Historically, only few of subgenera were described based on phylogeny but mostly on morphological or anatomical characters.

When a genus is too big to handle.

At this level (species within a genus) I would argue that categories like subgenus (or tribus, and there are others) are not very helpful. Not only is Dmitri right, but later molecular cladistics frequently clash with the conclusions drawn from morphology. Even genera turn out to be poly/paraphyletic in some cases. Nomenclatural rules mean that we must put a species in a genus, even if the decision is later reversed. We are not obliged to put species in subgenera, and I would always avoid it if possible.

If you are concerned with  a phylogeographical analysis, or want a time line for evolution within a clade, there is no substitute for a well-resolved cladogram. Within it, you will find many splits at different times or with differing degrees of difference. Any nomenclatural recognition will involve arbitrary cut-offs that will vary among workers

Like in discussing many other questions, we should not confuse phylogeny with taxonomy (classification)! Robert is right in saying that "If you are concerned with  a phylogeographical analysis, or want a time line for evolution within a clade, there is no substitute for a well-resolved cladogram" [boldface my: RBH] - but cladogram informs us only about the "branching pattern" and nothing else [as explained by GHISELIN (1997), “If we opt for paraphyletic grades such as Invertebrata and Pisces, then the system is ... at least noncommittal as to the branching sequences … of no use for anyone who needs that particular kind of information. On the other hand that particular kind of information can be clearly and unequivocally expressed in the form of a tree-like diagram. And since the tree does the job perfectly well, the arguments for a strictly genealogical arrangements are by no means compelling”]! On the other hand, classification should be informative not so much about the history of evolutionary diversification, but also (mainly!) about its results, i.e. the taxa (as units of natural classification, not necessarily as clades!) should be natural: informative in the sense of “the extent to which the (morphological, ecological, physiological, genetical, or any other) characteristics of an organism may be predicted from its placement in the system” - see e.g. .my papers on "Philosophy...", or recent "Fallacies and false premises..."!) - and this cannot be achieved with however well (and reliably!) resolved cladogram but only with adequate taxonomic subdivisions! And so, by the way, subgenera (like subspecies, subtribes, &c.) are useful not (contra Serban) "When a genus is too big to handle", but when the genus is too heterogeneous to be considered an undivisibly natural group.

Roman,

I see your point, but in the context of my field subgeneric categories are not much use. Of course, I may split up groups within a genus when writing an identification key, but I feel little need to give them names. I may be a bit prejudiced as some molluscan taxonomists keep rearranging/splitting genera, never mind subgenera, and the criteria for determining heterogeneity seem to vary even within a fairly narrow field.

Well, another frequent but - in my opinion - erroneous attitude it to look everywhere for rigid, universally accepted criteria. I am not aware of any strict criterion to discriminate between "young" and "old" people, and there may be differences in opinions concerning those born in, say 1970-s or 1980-s, but nobody would classify a school-boy as "old" or a II World-War veteran as "young". What is the strict criterion to distinguish between "bush" and "tree", or between "river" and "stream"? - in typical cases there are no serious doubts, and this is sufficient for the distinction being useful!

By the way, many of those who do not use the category of subgenus, "substitute" it by infinite splitting of genera - what is indeed harmful!

I personally think subgenera are useful in mantaining taxonomic stability, since they do not obligatory appear in the binomen (unlike subspecies, for example). When they are used in a taxonomic compendium, they also help to reproduce phylogenetic information, since two subgenera inside a genus are - in theory - more closely relaed to each other than to other genera.

Thank you everyone for their participation.

Concerning supraspecific groups and their nomenclature, I think we should keep in mind that the rank of all "categories"  of the Linnaean hierarchy is largely arbitrary. What I consider important is that any taxon is proposed, at any level in the classification hierarchy of Linnaean derivation, must be a monophyletic group - at least, until demonstrated otherwise.

Monophyletic - yes; holophyletic - not necessary! In other words, the term "monophyletic" was univocal up to half a century ago, but unfortunately since the publication of Hennig's work it is not, and now it should not be used (at least as long as we wish to clarify rather than confuse the matter of discussion...) without explanation what exactly we have in mind!

I agree with Roman,  the term “holophyletic” is an unnecessary junior synonym of "monophyletic" when it is applied just to  “crown groups”. Not so when you want to include all the elements - even hypothetical - included in a genealogy, starting from the common and exclusive ancestor . As for the misuse of the term “monophyletic” starting from the middle of last century, I invite everyone to (re) read Hennig - not to mention Daniele Rosa! The following image is taken from  the excellent Spanish version edited by Osvaldo Reig for E.U.D.E.B.A. (1968)- It seems to me extremely illuminating

I do not agree with Roman is several aspects!

1. “We should not confuse phylogeny with taxonomy (classification)!” I have always been told (and told my students) that classification should reflect phylogeny!

2. “Taxa (as units of natural classification, not necessarily as clades!) should be natural.” I think the term “taxon” has been introduced by Hennig for a group of species with a common ancestor and thus forming a monophyletic group. Moreover, the term “natural” was used by Remane (1954) to indicate groups sharing a common homology (which may be paraphyletic in Hennig’s system). So I do not really understand what you mean by “natural”.

On the other hand, Robert is right when saying that “there is no substitute for a well-resolved cladogram” and that the category “genus” is the only category to keep because of the nomenclatural rules. However, I think that we can give a name to some clades in the tree even though “any nomenclatural recognition will involve arbitrary cut-offs that will vary among workers”. This then reflects the “result of the evolutionary diversification”, wished by Roman and by most non-taxonomists working in the field. But we must get rid of the Linnean system since “the rank of all categories of the Linnaean hierarchy is largely (I would say completely!) arbitrary” like says Mario.

Another nice handbook explaining the Hennigian system is that of the late Peter Ax (a good friend of Hennig’s): “Das Phylogenetische System” (1984, G. Fischer Verlag). I think it has been translated in English not sure though). 

Wonderful !

The book by Peter Ax has been published in English as "The Phylogenetic system. The Systematization of Organisms on the basis of their Phylogenies" J. Wiley & Sons, 1987.

Cheers to all phylogenetic systematists, orthodox and heterodox!

Dear Harsimran,

I almost forgot your initial question! My advice: do not use the subgenus "level". If you have groups of species in the animals you are studying which are very different from other species, make a "new genus". This is as informative as a new subgenus and more fun. How different the species must be is your decision! (Someone else will "elevate" your subgenus to "genuslevel" anyway;). When you can indicate a good synapomorphy for this group of species, the better. As taxonomists, we must respect the "rules", but we must be pragmatic some time!

Succes,

Ernest

Dear Colleagues, please, if we wish to discuss (and wish our discussion to make any sense...), we should use arguments and reply (with counter-arguments) to the arguments presented by our opponents - simple repeating our standpoints is not constructive! Yes, Ernest, I "have also always been told ... that classification should reflect phylogeny" - the difference between us is that you understand the word "phylogeny" as a synonym of "branching pattern" apparently accepted  while to me branching pattern is only one of the aspects of phylogeny (tempo and direction of differentiation is another, at least as important aspect), and that you have apparently accepted Hennigian dogmas as a kind of "Holy Script" which must be accepted without discussion ("Roma locuta, causa finita") and would like me to accept them on the same basis, while I tried in several publications extensively and in my comment here summarily) to present the arguments showing that classification is a different thing, serves different purposes and - consequently - must not be simple "transliteration" of this or that hypothetical cladogram. Of course, you may consider my argumentation not convincing, but then I would expect to read from your comment not only "I have been told" but rather some explanation why do you think my reasoning wrong, e.g. what may be a function of a classification which provides no informwtion beyond that provided already (much more exactly!) by the respective cladogram?

No and, if we wish to discuss effectively, we must read the opponents' writings... How can I understand your statement "I do not really understand what you mean by “natural”" in reply to my comment in which it is expressis verbis explained: "natural: informative in the sense of “the extent to which the (morphological, ecological, physiological, genetical, or any other) characteristics of an organism may be predicted from its placement in the system”? Of course I cannot, in a necessarily short comment in RG, explain every detail so extensively as I did in my publications [e.g. those on "Philosophy..." (2005), "Taxonomy crisis..." (2008), "Taxonomy in changing world..." (2008), "Is paraphyly indication..." (2010), "Philosophy, evolution, and taxonomy..." (2011), or the recent "Fallacies ..." (2016) - all downloadable from ResearchGate], so if you wish to know my "way of thinking" better (of course, you need not to wish this, but if...) I would suggest to read at least one of them (perhaps the last one, as it is rather short and perhaps most concretely argumented).

To answer your other statements: "I think that we can give a name to some clades in the tree even though “any nomenclatural recognition will involve arbitrary cut-offs that will vary among workers”. This then reflects the “result of the evolutionary diversification”, wished by Roman and by most non-taxonomists working in the field". No, this does not reflect  the results of the evolutionary diversification, it reflects only the branching pattern, saying nothing (beyond what is already "readable" from cladogram) of "the (morphological, ecological, physiological, genetical, or any other) characteristics of the organism" "wished by Roman and by most non-taxonomists working in the field".

"we must get rid of the Linnean system since “the rank of all categories of the Linnaean hierarchy is largely (I would say completely!) arbitrary”" - yes, it is to some extent arbitrary (like any other system!), but it has proved (contrary to e.g. "PhyloCode") excellently efficient in providing the kind of information natural general purpose classification is expected to provide!

"If you have groups of species in the animals you are studying which are very different from other species, make a "new genus". This is as informative as a new subgenus and more fun". I do not know whether it is or is not "more fun", but it certainly leads to infinite, very harmful splitting of generic names, depriving them of their very important function: to serve as "signposts" showing what kind of animal (or plant) is being spoken about - this is just why each species name must include the name of the genus: not nearly a stupid Linnaean convention but a tool to enable orientation among millions of zoological and botanical names. Some decades ago, every person to whom the particular taxon could have been interesting, knew what is Parus, Mus, Lacerta, Helix, or Agrilus  (e.g. in "my" Buprestidae there were ca. 200 valid generic names in use, what anybody could remember and recognize). Now, with the spread of the "split first and think later (if at all)"attitude, I am frequently unable to say if the information found in a publication concerns a buprestid (and, so, is important for my work) or not: in some tribes the splitting has produced several tens of "genera" from what not long ago was one, and it is practically impossible to remember all of their names...

 Sorry, I have inadvertently "sent" the comments above before "proof-reading" it, and therefore it includes some potentially confusing errors - please discard it and read the corrected version below:

Dear Colleagues, please, if we wish to discuss (and wish our discussion to make any sense...), we should use arguments and reply (with counter-arguments) to the arguments presented by our opponents - simple repeating our standpoints is not constructive! Yes, Ernest, I "have also always been told ... that classification should reflect phylogeny" - the difference between us is that you understand the word "phylogeny" as a synonym of "branching pattern" while to me branching pattern is only one of the aspects of phylogeny (tempo and direction of differentiation is another, at least as important aspect), and that you have apparently consider Hennigian dogmas as a kind of "Holy Script" which must be accepted without discussion ("Roma locuta, causa finita") and would like me to take them on the same basis, while I tried (in several publications extensively and in my comment here summarily) to present the arguments showing that classification is a different thing, serves different purposes and - consequently - must not be a simple "transliteration" of this or that hypothetical cladogram. Of course, you may consider my argumentation not convincing, but then I would expect to read from your comment not only "I have been told" but rather some explanation why do you think my reasoning wrong, e.g. what may be a function of a classification which provides no information beyond that provided already (much more exactly - see e.g. the quotation from Ghiselin!) by the respective cladogram?

And, if we wish to discuss effectively, we must read the opponents' writings... How can I understand your statement "I do not really understand what you mean by “natural”" in reply to my comment in which it is expressis verbis explained: "natural: informative in the sense of “the extent to which the (morphological, ecological, physiological, genetical, or any other) characteristics of an organism may be predicted from its placement in the system”? Of course I cannot, in a necessarily short comment in RG, explain every detail so extensively as I did in my publications [e.g. those on "Philosophy..." (2005), "Taxonomy crisis..." (2008), "Taxonomy in changing world..." (2008), "Is paraphyly indication..." (2010), "Philosophy, evolution, and taxonomy..." (2011), or the recent "Fallacies ..." (2016) - all downloadable from ResearchGate], so if you wish to know my "way of thinking" better (of course, you need not to wish this, but if...) I would suggest to read at least one of them (perhaps the last one, as it is rather short and perhaps most concretely argumented).

To answer your other statements: "I think that we can give a name to some clades in the tree even though “any nomenclatural recognition will involve arbitrary cut-offs that will vary among workers”. This then reflects the “result of the evolutionary diversification”, wished by Roman and by most non-taxonomists working in the field". No, this does not reflect  the results of the evolutionary diversification, it reflects only the branching pattern, saying nothing (beyond what is already "readable" from cladogram) of "the (morphological, ecological, physiological, genetical, or any other) characteristics of the organism" "wished by Roman and by most non-taxonomists working in the field".

"we must get rid of the Linnean system since “the rank of all categories of the Linnaean hierarchy is largely (I would say completely!) arbitrary”" - yes, it is to some extent arbitrary (like any other system!), but it has proved (contrary to e.g. "PhyloCode") excellently efficient in providing the kind of information natural general purpose classification is expected to provide!

"If you have groups of species in the animals you are studying which are very different from other species, make a "new genus". This is as informative as a new subgenus and more fun". I do not know whether it is or is not "more fun", but it certainly leads to infinite, very harmful splitting of generic names, depriving them of their very important function: to serve as "signposts" showing what kind of animal (or plant) is being spoken about - this is just why each species name must include the name of the genus: not nearly a stupid Linnaean convention but a tool to enable orientation among millions of zoological and botanical names. Some decades ago, every person to whom the particular taxon could have been interesting, knew what is Parus, Mus, Lacerta, Helix, or Agrilus  (e.g. in "my" Buprestidae there were ca. 200 valid generic names in use, what anybody could remember and recognize). Now, with the spread of the "split first and think later (if at all)" attitude, I am frequently unable to say if the information found in a publication concerns a buprestid (and, so, is important for my work) or not: in some tribes the splitting has produced several tens of "genera" from what not long ago was one, and it is practically impossible to remember all of their names...

Yes, you are right: "we must be pragmatic some time"!

Best wishes!

Roman

Although genera have become very unstable, I agree absolutely that we should not abandon the Linnean system. It may creak (largely due to abuse), but the whole of our accumulated taxonomic knowledge is encoded in it. It is bad enough that my younger students find names hard to follow in old literature. A total discard would render much of it impenetrable to all but the most dedicated.

Dear Roman, when I say that "I have been told ... that classification should reflect phylogeny", that was in 1962, when I was a student, by the late Prof. Lucien De Coninck (famous nematologist) who was certainly NOT an adept of Hennig! On the contrary, he was more an adept of what was then called "evolutionary systematics" (cfr. Mayer and Simpson). Nevertheless, his conviction was that classification should reflect phylogeny. So, when our insights on the possible phylogeny (or evolution) - which will always remain a hypothesis! – the system should be adapted, independently of what method you may have used to come to that conclusion.

About “natural”. You state that a clade does not reflect the result of evolutionary diversification. The definition of a clade is a group of species with a common apomorphy (or synapomorphy) reflection the common ancentry. This apomorphy can be morphological, physiological or whatever kind of character… This, I think, tells you more than “just” the branching pattern! By the way, as far as I can see, phylogeny has always been seen as a branching pattern (what phylogeny actually means!), even in the 19th century. Of course, since today molecules are mostly used as “characters”, it has become a bit more difficult. So it is the task of the taxonomist to look for “visible” characters to define (some of) the clades. But when you do not want to think in the clastistic system – which is your good right – my reasoning is of course complete bull shit ;-). It is not my ambition to convert you to cladistics – nor do I want to be converted.

About the Linnean system. I agree with Robert that “the whole of our accumulated taxonomic knowledge is encoded in” the Linnean system. Our predecessors did a good work indeed! But when the system appears to be wrong (= when new and sound hypotheses on the evolution have been found), it should be corrected. What I mainly meant is that we should get rid of the categories, suggesting some kind of “fixed” hierarchy. A Classis or Order in e.g. Arthropoda is something completely different than in Platyhelminthes. And, by the way, nothing much is left of the “old” classification in this latter taxon!

About the genus. This is the category that should remain, for the good reasons mentioned by Roman and Robert. But once a species is “put into” a genus, it should remain there! I was indeed not clear in my advice to Harsimran. What I meant is, that when he finds a new species that is (very) different from other (closely related) species (how different is his decision), better make a new genus, not a subgenus. Do not split an existing genus. That is “harmful” splitting, where Roman and Robert complain about, and I agree. However, when you have a reliable phylogenetic analysis, different solutions might be considered.

"It is not my ambition to convert you to cladistics – nor do I want to be converted" - well, Ernest, I am not a missionary, so I do not wish to "convert" anybody - I have only entered a discussion in which you already participated, and (this is the aim of any discussion, isn't it?) try to convince you with my arguments and know your arguments which, potentially, might convince me. For the moment, neither I seem successful in having convinced you, nor have you presented any arguments having any potential of convincing me - mainly because none of your arguments addresses what in my opinion makes the main difference between our opinions. One of such is my - always carefully "steered clear of" by cladists... - question: what is the use of a classification which confers no information beyond that "readable" from the cladogram it is based on? Well, you say that the cladogram has been based on synapomorphy (or synapomorphies) which, you "think, tells you more than “just” the branching pattern", but 1) it is not any new information "readable" from the cladistic classification: cladogram is a graphic presentation of hypothetical branching pattern, which (at this particular "node") is based on the "synapomorphy", so that synapomorphy is already "included" in the branching pattern, and "deducible" from it - translation of the diagram (cladogram) into words (names of taxa) offers nothing beyond this; 2). Such "synapomorphies" are almost invariably a single, sometimes two-three, rarely more characters, often "important for phylogenetic analysis but usually trifling from any other point of view” – HOŁYŃSKI 2010)"; 3) even this information is much more exactly conveyed by cladogram than by resulting classification: cladogram "says" "the synapomorphy of the last common ancestor of the clade X was two pairs of legs (a result of transformation of two pairs of fins)", and this is a perfectly true and exact information; after making the clade X a taxon Tetrapoda the derivable information becomes "Tetrapoda have two pairs of legs" which is neither exact (very few "tetrapods" have legs even remotely resembling those of their last common ancestor) nor even factually true (several large subgroups of Tetrapoda (birds, bats, whales, snakes &c.) - altogether the majority of species! - do not have two or (snakes) even one pair of legs [well, cladists use to say that snakes or whales "do have legs, they have only been transformed" - but this is a religious belief or conjuring jugglery with words rather than description of reality...]. This is what I (and not only me) have in mind saying that rigidly cladistic classification is simply superfluous: the respective cladogram is at least as, and usually more, informative!

I do not, of course, know what your professor in 1962 had in mind saying that classification should "reflect" phylogeny: if, as you say, he was a follower of the "evolutionary systematics" of Mayr and Simpson, then he probably interpreted the word "reflect" so as Mayr or Simpson (and now, among many others, I) would have interpreted them: as a synonym of "being compatible with" [classification should be compatible with phylogeny - i.e. taxa should not be polyphyletic]; if, however, he interpreted "reflect" as "be identical to" then he evidently was "an adept of Hennig": the principal difference between Hennigian and Mayrian concepts of systematics was just this!

"as far as I can see, phylogeny has always been seen as a branching pattern (what phylogeny actually means!), even in the 19th century" - "phylogeny" means (and, I think, have always meant) "evolutionary development", one aspect of which is the branching pattern!

"What I mainly meant is that we should get rid of the categories, suggesting some kind of “fixed” hierarchy". Hierarchical categories are needed for the same reason (even if somewhat different form) what makes genera necessary: they serve as "signposts" informing what kind of animal or plant has been spoken about, with which other organisms it has (morphologically, ecologically, genetically and in any other respect) most in common; in other words, they convey most important - even if approximate - information about the organism in question.

Of course, I cannot repeat here what I have already written in several papers, so I can only repeat my suggestion from the previous comment: "if you wish to know my "way of thinking" better (of course, you need not to wish this, but if...) I would suggest to read at least one of them (perhaps the last one, as it is rather short and perhaps most concretely argumented)".

Best wishes!

The discussion pro and contra Hennig is going on since the publication of his book in 1954, so we are not going to come to a conclusion in one or the other direction here... It is interesting to see that some people are still fighting...

some people are still fighting, because their opponents carefully avoid to address the critical points and to answer the "unpleasant" questions... Such "discussion" is, indeed, likely to continue infinitely...

Indeed, there are still arguments, both in the pragmatic sphere of nomenclature and when the theoretical basis if the Hennigian model are concerned. And the issue becomes even harder when, as in my field, different molecular systematists come to radically different views as to what groups are monophyletic.

Ideas about phylogeny remain hypotheses (nobody was there to check) which can, by definition, change. When one accepts that classification should reflect phylogeny, it is inevitable that classification changes also. For some groups the arguments are so numerous and strong, that the hypotheses probably reflect reality (as e.g. in vertebrates to name one group), resulting in a rather stable taxonomy. In many taxa, our hypotheses are not so stable and systematicians search for new arguments. These can be (different!) molecules or other new arguments from thorough studies of morphology, ecology, physiology or whatever characters with whatever new technique that becomes available (e.g. EM many years ago). So Robert, I am afraid that we’ll have to live with a changing classification and with diverging views, whether you are using hennigian cladistics or any other method. That is exactly what makes taxonomy such an exiting science!

On the other hand, if you find that the classification should not necessarily reflect the phylogeny or evolution, then you can make “adequate taxonomic subdivisions”, stable and unchangeable, monophyletic or not. This may be an acceptable approach, but it is not mine.

I do not disagree in principle, and of course I have to change the nomenclature I use when there is a generally accepted revision. But I cannot help thinking that there should be an intermediate stage where a revision (and of course the evidence for it) is in the public domain for comment, but is only incorporated in terms of nomenclature if it survives scrutiny for some period. In my field some changes go back and forth like a rather arduous tennis match in which two equally matched players never reach break point.

And of course, some categories are more stable than others. Going back to the original question, this is why I feel that subgenera are not a particularly useful category, but I would always find a well-resolved cladogram useful.

I agree, Robert. We should not change our taxonomy each time a new cladogram is generated by someone (especially when it is not well-resolved), or his/her arguments must be particularly strong. In practice, new taxa (and names) only become of general use, when accepted by several colleagues in the field. As I said before, not much is left of the old classification of the Platyhelminthes, but most of the old taxa can still be recognized, but very much “reshuffled”. And you are also right to use only the “category” genus (because it is part of the name of a species). Subgenus or any other category is meaningless. In the most recent editions of the famous handbook of Barnes (Invertebrate Zoology) no categories are used and all taxa remain perfectly recognizable.

Well, I partially agree with both Robert and with Ernest. In essence, we must not neglect the nomenclature stability, to avoid the Tower of Babel (which is already there from the linguistic point of view). However, the use of intermediate rank taxa, such as the subgenus and not only it, can often be useful just for the purpose of a classification easier to consult. I insist, however, that any systematic group does not necessarily have to "reflect" the phylogeny, but can not be in conflict with it - at least, until proven otherwise. What gives reality to a systematic group are the exclusive genealogical relationships among its subordinate elements. And this, regardless of whether or not we are able to reconstruct these relationships. Sometimes ago I uploaded on ResearchGate two small,l more or less theoretical essays, written mainly from the perspective of a biogeographer (“I taxa parafiletici e la biogeografia”, 1991; “El concepto de especie y la biogeografía”, 1991). I'm uploading two more papers, on phylogeny and dichotomy.

I am particularly happy with this debate!