. 

Well said, Thomas! Personally I have no experience with the genetic approach, but I have some taxonomical experience. I have described some 200 species of free-living flatworms – not really an achievement in this group;-). In taxonomy one can use all kinds of characters and the more the better. A gene sequence can be one of them, but preferably not the only one. It is sometimes difficult to decide if you are dealing with a known species or not. Then I follow the advice of my professor (long time ago): when you have strong doubts you are dealing with a new or a known species, better name it a new species: it is much easier to put it in synonymity later on than to split the species when you appear to be wrong (when new evidence comes in). This is plain practical taxonomy. As Thomas says: taxonomy is an ongoing prosess. The future will tell who was right (supposing we shall ever know what a species is…).

P.S. I did not find the paper Willy attached to his question.

Thanks Ernest :) You can find the mentioned paper here:

http://subtbiol.pensoft.net/articles.php?id=1307

@Mark

" I believe we are falling victim to "sexy sciences". It seems much sexier to say "I've discovered a new species" rather than "Here I describe a new variant of species X". - Just my two cents."

Yes, we call it machismo. =) In my view describing new species is mainly how the zoology (and maybe other disciplines like botany, mycology etc.) was created. So this is something very basic and "ancestral". The desire of describing new species still conquers minds of researchers, especially those who suppose themselves to be taxonomists. They like to say: "I described 10, 200, 100500 species". It's like measuring some organs between men.

Those people want to have their surname after species name. And if they would be not that selfish they would never say: "when you have strong doubts you are dealing with a new or a known species, better name it a new species". What is left after such taxonomists? A headache for others. If one would like to describe a new form, create identification keys, even make list of regional fauna he would have to look on dozens of "species" described previously with little or even no clear evidences. Just because those guys wanted to describe "NEW SPECIES". They rudely transgress the principle of Occam's razor.

OK, here are some of my coins here =)

 I think Dr. Rusin missed the point of the discussion here... The question is: "What is a species and how can we recognize it?" That problem is a bit different for mammalians, or gastropods or meiofauna taxa (such as the majority of free-living flatworms). Describing (new) species is not a goal on itself: these species names will be used in other studies (biodiversity, phylogeny, biogeography, etc.). Wrongly identified species can lead to wrong conclusions in other disciplines. I would also advise Dr. Rusin to have a look on the ICZN, especially Art.23.

@Thomas: thanks for the link.

I think Dr Schockaert is absolutely right! I have experience with both DNA barcoding and Linean taxonomy. I have used morphology, DNA barcoding and nuclear locus sequences to assess the diversity of shipworms (Bivalvia: Teredinidae). I have strong evidence from the molecular data that I have cryptic species in "Lyrodus pedicellatus". However,  I haven´t yet described a new species, I am looking for more evidence to confirm my findings. I think a combined evidence approach is the best practice, as it has been mentioned in the answers above.

Luisa, this is a nice example how new facts may reveal new species! Often, after having used a new approach to distinguish species, one can also find morphological treats to distinguish the species. Most so-called "cosmopolitan species" appear to consist of several different species, even tough it is not always evident to decide "were one species ends and the other begins" - if you see what I mean :-).  These kind of studies may help to understand how speciation occurred.

Indeed Ernest, that´s my goal! I am also very interested  interested in Biogeography thus the correct identification of species is crucial as we all know.

ICZN says nothing about "when you are not sure whether you deal with a known or a new species describe it as a new one". Especially Article 23. In fact this Article is just about what I was talking about. After taxonomists who want to describe as many species as possible one should have a huge work, including identifying priority names between dozens of synonyms.

And finally, the question of the topic starter is not  "What is a species and how can we recognize it?", but whether is it correct to evaluate some form to species rank just using COI.

I thank all of you for the interesting answers and the passion you put into the topic.

All your comments have been of primary importance for helping me to get an impression about the discussed subject.  Definitively I think that the paper I posted as an example is NOT proper Taxonomy but a sort of "SHORTCUT" to validate the description of new taxa...which is very "cool" nowadays.

Taxonomy requires strong willpower, developing appropriate skills that take years of practice and excellent comparison capabilities...saying nothing of all the rest needed.

Thanks you all.

So far working on rats with mitochondrial COI,  Cytb I think molecular taxonomy (barcoding) is a relative and comparative studies of the specimens  with the sequences of known species and with this study we can come to know the identification of the speciemens upto specific level but not to subspecies status. in these whole procedure the initial sequences from the identity of the species was established should be first accurate generally done with help of morphotaxonmy , here the work of taxonomy is crucial. In short  identification of species could be done provided that the sequences that were compared were sure of the status of the species.