Talk:Cladogram

Not a cladogram!

The first figure of the article, the unrooted myosin supergene family tree, is NOT a cladogram. It's distance-matrix tree! (check the abstract of the original paper). I'm removing it.

I suggest replacing it with the second figure of this paper, the famous 1999 dinosaur phylogeny of Paul Sereno, which is nicely illustrated. I think it should qualify as fair use. --Earrnz (talk) 03:03, 7 January 2010 (UTC)

simple cladistics diagram

Isn't there a bug on the second diagram? Traits B,C are shown on the line pointing to species 3 which only has trait A, it should be on the line pointing to species 1 surely? --86.179.186.239 (talk) 12:44, 23 March 2010 (UTC)

You're right, anonymous, thanks for pointing it out. I will correct it ASAP.--Earrnz (talk) 03:32, 18 April 2010 (UTC)

How to read a cladistics diagram

The article contains a section on what a cladistics diagram is, and a section on how to generate one. What's missing is an intermediate section on how to read one.Christopher King (talk) 14:58, 28 October 2010 (UTC)

This article incorrectly defines a cladogram. There is no consensus about the exact meaning of the term, and the second sentence contradicts itself. The nodes are the ancestors and quite clearly indicate the relationships between the terminals and the nodes, and even other nodes.See:

http://www.ucmp.berkeley.edu/clad/clad3.htm

http://www.bu.edu/gk12/eric/cladogram.pdf

http://www.brooklyn.cuny.edu/bc/ahp/CLAS/CLAS.Clad.html

There are a bunch of others, but I think this proves the point. 66.168.116.107 (talk) 16:50, 15 June 2014 (UTC)

You are right that the term is used to cover two kinds of diagram. One, the most common, is when the terminals are known taxa, extant or extinct, and the branches show the hypothesised evolutionary relationships between them. The nodes can then be interpreted as hypothetical ancestors. Note that in the Hennigian approach, if a node is given the characters inferred from a cladogram and the construction process repeated, it will become a terminal. That's what I mean by the nodes being "hypothetical" ancestors. This appears to be the usage intended by the last two links you put above (the first doesn't seem to work). The second kind of diagram is when the nodes are also known taxa and some method other than cladistics is used to construct the diagram – such diagrams are sometimes called phylograms. Unfortunately many sources are not clear on these differences.

I don't think the second sentence is as clear as it could be, but it's not contradictory: a cladogram in the original Hennigian sense is not a full evolutionary tree: the arc lengths have no meaning, and the nodes are not known ancestral taxa. All it shows is the inferred branching order. Some thought is needed as to how the second sentence can be improved. Peter coxhead (talk) 00:27, 16 June 2014 (UTC)

Dear Author, I copied three sentences off of your article on Cladograms on Wikipedia. For that I deeply I apologize.I had used it for a question on one of my school tests because it had asked what a Cladogram was and I had no idea what it was so I had looked it up and I had copied it.I am really sorry. And I realize how bad what is I did. I will never do it again.I apologize for what I have done. And I feel really bad because of my actions.I aknowledge that I have plagarized and I am so very sorry. — Preceding unsigned comment added by 74.78.86.101 (talk) 00:56, 3 December 2014 (UTC)

What is and is not a cladogram

The following section was recently added; I have removed it pending further discussion.

What is a cladogram and what is not a cladogram?

As noted above, a cladogram is the diagrammatic result of a parsimony analysis, which groups taxa on the basis of synapomorphies alone. There are many other phylogenetic algorithms that treat data somewhat differently, and result in phylogenetic trees that look like cladograms but are not cladograms. For example, phenetic algorithms, such as UPGMA and Neighbor-Joining, group by overall similarity, and treat both synapomorphies and symplesiomorphies as evidence of grouping, The resulting diagrams are phenograms, not cladograms, Similarly, the results of model-based methods (Maximum Likelihood or Bayesian approaches) that take into account both branching order and "branch length," count both synapomorphies and autapomorphies as evidence for or against grouping, The diagrams resulting from those sorts of analysis are not cladograms, either.

Without sourcing, this is just a matter of opinion as to how terms should be used. Some authors restrict the term "cladogram" to phylogenetic trees produced in strict accordance with Hennigian principles, but the term is, in my experience of the literature, used much more loosely now. By all means discuss alternative usages of the term "cladogram" but the material must be attributed to a reliable source.

As an example of "loose" usage, The Kew Plant Glossary, published by the Royal Botanic Gardens, Kew, no less, gives the definition:

cladogram, in cladistics, 2-dimensional tree diagram showing relationships between taxa that are based on shared character states

This applies equally well to counting symplesiomorphies as synapomorphies, and says nothing about how the tree was constructed. Peter coxhead (talk) 18:40, 30 July 2015 (UTC)

Note that the Kew definition starts with "in cladistics". That would seem to restrict the domain of that definition of the word to the cladistic method, which (one hopes, at least) implies grouping by synapomorphy. One sloppy definition is not an excuse for another. It seems to me that the point of an encyclopedia is to provide precise definitions of terms - alternative ones, if necessary. For example, what is a "Yankee"? An American? A northerner? A baseball player? Vagueness leads to confusion or deliberate obfuscation.

If you want sources, here are some books that are clear on the definition: Nelson and Platnick (1981); Wiley (1981); Schuh (2000). Abrower (talk) 12:42, 5 August 2015 (UTC)

Please accept that I understand your concern, but Wikipedia works by presenting summaries based on reliable sources. If these sources use different definitions, we cannot choose between them and say that one is correct and the others wrong; we must report all those we find in reliable sources. This is the essence of WP:NPOV.

On the strict definition you favour, there cannot be a "molecular cladogram" (using the standard modern methods of obtaining one), and certainly not a "maximum likelihood cladogram" or a "Bayesian cladogram". Yet if you search using Google or Google Scholar, you will find hundreds of papers using each of these phrases. What "cladogram" meant to Hennig is not what it means to most biologists today. Peter coxhead (talk) 14:40, 5 August 2015 (UTC)

I understand the "no controversy" mission (although I think it is a vain aspiration to produce a document that has no voice - citing someone else's opinion to support a proposition does not make it any less controversial). You are wrong about molecular cladograms - I publish parsimony analyses of DNA on a fairly regular basis, and have been doing so for 20 years. However, I agree with you that there cannot be a ML or Bayesian "cladogram." The graphical representations of those sorts of analyses are dendrograms, phylograms or chronograms, but not cladograms (which was the point of my original post). The fact that the language of phylogenetics has degenerated into a sort of pidgin because people are not well-trained in systematics any more is unfortunate, but in my opinion that is not an excuse to let it deteriorate into babble. I guess it depends on whether you are a person who thinks that the meanings of words are important - which I would suppose somebody who spends his time editing an encyclopedia would - or a sort of Chomskyite who thinks that words cannot/should not be defined other than by popular consensus. Some people think that whales are fish. Does that mean that Wikipedia should devote space to that point of view? Or maybe if it is controversial, Wikipedia should not address the phylogenetic relationships of whales at all? Like the French say, not all fizzy white wine is Champagne! Abrower (talk) 12:12, 6 August 2015 (UTC)

Interesting; I've never seen a strict Hennigian analysis of DNA (can you supply a reference?). However, you're still missing the point. WP:5P2 applies here, so Wikipedia is committed to the view that terms are defined by usage in reliable sources. "Whales are fish" is irrelevant; reliable sources don't support this view.

A more relevant example is the use of the term "monophyly" and its relationship with "holophyly". As the Monophyly article says, these terms are now used by almost all sources as synonyms, although originally "monophyly" had a broader meaning. Speaking for myself, I regret the loss of precision involved, and the consequent denigration of "paraphyly" as distinct from "monophyly", but my regret is irrelevant to Wikipedia. Since reliable sources do describe "tree diagrams" derived from non-Hennigian methods as "cladograms", the meaning of "cladogram" has expanded. The article needs to say this, setting out both the original narrower meaning and the modern broader meaning.

WP:NOTTEXTBOOK applies; "[t]he purpose of Wikipedia is to present facts, not to teach". When I lectured, I didn't hesitate to select, with explanations of course, what I thought to be correct definitions and insist on them in my examinations. But this isn't what Wikipedia does. It summarizes what reliable sources say, and if they say things you or I consider to be wrong, tough. Peter coxhead (talk) 06:33, 7 August 2015 (UTC)

References: for my own papers, see https://www.researchgate.net/profile/Andrew_Brower/publications I would not exactly call these "strict Hennigian analyses," since Hennig did not use a computer and polarized his character states ahead of time, but they are "parsimony" in the sense of Fitch or Farris. The most recent cladistic analysis of DNA (and morphology) on that list is Garzón-Orduña et al. 2014. I have another one submitted now. However, just because I use parsimony does not mean much. Perhaps more pertinent is the Supplementary Table from Rindal and Brower 2010, which lists over 500 papers that include parsimony analyses of DNA data, that we examined in Mol. Phylogenet. Evol. from 2005-2009. This is 2015, but MANY people still do multiple analyses of their data that includes parsimony. I agree (and regret) that parsimony is no longer the most popular method, but I would argue that this is largely due to a lack of understanding of the ideas of statistical consistency and long branch attraction. Very briefly, model-based methods are only guaranteed to be consistent if the data actually evolved according to the parameters of the model. Of course, there is no way to know whether they did or not - most likely, not. For discussion of that, see my book, Schuh and Brower, 2009 (on the ResearchGate list).

I guess what I object to on this and associate pages is the cooptation of cladistic terminology and cladistic principles to describe a philosophical approach and set of methodologies that are antithetical to Hennig's ideas. I do not consider those things to be "facts."Abrower (talk) 23:08, 8 August 2015 (UTC)

Thanks for the references. I can only say that in writing about plants and their phylogeny I've never knowingly come across DNA analyses using parsimony.

On your last point, it is a fact that molecular phylogenetic studies are reported in the literature using cladistic terminology in a way inconsistent with the original usage of Hennig. Whether you like it or not, that's factual enough for Wikipedia.

To be constructive, what this article should do is to explain, with sources:

• the original Hennigian meaning of "cladogram"
• current usage, and how and why this differs.

The Cladistics article needs material on what I'll call the "Farris vs. Felsenstein et al. debate", which I'm sure you're familiar with, which could then be briefly cross-referenced here. Peter coxhead (talk) 07:09, 9 August 2015 (UTC)

Thank you for suggesting a reasonable way to "teach the controversy." I will see what I can do to modify my original section to reflect the evolution of the meaning of the term. I will also take a look at the Farris/Felsenstein issue and see if I can contribute anything there.

I took a look at the "early view" page of Cladistics, and found the following papers that did MP (as well as other methods) on plants:

Phylogeny and historical biogeography of the cocosoid palms (Arecaceae, Arecoideae, Cocoseae) inferred from sequences of six WRKY gene family loci. Alan W. Meerow, Larry Noblick, Dayana E. Salas-Leiva, Vanessa Sanchez, Javier Francisco-Ortega, Brett Jestrow and Kyoko Nakamura. Article first published online: 18 OCT 2014 | DOI: 10.1111/cla.12100

Phylogeny of Impatiens (Balsaminaceae): integrating molecular and morphological evidence into a new classification. Sheng-Xiang Yu, Steven B. Janssens, Xiang-Yun Zhu, Magnus Lidén, Tian-Gang Gao and Wei Wang. Article first published online: 11 MAR 2015 | DOI: 10.1111/cla.12119

Phylogeny of the Alismatales (Monocotyledons) and the relationship of Acorus (Acorales?). Gitte Petersen, Ole Seberg, Argelia Cuenca, Dennis W. Stevenson, Marcela Thadeo, Jerrold I. Davis, Sean Graham and T. Gregory Ross. Article first published online: 16 APR 2015 | DOI: 10.1111/cla.12120 Abrower (talk) 12:57, 10 August 2015 (UTC)

proposed addition on history of the term "cladogram"

I have prepared the following section which I would like to add to the "cladogram" page Abrower (talk) 16:31, 26 August 2015 (UTC)

History of the word and the concept

Willi Hennig (1966, p. 196) said, "Our contention is generally true , that "family tree" and "written fixation of the system" correspond exactly only when the family tree clearly shows recognized or presumed sister-group relationships, and also makes clear which groups are undoubtedly monophyletic, and which are doubtfully so ... . By no means all of the family trees scattered through the literature satisfy these criteria." Although Hennig's book was full of cladograms, Hennig did not describe them as such. Rather, it was Ernst Mayr (1965), and pheneticists Joseph Camin and Robert Sokal (1965) who independently proposed the term. Camin and Sokal said, "We suggest the term cladogram to distinguish a cladistic dendrogram from a phenetic one, which might be called a phenogram." Mayr (1965) did not provide a concise definition of the term, but in the glossary of his textbook (1969), Mayr defined a cladogram as "a dendrogram based on the principles of cladism; a strictly genealogical dendrogram in which rates of evolutionary divergence are ignored."

Three important post-Hennigian books about cladistics were published in 1980/81. Eldredge and Cracraft (1980) did not provide a succinct definition of cladogram, but offered the following considerations: "Cladograms, as defined and discussed here, are specific kinds of hypotheses about the history of life. They are hypotheses about pattern" (p. 20); "Cladograms depict nested sets of synapomorphies, thereby defining monophyletic groups and simultaneously presenting a hypothesis of relationships among the taxa" (p. 113); "... cladograms are hypotheses about the structure of that history, that is, not specifically about the history itself, but about the structure of the relationships of the organisms as expressed in their patterns of shared evolutionary novelties." (p. 212); and (notably), " ... cladograms, in themselves, are not phylogenies, but rather hypotheses about the pattern of nested evolutionary novelties." (p. 21). Wiley (1981, p. 97) defined cladogram as "a branching diagram of entities where the branching is based on the inferred historical connections between the entities as evidenced by synapomorphies", and also distinguished cladograms from phylogenetic trees. Nelson and Platnick (1981, p. 14) succinctly said, "phyletic trees depict aspects of evolutionary genealogies" while "cladograms depict structural elements of knowledge."

In more recent systematics textbooks, the following definitions of cladogram are found:

• "branching diagram[s] depicting the genealogical relationships between species or other suitable terminal taxa." (Minelli, 1993, p. 10) (Minelli also noted that "The differences between one tree-like diagram and another are not simply matters of aesthetics. Several conceptually different tree-like diagrams are possible and one
• "a dendrogram (tree diagram) specifically depicting a phylogenetic hypothesis and therefore based on synapomorphies. A cladogram generally only indicates the branching pattern of the evolutionary history" (Quicke, 1993, p. 263)
• "a branching diagram specifying hierarchical relationships among taxa based on homologies (synapomorphies). A cladogram includes no connotation of ancestry and has no implied time axis." (Kitching et al., 1998, p. 202)
• "an evolutionary tree that has no information on branch lengths" (p. 24) and "simply shows relative recency of common ancestry" (p. 10) (Page and Holmes, 1998)
• "a depiction of hierarchic relationships among taxa in the form of a treelike diagram, which shows relative recency of relationship and on which character-state transformations may be mapped, but without the connotation of the amount of difference or time since divergence." (Schuh and Brower, 2009, p. 259)
• "a parsimony tree where the weight of the edges is not relevant" (Wiley and Lieberman, 2011, p. 104)

The salient features of all of these definitions are that cladograms are dendrograms depicting relative recency of common ancestry, based on the presence of shared, derived characters (synapomorphies), and which do not take into account degree of difference or branch length.

It is notable that three widely-cited sources of model-based phylogenetics, Swoffored et al. (1996), Nei and Kumar (2000) and Felsenstein (2004) do not define and barely mention cladograms at all.

References

In terms of the material, I think it is a very useful addition. Platnick (1977) is I think useful as a source for a clear early distinction between "cladogram" and "phylogenetic tree". Although it's not well known, I find Podani (2010) gives a very clear account, especially in the Appendix.

In terms of style, it's a bit "academic" for Wikipedia, and not sufficiently "tertiary" or "encyclopedic", e.g. using many direct quotes rather than summarizing. You also need to be careful about text like "three of the most important foundational sources" – this will quickly get tagged "says who" ("three of the most important foundational sources"). We're not supposed to make editorial judgements here. Peter coxhead (talk) 20:38, 26 August 2015 (UTC)

References

All of the proposed references above other then Camin and Sokal and Mayr (1965) are textbooks or parts thereof (e. g. Swofford et al.). I am happy to include those you cited (I need to read Podani's).

I have changed "of the most important foundational" to "widely-cited".

Actually, I wasn't saying that you should include the two sources I gave, merely that I find them interesting as part of the history of the term and the controversies over cladistics/cladism, so you might too. If anything, I would remove some of the sources, replacing quotations by summaries, as I noted above.

Yes, "most cited" is definitely better. Peter coxhead (talk) 07:49, 27 August 2015 (UTC)

I have decided, in the end, that it is probably more suitable to put this in a journal. Thank you for your input. Abrower (talk) 21:29, 4 September 2015 (UTC)

Not an "evolutionary tree" - contradiction

The introduction claims that

"A cladogram is not, however, an evolutionary tree because it does not show how ancestors are related to descendants, nor does it show how much they have changed; many evolutionary trees can be inferred from a single cladogram."

citing 5 respectable sources. However, as far as the text of these sources is openly available, I don't find explicit statements of cladograms not being evolutionary trees. Moreover, evolutionary tree is a redirect to phylogenetic tree, treating these two terms as synonyms. This article listing cladograms as a special phylogenetic tree type: "A cladogram is a phylogenetic tree formed using cladistic methods." So we have a direct contradiction between these two articles, and I'm definitely sympathetic with the latter interpretation. Of course a cladogram is a special type of phylogenetic tree, what else? Otherwise I ask for clarification of the "not"-claim here and explicit quotes from the sources. --KnightMove (talk) 11:51, 4 November 2017 (UTC)

The statement is correct, but whether it's directly sourceable is another matter, because the literature uses terms like "cladogram", "phylogram", "phylogenetic tree", "evolutionary tree", etc. imprecisely and inconsistently.

(To see that it's correct, consider the following hypothetical example. Extinct species A (known from fossils) gave rise to extant species B and C. An evolutionary tree would have nodes for A, B and C, with A as an interior node joined to the terminal nodes B and C. Now suppose that a cladogram is constructed, whether by parsimony or statistical approaches, based on characters of A, B and C. It will have A, B and C as terminal nodes. The interior nodes in a cladogram are always strictly hypothetical, so never tell you what actual species were involved. The ancestor of B and C could be identical to A or different.) Peter coxhead (talk) 22:01, 4 November 2017 (UTC)

I see your point. Still, the prototype phylogenetic tree in this very article has this same approach as a cladogram, even though not being a cladogram in the strict sense. If there are broader, stricter, varying, inconsistent usages of the terms, Wikipedia is supposed to describe them - not to make selections, and that inconsistently among different articles. --KnightMove (talk) 19:54, 6 November 2017 (UTC)

I agree with you. Part of the problem comes from splitting up related topics into separate articles; one article covering "cladogram", "phylogram", "phylogenetic tree", "evolutionary tree", etc. could be made consistent and also explain the different usages. Multiple articles are harder to maintain and keep consistent. Feel free to work on them! Peter coxhead (talk) 07:35, 7 November 2017 (UTC)

"Hypothetical" ancestors

@Peter coxhead:According to the article, "These branching off points represent a hypothetical ancestor (not an actual entity) which can be inferred to exhibit the traits shared among the terminal taxa above it." You seem to agree concerning "hypothetical", writing "The interior nodes in a cladogram are always strictly hypothetical ..."

What is it for something to be hypothetical? Per Wiktionary, it's to be "based on a hypothesis" — on a view which may be true or false. The phlogiston theory was a hypothesis that turned out to be false; rather than being a hypothetical substance, phologiston simply does not exist. Neutron stars, on the other hand, were hypothetical once their existence was hypothesized in 1933 but ceased to be hypothetical in the '60s with the discovery of some such stars. The graviton is currently a hypothetical particle, and we can hope that its existence will turn out to be established or refuted as physics advances. As an open question, the endeavor to find gravitons or demonstrate their nonexistence provides a degree of guidance to physicists who attempt to discover whether or not gravitons are actual; once this is done, they will no longer be hypothetical. How can we say that an ancestor or an interior node is not an actual entity while still calling it hypothetical? If we can say that it's not actual, the associated hypothesis has been refuted and it's no longer hypothetical.

Up to this point, I've thought of the interior nodes as genuinely hypothetical, representing ancestors of which we might someday find fossils, and of paleontology as the science which stands a chance of making such finds. If "branching off points" in cladograms are not genuinely hypothetical, if they don't correspond to hypotheses that might be verified or refuted as paleontology progresses, what's the point of them? How can they provide paleontologists with guidance?

Peter Brown (talk) 22:45, 8 November 2020 (UTC)

@Peter M. Brown: well, I think this is yet another illustration of the difference between the way terms are strictly defined by those who care about these things and how they are actually used by biologists (who unlike, say, physicists, don't seem over-concerned by the precise use of terms). I can only repeat and amplify my earlier comments.

If you go back to Hennig and the original definitions and methods, cladograms were constructed by hand, based on deciding which characters were synapomorphies and which were not (and so not useful in constructing a cladogram). Clades were then defined by synapomorphies. If you had fossil evidence, you could manually put the fossil into the cladogram as an interior node. (Some authors tried to use "phylogram" rather than "cladogram" for such diagrams, but nowadays "phylogram" seems to be used for a cladogram in which the branch lengths are significant.)

The computer programs now used to create cladograms don't allow a unit to be placed at an interior node. Our {{clade}} template doesn't provide any way of truly labelling interior nodes, only the branches, because you won't find cladograms in modern literature that need to be represented with labelled interior nodes, only labelled branches, where the labels refer to the clade represented by that branch, not the ancestor node of the branch. (Thus the cladograms you see now are basically like File:Hominidae_chart.svg.) I have Kenrick & Crane (1997), The Origin and Early Diversification of Land Plants : A Cladistic Study open in front of me as I write this. It's full of cladograms based on the morphological characters of fossils. All the fossils appear as terminal nodes, and when the interior nodes are labelled, it's with things like "Embryophytes" or "Eutracheophytes".

So although you can, and people do, say that the interior nodes represent hypothetical ancestors, this is slightly odd given that, as I noted before, if you used the information in the cladogram to construct the character state of a hypothetical ancestor at a particular interior node, which you can do, added it to the program, and ran it again, that hypothetical ancestor would appear at a terminal node. In reality, in a modern computer-generated cladogram, the interior nodes are merely a way of showing the branching. If you want to put fossils into a tree as interior nodes, as far as I know, you have to do it by hand. I'm not aware of any program that does it.

Does this prevent hypothesis testing as you suggested? No. Consider this tree:

	Fossil 1 Extant species 1 Extant species 2

What this shows is that Fossil 1 was on the earliest branch of the lineage that led to the two extant species. It's "ancestral" in that sense, just as, say, Amborella, as the earliest diverging lineage in the extant angiosperms, is assumed to tell us something about the 'true' ancestor of angiosperms. All you can talk about with modern computer-generated cladograms is the order of divergence, the order of branching (plus the character states that change along the branch, although this isn't very meaningful with molecular characters). Peter coxhead (talk) 11:32, 9 November 2020 (UTC)

The lead section needs a major rewrite

As regards evolutionary history I am one of the much-targeted "general readers". In particular, I rely on the article lead to be coherent and to give me an idea as to to whether to read further or whether the matter is out of my depth. The lead of the article NAD(P)H dehydrogenase (quinone 1) leaves me with no doubt — it contains a host of technical terms whose meanings I don't understand, so I do not read further. The Cladogram lead, on the other hand seems to promise that, reading the article, I might actually understand what a cladogram is and what it can be used for. The promise is not fulfilled.

The first sentence of the lead tells me that cladograms are used in cladistics. While I have limited patience following link trails, I follow that one and read: "The outcome of a cladistic analysis is a cladogram – a tree-shaped diagram (dendrogram) that is interpreted to represent the best hypothesis of phylogenetic relationships." But the second sentence of the Cladogram lead says "A cladogram ... does not show how ancestors are related to descendants". Ancestry is a phylogenetic relationship, though; doesn't it need to be incorporated in the "best hypothesis of phylogenetic relationships"?. Anyhow, how an ancestor is related to each descendant is pretty trivial: it's related as its ancestor.

Further down, after introducing "branching off points", the lead says

These branching off points represent a hypothetical ancestor (not an actual entity) which can be inferred to exhibit the traits shared among the terminal taxa above it. This hypothetical ancestor might then provide clues about the order of evolution of various features, adaptation, and other evolutionary narratives about ancestors.

I have discussed this above, and I thank Peter coxhead for taking the time to respond at length. However, I do think I know what "hypothetical" means and, using the term as I understand it, one cannot say both that something is not actual and that it is hypothetical, i.e. that there is an unrefuted hypothesis according to which it exists. Further, though the evidence for its existence might nevertheless provide "clues" about this or that, the nonexistent entity itself cannot do so. I have to conclude that "hypothetical" is used in an obscure sense. For the sake of the general reader, this term should either be explained or abandoned.

(Admittedly, there is a Wikipedia article Hypothetical Axis victory in World War II, but the first sentence of the lead makes very clear that this is a sub-genre of fiction. Perhaps one could find a better title.)

The remainder of the Cladogram lead mentions RNA and DNA sequencing; I do know what those are. It also mentions computational phylogenetics; according to that article, its goal is "to assemble a phylogenetic tree representing a hypothesis about the evolutionary ancestry of a set of genes, species, or other taxa". As ancestry is not indicated in cladgrams, they are, at best, peripheral to the purposes of computational phylogenetics. Phylogenetic tree § Cladogram notes that a cladogram's "internal nodes do not represent ancestors" even nonexistent ones.

I still don't know what a cladogram is.

Peter Brown (talk) 22:43, 11 November 2020 (UTC)

The key problem is, I think, that the meaning of terms used in cladistics (including the term 'cladistics' itself) has changed over time. So if you merge material from different sources, particularly sources of different dates, the result is inconsistency. There are thorough discussions of the philosophy of cladistics as a science in the 1980s and 1990s – by strongly opposing camps at times. Since about 2000 most biologists just seem to use the computer programs available and not worry about what they do or what the definitions are. (Podani in 2010 is the last in-depth discussion I can find, but doesn't seem to have had any real influence.) In reality, it seems to me, a cladogram is now what is produced by a computer program whose purpose is to generate them. The older programs that use parsimony can be related to the manual methods that Hennig first put forward. Modern ones, especially those that use Bayesian methods, are harder to explain and quite remote from Hennig; certainly they do not rely in any straightforward way on finding synapomorphies.

I think the definition here is accurate: "a branching diagram depicting the successive points of species divergence from common ancestral lines without regard to the degree of deviation", except that it doesn't have to be "species" – "population" would be better, so I would say something like "in biology, a cladogram is a branching diagram showing the successive points at which populations diverge from common ancestral lines without regard to the degree of divergence". Representing time, i.e. the degree of deviation, makes the cladogram a phylogram. But, of course, the detail in my definition isn't sourced.

Another approach, which can also be sourced, is to start with "clade". Define a clade as "a monophyletic group of organisms" and then define a cladogram as a branching diagram that shows clades. But note that clade definitions can be apomorphy-based (the original Hennig method), node-based, stem-based, etc.

If anyone can find a good reliably sourced and up-to-date discussion it would be very, very useful! Peter coxhead (talk) 10:05, 12 November 2020 (UTC)

Podani 2010 is the latest in-depth discussion? Podani 2013, cited by the article, has a lot of depth, more than I am capable of plumbing. I guessed, above, that this article tries to refute the thesis that many phylogenetic trees are compatible with any given cladogram — is that correct?

Whoops, it's this paper that I meant. Peter coxhead (talk) 18:14, 12 November 2020 (UTC)

(The lead actually says that many evolutionary trees can be inferred from a single cladogram, but that can't be right, since an evolutionary tree contains information about descent that is not present in a cladogram.)

"Inferred" is correct, I think, but perhaps misleading; what is meant, I believe, is that many evolutionary trees are consistent with a single cladogram, in that the branch lengths are arbitrary in a cladogram, so given a cladogram you could draw many phylograms/evolutionary trees. Peter coxhead (talk) 18:14, 12 November 2020 (UTC)

Both your definition and the one you obtained from Random House mention "points" of divergence. What are points, in this usage? Do you mean approximate dates? It should be obvious that I won't accept that they are hypothetical ancestors, not without clarification of "hypothetical".

Peter Brown (talk) 17:27, 12 November 2020 (UTC)

Ok, here's another approach. Cladograms without actual interior nodes can be represented in bracket notation. Consider 3 terminals, A, B and C. Using bracket notation, there are three possible arrangements: (A, (B, C)), ((A, B), C) and ((A, C), B). These show, respectively, that A diverged before B and C, that C diverged before A and B, and that B diverged before A and C. Exactly the same information can be shown diagrammatically in a cladogram in which nothing is placed at the interior nodes. The bracketed expressions and the corresponding tree diagrams are precisely equivalent (indeed {{Clade}} accepts bracketed input). The points of divergence (interior nodes in the tree form) aren't dates because although they are ordered along a given branch, there's no ordering across branches. Thus in ((A, (B, C)), (D, (E, F))), we know that A diverged before B and C and that D diverged before E and F, but we don't know, for example, whether B and C diverged before or after D diverged from the lineage leading to E and F. So "points of divergence" are only partially ordered in time.

It's possible to give precise meanings to cladograms treated as mathematical trees by using mathematical terminology. I would be happier with such precision. But there's no evidence that I can see that most biologists agree; they just use cladograms without worrying over much about what they mean. This has allowed them to move apparently seamlessly from Hennigian cladograms to Bayesian cladograms, even though conceptually they are significantly different. Peter coxhead (talk) 18:14, 12 November 2020 (UTC)

That is a very serious objection. The article goes into criteria for judging cladograms and various kinds of data used to build them, but there is not a hint of conceptual differences. There needs to be a subsection on Hennigian cladograms and one on Bayesian ones within a section on conceptual differences. The article says that "many cladogram algorithms use a simulated annealing approach to increase the likelihood that the selected cladogram is the optimal one"; the general reader gets the impression that there is an optimal cladogram for a clade, that the problem is just how to find it and recognize it. If cladograms differ conceptually, the article is promoting a serious misunderstanding. Peter Brown (talk) 04:35, 13 November 2020 (UTC)

@Peter M. Brown: I'm struggling with language here, in particular unpacking "cladogram", and I don't think I'm being as clear as I want to be.

• If you view a "cladogram" as just a tree diagram showing hypothesized evolutionary relationships, then it doesn't matter how you arrived at those hypothesized relationships. A "cladogram" derived by following the original Hennig methods of determining (syn)apomorphies, and a "cladogram" derived by Bayesian statistics are the same.
• If you view a "cladogram" as a tree-based way of showing "clades" (i.e. monophyletic groups), then since there are different ways of defining clades, there are different kinds of "cladograms". A strict follower of parsimony won't accept that similarity or Bayesian methods should be used to define clades.

I don't think any of this bothers most biologists, as I've noted before. They just use computer programs to draw trees, call them "cladograms" (sometimes wrongly by some definitions because the branch length is shown), and discuss them as hypotheses of evolutionary relationships (often not explicitly saying they are hypotheses). Peter coxhead (talk) 07:56, 13 November 2020 (UTC)