Dear Damien,

I hope these article (including reviews) will be helpful for you:

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1692781/pdf/10905614.pdf

http://biology.kenyon.edu/courses/biol112/Biol112WebPage/Syllabus/Topics/Week%207/land%20plants.pdf

http://www.sciencemag.org.sci-hub.org/content/324/5925/353

http://onlinelibrary.wiley.com/doi/10.1111/nph.13247/full

Best regards, Anatoliy

Bryophyte diversity and evolution: Windows into the early evolution of land plants

doi: 10.3732/ajb.1000316

http://www.amjbot.org/content/98/3/352.full

Thanks for all of your suggestions.

If you could not find your appropriate answer than you please locate this book entitled "Cryptogamic Botany, Vol: 2 by GM Smith". 

https://archive.org/details/cryptogamicbotan030182mbp

Hello Damien,

If by an "evolutionary systematics" point of view you mean the recent specifically anti-cladist rhetoric, then I'm afraid your question is non-sensical. The extant bryophyte groups (mosses, liverworts and hornworts) and the other extant land plant group (tracheophytes) almost certainly share a common ancestor that evolved from within a group of algae that existed more than 400 million years ago, and from which the extant streptophyte (not chlorophyte) green algal groups are also derived. It doesn't make sense to ask which *extant* class of algae the ancestor of land plants evolved *from*, because whatever it evolved from doesn't exist any more, and taxa (such as classes) can only represent natural (non-artificial) groups if they are monophyletic.

It remains uncertain which extant streptophyte group (e.g. Charales, Coleochaetales, Zygnematales) is sister to embryophytes (land plants). E.g. see http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2707909/

The same applies when you ask "which class comes from which class among bryophytes". It may be the case that hornworts share a more recent common ancestor with tracheophytes than with the other two bryophyte classes, and that mosses share a more recent ancestor with (hornworts + tracheophytes) than either do with liverworts. Alternatively a range of other scenarios now seems possible, including a monophyletic bryophyte group (see e.g. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3926305/ and http://www.pnas.org/content/111/45/E4859.abstract ). Whatever, it is rather unlikely (although quite possible) that, for example, a common ancestor of mosses and (hornworts + tracheophytes) would have closely resembled either an extant moss or an extant hornwort, and so it is unlikely to make much sense to ask whether (e.g.) mosses evolved from hornworts or hornworts from mosses.

One ironic thing about the so-called "evolutionary systematics" narrative is that it's not impossible to envisage investigating what attributes the common ascestors of the various bryophyte groups might have had, and thus what sorts of organisms the extant groups have evolved "from". But this is an incredibly difficult problem given the time scales involved and the paucity of the fossil evidence, and the first critical step in solving it would be to figure out what the branching order of the major groups is with a high degree of confidence. Ask again in 20 years time... ;-)

Best wishes, Neil.

Damien,

Three papers were published last to infer the relationships among bryophytes based on transcriptome data, chloroplast or mitochondrial genome data. The answers are somewhat incongruent.

Wickett et al., nuclear data-transcriptome:http://www.pnas.org/content/111/45/E4859.short

Cox et al. (cp data): http://sysbio.oxfordjournals.org/content/63/2/272.full?ijkey=997321b9ddbbd4fe2d26d4fa9c60c74473e9b782&keytype2=tf_ipsecsha

Liu Y., C.J. Cox, W. Wang & B. Goffinet. 2014. Mitochondrial phylogenomics of early land plants: Mitigating the effects of saturation, compositional heterogeneity, and codon-usage bias. Systematic Biology 63: 862–878. (doi: 10.1093/sysbio/syu049) (http://sysbio.oxfordjournals.org/content/early/2014/07/28/sysbio.syu049)

Thank you for these insightfull suggestions.

@Neil : you are right, anti-cladism is exactly what I meant, but I don't find this point of view non-sensical. If you cut the whole tree of life into groups (phyla, classes, etc.), then every extant or extinct organism must belong to exactly one group (at each level), ancestors as well. If an ancestor had two daughters, then you could choose to classify the ancestor alone or with one of its daughters, thus naturally creating one (or two) paraphyletic group. Avoiding paraphyly is a logical illusion. For example Eukaryota come *from* Prokaryota, both extant groups, it does not matter that the very species from which Eukaryota come from is now extinct (it does not solve the problem to cut Prokaryota into Archaebacteria and eubacteria, the latter being also paraphyletic to neomurans). I may not be clear enough here, however I have a paper under review about a mathematical analysis of these concepts.

Damien - My opinion on this would be that the logical issues you describe derive from incompatibilities between strictly hierarchical classification systems and the way that evolution works. With some important and notable exceptions evolution produces nested monophyletic groups, but these are non-symmetrical and exist at far more levels than there are Linnean ranks.

Taking your example of an ancestor with two daughters. Let's call them A (ancestor) and X and Y. Another way of dealing with this would be to classify all three in one named group (Aphyta), X in another named group (Xphyta) inside Aphyta, and Y in another group (Yphyta), also within Aphyta . X belongs to Aphyta and to Xphyta, Y to Aphyta and Yphyta, but A only to Aphyta. All groups are monophyletic. Paraphyletic groups only come from the perceived need to classify A within a group at the same rank as Xphyta and Yphyta, which strictly hierarchical systems such as the Linnean one require.

Similarly, in a non-ranked system you would classify Eukaryota *within* Prokaryota, but you wouldn't feel the need to provide a name for the paraphyletic residue of Prokaryota just because it is useful to name Eukaryota. Eukaryotes are Eukaryotes *and* Prokaryotes, Prokaryotes are just Prokaryotes...

Of course this is a potentially confusing example because Eukatyotes are derived from more than one distantly related Prokaryotic group. A better example might be  that birds are dinosaurs as well as birds, while the other dinosaurs are just dinosaurs...

This is how alternative non-ranked classification systems such as the phylocode work. For practical and historical reasons we are stuck with the ranked Linnean system for the foreeable future and need to work with it. Usually it can be bent to fit monophyletic classification more or less (or in the areas that matter), but it's important to keep in mind that it is inherently non-optimal as a naming system for life and that where it doesn't work, that's the fault of the system, not evidence of paraphyletic named groups having any reality in nature...

I agree with this analysis, ranks and cladism are incompatible, and the cladification you describe is non-hierarchichal in the Linnaean sense. But I don't agree that paraphyletic groups have no reality in nature. For example extant species, roughly defined as interbreeding populations, can be paraphyletic (the so-called para-species). It is a natural consequence of peripatric/budding speciation and the change of ecological niche of one part of the original population, but not the other one.

In my view, the holophyletic principle - that every groups must be holophyletic - is much more artificial than the Linnaean categories. It is the fault of a system, yes, but the Hennigian one I would say. The empiricist recognition of adaptive zones for higher taxa, and so their reality, seems to me more natural than to be stuck to this axiomatic rule.

However I don't say that clades are not usefull, but their usefullness is different and complementary to that of taxa, and one should not try to replace the other one. Isn't it a little specious/weird to read everywhere the phrase "non-avian dinosaurs" to just talk about dinosaurs in the classical sense. I see this "non" as an evidence for the need of recognition of paraphyletic taxa in our scientific talk. Taxa should not be renamed to fit clades, this is useless and creates confusing as you noted. For example, one should not use the word Osteichthyes (bony fishes) if he wants to include Tetrapoda, because this creates confusion and the concept of "non-tetrapod osteichthyans" is a usefull one. So, a new name like euteleostomi should be used to speak of the entire clade.

Yes of course, monophyly is a hazy and at times inappropriate concept at the cutting edge of speciation and population dynamics. But if an interbreeding population "buds" off of a "paraspecies", and if full interbreeding is maintained among members of the "paraspecies" but not with members of the "budded" population, then with time the "paraspecies" lineage will become monophyletic and you will have two monophyletic sister groups, even if the phenotype of one remains very similar to the shared ancestor and the other becomes very different. But this is more a question of what the borders of monophyly are rather than whether monophyletic or paraphyletic classification is appropriate at higher taxonomic levels. There is no ambiguity about birds being a group that is derived from within dinosaurs and the remainder of dinosaurs only being recognisable as distinct from them on the basis of characters that they lack. If you "classify" dinosaurs as a group that excludes birds you are saying that the sister group of birds is more closely related to Stegosaurus than it is to birds, and this is simply not the case, either in terms of the actual events in evolutionary history or basic morphology. What's more, you actually lose valuable information that could be encoded in a monophyletic classification, that would classify birds and their sister group together as a more inclusive named group having shared characteristics that are not present in other dinosaurs.

The main reason that the term "non-avian dinosaurs" seems weird is that they are all extinct, and we mentally like to seperate these fabled ancient beasts from the real live ones we see around as all of the time, or at least some of us do.. ;-) If they were still extant I believe that the affinity with birds would be self evident, and that birds would have been classified as dinosaurs to begin with. An ostrich or a cassowary is superficially more like a "dinosaur" in the classical sense than some of those small bird-like dinosaurs are!

Classification is a matter of balance, that's why I prefer carefully delimited taxa without any a priori about topology (holophyly vs paraphyly) than the use of a blind principles. The phrase "more closely related to" must be interpreted to make sense, a "relation" is a very vague term. Crocodiles are more closely related to lizards and turtles than to birds, just because of adaptive zone (basic morphology, ecological niche, etc.). So many symplesiomorphies...

Of course species genealogy is a continuum, so any cut is artificial in a sense. So the sister species of the very first bird species may well be more closely related to birds than to other dinosaurs, but conversely this very first bird species is not more closely related to ostrich or cassowary than to the previously mentionned sister species. Otherwise this would mean that Charlemagne is more closely related to one of its extant descendants than to its extinct brothers and sisters.

I think this type of reasoning leads nowhere. As with paraspecies, the population may become theoretically holophyletic even if the phenotype does not change, but considering the species before or after the budding as different species is quite artificial. Let's imagine, if I take a time machine to the past and bring back a specimen with me, then it would still interbreed with its descendants... Species is not just a population in space, but also a population in time, and for me reasoning about evolution without the time is quite paradoxical.

There are any number of ways to classify things, and of course if someone prefers to lump all purple things together, or to classify Ichthyosaurs with dolphins on the basis of superficial similarity, it's impossible to argue that that's entirely irrational. And of course you are right that the nature of "relationship" needs to be specified. Phylogeneticists define relationship in terms of shared ancestry, as evidenced by synapomorphy. If you use that criterion, you can't exclude birds from dinosaurs just because they have a number of unique shared features (including losses of plesiomorphies) that the rest of dinosaurs don't have (any more than you would exclude Charlemagne's grandson from the Carolingians if he happened to be a dwarf), although of course you can recognise that uniqueness by naming them. It seems obvious to me that the most information is expressed by naming birds as birds, and also as dinosaurs, and that to name non-avian dinosaurs as a group can at best only be a shorthand recognition tool, not reflective of historical evolutionary processes. Would you want to name non-passerine birds as a taxon?

Yes, if you had the whole of life unpruned by extinction available to examine you would more or less have a continuum, and under any naming system *choosing which groups to name* is always going to be arbitrary to some extent. But that's a different question from whether to name monophyletic or paraphyletic groups. Your Charlemagne example is a little misleading because human ancestry is a reticulate process, but of course if that wasn't the case and the Carolingians were a clade, you would place Charlemagne, his brothers and sisters, and his extant descendants in one named group and perhaps some of the latter in a named sub-group, but no-one would suggest *naming* a group containing Charlemagne and his extant descendants to the exclusion of his brothers and sisters. In practice, we name clades that are distinct. But yes, in terms of *shared* ancestry the first bird species is closer to cassowaries (or even sparrows) than it is to the sister species of birds - BUT (and this is important) only infinitesimally so. If the latter existed, or we knew about it, it would doubtless be classified as a bird. We name groups that are *distinct and recognisable* on the basis of shared ancestry.

I do find it a little ironic that you "prefer carefully delimited taxa without any a priori about topology (holophyly vs paraphyly)" as contrasted with "the use of blind principles". So you would prefer not to have information about historical branching events available to inform your classification? Wouldn't that be a little blind? The principle of monophyletic classification simply says that in situations where bifurcating (not reticulate) evolutionary relationships dominate, we should name groups that reflect those relationships (and the processes underlying them), without artificial exclusion.

Anyway, this debate has been run at length on multiple occasions elsewhere, so I doubt we are going to convince each other.. ;-)

Yes you are certainly right, I may not convince you. As I said, I don't dispute the usefullness of clades, I just think that other types of groupings are usefull as well :-)

However, to come back to my initial question, you might have a personal opinion as a bryologist. Was the shared common ancestor of tracheophytes and bryophytes more moss-like in morphology ? more hornwort-like ? more liverwort-like ? or neither ? (maybe even more charophyte-like ?)

A good question, and I don't have a definitive answer! Most of those scenarios are possible. It is also even possible (especially if bryophytes turn out to be monophyletic after all) that the common ancestor of all *extant* land plants could have been polysporangiophyte-like (if not quite tracheophyte-like), i.e. something more like Horneophyton or Aglaophyton. This would seem unlikely, however, if polysporangiophytes are sister to hornworts (or to mosses) and derived from within a plesiomophically bryophyte-like clade. There are conflicting morphological characters that could support many different scenarios, and in the lack of strong fossil evidence and no consensus from the recent molecular data on branching order of the four embryophyte lineages (not to mention ongoing ambiguity about which streptophyte group is sister to embryophytes) it's difficult to do more than speculate.

The earliest land plant fossils (Ordovician) are liverwort-like spores (often termed cryptospores).The problem is that we can't say definitively that these spores were produced by liverworts just because no other *extant* plants produce spores like that.

The oldest definitive liverwort fossils are older than the earliest moss ones and much older than the earliest hornwort ones. However the oldest polysporangiophyte fossils are a little older than the earliest *definitive* liverwort fossils...

But if you wanted me to guess, I would probably put my money on vaguely liverwort-like. There are Devonian fossils of simple thalloid liverworts that are remarkably close to extant genera, and the likely sister groups to embryophytes would suggest the evolution of the sporophyte generation by interpolation (the antithetic theory), implying that the earliest land plants would have been gametophyte dominant.

These are indeed interesting scenarios. Thank you for this stimulating discussion.